Draft Regulatory Impact Analysis
and
Oxides of Nitrogen Pollutant Specific Study
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, Light-Duty Trucks, and
Heavy-Duty Engines; Particulate Emission Regulations
for 1987 and Later Model Year Heavy-Duty Diesel Engines
Environmental Protection Agency
Office of Air and Radiation
Office of Mobile Sources

-------
Draft Regulatory Impact Analysis
and
Oxides of Nitrogen Pollutant Specific Study
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, Light-Duty Trucks, and
Heavy-Duty Engines; Particulate Emission Regulations
for 1987 and Later Model Year Heavy-Duty Diesel Engines
Environmental Protection Agency
Office of Air and Radiation
Office of Mobile Sources

-------
Draft Regulatory Impact Analysis
and
Oxides of Nitrogen Pollutant Specific Study
Control of Air Pollution From New Motor
Vehicles arid New Motor Vehicle Engines:
Gaseous Emission Reaulations for 1987 and Later
Model Year Light-Duty Vehicles, Light-Duty Trucks, and
Heavy-Duty Engines; Particulate Emission Regulations
for 1987 and Later Model Year Heavy-Duty Diesel Engines
Environmental Protection Agency
Office of Air and Radiation
Office of Mobile Sources
Approved by:
Richard D. Wilson, Director
Office of Mobile Sources
Date:

-------
Table of Contents
Chapter	Page
1.	Introduction	1-1
I.	Background of the Regulations	1-1
A. Clean Air Act Requirements	1-1
3. Regulatory History 		1-2
II.	Description of the Proposed Action	1-4
A.	New Emission Standards	1-4
B.	Extension of Particulate Averaging to
1990 and Later Model Year HDDEs	1-5
C.	New Allowable Maintenance Regulations . .	1-5
D.	Miscellaneous Standards and Test
Procedure Revisions 		1-7
III.	Organization of the Regulatory Impact Analysis .	1-7
2.	Technological Feasibility 		2-1
I.	Introduction	2-1
II.	Light-Duty Trucks (LDTs) 		2-1
A.	Introduction	2-1
B.	Liqht-Duty Gasoline-Fueled Trucks (LDGTs).	2-2
C.	Light-Duty Diesel Trucks (LDDTs) 		2-19
D.	1.7 g/mi NOx Standard for Heavier LDTs . .	2-32
E.	Summary: 1987 NOx Standards for LDTs . .	2-37
III.	Heavy-Duty Gasoline Engines (HDGEs) 		2-37
A.	Introduction	2-37
B.	feasibility of a 1987 NOx Standard
of 6.0 g/BHP-hr	2-38
C.	Feasibility of a 1990 NOx Standard
of 4.0 g/BHP-hr	2-50
D.	Summary	2-57

-------
IV. Heavy-Duty Diesel Engines (HDDEs) 	 2-58
A.	Introduction	2-58
B.	Status of Current Technology 	 2-59
C.	Technoloa ica 1 Feasibility of the
Proposed 1987 NOx and PM Standards .... 2-62
D.	Feasibility of the Proposed 1990
NOx Standards	2-77
S.	Technoloaical Feasibilitv of the
Proposed 1990 Particulate Standard .... 2-Rd
F. Achievable Non-Trap Particulate
Levels in 1990 	 2-97
3.	Economic Impact	 3-1
I.	Costs	3-1
A.	Liaht-Duty Trucks (LDTs) - NOx Standards . 3-1
B.	Heavy-Duty Gasoline Engines (HDGEs) -
NOx Standards	3-28
C.	Heavy-Duty Diesel Engines (HDDEs) -
NOx and Particulate Standards	3-37
II.	Socioeconomic Impact 	 3-66
A.	Effects on Manufacturers 	3-66
B.	Regional Effects 		3-73
C.	National Effects 		3-73
4.	NOx Environmental Impact	4-1
I.	Introduction ....... 		4-1
II.	Health Effects	4-1
III.	Air Duality Analysis	&-2
h. Emission Factors 		4-3
B.	Der-Vehicle NOx Reductions		 .	4-4
-li-

-------
C.	NOx Emission Inventories	4-10
D.	Air Quality Effects	4-28
5.	Particulate Environmental Impact 		5-1
I.	Introduction	5-1
II.	Particulate Control Scenarios 		5-1
f III.y'Air Oualitv impact of Diesel
Particulate Control	5-1
A.	Relationship of Diesel Particulate
to Total Suspended Particulate and the
National Ambient Air Quality Standards . .	5-2
B.	HDDV Emission Factors and Resultant
Lifetime Emissions Per Vehicle 		5-3
C.	HDDV Particulate Emissions Inventory
and Comparison to Other Sources 		5-5
D.	Air Quality Projections for Urban Areas .	5-7
E.	High-Altitude Particulate Emissions . . .	5-9
IV. Health and Welfare Benefits of
Controlling Diesel Particulate 		5-10
A.	Non-Cancer Health Effects 		5-10
B.	rarcinoqen ic Health Effects	5-12
C.	Visibility Effects 		5-14
D.	Soiling Impacts	5-14
E.	Conclusions	5-16
6.	Cost Effectiveness	6-1
I.	Introduction	6-1
II.	Methodoloay	6-2
A.	Annual Approach	6-3
B.	Lifetime Anoroach 		6-11
-hi-

-------
III. HDDV Particulate Cost Effectiveness
and Comparisons to Other Sources 	 6-12
A.	HDDV Particulate Control	6-12
B.	Comoarison to Other Particulate
Control Strategies 	 6-15
IV. LDT and HDV NOx Cost Effectiveness and
Comparisons to Other Control Strategies .... 6-19
A.	LDT and HDV NOx Cost Effectiveness .... 6-19
B.	Comparison to NOx Control Strategies for
Other Mobile and Stationary Sources . . . 6-21
V. Conclusion	6-26
Appendix A to ChaDter 6	6A-1
I.	Introduction	6A-1
II.	Overall Methodoloay 		6A-1
III.	Annual Cost	6A-1
IV.	Annual Emission Reduction 		6A-2
V.	Cost Effectiveness	6 A-2
7. Alternative Actions 		7-1
I.	Introduction	7-1
II.	Statutory Requirements 		7-1
III.	Environmental Need	7-2
IV.	Prior Related Actions 		7-4
V.	Alternatives	7-4
A.	Heavy-Duty Diesel Particulate 		7-6
B.	Heavy-Duty Engine NOx	7-9
C.	Light-Dutv Truck NOx	7-11
Appendix A to Chapter 7	7A-1
I.	Derivation of 199S Emission Inventories . 7A-1
- iv-

-------
II.	Costs of Control Options	7A-2
III.	Derivation of Cost Effectiveness	7A-4
Appendix B to Chapter 7 - Discussion of
Alternatives for Heavy-Duty Enqines, from
the Preamble to the Proposed Rule	7B-1
8. Benefit-Cost Analysis 		8-1
I.	Introduction	8-1
II.	Benefits and Costs of HDV Particulate Controls .	8-2
A.	Particulate Benefit Analysis for HDDVs . .	8-2
B.	Particulate Cost Analysis for HDDVs . . .	8-21
C.	Benefit-Cost Comparisons for
Particulate HDDV Options 		8-29
-v-

-------
CHAPTER 1
INTRODUCTION
As required by Executive Order 12291 this document has
been orepared to summarize the results of all analyses
conducted in support of the proposed regulations. These
analyses include a consideration of the costs and the benefits
of the proposed action as well as a comparison of the proposed
action with alternative regulatory approaches. The primary
supporting analysis summarized in this document is the EPA
Diesel Particulate Study, which examined the costs and benefits
of a number of alternative diesel particulate regulatory
approaches. Supoorting analyses for the proposed NOx
regulations are contained in this document. The background air
quality data supporting the NOx environmental impact analysis
may be found in the public docket. The NOx Environmental
Impact Analysis contained in this document also serves as the
NOx oollutant-specific study reauired by Section 202 (a) (3) (E)
of the Clean Air Act.
I. Background of the Regulations
A. Clean Air Act Requirements
The Clean Air Act Amendments of 1977 created a statutory
heavy-duty vehicle (HDV) class and established mandatory
emissions reductions for that class. Under the language of the
amendments, all vehicles over 6,000 lbs oross vehicle weiqht
(GVW) were defined as "heavy duty" and were required to achieve
a 75 percent reduction in NOx emissions from uncontrolled
levels, effective with the 1985 model year.
The Act made no specific provisions for liqht-dutv trucks
(LDTs) , which at that time only encompassed LDTs between 0 and
6,000 lbs GVW (light LDTs). These LDTs were regulated by EPA
as a separate class under the general authority of the Clean
Air Act. Beainning with the 1979 model year, EPA expanded its
standards for the LDT class to 8,500 lbs GVW, thus encompassing
those heavy LDTs (6,001 to 8,500 lbs GVW) which are subiect to
the heavy-duty vehicle provisions mentioned above.
The Act also authorizes the Administrator to temporarily
establish revised standards for heavv-dutv enaines if the
statutory standards cannot be achieved without increasing cost
or decreasinq fuel economy to an "excessive and unreasonable
deqree."[l] The heavy-duty enqine NOx standards evaluated in
this document are being proposed under these Provisions of the
Act.

-------
1-2
The amendments of 1977 also require the "areatest deqree
of [particulate] emissions reduction achievable", given the
availability of control technology and considering cost,
leadtime and energy impacts.[21 These reductions were to beqin
during the 1981 model year. Although not specifically limited
as to applicability in the language of the Amendments of 1977,
it was recognized that the requirement was aimed at diesel
vehicles/engines. The heavy-duty diesel engine (HDDE)
particulate standards proposed in this NPRM are based on this
author ity.
B. Regulatory History
The first NOx standards antedated the Amendments of 1977.
Prior to the 1975 model vear, LDTs complied with the 3.0 a/mi
NOx standard that had been established two years earlier for
LDVs. With the splitting off of the LDT class for the 1975
model year, LDTs were required to meet a NOx standard of 3.1
g/mi, comparable in stringency to the LDV standard. Heavy-duty
engines (HDEs) will have no separate NOx standard until the
1985 model vear, however, there have been combined HC + NOx
standards in place for HDEs since the 1974 model year. Table
1-1 shows Federal NOx and particulate standards for LDTs and
HDEs from 1972 to 1986.
The current NOx standard for 1979 and later model year
LDTs is 2.3 q/mi, comparable in stringency to the 2.0 g/mi
standard established for LDVs of that year. Beqinning in 1979,
the LDT class was expanded to include vehicles between 6,001
and 8,500 lbs GVW. The current NOx standard for HDEs is 10.7
a/BHP-hr, established originally for the 1984 model year, but
later made optional until the 1985 model vear.
Turning now to more recent actions, an Advanced Notice of
Prooosed Rulemaking was promulgated for LDT and HDE NOx
emissions in January of 1981 (46 FR 5838) . Standards of 1.2
grams per mile for LDTs and 4.0 g/BHP-hr for HDEs were orooosed
effective for the 1985 and 1986 model years, respectively.
These standards did not correspond to the statutory 75 percent
reduction as noted above, but were proposed because they were
comparable in stringency to the existina 1.0 q/mi LDV NOx
standard in the case of LDTs and because they represented the
lowest practicable standard given the available technoloqv in
the case of HDEs.
The first diesel ©articulate standards were established
for LDVs and LDTs, effective beqinninq with the 1981 model
year. A standard of 0.60 q/mi was established for both LDVs
and LDTs, representing an achievable level for the (then)
available technology. More stringent standards (at 0.26 for

-------
Model
Year
1972
1973
1974
1975-78
1979-81
1982
1983-84
1985-86
Table 1-1
Exhaust Emissions for LDTs and HDEs
(Federal Standards)
LPT a/mi			
NOx
3.0
3.0
3 .1
2.3
2.3
2.3
2.3
Part
0.6
0.6
0 . 6
NOx
A*
B*
A*
B*
A*
B*
C* 10.7
10.7
HPE q/BHP-hr
HC+NOx
16
16
10**
5
10**
5
10**
5
Part
"A~ B~, and C designate requlatorv ODtions available for
HDEs .
If HC emissions do not exceed 1.5 g/BHP-hr.

-------
1-4
LDTs, and 0.20 for LDVs) were also promulgated effective
beqinning with the 1985 model year, but these have been delayed
and will now be effective for the 1987 model year (49 FR 3010,
January 24, 1984).
For HDDSs, a Notice of Proposed Rulemakina was published
in January, 1981 (46 FR 1910) which prooosed a standard of 0.25
q/BHP-hr for 1986 and later model years. Because of the
related technical issues that were raised during the comment
oeriods for both the MOx ANPRM and the oarticulate nprm and the
interrelationship between NOx and particulate emissions, EPA
has decided to issue a combined NPRM to address these issues
and insure that manufacturers could direct their efforts at
meeting a unified set of emission standards.
Also, this action represents the initial proposal of
high-altitude particulate standards for diesel-powered LDTs and
HDDEs.
II. Description of the Prooosed Action
A. New Emissions Standards
This proposal contains new NOx standards for LDTs and HDEs
and new particulate standards for light-duty diesel trucks
(LDDTs) operated under high altitude conditions and for HDDFS
operated under both high and low altitude conditions. The NOx
standard for LDTs is proposed at 1.2 g/mi for LDTs up to 6,000
lbs GVW or 3,999 lbs eauivalent test weiqht (ETW) to maintain
comparability with -the LDV standard of 1.0 g/mi that was
established for the 1981 model year. The standard for LDTs
over the above weiqht limits is 1.7 g/mi, primarily to
facilitate compliance with the particulate standard for the
heavier LDDTs. A staged NOx standard is prooosed for HDEs to
allow leadtime for further development of control technoloav.
The NOx standard for 1987-89 model vear HDEs is proposed at 6.0
q/BHP-hr, representing a level that is achievable given minimum
leadtime for engines currently in production, with a more
stringent standard of 4.0 a/BHP-hr to be effective for 1990 and
later model year engines.
A phased particulate standard is also prooosed for HDDEs.
Model year 1987-89 HDDEs operated under low-altitude conditions
would meet a standard of 0.60 g/BHP-hr. Por 1990 and later
model years, low-altitude urban bus engines will comply with a
standard of 0.10 g/BHP-hr, while the remainder of the
low-altitude HDDEs would meet a standard of 0.25 g/BHP-hr.
Both of these proposed 1990 standards will likely require the
use of trap oxidizers on a majority of applications. These
standards represent the approximate lower limit of feasibility

-------
1-5
given the above-mentioned NOx standard. HDDEs operated under
high-altitude conditions, including urban bus engines, would
comply with standards of 0.72 g/BHP-hr beginning in the 1987
model year and 0.30 g/BHP-hr (0.12 g/BHP-hr for urban buses)
effective for 1990 and later model years. Finally, a
particulate standard of 0.26 g/mi is proposed for LDTs ooerated
under hiqh-altitude conditions. These proposed standards are
summarized in Table 1-2.
B.	Extension of Particulate Averaging to 1990 and Later
Model Year HDDES
Diesel particulate averaging has been included for 1990
and later HDDEs as a means of reducing the cost of compliance
with the 0.25 g/BHP-hr standard. Under existing regulations,
manufacturers of 1987 and later model year light-duty diesel
vehicles (LDDVs) and light-duty diesel trucks (LDDTs) will have
the option of averaqing particulate emissions from their
products. This provision is designed to allow some vehicles to
exceed the standard as long as the manufacturer's corporate
average light-duty diesel particulate emission level is below
the applicable standard, or below a production-welqhted average
standard if the manufacturer's product line includes both LDDVs
and LDDTs. With this rulemaking, particulate averaging will
also be afforded to manufacturers of 1990 and later model vear
HDDEs. However, they would not be allowed to average HDDEs
with LDDTs or LDD^s if the manufacturer's product line also
includes these vehicle types. Similarly, averaging hiq'n
altitude or California engines with engines intended for sale
in low altitude non-California areas will not be permitted,
althouqh averaging within each of these areas is allowed.
Urban buses will be excluded from the averaging oroqram to
insure the maximum reduction in urban particulate emissions.
Because HDDE standards are expressed in mass oer unit of work
(q/BHP-hr) rather than mass per unit of distance travelled
(g/mi) and because HDDEs are divided into subclasses with
widely varying useful life periods, averaging will be limited
to within each of the existing subclasses (light-, medium-, and
heavy-heavy duty) and the calculation of average particulate
emissions must include weighting factors for brake horsepower
as well as for production volume.
C.	New Allowable Maintenance Regulations
The allowable maintenance provisions (§86.087-25) have
been simplified and streamlined to a large extent. The concept
of emission- and non-emission-related maintenance has been
extended from LDTs and HDEs to encompass LDVs as well.
Maintenance intervals have been changed, as outlined in the
preamble, and manufacturers will be required to demonstrate the
likelihood of in-use performance for certain critical emission-
related maintenance.

-------
1-6
Table 1-2
E f fect ive
Model Year
Low Altitude:
1987
1990
Proposed NOx and Particulate
Standards for LDTs and HDES
Vehicle Class
Aoplicable Standards
MOx	Particulate*
LDTl**
1.2
g/mi
LDT2**
1.7
g/mi
HDE
6.0
g/BHP-hr
LDTl**
1.2
g/mi
LDT2 * *
1.7
g/mi
HDE Urban Bus
4 . 0
g/BHP-hr
All Other
4.0
q/BH?-hr
* ~ ~
+ * *
0.6 a/BHP-hr
ir k k
kkk
0.1 g/BHP-hr
0.25 q/BHP-hr
High Altitude:
1997
LDTl**
1.2
g/mi
0.26
q/iTli

LDT2* *
1.7
g/mi
0.26
g/mi

HDE
5.0
g/BHP-h r
0.72
q/BHP-hr
1990
LDTl**
1.2
g/mi
0.26
g/mi

LDT2**
1.7
g/mi
0.26
g/mi

HDE
4.0
g/BKP-hr
0.30
g/BHP-h r
* Diesel-powered vehicles/enqines only
** LDTl - vehicles up to 6,000 lbs GVW or 3,999 lbs ETW.
LDT2 - vehicles 6,001-8,500 lbs GVW or 4,000 lbs or
greater ETW.
*** 1987 and later model year LDTs are currently required to
complv with a standard of 0.26 g/mi. No new standards are
beinq oroposed.

-------
1-7
D. Miscellaneous Standards and Test Procedure Rev is ions
The current idle CO standard for qasoline-powered LDTs has
been extended to cover vehicles intended for sale in designated
high altitude areas. This extension will allow the development
of more representative cutpoints for inspection/maintenance
proarams in high altitude areas and thus facilitate the
detection of failed catalytic converters.
Since there is currently no particulate standard (and
consequently no particulate test procedures) for HDDEs, it was
necessary to incorporate suitable test Procedures into Subpart
N. EPA is also proposing a humidity correction factor to be
applied in the calculation of NOx emission test results for
HDEs, and a correction factor to be used when HDGE NOx
emissions data are obtained through continuous sampling, as
opposed to the more common bag sampling, techniques.
III. Organization qP the Regulatory Imoact Analysis
The remaininq six chapters of this Regulatory Impact
Analysis will be divided as follows: Chapter Two will deal
with the technological feasibility of compliance with the
proposed standards. Chapter Three presents an analysis of the
economic impact of the proposed regulations. Chapter Four
analyzes the environmental effects of the proposed NOx
standards while Chapter Five performs the same function for the
proposed particulate standards. An analysis of the cost
effectiveness of the proposed regulations is presented in
Chapter Six, and the requlatorv alternatives considered are
discussed in Chapter Seven.

-------
1-3
References
1.	The Clean Air Act As Amended August 1977, Serial No.
95-11, Section 202 (a) (3) (B)&(C) .
2.	IBID; Section 202 (a)(3)(iii); p. 102.

-------
Chapter 2
Technological Feasibility
I.	Introduct ion
This chapter analyzes the technical feasibility of
nitrogen oxide (NOx) emission standards for 1987 and later
model year light-duty trucks (LDTs) , and the feasibility of NOx
and particulate emission standards for 1987 and later model
year heavy-duty engines (HDEs). Current emission levels are
discussed, as are the emission reductions which are judged to
be attainable by 1987 and by 1990. Specific technologies
required to achieve these emission reductions are identified.
Other factors considered in this analysis include the fuel
economy and performance implications of increased emission
control. After due consideration is given to other regulatory
requirements (i.e., compliance with standards over the full
useful life and assembly-1 ine testing), 1987 LDT, and 1987 and
1990 model year HDE emission standards are recommended.
The analytical methodologies used in this analysis will be
explained as presented.
II.	Liaht-Duty Trucks (LDTs)
A. Introduction
This analysis will address the feasibility of two LDT NOx
standards in 1987 : 1) 1 . 2 g/mi for all LDTs, 2) 1.2 a/mi for
lighter LDTs and 1.7 q/mi for heavier LDTs. Note that 1987 is
also the model year in which more stringent particulate
standards (.26 g/mi) go into effect for light-duty diesel
trucks (LDDTs). The feasibility of this particulate standard,
however, has already been addressed in the rulemakmg. [ 1, 2]*
This analysis will consider the particulate standard as aiven,
and will only address the feasibility of additional NOx
control. However, the effect of NOx control on the type of
technology needed to meet the particulate standard will be
addressed.
First of all, the feasibility of a 1.2 g/mi NOx standard
for 1987 light-duty gasoline-fueled trucks (LDGTs) will be
addressed. The analysis will consider the inherent similarity
between LDGTs and light-duty gasoline-fueled vehicles (LDGVs)
in assessing the feasibility of the 1.2 g/mi standard.
Available emission certification data will be extensively used,
and fuel economy effects of the standard will be estimated.
In determining the 1987 LDDT particulate standard of 0.26
g/mi to be feasible, EPA's analysis assumed a simultaneous
NOx standard of 1.2 g/mi.[2]

-------
2-2
Secondly, this analysis will examine the feasibility of
the 1.2 g/mi NOx standard for 1987 LDDTs. Again, a review will
be made of available emission certification data. A larqe part
of the discussion will address the specific compliance
capabilities of the heaviest trucks.
Thirdly, the feasibility of a 1.7 g/mi NOx standard for
the heavier LDTs {i.e., those LDTs with an equivalent test
weight (ETW) above 4,000 lbs or a qross vehicle weight above
6,000 lbs) will be discussed, along with a detailed
presentation of its fuel economy implications, as well as its
effect on diesel particulate control technology.
B. Light-Duty Gasoline-Fueled Trucks (LDGTs)
1. Low Mileage Target Derivation
Typically, the first step in determining the feasibility
of a given emission standard is to determine the low mileage
emission target levels. (Manufacturers must calibrate their
emission control systems on certification and production
vehicles somewhat below the actual emission standard to account
for in-use emissions deterioration and production line
emissions variability. This is done to assure that the averaqe
production vehicle will comply with the applicable emission
standard over its entire useful life.) There are three factors
which must be considered in develooinq the target emission
level: the emission standard itself expressed as the maximum
pass level (MPL), the Selective Enforcement Audit (SEA)
adjustment factor (AOL) necessary to account for
oroduction-line variability and to provide the manufacturer
confidence that no more than 40 percent of its vehicles will
fail a formal SEA, and the full-life multiplicative
deterioration factor (DF), The design target emission level
(TL) is, therefore, calculated as follows for LDTs:
TL ¦= mz x w 
Because, however, different emission-control technologies can
exhibit different rates of deterioration, this LDGT discussion
beqins with a brief description of likely 1987 LDGT
emission-control technologies.
LDGT emission-control technoloqies are identical in
concept and design to the systems used on LDGVs. Indeed, the
best predictor of likely 1987 LDGT emission-control strategies
are current LDGVs. Both vehicle classes are certified on the
same chassis dynamometer test procedure. Both vehicle classes
share the same technologies, and in many cases share the same
engines and/or components. Both vehicle classes are subject to
Corporate Average T^uel Economy (CAFE) standards. Current LDGVs
are required to meet a NOx standard of 1.0 q/mi; the 1987 LDT
1.2 g/mi standard was specifically derived to be comparable in

-------
2-3
stringency* to the 1.0 g/mi LDV standard. [41 In summary, the
emission control technologies present in current LDGVs will
most likely be adopted for 1987 LDGTs. As will be noted below,
42 percent of current LDGTs are already equipped with advanced
NOx control technology found on LDGVs.
The types of catalyst emission control systems present in
current LDVs are shown in Table 2-1. With respect to NOx
control, and especially with respect to NOx deterioration,
there are only two relevant systems: those systems which use
three-way reduction catalyst technology and those which do
not. (As shown in Tables 2-2 and 2-3, 38 of 40 1984 LDGT
engine families are equipped with EGR in both three-way and
non-three-way catalyst systems; as EGR is used to the same
extent in both, it need not be considered when discussing
differential deterioration rates between the catalyst
systems.) Deterioration for both types of catalyst systems
need to be characterized since both systems are applicable to
LDGTs, although three-way systems appear to be overwhelmingly
preferred in LDV applications (see Table 2-1).
Average NOx deterioration factors (DFs) for each catalyst
system can now be derived from existing emission certification
data. Tables 2-2 and 2-3 present 1984 certification data for
LDGT engine families equipped with reduction (three-way)
catalysts and non-reduction catalyst technology, respectively.
Note that the multiplicative deterioration factors presented in
both tables only address deterioration over 50,000 miles (half
life). To derive a "typical" full-life NOx DF for both
reduction catalyst and non-reduction catalyst technology, the
average half-life DF from Tables 2-2 and 2-3 must be adjusted.
Using the appropriate methodology,** average multiplicative
full-life NOx DFs of 1.286 for three-way and 1.070 for
non-threeway systems are derived. A larger DF for
catalvst-based NOx control (three-wav technology) is exoected
because of the inherently greater deterioration typically
observed in catalyst-based emission control systems.
A significant difference between the classes with respect
to emissions is that LDTs, on average, are heavier than
LDVs. Since more eneray (i.e., more fuel) is required to
push a heavier vehicle over the road, more exhaust
products are formed per mile of travel. Therefore, for a
given LDV standard, a slightly higher numerical standard
for LDTs provides equivalently stringent control.
Assumina linear deterioration over the full useful life,
the full-life multiplicative DF = 1 + 2.4 (half-life DF -
1), where 2.4 = 120,000 miles (full-life distance)/50,000
miles (half-life distance).

-------
2-4
Table 2-1
Market Share* by ECS for Liaht-Duty Vehicles
Emission Control
Svstem (ECS)
OLC + 0C
OLC + 3W
OLC + 3W + OC
FBC + 3W
FBC + 3W + OC
OLFT + OC
OLFT + 3W + OC
F^FI + 3W
FBFI + 3V + OC
Diesel
Total 3W;
Total Non-3w;
^otal Open Loop:
Total Closed Loop:
Total Non-Diesel:
1981
13.80%
2.39%
11.84%
21.33%
35 . Rl%
0.02%
0 . 01%
P. L 2 %
2. 68%
6.02%
30 .18%
13 . P2%
28.06%
65.94%
94.00%
1Q 82
11.55%
2.11%
IP. 63%
1?.82%
30.29%
0 .00%
0.02%
16.46%
2 .53%
7. 58%
80.86%
11.55*
30.31%
6? . 10%
92.41%
1983
8 .67%
0.32%
10.73%
3 . 50%
36. 8=;%
0.00%
O . 02%
22.10%
9.36%
8.45%
8?.88%
0.67%
19.74%
71.81%
91.55%
* The market share was calculated usinq *-he data hase from
Reference 3.
Abhr eviations
OLC	= Ooen-loop carburetor.
OLFI = Open-looo fuel injection (FT).
Ftb	= Feedback throttle bodv.
OC	= Oxidation catalvst.
3W + QC = Three-way-olus-oxidation catalvst.
F^C	= Feedback carburetor.
3W	= Three-wav catalyst.
FBFI = Feedback FI.

-------
2-5
Table 2-2
1984 Certification Data for LDG^s
Equipped With Three-way-Catalyst Technology
Emission Control '"ert. Low-wileaae
Manufacturer Fnoine Pamilv	Technoloav*	MOx DF NOx Level
American
Motors
FAMl^OTlHEA^
EAM258T,2"EAX
EAM360T2HLEX
FGR/PLS/3CL
EGR/PLS/3CL
EGR/PMP/3WY
1 .375
1.04^
1 . 100
0.35-1.1
0 . 7S-1.R
1.1-1.4
Du t ton
FDNfi.0T4HRAX
EGR/PMP/OXD/3CL
1.100
1 . 4
Ford Motor
ConDany
EFM2.3T1"KG°
EFM2.8T2HCG1
EFM4 . ^lHGGS
EFM5.0T2GAF2
EFM5.8T2HGG0
FFM5.8T24GAF3
FGR/PMP/OXD/3CL
EGR/PMP/OXD/3CL
EGR/PMP/0XD/3CL
EGR/PMP/0XD/3WY
EC-R/PMP/OXD/3CL
FGP/PMP/OXD/3WY
1.000
1. 4fi4
1. 000
1. 060
1.000
1.060
0.77-2.0
0	. 50
0.45-0.98
1	. 3-1 . S
0 . 57-1.9
1.7-1.8
Fu i i
PFJl.8T5FFH1
EGR/3CL
1.100
0 . 6Q
General
Motors
F1G1. 9T2'T,BA8
E1G2.8T2TRA2
FIGS.7T4TYA4
FGR/PMP/3CL
EGR/PMP/3CL
FGR/DMP/3CL
1.316
0 . 892
0 . Q 9 6
n .35
0.71
0.48
Mitsubishi
EMT?.6T2FCJ3
FGR/DLS/3CL/OTR
1.039
0.73-0 . 78
Suzuki
ESK1.0T1FSF6
3CL
1.085
1 . 4
Tovota
FTY2.0T5FER2
ETY2.4T2FBBO
ETY2.4T5FBR4
EGP/3CL/OTR
EGR/pwp/ 3WY/OrpP
EGP/3CL/OTP
1.308
1 .000
1 . 2Q3
0.18-0.20
0.30-0.56
n.]2-0.76
Vo 1 !< swaaon
FWl. 9m5CVA 2
TCL
1. oon
0.3"*


Averaae:
1 . 110

EGR = Exhaust gas recirculation
= 3-wav catalvst, closed looo fuel control
dmp = Air pump
O^R = Other
PLS = Pulse air injection
3wy = 3-way catalvst, ooen-loop fuel control
OXD = Oxidation catalyst

-------
2-6
Table 2-3
1984 Certification Data for LDGTs
Fquipped Without Three-Wav Catalvst Technology
wanufacturer
Enaine Familv
Emission Control
Technoloav
Cert.
VOX DF
Low-Mi leaai
MOx Level
Amer ican
Motor s
EAM2.8^2AXE 2
EGR/PMP/OXD
1 .079
1.5-1.7
Chcvsler
ECR2.2T2AAB5
ECP2 . 5T2AAC8
ECR3.7T1BBB0
ECR5.2T2BBF5
ECR5 . 2rr14RAFO
ECR5-9T4BAF9
FGR/PMP/OXD
EGR/PLS/OXD
EGR/PMP/OXD
FGP/PMP/OXD
FGR/P^P/OXD
FGF/PMP/OXD
1 .000
1. 006
1.000
1 . 000
1 .000
1.000
1.7-2.0
2.0-2.3
1.^-1.8
1 . 7-1.P
2.1-2.2
1 .3-1.5
Fu} 1
EFJl-8T2AFD3
EGR/PLS/OXD
1 . 000
1.2-1.3
Gene ra1
Motors
F 1G1. 9T2HPC9
E1G2.0T2HJCX
^102. 8rT,2»Tir7
E1G4 . 1T-2HHD3
FIGS.7TdHHCX
FGR/p^P/OXD
FGR/PMP/OXH
PGR/pmd/OYD
ECR/PMP/Ovd
FGR/PM^/Ovn
1 .onn
1.000
1 .079
1 .OQn
1.000
1.7-2.0
1 . o-2 . 1
1.7-1 . 9
1.6-1.7
1.6-2 . 2
TSUZU
ESZ119T2AAG1
PGR/PMP/OXD
1.000
1.6-1.8
Mi tsubish i
EMrr,2 . OT2AFOS
FMT2.0T2BCAV
FMT2 . 6rT,2AFD8
FMT2.6T2PCA?
FMT?.6T2BFJ8
EGR/PLS/OXD
FGP/PLS/OXD
FGR/PLD/OXD/O^P
PGR/pLS/OXD/Ottr
EGR/PLS/OXD/OTR
1 .000
1.102
1.158
1.046
1.046
1.4-1.7
0.67-0.73
1.2-1.9
0.46-0 . SR
1 . 6-1 . 7
Nissan
EMS2.4T2AAF0
EGR/P LP/OXD
1.095
1.3-1.5
^ovo Kociyo
E^K 2.0T2AFL7
EGR/PLS/OXD
1.052
1.6-1. 8
Tovota
ETY2.4T2AFFO
FTY4 . 2'TT2AFF4
EGR/PLS/OXP/OTR
/dmd/OXD/O^P
1 .000
1 . 000
1.2-1.8
0.99
Averaoe:
1.029

-------
2-7
The next step in low-mileage target determination is the
derivation of the production variability (AQL) factors for NOx
emissions. An elaborate statistical analysis of LDT production
variability will not be performed here, primarily because such
precision is not important to this particular discussion,* and
also because sufficently accurate estimates have already been
derived.[5] For this analysis, a range of likely AOL factors
will be used for all LDGTs. A typical range of AQL factors for
a variety of engines and vehicles is 1.1 to 1.3. (Numbers in
this range have been used in previous analy ses [ 5,61 , and are
used elsewhere in this analysis.) These values represent
conservative maximum and minimum factors likely to be observed
in production.
Using the full-life DFs derived above, the full range of
likely AQL factors and a MPL of 1.2499 g/mi, LMTs of 0.8-0.9
g/mi for three-way systems and 0.9-1.1 g/mi for non-three-way
systems are calculated. The difference in target levels
entirely reflects the difference in NOx deterioration rates
between the three-way and non-three-way systems.
2. Current Technology LDGTs
Tables 2-2 and 2-3 also present a summary of available
1984 model year Federal certification test data for LDGTs.
Table 2-2 presents engine families equipped with three-way
technology which, according to manufacturers' confidential
sales projections, will account for about 42 percent of Federal
1984 LDGT sales. Two significant observations can be made from
this table. First of all, 15 of the 20 families have one or
all vehicle configurations below the 0.8-0.9 q/mi tarqet ranqe
(i.e., less than 0.8 g/mi). In short, these families have
already demonstrated the ability to comply with the 1987 1.2
q/mi standard. (These 15 families represent about 30 percent
of total projected 1984 LDGT sales as well as a wide range of
LDGT sizes (2875-4925 lbs ETW).) The second major observation
involves the five families above the target level. Three of
these families are equipped with open-loop fuel control, (i.e.,
there is no exhaust oxygen sensor to provide feedback to the
microprocessor-based feedback fuel control system. Since
precise fuel metering is essential to the efficient operation
of a three-way catalyst, the adoption of closed-looo fuel
control for these families would almost certainly allow some
It will be readily apparent from this analysis that the
NOx reducing ability of available LDGT technology is
sufficiently great that any uncertainty induced by an
imprecise AQL factor is immaterial to a conclusion of
feasibility.

-------
2-8
degree of reduction in NOx emissions. (These three families
represent about 11 percent of total projected 1984 LDGT
sales.) The emissions performance of the engine families
presented in Table 2-2 demonstrates the overall feasibility of
the 1.2 standard.
Table 2-3 presents the remaining 1984 Federal engine
families certified for sale without three-way technology.
These families represent almost 58 percent of projected 1984
LDGT sales. Three notable points can be made from this data.
First, the range of average NOx emissions for these families is
1.4 to 1.7 g/mi, well below the 2.3 standard of course, but
well in excess of either the .8 - .9 or .9 - 1.1 target ranoes
for the 1.2 standard. Obviously, recalibrat ions or additions
of hardware will need to be made in the majority of these
families to meet a 1987 1. 2 g/mi standard. Second, three of
the families (representing about .2 percent of 1984 sales)
exhibit NOx emissions below 1.0 g/mi, and can be judged to
already exhibit the ability to comply with a 1.2 standard with
very little change. Finally, an additional five families
(representing about 10 percent of 1984 sales) exhibit NOx
emissions at or below 1.3 g/rai in at least one configuration,
and at least hint at the possibility that some families through
minor changes (i.e., EGR and timing recalibrations) could
complv with a 1.2 g/mi NOx standard usina the existing
oxidation cataly-st system.
These observations from Tables 2-2 and 2-3, along with
other information, represent the	basis for the following
analysis of the feasibility of a 1.2	g/mi NOx standard for 1987
and later model year LDGTs.
3. Implications of a 1.2 a/mi 1987 LDGT NOx Standard
The data presented in Table 2-2 and discussed above, along
with the demonstrated ability of substantially identical LDG^s
to comply with a NOx standard of comparable stringency,
establish for all practical intents and purposes the inherent
feasibility of a 1.2 g/mi NOx standard for all 1987 LDGTS, if
closed-loop three-way technology is assumed. Indeed, the 1984
three-way fleet exhibits NOx emissions 40 - 50 percent lower,
on average, than the non-three-way fleet; many of these
vehicles are already complying with a 1.2 g/mi standard. This
conclusion of feasibility has not been disputed by any
manufacturer, and warrants no further discussion.
The remainder of the LDGT analysis will concern itself
with the possibility that some LDGT families may not need to
adopt complete closed-loop three-way systems. An attempt will

-------
2-9
be made to predict the likely 1987 fleetwide emission control
system product mix under a 1.2 NOx standard, taking into
account product performance and fuel economy tradeoffs.
Finally, a projection of the fleetwide fuel economy impact of
the 1.2 standard will be made.
Table 2-4 presents our predictions for market shares of
emission control technologies for 1987 LDGTs meeting a 1.2 NOx
standard, along with the assumptions made in generating these
predictions. As indicated by the results, it is reasonably
certain that 90 percent or more 1987 LDGTs will be equipped
with closed-loop three-way catalyst technology. The remaining
vehicles may or may not, primarily depending upon the
individual product choices of the individual manufacturers.
To evaluate the fuel economy implications of a 1.2
standard with the projected market mix of Table 2-4, a more
detailed description of current LDGTs is necessary than the one
presented above (which concerned itself only with emission
control systems).* Tables 2-5 and 2-6 present more detailed
vehicle data for the engine families presented earlier in
Tables 2-2 and 2-3; Table 2-7 presents average values for the
important parameters listed in Tables 2-5 and 2-6.
The current emission control product mix itself is an
important clue to the relative fuel economies of the two
emission control technologies. For any given gasoline engine,
there is a point beyond which increasing EGR or recalibrating
ignition timing for lower NOx will begin to impair fuel economy
and performance (see Figures 2-1 and 2-2). Since EGR and
timing recalibrat ions are less expensive than three-way
systems, and assuming the rational allocation of funds by the
manufacturers, the existence of three-way systems on current
vehicles strongly suggests that a point of unacceptable fuel
economy or performance tradeoffs has already been reached at a
2.3 g/mi NOx level for 42 percent of the vehicles.** Therefore,
for any level of standard below 2.3, one would expect three-way
The baseline for this comparison is the 2.3 g/mi standard.
There are several conceivable determinants of emission
control system configuration. Factors such as cost,
vehicle weight, fuel economy, vehicle application,
product-line standardization (including standardization
between LDGVs and LDGTs), the ratio of engine displacement
to vehicle weight, individual engine characteristics,
subjective driveability assessments, etc. are taken into
account. Manufacturers have some flexibility, even with
emission constraints, in configuring a total vehicle
package.

-------
2-10
Table 2-4
Technology
Non-^hree-way
(open-loop)
Th ree-Wav
(open-loop)
Three-Way
(closed-loop)
Present and Projected Market Shares
of Emission Control Technology
1984 LDGTs	19 8 3 LDGVs
(NOx= 2.3 q/mi) [11 (MQx = l.0 q/mi) [21
58%
11%
31%
10%
12%
78%
19 8 7 LDGTs
(NOx=1.2 q/mi)
0.2 - 5% f31
0.4 - 3 . 7% f 41
91 - 9 9 % f 5 ]
[1] Based upon Tables 2-2 and 2-3, and manufacturers' confidential
proiected 1984 ldgt sales estimates.
[21 From Table 2-1, these oercentaqes have been changed to exclude
diesels from the total.
f 31 Lower limit is based upon the three ooen-loop oxidation
catalvst systems already emittinq below 1.0 q/mi. MOx. (See
Table 2-3.) Upper limit is based upon the additional five
ooen-looo, oxidation catalyst svstems with current emission
levels below 1.3 q/mi NOx, and the assumption that 50 percent
of these vehicles can meet the 1.2 standard with increased fgp
and recalibrations instead of with three-wav technology (see
Table 2-3) .
[41 The minimum proiection assumes that all current three-wav,
ooen-loop Ford families adoot closed-loop technology for fuel
economy reasons, while the one three-way ooen-looo Tovota
family remains open loop. Upper limit is based upon the
assumption that onlv manufacturers which todav use ooen-looo
three-way svstems in LDGVs would consider their use for future
LDGTs. AMC, Tovo Koqvo, and Tovota are the onlv LDGT
manufacturers in this cateqory with current ooen-loop-oxidat ion
catalvst LDGTs. However, since Tovo Kovqo has chosen to qo
completely closed loop in a 1983 LDGV engine of identical
displacement as their currently open-loop oxidation catalvst
LDGT family, we project that it will also qo closed looo in the
1987 LDGT.
[51 Remainder.

-------
2-11
fable 2-5
Vehicle Data for 1984 Federal LDGTs Eauipped
with Three-Way Catalvst technology (see Table 2-2)





Mean





Mean
Veh icle



Ena.
Mean
Vehlcle
Equlvalent
Mean

F,na ine
Diso.
N/V
Road Load
Test Weiabt
ritv
Manufacturer
Familv
(1 iters)
ratio
(hp)
(lbs)
mDq
Atner ican
EAM150TlFEA 5
2 . 5
44 . 3
13 . 3
3 ,400
20 . 3
Motor s
EAM258T2HEAX
4 . 2
34.fi
14 . 1
4,000
18.0

EAM3 60T2HLEX
5.9
38 . 4
16.4
4 ,750
11.4
Dutton
EDN6.0T4HRAX
6.0
38.2
19.0
6 , 000
11.5
Ford Motor
FFM2 . 3Tl^KG9
2 . 3
47.8
10 . <3
3 ,200
21.7
Companv
FFM2.8T2HCG1
2.3
45.0
13.0
3,375
18.4

EFM4.9T1HGG5
4 . 9
29.8
1*. 6
4 , S75
15.8

EFMS.DT2GAF2
5.0
41.5
.0
5,125
12.8

fttM5 . 8T,2HGGn
5.8
38.9
17.0
4,925
12 . 5

EFM5.PT24GAF3 S.8
42.5
17.4
4,875
11.6
Fuii
EFJ1.8T5FFW1
1.8
47.0
9 . 6
2 ,875
24 . 1
General
Elnl.9T?TBA8
1.9
47.0
10.4
3,000
24 .2
Motors
E1G2 . 8T2TRA2
2 . 8
38 . 2
12.0
3 , 500
19 . 5

E1G5 . 7T4TYA4
5 . 7
25.4
16 . 6
4 ,750
14 .6
Mitsubishi
FMT2 . 6T,2FCJ3
2 . 6
51 . 6
15 . 8
3,550
10.n
Suzuki
FSKI.Of1FSF6
1.0
84 . 1
13.0
2 ,250
28.1
Toyota
Fty2 . 0'r5FBB2
2.0
41.7
14.9
3 ,375
24 . 6

ETY2.4T2EBB0
2 . 4
an. 7
14 .9
3 ,300
20.8

ETY2.4T5FBB4
2.4
39 . 0
12 . 7
3 ,000
24 . 5
Volkswagen
EVW1 . 9rT,5CVA2
1.9
5S.9
17.6
4, nno
16.8

-------
2-12
Table 2-6
Vehicle Data for 1984 Federal LDGTs Eauipped
With Oxidation Catalvst Technology (see fable 7-1)





Mean





Mean
vehicle



Ena.
wean
veh icle
Eauivalent


Fnqine
Diso.
N/V
Poad Load
^est Weiaht
City
vanu facturer
Familv
(liters)
ra t io
(hp)
(lbs)
"lOCJ
Amer ican
EAM2.8T2AXE2
2.8
41. 6
14 . 0
3, 500
18 . 7
Motors






Chrysler
ECF2.2T2AAB5
2.2
39.5
10 .1
3 ,500
22 . 7

ECP2.6T2AACP
2. fi
44 . 2
10. 1
3 ,450
21.2

ErR3.7T1BBB0
3.7
42.6
16 . 8
4 , 500
15.5

FCP5.2T2BBF5
5. 2
39. 2
IB. 1
snoo
12.5

ECR5.2T4BAF0
5.2
39.7
18.4
4 , 750
12 . 2

ECR5.9TABAF9
5.9
40 . 0
19.7
5,375
10 . 2
Fu j i
EFJ1.8T2AFD3
1.8
48 . 5
9 . *
2 ,800
24 . 8
General
E1G1.9T2HEC9
1.9
42.8
10 . 4
2,925
26.4
Motor s
E1G2.0T2WJCX
2.0
4 9 . 9
12.0
3 ,700
19 . 7

E1G2.8T2FTC 7
2.8
42 . 8
13 . 5
3,875
15.A

E1G4.1T2HHD3
4 . 1
36.6
17.0
4 ,750
15.0

F1G5 . 7TMHHCX
5.7
30 . 3
lfi. 7
4 ,725
14 . 8
Isuzu
FSZ11QT'2AAG1
2.0
31.0
13.0
3 ,250
22.2
Mitsubishi
EMT2. OT2AFD5
2.0
48.7
11. o
2,^50
?5.3

FMT2 . 0T2 BCA,v
2.0
48 . 7
11.9
2,950
23.9

FMT2.6T2AFD8
2.6
46. 3
12 . 4
1,200
22 . 2

EMT2.6T2BCA2
2.6
46.3
12.4
3,200
21.4

FMT2.6T2BFJ8
2.6
27 . 2
15.9
3 , 625
19 . 6
Nissan
ENS 2.4^ 2AA FO
2 . 4
48 .1
13.0
3 , 300
21.4
^ovo ffoavo
FTK2.OT2AFL7
2.0
44 . 5
11.4
3 ,000
26.1
Tovota
ETY2.4T2AFF0
2-4
39 . 2
14 .8
3 , 200
22.0

ETY 4.2T2AFF4
4. ?
45.2
17.0
4 ,500
11.8

-------
2-13
Table 2-7
Summary Values From Tables 2-5 and 2-6
For Federal 1984 LDGTs
SPA	Fnaine
City	Disp.
nroq	(liters)
Fqu ivalent
Test	Road
Weioht	Load
(lbs)	hp
Three-Way	18.5
Familv Averaqe
Non-Three-Way 19.4
familv Averaae
3.49	3,891 14.7
3.07	3,740 13.9

-------
2-14
Figure 2-1
EGR Effect on Gasoline Enqine NOx
1- 0+
0.9--
Q. St
0.4-


10. 15. 20.
EGR RATE (%)
A/F = 15.0
3C.
NOTE: This graph was assembled from data presented in several
references,[9,10,11,12] and is a representative depiction o£
the NOx reduction capabilities of EGR. For example, if a
current LDGT is assumed to be calibrated at an average EGR rate
of 10 percent, with emission rates from 1.4 to 1.7 g/mi NOx
(see Table 2-3), then to reduce NOx emissions through EGR by 50
percent (i.e., to levels of 0.7 to 0.9 g/mi), NDx will need to
be reduced from .5 to .25 on the relative scale above,
requiring an increase in EGR from 10 percent to 20-25 percent.

-------
2-15
Figure 2-2
The Effect of EGR Rate on BSFC
in a Conventional Gasoline Engine
EGR RATE C=ERCENT)
MOTE: This graph was assembled from data presented in several
references.[9,10,11,12] Continuing the example begun in Figure
2-1, engines which have EGR rates increased from 10 percent to
20 - 25 percent to lower g/mi NOx by 50 percent will experience
an increase in BSFC, and presumably a decrease in mpg, of three
to seven percent (from .980 to 1.01 - 1.05 relative BSFC).
Vehicle dnveability and performance implications of such EGR
rates are unknown, being by necessity determined by subjective
driver evaluations. However, at least one reference [ 10]
reports significant engine misfire beginning at EGR rates
between 25 and 30 percent, corresponding to an approximate fuel
economy penalty of about 10 percent (relacive to a 10 percent
EGR rate) ; those values are assumed to be the very worst case
(i.e., vehicle drivers will report unacceptable losses in
dnvaability well before this point).

-------
2-16
catalyst systems to provide better fuel economy and/or
performance than oxidation catalyst systems with increased EGR
and timing retard, on an average vehicle.
This inference is substantiated by engineering logic.
With a three-way reduction catalyst, significant NOx control
takes place in the exhaust system. This strategy reduces the
need for NOx control in the combustion chamber (i.e., reduces
the need for alteration of the combustion process itself).
This permits the amount of EGR and the ignition timing to be
calibrated to oDtimize engine performance (fuel economy,
driveability, etc.), and not to be calibrated to a large extent
for NOx emission control purposes. Some EGR is usually
advantageous in conventional gasoline-fueled engines (see
Figure 2-2) ; beyond this level of EGR, however, a three-way
catalyst system will provide more fuel efficient NOx control
with better vehicle driveability than increased EGR (at
constant air/fuel ratio and ignition timing). When one also
considers the enhanced calibration, engine control, and precise
fuel metering made possible by the electronics used in
closed-loop three-way systems, the fuel economy advantage of a
three-way system should be even greater. Indeed, the
additional electronics probably contribute more to enhanced
fuel economy than the deletion of excess EGR, considering the
relative size of the fuel economy tradeoffs presented in Fiaure
2-2. If these relative tradeoffs are representative, then the
ultimate determinant of maximum EGR is likely its effect on
vehicle driveability and performance.
As can be seen in Table 2-7, both the interpretations of
current product mix and engineerinq loqic are not at face value
substantiated by available test data: fuel economies for the
three-way equipped vehicles appear lower than those of
non-three-way vehicles. (It is assumed that all vehicles
deliver commercially acceptable driveability and performance.)
However, there are significant vehicle differences aside from
emission control systems which must be taken into account
before relative fuel economies can be accurately compared.
Referring again to Table 2-7, the vehicle difference
between the three-way and non-three-way fleets, in all cases,
made the three-way technology appear less fuel efficient than
it actually is. For example, the three-way fleet is, on
averaae, heavier than the non-three-way fleet. Because heavier
vehicles require more energy to move over the road, they are
less fuel efficient. Mso, the road load horsepower (hp)
(i.e., the power required to overcome rolling resistance and
aerodynamic drag while the vehicle moves at a constant 55 mph
over smooth, level road) is hiqher for the three-way fleet
(14.7 hp vs. 13.9 hp), again lowering its measured fuel economy.

-------
2-17
Perhaps the most important difference between the two
fleets is the difference in average engine displacement (3.49
liters vs. 3.07 liters). Greater displacement gives engines
more reserve power, as one would expect to find in engines used
in heavier vehicles. This reserve power, however, is costly:
higher displacement engines run less efficiently, all other
factors being equal. (At :he typical cruising road loads
observed in both the FTP and on-road operation, an engine will
operate at some fraction of its maximum available power. For a
gasoline engine, this means that the inlet air flow is
throttled and that the engine is using more energy to pump air
into the combustion chambers. For this reason, Otto cycle
engines are more efficient at wide-open throttle (assuming
constant air/fuel ratio and engine speed) where pumpinq losses
are minimum. The greater the engine displacement, however, the
more air throttling is required at a given road load horsepower
and engine speed. This higher throttling makes the engine work
harder to deliver the same amount of output power, and hence,
is less efficient.)
This throttling effect also comes into play when the fleet
engine rpm/vehicle speed (N/V) ratios are considered. (The N/V
ratio is determined by the amount of speed reduction designed
into a vehicle's transmission and rear axle aears.) For a
aiven road speed, vehicles whose engines are turning faster
(i.e., have higher N/V ratios) are less fuel efficient. This
is due to more energy being required to pump engine intake air
at a hiaher speed, and also due to increased frictional losses
associated with hiaher speed. (These losses are, to some
extent, reduced by an engine's propensity to lose less energy
at higher speeds through heat transfer.) Again, the higher M/V
ratios of the three-way fleet reduce its measured fuel
economy. when all of these fleet differences are taken into
account (i.e., when the fleet fuel economy is corrected to a
value which would be seen if all vehicle parameters between
fleets were identical), the true relative fuel economies of the
three-way and non-three-way technologies can be compared.
Correction factors have been developed from EPA' s
certification and fuel economy data bases to allow a direct
comparison of emission control system fuel economies despite
dissimiliarities in vehicle parameters.[7,61 These correction
factors are used by this analysis and the results of this
correction exercise are presented in Table 2-8. When vehicle
and engine displacement effects on measured fuel economy are
eliminated, three-way vehicles are indeed found to be more fuel
efficient at the 2.3 g/mile standard than non-three-way
vehicles. This confirms our earlier interpretation of the
manufacturers' product mix, i.e., that some tangible benefit is
received from the additional investment in three-way systems at

-------
2-13
Table 2-8
Corrected and Uncorrected vehicle Data
For 1984 LDG^s
Non-Three-Way
Vehicle Averaae
Corrected* Non-
Three-Way
Vehicle Ave.
Actual Three-Way
vehicle Average
Percent Difference
(actual three-wav
relative to
corrected non-
three-wav):
EPA
City
Tipg
19 . 4
17 . 1
1R . 5
DlSP,
3 .07
3 .49
3 .49
Eau ivalent
Test
Weiqh t
(lbs)
3 , 740
3 ,891
3 ,891
Road
Load
hp
13.9
14 .7
14 . 7
N/V
Ra t io
41.9
43 . 6
43 . *
+ 8 . 2%
* The correction equation, adaDted from Reference 7 (for
Etw, ho, and N/v) and from Reference 8 (for displacement),
is:
p ID - CTD	FTW - tr T\\j
Corrected hdq - moa 1-a	mnr new	old, „ „-i r" new - old,
	:	°Id = ~ r—rjD 	1 -.441 [	ETw 	1
old	old
HP - HP ,, N/V ,, - N/V
_.2n9[_n^	2ld1 _-138[_old 	new,
old	old
where rnpq0-Ld, CID0ld' ^TWold' HPold» and	are
non-th r ee-wa v parameters and ETWnew, CIDnew, HPnewi anc1
N/vnew are three-wav based parameters. Corrected mpq is thus
the fuel economy non-three-way vehicles would exhibit _i_f their
CIO, ETW, HD, and N/V oarameters were the same as those in
three-way vehicles, i.e., if the only difference between the
vehicles were the emission control system.

-------
2-19
a 2.3 NOx standard. This result is also consistent with
enqineering logic. At any NOx standard at or below 2.3 q/mi, a
three-way catalyst system should be inherently more fuel
efficient and/or give better performance for the "average"
LDGT. Based upon the above analysis, the current three-way
LDGT fleet exhibits NOx emissions 40-50 percent less than the
non-three-way fleet, and would simultaneously get about eight
percent better fuel economy than identical vehicles equipped
with non-three-way technology.
This leads us to an evaluation of the net fuel economy
effects associated with a 1.2 NOx standard, assuminq the
product mix presented in Table 2-4. This evaluation is
summarized in Table 2-9, which weights the expected fuel
economy chanaes for each potential technology with the expected
market share for each technology. Taking into account the
worst case and most likely fuel economy penalties and benefits
associated with our projected manufacturers' product mix, a
fleetwide fuel economy benefit should be realized through
compliance with the 1.2 q/mi NOx standard. Our analysis
indicates that this would probably represent a 2-4 percent
imnrovement in overall fleet fuel economy.
4. LDGT Summary
A 1.2 q/mi NOx standard for 1987 LDGTs is easily
feasible. Many 1984 LDGTs are today usinq the required
technoloav and exhibit emissions at or below 1987 target
levels. The entire class of LDGVs have been meeting an
essentially equivalent NOx standard with the same technology
since 1981.
The fuel economy impact of this standard on LDGTs is
dependent upon the manufacturers' individual choices in
establishing product mix. However, our analvsis indicates very
little likelihood that a fuel economy penalty will be observed
on a fleet-wide basis. Based upon available data, a fleetwide
fuel economy benefit of 2-4 percent may result.
C. Light-Duty Diesel Trucks (LDDTs)
1. Low-Mileage Target Derivation
The low-mileage NOx emission target for LDDTs is derived
using the same methodology used for LDGTs. Based upon the 1987
certification data in Table 2-10, a full-life low-mileage
target range of .9 to 1.1 g/mi NOx is necessary to comply with
a standard of 1.2 g/mi. (This target range was derived
assumina an AOL factor range of 1.1-1.3,* and a calculated
full-life multiplicative DF of 1.007.)
Again, the imprecision involved in using a ranoe of AOL
factors will be shown to be immaterial to this analysis'
assessment of technological feasibility.

-------
2-20
Table 2-9
Fuel Economy Implications of a 1.2 a/mi
NOx Standard for 1987 LDGTs
1984
Emission
Cont rol
System [ 11
3 WY - C L
3WY-CI;
3WY-0L
1WY-0L
OX-CAT
OX-CA^
OX-CAT
Projected
19 8 7
Emission
Control
Percent
of Total Fuel vaanitude of Fuel
LDGE Econorav fconomv Effect
Fvstem[ll Other Chanaes Sales
Effect Worst Case Likelv
3WY-CL None
15% f 21 None
3WY-CL Increased EGR, 16% T 21 Penalty
recallhration
3WY-CL Calibration	11%
opt im izat ion
None
0% [31
-10%[41
0% [5]
3WY-OL Increased EGR, 0-.7% Penalty -10%[dl
reca1ibra t ion
3WY-OL Calibration
optim ization
3VTY-CL Calibration
ODtinlzat ion
3%	Benefit +8.2%[6]
<19-55% Benefit +8.2*[61
OX-CAT Increased EGR, 0.2-5% Penaltv -*%[71
r ecalibration
Sales-Welahted Fuel Economv Chanqe: +2 to 3%,
0% [31
-5%[41
0% f 51
-*%[41
+ 8 .2% r61
+ 8 . 2* fhi
-6% r 71
J-3 to 4%
[ 11 From "'"able 2-4 .
[21 Assumes that 50 percent of 1984 LDGT 3WY-CL conf iou rat ions meet the
1.2 NOx standard with absolutely no effort; the remainina 50 percent
have emissdions above 1.2 (see Table 2-2) and are assumed to reauire
further non-catalyst NOx control (i.e., increased EGr and timing
retard).
[3] No emission control svstem changes are projected; hence, no fuel
economy effect is projected.
[41 A 10 percent penalty is considered an absolute worst case value (see
Figures 2-1 and 2-2) ; the 5 percent penaltv is nudaed to he more
likely for an "average" vehicle, representing the penalty observed
with EGR induced NOx reductions of 50 percent from the NOx level
seen at the most fuel efficient EGR calibration.
f 5 "I Minimal fuel economy benefits are presumed attributable to the
adoption of closed-loop technology; NOx reductions are attributable
to more efficient reduction catalyst operation. This is a verv
conservative estimate, whereas it lqnores the likelv significant
improvements attributable to the addition of electronic enqine
controls.
[51 From Table 2-8.
f71 A 5-7 percent oenaltv is ludaed the most severe any manufacturer
would accept; if the penalty were higher, closed-loop and/or
three-way technology would be adopted.

-------
2-21
Table 2-10
Certification Data for 1984 LDDTs
Emiss ion


Control
Half-Life
Low-Mi
Manufacturer
Enqine Familv
System**
NOx DF
NOx (¦
Small Enqines
With EGR:



I suzu
ESZ137K6JBD2
EGR
.968
.85
I suzu
ESZ137K6JBD2
EGR
.968
. 76
Mitsubishi
EMT2.3K6JCA3
EGR
.946
. 78
Mitsubi sh i
EMT2. 3K6JCA3
EGR
.946
.67
Toyota
ErpY2 .4K6JCC2
EGR/OTR
.976
.76



Average:
. 76
Small Enqines
Without EGR:



Ford
F.FM2 . 2K6JAG6
EM
.992
1. 18
CM
E1G2.2K7ZZ97
EM
.948
2.08
GM
E1G2.2K7ZZ97
EM
.948
2.03
Grumman-Olsen
EGRl.6K6JAA5
EM
1.000
1.23
Grumman-Olsen
EGRl.6K6JAA5
EM
1.000
1.03
I suzu
ESZ137K6JCD4
EM
.948
1.73
Tsuzu
ESZ137K6JCD4
EM
.948
1.72
Nissan
ENS2.5K6JAF6
EM
1.021
1. 50
Mitsubishi
EMT2.3K6JFD1
EM
1.002
1. 54
Mitsubishi
F.MT2 . 3K6JFD1
EM
1.002
1.57
Tovo Koqvo
ETK2.2K6JCS1
EM
.992
. 85
Toyo Koqyo
FTK2 . 2K6.TCS1
EM
.992
.83
Toyo Koqvo
ETK2.2K6JFR3
F.M
.992
1.20
Toyota
ETY2. 4K6JFF0
EM
1.003
1 .56



Averaqe:
1.43
Larae Enqines
:



GM
E1G6.2K7ZZ41
EGR
1.012
1.85
GM
E1G6 .2K7ZZ41
EGR
1.012
1.P1
GM
E1G6.2K7ZZ74
EGR(elec.)
1.012
1. 79
GM
E1G6.2K7ZZ74
EGR(elec.)
1.01?
1.32

Averaae (all
LDDTs):
1 .003***
1.33
* Some enqine families are duDlicates, but represent the
same enqine family in a different vehicle confiquration.
** EGR = Exhaust Gas Recirculation
EM = Enaine Modification
OTR = Other
*** All DFs less than 1.000 are assumed to be 1.000 for
purposes of average DF calculation.

-------
2-22
2. Current LDDT Emission Levels (small engines)
With respect to emissions, the LDDT fleet is best
discussed in two separate groupings: 1) those vehicles with an
ETW below 4,000 lbs, and 2) those with an ETW above 3 ,999 lbs
or with a GVW above 6,000 lbs. The majority of LDDT enqine
families are of smaller displacement and are found in the
lighter vehicles; indeed, only a single engine model, GM' s
6.2-liter engine, is currently found in the heavier vehicles.
(The 6.2-liter engine, however, reoresents about 60 percent of
total LDDT sales; its unique market niche and its high sales
make it an important engine family to consider. Also, given
its pooularity, other similar engines are likely to follow.)
Thus, the current LDDT class could be split semantically either
by ETW/GVW or by engine displacement. For convenience,
reference to engine displacement (i.e., small vs. large will be
used here).
For these liqhter vehicles, another segmentation is
necessary to properly evaluate NOx emissions performance. This
is accomplished in Table 2-10, where engines are grouped by
emission control system. (The 6.2-liter engine families are
also included for completeness.) As is readily apparent, all
small engines equipped with EGR have current NOx emissions
below the 0.9 to 1.1 g/mi taraet range. These are California
vehicles meeting a 1.0 g/mi NOx- standard. These families all
have Federal counterparts which are not equipped with EGR, and
whose emissions lie well above the 1.2 g/mi NOx emission level,
as do the emissions of the majority of other small engines not
equipped with EGR. This data substantiates the ability of EGR
to reduce NOx emissions, and also leads one to conclude that
all small LDDT engines, if equipped with EGR, could also meet a
1.2 g/mi NOx standard. The enaine and vehicle size of those
small LDDTs at or below 0.9 g/mi NOx are representative of the
small LDDT fleet.
This conclusion is supported by the presentation in Table
2-11. Again, small displacement LDDT engines are segmented by
emission control technology, but a clearer distinction is made
between those engines not equipped with EGR on a Federal basis
(but equipped with EGR in their California counterparts), and
between those remaining non-EGR engines. By equipping
virtually identical engines with EGR, NOx emissions were
reduced, on the average, from 1.62 to .76 g/mi--a reduction of
53 percent. As can also be seen in Table 2-11, the remaminq
small engines without EGR would not need as great a reduction
to comply with a 1.2 g/mi NOx standard. At worst (i.e.,
assuming the lower end of the target range would need to be
met), emission reductions of 0 to 40 percent would be required;
on the average a modest reduction of 20 percent (1.12 g/mi to

-------
2-23
Table 2-11
Direct Comparison of Vehicle Data
for Similar 1984 LDDT Engine Families



EPA

Ecru i va lent
Road


4K
4K
City
Displacement
Test Weiqht
Load

Manufacturer
NOx
HC
mpg
(liters)
(lbs)
hD
N/V
Small Enaines
with
EGR





I suzu
.85
.28
34 . 6
2.25
3 ,125
10.8
43 . 6
I suzu
.76
.36
30.9
2.25
3 ,375
13.0
55.0
Mitsubishi
.78
. 19
27.8
2 . 30
3 , 500
13 . 5
42.5
Mitsubish i
.67
. 16
31.2
2 . 30
3,125
11.4
42 . 5
Toyota
.76
.13
33 .1
2.40
3 ,125
12.0
45.5
Averaqe:
.76
.22
31.5
2 . 30
3 ,250
12.1
45.8
Similar Small
Enqines Without EGR*:



T suzu
1.73
.23
30 . 6
2.25
3 ,375
13.0
55.0
T suzu
1.72
.22
30.3
2.25
3 ,125
10.8
56. 1
Mitsubishi
1. 54
.09
30 .1
2 . 30
3 , 500
13 . 5
42 . 5
Mitsubishi
1.57
.15
32.1
2 .30
3 ,125
11.4
41.5
Toyota
1.56
. 12
31.5
2.40
3 ,625
13.8
46.0
Averaqe:
1.62
.16
30.9
2 .30
3 ,350
12.5
48.2
Other Small Engines
Without EGR
:



Ford
1.18
.27
32.4
2 . 20
3 , 375
11.0
52.0
Grumman-Olsen
1.23
.26
35.3
1.60
2 , 375
12.7
51.8
Grumman-Olsen
1.03
.24
40.4
1. 60
2 ,375
13 .7**
39 . 3
Nissan
1. 50
.23
33. 6
2 . 50
3,250
11.4
42.6
Tovo Koqvo
.85
. 16
31.4
2 .20
3 ,250
12 . 5
46.0
Toyo Koayo
.83
.14
32 . 5
2 . 20
3 ,125
12 .5
46.0
Tovo Koqyo
1.20
. 12
33.1
2.20
3 ,250
12 .5**
45.7
Average:
1.12
.20
34 .1
2 .07
3 ,000
12 . 5
46.2
Non-EGR engine families which have a counterpart enqine
equipped with EGR.
Value assumed (actual data unavailable).

-------
2-24
0.9 g/mi) is necessary. As demonstrated by three families
(five vehicle configurations) , these percentage reductions are
well within the capability of existing EGR systems.
Three potential tradeoffs
diesel EGR: increased HC
durability, and decreased fuel
increased particulate, is not
presented at the end of the sect
standard for heavier LDTs.)
are typically associated with
emissions, decreased engine
economy. (A fourth tradeoff,
discussed here; it will be
ion addressing the 1.7 g/mi NOx
With respect to HC, there appears to be very little
likelihood that the 0.8 g/mi HC standard will be exceeded. Of
the engines already equipped with EGR, HC increased jv 38
percent relative to their non-EGR counterparts, but at levels
still well below the 0.8 standard. (For example, HC increased
from 0.16 to 0.22 g/mi, on the average; HC emissions of the
highest emitting EGR-equipoed engine were only 0.36 g/mi.)
Assuming the worst case scenario, that a 38 percent increase in
HC would occur on the highest emitting of the remaining small
engines in Table 2-11 (i.e., the Ford 2.2-liter engine
experienced an increase in HC from 0.27 to 0.37 g/mi),
compliance with the 0.8 g/mi HC standard is still assured.
Although small HC tradeoffs are expected, they should in no way
jeopardize compliance.
As for engine durability, the major concern lies with the
effect of recirculated exhaust particulate on turbochargers and
on the lubricating Qualities of the enaine oil. Since no
current LDDT is equipped with a turbocharqer, this aspect of
the oroblem does not exist. With respect to engine oil, more
frequent oil change intervals are the simplest solution to
potential oil degradation problems. However, some engines are
alreadv equipped with EGR and, to EPA1 s knowledge, are not
experiencing decreased durability, nor utilize shorter oil
change intervals. At any rate, oil degradation, if observed at
all, is a problem easily dealt with using techniques alreadv
developed for LDDV NOx control.
To assess the fuel economy imDact of EGR on these engine
families, an analysis similar to that performed on LDGTs is
necessary. In short, a direct comparison is made of those
engine families with and without EGR, and when all
vehicle-to-vehicle differences are eliminated, the differential
fuel economies are assumed to be attributable to the increased
emission control. This is presented in Table 2-12. Unlike
LDGTs, however, two of the correction terms (i.e., those for
eng ine displacement and N/V ratio) may be somewhat
inappropriate for diesel engines.* Since diesel enqmes are
unthrottled, i.e., have high volumetric efficiencies at all
The correction factors were developed from EPA's
certification data base, which is overwhelmingly comprised
of gasoline-fueled vehicles.

-------
2-25
Table 2-12
Corrected and Uncorrected Vehicle Data
	For Specific 1984 LDDTs	
Road
EPA	Equivalent	Load	N/V
City mpa	Test Weight	hp	Ratio
Non-EGR Enqine	30.9	3,350	12.5 48.2
Average*
Corrected**	31.5-31.8*** 3,250	12.1 45.8
Non-EGR Enqine
Average
EGR Engine 31.5	3,501	12.1 45.8
Average	
Fuel Economy
Penalty Attributed
to NOx Emission
Control:	0-1%
* From Table 1-11.
** Corrected using the"equation from Reference 7:
mpq - mpa , ,	ETW - ETW , ,
^new e old	.._ r n®w	old,
	ETH^j	1
HP - HP , ,	N/V „ - N/V , ,
• 209 [ ne^p °ld) - .1881	ne
old	old
*** The range in corrected mpa values was derived by first
excludmq the N/V correction, and then including the h'/v
correction (see text).

-------
2,-26
loads, much of the increased pumpinq losses seen in throttled
gasoline engines due to larger displacement or higher enaine
speeds (at constant road load) are not observed in diesel
engines. Since Table 2-12 only compares similar engines,
however, the issue of displacement disappears. As for N/V
ratio, since enaine friction increases with increasing engine
rpm, at least some fraction of the N/V correction is
appropriate; for estimation purposes, the N/V correction is
both included and excluded in our total correction, yielding
the maximum and minimum likely corrections. The results of
this direct comoarison in Table 2-12 indicate that, in
achieving 53 percent reductions in NOx, vehicle fuel economy
decreased 0-1 percent. When one considers that the remaining
small LDDT engine families require, on the average, only about
20 percent reductions in NOx to meet a 1.2 standard, the fuel
economy impact of the commensurately smaller amount of EGR
should be negligible. In summary, the entire fleet of small
Federal LDDTs--representing about 40 percent of projected 1984
Federal sales—should experience no measurable fuel economy
penalty on account of the 1.2 g/mi NOx standard.
3. General Motors' 6.2-Liter Engine
As can be _seen in Table 2-13, GM's 6.2-liter engine is
found in LDDTs almost twice as heavy as those of other LDDT
engine families. The 6.2-liter engine is already equipped with
EGR, and in their California family, with electronically
controlled EGR. In the heaviest vehicles (6,000 lbs ETW), the
6.2-liter engine will require nearly a 40-50 percent reduction
in NOx emissions (from 1.3 to 0.9-1.1) to comply with the 1.2
g/mi standard. In the lighter vehicle (5,000 lbs ETW), NOx
levels of 1.32 have already been achieved with electronically
controlled EGR (see below). For this particular engine and its
various vehicle conflaurations (representing 60 percent of
projectd 1984 Federal sales), the feasibility of a 1.2 NOx
standard depends upon the ability of the 6.2-liter engine,
through either increased EGR, injection timing retard, or other
eng i ne mod ifications, to achieve the required NOx reductions
without unacceptably compromising engine performance or other
emissions.
Typically, diesel engines, because of their inherently
lean heterogeneous combustion, are capable of tolerating larger
amounts of EGR than conventional* gasoline-fueled engines.
"Fast-burn" combustion chambers can substantially improve
the EGR tolerance of gasoline-fueled enqines, although
this technology has yet to be widely adopted even amonq
passenger cars.

-------
2-27
Table 2-13
Vehicle Data For 1984
GM 6.2L LDDT Enaine Families
GM 6.2L Enaine
4K
NOx
4K
HC
EPA
City
mpg
Equ ivalent
test
we iqht
(lbs)
Road
Load
hp
N/V
EGR
1.85
.22
18.2
6,000
15.1
31.4
EGR
1.81
.33
17.8
6,000
15.1
28.8
EGR (electronic)
1.79
. 27
17.6
6,000
15.1
31.4
EGR (electronic)
1.32
. 18
21.0
5,000
17.2
24. 5
Averaqe
1. 69
.25
18.7
5,750
15.6
29.0

-------
2-28
Since the smooth and Cast propagation of a flame front in
gasoline engines is impaired by too much oxygen-deficient EGR,
there is a limit to the amount of EGR gaso 1 ine-fueled engines
can accept before performance suffers. On the other hand,
diesel engines run very lean (oxygen rich) , and even large
amounts of EGR fail to dilute the oxygen sufficiently to impair
turbulent heterogeneous combustion at lighter engine loads.
Above some level of EGR, however, even diesel engine
performance is impaired. This can be inferred from Figure 2-3,
in which the effect of EGR on several parameters is plotted for
a small (2.2-liter) diesel engine. As fuel flow is increased
(i.e., as the air/fuel ratio becomes less and less lean, and
less oxygen becomes available for any given fuel droplet), a
point is reached beyond which more EGR penalizes fuel
consumption, and drastically increases HC and smoke emissions.
Quite simply, complete combustion is being impaired. The same
effects would be observed at lighter engine loads (i.e., lesser
fuel flows} if even more EGR were added. Again, the ma]or
mechanism is the reduction of the amount of oxygen relative to
the amount of fuel to be burned.
From the above discussion, it can be deduced that at
lighter engine loads a greater percentage of EG"R can be
tolerated because of the greater amount of excess oxygen.
Indeed, this is typically observed in EGR calibrations - as the
load is increased, the relative percent of EGR decreases. This
is in fact the calibration scheme for GM' s 6.2L engine (see
Figure 2-4). [14] m almost all cases, this calibration is
maintained by simple mechanical means, e.g., by mechanically
controlling EGR flow as a function of throttle position.
However, mechanical control systems are not capable of
precisely controlling EGR systems to prede te rm ined" optimum
mode-specific* calibrations; only a general calibration is
possible. ror this reason, electronic control systems are far
more effective than mechanical systems because of their
potential for monitoring more than one control parameter (e.q.,
throttle position, engine speed, and EGR valve control vacuum),
and their ability to set the optimum EGR flow for predetermined
combinations of control parameters. The 1984 California
version of the 6.2L diesel is equipped with electronically
controlled EGR of this type. For this particular enqine, GM
stated that "with more stringent NOx standards, the requirement
for optimization of the inverse NOx/particulate relationship
can be achieved by the electronic EGR control system."[141
As determined by specific modes observed over the
applicable test cycle.

-------
'-3
0)
?r
0)
D
i-h
f-t
O
3
IT)
tn
Q
'i
(B
D
O
n>
FUEL CQNSLHPTlON
(G/BHP HR) -COfPPMl HC(PPM_] SMOKE (BOSCH No.) NOx(ppm)
O
c —
mo
r
r
O |0
^oi
u) O
-I
o
(\) l\) f\) fO
o (\) 1^ (D
o o O o o
>
N
_N
V

V
c-"

V
oj rv)
o o
V® VP
c>
o O

:o
)>
J
o
LJy ^
;(,r"
IvX
5.« K
(!) r
qo m
~n ~n
c m
m C|
>n
c
f-s
n>
M
CO
o
2:
CO
K)
t
K)
VO

-------
2-30
Fiqure 2-4
loot

DTEStTL EG* KEOUIRTMEVT
At a Given bev«l of
HO* and Particulate
Control
EXCTSS eg a
•	Decreased r*Ox
•	Increased Particulates
insurfTctew re*
•	Increased NO*
•	Decreased Particulates
KH
\ of
Zntak*
Air
0*
Closed
folly
Open
LOAD/THTOTrLK AWGT.Z
AT A GIVEN RP*
(Taken from Reference Id)

-------
2-31
(The applicable 1984 California NOx standard is 1.5 g/mile.)
This mode specific NOx control maximizes NOx reductions while
minimizing adverse tradeoffs.
Since GM's 6.2L engine in California is already equipped
with electronic EGR, the main question to be addressed at this
point is whether or not NOx emissions can be further reduced to
permit compliance with a 1.2 g/mile NOx standard. To answer
this directly, one would need to determine whether or not the
EGR rate could be increased. Sufficient information about this
engine does not exist to make this determination. Thus, more
indirect means must be used.
This alternate approach compares the stringency of the 1.2
g/mi NOx standard for the 6.2L engine to that of the 1.0 g/mi
NOx standard for GM's largest LDDV engine, the 5.7L. Large
LDDVs with this engine meeting the 1985 1.0 g/mi Federal NOx
standard provide the closest analogy to the compliance of the
5.2L engine to a 1.2 g/mi NOx standard. The primary
differences in the two situations is that the LDDT is heavier
and boxier than the LDDV and requires more power to operate
over the EPA city cycle, thus producing more NOx emissions.
It can reasonably be assumed that the increase in NOx
emissions is proportional to the amount of fuel consumed over
the cycle (or inversely proportional to fuel economy). Engine
testinq has shown NOx emissions to be relatively constant on a
brake-specific basis over a wide variety of cycles. Given that
LDDV and LDDT engines have the same basic designs and, thus,
the same efficiencies, NOx emissions that were constant on a
brake-specific basis would also be constant on a fuel-specific
basis. Thus, the relationship between LDDV and LDDT NOx
emissions can be approximated by the following equation:
NOx (LDDT) = mpg (LDDV) 
-------
2-32
of NOx deterioration. Using the full-life (120,000 miles) DF
of 1.007 estimated previously, the last 70,000 miles should
account for an additional DF of 1.004. As can be seen,
however, this adjustment is negligible and will be ignored.
Also, since the SEA requirements for LDDVs are the same as
those for LDDTs, the SEA factors can be ignored and the two NOx
standards can be compared directly. Thus, the 17 percent
greater fuel consumption of the 6.2L engine should produce 17
percent greater NOx emissions and a standard of 1.17 g/mi NOx
should be achievable. As this is slightly below 1.2 g/mi, one
can conclude that the same technology which will allow the 5.7L
engine to comply with a 1.0 g/mi NOx standard in 1985 should
provide compliance with a 1.2 g/mi NOx standard in 1987. Of
course, this is an approximation since there are other more
subtle differences between LDDVs and LDDTs. However, it
provides a strong indication that a 1.2 g/mi NOx standard is
achievable.
The effect of meeting the 1.2 g/mi NOx standards on fuel
economy and particulate emissions will not be addressed here.
These topics will be addressed at the end of the following
discussion of the feasibility of the 1.7 g/mi standard for
heavier LDDTs.
4. LDDT Summary
For 40 percent of total LDDT sales, i.e., those engine
families of smaller displacement in the lighter vehicles,
compliance with a 1987 NOx standard of 1.2 g/mi should be
easily feasible. EGR will be necessary in probably all but one
engine family to permit compliance and, overall, any fuel
economy penalty should be negligible. The demonstration of
technological feasibility for the GM 6.2L engine (i.e., larger
LDDTs) is less straightforward; feasibility depends on
extrapolation of data from large LDDVs. However, all
indications point toward the 1.2 g/mi standard being feasible
for these large LDDTs.
D. 1.7 g/mi NOx Standard for Heavier LDTs
There are three reasons for considering a 1.7 g/mi NOx
standard for heavier LDTs. First of all, while apparently
feasible, a 1.2 g/mi NOx standard for heavier LDDTs will affect
the stringency of the 0.26 g/mi particulate standard, primarily
the number of vehicles requiring trap-oxldizers to meet the
standard. Secondly, while a 1.2 g/mi NOx standard is clearly
feasible for heavier LDGTs, a more stringent NOx standard for
LDGTs than LDDTs would economically encourage the purchase of
LDDTs, which the Agency does not wish to do. Thus, equivalent

-------
2-33
standards are desirable. Thirdly, 1995 LDT NOx emissions can
be maintained at current levels with a 1.7 g/mi NOx standard
for heavier LDTs, gasoline-fueled and diesel.
In splitting the LDT class at 4,000 lbs ETW and 6,000 lbs
GVW, EPA is following the precedent already set by California
and the Clean Air Act. There is also a natural break in LDT
sales at this point, oarticularly for diesels.
The LDT engine families certified in vehicle
configurations over 3,999 lbs ETW are presented in Table 2-14.
Also presented are current emission control systems, and
orojections (and assumptions) for likely compliance strategies
for a 1.7 standard.
The analysis presented in Table 2-14 is easily
summarized. For LDGTs, a 1.7 standard still appears to require
the majority of engine families to adopt closed-loop three-way
technology, though a few additional oxidation catalyst equipped
families should be able to remain non-three-way. However,
since less severe engine calibrations are necessary, there is
the potential for significant (5-10 percent) fuel economy
improvements for selected vehicles. At the same time, a couple
of familes could see a decrease in fuel economy by avoiding a
change to closed-looo, three-way technology. However, this
would be the manufacturer's choice to trade-off first cost and
fuel economy. In any event, on a sales-weighted basis, no net
fuel economy chanqe is projected relative to a 1.2 g/mi
standard. Relatively minor changes to fleetwide emission
control system mix are expected relative to a 1.2 g/mi NOx
standard; these are summarized in Table 2-15.
With respect to heavier LDDTs, the primary effect of a
more relaxed NOx standard would be reduced EGR rates. Given
that GM has already gone to electronics for California's 1.5
g/mi standard, electronics would still likely be used at 1.7
g/mi. Given this use of electronics, there should be no fuel
economy penalty with respect to the current 2.3 g/mi standard
(i.e., any penalty associated with the 1.2 q/mi standards
should disappear).
This finding is confirmed by the LDDV data contained in
Table 2-16. The fuel economies of these LDDV engine families
are shown under various NOx standards. As can be seen, the
fuel economy effect of either the 1.5 or 1.0 g/mi NOx is very
small, and in no case more than one percent.

-------
Table 2-14
Implications ot a 1.7 NOx Standard for tJTs Over 4,000 Lbs Em
Manufacturer Engine Family 111
Gasoline Engines
Mean	Low-
fci}uivalent	Mileage
'Itest Weight	NOx
(lbs)	(g/mi)
Low-
Mileage
Targets
for
1987 1.7
tg/mi) 121
1984 tlnission
Control Systems
Proiected Compliance Strategy
for 1987 1.7 q/mi NOx Standards	
Incremental
Fuel Economy
Change
Relative to
1.2 NOx
Standard(41
Additions of
Technology(3)
Changes to
Calibrations:
BGR, tuiung|3|
AMC
EAM258T2HEAX
4,250
1.17
1.0-1.2
a^R/PLS/SCL
None
None
+5%

EAM258T2HEAX
4,500
1.79
1.0-1.2
UGR/PL3/3CL
None
Yes
+ 10%

EAM360T2HUX
4,750
1.27
1.0-1.2
EER/mP/3WY
None
Yes
0%
Chrysler
BCRJ.7T1UBU0
4,500
I.6L
1.2-1.4
fcUH/ rwp/oxD
3CL[5|
Yes
omi

fcCRS. 2T2QBF5
5,000
1.77
1.2-1.4
EEH/MP/OXD
3CL
Yes
0%

ECK5.9T4UAF9
5,250
1.37
1.2-1.4
EEiyrWPAXD
None
(¦tine
-8%
Ford
EfM4.9TbUuS
4,583
.72
1.0-1.2
EGfVFMP/CKD/3CL
None
None
0*

EFM5.0T2GAF2
5,125
1.38
1.0-1.2
0GJVrW/CKD/3WY
CL
Yes
0%

EFM5. BT2HGG0
4,625
.69
1.0-1.2
KH/H-1P/OXD/3CL
None
None
ot

EIM5.8T2HGG0
5,375
1.86
1.0-1.2
ECVFMP/0XD/3CL
None
Yes
+ 10*

E1-M5.8T4UAK3
4,625
1.76
1.0-1.2
EGIVFMP/OXD/3WY
CL
Yes
+5%
CM
K1G5.7T4HIICX
4,750
1.61
1.2-1.4
tUR/FMP/OXD
3CL| bl
Yes
0%f 51

LLG5.7T4HI1CX
5,500
2.20
1.2-1.4
tr.lVFWP/OXD
3CL
Yes
+5%
Out ton
EDN6.0T4HKAX
6,000
1.40
1.0-1.2
tiGH/mP/OXO/3CL
Hone
Yes
+5%
'Itoyota
ETY4.2T2AFF4
4,500
.99
1.2-1.4
KR/miyoxD/orH
None
None
0%






Sales-Weighted Average[6|:
0%
Diesel tjigines








GM
E1G6.2K7Z241
6,000
1.83
1.3-1.5
BGR
Electronic EER
Yes
0%






Sales-Weighted Average(7|:
Oft
(1| Where significant differences in NOx emissions arise between configurations of the same family tested at different
tflVIs, all relevant ElWs are presented.
121 Based upon full-life DFs of 1.286 (tliree-way catalysts), 1.070 (oxidation catalyst), and 1.007 (diesels), and an AQI,
range of 1.1-1.J; also based upon the technology in the 1984 vehicle.
(31 Judgments are made whether the addition ot technology is needed to lower 1984 low-mileage NOx levels to 1987 target
levels. In general, additions of technology are presumed necessary for any current three-way vehicle with NOx
emissions gieater than 1.3 y/mi, and for any non-three-way vehicle with NOx emissions greater than 1.5 g/mi. ('Hie
difference arises primarily from the difference in LMTs for the specific technologies. I Where technology diaries are
expected, ot wl*ere technology changes are not expected but emissions must still be lowered to meet target levels, then
recalibrations are assumed necessary.
14] 'Ihese values are drawn from Table 2-9. Note that the estimates are still conservative with respect to the addition of
electronic controls: no fuel economy benefit is assumed where open-loop three-way systems are converted to closed loop.
i5| It is also possible that manufacturers would letam open-loop, oxidation catalyst technology for these vehicles. If
so, then a fuel economy penalty uf up to 8 peicenl could occur.
161 liased upon inanutactuierb' confidential sales pro k-cL ions for 1984.
ba~ igh »r a lis, .nd	t.

-------
2-35
Table 2-15
Approximate Projected Market Shares
of Emission Control Technoloqies for 1987
LDGTs Meeting NOx Standards of 1.7/1.2 g/mi
Technology	Market Sharefl]
Non-Three-Way	1 - 10%
(open-loop)
Three-Way	.7 - 3.7%
(open-looD)
Three-Way	86 - 98%
(closed-loop)
Derived from Tables 2-4 and 2-14. The major differences
between these projections and those in Table 1-4 are that
one additional engine family (AMC EAM360T2FLFX) will
probably remain open-loop three-wav, two additional
families (Chrysler ECR3.7T1BBBO and GM F1G5.7T4HHCX), may
remain open-loop non-three-way, and one family (Chrysler
ECR5.9T43AF9) will probably remain open-loop non-three-way.

-------
2-36
Table 2-16
Fuel Economy Effect of
Various LDDV NOx Standards
NOx	EPA City	Change
Model Year	Standard	MPG	in MPG
General Motors 5.7L Engine Family - Federal
1980	2.0	19.90
1981-82	1.5	19.85	-0.05 (0.3%)
Mercedes-Benz 3.PL Engine Family
Non-Turbo - Federal
1980	2.0	23.0
1981	1.5	23.3	+0.3 (+1.3%)
Turbo - Federal
1980	2.0	24.4
1981-82	1.5	25.2	+0.8 (+3.3%)
Turbo - California
1981-82	1.5	25.71
1984	1.0	25.54	-0.17 (-.7%)
General Motors &.3L Engine Family - California
1982-83	1.5	24.14
1984	1.0	23.98	-0.16 (-.7%)

-------
2-37
The primary effect of the reduced EGR rates would be lower
engine-out particulate emissions. This is important, because
the 1987 particulate standard of 0.26 g/mi does not appear to
be feasible without the use of trap-oxidizers on some
vehicles. Given that particulate emissions averaging is
allowed, any increase (or decrease) in engine-out particulate
levels will increase (or decrease) the percentage of vehicles
requiring trap-oxidizers to meet the particulate standard.
While the full cost of meeting the 0.26 g/mi NOx standard was
considered when the particulate standard was implemented, it is
still appropriate to consider any incremental savings here of a
1.7 g/mi NOx standard since it is the direct result of the
decision to propose a 1.7 g/mi NOx standard for heavier LDDTs.
The effect of the NOx standard on the percentage of
trap-equipped LDDTs has been analyzed elsewhere.[ 2] Under a
1.2 g/mi NOx standard for all LDDTs, 41-56 percent of all LDDTs
would require trap-oxidizers. With a 1.7 g/mi NOx standard for
heavier LDDTs, roughly 16-24 percent of all LDDTs would require
trap-oxidizers, a difference of 32-36 percent.
E. Summary: 1987 NOx Standards for LDTs
A 1.2 g/mi NOx standard appears feasible for all LDGTs and
lighter LDDTs; some heavier LDDTs may experience some
difficulty but the 1.2 g/mi standard should be feasible for
these vehicles as well. Overall, a 1.2 g/mi NOx standard
should result in a net fuel economy improvement for the LDT
fleet, primarily due to the increased use of electronics.
For reasons of reduced engine-out particulate emissions
and market equity, a 1.7 g/mi NOx standard is being considered
for heavier LDTs. This standard is quite feasible for both
heavier LDGTs and LDDTs.
III. Heavy-Duty Gasoline Engines (HDGEs)
A. Introduction
Unlike the majority of light-duty trucks, heavy-duty
engines (HDEs) as a class, especially diesels, will encounter
substantial technological difficulty in meeting the statutory
NOx standard which represents a 75 percent reduction from
uncontrolled levels, i.e., 1.7 g/BHP-hr. As is evident for the
discussion in the following section (Section IV), no proven
technology is available today, nor is expected to be available
by 1990 which will allow heavy-duty diesel engines (HDDEs) to
comply with a 1.7 g/BHP-hr standard. On the other hand, it is
conceivable that heavy-duty gasoline engines (HDGEs) could meet
a 1.7 g/BHP-hr standard using three-way catalyst technology.

-------
2-38
(This possibility was extensively discussed in support
documents to the January 19, 1981 ANPRM.)[15] However, if a
single NOx standard for all HDEs is promulgated, then the
ultimate determinant of NOx standard stringency in the near
future must be the heavy-duty diesel engine.
This section will address the feasibility of both the 1987
6.0 g/BHP-hr and the 1990 4.0 g/BHP-hr emission standards for
HDGEs. Current HDGE emission levels will be discussed as part
of the analysis of the 1987 standard, as will the effects of
leadtime constraints and available emission control
technologies. The analysis for the 1990 standard will consider
the likelihood of new and more refined emission control
technologies. In both cases, fuel economy and performance
implications of the standards will be considered.
3. Feasibility of a 1987 NOx Standard of 6.0 g/BHP-hr
In 1987, there will be two classes of HDGEs:
catalyst-equipped and non-catalyst engines. Approximately 70
percent of HDGEs, those intended for use in most Class lib and
III vehicles, will be equipped with oxidation catalysts and
will be complying with emission standards of 1.1 g/BHP-hr HC,
14.4 g/BHP-hr CO, and 10.6 g/BHP-hr NOx.* The remaining HDGEs
will be complying with non-catalyst standards of 1.9 HC, 37 CO,
and 10.6 NOx.**
All 1987 HDGE enqines will likely be equipped with large
dual air pumps for HC and CO control. Carburetor calibrations
will be somewhat leaner than current technology engines at both
part- and wide-open throttle. Combustion and exhaust system
temperatures will be higher than those of current technology
engines because of these leaner air/fuel ratios, and also
because of increased post-combustion oxidation of partially
burned fuel. [16] About 70 percent of 1987 HDGEs will be
equipped with oxidation catalysts. Most 1985 and later model
year HDGEs will also be equipped with at least some EGR.
(Manufacturers have reportedly chosen to do this for two
reasons: to standardize vehicles sold under different Federal
Based upon the MVMA HDGE transient test cycle.
Also based upon the MVMA cycle. Some non-catalyst engines
may also be certified on the EPA cycle prior to 1987, but
for certification to the 1987 NOx standard, only an MVMA
cycle-based standard will be provided, carryover
notwithstanding.

-------
2-39
and California NOx standards, and also in anticipation of a
future Federal NOx standard.)[ 17]*
As a first step in evaluating the feasibility of a 6.0
g/BHP-hr standard for 1987, the low mileage emission targets
must be derived to account for both deterioration and
production variability.
To account for deterioration, two very important points
must be considered. First of all, for HDGE engines to-date,
NOx deterioration has been reported to be typically minimal or
non-existent. For example, 13 of 16 HDGE families certified in
1983 had additive NOx deterioration factors of 0.00; these 13
families include seven of the eight families certified with
EGR. Therefore, for this analysis it will be assumed that an
additive NOx deterioration factor of 0.00 will adequately
characterize both the average oxidation catalyst and
non-oxidation catalyst engine. Secondly, unlike the
methodology used for LDTs, oxidation catalyst-equipped HDGEs
will not be considered to have multiplicative NOx DFs for
purposes of this analysis. Although contrary to standard
methodology (i.e., vehicles with after treatment devices use
multiplicative DFs and those without use additive DFs), this
approach is consistent with engineering logic in the sense that
oxidation catalysts (the aftertreatment device present) do not
directly affect NOx emissions or their deterioration. The
assumption also permits a single low-mileaqe tarqet level to be
derived for all HDGEs, rather than forcinq the derivation of
different tarqet levels for engine families which will be
equipped with catalysts.
Some NOx control is probably already in effect on current
and prototype HDGEs. This is likely not so much because
of the presence of EGR (manufacturers report that EGR flow
rates for 1985 Federal calibrations will be minimal) , but
primarily because NOx levels have remained essentially
constant despite the increased control of other
pollutants. In particular, over the last twelve years
HDGE HC and CO emission levels have decreased dramatically
from uncontrolled levels, whereas NOx emissions have
remained essentially constant (see Table 2-17) . These
dramatic reductions in HDGE HC and CO emissions have to
some extent been achieved via leaner air/fuel ratios, a
strateqy which typically increases NOx (see Figure 2-5).
(Leaner air/fuel mixtures promote higher maximum
combustion temperatures, thereby promoting increased NOx
formation.) Therefore, one would have expected HDGE NOx
emissions to have increased somewhat as more stringent HC
and CO standards went into effect. The fact that NOx
emissions have remained essentially constant suggests that
some degree of NOx control is already in effect.

-------
2-40
Table 2-17
Heavy-Duty Gasoline Engine NOx Emission Trends
Emission Levels
(EPA transient cvcle-
based g/BHP-hr)
HDGE Model Year Technoloay
NOx
HC
CO
1969 (uncontrolled)*
6.1
12 . 7
155
1972-73 (uncontrolled NOx)*
6.9
6.8
116
1979 (some HC/CO control)*
5.9
3.0
88
1985-87 Prototype
6.0***
1.7***
26
Applicable Emission
Standards (g/BHP-hr)
(EPA transient cycle-
based q/BHP-hr)
NOx
HC
CO
~ ~ ~
10.7
	None-

27 5 ppm
1.5%
1.5**
25**
2.5
40
(significant non-catalyst
HC/CO control)
* Sales-weiqhted averaqe.
** Based upon the 9-mode test. The NOx standard was actually
an HC + NOx standard; the range 8.5 to 10 g/BHP-hr
represents the possible range of NOx emissions for
complying engines.
*** Arithmetic mean of available data from Table 2-18.

-------
2-41
Figure 2-5
2500 R PM, 4 Hg Voc , 40° BTDC
mass specTROMcrtu oat*
	ViC MM TIMM
92SS i*2U	itt&c)
2900 «0
1500 i« 0
990 if 0
40
SO
550 R P M , I9"hc.,60 STDC
•4	I '
W IS W
AIA-FUCL RATIO
Effect of air-fuel ratio on exhaust NO concentrations for
various speed-load combinations. (From Reference 12.)

-------
2-42
With respect to production variability, no elaborate
statistical derivation will be attempted here, primarily
because reasonable estimates have been provided in the past.
In response to earlier rulemakings,[18] Ford estimated
non-catalyst HDGE 40 percent AOL factors of 1.2 for HC and 1.3
for CO, whereas GM estimated factors of 1.1 for all emissions.
For purposes of this analysis, a mean value of these three
estimates will be used (i.e., a NOx AOL factor of 1.2). This
assumption presupposes that NOx emission variability from one
enqine to the next will be no larger than that of non-catalyst
HC and CO. (In fact, NOx variability should be less than that
of HC and CO because of the greater sensitivity of HC and CO
emissions in gasoline-fueled engines to variations in
calibration parameters.)
Using these estimates for deterioration and production
variability, a low-mileage target of 5.0 g/BHP-hr for a 6.0
standard is derived. This target agrees closely with the
estimate provided by Ford (4.9 q/BHP-hr) in their comments to
the January 19, 1981 ANPRM. As will be seen, a precise
estimate of the target level is not critical to the outcome of
the analysis.
The second step in feasibility determination is to
consider leadtime. In 1987, non-catalyst HDGEs will have last
undergone certification to new standards in the 1985 model
year.- Two years of development time are thus available before
recertificat ion to a more stringent 1987 NOx standard, if
recer t i f ication is necessary at all (.i.e., carryover of
low-emitting non-catalyst engine families may be possible).
This two-year minimum time does not include development efforts
undertaken prior to 1985; NOx control-related development
activities actually began in 1981 in response to the ANPRM.
Also, the combustion process in gasoline engines is well
understood, and a wealth of relatable experience is available
from LDV research.
However, NOx standards for HDGEs cannot be considered
independently from other emission requirements. Substantial
industry development efforts are already underway prior to 1987
as HDGE manufacturers prepare to equip the majority of their
engines with oxidation catalysts to meet 1987 HC and CO
standards. Promulgation of a new NOx standard for 1987 will
require that NOx development work be added to this ongoing
effort, especially that which requires interactive calibrations
for HC, CO, and NOx control in the new catalyst-equipped engine
families.
Obviously, some development testing has already been
performed (as noted in the manufacturers' public comments to

-------
2-43
the January 19, 1981 ANPRM) , and further development testing
will no doubt be initiated during the course of this
rulemaking, as well as that which will occur as a matter of
course during the development of catalyst technology. However,
the substantial ongoing development efforts and the leadtime
situation for 1987 HC and CO control will limit the level of
development effort, and hence the degree of NOx reductions,
which can be accomplished by 1987 for HDGEs. Thus, somewhat
less than the equivalent of two years leadtime is available for
NOx control development. Because of this, it is unreasonable
to expect major engine or hardware changes to be available for
production within this timeframe. The analysis below takes
this finding into account in assessing the degree of feasible
NOx reductions.
The final step in feasibility determination, given the
factors discussed above, is to evaluate the feasibility of the
6.0 g/BHP-hr standard if only easily achievable emission
reductions from current levels are assumed feasible by 1987.
The latest available emission data for 1985 prototype
HDGEs are presented in Table 2-18. Four of these eleven
families already meet the target level associated with a 6.0
g/BHP-hr standard. As presented in Table 2-19, the remaining
families will require reductions in NOx emissions from 7 to 34
percent; for the "average" engine family, a 15 percent
reduction will be necessary. In short, to meet a 1987 6.0
g/BHP-hr NOx standard, only modest NOx reductions are required
in only two-thirds of the fleet.
There are several proven emission control strategies
available to reduce NOx emissions in the fraction of engine
families which require reductions to meet a 6.0 g/BHP-hr
standard in 1987. These emission control strategies, discussed
at greater length below, do not involve the addition of new
hardware or technologies to the engine. Rather, they involve
optimization and recalibration of existing components.
Specifically, they include optimization of ignition timing, EGR
rates, and air/fuel ratio calibrations. In all cases, these
strategies have already been applied to LDV, LDT, and HDG
engines, and a large base of technical experience is available
to guide development work. In fact, as illustrated in Table
2-20, Ford quickly achieved NOx levels below the target levels
for the 1987 standard in their early development work in 1981.
For non-catalyst engine families, strategies as simple as
rich air/fuel calibrations may be used for NOx control. Rich
mixtures promote cooler combustion, thereby reducing NOx
emissions. At the same time, the resulting increased
engine-out HC and CO emissions must be afterburned in the

-------
2-44
Table 2-18
1985-87 Non-Catalyst HDGE
Prototype Emission Levels[17]
Enqine Familv	NOx*	HC*	CO*
Ford 4.9L-1V	6.6	(6.5)	1.7	(1.2)	28	(25)
Ford 6.1L-2V	7.4	(7.3)	1.6	(1.1)	27	(23)
Ford 6.1L-4V	5.5	(5.4)	2.3	(1.7)	32	(29)
Ford 7.0L-4V	7.7	(7.6)	2.1	(1.5)	21	(18)
Ford 7.5L-4V	6.5	(6.4)	1.9	(1.4)	27	(24)
GM 292	6.8	(6.7)	2.2	(1.6)	25	(22)
GM 350-2V	6.1	(6.0)	1.6	(1.1)	28	(25)
GM 350-4V	5.1	(5.0)	2.0	(1.5)	27	(24)
GM 366	3.6	(3.4)	0.8	(0.4)	18	(15)
GM 454	4.3	(4.2)	1.0	(0.6)	22	(19)
IHC 345**	4^	(4.6)	2 .0	(1. 5)	37	(29)
Arithmetic Mean:	6.0	(5.9)	1.7	(1.2)	26	(22)
Emissions are expressed in q/BHP-hr, EPA cycle based.
Numbers in parentheses are MVMA cycle based. The derived
correlation between test cycles is:
HC: MVMA = .886 (EPA) - .318,
CO: MVMA = 1.03 (EPA) - 4.04, and
NOx: MVMA = 1.01 (EPA) - .193, as derived in Reference B.
These equations were used where emission results were
available only in terms of one test cycle.
1980 California production version.

-------
2-45
Table 2-19
Low Mileage NOx Emission Reductions From
1985-87 Prototype Levels** for Various NOx Standards
1985-87	Potential NOx Standards**, cj/BHP-hr (associated
Prototype	target levels* are presented in parentheses)
Engine Family
NOx Level,**
g/BHP-hr
3.0
(2.5)
4.0
(3.3)
5.0
(4.2)
6.0
(5.0)
7.0
(5.8)
8.0
(6.7)
9.0
(7.5)
10
(8
.7
.9)
Ford 4.9L-IV
6.5
-62
-49
-35
-23
-11
0
0

0
Ford 6.1L-2V
7.3
-66
-55
-42
-32
-21
-8
0

0
Ford 6.1L-4V
5.4
-54
-39
-22
-7
0
0
0

n
Ford 7.0L-4V
7.6
-67
-57
-45
-34
-24
-12
-1

0
Ford 7.5L-4V
6.4
-61
-48
-34
-22
-9
0
0

0
GM 292
6.7
-63
-51
-37
-25
-13
0
0

0
GM 350-2V
6.0
-58
-45
-30
-17
-3
0
0

0
GM 350-4V
5.0
-50
-34
-16
0
0
0
0

0
GM 366
3.4
-26
-3
0
0
0
0
n

0
GM 454
4.2
-40
-21
0
0
0
0
0

0
IHC 345***
4.6
-46
-28
-9
0
0
0
_0

_0
Arithmetic Mean
: 5.9
-55
-39
-25
-15
-7
-2
0

0
* Assuminq an additive DF of 0.00, and an AOL factor of 1.2.
** MVMA cycle based, from Table 2-18.
*** 1980 California production version.

-------
2-46
Table 2-20
Ford's Best-Effort HDGE
NOx Development Data in 1981 [17]**
Ford	Emissions (q/BHP-hr)*
Engine Family	HC	CO	MOx
6.1L-4V	3.15**	37.1	4.60
7.0L-4V	1.0	28.7	4.85
7.SL-4V	1.31	23.8	4.41
Non-catalyst, MVMA cycle based.
Mote that this data was submitted before substantial
improvements in the HC and CO performance of this engine
was achieved during 1982 and 1983 (see Table 2-19).

-------
2-47
exhaust system. Increased air injection and stainless steel
exhaust systems, as well as vehicle chassis designs to preclude
overheating problems, are necessary. The drawback associated
with this strateqy, however, is an increase in fuel consumption
associated with rich mixture calibrations. This strategy is
nevertheless being used today in California for IHC's 345-CID
family, data from which are presented in Tables 2-18 and 2-19.
Note that this 1980 model year engine already meets MVMA
cycle-based non-catalyst emission standards of 1.9 g/BHP-hr HC,
37.1 g/BHP-hr CO, as well as 6.0 g/BHP-hr NOx.
At this point, it should be stated that the presence of an
oxidation catalyst should have little impact on the feasibility
of a given NOx standard. With respect to NOx control, the
emission control strategies noted below will be equally
applicable to both non-catalyst and catalyst-equipped engines.
In fact, greater calibration flexibility is provided with
oxidation catalysts, because EGR or other calibration settings
which may increase engine-out HC emissions are far less prone
to increasing catalyst-out HC emissions. (Heavy-duty catalysts
will be sized for CO control, and will contain a large reserve
capacity for non-cold start HC oxidation.) Other HC and CO
control techniques, notably air injection and cold-start
management, will also have^ small effects on NOx emissions. For
these reasons, all 1987 HDGEs will be discussed as a single
group, as far as NOx control is concerned, for the remainder of
this analysis.
For all engine families, the retarding of ignition timing
will reduce NOx emissions. Retarded ignition timing decreases
combustion temperatures (thereby decreasing NOx formation), and
also increases post-combustion temperatures, post-combustion
oxidation of HC and CO (presuming sufficient air injection) ,
but can increase fuel consumption if sufficiently retarded.
Thus, the most likely strategy applicable to all 1987
HDGEs is the use of increased EGR. EGR reduces the maximum
combustion temperatures in the engine cylinders, primarily by
introducing a higher heat capacity fluid into the combustion
process. (This fluid can then absorb more heat energy per unit
temperature rise, thereby reducing maximum combustion
temperatures.) EGR is extensively used in internal combustion
engines, most notably current HDGEs. As noted above, current
Federal HDGEs are calibrated to very low EGR flow rates, while
in California — to meet a 1985 5.1 g/BHP-hr transient NOx
standard--identical hardware but greater EGR flow rates will
probably be necessary. The most likely change to current
Federal HDGE EGR systems will simply be an increase in EGR flow
rates.

-------
2-48
At this time, there is ample evidence that achieving a 6.0
g/BHP-hr NOx standard for HDGEs by 1987 is easily feasible with
modest hardware and calibration changes. This judgment was
confirmed by Ford and GM in their comments to the January 19,
1981 ANPRM. Ford commented that ..."a [1986] NOx standard of
about 6.0 g/BHP-hr may be achievable without experiencing
unacceptable tradeoffs."* General Motors claimed that "a
[1986] NOx standard of 7 g/BHP-hr may be feasible with no
increase in fuel consumption".** GM also claimed that
standards more stringent than 7 g/BHP-hr are feasible, but
result in increased fuel consumption (see Figure 2-6). For
example, a NOx standard of 5.1 g/BHP-hr was projected by GM to
cause a two percent increase in fuel consumption.**
Extrapolating from GM's submission, GM would expect about a one
percent increase in fuel consumption to arise from a 6.0
standard. Ford's 1981 development work (Table 2-20), in which
NOx levels well below 1987 target levels were readily achieved,
further supports this judgment of feasibility. Again, there is
ample evidence that a 6.0 g/BHP-hr NOx standard can be met by
HDGEs with existing technology.
Furthermore, at the 6.0 g/BHP-hr level, it is not likely
that the use of existing technology will measurably impact
fleet-wide fuel economy, or preclude certification with other
emission standards. The manufacturers' own predictions for
1986 indicated maximum fuel economy penalties of only zero to
one percent. Their comments were based upon data collected in
1981, and could not include the substantial progress made by
1983 in reducing engine-out HC and CO emissions from HDGEs.
Therefore, the manufacturers' 1981 assumptions with respect to
the feasibility of a NOx standard at given levels of HC and CO,
as indicated by the latest available data (Table 2-18) , have
proven to be overly pessimistic. In addition, the presence of
a heavy-duty oxidation catalyst would more than eliminate HC
tradeoff problems for the majority of engines in the very
unlikely event that problems would be seen at a 6.0 NOx level.
Finally, if the NOx/brake-specific fuel consumption (BSFC)
tradeoff curves presented in Figures 2-1 and 2-2 are at all
representative of current Federal HDGEs, which are now
calibrated with minimal EGR, then large percentage reductions
in NOx emissions should be possible with little or no negative
fuel economoy effect. As presented in Table 2-19, only minor/
NOx reductions will in fact be required.
Ford assumed half-life non-catalyst standards of 3.3
g/BHP-hr HC and 42 g/BHP-hr CO, no idle CO test, and an
implementation date of 1986.
GM assumed an HC standard no more stringent than 2.9
g/BHP-hr in 1986.

-------
2-49
Figure 2-6
HDG Engine Emission/Economy Comparison
1984 350 V3 Prototype
Transient Test Cycle

. L I ' | s i 1. (.
3 -j ¦ 1-1
i ill1 * 1 i ;
TT'"i]: 1 :.r;;; \ii
< i I !
•[IfliuiXTUxp^
' i "t I t 1 I I )'- ; ' I l ¦)
! I 1-1 ! ' ! ; I (Uii-LLj.Et i.
r - -
'' j | 1	1 ' I I L ' 1 l ' 	L— - _lJ * ¦ _
LJ*.IITI'_!	J LJ llliZllXlSZ

1^?
	 _i._

- i . < I •
i •
4i

51,1	6. ;
1 i ¦-r ; it-
- 1 ; r- r r ---gs^riOx	"r:"~ '"IITZ"
i " ~ r r r * * i r* i ~! r ' ¦ i "> ¦ , ' i ! * - * i r* •" r i i ~ *
""" " : "¦ r!-i rr.i ;; i! J".~" i t: t! z:.:
1- - •;
i
: ; -BCGlY
;. /OR
¦4.0 5TD
BOG£7 ' j , ' 1 BOGl't . .
! for-¦ ¦ for
5/11 5TD 1 ' • 7.0 5TD '
GSNEHAL ^.Or
FE-29 54
9/13/32

-------
2-50
In summary, a 6.0 g/BHP-hr NOx standard for 1987 HDGEs is
easily feasible. As noted in Table 2-19, a standard of 6.0
g/BHP-hr would require modest reductions (i.e., less than
one-third) in about two-thirds of the 1985 prototype engine
fleet; the other one-third of the engine fleet are already in
compliance. EPA notes that compliance with a transient 5.1
g/BHP-hr NOx standard is required in California beginning in
1985, and that both future Federal and California engines will
probably be equipped with identical EGR hardware but different
calibrations. (Actual certfication data from 1985 California
HDGEs will be available for review in the course of this
rulemaking, at which time a precise quantification of tradeoffs
can be made.) EPA also notes that Ford was readily able to
achieve NOx emission levels well below the 1987 target level in
their 1981 development work. Given the specific experience
which will be gained with these heavy-duty engine families,
given the decade of existing experience with NOx control in
light-duty vehicles, given the relatively modest reductions
necessary for a fraction of the fleet, and qiven the
availability of well understood NOx control technologies for
qasoline-fueled engines, a 1987 NOx standard of 6.0 g/BHP-hr is
easily feasible for HDGEs within leadtime constraints, and
should result in no fuel economy, performance, or driveability
penalties.
C. Feasibility of a 1990 NOx Standard of 4.0 g/BHP-hr
A 4.0 g/BHP-hr NOx standard will not be stringent enough
to require the adoption of three-way catalysts. This was not
disputed by any manufacturer in comments to EPA's January 19,
1981 ANPRM, and will be readily apparent from the discussion
which follows. Therefore, 1990 HDGE NOx control technologies
should exhibit deterioration and production emissions
variability similar to those of 1987. Therefore, using the
earlier estimates for deterioration and production variability,
a target level of 3.3 g/BHP-hr is derived as the low-mileage
emission level needed to comply with a 4.0 g/BHP-hr NOx
standard. (This is identical to Ford's projected target for a
4.0 NOx standard, as noted in Ford's comments to the ANPRM.
GM, however, projected that a target of approximately 3.0
g/BHP-hr would be necessary.)
The emission reductions necessary to comply with a 4.0
g/BHP-hr standard are presented in Table 2-19. One 1985
prototype engine appears very close to the 3.3 a/BHP-hr level.
On the other hand, the majority of engines would still require
reductions of 21 to 57 percent to meet the 1990 standard (based
upon prototype data available to EPA in 1983).

-------
2-51
To achieve the required emission control, the simple
techniques used in 1987 may not be sufficient in themselves to
yield NOx levels approaching 3.3 g/BHP-hr in the majority of
engines without the risk of significant penalties on engine
performance and fuel consumption. There is evidence that
non-catalyst HDGE NOx emissions can be further reduced by
increased EGR. Most of the manufacturers' initial data,
collected in 1981 in preliminary characterization projects,[19)
indicate that HDGE engine-out NOx emissions in the 3-4 g/BHP-hr
range can be achieved. In most cases, the achievement of these
low NOx emission levels also resulted in adverse tradeoffs of
fuel consumption, power, driveability, and other emissions.
However, for example, Ford increased the exhaust system
backpressure on a test engine as a quick and easy method of
inducing increased EGR. While NOx emissions were reduced from
5.1 to 3.8 g/HP-hr, the increased backpressure decreased engine
power and increased fuel consumption. As these observed
adverse tradeoffs were largely attributable to the method bv
which EGR was induced and there are far less punitive methods
of inducing EGR into production enqines, these data cannot be
used to quantify these effects.
However, as can be seen in Figure 2-2, too much EGR in any
given engine can increase fuel consumption. Other
manufacturers' test results also indicate that too much EGR can
also increase other emissions, and impair driveability and
maximum power. For example, some of GM's early test data are
reproduced in Figure 2-3, in which NOx emissions in the 3.0
a/BHP-hr range were achieved, but with substantially increased
HC. Thus, the preliminary test results submitted by the
manufacturers in 1981 illustrate the limitations of simply
increasing EGR rates and changing enqine calibrations without
performing compensatory engine development.
In their comments to the January 19, 1981 ANPRM (which
proposed a 1986 NOx standard of 4.0 g/BHP-hr), Ford stated
which technologies it believed held promise to reduce NOx
emissions below 6.0 g/BHP-hr. These technologies included burn
rate improvements, other combustion chamber revisions
(including reductions in surface to volume ratio and
compression ratio), camshaft revisions, friction reductions,
EGR system improvements (including electronic controls and an
EGR pump for hiqh engine loads), thermactor air improvements,
adoption of fuel injection, and adoption of electronically
controlled high energy ignition. Ford arqued that the
interactive nature of these strategies, and also the need to
maintain engine power and durability, necessitated detailed
design and development programs with adeauate time for
dynamometer and vehicle demonstration. Ford concluded its
comments with the followina statement: "The leadtime to

-------
2-52
research, design, develop, and implement these advanced HDGE
emission technologies is also an open issue, but it is Ford's
judgment that production feasibility prior to the early 1990s
is problematical and unlikely."
Ford's comments are consistent with experience; all of the
above technologies indeed hold promise for HDGE NOx control.
Before a judgment can be made, however, whether all or only a
few would be required to allow compliance with 4.0 q/BFP-hr
standard, the task of compliance needs to be more carefully
examined, as does the emission control potential of the most
widely adopted non-catalyst NOx control technology for
gasoline-fueled engines--EGR.
The transient test for HDGEs is still relatively new.
While NOx emissions are not at all as sensitive to the specific
test cycle as other emissions,* there are variations in
brake-specific emission rates between specific modes of the
test cycle. For this reason, a thorough characterization
(e.g., emissions "mapping") of engines is desirable, and
perhaps essential if effective emission control strategies are
to be developed for a qiven engine on a given test cycle.
To some extent, this has already been done for liqht-duty
gasoline-fueled trucks, and also to a limited extent in current
HDGEs. Many production EGR systems are controlled as a
function of engine backpressure. As backpressure increases
(i.e., the amount of combustion increases as measured by
exhaust flow) , the EGR rate is also increased to maintain EGR
at a relatively constant percentage of total intake air. At
wide-open throttle (as sensed by throttle position or manifold
vacuum), the EGR is reduced; to some extent this may be
desirable to maintain maximum power, but EGR is also not as
necessary at WOT because the richer air/fuel ratios typically
observed at WOT automatically reduce brake-specific NOx
emission rates (see Figure 2-5). This type of EGR system is
typically mechanically controlled, and is far removed in
complexity and capability from the electronically controlled
EGR systems present on current LDGVs, in which many other
engine parameters (rpm, temperature, rate of change of load,
etc.) can be used for precise mode-specific EGR control. As a
first step, however, engines need to be carefully characterized
to permit the optimization of emission control strategy,
including the minimization of adverse tradeoffs.
See Chapter C of Reference 16.

-------
2-53
It is not likely that the manufacturers' early test data
on HDGE EGR, gathered in 1981, represented an elaborate
matching of EGR systems to engines. They represented very
quick characterization programs conducted on existing engines
with little design optimization. Nevertheless, both the
manufacturers' test data and other research (presented in
Figure 2-1) indicate that increasing EGR decreases NOx. What
remains is that adverse tradeoffs be minimized (i.e., that EGR
systems be more carefully matched to an engines' operational
requirements) , and also that the engine be made more tolerant
of increased EGP. One example of improved design to limit or
eliminate these tradeoffs will be discussed below.
The tolerance of a spark-ignition engine for EGR can be
dramatically increased by the use of "fast-burn" technology.
Fast-burn technology involves modification of the combustion
chamber to both increase turbulence and to reduce the distance
the flame must travel through the combustion charge. Quite
simply, the combustion event takes place in a shorter period of
time than in other chambers. The use of this technology has
been observed to improve engine thermal efficiency, EGR
tolerance, and combustion stability with lower octane fuels.
This technique is already in limited production in LDV
engines, one of the more well-known examples being the Nissan
NAPS-Z engine. This engine uses...
"...a swirl system for increasing the turbulence of the
mixture in the cylinder and a two-point ignition system
for shortening flame travel. Good driveability can be
secured even with the quantities of EGR utilized by the
adoption of fast-burn combustion... Thus, in the vehicles
with the NAPS-Z engine, low NOx emissions and high fuel
efficiency are ach ieved....In the fast-burn combustion
engine, fuel efficiency improves as the quantity of EGR
increases. The optimum point of fuel efficiency appears
with an EGR of about 20 percent."[201
Figures 2-7 and 2-8 illustrate actual improvements in LDV
engine performance using "fast-burn" techniques. In Figure
2-7, the EGR rate corresponding to minimum fuel consumption for
the NAPS-Z engine was raised from 10 to 20 percent when
fast-burn improvements were adopted; at the same time, brake
specific NOx emissions were reduced by more than 50 percent.
Similarly, Figure 2-8 presents performance data from the new
Ford 2.3L High-Swirl-Combustion (HSC) engine also intended for
use in LDVs. The "Phase I" (fast-burn) configuration produced
more than a 66 percent NOx reduction from the "base"
(non-fast-burn) configuration, while at the same time improving
overall fuel efficiency and eliminating adverse tradeoffs with
HC emissions. This performance is typical of the capabilities
of fast-burn technology.

-------
2-54
Figure 2-7
CO
Q_
\
CJ
x
O
15
10
2000RPM
2.2 KG-M
MBT
A/F 15.0
>
vX
V^Ss. _
"l^ 1 , 1
i I


oo
Q_
a
a
|§ 380
10	20
EGR RATE (%)
Comparison of NAP(s)-Z
Conventional and Fast-Burn
from Reference 20.)
Engine Performance with
Combustion Chamber (adapted

-------
2-55
Figure 2-8
L2L HSC: 1500 RPM, 38 3MEP, 14.6 A/F
B
S
H
C
G
R
/
H
?
H
n
4	6
SSNX GR/MP-HR
2.3L HSC: 1500 RPM, 33 BME?, 14.6 A/F
0	2	4	6	3	10
BSNX GR/HP-HR
Comparison of Ford HSC Engine performance with
Conventional ("base) and Fast-Burn ("Phase I") combustion
chamber (adapted from Reference 21)

-------
2-56
This level of burn rate improvement, however, probably
will riot be necessary to allow compliance with a 4.0 standard.
From current NOx levels, percentage reductions ranginq from 3
to 57 percent, with an average of 39 percent, will be necessary
(see Table 2-19). Again referring to Figure 2-1, to achieve
NOx reductions of this size from levels where little EGR is
already used requires average EGR rates of up to 10-15 percent
for the highest emitting enqines, and somewhat less than 10
percent for the "average" engine. Since modal requirements of
the transient test and in-use operational requirements may
dictate more mode-specific calibrations, both more and less EGR
may be used at different modes. (For example, the rich
air/fuel ratio calibrations at wide-open throttle produce less
brake-specific NOx than other modes, and also are those modes
when maximum available power is demanded. Less EGR may be used
in these modes, to be compensated by more EGR at lighter
loads.) Since sophisticated fast-burn technology can permit
the use of 20 percent EGR while maintaining optimum engine
efficiency (see Figure 2-7), the limits of NOx control for this
technology probably will not be reached for a 4.0 standard. In
fact, burn rate improvements, or any other NOx control
technology, will probably be adopted to whatever extent is
necessary to allow each individual engine family to comply with
the standard. In the worst case, it is possible, given
mode-specific calibrations and the highest emitting enqines in
Table 2-19, that substantial combustion chamber changes may be
required on one or two engine families, and perhaps even
electronically controlled EGR systems. For the most part,
however, given the size of reductions which will be necessary
for the majority of engines, only moderate burn rate
improvements and other less sophisticated modification (see
earlier page) should be necessary for most, assuming careful
matching of EGR systems to engines. Even in the worst case,
all of the technoloqies noted by Ford remain available for
increased NOx control.
In summary, EPA agrees with Ford's list of potential NOx
control technologies for HDGEs. However, most of these
technologies should not be required for compliance with a 4.0
g/BHP-hr NOx standard. The degree of NOx reductions required
from prototype levels (see Tables 2-19 and 2-20) are not very
great (an average of 39 percent for the fleet), especially when
compared to the potential reductions of 50 and 66 percent
achievable with fast-burn technology. Burn rate improvements
to combustion chambers, optimized EGR systems for mode-specific
calibrations, and systematic optimization of the combination of
air/fuel ratio, spark timing, and EGR rates to match the modal
requirements of the transient test should be sufficiently
effective to control NOx to a 4.0 standard.

-------
2-57
As for leadtime, even the "maximum technology" approach
discussed by Ford was speculated to be achievable by perhaps
"the early 1990s"; the adoption and demonstration of the lesser
and more likely level of technology should indeed be feasible
for 1990. Unlike the heterogeneous fleet of HDDEs, HDGEs are
technologically identical in almost all aspects except
displacement. Their inherent similarity, both within their
class and to other gasoline-fueled engines, makes the transfer
of experience and technology possible. As was noted above, the
types of technology required to meet a 4.0 standard have
already appeared in production in LDVs. When compared to the
almost revolutionary technological changes which HDDEs will
undergo between 1987 and 1990, the technological difficulty
associated with HDGEs meeting a 4.0 g/BHP-hr standard appears
small.
Given the available development time, and given the
availability and transferability of a variety of proven
technologies, a 4.0 g/BHP-hr HDGE NOx standard will be feasible
by 1990. Indeed, if HDGEs follow the same trend as LDVs with
the implementation of burn rate improvements, then simultaneous
reductions in NOx and fuel consumption would be expected. Only
in the case of "quick-fix" development efforts, i.e., those in
which little or no compensating engineering changes are made to
HDGEs to improve EGR tolerance, would fuel economy or o-ther
penalties be observed on account o£ this standard.
While the 4.0 g/BHP-hr standard appears feasible via
engine-related changes, it is possible that some manufacturers
would apply three-way catalyst technology. Class lib and III
HDGEs will be equipped with oxidation catalyst in 1987 and
their LDGT counterparts will likely be equipped with
closed-loop, three-way catalyst technology. Thus, the step to
three-way catalyst technology may be a natural one. While this
would likely be more expensive with respect to its effect on
initial vehicle cost, fuel economy and performance should
improve. Overall, a manufacturer should only choose this
option if three-way catalyst technology minimizes its total
costs while meeting fuel economy and performance design goals.
D. Summary
The feasibility of two NOx standards for HDGEs	was
discussed: 6.0 g/BHP-hr in 1987 and 4.0 g/BPP-hr in 1990.	The
former appears feasible via increased use of EGR with no	net
effect on fleetwide fuel economy, performance,	or
driveability. The latter appears feasible via advanced
combustion chamber modifications in addition to increased	EGR
rates. &igain, no net adverse effects on engine operation	are
expected,

-------
2-58
IV. Heavy-Duty Diesel Engines (HDDEs)
A. Introduct ion
There are a couple of factors which make a
straightforward, quantitative assessment of the emission
control capabilities of HDDEs difficult. First of all, many of
the advanced technologies discussed here are currently in their
infancy or have yet to be put in widespread production, which
makes it difficult to quantify their effectiveness. Second,
the technological heterogeneity of engines in the HDDE class
makes it difficult to draw fleet-wide conclusions as to the
viability of a given technology. These factors constrain one's
ability to analytically address future emission control, and
one's ability to confidently determine standards representing
the most stringent control achievable without unacceptable
cost. In short, unlike the earlier analysis for LDTs and HDGEs
where specific production technologies and emission data were
easily reviewed and conclusions drawn in a straightforward
manner, this HDDE analysis must rely to a great extent on
engineering judqment.
On the other hand, as will be reviewed in greater detail
below, a large body of research is currently underway worldwide
to characterize and improve the emissions performance of diesel
engines. Equally important, market pressures are continually
stimulating increased fuel economy in HDDEs. In many cases,
the same technologies which improve fuel economy will also
reduce HDDE emissions.
This analysis will begin with a review of the current NOx
and particulate emission levels for HDDEs. It will then
examine the technological feasibility and other impacts of
HDDEs meeting two sets of proposed NOx and particulate
standards: 1987 standards of 6.0 and 0.60 g/BHP-hr,
respectively, and 1990 standards of 4.0 and 0.25 g/BHP-hr,
respectively. The interaction between the control of these two
pollutants requires that they be addressed together. The one
exception is the discussion of trap technology for 1990, the
trap having no effect on NOx emissions. Thus, the discussions
of the 1990 NOx and particulate standards will be presented
separately. Also, as indicated above, it should be noted that
the outcome of this analysis with respect to the feasibility of
NOx control determines the proposed NOx standards for HDGEs as
we 11.
In addition, the analysis will examine the maximum
feasible NOx and particulate levels achievable in 1990 and
later years without the use of trap-oxidizer technology. The
impact of these standards and control techniques on other

-------
2-59
important areas such as fuel consumption and HC emissions will
also be addressed in each case.
B. Status of Current Technology
This portion of the analysis will examine the NOx and
particulate emission levels of current HDDEs. In addition, it
will also describe some of the characteristics of these engines
which might explain their current NOx and particulate levels.
Table 2-21 is a summary of the HDDE NOx and particulate
transient emission data currently available to EPA. The data
comes from three sources: in-house EPA testing, testing
conducted at the manufacturers' labs, and testing conducted as
part of the EPA/EMA round robin program. When multiple tests
on the same engine were available, the test results were
averaged, except for tests conducted as part of the round robin
program, where only those tests conducted at the engine's home
lab were used.
The engines tested represent approximately 66 percent of
all domestically produced FDDE sales. In most cases, the
engines tested are actual production models. However, in a few
cases the engines are viable production prototypes or
production models with modified calibrations. While some
caution should be used since every engine family sold in the
U.S. is not represented, the engines in Table 2-21 adequately
represent all HDDE technology.
The engines in Table 2-21 are grouped together in three
subclasses, established based on their emission-related design
and durability characteristics. The three subclasses are: 1)
HDDEs of low horsepower and short useful life (which also
happens to be those with indirect injection (IDI)), 2) direct
injection (DI) HDDEs of moderate horsepower and useful life,
and 3) DI HDDEs of high horsepower and lonq useful life. Since
these three subclasses are rouqhly equivalent to the light,
medium, and heavy HDDE subclasses established for useful-life
purposes,[24] these common labels will be used here.
The NOx and particulate emission results shown in Table
2-21 are generally indicative of low mileage NOx and
particulate emission levels of current engines. Thus, these
levels represent a viable starting point for our assessment of
the technological feasibility of future standards. The
emission results from the one prechamber engine family are 3.01
and 0.46 g/BHP-hr NOx and particulate, respectively. (Only one
other light HDDE is currently marketed, the IHC 6.9L diesel.
While no transient emission data are currently available on
this engine, its emissions are expected to be very near those

-------
2-60
Table 2-21
Summary of NOx arid PM Emission Results for Current HDDEs
Model
CID
J?P
Aspiration
Timinq[ll
NOx
Part.
Source
Light HDDE: (all indirect injection)




Detroit Diesel
(DDA)






V8-6.2
378

N[21

3.01
.46
EPA
Medium HDDE: (all direct injection)




Caterpillar (CAT)






3208DIT
636
250
T
16.00
10.00
.59
M
3208DINA
636
210
N
16.00
8.49
.63
RP
3208DINA
636
200
N/EGR
16.00
4.30
.88
M
3208DTNA
636
200
N/EGR
16.00
4.13
1.09
M
Cummins







MH250
855
225
N
.029
8.11
.72
SwRI
NH250
855
225
N
.029
6.80
.83
M
Detroit Diesel
(DDA)






8.2T
500
200
T
4.00
5.99
.43
M
8.2T
500
200
T
6.00
6.80
.43
M
8.2T
500
205
T
6.00
6.88
.44
N
8.2T
500
200
T/EGR
4.00
5.04
.67
M
8.2T
500
200
T/EGR
6.00
5.60
.69
M
8.2T
500
200
T/EGR
3.00
4.17
1.06
M
B.2T
500
200
T/EGR
6.00
4.73
1.42
M
International Harvester
(IH)





DTI466B
466
210
TA
	
4.05
.81
RR
DTI466C
466
210
TA
17.00
4.64
.62
M
DTI466C
466
210
TA
17.00
4.62
.62
M[31
DTI466C
466
210
TA
20.00
5.50
.57
M[3]
DTI466C
466
210
TA
23.00
6.95
.57
M[31
DT466C
466
210
T
20.00
5.50
.56
Mf31
DT466C
466
210
T
23.00
7.05
.53
M[3]
OT466C
466
210
T
26.00
9.65
.43
M[31
[1] Degrees
before
top
dead center,
except
Cummins,
which
is in
inches.
[2] N Naturally aspirated
T - Turbo charged
TA = Turbocharged and after
EGR = Exhaust gas recirculation
[31 Hot start only.
[EPA] rj.s. EP^.
TM] Manufacturers tests.
FSwRI] Southwest Research institute.
[RR] EPA/ENR Roundrobin Test Program.

-------
2-61
Table 2-21 (cont'd)
Summary of NOx and PM Emission Results for Current FTPDEs
Model
CID
Hp
Aspiration
Timing [1]
NOx
Part.
Source
Heavy HDDE
(all direct injection)




Caterpillar
(CAT)






3306DITA
638
270
TA [2]
18.00
9.02
.73
M
3406DITA
893
377
TA
19.50
4.78
.81
RR
3406DITA
893
310
TA
22.00
8.20
.58
M
Cummins







NTCC240
855
240
T
.063
4.78
.77
M
MTC350
855
350
TA
—
9.00
.70
M
NTC350
855
350
TA
.065
7.23
.58
M
NTCC400
855
400
TA
.062
5.31
.85
M
VTR903
903
275
T
.101
5.27
.72
RR
Detroit Diesel (DDA)






6V92TA
552
335
TA
7.00
A. 92
.83
M
6V92TA
552
335
TA
7.00
4.59
.83
M
8V71N
568
262
N
Id. 00
6.31
.69
SwRI
8V71TA
568
350
TA
14.00
7.03
.42
RR
8V92TA
736
435
TA
7.00
8.68
.46
M
8V92^A
736
435
TA
10.00
9.53
.41
M
8V92TA
736
435
TA
12.00
11.65
.34
M
8V92TA
736
335
TA
7.00
5.61
.52
M
8V92TAS
736
335
TA
14.00
9.24
.44
M
Mack
EC6-235
672
235
T
20.00
9.73
.81
M
EM6-250
672
250
TA
23.00
9.11
.40
M
RM6-250R[31
672
250
TA
24.00
8.81
.49
M
EM6-2B5
672
285
TA
22.00
6.97
.61
RP
EM6-300
672
300
T
24.00
8.20
.55
M
EM6-300R[31
672
300
TA
23.00
7.98
.55
M
E6-350[3]
672
350
TA
23.00
8.61
.36
M
[1]	Degrees before top dead center, except Cummins, which is in
inches.
[2]	N = Naturally aspirated
T = Turbo charqed
TA = Turbocharged and after
BGR = Exhaust gas recirculation
[31 "R" refers to reduced rated engine speed.
[EPA] U.S. EPA.
[Ml Manufacturers tests.
[SwRl] Southwest Research Institute.
[RR] EPA/EMA Roundrobin Test Program.

-------
2-62
of the 6.2L GM engine.) The medium HDDEs range in NOx from
4.05 to 10 g/BHP-hr and range in particulate from 0.43-1.09
g/BHP-hr. Similarly, the heavy HDDEs range in NOx from 4.78 to
11.65 and in particulate from 0.34 to 0.85 g/BHP-hr.
Beginning in the 1984 model year, the state of California
implemented optional transient test based emission standards
for HDEs. Included in this change was a NOx standard of 5.1
g/BHP-hr. All five domestic manufacturers have certified at
least one engine family under this option and their
certification levels (full-life) are shown in Table 2-22.
There is no California particulate standard, so the particulate
emission levels of these engines are not available, except
insofar as the California designs of Table 2-22 are represented
by engines shown in Table 2-21, which can be deduced from
similar NOx levels.
Before going on to a discussion of conventional and
advanced NOx and particulate emission control approaches one
final point needs to be made. Even though the HDDEs shown in
Tables 2-21 and 2-22 are representrative of current engines,
the manufacturers' HDDE product offerings are constantly
changing and evolving. This is being driven primarily by
durability and fuel economy copcerns in the marketplace.
Therefore, it is expected that some of the engines shown in
Table 2-21 will not be produced in the future. Those engines
which are less fuel efficient or do not have the durability
demanded by the marketplace will be replaced by engines which
are more fuel efficient, durable and which are designed for
inherently lower emissions. The HDDEs in Table 2-21 represent
only current technology engines, not future technical
capabilities. Even absent more stringent emission standards
there would be a trend to more efficient engines. The
application of turbocharging, aftercooling and electronic
engine controls will all increase, bringing with them improved
emission control as well.
The next three sections of this analysis	turn to
discussions of the technical feasibility of the proposed NOx
and particulate standards for 1987 and 1990. In	the last
section, feasible 1990 non-trap particulate levels	are also
examined.
C. Technological Feasibility of the Proposed	1987 NOx
and PM Standards
The discussion of the feasibility of the proposed 1987 NOx
and particulate standards falls into three sections. The first
section estimates design target levels associated with the
proposed standards. This is necessary because the emissions

-------
2-63
Table 2-22
1984 Model Year HDDE
California NOx Certification Levels--Transient Test
Manufacturer
Catepillar
Model
3406T
3208T
3208E
CID	Aspirat ion
893	T
636	T
636	NA
Deter iorated
NOx Level*
5.02
4.5
5.0
Cummins
Mack
9031
10
12
855
672
672
T
T
4.7
4
4
Inter. Harvester
DT-466
DTI-466
9 . 0L
466
466
551
T
TA
NA
General Motors
8. 2T
8.2
6V-92TA
500
500
552
T
NA
T
4,
5,
4 .
Multiolicative DFs used, q/BHP-hr.

-------
2-64
data in Table 2-21 are low-mileage levels and cannot be
directly compared to a standard. The second section contains
the bulk of the technical feasibility analysis. There the
analysis first addresses IDI engines and then DI engines. The
final section addresses the impact of the proposed standards on
HC emissions and fuel consumption.
1. Establishment of Design Target Levels
The equation for determining the design target level (TL)
is:
= MPL-DF
AQL
Where:
MPL, DF, and AOL are as defined above.
One difference between the design target equation for HDDEs and
that for LDTs is that the DF is expressed as an additive factor
and not as a multiplier, due to the absence of an
after treatment device in the exhaust.
As the technology feasibility analysis which follows is
not able to have the precisions of those for LDTs and HDGEs, it
is not necessary to precisely determine the TL. In fact, as
the various inputs to the above equation do vary from engine to
engine, the ranges indicated below may be more indicative of
actual TLs than a single-point estimate. In addition, the
process of establishing TLs is intended here to simulate how
manufacturers take into account engine variability and
deteriorations. However, for any given engine, a manfacturer
may choose to reduce deterioration variability or the
confidence in passing an audit rather than reduce emissions to
the "standard" target level described here. Thus, it may not
be necessary to reduce emissions to the degree shown here.
For a NOx standard of 6.0 g/BHP-hr, the MPL	would be
6.0499 g/BHP-hr or essentially 6.05 g/BHP-hr.	For a
particulate standard of 0.6 g/BHP-hr, the MPL would	be 0.605
g/BHP-hr.
The equation used to compute the AQL factor has been
developed in past analyses and is:[18]

-------
2-65
AOL = [(1 + z l X 1 y32 + (z.4 X CV)1*
Where:
z = value of the z statistic for alpha (one-sided) = 0.1
(1.28) and alpha = 0.4 (0.25)
CV = within-lab coefficient of variation between engines
for emissions of a given pollutant
The CV for NOx will be assumed to be approximately 10
percent. This is supported by comments made by Cummins and the
results of the EPA/EMA HDDE round-robin test program, which
showed an average within-lab CV of 4 percent for a single
engine. [22 , 23] Using the above equation, the AQL factor is
1.1. for NOx.
The CV for particulate was estimated to be 10 percent by
Cummins, and to be roughly the same as that for HC by
Caterpillar, which is 16 to 20 percent. As the EPA/EMA
round-robin testing indicates that variability for particulate
should be somewhat below that for HC,[23] a range of 10 to 15
percent will be used here, yielding an AOL factor of 1.1-1.15
for particulate.
The next step is the determination of the full-life
deterioration factors. As there is a relationship between the
full-life period and the durability characteristics of an
engine (i.e., long-lived engines are designed more durably and
their emissions increase more slowly), engines with different
full-life periods tend to have similar DFs. Thus, DFs will be
estimated for all HDDEs as a single class here.
The "z" statistic is preferred here over the "t" statistic
since the standard deviation of the population can be
accurately estimated based on past experience. Also,
1.732 is the square root of 3, the number of prototype
engines assumed to be developed.

-------
2-66
A review of the 1983 HDDE certification results shows that
more than one-third of the families had zero DFs, with the
overall mean and median DFs being 0.24 and 0.20 g/BHP-hr,
respectively. Since these values are roughly half life DFs,
mean and median full-life DFs would be approximately 0.48 and
0.40 g/BHP-hr, respectively. Given the frequency of a zero DF
and the lack of historical pressure to control NOx, it will be
assumed that the higher DFs can be reduced to the current mean
DF. This results in an overall range of 0-0.48 g/BH^-hr for
the NOx DF.
No actual particulate durability data are available,
however, it is reasonable to expect that particulate should
deteriorate no more than NOx or FC on a percentage basis. Most
HDDEs certified in 1983 have half-life HC DFs equal to five
percent or less of the low mileage level, and doubling this
would lead to a full-life DF of approximately 10 percent. As a
NOx DF of 0.48 g/BHP-hr is equivalent to about six percent, a
range of 6-10 percent, or 0.04-0.06 q/BHP-hr at the 0.6
a/BHP-hr level, will be used here.
Substituting these fiqures into the expression for the
target level given above, the design target level for a 6.0
g/BHP-hr NOx standard is 5.1-5.5 g/BHP-hr. The particulate
tarqet level is estimated to be 0.47-0.51 q/BHP-hr.
2. Technological Feasibility by HDDE Subclass
As indicated above, HDDEs can be divided into three
subclasses: light, medium, and heavy. While the basic
combustion technology of current light HDDEs is distinct (i.e.,
use of prechamber, or indirect injection (IDI) technology),
that of the medium and heavy HDDEs is more similar. All medium
and heavy HDDEs use direct injection (DI) desiqns.
Furthermore, most medium HDDEs are naturallv aspirated (roughly
60 percent) , while nearly all heavy HDDEs are turbocharged and
aftercooled (roughly 90 percent). However, some medium HDDEs
are both turbocharged and aftercooled (2-3 percent) . As the
techniques available to both medium and heavy HDDEs to reduce
emissions are very similar, they will be treated together below
as DI engines.
a. IDI HDDEs
Two IDI HDDEs are currently produced: a 6.2L engine
produced by Detroit Diesel (DDA) and a 6.9L enqine produced by
International Harvester (IH). The DDA engine is also currently
used in LDTs.

-------
2-67
EPA currently has full transient emission data only on the
DDA engine (see Table 2-21). As can be seen, the 6.2L engine
currently emits 3.01 g/BHP-hr NOx and 0.46 g/BHP-hr
particulate, both of which are well below the target emission
levels of 5.1-5.5 g/BHP-hr NOx and 0.47-0.51 g/BFP-hr
particulate. No transient emission data is available on the IK
6.9L engine. However, the 1983 13-mode NOx certification level
of 4.25 g/BHP-hr indicates that the 6.9L can also meet the NOx
target level in its current configuration, as transient and
13-mode NOx emissions agree fairly well.
Any additional light HDDEs introduced prior to- 1987 should
also be able to comply with both standards, due primarily to
the inherently lower emissions of IDI engines. It is possible
that DI engines could be introduced into the light HDDV class
at some future date. If that occurred, their emissions should
be similar to those of current medium HDDEs, and the control of
their emissions would be similar to that described below.
b. DI HDDEs
Two factors make the discussion of emission control more
difficult for DI engines than for IDI engines. One, DI engines
tend to have inherently higher NOx and particulate emissions
than IDI engines and more control is needed. Switching to IDI
technology is not feasible due to the fuel economy penalty
involved (10-15 percent). Two, a wide variety of DI engines
exist. Some are naturally aspirated, some are turbocharged,
and some are turbocharged and aftercooled. Also, the
combustion chamber and fuel injection technology of each
manufacturer tends to be somewhat unique, each having its
relative advantages and disadvantages. Thus, what reduces
emissions on one engine may not be applicable to another. On
the other hand, some techniques are transferable.
To aid in this examination, the following discussion will
be divided into three parts. The first part will address
non-emission related engine changes expected in the 1987
timeframe that will also lower NOx and particulate. These are
mainly improvements to brake specific fuel consumption (BSFC) .
The second part will address NOx control technology. The last
part will address particulate control. Where data are
available, they will be referenced. However, in many cases,
data are not yet available or the data are still proprietary;
therefore, engineering judgment has to be relied upon.
i. BSFC Improvements
Reductions in BSFC reduce brake-specific emissions
proportionately because emissions tend to remain constant on a

-------
2-68
fuel-specific basis. Lower BSFC	means less fuel burned and
thus, lower absolute emissions.	However, the work generated
over the cycle (BHP-hr) remains	constant and brake-specific
emissions (g/BHP-hr) decrease.
The BSFC of DI engines is expected to improve steadily
over the foreseeable future due to the high mileage of the
vehicles they power and the constant pressure to cut operating
costs. One analyst's recent estimates of BSFC improvements
over the 1982-87 timeframe for medium and heavy HDDEs are shown
in Table 2-23. [24] As can be seen, overall improvements of 7-8
percent are expected. Also, QDA recently " published a history
of recent BSFC improvements on their 6V-92TA showing a 14
percent improvement from 1976 to date and an additional 4-6
percent in the near future. [25] While it is not possible to
independently confirm the projections contained in Table 2-23,
the GM experience and the fact that every major HDD^
manufacturer is known to have programs in place to improve
engine efficiency, argues for lower BSFCs in the future.
The result of this projection is that the emission data of
Tables 2-21 and 2-22, which represent pre-1983 technology,
should be reduced by some amount (e.g., 5-10 percent) prior to
the addition of specific emission controls. This will be
discussed further in the next two sections.
ii. NOx Control
The most common approach currently used to reduce NOx
emissions is static fuel injection retard. It is simple and
inexpensive to implement and its effect on NOx emissions can be
easily characterized. Those DI engines of Tables 2-21 and
2-22, which meet the NOx target level of 5.1-5.5 g/BHP-hr, do
so for the most part with injection retard. These engines
comprise a representative sample of DI HDDEs, as three of the
engines are naturally aspirated, three are turbocharged and six
are turbocharged and aftercooled, and roughly half are medium
HDDEs and half heavy HDDES. Thus, there is little doubt that
all DI engines can meet these levels by simply applying
injection retard. And because of its low cost--it is
essentially an engine adjustment--, it receives the most
attention when NOx controls are discussed. The drawback with
this approach to NOx control is that large amounts of injection
retard can increase particulate emissions and BSFC. The effect
of timing retard on particulate is indicated by the data in
Table 2-21.
Other techniques are available to reduce NOx emissions
without the particulate and BSFC tradeoffs, such as
aftercooling turbocharged engines. Because of the short

-------
2-69
Table 2-23
Projected Improvements	to HDDE Efficiency (%)*[24]
Source of Liqht HDDE	Medium HDDE Heavy HDDE
Improvement 1982-87 1987-92	1982-87 1987-92 1982-87 1987-92
General 2 3	2 3 5	5
Improvements
Direct 2 5	--
Injection
Turbocharging --2	2 1 - -
RPM Reduction -- 2	1.8 1.4 0.9
Lubricants 11	11 11
Aftercooling -- --	0.5 1
2 to 4 Stroke -- --	-- -- --	0.7
Turbocompound -- --	-- -- --	1.5
Reduced HP _~z ZZ	ZZ ZZ 0.3	0.5
5% 13%	7.3% 7.4% 7.2% 8.7%
Fleetwide averages for new engines.

-------
2-70
leadtime available these techniques are unlikely to be applied
to many HDDEs in the 1987 timeframe unless already planned.*
Overall, because of its low cost and easy application, timing
retard appears to be the most likely approach to meeting a 6.0
g/BHP-hr NOx standard.
There are two ways one can examine the effect of timing
retard on particulate emissions. (The effect of timing retard
on BSFC will be examined below in Section 3.b.) The data in
Tables 2-21 and 2-22 can be used in both cases.
One approach is to measure the increase in particulate
emissions on a given engine per unit decrease in NOx
emissions. Restricting the analysis to situations where at
least three timing settings were examined and all of the data
were taken at the same laboratory, one sees that particulate
increases by 0.02-0.04 g/BHP-hr per 1.0 g/BHP-hr reduction in
NOx (or from a slightly different angle, particulate increases
0.26-1.38 percent for each 1.0 percent reduction in NOx).
The other approach is to examine absolute particulate
levels at NOx levels within 5-10 percent of the NOx target
levels, which have been achieved via timing retard. For
example, at these low NOx levels, both the DDA and IH medium
HDDEs are still very near the particulate target, as is the DDA
8V-92TA (335 hp). On the other hand, other engines (e.g.,
Cummins NTCC400 and Mack EC6-235) appear to require up to 50
percent reductions in particulate levels at those NOx levels.
Thus, the degree of particulate control required at NOx levels
of 5.1-5.5 q/BHP-hr varies significantly among engines.
iii. Particulate Control
To date, no particulate standards have been applied to
HDDEs. Thus, while HDDE manufacturers have some experience
related to controlling diesel smoke, they have never been
required to seriously address the control of particulate
emissions from each of their engines. This is particularly
true of particulate formed during transient engine operation,
since the focus on transient emissions only began with the
development of the EPA transient engine cycles in the late
1970's; the exception again being the modest control of smoke
under three specific transient conditions. Thus, one would
expect that some reduction in the current level of diesel
Variable injection timinq via mechanical or electronic
controls can reduce NOx with reduced tradeoffs and may be
able to be introduced by 1987. However, it will be
discussed as a particulate control technique below.

-------
2-71
particulate could be achieved with relative ease via the
standard application of the techniques which have already led
to low particulate emissions on some current engines. However,
due to the heterogeneity of DI HDDEs, it would be inappropriate
to assume that the straightforward transfer of these techniques
could reduce the particulate levels of all engines to those of
the lowest emitting engines of Table 2-21- At the same time,
such levels should be approachable within subclasses of 01
HDDEs (i.e., naturally aspirated, turbocharged, and
turbocharged and aftercooled engines) . As described at the
close of the previous section, there are examples of
turbocharged and turbocharged/aftercooled engines which are
currently at or near both the NOx and particulate design target
levels. Naturally aspirated engines may have a more difficult
time as none currently approach the two design target levels
simultaneously.
In addition to incorporating current design features shown
to produce low particulate levels, there are several, more
generic approaches which can be used to reduce particulate
emission ievels in the 1987 timeframe without increasing NOx
emission levels. The approaches include fuel injection system
modifications, combustion chamber modifications, transient
air-fuel ratio control, and variable injection timing. Fuel
injection system modifications primarily include increased fuel
injection pressure and changes in the fuel injection nozzle
design. Turbocharginq, while promising particulate reductions,
is probably not available for widespread application for 1987
due to leadtime constraints, unless already planned for other
reasons by the manufacturer.
Increased fuel injection pressure has been demonstrated in
a past EPA program to decrease particulate emission levels by
50 percent on a Mack engine [25a] and in the results of
manufacturer experimental programs reported to EPA. However,
high-pressure injection also appears to increase NOx
emissions. The higher injection pressure increases the fuel
injection rate and speeds up the combustion process, much like
injection timing advance. After retarding timing to account
for this, a net particulate decrease should result. while
increasing injection pressure is a substantial modification to
the engine considering the leadtime available, for 1987, a
number of manufacturers are known to have programs in place to
modestly increase fuel injection pressure (20-30 percent).
These modest increases could be expected to provide net
particulate reductions on the order of 5-10 percent for those
manufacturers already planning to do so. It also appears that
increased fuel injection pressure reduces BSFC, oroviding
another incentive for its implementation.

-------
2-72
Modifications to the fuel injection nozzle include zero
sac nozzle volumes to reduce fuel dribble at the end of
injection, damped nozzle checks to prevent secondary injection,
and optimized spray nozzle angle, including the number of spray
holes, and size of holes. In addition to having potential to
reduce carbonaceous particulate, these modifications have been
shown to be effective in the past in controlling HC
emissions,[16,18] which affects particulate as well by reducing
the amount of soluble organics absorbed. The degree of control
achievable will vary from engine to engine and cannot be
precisely quantified here. However, HDDE manufacturers, such
as Caterpillar, have identified such areas for improvement and
expect measurable particulate reductions.[25c]
Combustion chamber modifications include changes in piston
and ring design to reduce dead volume within the combustion
chamber, as well as air inlet port redesign to affect the rate
of swirl and in turn improve air-fuel mixing. Reducing oil
flow past the piston rings can also reduce particulate
emissions, as it has been shown that a portion of the
particulate emitted comes from the oil. For naturally
aspirated engines, the amount of air available in the cylinder
can be increased by adding an additional intake and exhaust
valve per cylinder (a 4-valve engine). *\gain, the net effect
on particulate emissions is difficult to quantify due to
differences between engines.
Preventing overfueling throughout all engine operating
conditions, specifically the prevention of overly low air-fuel
ratios typically occurring during transient engine operation
(e.g., during an acceleration) and more flexible control of
injection timing, probably hold the most promise of any of the
techniques discussed thus far for the 1987 timeframe. This
potential is related to the fact that the engine designs
represented in Tables 2-21 and 2-22 include no improvements
related to optimizing particulate and NOx emissions. 3y
examining instantaneous NOx and particulate emissions over the
transient cycle, periods of high emissions can be identified
and addressed. Particularly, the allowable fuel rate can be
limited and timing adjusted. While full optimization of these
parameters will not occur until the advent of electronics
(which will be discussed below), partial control can be
achieved via mechanical means.
An estimate of the potential for particulate reduction via
better transient air-fuel ratio control can be obtained from
the difference between 13-mode and transient emissions, since
the former includes no transient effects and the latter does.
This difference ranges between 0.1 and 0.3 g/BHP-hr on current
engines. Most of this difference should be controllable via

-------
2-73
improved control of fuel governing so that transient emissions
will eventually match steady-state emissions. By 1987, a
portion of this difference should be controllable by the use of
improved mechanical fuel governors.
An indication of the potential for variable injection
timing can be obtained from analyzing 13-mode NOx and
particulate emissions. In general, NOx emissions on a
brake-specific basis are relatively constant from mode to mode,
while particulate emissions are heavily weghted toward the four
high-power modes (Nos. 5, 6, 8, and 9). This indicates an
ability to retard timing at light loads to reduce NOx without
commensurately increasing particulate emissions. As roughly
half of the transient cycle occurs at such light loads, one
could surmise that roughly half of the 2-4 g/BHP-hr reductions
in NOx emissions occurring via static timing retard in Table
2-21 may be achievable with little increase in particulate. As
particulate emissions otherwise increased 0.05-0.6 g/BHP-hr,
this potential could be quite large.
Mechanical timing controls were described in
manufacturers' comments to the rulemaking. Timing is
controlled by changing the position of a cam follower roller on
the camshaft which is actuated by an air actuator cylinder.
The air cylinder piston motion is based on fuel rail pressure.
Although the widespread use of electronic engine controls
(EECs) is not expected until later in the decade, some limited
application of EECs is expected prior to 1987. For example,
DDA is using electronic control of injection timing, bypass
ratio, and fuel governing on a prototype methanol-fueled HDDE
currently operating in a transit bus in California.[25b] DDA
plans to introduce EEC technology in 1985, so at least some
models will have access to this technology by 1987. While it
is expected that HDDEs should be able to comply with the 0.6
g/BHP-hr particulate standard without electronics, electronics
should be available for those models experiencing difficulty.
To place the need for the modifications discussed above in
some context, the medium HDDEs of Table 2-21 will be examined
in some detail. Overall, the modest emission reductions needed
by the DDA and IH engines should be easily achievable by any
number of the above techniques. The Cat 3208, which
Caterpillar is known to desire to keep in the market as a
naturally aspirated engine, will require most of the
above-mentioned techniques related to the combustion chamber
and fuel injection system. Caterpillar is known to be actively
working on techniques like these for this engine.

-------
2-74
Caterpillar has indicated the types of improvements they
expect to make to their engines and these are shown in Table
2-24. [25c] Caterpillar was the most optimistic of all the HDDE
manufacturers about complying with the particulate and NOx
standards of 0.6 and 6.0 g/BHP-hr, respectively, in the near
term, and went so far as to recommend these levels as feasible
standards for 1986. While it is known that Caterpillar has
found the task somewhat more difficult than it believed at that
time, Caterpillar is still targeting their efforts toward these
levels and expects to reach them.[25c] Thus, while the Cat
3208 may require many of the above-mentioned techniques, it is
expected to be able to comply with the 1987 standards.
The degrees of modification required by heavy HDDEs fall
into similar categories; some, like the DDA and IHC medium
HDDEs, require little modification, while others, probably the
majority, require greater modification like that described for
the Caterpillar engine. In addition, several manufacturers,
including Mack, [22] have indicated that in the short term,
production of one or more models may have to be dropped.
However, it would not be appropriate to attribute this solely
to the emission standards. While more stringent emission
standards do tend to force aside lagging technology, it is also
clear that HDDE models which manufacturers decide to drop are
also likely to be deficient in other aspects such as
durability, fuel economy, and market penetration, and that
manufacturers have decided that the investment required to meet
the standards is not justified by the potential payback in
retaining the model in its product line. A number of HDDE
models are over 20 years old and currently have very low sales
(e.g., below 1,000 engines). The equipment to produce these
engines has been fully depreciated long ago, justifying
continued production despite low sales. These NOx and
particulate emission standards represent the first major
requirement for new investment on these models to come along in
some time. Overall, the lost sales from these engines should
be made up elsewhere in the manufacturer's product line.
3. Other Impacts of the Proposed 1987 NOx and
Particulate Standards
a. Hydrocarbon Emission Levels
It is generally recognized that injection retard tends to
increase HC emission levels. However, techniques which are
effective in controlling particulate emissions tend to reduce
HC emissions. Given that a net decrease in particulate
emissions will occur, we expect no net increase in HC levels
should occur.

-------
2-75
Table 2-24
Modifications Beinq Made to
Caterpillar Development Engines [241
Improved Fuel System
Higher Injection Pressures
Faster Injection Rates
Damped Checks to Eliminate Secondary Infection
Lower Sac Volume Fuel Nozzles
Reduced Crevice Volumes
Improved Turbocharqer Efficiency
Improved Charge Air Cooling

-------
2-76
b. Fuel Economy
There is a well established, adverse tradeoff between NOx
emission control via retarded injection timing and fuel
consumption. However, no fuel consumption increase is expected
for IDI engines since their IDI technology provides compliance
with the proposed NOx standard without significant timing
retard. The potential fuel consumption effect for DI HDDEs is
discussed below.
In their previous comments on the fuel economy impact of
further HDDE NOx control, several manufacturers provided
estimates of the fuel consumption impact of NOx standards of
approximately 6.0 g/BHP-hr. [22) Mack, Cummins, and Caterpillar
estimated fuel consumption penalties of 2-11 percent for NOx
standards of 5-6 g/BHP-hr.
In a follow-up presentation to EPA in October 1983,
Caterpillar estimated that, with NOx and particulate standards
of 6.0 g/BHP-hr and 0.60 g/BHP-hr respectively (those levels
recommended by Caterpillar in the hearing on the initial
proposal of the particulate standard), fuel consumption
penalties at or near the lower end of the above range were
likely.[25c] This is based largely on the results of their
newly released 3406B model which employs many of the emission
control techniques discussed above (California comparisons).
In their 1981 report on NOx control from HDEs,[25d] the
National Academy of Sciences (NAS) estimated fuel consumption
penalties of 2.5 to 4 percent for a 6 g/BHP-hr target level and
7 to 12 percent for a 4 g/BHP-hr target level. By simple
interpolation using a NOx target level of 5.2, the NAS
projections provide fuel consumption penalty estimates of about
4 to 7 percent (the NAS projection was based on 13-mode
relationships and covered only DI HDDEs; IDI HDDEs were not
available when the NAS study was conducted).
Except for the latest Caterpillar estimate, the above
estimates of fuel penalty do not take into account the
particulate control techniques which will be used to comply
with the 0.6 g/BHP-hr standard. Thus, they likely overestimate
the effect of the 6.0 g/BHP-hr NOx standard.
Given this information, a fuel consumption increase of at
most two percent is projected for medium and heavy HDDEs. If
timing retard were the only NOx control technique used, the
full two percent penalty would be expected. The use of other
techniques, such as advanced charge air cooling, electronics
etc., should all reduce this effect, to the extent they are
applied prior to 1990. The application of this advanced NOx

-------
2-77
technology in 1990 (which is discussed below with respect to
the 1990 standards) should eliminate this 1987 impact. Thus,
any impact of the 1987 standards should be temporary and not
long lasting.
D. Feasibility of the Proposed 1990 NOx Standards
1.	Establishment of Design Target Level
As was discussed previously, determination of the tarqet
level requires estimation of the MPL, the AOL factor, and
full-life DF. For a 4.0 g/BHP-hr NOx standard, the MPL is 4.05
q/BHP-hr.
Estimating the AOL factor and the full-life DF is more
complicated since they can vary somewhat with the technologv
used to comply with the emission standard. As will be
discussed below, it is expected that compliance with a 4.0
q/BHP-hr NOx standard (beginning from a 6.0 g/BHP-hr level)
will require application of various combinations of electronic
engine controls (EECs) , advanced charge air coolinq, and
possibly in some cases some additional injection timing retard
and exhaust gas recirculation (EGR).
For the 6.0 g/BHP-hr NOx standard, the AOL factor was
determined to be 1.1 and the full-life DF was 0-0.48 g/BHP-hr.
The net effect of adding the predominant technologies mentioned
above (i.e., electronics, EGR, and improved charqe air cooling)
is difficult to ascertain due to a complete absence of data.
Electronic controls should decrease variability and
deterioration. Improved charge air coolinq should have no
effect on variability or deterioration. However, EGR (if used)
may increase variability and deterioration, since the EGR rate
may vary between engines and with time. while there is some
indication that variability and deterioration may decrease
overall, the effect cannot be quantified. Thus, it will be
assumed that the AOL factor and DF will remain constant at 1.1
and 0-0.48 g/BHP-hr, respectively, with a 4.0 g/BHP-hr
standard. Using these factors yields a NOx tarqet level of
3.2-3.6 g/BHP-hr.
2.	Technological Feasibility by HDDE Subclass
This section will follow the same general outline as that
above for the 1987 HDDE standards. IDI HDDEs will be addressed
first, followed by DI HDDEs. The DI HDDE discussion will be

-------
2-78
divided into two sections, BSFC improvements and NOx controls.
Following all of this will be a discussion of the effect of NOx
control on HC emissions and BSFC.
a. IDI HDDEs
As was discussed above, there are currently only two IDI
HDDE models. Engines with IDI technology qenerally have verv
low NOx emission levels. The limited amount of transient
emission test data available (Table 2-21) indicates that levels
of about 3.0 g/BHP-hr are already achievable. This one piece
of emission test data is not conclusive evidence that all IDI
HDDEs can meet the 3.4 g/BHP-hr NOx target level in their
present configuration. However, it does provide evidence that
IDI HDDEs require at most small reductions to meet a 3.4
g/BHP-hr target level, which should be achievable through minor
steps such as a small amount of injection timing retard or
perhaps the use of electronic engine control (EEC). Thus,
compliance with a 4.0 g/BHP-hr NOx standard appears to be
relatively straightforward for IDI HDDEs.
b. DI HDD^.s
i. BSFC Improvements
As mentioned above in Section C.2.b.ii with respect to the
1987 standards, general fuel efficiency improvements cause
proportional NOx emission reductions by reducing the amount of
fuel consumed over the test cycle. Table 2-23 of the above
mentioned section shows one analyst's projections of engine
fuel consumption improvements in the late eighties and early
nineties, yielding some idea of the magnitude of the values
being discussed. As can be seen from Table 2-23, efficiency
improvements between 1987 and 1992 are expected by EEA to be
7-8 percent for medium and heavy HDDEs. This would represent
roughly one-fifth of the 30-40 percent NOx reduction required
by the 4.0 g/BHP-hr standard. A similar reduction should occur
in particulate emissions as well, as will be mentioned in
Section F below. DDA experience with its 6V-92TA appear to
roughly confirm these projections, at least for two-stroke
engines. [251
It should be noted that use of EEC is not mentioned m
Table 2-23. This technology was not considered because EEA
assumed that it's use would be optimized for NOx control rather
than improved BSFC, which is also the orientation being taken
here.

-------
2-79
ii. NOx Control
To comply with a 4.0 g/BHP-hr NOx standard, manufacturers
of DI HDDEs are expected to use various combinations of the
following technologies: advanced charge air cooling,
electronic engine controls (EECs), and possibly small amounts
of exhaust gas recirculation (SGR) and/or timing retard. These
are discussed below.
NOx emissions are a strong function of peak combustion
temperature, which, in turn, depends in part on the initial
temperature of the combustion chamber air. Turbocharging,
which is used on most DI HDDEs, increases the inlet air
temperature significantly above ambient levels. The actual
temperature increase is a function of the turbocharger design
and operating condition, but a range of 100-200°F is fairly
representative. Charge air cooling reduces the temperature of
the turbocharged air, increasing the mass of air able to be
pumped into the cylinder. Charge air cooling using engine
radiator water as the cooling medium (referred to as connected
circuit) can cool inlet air to about 185-215°F. Similarly,
charge air cooling using a separate radiator system (separate
circuit) can reduce inlet air temperatures to about 140-170°F,
while using ambient air as the cooling medium (air-to-air) can
reduce inlet air temperatures to 110-120°F.
Data published by DDA can be used to estimate the effect
of separate circuit and air-to-air charge air cooling.[25] DDA
data indicate that at 5.0 g/BHP-hr NOx, separate circuit charge
air cooling provides a 2.5 percent fuel economy improvement
over connected circuit charge aircooling at the same NOx
level. Air-to-air charge air cooling provides a six percent
improvement under the same NOx level. While the author did not
indicate the reduction in NOx level at constant fuel economy
(i.e., BSFC) , this reduction can be estimated from the fuel
economy data.
As discussed in Section D.3.b. below, a reduction in NOx
from 6.0 to 4.0 g/BHP-hr via timing retard has been estimated
to cause roughly a 4.5 to 8 percent increase in BSFC*.[25d]
It should be noted that NAS labeled the lower limit of
this BSFC penalty as that indicative of "advanced" NOx
control technology. However, this advanced technology did
not include EEC or improved charge air cooling. It only
represented the lowest BSFC increase derivable from data
available in 1980, which probably meant combustion chamber
and fuel injection designs which minimized the impact of
timing retard on BSFC. Therefore, the entire range is
applicable here.

-------
2-80
Since DDA held NOx levels constant, timing must have been
advanced to increase NOx emissions after the addition of
improved charge air cooling. Retarding timing using the above
NOx/BSFC relationship to the point where BSFC is at its
original level can be used to determine the equivalent NOx
reduction of the charge air cooling at constant BSFC.
For separate circuit charge air cooling, the 2.5 percent
fuel economy increase at constant NOx translates into a 0.5-1
g/BHP-hr NOx reduction at constant BSFC, or an 8-17 percent
reduction in NOx emissions at 6 g/BHP-hr. For air-to-air
charge air cooling, the six percent fuel economy increase is
equivalent to a 1.5-2.6 q/BHP-hr reduction in NOx, or a 25-40
percent reduction at 5 g/BHP-hr NOx. These figures must be
considered rough estimates since the actual NOx levels were not
published (and no other emission data on the effect of such
technology are publicly available) and the NOx/BSFC
relationship resulting from timing retard may be quite
different for the two-stroke 6V-92TA than the group of HDDE
engines examined by NAS. However, they do indicate that
improved charge air coolinq has the potential for providing
much and possibly all of the NOx control required by the 4
g/BHP-hr standard.
EEC technology was briefly discussed above with respect to
the 1987 standards. The various HDDE manufacturers appear to
be at different stages of developing EEC technology. However,
by 1990 all HDDE manufacturers are expected to have EECs on
some or all of their engines.
DDA appears to be the closest to implementing EECs. By
mid-1983, DDA had already accumulated the equivalent of 5.5
million miles on EEC-equipped prototype engines and 75 trucks
will test this technology on the road during 1984. Limited
production is expected to begin in 1985. Among other
parameters, this system precisely controls fuel governing and
injection timing at any engine speed, load, and coolant
temperature.f 2 5]
The data published by DDA again gives some indication of
the NOx control potential available with EEC technology.[ 25]
There, EEC technology is shown to improve BSFC by eight percent
on a connected circuit cooled engine at 6.0 g/BHP-hr NOx
relative to a mechanically controlled engine. Aqain using the
relationship between BSFC and NOx level for timing retard, EEC
could be expected to be able to reduce NOx 2-3.5 g/BHP-hr, or
by 30-60 percent at 6.0 g/BHP-hr NOx. These figures may be
somewhat overstated, since the presence of EEC could affect the
effect of an across-the-board timing retard on BSFC. Still,
they indicate a large NOx control potential.

-------
2-31
The percentage reductions estimated for both improved
charge air cooling and EEC would not likely be directly
additive, but would be somewhat lower due to interactions
between the two techniques. It should be noted that neither of
these two technologies would be expected to increase
particulate emissions, which should follow BSFC and remain
roughly constant.
The use of EGR for NOx control is a well proven technology
for LDDs, but has not been widely used on HDDEs*. Although
publicly available data on the effectiveness of EGR on HDDEs is
limited, a reasonably good example of the effectiveness of EGR
on medium HDDEs is given in Table 2-21. For example, EGR on
the naturally aspirated Caterpillar 3208 reduces NOx about 60
percent. Also, although not quite as dramatic, a reduction of
16 percent can be seen for comparable versions of GM's 8. 2L
turbocharged engine.
One drawback of EGR is that LDD emission test data and a
limited amount of HDDE emission test data indicate that
controlling NOx with EGR increases particulate emission levels,
often by a greater percentage than the percentage reduction in
NOx emission. [15,25a] Recent LDD data show this effect to be
decreasing, [2], but the carryover of this improvement to ni
engines may not be straightforward. Electronically controlled
EGR would certainly mitigate some of this increase in
particulate, but whether or not it can be completely eliminated
is unknown at this time.
Unlike improved charge air cooling and EEC, which
manufacturers are moving toward for market reasons, EGR would
only be added for emission control purposes. While EGR should
be technically feasible for HDDEs, its implementation faces a
number of practical hurdles.
For example, the recycled gas will contain particulate
matter which will be spread throughout the intake air in the
cylinder. This will increase the contact of particulate with
Several manufacturers have suggested that for full
efficiency an EGR system will require the use of an EGR
cooler, performing the same function as charge air
cooling.[22] The need for such a cooler would likely
depend on the overall EGR rate being used; the higher the
rate, the greater the need. As will be seen below, EGR is
not expected to be used on many HDDEs to meet a 4.0
g/BHP-hr standard, and even then at low overall rates.
Therefore, at this time, EPA does not have adequate data
on hand to judqe whether an EGR cooler is necessary.

-------
2-82
the cylinder walls and could raise particulate levels in the
engine oil and reduce oil change intervals. Also, since the
recycled gas would have to be introduced upstream of the
turbocharger on a turbocharged engine, increased turbocharger
wear may occur. If an aftercooler is present, it's surfaces
could be fouled by recycled particulate. A system which
eliminated the presence of particles in the recycled exhaust
would solve these problems, but such a system could be
difficult to design. Some possibilities are: 1) an oil bath
filtering system, 2) placing a small trap-oxidizer in the EGR
circuit, or 3) taking the exhaust from a point after the
exhaust trap-oxidizer.
The one production EGR system on a HDDE is on the Cat
3208. However, this system provides EGR primarily at light
loads for the control and the enqine is naturally aspirated.
Thus, it does not address many of the above-mentioned
concerns. Mercedes-Benz uses EGR on their turbocharged 300D
passenger car, but the useful life of this vehicle is well
below that of DI HDDEs.
Overall, EGR has more drawbacks associated with it
compared to the other two techniques mentioned above. Thus,
its use will likely be limited, particularly on turbocharged
and aftercooled engines. Even when used, the overall EGR rate
will likely be low, to minimize its effect on particulate
emissions and durability.
In summary, improved charge cooling and EEC, coupled with
general BSFC improvements, appear to have the potential to
provide more than the necessary 30-40 percent reduction in NOx
emissions required by the 4.0 g/BFP-hr. Where one or both of
these technologies is not sufficient, additional timing retard
and/or EGR may be used. The use of an efficient trap-oxidizer
for particulate control should mitigate any particulate
increase resulting from use of these latter two techniques.
However, use of either of these two less desirable technologies
is expected to be the exception in the short-term and
negligible in the long-term.
3. Other Impacts of the Proposed 1990 NOx Standard
a. Hydrocarbon Emissions
There is some indication that charge air cooling may very
slightly increase HC emission levels. Therefore, advanced
charge air cooling might lead to slightly higher HC emissions
as well, apparently due to the fact that the air entering the
cylinder is cooler.

-------
2-83
Turning to EGR, LDD data indicate that relatively high EGR
rates will also lead to higher HC emission rates. However, the
EGR rates used by DI HDDEs should be well below those used by
LDDs, given the practical problems of using EGR in DI engines
mentioned above. Also, as evidenced by Caterpillar's use of
light-load EGR on one of its 3208 models, EGR can actually be
used as an explicit HC control strategy. Therefore, little, if
any, adverse effect on the HC emissions should occur due to the
small amount of EGR expected to be used.
Finally, electronic controls should decrease HC emissions,
if affecting them at all, due to the increased flexibility
provided.
Overall, these should be little net change in HC emissions
due to a 4.0 g/BHP-hr NOx standard.
b. Fuel Consumption
In a situation where fuel consumption could be optimized,
techniaues such as advanced charge air cooling and EEC would be
expected to provide significant BSFC improvements.[25]
However, when optimized for NOx emission control, these BSFC
improvements will likely be diminished.
NAS estimated that a 4 g/BHP-hr NOx level would cause a
4.5-8 percent BSFC increase over a 6 g/BHP-hr level. This
assumes that timing retard is the primary control technique,
though "advanced" (1980-81) combustion chamber designs which
reduce the effect of timing retard are also considered.[25d]
In their recent report on light-duty diesels, the National
Research Council estimated that a 10-15 percent increase in
fuel economy would occur with electronic controls.[25e] Also,
in a recent paper, GM reported that EEC and improved charge air
cooling would improve BSFC 13-16 percent at 4-6 g/BHP-hr NOx
levels.[25]
Overall, then, it would appear that no fuel consumption
penalty should occur if advanced NOx control techniaues are
used to meet the 4 g/BHP-hr NOx standard. However, for a
variety of reasons, such advanced techniques may not be applied
to all engines immediately in 1990. Cost considerations,
leadtime constraints, and the timing of model
replacement/redesign could lead a manufacturer to choose
further timing retard over electronics and improved charge air
cooling. If this occurred, then a short-term BSFC penalty of
up to 1-2 percent could occur.

-------
2-84
E. Technological Feasibility of the Proposed 1990
Particulate Standard
1.	Introduction
In addition to the engine-related emission control
techniques discussed above with respect to the 1987 particulate
and those discussed below with respect to 1990/ another method
of particulate control is available. It involves the
application of an after treatment device to a heavy-duty diesel
engine. The most developed after treatment device is the
particulate trap-oxidizer. Other devices, such as
electrostatic filters, cylone filters, thermal precipitators,
liquid scrubbers, fluidized bed scrubbers and granular filter
beds, have been suggested, but their state of development is
not well advanced, and need much more development before they
could be considered commercially feasible. A trap-oxidizer
works by collecting particulate matter from the diesel
enqine-out exhaust and oxidizing the trapped particulate
matter. The oxidation phase (also known as regeneration) is
necessary because it is not possible for a trap to collect all
the particulate matter emitted by a diesel enqine over its
lifetime. (A trap designed to hold all of the accumulated
particulate would not be viable because of its very large size
and because the resulting increased exhaust gas backpressure
would be detrimental to engine operation.) In this section,
EPA will evaluate the feasibility of this control technology
for application to heavy-duty engines.
The initial research and development of diesel particulate
trap oxidizers was done by the light-duty diesel vehicle
industry. Much of the technology developed by and for this
industry has formed the basis for the development of similar
technology for the heavy-duty diesel engine industry.
Therefore, the discussion of heavy-duty diesel particulate
trapping technology will start by summarizing the status of the
light-duty trap-oxidizer. This will be followed by a
discussion of the applicability of light-duty diesel
particulate trapping technology to heavy-duty engines. The
third part of this discussion will examine the status of
heavy-duty diesel particulate trap technology development and
the additional development that is necessary for a particulate
trap-oxidizer system to be available for use on heavy-duty
diesel engines. The ability of trap-equipped heavy-duty diesel
engines to meet a .25 g/BHP-hr trap standard will also be
examined.
2.	Status of Light-Duty Trap Oxidizer Development
The technical feasibility of liqht-duty trap-oxidizer use
was recently reviewed in two analyses: "Trap-Oxidizer

-------
2-85
Feasibility Study"[261 an EPA document, and "Trap-Oxidizer
Technology for Light-Duty Diesel Vehicles: Feasibility, Costs,
and Present Status,"[27] a document prepared for EPA by an
independent contractor. The other major source of information
available to the Agency came from the public participation
(comprised of public comments and data) that resulted from the
proposed standard of 0.2 grams of particulate per mile for
light-duty vehicles produced for the 1985 model year (delayed
to 1987). Based upon this information, the Agency concluded
that light-duty vehicle trap-oxidizers would be technically
feasible no later than the 1987 model year. The following is a
summary of the data that led to this conclusion. (An EPA
regulatory support document, "An Updated Assessment of the
Feasibility of Trap-Oxidizers", contains a more detailed,
complete report. [28])
To summarize the status of the light-duty trap oxidizer,
there are two basic trap-oxidizer designs that can meet the
required collection efficiency, backpressure, and durability
characteristics. One of them is the ceramic wall-flow monolith
trap and the second design is an alumina-coated wire mesh
trap. Both trap designs are capable of satisfactory collection
efficiencies (70-90 percent for the ceramic wall-flow monolith
trap and 50-80 percent for the wire mesh trap) and both can
maintain acceptable backpressure over durability accumulations
of up to 50,000 miles with properly controlled regeneration.
The wall-flow monolith diesel particulate trap is based on
a design similar to the ceramic honeycomb substrate used in
light-duty gasoline-fueled-engine catalytic converters. As
such, it is a porous, ceramic honeycomb matrix of tubes of
which alternate cells of the upstream face are plugged with
ceramic. The complementary, alternate cells of the downstream
face are also plugged with ceramic and thus there is no direct
path for exhaust to flow through the monolith. Rather, the
exhaust flow must enter the open cells on the upstream face and
then pass through the porous wall (and then out the open
alternate cell on the downstream face) which filters the
exhaust stream of its particulate matter. The particulate
matter collects on the walls (which increases the backpressure
through the trap) and must be periodically removed by oxidizing
it, which serves to "regenerate" the trap. This oxidation
procedure is very important for proper long term trap usage.
It will be discussed in more detail in a subsequent paragraph.
This type of trap usually contains no catalytic material to
alter the temperatures at which oxidation begins.
On the other hand, the wire mesh trap usually contains
wire which is coated with alumina which, in turn, is coated
with a washcoat of catalytically active metal. The tightly

-------
2-86
packed mesh traps the particulate and then oxidizes it, thus
regenerating the trap more or less continuously.
The second part of trap-oxidizer development concerns the
oxidation of the trapped particulate. For non-catalyzed traps,
exhaust temperatures in the range of 500-650°C and oxygen
levels of 3-4 percent are required for a duration of 2-10
minutes in order for regeneration to occur. A catalyst-based
trap will require lower pre-trap exhaust temperatures of
approximately 310-350°C. There are two general approaches to
regeneration: positive regeneration and self-regeneration.
Positive regeneration systems regenerate the trap at a specific
point in time by raising the temperature of the particulate in
the trap-oxidizer. Methods of positive regeneration include
external heat sources (e.g., fuel burner), modification of
engine parameters (e.g., air intake throttling), or raisinq the
HC and CO levels in the exhaust for catalyzed traps {e.g,
exhaust-stroke fuel injection). On the other hand,
self-regeneration relies on attaining regeneration conditions
during normal vehicle operation without the activation of any
special mechanism. An example of self-regeneration is the
continuous use of fuel additives to lower the particulate
ignition temperature. Technically feasible regeneration
systems for non-catalyzed ceramic wall-flow monolith traps
include the fuel burner, electrically heated traps and a trap
that is close coupled to . the exhaust manifold that uses
throttling and EGR for elevating the upstream temperatures.
The fuel burner provides the additional heat necessary for
regeneration by igniting an air/fuel mixture immediately in
front of the trap inlet. The burner requires a fuel supply and
fuel injector, a glow plug or spark plug to initiate ignition,
and an oxygen supply. The oxygen may be supplied by a
low-pressure air pump (air-fed burner) or by residual oxvaen
present in diesel exhaust (exhaust-fed burner). Alternatively,
the burner can heat the entire exhaust gas flow (in-line
burner), or can heat just the trap itself while the exhaust gas
is routed away from the trap (bypass burner). All of these
options have been investigated by various manufacturers.
For the catalyzed wire mesh traps, there are several
technically feasible regeneration approaches. These include
exhaust-stroke fuel injection, limited throttling with excess
EGR and limited throttling without excess EGR.
The basic concept of self-regeneration through
exhaust-stroke fuel injection is to create the diesel
equivalent of a misfire by injecting additional fuel into one
cylinder at bottom dead center at the beginning of the exhaust
stroke. This fuel is "cracked" into lighter HC and CO

-------
2-87
molecules by the heat and pressure which still exists in the
cylinder. The HC and CO are then exhausted on the piston
upstroke and oxidized by the precious metal catalyst on the
wire mesh trap. Johnson-Matthey has stated that the exothermic
reaction of the HC and CO produces a temperature rise of
approximately 150-200°C. The catalyzed wire mesh trap oxidizes
particulate at approximately 350°C, so exhaust-stroke fuel
injection can initiate regeneration for catalyzed wire mesh
traps at exhaust temperatures as low as 200°C. Limited
throttling (one or two cylinders) with or without excess EGR
are two other methods of delivering additional quantities of HC
and CO to the catalyzed trap. (These systems would obviously
not work for non-catalyzed traps.)
The utilization of fuel additives to induce regeneration
has also been examined. This concept involves the introduction
of a catalytically active fuel additive (usually a metal
compound) upstream of the trap, thus catalyzing oxidation of
accumulated particulate. This fuel additive introduction could
be continuous or intermittent. An example of a continuous
additive introduction system would be one that had the additive
mixed with the fuel, either during refueling or earlier in the
fuel distribution system. An example of an intermittent
additive system would be a mechanical injection system that
would place the additive either into the onboard fuel after it
leaves the fuel tank or into the exhaust stream.
Because additive systems can be mechanically simpler,
requiring fewer vehicle/engine modifications than the other
regeneration systems, they may be preferred by some
manufacturers. The eventual use of fuel additives, however, is
conditional upon the resolution of trap plugging and
environmental concerns. For example, certain test results
indicate that the trap may retain most of the catalyst additive
material. But if there are catalyst emissions, the question of
whether they would represent an unacceptable environmental
hazard would need resolution.
The problem of how the additive will be introduced to the
fuel must also be resolved. Mixing the additive into the fuel
at the refinery raises concerns about the fuel distribution.
For example, will all diesels use the fuel with additives? If
so, the non-trap equipped vehicles would emit all of the
catalyst material. Will a more costly and complex method of
separate fuel distribution be used for trap and non-trap
equipped engines? The other alternative, on-board introduction
of the additive, is less complex from a fuel distribution
standpoint, and is presently being explored by various
manufacturers. It too presents some unanswered questions such
as potential toxicity of an additive that would have to be
handled and stored onboard a vehicle.

-------
2-88
Positive regeneration requires a control technique to
determine when regeneration is required, so as to initiate the
regeneration process and to confirm that regeneration is
complete. To date, Johnson-Matthey has designed a regeneration
control algorithm for HC and CO enrichment with its catalzed
wire mesh trap. A manually controlled fuel burner regeneration
system has also been developed by Ford. The primary work
remaining will be the development of a fully automated system
and the additional testinq of prototype systems needed for
system refinement. None of the tasks associated with positive
regeneration control systems are judged to be technically
infeasible.
The Agency considers the manufacturers to be in the final
stage of light-duty trap-oxidizer development. Although a
large amount of the information about trap-oxidizers is
proprietary, what information is publicly available indicates
that final development and optimization are underway.
Daimler-Benz's 1985 certification plans confirm this.[29]
Daimler-Benz plans to start certification testing of its 3L
5-cylinder turbo diesel equipped with a particulate trap to
meet California's 1985 emission standards. The full details of
Mercedes' trap technology have not been publicly released, only
the general design. According to this general information
there will not be a separate high-temperature burner; the
accumulated particulate will be oxidized by self regeneration,
most likely through partial throttling at designated times.
Although the other vehicle manufacturers have not yet
identified a trap/regeneration control system to be committed
to production, EPA believes they are not too far behind Daimler
Benz's trap development to jeopardize compliance for 1987.
The trap-oxidizer analysis by EPA's independent
contractor[27] confirms the Agency's belief that light-duty
trap technology is at a very advanced stage. The analysis
identified three trap-oxidizer systems which it considered to
be at an advanced stage of development: 1) ceramic wall-flow
monolith trap/regeneration with fuel burner, 2) catalyzed wire
mesh trap/regeneration by HC and CO enrichment (such as
exhaust-stroke fuel injection), and 3) ceramic wall-flow
monolith trap/self-regeneration using fuel additives. This
study concluded that there are no technical uncertainties
regarding the feasibility of the ceramic monolith/fuel burner
system. It considered successful development of the catalyzed
wire mesh/HC and CO enrichment system to be "highly probable"
(although the occurrence of increased sulfate emissions has not
been resolved) and development of the ceramic monolith/fuel
additive system to be "probable" in the near term. Both of
these systems would be preferred if feasible, due to their
greater simplicity and lower costs. If these systems did not

-------
2-89
prove to be feasible, then the fuel burner could be adopted
industry-wide. EPA agrees with these conclusions and believes
that successfully functioning trap-oxidizers will be available
for 1987 LDV use.
3. Technical Evaluation of Light-Duty/Heavy-Duty
Differences
Much of the particulate trap technology described in the
previous section can be adapted to heavy-duty engines.
However, while the heavy-duty industry can make use of existing
trap technology, there are specific differences between light-
and heavy-duty applications that must be considered in
designing a heavy-duty trap oxidizer. What these differences
are and how these unique design needs can be resolved will be
discussed in the following paragraphs.
An obvious light-duty/heavy-duty difference is the engine
size and load factor. The larger and more heavily loaded HDE
has a higher exhaust volume and mass flow rate, thus increasing
the amount of particulate in the exhaust. In order to
compensate for the increased flow rate, the heavy-duty trap
size must be increased to ensure that the backpressure doesn't
rise too quickly, or regeneration occur too frequently. This
larger trap size can be accomplished by using larger traps
(which could present mechanical durability problems due to
greater stresses) or by using more than one trap. The larger
trap size also makes auxiliary heating for trap regeneration
more difficult on HDEs than on LDEs; heating evenly over a
larger surface can be a problem. However, the larger size of a
HD vehicle does ease the housing problem of the larger
heavy-duty trap. The larger HDV also allows for more variation
in the trap geometry. Another problem is with temperatures.
Temperature considerations must be divided into two areas of
discussion that relate to turbocharqers. Exhaust gases expand
in the turbine portion of a turbocharger in an engine so
equipped, which lowers the temperature of the exhaust stream.
Therefore, non-turbocharged engines and exhaust systems
upstream of a turbocharger have higher temperatures than found
in exhaust systems downstream of a turbocharger. Most
heavy-duty engines currently produced are equipped with a
turbocharger.
For the case of a non-turbocharged engine and in exhaust
system locations upstream of the turbocharger, HDDEs yield
higher exhaust temperatures than LDDEs due to sustained high
load/high speed operating conditions typical of HDDEs. A trap
may melt when exposed to these sustained higher temperatures
and thus, heavy-duty traps must be designed to withstand them.
During operation at high temperature modes, it may be possible

-------
2-90
for self-regeneration to occur because the required temperature
range of 560°-650°C may be reached. However, because the
operating conditions which yield these high temperatures may
not occur routinely, self-regeneration cannot be relied upon; a
regeneration system must also operate when the heavy-duty trap
is exposed to routine cooler exhaust temperatures. For
example, HD trucks frequently idle for long periods of time,
yielding lower exhaust temperatures.
In the case of a turbocharged engine, in locations
downstream of the turbocharger, the necessary regeneration
temperature may not be attained as required, and the
accumulating particulate would increase the backpressure to the
detriment of engine operation. The regeneration technique must
therefore compensate for the low-temperature periods by either
increasing the exhaust-gas temperature (e.g., using a fuel
burner) or decreasing the required oxidation temperature of the
particulate (e.g., using fuel additives or a catalyzed trap).
In short, although HDE regeneration problems are similar to
those observed in light-duty, trap sizing and temperature
ranges are special problems that need to be accommodated.
Another major difference between light-duty and heavy-duty
diesel operation is the useful life of an engine. The useful
life of a LDDV is approximately 100,,000 miles while a HDDE' s
useful life ranges from 110,000 to 290,000 miles, depending
upon the service usage. In-use engines can conceivably run
much further. The longer on-road life of a heavy-duty diesel
engine requires that substantially more durability be designed
into the trap. This will require that the heavy-duty industry
perform sufficient durability testing under conditions which
simulate in-use HDE operating conditions.
Another difference that needs to be addressed is the
potential increase in ash accumulation in the trap because of
both the heavy-duty diesels1 longer useful life and higher
exhaust flow rate through the trap. Diesel fuel and
lubricating oil contain metallic additives that are
non-combustible and that can accumulate in the particulate
trap. The accumulated ash does not oxidize, and thus
contributes to increased backpressure and, in extreme cases,
can cause trap clogging-.' Cummins researched this problem[301,
conducting accelerated bench tests to simulate actual engine
operation. Their research yielded an exhaust ash emission rate
of 17 g/1,000 km for a typical HDDE. Further research is
necessary because the bench work of Cummins does not
necessarily characterize an actual heavy-duty diesel engine.
If actual heavy-duty tests show ash accumulation to be a real
problem, possible solutions include changing the trap design
and flow path, or reducing or changing the metallic additives
in the fuel and oil.

-------
2-91
In conclusion, conditions which are specific to the HDDE
environment must be considered and dealt with when extending
light-duty trap technology to heavy-duty applications. ^"hile
EPA does not deny that considerable development eEfort will be
necessary on the part of the heavy-duty industry, EPA does not
consider the problems to be without solutions. The feasibility
of light-duty traps, combined with the ongoing research and
development of a heavy-duty trap, lead EPA to the conclusion
that traps should be feasible for heavy-duty use.
In fact, early HDDE trap development work is now underway,
and is discussed next.
4. Status of Heavy-Duty Trap Technology Development
Data discussed in this section comes from several
sources. After the first HPRM for HDDE particulate emission
standards was published (46 FR 1910, January 7, 1981), EPA
received comments from the heavy-duty industry describing the
status of their current trap work. Representatives of most
HDDE manufacturers have also recently met with EPA staff to
furthec discuss their development programs. (Much of the
information about test results and the specific trap-oxidizers
being tested is deemed by the manufacturers to be of a
proprietary nature; only that information that is publicly
available can be reviewed here.) Finally, EPA has contracted
with an independent researcher to study the problems of FIDDE
trap development.
The January 7, 1981 NPRM proposed a trap-based standard of
.25 g/BHP-hr for the 1986 model year. In their comments, the
manufacturers were in agreement that heavy-duty "trap-oxidi2er
technology is still in the development stage, and it will
certainly not be ready for introduction in production
heavy-duty vehicles by the 1986 model year."(31] The
significant question addressed by this analysis is when traps
will be ready. A leading potential trap manufacturer stated to
Mack that "traps will not be available for heavy-duty engines
until 1989."(32] The traps being tested for heavy-duty
application are similar to the promising light-duty traps with
an exception that will be discussed in the following paragraph;
they include both catalyst and non-catalyst ceramic monolith,
metal mesh, ceramic foam, ceramic fiber, and coil-type
filters. Due to the high exhaust flow, the heavy-duty traps
being evaluated are typically larger than a light-duty trap,
increasing its total capacity. The collection efficiency of
the trapping device, however, is fairly constant for both
light-duty and heavy-duty. Although Cummins claims that
current traps are not achieving consistent collection
efficiencies, it is generally accepted that the basic designs
can meet the required collection efficiency characteristics.

-------
2-92
In addition to these light-duty trap designs that are
being adapted to heavy-duty requirements, there is a
solica-fiber candle/catalyst design that was developed
especially for heavy-duty use. This system is produced by
Damiler-Benz; the limited available data on its is presented in
a report to the EPA by a independent contractor, [33} , and is
summarized here. The trap consists of a number of candle-like
filtering elements made of treated silica-fiber yarn woven on a
porous metal tube. The "candles" are arranged axially in a
container. The exhaust gas flows into the container, through
the candles radially (outside to inside) , depositing the
particulates in the yarn as it exits through the centers of the
tubes. Daimler-Benz reports[34] that this trap has a high
collection efficiency and low backpressure. Regeneration is
accomplished by injecting a powdered oxidation catalyst into
the exhaust stream at a temperature high enough for oxidation
to occur. The control and actuator system includes an
electronic control system, a backpressure sensor, an exhaust
temperature sensor, and an engine rpm sensor.
The second and third parts of a trap-oxidizer system,
regeneration and its control technique, are the major focus of
industry's research. The heavy-duty operating conditions,
especially the frequent and oft times lengthy idling periods,
prevent self-regeneration from occurring on a regular basis
because of insufficient exhaust temperatures. Therefore,
either a positive regeneration system is required or a means
for lowering the temperature of self-regeneration is
necessary. The regeneration methods being evaluated are
similar to those discussed in the previous section for
light-duty, plus the Daimler-Benz method. Fuel additives are
being investigated as an appealing method of regeneration due
to the simplicity of the system and the resulting low cost to
the engine manufacturer. The use of fuel additives should
lower the temperature at which the trapped particulates
oxidizes, but Daimler-Benz reports that testing has not yet
produced successful or reproducible results. [34) Also, the
same unresolved questions of light-duty fuel additive use apply
to heavy-duty use: the method of introduction of the additives
to the fuel, potential environmental consequences, and the
possibility of trap plugging.
The use of catalyzed traps has been examined and/or tested
by several manufacturers. A catalyzed trap's regeneration
system, such as late fuel injection, inlet throttling, or
exhaust restriction, raises the HC and CO levels in the
exhaust. The catalyst oxidizes the HC and CO and the
additional heat burns the collected particulate. The drawback
of a catalyzed trap is the formation of sulfate by the
catalyst: heavy-duty operating conditions have been observed

-------
2-93
to significantly increase sulfate production from a catalyzed
trap. However, the increased sulfate production observed
during heavy-duty testing may be a function of the type of
catalyzed trap being used. Presently the traps being evaluated
for heavy-duty are larger light-duty traps that haven't yet
been fully adapted foe heavy-duty usage. For example,
Johnson-Matthey believes that the high-sulfate problem is
associated with the catalyst noble metal loading. Work is now
underway to create a reformulated trap to eliminate the problem
of sulfate production. The heavy-duty trap developed by
Daimler-Benz has a tendency to retain the catalytic additive,
resulting in plugging and increases in backpressure after
100,000 miles. This design would require a replacement of the
trap if the plugging problem could not be solved.
The most developed regeneration system continues to be
burner regeneration, in conjunction with uncatalyzed traps.
Although complex, a burner system appears to be the type of
regeneration that will be viable at the nearest future date.
The heavy-dutv burner system is similar to light-duty burner
regeneration systems, although a heavy-duty vehicle would
likely require a larger system.
The silica-fiber candle/catalvst system also appears to be
a well developed technology. Daimler-Benz has accumulated more
than 250,000 kilometers of testing on it and projects that it
will be ready for production in 1989.[331
Development of an automatic regeneration control system
remains the next major step. Except for the instances where
self-regeneration occurred, all of the traps tested to date
with heavy-duty engines have had their regenerations controlled
manually. Industry's work in this area is purportedly
constrained by the fact that "automotive trap technology is in
its infancy."[35] GM in particular believes that "optimization
of a control system cannot be done until sufficient data are
available from a suitable trap operated under all anticipated
types of driving conditions31]	has tested numerous
trap-oxidizers on both engine dynamometers and in vehicles, and
has observed trap failures involving thermal cracking of the
trap material. GM claims that until a trap can be designed
that will fulfill all the necessary performance requirements
such that the system can be completely characterized, a control
system cannot be developed. This conclusion, however, probably
represents the results of worst case durability testing, and
fails to consider the moderating effects of a properly
functioning regeneration control system. The Agency agrees
that the heavy-duty engine/trap/regeneration system's
operational limits must be thoroughly defined. However, a
functioning regeneration system should help prevent thermal

-------
2-94
failure of the trap, as certainly will other design and
material modifications.
Any production regeneration control system must also
include sensors, actuators, and a programmed central processing
unit for a fully automated system. This equipment is already
being developed for light-duty trap regeneration control
systems, and should be easily applicable to the heavy-duty
trap. In addition, the heavy-duty industry is also evaluating
the application of electronically controlled fuel-injection
systems (see above), which would facilitate the application of
the control logic for the regeneration system.
In summary, a heavy-duty trap regeneration system can
probably be designed to meet future regulatory needs. As the
light-duty industry progresses towards the complete development
of a regeneration control system, the heavy-duty industry can
probably adapt this technology to its own needs. However,
there is a definite lack of data from the heavy-duty diesel
engine industry on particulate traps, especially relative to
the light-duty diesel industry. This is probably due to the
fact that the light-duty industry has had to work towards a
particulate standard that essentially required traps whereas
the heavy-duty industry has not had that incentive.
To obtain some HDDE trap data the Agency contracted with
Southwest Research Institute (SwRl) to test a trap-oxidizer
system on a heavy-duty engine. The objectives of the testing
were: 1) to evaluate the effectiveness of a low-mileage trap
installed on a HDE, 2) to develop a method of regeneration, and
3) to characterize the emission levels during trap use compared
to the baseline levels. An in-depth results of the
investigation are available in the final report by SwRI[361;
the summary of the testing results is included here.
A Corning catalyst was tested on a DDAD 6V71 diesel coach
engine and also in a 1980 GMC RTS-II coach vehicle. The engine
was a 2-stroke direct-injected diesel engine which used a
blower for scavenging. The trap was a ceramic wall-flow
monolith type and regeneration was accomplished with an
in-exhaust-pipe burner. The regeneration was manually
performed at idle, with the burner capable of raising exhaust
temperatures from 120 to 700°C.
The trap was found quite effective in reducing particulate
emissions. During transient testing of the engine, the trap
reduced total particulate emissions by about 65 percent;
chassis testing of the coach vehicle equipped with a trap
yielded a 92 percent reduction in particulate emissions. The
trap reduced steady-state particulate emissions from the engine

-------
2-95
by an average of 79 percent. Smoke emissions were reduced by
almost 100 percent under all modes of operation. The trap was
effective in reducing carbonaceous particulates, although it
had a variable effect in reducing the soluble organic fraction
(SOF) of the total particulate. Differences between observed
brake specific fuel consumption (BSFC) with and without the
trap were minimal. There was, however, a thermal cracking
failure of the trap following a chassis test when the trap exit
temperature peaked at 990CC (occurring 66 seconds after the
burner was shut off).
The SwRI test results indicate that sufficient trap
efficiency can be obtained with a heavy-duty trap, for both
steady-state and transient testing. A method of regeneration
was also developed; some problem was seen with thermal
cracking, but as noted above, further development and
refinement of an automatic regeneration control system will
help prevent thermal trap failures.
Finally, to date there has been some concern as to the
fuel economy effects of trap-ox idizer use. As particulate are
collected, an increase in the exhaust back pressure occurs,
which may result in fuel economy losses. Since minimal
differences in BSFC were noted by SwRI with the trap as
compared to the baseline levels, EPA believes that a zero to 2
percent fuel consumption penalty is the very worst penalty
which will be observed with HD traps. Carefully controlled
regeneration of the trap should minimize the fuel economy
impact. For example, Cummins notes that "if the trap can be
regenerated at 100 mile intervals, a 1.6 percent fuel
consumption penalty will be incur red35] As with the thermal
trap failures, the refinement of an automatic regeneration
control system is critical to minimize increased fuel
consumption associated with trap oxidizers.
5. Design Target and Degree of Requisite Trap
Application
A trap standard of .25 g/BHP-hr has been proposed for 1990
HDDEs. This emission level is dependent on several factors and
can be calculated from the following formula, which has already
been described in a slightly different form.
MPL = ETL X DF x AOL X (1 - Eff)
Where:
MPL = maximum pass level
ETL = engine out design target level

-------
2-96
DF = deterioration factor
AQL = SEA adjustment factor
Eff = trap efficiency
The target level, 0.47-0.51 g/BHP-hr, and the AQL factor,
1.10-1.15, apply to an engine-out emission level of ,6
g/BHP-hr, as previously discussed. The deterioration factor
represents the deterioration for engine-out particulate and
also the deterioration for the trap. The DF for the engine-out
PM was determined to be 1.1 on a multiplicative basis. To
determine the DF for particulate traps, it is necessary to
perforin durability testing. Although no such testing has been
performed for heavy-duty traps, the deterioration rate for
light- and heavy-duty traps is assumed to be similar. Swri
tested two types of traps, a ceramic trap and a wire mesh trap,
operating both traps in LDVs over 50,000 miles of service
accumulation. No significant deterioration of particulate
emissions occurred over the 50,000 miles of testing. Thus, a
DF of 1.0 is assumed for traps. This is reasonable,
particularly for wall-flow ceramic traps which apparently are
currently the most preferred, since there is no mechanism for
small reductions in efficiency. If cracks develop the
efficiency will effectively be zero. However, this is a
reliability problem which is expected to be solved by 1990.
Deterioration thus appears to be zero.
The trap efficiency is the most variable factor that
effects the emission level. The ceramic wall-flow monolith
trap is the more efficient trap with a collection rate ranging
from 70 to 90 percent: this is a very conservative range since
most of the testing with this type of trap has yielded
efficiencies of greater than 90 percent. The wire mesh trap's
collection efficiency ranges from 50 to 80 percent. The
collection efficiency of the silica-fiber candle trap is not
presently available. The differences in trapping efficiencies
yield large differences in particulate emission levels. An 60
percent efficiency has been chosen to represent a trap
efficiency that is in this range, but would be quite low if
wall-flow traps were used.
Applying these factors to the above formula, the
calculated emission level is 0.14-0.17 g/BHP-hr. Thus, traps
will not be required on all engines; the technically most
difficult applications will be able to be excluded from trap
usage, which is desirable given the new nature of this
technology. With averaging, the number of trap-equipped
heavy-duty engines can be reduced to roughly 70 percent, thus
increasing the averaged particulate emission level to 0.24-0.26
g/BHP-hr, bracketing the 0.25 g/BHP-hr standard proposed.

-------
2-97
Limiting consideration to only highly efficient, ceramic
wall-flow monoliths, even lower levels would be achievable.
Using a reduction efficiency of 90 percent in the above
calculations would yield achievable levels of 0.07-0.09
g/BHP-hr, or low enough to comply with a 0.10 q/BHP-hr
standard. Of course, under this standard, essentially all
vehicles would require trap-oxidizers.
6. Summary
Trap technology, while not presently available for
production heavy-duty application, should be technically
feasible by 1990. To date, heavy-duty trap development efforts
have been limited. In light-duty, substantial development was
only stimulated by the proposal and promulgation of trap-based
standards. Furthermore, with industry's vigorous pursuit of a
trap-oxidizer system, a 1990 HDDE trap-based particulate
standard should be meetable. This follows from the fact that
trap-oxidizers will be in production on many light-duty diesels
no later than 1987. Secondly, as discussed above, light-duty
trap technology can be adapted for heavy-duty use. Additional
design effort is necessary but the problems and differences
from light-duty applications are easily identified, and should
be solvable by system design modifications, optimizations, and
the application of known technologies. The heavy-duty industry
has six years of leadtime until 1990 to refine technologies and
incorporate these changes. At least three of these years will
follow widespread production of light-duty traps. It stands to
reason, however, that a 1987 implementation date, the year
traps will be produced on LDVs, is too soon to be feasible for
heavy-duty given the unique problems to be solved.
In summary, based upon the current state of trap
development, there appears to be sufficient time for the
manufacturers to design, develop, and prepare trap-oxidizers
for 1990 model year HDDEs. Development of an automatic
regeneration control system and solution of potential ash
accumulation problems remain the critical tasks left for the
heavy-duty industry.
F. Achievable Non-Trap Particulate Levels in 1990
To provide additional pertinent information to the
rulemaking process, the feasible level of a 1990 non-trap
particulate standard, in conjunction with the proposed 1990 NOx
standard of 4.0 g/BHP-hr, was evaluated. That analysis appears
below.
1. Effect of a 4.0 g/BHP-hr NOx Standard
The first step in this analysis is to determine the
effect, if any, that a 4.0 g/BHP-hr NOx standard would have on

-------
2-98
1987 particulate levels. This primarily means addressing
improved charge air cooling and EEC, since EGR and additional
injection retard are expected to be needed on only a few
engines and their impact can be ignored here.
As mentioned in Section D.2.b.ii. above, when used to
their fullest NOx control extent (i.e., at constant BSFC),
neither improved charge air cooling or EEC is expected to
reduce particulate emissions. However, it is possible that
this extent of NOx control will not be needed, and BSFC and
particulate emissions could both improve. As it is unknown how
the size of the particulate reduction would compare to the size
of the BSFC reduction, it is not possible to quanti tatively
estimate the potential particulate reduction. However, given
the size of the initial BSFC improvements, any net particulate
reduction is likely to be small (i.e., less than 10 percent).
In addition to the effect of these specific NOx controls,
there is the general effect of BSFC improvements on all
emissions. In Section D.2.b.i. above, one analyst's estimates
were described, showing a 7 to 8 percent BSFC improvement
between 1987 and 1992. Using this as a rough approximation, a
5 to 10 percent reduction in particulate emissions should be
achievable via use of fuel efficiency techniques,
2. Engine-Related Particulate ControLs
Engine-related particulate controls in the 1987 timeframe
were already discussed in Section C. 2. above. Further
reductions beyond these could come from five areas: 1) further
improvements to general combustion chamber design, 2) advanced
fuel injection techniques, including high pressure injection,
3) cylinder heat retention technology, and 4) fuel
modification, particularly use of methanol. The particulate
reduction potential in each of these areas is dicussed below.
Another technique, turbocharg ing, will riot be discussed simply
because the majority of DI FTDDEs are already turbocharged.
a. Combustion Chamber Design
Modifications to basic combustion chamber designs were
already discussed previously with respect to compliance with
the 1987 particulate standard. While three additional years of
leadtime would certainly provide the opportunity for further
improvement, the general evidence available appears to indicate
that across-the-board particulate reductions are fairly small.
Parametric studies, like that conducted at Michigan Technical
University (MTCJ) for EPA, [37] have simply not discovered
standard design criteria which could lead to significant
particulate reductions by 1990. At the same time, a large

-------
2-99
number of researchers are investigating the formation of
particles in diesel engine on a fundamental level and the
results of this work may lead to reductions in the future.
Thus, at the present time, this area does not appear to hold
significant promise in providinq large particulate reductions
in this decade.
b. Fuel Injection System Design
Two aspects of the HDDE fuel injection system appear to
hold some promise in reducing particulate emissions. One is
increased injection pressure (i.e., injection rate) and the
other is the shape of the fuel injection rate over its
duration. The reader is referred to Reference 37 for a
detailed description of these modifications.
With respect to increased injection pressures, Figure 2-9
shows its effect on emissions and BSFC on 6 modes of the
13-mode cycle (3, 4, 5, 9, 10, and 11). The peak injection
pressure of the standard (low pressure) pump was 40.0 MPa,
while that of the high pressure pump was 46.5 MPa. As can be
seen, significant reductions in total particulate matter (TPM)
occurred at the high injection pressure. Focusing on NOx
emissions first, a timing of about 14° BTC with high injection
pressure would have held NOx emissions constant. At this
timing, BSFC should also be roughly constant between the two
injection pressures. However, particulate emissions (both
solids and soluble organics) decrease 25-60 percent, depending
on the mode.* While these reductions may not be repeatable on
every engine design, they are indicative of the potential
reductions available via high pressure injection. Also, a peak
pressure increase of 16 percent is not large and should be
achievable by 1990.
With respect to the shape of the fuel injection rate curve
over the duration of injection, the work performed at MTU
suggests three changes to further lower particulate. [37] One,
use of a zero sac volume should lower soluble organics. Two,
injecting the fuel as rapidly as possible should reduce the
carbonaceous portion of total particulate by enhancing rapid
air-fuel mixing. Three, the initial injection rate should be
only high enough for a time to avoid poor mixing, since too
high an initial injection rate can lead to increased solids
formation. These changes, coupled with high injection
pressure, are projected by MTU to reduce carbonaceous
particulate by 70 percent overall, with lower but still
significant reductions in the soluble organics.
A third component of total particulate, sulfates, is not
shown in the figure.

-------
2-100
Figure 2-9
Effect of Increased Rate of Injection
and Retarded Injection Timing. (1)
a"? 100
h 1 50
"• I
SI
o»«
9 H
-J *t
so
100
so
(A « 0
1000
o*? 500
MM. BOWL -
• LOW MAT! ¦¦¦ WlOW NATS
^frtre
l a	
17*4 1*
II* 11?
ij» rrc
A iriTC
1000
BMEP (KPA)
1270 rpm
m
M
-	ISO
100
-	so
0
*1 30
(A O
1 0
ISO
§|»
«• i
Ms
O £1000
x 5
~ soo
_ 0
u. »
S5 23
o
tt .20
1
&	^	.lIH-ao J
ir»ir «TC"
1
trc
il»«T
k		n»«Te__
p ~~—-a-	^ r «tc J
X
400 SOO 800 1000
3MEF (KPA)
2100 rpn
1200
(1) Taken from Reference 37
TPM = Total Particulate Matter.
SQF = Soluble Organic Fraction.

-------
2-101
While increases in injection pressure of 15-20 percent
should be achievable in the 1990 timeframe, affecting the shape
of the injection rate curve would appear to require advanced
unit injectors.[37] At the present time, two of the domestic
manufacturers use distributor pump systems and conversion to
unit injectors would be a major task. Even for those currently
employing unit injectors, such a change would not be minor and
could require at least until 1990 to implement. Thus, these
changes to the shape of the injection rate may be feasible for
some manufacturers and some HDDEs by 1990 but could be quite
difficult for others. Again, the reductions projected based on
data for this one engine may not be the same as that for
another. However, significant reductions in particulate
emissions appear achievable with such technology.
c. Reduced Cylinder Heat Rejection
A significant portion of the heat generated by fuel
combustion (20-30 percent) is rejected to the engine coolant in
order to maintain acceptable temperatures in cylinder walls,
heads, and pistons. For a variety of reasons, such as
efficiency, size, weight, and wider fuel tolerance, research is
being focused on high temperature materials in order to reduce
the amount heat rejected to the coolant. This technology
appears to have a great potential for affecting emissions as
well. Without going into the detail of such "adiabatic"
technology, this emission control potential is described below.
Cummins and the Tank Automotive Command (TACOM) have been
jointly working on an uncooled HDDE for a number of years. In
their latest paper, [38] the emissions of this engine are
described. At constant BSFC, NOx emissions decreased 11-15
percent over the 13-mode cycle, which provides results very
comparable to the NOx emissions of the transient cycle. Or,
put another way, BSFC decreased 1-2 percent at constant NOx
emissions.
Particulate was measured via smoke opacity using
smoke/particulate correlations developed previously. The
results over the three modes tested showed dramatic reductions
of 83-86 percent. While some potential error is induced by the
use of smoke as a surrogate for particulate, it is relatively
certain that the carbonaceous portion of the particulate
decreased by this amount, since it is the carbon that causes
the opacity. While soluable organic emissions are uncertain,
they likely decreased significantly as well, since gaseous HC
emissions decreased dramatically.
These results were those for a completely uncooled
engine. However, it was noted that the reduction in smoke

-------
2-102
opacity dropped consistently as the degree of heat retention
increased. Therefore, one can expect that the initial stages
of heat retention will provide particulate reductions in rouqh
proportion to their degree of heat retention.
The authors present an aggressive timetable Cor the
potential development of this technology. An uncooled engine
is projected to be producible by 1987, based in part on the
successful demonstration of such an engine in an army 5-ton
truck in 1932. Whether this date is for military application
alone or for both military and commercial application is
unknown. it appears unlikely, however, that such technology
would be widely available prior to 1990. For example, EEA only
projects its use after 1992 and then only on heavy HDDEs.[241
Overall, this technology promises very large particulate
control potential and in so far that it is available by 1990,
large particulate reductions could result,
d. Fuel Modification
It is well known that diesel fuel quality has been
deteriorating steadily over the past seven to ten years (eg,
lower cetane and higher 90 percent boiling point) . Given the
continued pressure to obtain as much transportation fuel from a
barrel of crude oil as possible, it is unlikely that this trend
will reverse itself. Approaches to lowering particulate
emissions via improvement of a better quality hydrocarbon fuel
appear too expensive for the control obtained. However, the
use of a markedly different non-petroleum Euel, methanol,
appears to offer large emission reductions, as well as being
feasible in HDDEs.
The testing results of three methanol-fueled HDDEs are
publicly available: 1) a MAN D2566 FMUH enqine{39], a DDA
6V-92TA[25b], and a Volvo TD-100A[391. The Volvo enqine is a
dual-fueled engine, using a pilot injection of diesel fuel to
ignite the primary methanol fuel. As its emission
characterstics are a blend of those expected from use of solely
methanol and solely diesel fuel, they will not be discussed
here. Instead, this discussion will focus on the two engines
operatable on solely methanol, as they reveal most clearly the
effect of this fuel.
The emissions and BSFC of these two engines are shown in
Table 2-25. As can be seen, the catalyst-equipped MAN engine
had extremely low emissions of EC, CO, aldehydes, and
particulate, while the presence of an automotive-type catalyst
had much to do with the meagerness of the former three
pollutants, the particulate emissions were likely low prior to

-------
2-103
Table 2-2S
Emissions and Fuel Consumption of Two
Methanol-Fueled HDDES
Catalyst-Equipped
MAN D2 566 FMUH
(198 hp)[25gl
EPA Transient Test
HC (a/BHP-hr)[1]
Aldehydes
CO (q/BHP-hr)
NOx (q/BHP-hr)
Particulate (q/BHP-hr)
BSFC (lb/BHP-hr)[2]
n.04
0.001
n.3i
6.61
0.0^3
0 . 536
DDA 6V-9 2TA
(270 hp) f25b1
13-Mpde Test
1.28
0.15
2.20
0.17
(II Standard HFID measurement.
[2] Diesel equivalent basis using lower heatinq values.

-------
2-104
the catalyst. The BSFC of this engine is roughly 10-15 percent
higher than that of typical HDDEs of its type and size (eg.,
Cat 3208 and DDh 6V-71N) .
The DDA engine was tested without a catalyst. While
aldehydes were higher than those of a typical HDDE, NOx
emissions were extremely low and particulate emissions were
below the 0.25 g/BHP-hr standard, but not below the 0.10
level. While the extremely low levels of P/Ox should allow
timing advance to reduce particulate further, there is reason
to believe that the 0.17 g/BHP-hr level may be overstated.
Only one particulate test was run in a test cell used just
prior for diesel testing. Thus, both measurment error and
contamination by previously stored diesel particulate could
explain the unexpectedly high results. Oil consumption could
also explain part of the particulate level, but not all of it.
It should be noted that the BSFC of this engine was better than
that of its diesel-fueled counterpart, though some advanced
features were used on the methanol-fueled enqine that are not
on the current production 6V-92TA.
Neither of these engines has yet to undergo extensive
development for emissions control. Thus, additional reductions
are possible . However, both engines appear to be newly ready
for production {i.e., two to three years). Both are already
being tested by. the State of California in in-use transit
situations in the San Francisco Bay area. The early
indications are that they are fully satisfactory as bus engines.
One problem with the use of methanol that is outside the
scope of this analysis is fuel availabili ty. Producing and
distributing sufficient quantities of methanol for all HDDEs
would be a significant challanqe. However, for urban HDD
fleets and particularly public transit fleets, the purchase and
distribution of methanol would be relatively easy due to the
use of central fuel depots. Also, the methanol market
projection is that excess capacity will be available for
sometime to come. Thus, the use of methanol to control HDDE
oarticulate in these situations could easily be realistic for
1990.
3. Summary
The above discussions indicate a large long-term potential
to reduce particulate emissions via engine-related means.
However, a high degree of uncertainty exists with respect to
the applicability of all of these techniques in the 1990
timeframe. Among the more certain effects, NOx controls
related to the 1990 standard should not negatively affect
circa-1987 particulate levels and may slightly reduce them

-------
2-105
(i.e., up to 10 percent). Reductions in RSFC caused by general
engine improvements should have a definite positive effect on
particulate emissions, reducing them by roughly 5-10 percent.
Increasing injection pressure showed significant particulate
control potential on a Mack engine (25-60 percent), but then
whether these reductions would occur on all engines is
unknown. However, some degree of particulate reduction
potential would appear fairly certain (e.g., 10-15 percent).
On the other hand, straight-forward optimization of the
combustion chamber appears to hold little promise for
substantial particulate control beyond that already considered
for 1987. The net effect of these controls should be
particulate reductions of roughly 25-35 percent, with higher
injection pressures possibly providing greater reductions.
Adiabatic technology promises great particulate reduction
potential, upwards of 90 percent. However, its widespread use
by 1990 is highly questionable. Hiqhly advanced unit injectors
also promise significant potential, but again their
availability by 1990 is uncertain. Finally, use of methanol as
fuel promises particulate control well below 0.05 g/BHP-hr and
appears practical from an engine standpoint for 1990. However,
fuel availability is an issue and except for urban fleets,
particularly transit bus fleets, is probably not practical for
1990.
Given this, it appears that the largest degree of
particulate reduction that can be firmly projected for the
entire fleet is roughly 35 percent. Starting with particulate
levels complying with a 0.6 g/BHP-hr standard, this roughly
translates to levels low enough to meet a 0.4 q/BHP-hr
particulate standard. Certainly larger reductions could be
feasible for some engines. However, even the 35 percent figure
involves a significant degree of uncertainty when the entire
HDDE fleet is concerned. Therefore, it does not appear
reasonable to project more than a 35 percent reduction for the
1990 timeframe.

-------
2-106
References
1.	Regulatory Support Document "An Updated
Assessment of the Feasibility of Trap Oxidizer", Alson, J.
and Wilcox, R., June 1983.
2.	"Diesel Particulate Study," U.S. EPA, OAR, OMS,
ECTD, SDSB.
3.-	"Light-Duty Automotive Fuel Economy... .Trends
thru 1983," Murrell, J. D. et al., SAE Paper No. 830544,
March 1983.
4. See the Preamble to this NPRM.
5.	"Draft Regulatory Analysis, Environmental Impact
Statement, and NOx Pollutant Specific Study for Proposed
Gaseous Emission Regulations for 1985 and Later Model Year
Light-Duty Trucks and 1986 and Later Model Year Heavy-Duty
Engines," (Table III-A of support document for January 21,
1981 ANPRM), U.S. EPA, OANR, OMSAPC, 1980.
6.	"Summary and Analysis of Comments on the NPRM
for Revised Gaseous Emission Regulations for 1984 and Later
Model Year Light-Duty Trucks and Heavy-Duty Engines," Issue
1, U.S. EPA, OAR, OMS, ECTD, July 1983.
7.	"Analysis Memorandum: Design Factor Update,"
prepared by Energy and Environmental Analysis, Inc., for
the U.S. Department of Energy's Office of Policy Planning
and Analysis, October 1, 1982.
8.	"Passenger Car Fuel Economy - EPA and Road," EPA
Technical Report 460/3-80-010, September, 1980, p. 52.
9.	"Low NOx Emissions from Automotive Engine
Combustion," Hansel, J.G., SAE Paper No. 40104, 1974.
10.	"Optimizing Engine Parameters with Exhaust Gas
Recirculation," Gumbleton, J., et al., SAE Paper No.
740104, 1974.
11.	"Nissan NAPS-7 Engine Realizes Better Fuel
Economy and Low NOx Emission," Harada, et al., Nisson Motor
Co., Ltd. SAE Paper No. 810010, 1981.
12.	Emissions from Combustion Engines and Their
Con trol, Patterson, D.J., and N.A., Henein, Ann Arbor
Science Publishers Inc., 1972.
13.	"Exhaust Emissions Research on Light-Duty
Engines," submission to EPA by Isuzu Motors, Ltd.

-------
2-107
References (cont'd)
14.	Letter to Mr. B. Burgess, U.S. EPA from T. F.
Watrous, General Motors Corporation, Request to Carryover
Emission Certification Data from 1982 Drivability Vehicle 2893
(C1G6.2K7ZZ4X) to 1984 Chevrolet Light-Duty Truck Engine Family
E1G6.2K7ZZ7 4 , dated January 27, 1983.
15.	"Draft Regulatory Analysis, Environmental Impact
Statement, and NOx Pollutant Specific Study for Proposed
Gaseous Emission Regulations for 1985 and Later Model Year
Liqht-Duty Trucks and 1986 and Later Model Year Heavy-Duty
Engines," U.S. EPA, OANR, OMSAPC, 1980.
16.	"Summary and Analysis of Comments on the NPRM for
Revised Gaseous Emission Regulations for 1984 and Later Model
Year Liqht-Duty Trucks and Heavy-Duty Enaines," U.S. EP*\, OAR,
OMS, ECTD, SDSB, July 1983.
17.	Public comments received and incorporated in Public
Docket No. A-81-11.
18.	"Revised Gaseous Emission Regulations for 1984
Heavy-Duty Engines and Light-Duty Trucks, Summary and Analysis
of Comments to the NPRM," U. S. EPA, OAR, OMS, ECTD, SDSB, July
1983.
19.	Public comments received and incorporated in Public
Docket No. A-80-31.
20.	"Nissan NAPS-Z Enqine Realizes Better Fuel Economy
and Low NOx Emissions," Harada, M., et al., SAE Paper No.
810010, February 1981.
21.	"The New Ford 2.3L High-Swirl-Combustion (HSC)
Engine," Lenox H., and Scussel A., SAE Paper No. 831009, June
1983 .
22.	Public comments received and incorporated in EPA
Public Docket No. A-80-18.
23.	Written comments by EMA to EPA's proposed 0.25
g/BHP-hr particulate standard for 1986, Appendix A, September
10, 1982, Public Docket A-80-18 (Document IV-D-7).
24.	"Historical and Projected Emissions Conversion
Factor and Fuel Economy for Heavy-Duty Trucks - 1962-2002,"
Energy and Environmental Analysis, Inc., for the Motor Vehicle
Manufacturers Association, December 1983.

-------
2-108
References (cont'd)
25.	"Series 92--Fuel Economy Improvements," Henriksen,
Bian S., SAE Paper No. 831202, August 1983.
25a. "Draft Regulatory Analysis, Heavy-Duty Diesel
Particulate Regulations," U.S. EPA, OANR, OMSAPC, December 23,
1980.
25b. "Development of Detroit Diesel Allison 6V-92TA
Methanol-Fueled Coach Engine," Toepel, R. , et al., SAE Paper
No. 831744, October 1983.
25c. Briefing of EPA by Caterpillar - Heavy-Duty Diesel
Particulate Control Technology, Memo from Richard A. Rykowski,
SDSB, to the Record dated November 1, 1983.
25d. "NOx Emission Controls for Heavy-Duty Vehicles:
Toward Meeting a 1986 Standard," National Research Council,
National Academy Press, 1981.
25e. "Diesel Cars: Benefits, Risks, and Public Policy,
Impacts of Diesel-Powered Light-Duty Vehicles," National
Research Council, National Academy Press, 1981.
25f. "Tacom/Cummins Adiabatic Engine Program," Bryzik, W.
and R. Kamo, SAE Paper No., 830314, February 1984.
25g. "Emissions from Direct-Injected Heavy-Duty
Methanol-Fueled Engines (One Dual-Injection and One
Spark-Ignited) and a Comparable Diesel Engine," Ullman, T. L. ,
Hare, C. T., and T. M. Baines, SAE Paper Mo. 820966, August
1982.
26.	"Trap-Oxidizer Feasibility Study", U.S. EPA, OANR,
OMS, ECTD, SDSB, February 1982, Public Docket No. A-82-32.
27.	"Trap-Oxidizer Technology for Light-Duty Diesel
Vehicles: Feasibility, Costs and Present Status", Energy and
Resource Consultants, Inc., Final Report for U.S. EPA, Contract
No. 68-01-6543, Public Docket No. A-82-32.
28.	"An Updated 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.
29.	"Mercedes Set To Try Traps For Diesels", Ward's
Engine Update, Volume 9, Number lb, August 15, 1983.

-------
2-109
References (cont'd)
30.	"Effect of Ash Accumulation on the Performance of
Diesel Exhaust Particulate Traps", R. Sachdev, v. W. Wong, S.
M. Shaked, SAE Paper No. 830182, March 1983.
31.	General Motors Comments on the 1986 NOx ANPRM and
1986 Particulate NPRM submitted to the U.S. EPA, September 13,
1982.
32.	Mack Trucks Inc. Response to the Environmental
Protection Agency's NPRM "Particulate Regulations for
Heavy-Duty Diesel Engines", Public Docket A-80-18 and the ANPRM
"Gaseous Emissions Regulations for 1985 and later Model Year
Light-Duty Trucks and 1986 and Later Model Year Heavy-Duty
Engines," Public Docket A-80-31, November 10, 1982.
33.	"Particulate Control Technology and Particulate
Standards for Heavy-Duty Diesel Engines," Weaver, C. S., SAE
Paper No. 840174, February 1984.
34.	Submission of Daimler-Benz A.G. to the U.S.
Environmental Protection Agency, September 13, 1982.
35.	Testimony on Control of Air Pollution from New Motor
Vehicles and New Motor Vehicle Engines: Particulate Regulation
for Heavy-Duty Diesel Engines (46 FR 1910, January 7, 1981) and
Gaseous Emission Regulations for 1985 and Later Model Year
Light-Duty Trucks and 1986 and Later Model Year Heavy-Duty
Engines (46 FR 5838, January 19, 1981), Cummins Engine Co.,
Inc., July 12, 1982.
36.	"Emission Characterization of a 2-Stroke ^eavv-Duty
Diesel Coach Engine and Vehicle With and Without a Particulate
Trap," Southwest Research Institute, Final Report for U.S. EPA,
Contract No. 68-03-3073.
37.	"Study of Aftertreatment and Fuel Injection
Variables for Particulate Control in Heavy-Duty Diesel
Engines," Scholl, Jackson P., et al, for EPA, Contract No.
68-03-2394, EPA-460/3-83-002, November 30, 1982.
38.	"TACOM/Cummins Adiabatic Engine Program," Bryzik,
W., and R. Kamo, SAE Paper No. 830314, February 1983.

-------
Chapter 3
ECONOMIC IMPACT
This chapter examines the economic impacts of the
compliance costs associated with the manufacturers' efforts
toward complying with the proposed light-duty truck (LDT)
nitrogen oxides (NOx) regulations for 1987 and later model
years, the heavy-duty gasoline engine (HDGE) NOx regulations,
and the heavy-duty diesel engine (HDDE) NOx and particulate
regulations, applicable to the 1987 and 1990 model years.
Fixed or pre-production costs (research, development, and
testing (RD&T) , including certification testing) and variable
costs (emission control hardware) will be considered.
Certification costs are treated as a one-time fixed cost
occurring prior to the first model year in which the standards
are applicable, with subsequent model year certification being
based on data carryover. The estimates of emission control
hardware costs given in this chapter are for the systems
expected to be used in complying with the regulations, as
described in Chapter 2, Technological Feasibility. Possible
changes in operating costs will also be evaluated.
The analysis is divided into two general sections. These
sections address costs and socioeconomic impacts. The first
section of this chapter identifies the costs to manufacturers
and users and is subdivided by vehicle class and regulated
pollutant. The socioeconomic section of the analysis addresses
the effects of the costs on manufacturers, users, energy,
inflation, balance of trade and local and regional effects.
I. Costs
This portion of Chapter 3 presents the analysis of the
potential cost impact (in 1984 dollars) of complying with the
NOx and particulate emission standards, and will cover
research, development and testing costs (fixed costs), emission
control hardware costs (variable costs) and operating costs
(fuel and maintenance). Work in the areas of RD&T is subject
to unforeseen problems and delays which can cause the cost to
be higher than originally expected. A 10 percent contingency
factor has, therefore, been included throughout this analysis
in all estimates of RD&T costs.
A. Light-Duty Trucks (LDTs) - NOx Standard
1. Cost to LDT Manufacturers of NOx Standard
a. pixed Costs (Research, Development and Testing)
Light-Duty Gasoline Trucks (LDGTs): The proposed LDT NOx
emission standards are comparable in stringency to the NOx

-------
3-2
standard for light-datv vehicles (LDVs). As detailed in
Chapter 2 it is reasonable to expect that similar technology
will, therefore, be employed, i.e., reduction catalyst and
exhaust gas recirculation technologies will predominate.
Because similar technology is expected to be used, it is
reasonable to expect that there will be little or no need for
research. It is inevitable, however, that some design,
development and testing costs will be incurred. In the case of
exhaust qas recirculation (EGR), costs will be limited to those
associated with system recalibrations, because EGR is already
applied to all 1984 model year LDGTs.
Design, development and testing work would consist of the
tasks to cover: 1) recal ibrat ions of the EGR systems, 2)
changes to the exhaust manifold to accommodate the installation
of an O2 sensor, electronic control module installation and
wiring harness modifications where closed loop systems are
employed for the first time, 3) modifications to the exhaust
system to accomodate the new catalyst and necessary chanqes to
the secondary air plumbing, and 4) testing to establish the
mechanical integrity of the systems as well as certification
testing. The costs for these tasks are estimated to total
$19.77 million as detailed below.
Fifty-four LDGT engine families were certified for. the
1984 model year with an average of approximately eight system
calibrations per family. It is reasonable, however, to expect
that there would be fewer EGR calibrations than total system
calibrations. Four EGR calibrations per engine family were,
therefore, chosen for purposes of this cost analysis. In
developing each final EGR calibration it is reasonable that
manufacturers will evaluate two calibrations and use the most
effective calibration. While some engine families may not
require cecalibration, the conservative approach used was to
assume that all engine families would require this work. The
number of enqine families and calibrations per enqine family
was, therefore, assumed to remain constant. The cost for EGR
recalibrations will be about $2.85 million based on fifty-four
enqine families and the evaluation of eiqht calibrations per
engine family, three weeks of effort per calibration, $50 per
hour to cover labor, overhead and parts and a ten percent
contingency factor.
The second set of tasks cover the redesign of exhaust
manifolds and the addition of electronic control modules. The
estimates given in Chapter 2 for the catalyst technology mix
which will be employed in 1987 were; between 0.7 and 3.7
percent open loop three-way catalyst, between one and ten
percent oxidation catalyst and between 86 and 98 percent closed

-------
3-3
loop three-way catalyst systems. For this cost analysis, the
catalyst technology mix will be assumed to be two percent open
loop three-way catalysts, five percent oxidation catalysts and
93 percent closed loop three-way catalysts. It was also shown
in ChaDter 2 that the catalyst technology mix by vehicle sales
in 1984 is expected to be about 42 percent three-way catalyst
(31 percent closed loop three-way, 11 percent open loop
three-way) and 58 percent oxidation catalyst. On an engine
family basis, the catalyst technology mix is 43 percent
three-way catalysts and 57 percent oxidation catalysts, i.e.,
essentially the same as that determined on the basis of vehicle
sales. Assuming that the present catalyst technology mix
carries through 1986, 38 percent of the vehicles will not
require the application of new catalyst technology in 1987.
3ased upon EPA certification records, it is estimated that
one-half of the remaining 62 percent can use California
specification systems, so that design work, exclusive of
exhaust systems, will be required for approximately 31 percent
of the vehicles or engine families. With a level of effort of
four person-months per family redesign, excluding exhaust
system redesign, and $50 per hour to cover personnel, test part
procurement and overhead plus a 10 percent contingency factor
the cost is approximately $1.23 million.
The cost for the third task, i.e., exhaust system
redesign, was develooed by identifying the number of exhaust
system designs required to accommodate such vehicle differences
as wheel base, number of driving wheels, type of transmission
and vehicle body configuration (van vs. pick-up for example).
For the 1983 model year, 253 probably exhaust system designs
were identified. Applying the 31 percent redesign requirement
which was previously developed, 4 person-months per redesign
and an hourly cost of $50, this cost totals to approximately
$3.00 million including a ten percent contingency factor.
The fourth development cost component is the cost to prove
the mechanical integrity of the exhaust systems. This cost is
approximately $2.59 million based on 60,000 miles of testing at
30 mph average speed, $30 per hour (labor and overhead costs
are lower for mileage accumulation than for design) and test
cost shared equally with other mechanical durability testinq
programs plus a ten percent contingency factor.
The final component of the fixed costs is the cost of
certification testing. This cost is estimated to be
approximately $10.10 million based on the continued production
of 54 engine families. The basis for the certification cost
estimate per engine family is one durability vehicle operated
for 120,000 miles and two data vehicles operated for 4,000
miles each at average speeds of 30 mph with an hourly cost of

-------
3-4
$30 for labor, vehicles, fuel, etc. and 28 emission tests per
family at a cost of $1,500 each, plus a ten percent contingency
factor ($187,000/enqine family).
In summary, the total fixed cost of $19.77 million would
cover the application and optimization of reducing catalyst
technology, exhaust gas recirculation (EGR), and required
electronic controls ($9.67 million) plus certification to
establish compliance with the standards ($10.1 million).
If it is assumed that these costs are spread over the five
model years following introduction of the standards, the
initial effort would represent expenditures of approximately
$0.74 for design and development and $0.77 for certification
per gasoline LDT expected to be produced during those model
years (Table 3-1) . The sales projections contained in Table
3-1 were derived from references 1, 2, and 3 and from
manufacturers comments in response to an NPRM published on
January 13, 1982 (Revised Gaseous Emission Regulations for 1984
and Later Model Year Light-Duty Trucks and Heavy-Duty Enqines).
Much of the practical application work ($9.67 million)
would be accomplished in 1985 with the longer leadtime portions
of the Certification testing program being initiated in 1985.
The final year, 1986, would involve continued optimization,
driveability, fuel economy testing, and completion of the
Certification testing. On this basis, the fixed costs are
apportioned over the two years which the manufacturers and
vendors will have available for development, optimization, and
testing according to the following schedule:
Non-Certification Certification	Total
Light-Duty Diesel Trucks (LDDTs): As was shown in Chapter
2 (Technological Feasibility) , it is expected that EGR will be
the technology used on LDDTs in meeting the standards. One
manufacturer (GM) is presently using (1984 model year) EGR on
its engines and electronic controls are expected to be applied
to these engines for compliance with the 1987 NOx standard. As
was shown in Chapter 2 these engines account for approximately
60 percent of sales. Design and development costs are
estimated to be approximately $38,000. This cost is based on
four person months of effort at $50 per hour to cover
personnel, overhead, and materials for three calibrations and a
ten percent contingency factor. Little effort is expected to
be necessary for the development of experimental and production
Costs
Costs
Fixed Costs for LDGTs
1985	$7.0 million
1986	$2.67 million
$ 1.0 million $ 8.0 million (40%)
$ 9.1 million $11.77 million (60%)

-------
3-5
Table 3-1
Projected Liqht-Duty Truck
and Heavy-Dutv Engine Sales (in thousands)*

Liaht-Duty
Trucks

Year
Gasoline
Diesel
Total
1987
2 ,680
710
3,390
1988
2,670
820
3,490
1989
2,620
940
3,560
1990
2,son
1,040
3,540
1991
2,750
1,100
3,670
Totals
13,040
4,610
17,650

Heavv-Duty
Engines

Year
Gasoline
Diesel
Total
1987
385
351
736
1988
382
366
748
1989
382
384
766
1990
371
392
763
1991
367
41 1
778
1992
364
435
799
Totals
2,251
2,339
4,590
Total sales and distribution between qasoline and diesel
enaines derived from References 1, 2, and 3 and from
manufacturers comments in response to NPRM published
January 13, 1982 (Revised Gaseous Emission Regulations for
1984 and Later Model Year Light-Dutv Trucks and Heavy-Dutv
Engines).

-------
3-6
hardware because the necessary
already in use on counterpart
standards.
electronic components are
engines meeting California
Application of EGR to engines which are presently not
equipped with this system (approximately 40 percent of sales),
can reasonably be expected to incur RD&T costs for redesign of
air intake and exhaust manifolds, EGR valve and actuator
redesign (assuming that EGR valves already in production for
gasoline fueled LDTs form the basis for the design), design of
plumbing which will transfer the exhaust gases to the EGR
control valve and testing to establish the mechanical
durability of the system. The cost for this work per engine
family is expected to be approximately $90,000 based on six
person months for redesign of the manifolds and the EGR valve
and actuator, design of the plumbing plus one-half of the cost
of accumulating 60,000 miles for confirming mechanical
integrity and a ten percent contingency factor. At present, 24
engine families do not use EGR and one engine family does. The
cost for the design and development work for the 24 engine
families which are expected to use EGR for the first time is
approximately $2.16 million. The total design and development
cost is approximately $2.2 million when the GM engine familv is
included. When spread over the projected total diesel LDT
production (Table 3-1) for the five model years following the
introduction of the standards the expenditure per vehicle is
approximately $0.48. The final component of the pre-production
fixed cost is the cost of certification. This cost was
estimated to be $187,000 per engine family for gasoline LDTs.
The cost for certification for diesel LDTs will be sliahtly
higher than that for gasoline LDTs because of the requirement
for measurement of particulate emissions. A reasonable cost
per diesel LDT family for certification is, therefore,
$200,000. Total costs for certification of 25 diesel engine
families is therefore $5.0 million.
Summing the component parts of the fixed costs for LDDTs
results in an overall fixed cost of $7.2 million. Using the
same overall percentages as were used for LDGTs for allocating
the costs between 1985 and 1986 results in the following:
Non-Certification
Costs
Certification
Costs
Total
Fixed Costs for LDDTs
1985
1986
$2.0
$0.2
million
mi 11 ion
$0.9
$4.1
million
million
$2.9
$4.3
million
million
(40%)
(60%)
These RD&T costs,	together with the RD&T costs for LDGTs, are
carried forward	and will be combined with the estimated
hardware costs to	develop the total estimates of manufacturers'
costs.

-------
3-7
b. Emission Control System Hardware Costs (variable
cost)
Light-Duty Gasoline Trucks; As was shown in Chapter 2, it
is reasonable to expect that the proposed standards can be met
by all LDGTs, but that a variety of emission control strategies
could be used based on manufacturer preference and on the
emission characteristics of each LDGT family. An estimate of
the increase in hardware costs can be developed based on past
trends in the apolication of technology, present technology
applications and the technology projections developed in
Chapter 2. In developing the cost estimate it is necessary
that the overall LDGT application rates for close-loop and
open-loop systems be developed as well as the application rates
for three-way systems and three-way plus oxidation catalyst
systems. In addition to these overall LDGT technology
application characteristics, identification of the catalyst
size and loading is required. This was done by dividing the
LDGTs into three representative engine sizes based on the
number of cylinders in the engine. The estimates are developed
below.
In the 1982 and 1983 model years, oxidation catalyst
technology was used on approximately 85 percent of the
light-duty gasoline trucks expected to be sold under Federal
emission standards. Non-catalyst and three-way catalyst
technologies were used on the remaining 15 percent of the
LDGTs. Technology application for LDGTs was distributed
approximately as shown in Table 3-2. While not shown in Table
3-2, the ratio between closed-loop three-way-plus-oxidation
catalyst and closed-loop three-way catalyst systems was
approximately	8:1.	Open-loop	three-way	and
three-way-plus-oxidation catalyst systems were essentially not
used. The trend for technology application between the 1982
and 1983 model years was away from non-catalyst and oxidation
catalyst systems and to three-way catalyst systems (Table 3-2).
In the 1983 model year, the ratio between closed-loop
three-way-plus-oxidation catalyst and three-way catalyst
technology for light-duty vehicles (LDVs) was approximately
3:2. This ratio is significantly lower than the 8:1 ratio
which applied to LDTs in the 1982 and 1983 model years.
Because of the generally lower cost of three-way catalysts
relative to three-way plus oxidation catalysts, it will be
assumed that the 3:2 ratio will be representative of
application rates on future LDTs when the LDV and LDT standards
will be essentially equal in stringency.

-------
3-8
Table 3-2
Liqht-Dutv Trucks,
Percent Technology Usage by Model Years for all LDGTs
Percent Technology Used
No	Oxidation	Open-Loop Closed-Loop
Model Year	Catalyst	Catalyst	3-wav	3-Wav
1982*	4.2	91.2	0.2	4.4
1983*	1.4	80.3	0	18.3
1984*	0	58.0	11.0	31.0
1985 & 1986	0	58.0	11.0	31.0
(orojected)**
1987 and later	0	5.0	2.0	93.0
(projected)
Based on data provided by manufacturers as part of the
certification procedure.
If the increase in the usage of three-way systems which
occurred between the 1982 model year and the 1984 model
vear were to continue to the 1986 model vear, the
projected catalyst technology utilization rate in the 1986
model year would be about 22 percent oxidation catalyst,
20 oercent open-loop three-way and 58 percent closed-looo
three-way.

-------
3-9
Beginning with the 1984 model year, the stringency of the
hydrocarbon (HC) and carbon monoxide (CO) standards for LDTs
was increased without any change in the NOx standards. As was
shown in Chapter 2 information provided by manufacturers for
the purpose of 1984 model year certification shows that the
sales split by technology is expected to be approximately 31
percent closed-loop three-way, 11 percent open-loop three-way
and 58 percent oxidation catalyst. These sales splits by
technology are presented in Table 3-2. It should be noted that
the trend away from oxidation catalyst systems which occurred
in the 1982 and 1983 model years has carried on into the 1984
model year. Because of the lower cost of oxidation catalyst
systems relative to three-way systems, it is reasonable to
expect that oxidation catalyst systems will continue to
predominate until the 1987 model year.*
In summary, it is likely that the technology application
rates in the 1986 model year LDGT fleet will be approximately
as follows: 58 percent oxidation catalyst, 11 percent
open-loop three-way catalyst, 20 percent closed-loop three-way
plus oxidation catalyst and 11 percent closed-loop three-way
catalyst. These percentages form the starting point from which
the hardware cost increase will be calculated.
The previously developed estimate (Chapter 2) of the
catalyst technology mix for 1987 did not distinguish between
three-way plus oxidation catalyst systems and three-way
systems. Applying the 3:2 ratio for these technologies results
in an overall technology mix in 1987 of 5 percent oxidation
catalyst, two percent open-loop three-way catalyst, 56 percent
closed-loop three-way plus oxidation catalyst and 37 percent
closed-loop three-way catalyst. As was noted previously, 38
percent of the LDGTs will not be changed between 1986 and
1987. Changes in catalyst technology would be apportioned as
follows: 36 percent would be converted from oxidation catalyst
systems to closed-loop three-way plus oxidation catalyst
systems, 17 percent would be converted from oxidation catalyst
to closed-loop three-way catalyst systems and nine percent
would be converted from open-loop to closed-loop three-way
catalyst systems. These percentile changes in technology
application on a fleet-wide basis are converted to a fleet cost
through the development of technology cost estimates for
different size engines and estimates of sales by engine size.
If the increase in the usage of three-way technology which
occurred between the 1982 model year and the 1984 model
year were to continue to the 1986 model year, the
projected catalyst technology utilization rate in the 1986
model year would be about 22 percent oxidation catalyst,
20 percent open-loop three-way and 58 percent closed-loop
three-way.

-------
3-10
In preparing the estimate, LDGTs were divided into
subclasses based on the number of cylinders in the engine. Use
of these subclasses permits development of representative
engine sizes for each subclass and the subsequent development
of the costs of the catalysts which are expected to be
employed. Certification records for the 1980-84 model years
formed the basis for the follwing size versus number of
cylinder groups. In the case of 4-cylinder engines, the
displacements ranged between 1.8 and 2.4 liters with the
majority being in the 2.3 to 2.4-liter size. A representative
size of 145 cu.in. (2.375 liter) was chosen. The
representative size for 6-cylinder engines which was selected
to cover their range of 2.8 to 4.9 liters was 240 cubic inches
(3.9 liter). Eight-cylinder engines were 5 liters and larger
and the representative size selected was 325 cubic inches (5.3
liter).
burning to the issue of future sales splits, it is
expected that a general downsizing of LDTs will occur over the
next several years, due primarily to fuel economy pressures,
with emissions as a secondary consideration. In addition to
the introduction of more fuel efficient truck lines with
smaller engines, domestic manufacturers can be expected to
adjust their sales mixes such that they sell fewer of their
large CID engines, and more of the smaller, more fuel efficient
engines. The 1980, 1982, 1983, and 1984 sales orojection data
submitted by the manufacturers for certification, broken into
the three cylinder and CID groups, is shown below. The shift
from 8-cylinder to 4-cylinder and 6-cylinder engines is
expected to continue and was used in developing the 1987-91
sales projections shown below.
Sales Splits - Gasoline Engines
Approximate Market Percentages


Manufacturer-
Furnish
ed
EPA Market
Number of
Engine

Data


Projection
Cylinders
CID Range
1980
1982
1983
1984
(1987-91)
4
under 170
11%
22%
41%
30%
40%
6
170-300
19%
44%
34%
42%
40%
8
over 300
70%
34%
25%
28%
20%
The past, present and projected technology applications by
engine size are shown in Table 3-3. The application rates by
engine size for the 1982 through 1984 model years are based on
manufacturer provided sales projections. The EPA projections
for 1986 are based on the 1982 through 1984 model year trends
and the assumption that technology application rates will be
stable from the 1984 through 1986 model years. Projections for

-------
3-11
Table 3-3
Liqht-Duty Trucks, Percent
Technoloav Usaqe in Model Years by
Number of Cylinders for Gasoline-Fueled Vehicles*
Percent Technology Used



Open Loop
Closed Loop

No Cat
Ox Cat
3-Way
3-Way 3-Wav + Ox
4 Cvlinder




1982MY
8.7
87.6
1.0
2.7 0
1983MY
0
74.9
0
1.5 23.6
1984MY
0
74.5
1.2
1.4 22.9
1986MY, Projected
0
75
1
1 23
87+MY, Projected
0
5
1
37 57
6 Cylinder




1982MY
0
92.5
0
3.3 4.2
1983MY
4.2
8 6.3
0
8.7 0.8
1984MY
0
49.5
0
20.2 30.3
1986MY, Projected
0
50
0
20 30
87+MY, Projected
0
5
0
38 57
8 Cylinder




1982MY
6.7
91.8
0
0 1.5
1983MY
0
80.9
0
0 19.1
1984MY
0
54.5
39. 5
0 6.0
1Q86MY, Projected
0
55
30
0 15
87+MY, Proiected
0
5
8
35 52
Fleet Average




1987, Projected
0
5
2
37 56
* Percentages for 1982 through 1984 model years derived from
data provided by manufacturers as part of the
Certification orocedure.

-------
3-12
1987 were based on the overall fleet technology rates as
developed in Chapter 2 and the trends for each size of engine
for the years 1982 through 1986. In the case of oxidation
catalysts it was assumed that the five percent overall fleet
application rate would be achieved by an equal application rate
of five percent on each engine size class. For the open loop
three-way systems it was assumed that the one percent
application rate for 4-cylinder engines would continue and that
these systems would not be introduced on 6-cylinder engines as
indicated by the previous trend. Achieving an overall fleet
average installation rate of two percent necessitated,
therefore, that eight percent of the 8-cylinder engines use
open loop three-way systems in 1987. For the closed loop
systems, the percentile splits between three-way and
three-way-plus-oxidation catalyst systems were derived on the
basis of the overall fleet installation rates, the sales
distribution between the three engine cylinder groups and the
previously identified 3:2 ratio between three-way-plus-
oxidation catalyst systems and three-way systems.
On the basis of the sales and technology rate projections
for the 1986 and 1987 model years fleets, conversions from
oxidation catalysts to closed-loop three-way or closed-loop
three-way-plus-oxidation catalyst will involve 70 percent of
the 4-cylinder LDGTs, 45 percent of the 6-cylinder LDGTs, and
50 percent of the 8-cylinder LDGTs. Conversions from open-loop
to closed-loop three-way systems will involve 22 percent of the
8 cylinder LDGTs. applying the 3:2 ratio previously identified
for closed-loop three-way plus oxidation catalyst to
closed-loop three-way catalyst usage results in the application
rates for these technologies as shown in Table 3-3. These
percentages, coupled with the costs for each technology, will
permit the development of an average cost once the cost
differentials for 4-, 6-, and 8-cylinder LDTs are estimated.
Certification records for the 1983 and 1984 model years
were used to develop estimates of catalyst size as a function
of engine size as well as estimates of catalyst loading as a
function of catalyst type. This information was used in
conjunction with the cost estimation procedures given in
Reference 4 to develop the catalyst cost estimates shown in
Tables 3-4 through 3-6.
Tables 3-4, 3-5, and 3-6 show the estimated increase in
hardware costs for each configuration applicable to 4-, 6-, or
8-cylinder engines. The credits taken are for the emission
control system which are expected to be used to comply with the
emission standards through the 1986 model year. As can be seen
in Tables 3-4 through 3-6, the hardware cost estimates range
from a low of $65 for 8-cylinder engines where closed-loop

-------
Table 3-4
Light-EXjty Gasoline Truck, Qnission
Control System Cost, 4-Cylinder 145-CID Engines[4, 5 ] *
Hardware Added	Hardware Removed
Three-Way-	Feedback
Three-Way Plus-Oxidation Carburetor	Closed-Loop	Oxidation	Air	Open-Loop
Configuration CStalyst Catalyst	Modifications	Control	Catalyst	Injection	EGR Cbntrol Total
1	4111 -	$11	$126	$70	-	- $72 $106
2	- $134	$11	$126	$70	-	- $72 $129
Cost Estimates developed from Reference 1 and account for inflation using new car price inflation indices
provided by the Bureau of Labor statistics for 1978, 1979, 1980, 1981, 1982, and 1983 of 6.2 percent, 7.4
percent, 7.5 percent, 6.8 percent, 1.6 percent and approximately 3.6 percent respectively.

-------
Table 3-5
Light-Duty Gasoline Truck, Emission
Control System Cost, 6-Cylinder 240 CID Engines[4,5]*
Hardware Added	Hardware Removed
Three-Way-	Feedback
Three-Way Plus-Oxidation	Carburetor	Closed-Loop	Oxidation	Air	Open-Loop
Configuration Catalyst Catalyst	Modifications	Control	Catalyst	Injection	EGR Control Tbtal
1	$169 -	$11	$126	$101	-	- $72 $133
2	- $216	$11	$126	$101	-	- $72 $180
Cost Estimates developed from Reference 1 and account for inflation using new car price inflation indices
provided by the Bureau of Labor statistics for 1978, 1979, 1980, 1981, 1982, and 1983 of 6.2 percent, 7.4
percent, 7.5 percent, 6.8 percent, 1.6 percent, and approximately 3.6 percent respectively.

-------
Table 3-6
Light-Duty Gasoline Truck, Emission
Control System Cost, 8-Cylinder 325 CID Engmes[4,5]*


Hardware Added


Hardware
Removed

Configuration
Three-Way
Catalyst
Three-Way-
Plus -Oxidation
Catalyst
Feedback
Carburetor
Modifications
Closed-Loop
Control
Oxidation
Catalyst
Air
Injection
Open-Loop
EGR Control
Total
1
$221
-
$11
$126
$129
-
$72
$157
2
-
$289
$11
$126
$129
-
$72
$225
3
_
—
$11
$126
—

$72
$65
Cost Estimates developed from Reference 1 and account for inflation using new car price inflation indices
provided by the Bureau of Labor statistics for 1978, 1979, 1980, 1981, 1982, and 1983 of 6.2 percent, 7.4
percent, 7.5 percent, 6.8 percent, 1.6 percent, and approximately 3.6 percent respectively.

-------
3-16
three-way catalyst systems replace open-loop three-way catalyst
systems to $225 for 8-cylinder engines where oxidation
catalysts are replaced by closed-loop three-way plus oxidation
catalyst systems.
Combining the hardware costs and sales projections which
have just been developed permits the development of an average
hardware cost for LDGTs as follows:
0 Forty percent of sales will be 4-cylinder engines. Of this
total, 70 percent will utilize new technology with 36
percent being converted to closed-loop three-way catalyst
systems and 34 percent being converted to closed-lop
three-way plus oxidation catalyst systems. The average
hardware cost increase for a 4-cylinder truck which
receives new catalyst technology is, therefore, [($106 x
0.36) + ($129 x 0.34)1/0.70 = $117.17. The contribution by
4-cylinder LDGTs to the average hardware cost for all LDGTs
is:
0.4 x 0.7 x $117.17 = $32.31.
° In the case of 6-cylinder engines, 45 percent will use new
technology with 18 percent receiving closed-loop three-way
systems and 27 percent receiving closed-loop three-way plus
oxidation catalyst systems. The average hardware cost
increase for the new systems will be: [($133 x 0.18) +
($180 x 0.27)]/0.45 = $161.20. With 6-cylinder sales
representing 40 percent of total sales, the contribution by
6-cylinder LDGTs to the average hardware cost for all LDGTs
is:
0.4 x 0.45 x $161.20 = $29.02
0 Eight-cylinder engines represent 20 percent of sales, with
50 percent of the 8-cylinder engines being changed from
oxidation catalyst technology to closed-loop systems. An
additional 22 percent will be changed from open-loop to
closed-loop three-way systems. The average hardware cost
increase for new systems on 8-cylinder trucks will,
therefore, be: [($157 x 0.13) + ($225 x 0.37) + ($65 x
0.22)1/0.72 = $163.83. The contribution by 8-cylinder
engines to the average hardware cost of the LDGT fleet is:
0.20 x 0.72 x $163.83 = $23.59

-------
3-17
0 The average hardware cost per LDGT in the fleet is the sum
of the contributions by the three groups of engines and is:
$32.81 + $29.02 + $23.59 = $85.42 or approximately $85*
These costs will be carried forward in developing estimates of
the average hardware costs for all LDTs.
The average hardware cost for LDGTs which require new
hardware is calculated using the same procedure as was employed
.above but without distributing the cost over those vehicles
which do not require new technology. This cost is $137. 77 or
approximately $138.
Light-Duty Diesel Trucks: The hardware cost for LDDTs
would come from the application of non-electronically
controlled EGR to 40 percent of the vehicles and the conversion
of 60 percent of the vehicles to electronically controlled
EGR. The estimated cost of the EGR system which is expected to
be added to LDDTs not previously equipped with EGR is $20. [4]
The hardware cost for the addition of electronic controls is
estimated using the system design employed by GM on their
engines designed to comply with the California standard. In
Reference 1, EPA estimated the cost of an electronic control
module for use on LDDTs as being $19. The cost of a pressure
sensor for determining particulate trap backpressure was
estimated as being $2. These costs will be used as the basis
for the estimate of the cost for electronic controls for LDDT
EGR systems. Sensors which will be employed as part of the
electronic EGR control system are expected to monitor engine
speed, load, and absolute pressure. The load sensor (throttle
position) is expected to be a proportional unit replacing an
on/off unit used on the non-electronic systems. The
incremental cost can be expected, therefore, to be about $3.
The engine speed sensor will be required to deliver a
proportional signal and can be an electrical pulse generator.
The cost is estimated to be $3. The cost of the absolute
pressure sensor is estimated to be $10 for an aneroid and the
A hardware cost of $42, developed as follows, could be
projected if the trend in the use of three-way technology
which occurred between the 1982 model year and the 1984
model year were to continue to the 1986 model year:
Hardware Cost = Seventeen percent of LDTs are changed from
oxidation to closed-loop three-way catalyst
at an average cost of $178 per vehicle
($129 + $180 + $225)/3 plus 18 percent of
LDTs are changed from open-loop to
closed-loop systems at a cost of $65 per
vehicle.

-------
3-18
electrical position signal generator. The final major
component of the electronic control system is the actuator
which will modulate the EGR valve as directed by the control
module. On the California specification engines, GM controls
the EGR flow by the use of a pulsed solenoid. The cost for
this component is estimated to be $5. The remainder of the
system consists of electrical connectors, a filter, and small
vacuum liner. The cost of these components is estimated to be
$2. Summing the costs of the component parts of the elctronic
control system gives the estimate of the cost of the total
system as follows:
$19 for the control module + 1 sensor at an incremental
cost of $3 + 2 sensors at a total cost of $13 + $2 for the
connectors, etc* + $5 for the actuator for a total of $42.
Combininq the cost estimates and the sales split between
the first time application of EGR and the conversion to
electronic EGR systems results in an average hardware cost for
LDDTs of:
0.60($42) + G.40($20) = $33.20, or approximately $33
These costs will be carried forward to develop estimates of the
hardware cost for all LDTs.
LPT Fleetwide Average Emission Control Hardware Costs:
Having estimated control system costs for gasoline and
diesel-powered LDTs, all that remains is to determine the
fleetwide average cost. hs shown in Table 3-1, sales of LDTs
in the 5-year period, 1987-91, are expected to consist of
13,040,000 gasoline-powered vehicles and 4,610,000
diesel-powered vehicles. Gasoline-powered vehicles are
expected, therefore, to represent 73.9 percent of sales and
diesel-powered vehicles to represent 26.1 percent of sales.
Weighting the average hardware costs for gasoline-powered
vehicles ($35.42) and diesel-powered vehicles ($33.20) by theic
respective sales fractions results in an average hardware cost
for all LDTs of approximately $72 as shown below:
0.261 ($33.20) + 0.739 {$85.42) = $71.79 or approximately
$72
c. Total Costs to Manufacturers
The major costs to manufacturers of RD&T and emission
control hardware are summarized in Table 3-7. The costs
presented are shown in the year in which they are expected to
occur, and include profit and overhead. The total undiscounted
cost, $1.29 billion shown in Table 3-7, should provide
sufficient funds to comply with all aspects of this proposed
rulemaking action. The discounted value ($1.08 billion) is
shown in Table 3-8. The costs are also shown in Tables 3-9

-------
3-19
Table 3-7
Liqht-Duty Trucks, Total
Cost (undiscounted) to Manufacturers
of LDTs Manufactured in 1987-91 Model Years
Year	RD&T	Emissions Hardware
1985	S10.90M
1986	16.07M
1987	-	$243.37M
1988	-	250.55M
1989	-	255. 57M
1990	-	254. 14M
1991	-	263.47M
$2 6.97M	$1,26 7.10M
Total undiscounted cost: $1,294.07M (1984 dollars)

-------
3-20
Table 3-8
Light-Duty Trucks,
Present Value in 1987 of the
Costs of Compliance for the 1987-91 Model Years
Caoital and Emission Control Systems:
Present Value
Year[11
Cost[21
in 1987 [31
1985
$ 10 , 900K
$ 13,190K
1985
16 ,070K
17,680K
1987
243 , 370K
243 , 370K
1988
250,550K
227,770K
1989
255,570K
211,215K
1990
2 54,140k
190,940K
1991
263, 470K
179,955K
Total:	$1 , 294 , 070K	$1,084 , 120K
Operating Costs for Fuel:
Chanae in cost per	+$855,592K
1 percent chanae
in fuel economv
[1] Costs are assumed to occur at the start of each year and
are allocated according to Table 3-7.
r 21 1984 dollars, includinq profit and overhead.
[3] 10 percent discount rate.

-------
3-21
Table 3-9
Light-Duty Trucks, Aggregate Costs of
Compliance for LDTs Produced During Model Years
1987-91 (discounted at 10% to January 1, 1987)
Development and Testinq	$30,870K
Emission Control Systems	1,053,250K
$1,084,120k
Operating Costs
Chanqe in aqqregate cost	+$855.6M
per one percent chanqe
in fleetwide fuel economy

-------
3-22
through 3-12 expressed as discounted total costs foe RD&T and
hardware, discounted and undiscounted costs per vehicle and as
RD&T costs per fuel type.
2. Cost to Users (LDT-NOx)
a. Increases in First Costs
The added cost to manufacturer for RD&T and emission
control system hardware is expected co be passed on to the
purchasers of LDTs. The amount a manufacturer must increase
the price of its vehicles to recover its expenses depends on
the timing of the costs, the revenues from sales, and on the
cost of capital to the manufacturer. Table 3-7 showed the
manner in which the manufacturers' costs are distributed over
the period 1985-91. It is expected that manufacturers face a
10 percent cost of capital over the long run, and that they
increase the vehicle prices to recover their pre-production
investment in five model years, 1987-91. The first price
increase of a vehicle would be the sum of the discounted RD&T
costs allocated on a per vehicle basis plus the cost of the
hardware.
The first price increase for LDTs in 1987 is shown in
Table 3-13 in terms of vehicles which employ new technology and
as average price increases. The expected sales-weighted
average first price increase for LDTs requiring the application
of new technology is approximately $140 for LDGTs and $35 for
LDDTs. The sales weighted average price increase for all LDTs
is approximately $74. The first price increase for LDTs using
new technology would consist of development costs (discounted)
plus hardware costs and would range in the case of
gasoline-powered LDTs from a low of approximately $67 for
8-cylinder LDGTs which are converted from open-loop to
closed-loop three-way catalyst systems, to a high of $227 for
8-cylinder LDGTs which are converted from oxidation catalyst
systems to three-way-plus-oxidation catalyst systems. LDDTs
are expected to increase in price by an average of $35.
b. Fuel Economy
As was shown in Chapter 2, it is reasonable to expect that
some small increase could occur in the fuel economy of some
LDGTs and some small decrease could occur in the fuel economy
of some LDDTs. The potential fuel economy improvements on
LDGTs could be available in a couple of areas. The first area
of potential savings is related to the use of electronic engine
controls and feedback carburetors (closed-loop systems). These
components would allow the air/fuel ratio to be maintained
nearer to stoichiometric conditions, and would allow better use

-------
3-23
Table 3-10
Liqht-Duty Trucks,
Undiscounted Cost of Compliance per LDT
when Distributed Over All LDTs Produced During 1987-91
Development and Testing	$ 1.52
($1.51 for LDGT, $1.56 for
LDDT)
Emission Control Hardware	72.00
(sales-weiahted averaqe)
Undiscounted Cost per Vehicle	$73.52
Operating Cost
Chanqe in operating	+$ 74.63
cost per one percent
change in fuel economy
(non-discounted)

-------
3-24
Table 3-11
Light-Duty Trucks, Discounted Cost of
Compliance per LDT[11 Cor LDTs Produced During 1987-91
Development and Testing	$ 1.75
($1.74 for LDGT, $1.79 for
LDDT)
Emission Control Hardware	59. 67
(sales weighted average)
Discounted Cost Per Vehicle	$61.42
Operating Cost
Change in operating	+ $48.48
cost per one percent
change in fuel economy
[1] Discounted to 1987 at 10 percent,
I

-------
3-25
Table 3-12
Light-Duty Trucks, Pre-1987 RD&T Costs
	(1984 dollars undiscounted)	
LDGT	$ 19,770K
LDDT	7 ,200K
Vendor:
Gasoline-powered LDT:	126,390K
(tooling and equipment for
catalytic converters, closed
loop electronic controls, and
other emission-related hardware) 	
$153,360K

-------
3-26
Table 3-13
Light-Duty Trucks, First Price
Increases for Trucks with New Technology
and Sales Weighted Averages (1984 dollars)
4 Cylinder Closed-Loop Three-Way Catalyst	$108
from Oxidation Catalyst
4 Cylinder Closed-Loop Three-Way Plus	131
Oxidation Catalyst from Oxidation Catalyst
6 Cylinder Closed-Loop Three-Way Catalyst	135
from Oxidation Cataiyst
6 Cylinder Closed-Loop Three-Way-Plus-	182
Oxidation-Catalyst from Oxidation Catalyst
8 Cylinder from Ooen-Loop to Closed-Loop	67
Three-Way Catalyst
8 Cylinder Closed-Loop Three-Way Catalyst	159
from Oxidation Catalyst
8 Cylinder Closed-Loop Three-Way-Plus-	227
Oxidation Catalyst from Oxidation Catalyst
Light-Duty Diesel Trucks	35
Sales-Weighted Average price increase for
LDGTs recruiring new technology is approximately 140
Sales-Weighted Average price increase for all
LDGTs is approximately	87*
Sales-Weighted Average price increase for LDDTs
requiring new technology is approximately	35
Sales-Weighted Average price increase for all
LDTs is approximately	74**
$44 if the trend in the use of three-way technology which
occurred between the 1982 model year and the 1984	model
year were to continue to the 1986 model year.
$42 if the trend in the use of three-way technology which
occurred between the 1982 model year and the 1984	model
year were to continue to the 1986 model year.

-------
3-27
of the energy available in the fuel. A second potential fuel
economy improvement lies in the substitution of "pulse air" air
injection for air pump supplied air injection in some LDGTs.
The fuel savings attributable to these changes could be in the
two percent to four percent range (Chapter 2) . It is very
difficult, however, to estimate the number of vehicles which
could be involved in these savings because manufacturers will
have to evaluate each engine family on the basis of its
emission characteristics. While the fuel economy changes which
may occur cannot be quantified precisely at this time, a fuel
economy improvement of 1 percent for all gasoline-powered LDTs
would mean a lifetime reduction in operating costs of
approximately $51 (20 mpg base, 120,000 miles in 11 years,
$1.30/gal, 10 percent discount rate).
In the case of LDDTs some vehicles may incur a small fuel
economy penalty as a result of these regulations. As was shown
in Chapter 2, the effect is not expected to be measurable on
small LDDTs and in the case of large LDDTs the effect is
expected to be less than one percent. It is possible, however
that any fuel economy penalty attributable to increased EGR
rates may be avoided by the use of electronically controlled
EGR systems. For LDDTs, a 1 percent change in fuel economy
means a $41 (23 mpg base, 120,000 miles in 11 years, $1.20/gal,
10 percent discount rate) change in lifetime operating costs.
c.	Ma intenance
Changes in the costs of maintenance attributable to the
NOx standard for LDTs, are not expected to occur. User costs
in this area are, therefore, not expected to change.
d.	Total Costs to Users
To summarize, it is reasonable to expect that users of
LDTs can expect to pay an average of about $74 more for 1987
model year LDTs than for comparable models purchased in 1986
(1984 dollars). Operating costs of LDDTs could increase as a
result of these regulations if losses in fuel economy occur.
Conversely, operating costs of gasoline-powered LDTs may
decrease as a result of improvements in fuel economy.
3. Aggregate Costs (LPT NOx)
The aggregate cost to the nation of complying with the
proposed 1987 Federal LDT NOx emission regulations consists of
the sum of fixed costs for RD&T and new emission control
hardware. These costs are calculated based on sales
projections for the 5-year period following introduction of the
standard.

-------
3-28
The 5-year costs of compliance are dependent on the number
of LDTs sold during the period. The accuracy and validity of
projecting vehicle sales as far into the future as 1991 is
problematic, so cost estimates based on such projections are
subject to some qualification. However, because the largest
portion of the costs in this analysis are variable costs
(hardware), and not fixed costs (RD&T), the accuracy of the
sales projections are not as critical as might be the case in
some other rulemaking actions. Future sales of LDTs for the
analysis period were discussed previously, and were shown in
Table 3-1.
The various costs associated with this rulemaking action
will occur in different periods. In order to make all costs
comparable, the present value at the start of 1987 of the
aggregate costs has been calculated, based on a discount rate
of 10 percent. The calculations and the assumptions required
for the calculations, are shown in Table 3-8. The aggregate
cost of complying with the new regulations for the 5-year
period is estimated to be equivalent to a lump sum investment
of about $1.08 billion (1984 dollars), made at the start of
1987. Expressed in other terms, the aggregate cost of
compliance is equivalent to an investment of $61 per LDT, made
at the start of the year these NOx standards become effective.
The aggregate cost changes by about $856 million for each 1
percent change in LDT fuel economy.
For ease of reference, the components of the cost of
compliance and the different methods of expressing it are
summarized in Tables 3-9 through 3-12.
B. Heavy-Duty Gasoline Engines (HDGEs) - NOx Standards
1. Cost to HDGE Manufacturers of the NOx Standards
a. ^ixed Costs (Research, Development and Testing)
The technologies which can be expected to be required on
HDGEs to meet the 1987 and 1990 NOx standards were identified
in Chapter 2. That analysis showed that HDGEs can be expected
to be brought into compliance with the 1987 standard through
recalibrations of the fuel, ignition, and EGR systems. For the
1990 standard, combustion chamber redesign for the application
of "fast-burn" technology coupled with the development of
appropriate calibrations for the fuel, ignition and EGR systems
were the most probable approaches which manufacturers are
expected to employ. Other technologies which were identified,
but which were judged not to be required for compliance with
the 1990 standard were: electronic controls, fuel injection,
and three-way catalysts. For this analysis, it was assumed
that manufacturers would use "fast-burn" combustion chambers
which employ a single spark plug and that, as a consequence,
substantial ignition system redesign would be avoided.

-------
3-29
Development and testing costs are not expected to be large
for the recalibration of the fuel, ignition, and EGR systems.
Because essentially similar work is expected to be required for
the recalibrations for the 1987 standard and for the 1990
standard, a single calculation will be performed to develop the
recalibration costs. This cost will then be applied to both
the 1987 and the 1990 standards. Similarly, the costs of
certification will be calculated once and the applicable* cost
will be applied to each of the standards. As a first estimate,
the costs of the calibrations and the certification testing,
per engine family, are estimated as follows:
0 Development of three calibration combinations of EGR, fuel
and ignition systems per engine family with the optimum
calibration selected for production. Six person weeks of
effort per calibration at $50 per hour results in a cost
of $36,000.
° Certification testing involving one durability and three
data engines per engine family. Dynamometer operation
time of 3, 500 hours** for the durability engine, and 130
* New split-class standards for HC and CO are effective
beginning with the 1987 model year. The new, more
stringent statutory standards will apply to 70 percent of
the HDGEs while the 1985 standards would remain in effect
for the remaining 30 percent. As a result, 70 percent of
the HDGEs will require certification testing to establish
compliance with the HC and CO standards and 30 percent
will utilize carryover HC and CO certification.
Additional costs for certification testing for the 1987
NOx standard would, therefore, not be required for 70
percent of the HDGEs. The cost allocated in this action
for certification testing associated with the 1987 model
year NOx standard is, therefore, limited to 30 percent of
the HDGEs.
** Prior to the 1984 model year, 1,500 hours of dynamometer
operation of a HDGE was used to represent 50,000 miles of
vehicle operation. Direct proportioning of these values
to a lifetime mileage of 110,000 miles would result in a
dynamometer operating time requirement of 3,300 hours. To
be conservative, EPA increased the 3,300-hour value by a
contingency factor of approximately six percent to arrive
at the value of 3, 500 hours which is used in the cost
estimate. Present regulations require durability data
engines to be operated in such a fashion as to represent
real-world conditions but allow manufacturers to select
the operating conditions. It is possible, therefore, that
manufacturers may identify procedures which would allow
some reduction in the dynamometer-hour requirements for
durability engines. Under these conditions, the cost of
collecting data on durability engines could be reduced.

-------
3-30
hours per emission-data engine at a cost of $30 per hour
results in a cost of $116,700.
Emission tests at 150-hour intervals for the durability
engine for a total of 23 tests and two tests per data
engine at a cost of $2,000 per test results in a cost of
$58,000.
Application of a 10 percent contingency factor results in
a total cost per engine family of approximately $232,000 for
recalibrations and certification. These are the only fixed
costs which apply to the 1987 standard. In addition to the
recalibration and certification costs, the fixed costs for the
1990 standard would include work necessary to redesign the
combustion chamber. For HDGEs, combustion chamber redesign is
expected to be achieved through a redesign of the cvlinder
head. This task would involve the design effort, prototype
part procurement, testing to establish the performance of the
design, mechanical durability testing, and necessary design
modifications to incorporate the results of the testing
programs. ^s a first estimate, two person-years of effort will
be assigned for the design, performance testing, and any
necessary redesigns. At an hourly cost of $50, including the
cast of prototype parts, this effort comes to $208,700 per
engine family. Dynamometer test , time to prqve mechanical
durability will be assumed to cover full engine life
(110,000-mile equivalent) and is assumed to be performed under
higher than average mechanical and thermal load conditions.
These conditions are assumed to be reflected in engine speeds
which correspond to a vehicle speed which is 50 percent hiqher
than that used for certification testing, i.e., at 45 mph
equivalent. These assumptions lead to a mechanical durability
dynamometer test time of 2,333 hours and a corresponding cost
of $70,000. The fixed cost per engine family for the cylinder
head redesign is, therefore, estimated to be approximately
$279,000. The total fixed cost per engine family for the 1990
standard is the sum of the component parts of the fixed cost
and is approximately $539,000, including a 10 percent
contingency factor.
Being conservative, it is possible that there could be as
many as 22 heavy-duty gasoline engine families in production in
the 1987 and later model years (assuming that each of the 11
existing basic engine groups are divided into two enqine
families). The total fixed or pre-production cost for each of
the NOx standards would, therefore, be $1.5 million ($232,000 X
22 X 0.3) for the 1987 standard and $11.9 million ($539, 000 X
22) for the 1990 standard (Table 3-14). Expressed on a
per-engine basis for engines produced in the model years
1987-89 and 1990-92 covered by the NOx standards (Table 3-1) ,

-------
3-31
Table 3-14
Heavy-Duty Gasoline
Engines; Costs and Present Values* of
Complying with the 1987 and 1990 NOx Standards (1984 dollars)
Cost	Present Value Present Value
Year
Fixed
Hardware
In 1987
In 1990
1985
$766K
--
$9 2 7K
--
1986
766K
--
823K

1987

--
--

1988
8 ,00OK

--
9,680K
1989
3,858K

--
4 ,244K
1990
--
1,855K
--
1,85 5K
1991
--
1, 835K
--
1,668K
1992
	
1,820K
	
1,504K
TOTAL
$13,390K
$5,510K
$1,750K
$18,9 51K
10 percent discount rate to year standard is introduced.

-------
3-32
this cost is $1.33 for the 1987 standard and $10.76 for the
1990 standard (Table 3-15) . These costs are expected to be
spread over the two years preceding the introduction of the
1987 and 1990 model year engines (i.e., during 1985 and 1986
for the 1987 standards and during 1988 and 1989 for the 1990
standards). Distribution of the costs between the years can
reasonably be expected to be as shown in Table 3-14.
b.	Emission Control Systems Hardware Costs (Variable
Costs)
For the 1987 standard, it is probable that manufacturers
of gasoline-fueled HDEs will recalibrate existing systems and
will, therefore, not incur any significant hardware costs.
For the 1990 standard, hardware costs will probably be
mainly attributable to the new cylinder heads. While it could
be assumed that the redesigned cylinder heads would not cost
any more than those which are being replaced, a conservative
approach was used and a cost increase of $5 per engine was
assigned. For the engines which are expected to be produced,
this per engine cost of $5 represents totals of $1,855,000 in
1990, $1,835,000 in 1991, and $1,820,000 in 1992 (Table 3-14).
If manufacturers choose to use one or more of the other
technologies which were identified in Chapter 2, the costs can
reasonably be expected to increase appropriately.
c.	Total Costs to Manufacturers
The total cost to HDGE manufacturers of complying with the
1987 NOx standards is estimated to be approximately $1.5
million coming exclusively from fixed costs. For the 1990 NOx
standard, both fixed and variable costs are involved and total
approximately $17.4 million. On a per-vehicle basis, the total
cost is estimated to be $1.33 for the 1987 standard and $15.76
for the 1990 standard (Table 3-15). The discounted costs per
engine are shown in Table 3-16 and are $1.52 for the 1987
standard and $17.20 for the 1990 standard.
2. Costs to Users (HDGE - NOx)
a. Increase in First Cost
For the 1987 standard, the first price increase will
consist only of the discounted-fixed costs of $1.52 or
approximately $2. For the 1990 standard, the first price
increase will consist of the discounted fixed cost ($12.64)
plus the hardware cost ($5.00) for a total of $17.64 or
approximately $18.

-------
3-33
Table 3-15
Heavy-Duty Gasoline Engines, Undiscounted
Costs of Compliance with the Standards per
Engine Cor Engines Produced in 1987-89 and 1990-92
Manufacturer Cost	1987-89 1990-92
Research, Development & Testing	$1.33 $10.76
Emission Control Hardware	--	5.00
$1.33 $15.76
Operating Cost
Change in Fuel Cost per one	$143	$143
Percent Change in Fuel Economy

-------
3-34
Table 3-16
Heavy-Duty Gasoline Engines, Discounted
Costs of Compliance per Engine for
Engines Produced in 1987-89 and 1990-92*
Manufacturer Cost	1987-89 1990-92
Research, Development & Testing	$1.52 $12.64
Emission Control Hardware	--	4.56
$1.52 $17.20
Operating Cost
Change in Fuel Cost per one	$104.92 $104.92
Percent Change in Fuel Economy
10 percent discount rate to year standard is introduced,

-------
3-35
b.	Fuel Economy
While changes in the calibrations of fuel, ignition, and
EGR systems can have some impact on fuel economy, it is
expected that the system recalibrations should not result in
any fuel economy penalty. In fact, other measures to improve
fuel economy will probably be taken in this period because of
market pressure in that direction.
While a fuel economy effect is not anticipated, an
estimate has been prepared of the cost of a fuel economy change
for purposes of completeness in this document. This value was
computed by using a fuel cost of $1.30 per gallon, a fleet
average fuel economy of 10 miles per gallon, a 110,000
mile/8-year lifetime, and a 10 percent discount rate. These
computations show that each 1 percent change in fuel economy
yields a $105 discounted change in operating costs for fuel
over the lifetime of the engine (Table 3-16) .
c.	Maintenance
The cost of maintenance is not expected to be impacted by
either the 1987 model year or the 1990 model year NOx standards
for HDGEs.
d.	Total Costs to Users
To summarize, owners of gasoline-fueled HDVs can, as a
result of the proposed NOx standards, expect an increase in
first price of approximately $2 in 1987 and $18 in 1990 with no
change in operating cost.
3. Aggregate Costs
For HDGEs, the aggregate cost to the nation of compliance
with the 1987 and the 1990 NOx standards would consist of the
fixed costs and the hardware costs discounted to the years in
which the standards become effective. As in the case of LDTs,
a 10 percent cost for capital is employed in determining the
lump sum investment which would be equivalent to the fixed
costs for the two standards. The lump sum investments are
approximately $1.8 million in 1987 and $19.0 million in 1990
and reflect the aggregate costs for the standards. Expressed
in other terms, the aggregate cost of compliance is equivalent
to an investment of just under $2 per HDGE for the 1987
standard and just over $17 for the 1990 standard made in the
years that each of the standards become effective. For ease of
reference, the aggregate costs are shown in Table 3-17.

-------
3-36
Table 3-17
Heavy-Duty Gasoline Engines,
Agqreqate Cost of Compliance for
the 1987 and 1990 Standards (10 percent
discount to January of year standard is introduced)
1987	1990
Research, Development & Testing $1,750K $13,924K
Emission Control Hardware		--	5,Q27K
Total
Change in aggregate cost per
1 percent change in fuel economy
of the fleet
$1,750K
$121M
$18,9 51K
$115M

-------
3-37
C. Heavy-Duty Diesel Engines (HDDE) - NOx and
Particulate Standards
In the case of HDDEs, four new standards are proposed:
NOx and particulate standards for the 1987 model year and NOx
and particulate standards for the 1990 model year.
Manufacturer costs are developed for each standard by the
format previously used for LDTs and HDGEs. The technologies
which EPA expects that manufacturers will employ in meeting the
standards were identified in Chapter 2 (Technological
Feasibility). Application of these technologies forms the
basis for the cost estimates developed below.
It should be noted that a relatively large number of
technology combinations could be employed in meeting the
proposed standards. For purposes of cost estimation, the
approach used was to address the overall application picture
rather than to attempt to identify specific combinations of
technologies for each power rating within an engine family.
Because the exact mix of technologies to be used is uncertain,
there is also uncertainty in the cost estimates developed in
this chapter. However, these estimates are believed to be
representative of the actual costs manufacturers will incur.
1. Cost to HDDE Manufacturers of the 1987 Standards
a. Fixed Costs (research, development and testing
(RD&T))
NOx Standard: HDDE manufacturers can be expected to meet
the 1987 NOx standard by employing such techniques as injection
timing retard and the addition of aftercooling* to some of the
turbocharged engines which presently are not so equipped.
The work involved in implementing retarded injection
timing can be viewed as being the equivalent of a calibration
change. Because of the high priority which is placed on the
fuel economy of HDDEs, it is reasonable to expect that
manufacturers will evaluate several injection timing
Aftercooling means cooling of the intake air after it is
compressed by the turbochacger and prior to its entering
the engine. Cooling of the pressurized intake air can be
achieved by passing it through a heat exchanger which is
cooled either by engine coolant or by ambient air or by a
separate liquid coolant system. Aftercooling systems
which use ambient air are referred to as air-to-air
systems while the other systems are referred to as
air-to-liquid systems.

-------
3-38
calibrations before selecting the calibration which results in
minimum fuel economy penalty while complying with the emission
standard. For purposes of developing this cost estimate it was
assumed that three calibrations would be evaluated per engine
family in the development of the final calibration. Assuming
160 hours of effort per calibration for the calibration
changes, data collection and analysis plus a ten percent
contingency factor, the development cost per engine family
would be:
3 calibrations x 160 hours/calibration x $50/hour x 1.1
contingency factor = $26,400.
The second design and development effort which was identified
as being necessary for compliance with the 1987 model year NOx
standard involves the application of aftercooling to some
engines. The work required to accomplish this application
would involve aftercooler sizing calculations, preparation of
design and installation drawings, procurement of experimental
parts, and some testing to confirm the performance of the
aftercooler. Approximately six person-months of effort should
suffice to accomplish this work. Using a cost of $50/hour to
cover personnel, experimental parts and confirmatory testing
plus a ten percent contingency factor, the cost per engine
family would be:
6 person-months x 174 hours/person-month x $50/hour x 1.1
contingency factor = $57,400.
For the 1982 and 1983 model years, the sales mix between HDDEs
which were turbocharged with aftercooling, turbocharged without
aftercooling and non-turbocharqed were approximately 41
percent, 21 percent and 38 percent respectively. As was stated
previously, EPA expects that some, but not all, of the
non-aftercooled engines will have aftercooling added. As a
first estimate, it is assumed that approximately one-half of
the non-aftercooled engines will have aftercooling added. This
change would, therefore, impact approximately 10 percent of
total engine sales.
Certification records show that 86 engine families were
certified for the 1983 model year. While it is possible that
some manufacturers may add engine families and that other
manufacturers may enter the HDDE market, it is also possible
that some engine families may be discontinued. Because it is
not possible to quantify precisely what changes may occur, it
was assumed that there will be no significant changes from
1983, in the number of engine families which will be certified
in 1987 and 1990. Combining the cost estimates for each task
which has just been developed, with the number of engine
families involved, gives the development cost for the 1987 NOx
standard as follows:

-------
3-39
$26,400/engine family for recalibration x 86 engine
families + $57,400/engine family for aftercooling x 0.1 to
account for the number of families which are involved x 86
engine families or approximately $2.8 million.
The third component of the fixed cost is the cost of
certification. This cost will be estimated first in terms of
the total cost of certification, followed by equal
apportionment between the NOx and particulate standards.
Continuing the conservative approach for estimating
dynamometer-hour requirements addopted for HDGEs, but employing
the relationship of 1,000 dynamometer hours as being equivalent
to 100,000 miles of operation for HDDE (regulations applicable
prior to 1984), the dynamometer-hour requirements are as
follows. For HDDEs, the three lifetime mileages (110,000 miles
for LHDDEs, 185,000 miles for MHDDEs and 290,000 miles for
HHDDEs) are proportional to 1,100, 1,850, and 2,900 dynamometer
hours. Applying the same contingency factor of ten percent as
was used previously results in dynamometer requirements of
1,210 hours, 2,035 hours and 3,190 hours for the three groups.
Combining these dynamometer-hour estimates with the appropriate
distribution between engine families of the three groups
permits development of an average dynamometer time requirement
per engine family. Inspection of the engine descriptions for
the 86 engine families certified for the 1983 model year
shows: 1) that approximately 14 families represent
110,000-mile useful-life engines, 2) that approximately 16
families represent 185,000-mile useful-life engines, and 3)
that approximately 56 families represent 290,000-mile useful
life engines. Weighting the dynamometer hour requirements by
these engine family values results in an average dynamometer
hour requirement as follows:
(14/86 x 1,210 hours) + (16/86 x 2,035 hours) + (56/86 x
3,190 hours) = 2,653 dynamometer hours per engine family.
Following the same methodology for HDDEs as was used for
HDGEs in developing the certification cost estimate results in
the following:
° One durability and three data engines per engine family,
dynamometer operation time of 2,653 hours for the
durability engine, and 130 hours per emission-data engine
at a cost of $30 per hour results in a cost of $91,290.
° Emission tests at 150-hour intervals for the durability
engines for a total of 17 tests and two tests per data
engine at a cost of $2,000 per test results in a cost of
$46,000.

-------
3-40
Applying a 10 percent contingency factor to the sum of
these costs and multiplying by the 86 engine families involved,
results in a total cost for certification of approximately $13
million. Apportioning the certification costs equally between
the ?JOx and the particulate standards, results in a per
pollutant certification cost of $6.5 million.
The total fixed cost for the 1987 model year NOx standard
would be the sum of the costs for injection timing
calibrations, additional applications of aftercooling and
certification and equals approximately $9.3 million. This
value is shown in Table 3-18.
Particulate Standard: The technological changes which are
expected to be employed by manufacturers in complying with the
1987 model year particulate standard were detailed in Chapter
2. In summary, the changes included: 1) modifications to the
combustion chamber (probably through changes in the piston] , 2)
changes in injectors and increased injection pressure, 3)
changes in the fuel delivery system so as to refine air/fuel
ratio control during transient operation, and 4) changes to the
turbocharger to improve air delivery characteristics during
transients. A small (but not quantifiable) increase in the use
of turbochargers which could benefit particulate control was
also identified. Marketing forces will likely be, however, the
orimary reason for any increase in turbocharger use. Costs
associated with this change are, therefore, not attributable to
this regulation and are not included.
Changes in combustion chamber configuration through
changes in the piston crown will involve some design effort,
procurement of redesigned pistons and testing to establish the
effects of the redesign. Changes in the injectors, injection
pressure and the air/fuel controls would involve the same type
of tasks as those required for the combustion chamber changes
(i.e., some design effort, test part procurement and testing to
establish the effects of the changes).
The costs associated with these four changes are derived
on the basis of four person-months per design per component
changed, assuming that one-half of the engines will require
changes and allowing for two design changes per component to
facilitate selection of the optimum design, as follows:
86 engine families x 1/2 (to account for half of the
families being changed) x (4+4+4+4)
person-months/change x 2 designs/change x 174 hours/month
x $50/hour for personnel, parts and testing plus a ten
percent contingency factor = $13,168,000 or approximately
$13.2 million.

-------
3-41
Table 3-18
Heavy-Duty Diesel Enqines, Fixed
Costs of the 1987 NOx and Particulate Standards (1984 dollars)
Non-Certifi-
cation Costs Certification Costs Total Fixed Costs
Year NOx	Part.	NOx	Part.	NOx	Part.
1985	$2,000K $12,000K $1,000K $1,000K $3,000K $13,000K
1986	800K 1,200k	5,500K 5,500K	6,300K	6,700K
Totals$2,800K $ 13,200K $6,500K $6, 500K $9, 300K $19,700K

-------
3-42
Combining the estimated cost for development with the portion
of the total certification cost previously developed for the
1987 model year particulate standard ($6.5 million) results in
the total RD&T cost for the particulate standard of
approximately $19.7 million. This cost is shown in Table 3-18.
Much, if not all, of the design and development work
necessary for compliance with the 1987 model year NOx and
particulate standards can reasonably be expected to be expended
durinq 1985 with some of the longer leadtime certification
testing also being performed in 1985. The apportionment of the
design, development and certification costs between 1985 and
1986 together with the totals are shown in Table 3-18.
b. Emission Control System Hardware Costs for the 1987
Standards (Variable Cost)
NOx Standard; It is expected that changing fuel injection
timinq can be accomplished without any change in hardware.
There are, therefore, no hardware costs associated with this
change.
In developing the cost estimate for the aftercooler, it
was assumed that these units would be air-to-liquid systems
employing engine coolant as the medium for cooling the intake
air and that the engine (truck) radiator does not have to be
enlarged to accommodate the extra cooling load imposed. The
components required for the aftercooler application would,
therefore, be: 1) the heat exchanger and its housing which is
introduced into the engine intake, modified engine intake
system to provide olumbinq for the intake air from the
turbocharger to the aftercooler and from the aftercooler to the
intake manifold, and 2) coolant hoses to and from the
aftercooler. Using the cost estimate for an LDV radiator given
in Reference 4 as the basis and assuming that the heat
exchanger will be smaller than an LDV radiator, the estimated
cost of the heat exchanger component of the aftercooler system
is estimated to be approximately $41.* Reference 4 does not
contain cost information which would allow direct development
of an estimate for the costs of the aftercooler housing, intake
air plumbing and coolant plumbing for the aftercooler. As a
first estimate, a cost was assigned for these components which
$37.09 (1977 dollars) for the radiator assembly x 0.8 to
account for reduced materials usage because of size
reduction x (product of price inflation indices for 1978
through 1983 of 6.2, 7.4, 7.5, 6.8, 1.6, and 3.6 percent,
respectively) = $40.90.

-------
3-43
is approximately equivalent to one-half of the cost of the
aftercooler heat exchanger, or $20. The total estimated cost
for the aftercooler system is, therefore, $61 per engine. New
engine list prices were checked as a method of evaluating this
estimate. In all cases where aftercooled and non-aftercooled
but otherwise similar engine pairs were compared, the
aftercooled engine had a higher horsepower rating and a
substantially higher price. Because marketing forces
interrelate horsepower and price, the incremental price
difference could not be attributed to the aftercooler. EPA
believes that the cost estimate shown is a reasonable estimate
of the aftercooler cost. With an aftercooler installation rate
of 10 percent of sales, the total undiscounted costs are
$2,141,000 in 1987, $2,233,000 in 1988, and $2,342,000 in
1989. These values are shown in Table 3-19.
Particulate Standard: Little, if any, difference exists
between the hardware costs of original design and redesigned
components such as pistons, injectors, fuel metering and
delivery systems and turbochargers. To be conservative,
however, a per engine cost of $5 was assigned for each of the
modified components (Distons, injectors, fuel metering and
delivery and turbochargers) which are expected to be employed
in meeting the 1987 model year particulate standard. This
approach will also account for any additional hardware needed
for improved air/fuel ratio control where none exists today.
The total per engine cost for hardware would, therefore, be
$20. Continuing the previously stated assumption that one-half
of the engines produced could require these hardware
modifications, results in hardware costs of $3,510,000 in 1987,
$3,660,000 in 1988 and $3,840,000 in 1989. These costs are
shown in Table 3-19.
c. Total Costs to Manufacturers for the 19 S 7 hdde
Standards
The major undLscounted costs to manufacturers of 3D&T and
emission control hardware are $16.0 million for the
standard and $30.7 million for the particulate standard. The
overall total is approximately £46.7 million as shown in Table
3-19. The discounted costs (1987) of $16.7 million, $33.1
million and $49.8 million for the NOx standard, the particulate
standard, and the total respectively are shown in Table 3-20.
The per engine discounted costs are also shown in Table 3-20
and are $15.14 for the NOx standard, $30.07 for the particulate
standard and $45.21 for both standards.

-------
3-44
Table 3-19
Heavy-Duty Diesel Engines;
Manufacturers Undiscounted Costs of
Compliance for the 1987 MY NOx and Particulate Standards
	RD & T		Emissions Hardware
Year NOx	Particulate	NOx	Particulate
1985	$3,000K	$13,000K
1986	6,300K	6,700K
1987	--	--	$2,141K	$3,510K
1988	--	--	2,233K	3,660K
1989	--	--	2 , 342K	3,840K
$ 9,3 OOK	$19,700K	$6,716K	$11,010K
Total Undiscounted Cost:	NOx: $16,016K (1984 dollars)
Particulate: $30,710K (1984 dollars)
Total: $46,726K

-------
3-45
Table 3-20
Heavy-Duty Diesel Engines,
Discounted Cost of Compliance for
the 1987 Model Year NOx and Particulate Standards
Cost	Present Value in 1987
Year
NOx
Par t.
Total
NOx
Part.
Total
1985
$3,000K
$13,000K
$16,000K
$3,630K
$15,730K
$19,360K
1986
6,300K
6,700K
13,000K
6,930K
7,370K
14,300K
1987
2,141K
3,510K
5,651K
2,14 IK
3,510K
5,651K
1988
2,233K
3,660K
5,89 3K
2 , 0 30K
3,327K
5,357K
1989
2,342K
3,840K
6,18 2K
1,936K
3,174K
5,110K
Total
$16,016K
$30,710K
$46,726K
$16,667K
$33,111K
$49,778K
Lump
sum investment cost
per enqine
made at
the start of
1987
for the NOx standard:	$15.14
for the Particulate standard: $30.07
for both standards:	$45.21

-------
3-46
2. Cost to HDDE Manufacturers of the 1990 Standards
a. Fixed Costs (research development and testing (RD&T))
NOx Standard: In Chapter 2 (Technological Feasibility),
technologies which may be employed in complying with the 1990
NOx standard were identified. As was pointed out in Chapter 2,
some uncertainty exists with respect to these technologies.
The estimate of cost developed below is, therefore, also
subject to some uncertainty. These technologies were: 1)
possibly some additional injection timing retard, 2)
application of aftercooler systems which are designed to
maximize the cooling of intake air through the use of either
air or an auxiliary radiator for cooling, 3) application of EGR
to prechamber engines, 4) modification of the compression ratio
on some engines, and 5) the use of electronics for fuel
injection and EGR control.
It is expected that manufacturers will employ essentially
the same procedures in development of the optimum injection
timing retard for the 1990 NOx standard as was employed for the
1987 NOx standard. The cost for this work was previously
estimated as being $26,400 per engine family and will be used
here.
Two methods of maximizing the cooling of the intake air
charge have been identified. These methods are: 1} an
air-to-air aftercooler system, or 2) an air-to-liquid
aftercooler system similar to that which is presently used plus
an auxiliary radiator for cooling of the aftercooler-system
coolant. It appears reasonable to assume that manufacturers
would avoid the use of fans whose sole function would be
cooling of the aftercooler system because of their cost and
parasitic losses. Expectations are, therefore, that
manufacturers will design the aftercooler systems to employ
either ram-air cooling or cooling by the existing radiator fan.
In the case of air-to-air aftercooler systems, HDDE
manufacturer design and development costs would consist of
development of the performance specification for the
aftercooler system, design of the intake air plumbing from the
turbocharger to the aftercooler heat exchanger and back to the
engine, procurement of test parts, and testing to establish the
performance of the system. Two factors can be expected to bear
heavily on the level of effort which would be associated with
this design approach. First, redesigns of the intake system
can be expected. Second, this design approach has not been
extensively used to date on HDDEs and manufacturers may tend to
perform more development testing than they would otherwise.
The level of effort associated with this task is estimated,

-------
3-47
therefore, to be 1.5 person-years per engine family at a cost
of $50/hour to cover personnel, equipment and test cost. The
cost per engine family is estimated, therefore, to be $172,200
including a 10 percent contingency factor.
For the air-to-liquid aftercooler system, it is expected
that the design and development costs would be lower than those
for the air-to-air system because the aftercooler heat
exchanger and the engine intake system should not require
modification. The new components will be the aftercooler
radiator, coolant hoses and a pump for circulating the
coolant. Testing requirements to establish the performance of
the system could be less than with the air-to-air system
because the lack of change in the intake system design would
oermit greater dependance on aftercooler radiator performance
tests in place of total system tests. It is expected that the
design and development costs for the air-to-liquid aftercooler
system could be approximately one-third of that for the
air-to-air system (i.e. $57,400 per engine family).
The third area where RD&T work can be expected is in the
application of EGR to those LHDDEs which use prechamber
design. One of these prechamber engines (GM product) is very
similar to an EGR equipped LDT engine sold by that
manufacturer. Direct technology transfer can be expected to
occur which would preclude significant 'design and development
costs for these engines. In the case of the other prechamber
engine, design costs would be incurred for the application of
EGR. The cost per engine family for this work can reasonably
be expected to be similar to that which was previously
estimated for LDDTs and is $86,000 including the cost of
mechanical durability testing.
Changes in the compression ratio of some HDDEs constitutes
the fourth area where RD&T costs may be incurred. This change
could be accomplished through changes in the pistons and would
incur costs similar to those previously estimated for
combustion chamber changes; i.e., 2 designs/change x 4 person
months/change x 174 hours/month x $50/hour plus 10 percent
contingency = $76,600.
The application of electronic controls constitutes the
final design and development task which was identified. The
basic functions associated with the application of electronic
controls to HDDEs can be identified as follows: the
installation of sensors, the installation of actuators,
installation of the control module, and the wiring necessary to
interconnect the sensors, actuators and electronic module.
Development of cost estimates for these functions pose
significant problems as they relate to this regulation. Some

-------
3-48
manufacturers use vendor supplied fuel-injection systems and
may, therefore, be able to purchase completely designed
electronic control systems from vendors. Other manufacturers
design their own fuel-injection systems and, as a result, may
have to design much of the electronic-control system for their
engines. In addition, manufacturers are known to be working on
the development of electronic controls for HDDEs and may have
complete systems, including the components necessary for NOx
and particulate control, in production prior to the effective
date of this regulation. The most probable reasons for the
ongoing development work on electronic controls for HDDEs are
improvements in fuel economy which may be realized and
manufacturers' desires to increase their capabilities for
compliance with stringent emission standards. Only a part of
the cost of developing electronic controls for HDDEs should
reasonably be allocated, therefore, to the control of
emissions. Based upon this analysis, a design, development and
test cost of $208,800 (two person-years at $50/hour) per engine
family was assigned for the work of developing electronic
controls for both the fuel injection and the EGR systems. An
equal division of the costs between the NOx and particulate
standards was also chosen.
The cost of certification testing was previously estimated
as being $6.5 million per emission standard for the 1987
standards. This cost will also be used for the 1990 standards.
The overall fixed cost for the 1990 NOx standard is
developed by weighting the individual cost estimates by the
number of engine families which are expected to require each of
the technologies as follows:
0 86 engine families x $26,400/engine family for injection
timing modifications = $2,270,000.
0 Application of aftercooling to all families except the 14
LHDDE families: (86-14) x [($177,200 + $57,400)/2],
assuming equal distribution between the two aftercooler
design approaches = $8,266,000.
0 Application of EGR to one engine family = $86,000.
0 Changes in piston design on 25 percent of the engine
families: (86/4) x $76,600 = $1,685,000.
0 86 engine families employing electronic controls x
($208,800)/2 per engine family x 1.1 (10 percent
contingency factor) = $9,876,000.

-------
3-49
0 86 engine families being certified at a cost of $6,500,000.
The total fixed cost estimate is the sum of the individual
components and is $28,683,000, or approximately $28.7 million.
RD&T work directed to meeting the 1990 NOx standard can
reasonably be expected to begin in 1987 and to continue into
1989. The projected distribution of these costs is shown in
Table 3-21.
Particulate Standard: Particulate trap technology is
presently Tn the transition stage from research to relatively
low volume production for some 1985 model year LDDVs. It is
reasonable to expect, therefore, that a relatively high level
of effort would be expended by each of the larger manufacturers
of HDDEs, with appropriate interfacing with trap manufacturers,
in the design and development of particulate trap systems for
HDDEs. These efforts are expected to focus first on the
development by each of the larger manufacturers of a general
system design which possesses the desired durability and trap
efficiency and then on the design and development of engine
family specific systems. In the case of HDDE manufacturers
which are either relatively small or which have low sales
volumes in the United States, it is reasonable to expect that
there would be a dependence upon the trap manufacturers for the
general system design, followed by the engine manufacturer's
development of engine family specific designs.
With the exception of Ford, all larger manufacturers of
heavy-duty trucks which also produce cars and/or lighter duty
trucks are manufacturers of HDDEs. To be conservative, it was
assumed that Ford could enter the HDDE market by 1990 with the
result that there would be seven HDDE manufacturers which would
develop their own general system designs. The other
manufacturers would be expected to rely on particulate trap
suppliers for the general system designs. As a first estimate,
it is expected that each of the larger manufacturers could
expend 20 person-years of effort in the 1986-87 timeframe in
the initial development of a general system design. This work
would entail higher expenditures for experimental parts, test
engines and fuel than would be expended in the development of
engine family specific designs. The cost of $50 per hour which
has been used previously in developing cost estimates for
design and development will be increased to $60/hour for this
effort. On a per manufacturer basis, the cost of developing
the general design is estimated to be 20 person-years x 2088
hours per year x $60 per hour or $2,506,000.

-------
3-50
Table 3-21
Heavy-Duty Diesel Engines, Fixed
Costs for the 1990 NOx and Particulate Standards

Non-Certifi-
cation Costs
Certification Costs
Total
Fixed Costs
Year
NOx
Part.
NOx
Part.
NOx
Part.
1986

$8,000K
--


$ 8 ,000K
1987
$7,COOK
20,000K


$ 7,0 0 OK
20,000K
1988
14,000K
13,000K
1,00 OK
1,000K
15,000K
14 ,000K
1989
1,200K
2,000K
5,500K
5,500K
6,700K
7,500K
Total
22,200K
4 3,000K
6,500K
6,500K
28,700K
$49 , 500K

-------
3-51
Design, development and system performance testing for the
engine family specific designs is estimated to require two
person-years of effort per design. The cost of this effort is
estimated to be 2 person-years x 2088 hours per year x $50 per
hour or $208,800 per engine family.
It is anticipated that 30 percent of the HDDE families may
not require the use of particulate trap systems as a result of
averaging in determinations of compliance with the particulate
standard. The total, industry wide, fixed cost for the design
of the particulate trap system is, therefore, estimated to be
$30,112,000 (($2,506,000 x 7 manufacturers) + ($208,800 x 86 x
0.7 engine families)) exclusive of any contingency factor.
With a 10 percent contingency factor, the total cost becomes
$33,123,000.
The first cost for applying electronic controls to HDDE
attributable to these regulations has previously been estimated
at $19,752,000, equally divided between the NOx and particulate
standards, (i.e. $9,876,000 for the particulate standard).
The final component of the first cost of the particulate
standard is the cost of certification which was previously
estimated at $5.5 million per emission standard.
Summing the component parts of the first cost results in a
total first cost estimate of $49,499,000 or approximately $49.5
million. These RD&T costs can reasonably be expected to be
expended between 1986 and 1989 as shown in Table 3-21.
b. Emission Control System Hardware Costs (Variable
Cost) for the 1990 Standards
NOx Standard: Hardware costs attributable to the 1990 NOx
standard for HDDEs would accrue from the use of such
technologies as aftercooling systems, electronic engine
controls, modified pistons and the application of EGR. The
estimates for the cost of each of these hardware changes will
be developed separately followed by appropriate sales
weightings to develop the total hardware cost estimate.
/
As was discussed previously, two aftercooler design
approaches can be employed, (i.e. air-to-air systems or
air-to-liquid systems). It was also indicated previously that
aftercooler designs are expected to avoid the use of an
additional cooling fan. This approach would result in the
placement of either the air-to-air aftercooler or the radiator
section of the air-to-liquid system in front of the engine
radiator. While some small increase in the temperature of the
cooling air will occur as the air passes over the aftercooler,
it is expected that the increase will not be great enough to
necessitate an increase in the size of the engine radiator.

-------
3-52
For the case of the air-to-air aftercooler system, the
aftercooler would be very similar in appearance and
construction to an engine radiator but would have the
pressurized engine intake air passing through it rather than
engine coolant as in the case of a radiator. The cost of this
component can, therefore, be derived from the cost estimate for
a LDV radiator as given in Reference 4. The size of the
aftercooler and, therefore, the amount of materials used in its
construction are expected to be larger, however, than that of a
radiator on a LDV. A sizing factor of 30 percent was assigned
to account for the size differential. The resulting cost of
the air-to-air aftercooler is approximately $67. As was
pointed out previously, Reference 4 does not contain data upon
which the cost estimate of the intake air plumbing can be
based. In this case, the extent of the plumbing change will be
greater than that required in the aftercooler design for the
1987 standard. In addition, the plumbing will have to include
flexible components to allow for motion of the engine relative
to the aftercooler. As a first estimate, the cost chosen for
the plumbing equals that of the aftercooler, (i.e., $67). The
total estimated cost for the air-to-air aftercooler system is
the sum of these two costs (i.e. $134 per engine).
In those cases where a manufacturer chooses to replace an
existing air-to-liquid system which is cooled by engine coolant
with an air-to-air system, the hardware cost would be the value
just calculated less the cost of components removed. For the
1987 standard, the cost of adding an air-to-liquid system which
used engine coolant was estimated to be $61. The differential
hardware cost for the system replacement design approach would,
therefore, be $73 ($134 - $61).
In the case of air-to-1iquid systems which chanqe from the
use of engine coolant as the cooling medium to a separate
aftercooler radiator, the anticipated new components are: 1)
the aftercooler radiator, 2) hoses to connect the two heat
exchangers, 3) a coolant circulation pump, 4) belt drive from
the engine to the pump, and 5) support bracketry for the pump.
The aftercooler radiator would be very similar to the engine
radiator of a, LDV and its cost can be derived directly from
Reference 4, (i.e. $51). The cost estimates for the other
components are assigned values because applicable data is not
contained in Reference 4. The estimated costs are: $25 for
the pump and support bracketry, $10 for the hoses, and $5 for
the belt drive. The total for the components is $40, which
when combined with the cost of the aftercooler radiator, gives
a system cost of $91.

-------
3-53
The cost of applying an air-to-liquid system to an engine
which was not previously intercooled would consist of the cost
just developed ($91) plus the cost of the aftercooler at the
engine which was previously estimated as being $61. The total
cost would, therefore, be $152. Because this cost is higher
than the cost of using an air-to-air system ($134) it was
assumed that this approach would not be extensively employed on
engines which were not intercooled.
As was previously shown (Fixed Costs for the 1987 HDDE NOx
standard) the sales split between engines which are
turbocharged with aftercooling, turbochared without
aftercooling and non-turbocharged was 41 percent, 21 percent
and 38 percent respectively. As was also shown previously,
approximately one half of the non-intercooled turbocharged
engines can be expected to receive aftercooling as a result of
the 1987 NOx standard (i.e., 10 percent of sales will be
intercooled for the first time) . ^s a result of the 1990 NOx
standard, it is expected that aftercooling may be applied to
all turbochared engines and that a 50/50 split could occur
between the use of air-to-air systems and air-to-liquid
systems. The 11 percent of sales which receive aftercooling
for the first time (21 percent - 10 percent) as a result of the
1990 standard are assumed to use air-to-air systems and
represent 18 percent (11/62) of turbocharged engine sales.
With a 50/50 split between the two aftercooler systems, 32
percent of the existing aftercooler systems would be modified
to air-to-air systems. The remaining 50 percent of the
existing systems would be converted from an air-to-liquid
system using engine coolant to an air-to-liquid system using a
separate radiator. The average hardware cost per turbocharged
engine for improved aftercooling is calculated as follows:
0 18 percent of all turbocharged engines receive
aftercooling for the first time and use air-to-air systems
which cost $134 for a per turbocharged engine cost
contribution of $24.12.
0 32 percent of all turbocharged engines are converted	from
air-to-liquid systems to air-to-air systems at a	cost
differential of $73. The per turbocharged engine	cost
contribution is $23.36.
0 50 percent of all turbocharged engines are converted from
air-to-liquid systems using engine coolant to
air-to-liquid systems using separate aftercooler
radiators. The system hardware cost for this design
approach is $91 for a per engine cost contribution of
$45.50.

-------
3-54
The average cost per turbocharged engine is the sum of the
three components, or approximately $93. It was assumed that
the sales split between turbocharged and non-turbochared HDDE
remains relatively stable at the 1982 level (i.e., 62 percent
turborcharged). With 62 percent of HDDEs being turbocharged,
the per engine cost expressed in terms of all HDDEs is
approximately $58.
The cost of adding EGR to a LDDT was previously estimated
as being $20. This cost will be used for the engines of the
LHDDE family which are expected to have EGR applied for the
first time. This cost is considered to be reasonable because
of the general similarity which exists between the engine of a
LDDT and a LHDDE. Between five and ten percent of the
production engines are expected to have EGR fitted for the
first time. Using the larger percentage value, the per engine
cost for all engines is $2.
For those engines which employ redesigned pistons, the
previously adopted conservative approach will be continued and
a hardware cost increase of $5 per engine is assigned.
Assuming that 25 percent of all engines are equipped with
redesigned pistons, the per engine cost for all engines is
$1.25.
In Reference 1, the cost of an electronic control module
was estimated as being $75 with equal prorating of the cost
between NOx and particulate control. The cost of a sensor for
determining trap temperature was estimated as being $12. This
cost will be used as the basis for developing the estimates of
costs of sensors for NOx control systems. Sensors which will
be employed as part of the electronic control system for NOx
control will monitor engine speed, load, rate of change of
load, coolant temperature, intake air temperature and
pressure. The temperature sensors, while being essentially
on/off units similar to those used on the particulate trap,
will operate in a less thermally demanding environment than
that of the particulate trap. Their cost will, therefore, be
lower and is estimated to be $6 each. The other sensors will
be required to deliver proportional signals rather than on/off
signals and will, therefore, be more costly. Their cost is
estimated to be 50 percent higher than that of the trap
temperature sensor or $18 each. The cost of the wiring harness
connecting the sensors, electronic control module and the
actuator is estimated to be $15[4], The final components of
the electronic control system are the actuators which will
modulate fuel delivery and EGR as directed by the control
module. While Reference 4 does not contain any data on the
cost of an actuator, it is reasonable to expect that the
actuator will be an electrical stepper motor which very

-------
3-55
precisely and rapidly translates the control module signal into
either linear or circular motion. As a first estimate, a cost
of $40 has been assigned to the actuator for the fuel delivery
system and $25 to the actuator for the EGR system. Summing the
costs of the component parts of the electronic control system
gives the estimate of the cost of the system for NOx control as
follows:
$37. 50 for the control module + 2 sensors at $6 each + 4
sensors at $18 each + $15 for the wiring harness + $40 for
the fuel actuator + $25 for the EGR actuator for a total
of $201.50 or approximately $202.
The total hardware cost per engine for the NOx standard is
the sum of the components, (i.e. $58 for af tercooling, $2 for
the application of EGR, $1.25 for redesigned pistons and $202
for electronic controls). The total hardware cost per HDDE for
the NOx standard is therefore, approximately $263. Because of
uncertainties with respect to the exact mix of technologies
which will be used by manufacturers, the cost which actually
occurs could be either lower or higher than this estimated
value.
Particulate Standard: In the Diesel Particulate Study,[1]
estimates were developed for the hardware costs of particulate
trap systems for both catalyzed and non-catalyzed traps. The
system hardware costs developed in Reference 1 are reproduced
and summarized in Table 3-22.
Assuming equal performance for the two system types
(catalyzed and non-catalyzed) with respect to durability and
trap efficiency, differences in first price and fuel economy
will dominate manufacturers decisions on system selection. To
overcome the first cost differential, catalyzed systems would
have to exhibit lifetime fuel economy benefits relative to
non-catalyzed systems on the order of 1 percent, 1.5 percent
and 4.5 percent for HHDDE, MHDDE, and LHDDEs, respectively.*
EPA expects that there will be little, if any, difference in
fuel economy between the two system types and has concluded
that manufacturers will probably select the non-catalyzed
system. Combining the costs for the non-catalyzed systems with
the expected sales distributions between LHDDEs (33 percent)
MHDDEs (25 percent) and HHDDEs (42 percent)[1] gives the
average hardware cost per engine of $533 for engines equipped
System cost differential divided by the undiscounted,
lifetime fuel economy cost differentials of $900, $390,
and $86 per one percent change in fuel economy for HHDDEs,
MHDDEs, and LHDDEs, respectively.

-------
3-56
Table 3-22
Heavy-Duty Diesel Engines,
Particulate Trap System Hardware Costs per Engine[3]
Non-Catalyzed Trap		Catalyzed Trap
HDDE
Group
Trap
Regeneration
Svstem
Total
System
Trap
Regenerat ion
System
Total
System
LHDDE
$207
$156
$363
$636
$118
$754
MHDDE
343
213
556
1,051
156
1,207
HHDDE
403
249
652
1,252
192
1,444

-------
3-57
with particulate-trap systems. Allowing for the reduced
installation rate of particulate-trap systems which results
from averaging (30 percent are not expected to use trap
systems), the per engine cost for all HDDEs is $373.
c. Total Cost to Manufacturers of the 1990 Standards
The total RD&T and emission hardware costs (undiscounted)
are shown in Table 3-23 for the years in which they are
expected to occur. The undiscounted costs to manufacturers of
RD&T and emission control hardware are estimated to be $354.3
million for the NOx standard and $511.3 million for the
particulate standard. The projected overall total for the 1990
NOx and particulate standards is $865.6 million as shown in
Table 3-23. The discounted costs are shown in Table 3-24 and
are $330.7 million for the NOx standard, $483.2 million for the
particulate standard for a total of $813.9 million.
3. Costs to Users (HDDE - NOx and particulate)
a. Increases In First Cost
The amount that a manufacturer must increase the price of
i,ts HDDEs to recover its expenses depends on the timing of the
costs and on the cost of capital to the manufacturer. In the
case of HDDEs, new NOx and particulate standards will become
effective in the 1987 and again in the 1990 model years. It is
expected that manufacturers of HDDEs face a 10 percent cost of
capital and that they will increase the engine prices to
recover their pre-production costs in the three model years
following introduction of each of these new standards. The
increase in first cost is determined below on the basis of each
model year in which new standards become effective and for the
contribution by each of the standards as well as for the total
increase.
1987 Model Year Standards: The average increase in first
cost of HDDEs would consist of the sum of the hardware cost as
shown in Table 3-25 plus the discounted fixed cost as shown in
Table 3-26. For the NOx standard, these costs are $6.10 for
hardware and $9.59 for the fixed costs for a total of $15.69 or
approximately $16. In the case of the particulate standard,
the contributions to the total first cost increase of $30.98 or
approximately $31 are $20.98 from fixed costs and $10.00 from
hardware costs. The total increase in first cost is the sum of
the costs from the two standards and is $46.67 or approximately
$47.

-------
3-58
Table 3-23
Heavy-Duty Diesel Engines,
Manufacturers Undiscounted Compliance
Costs for the 1990 MY NOx and Particulate Standards
Year
1986
1987
1988
1989
1990
1991
1992
RD & T
NOx
7,0 00K
$ 15,OOOK
$6,700K
Particulate
$ 8, OOOK
20,000k
$14,000K
$7,500K
$28,700K
Total Undiscounted Cost:
$49, 500K
Emissions Hardware
NOx
Particulate
$ 10 3,096K $146,216K
$108,093K $153,303K
$114,405K $162,255K
$32 5,594K $461,774K
NOx: $3 54,294K (1984 dollars)
Particulate: $511,274K (1984 dollars)
Total: $86 5,568K

-------
3-59
Table 3-24
Heavy-Duty Diesel Engines,
Discounted Cost of Compliance for
the 1990 Model Year NOx and Particulate Standards
Year

Cost

Present Value
in 1990
NOx
Part.
Total
NOx
Part.
Total
1986
--
$8,000K
$8 ,000K

$11,713K
$ 11,713K
1987
7,000K
20,000K
27,000K
9,317K
26,620K
35,937K
1988
15,000K
14,000K
29,000K
18,150K
16,940K
3 5,090K
1989
6,700K
7, 500K
14,200K
7,370K
8 ,2 50K
15,620K
1990
103,096K
14 6,216K
249 ,312K
103,096K
146,216K
249,312K
1991
108,093K
153,303K
261,396K
98,266K
139,366K
237,632K
1992
114,405K
162,255K
276,660K
94,550K
134 ,095K
228,645K
Total
354 ,294K
511 ,27 4K
86 5,568K
330,749K
483 , 200K
813,949K
Lump
sum investment cost
per engine
made at
the start <
of 1990:

for the NOx
standard:

$267.16

for the particulate standard: $390.31
for both standards:	$657.47

-------
3-60
Table 3-25
Heavy-Duty Diesel Enqines, Undiscounted
Costs of Compliance with the Standards per
Eriaine Cor Engines Produced Purina 1987-89 and 1990-92
1987-89
1990-92
Manufacturer Cost
NOx
Part.
Total
NOX
Part.
Total
Research, Development &
Test inq
Emission Control Hardware
Total
	Operating Cost	
$8.45 $17.89 $26.34 $23.18 $39.98 $63.16
6.10 10.00 16.10 263.00 373.00 636.00
$14.55 $27.89 $42.44 $286.18 $412.98 $699.16
Change in Fuel Cost per one
Percent Change in Fuel Economy for	LHDDE
for	MHDDE
for	HHDDE
$ 86.02
$389.61
$900.43
for all HDDE: $503.97

-------
3-61
Table 3-26
Heavy-Duty Diesel Enqines, Discounted
Costs of Compliance per Enqine for
Enqines Produced During 1987-89 and 1990 — 92 Tl]
Manufacturer Cost
Research, Development &
^estinq
Emission Control Hardware
Total
ODeratinq Cost
1987-89
1990-92
NOx
Part.
Total
NOx
Part.
Tota 1
$9.59 $20.98 $30.57 $28.14 $51.32 $79.45
5.55 9.09 14.64 239.02 339.00 578.02
$15.14 $30.07 $45.21 $267.16 $390.31 $657.47
Chanqe in Fuel Cost per one
Percent Chanqe in Fuel Economy for LHDDE
for	MHDDE
for	HHDDE
$ 58.13
$274.18
$660.55
for all HDDE: $365.16
[d] 10 percent discount rate to year standard
is introduced.

-------
3-62
1990 Model Year Standards: The components of the first
cost increase for the 1990 standards are shown in Tables 3-25
and 3-26. The first cost components of the NOx standard per
engine for all engines built are estimated to be $28.14 for
fixed costs and $263 for hardware for a total first cost
increase attributed to the NOx standard of $291.14 or
approximately $291. On the basis of all engines built, the
first cost increase resulting from the particulate standard is
$424.31 or approximately $424 consisting of $51.31 fixed cost
and $373.00 hardware cost. The overall first cost increases
for the NOx and particulate standards is estimated to be
$715.45 or approximately $715. For the three HDDE classes, the
first price increases are $597 for LHDDEs, $732 for MHDDEs and
$799 for HHDDEs on the basis of all engines. The first cost
increases for engines using particulate trap systems are; $705
for LHDDEs, $898 for MHDDEs, $994 for HHDDEs, and $875 for the
sales weighted average.
b. Fuel Economy
In Chapter 2, Technological Feasibility, it was estimated
that fuel economy penalties associated with the 1987 and 1990
model year NOx standards would be in the range of 0 percent to
2 percent for each of the standards in the short term. These
penalties should both tend to disappear with time as further
engine and vehicle improvements are made. In the case of the
particulate standards, a long term fuel economy penalty of
between 1 and 2 percent may be associated with the 1990
standard. \ fuel economy penalty associated with the 1987
particulate standard is not expected.
In Reference 1, fuel economy estimates for LHDDEs, MHDDEs,
and HHDDEs of 15.5 mpg, 8.4 mpg, and 7.0 mpg, respectively were
developed. Combining these fuel economy values with a fuel
cost of $1.20 per gallon and with annual mileages and lifetimes
of 11,000 miles per year for 10 years, 30 , 000 miles per year
for 9 years, and 65,000 miles per year for 8 years for LHDDEs,
MHDDEs, and HHDDEs, respectively, allows development of fuel
economy cost estimates. A 10 percent discount rate is employed
and the fuel economy cost estimates for each engine class were
sales weighted in arriving at the average cost per engine and
total average lifetime cost. The sales fractions used are 33
percent LHDDE, 25 percent MHDDE, and 42 percent HHDDE.
Based on the above figures, each one percent reduction in
fuel economy corresponds to an annual increase in fuel usage of
7.17 gallons for LHDDEs, 36.07 gallons for MHDDEs, and 93.80
gallons for HHDDEs. These increases in fuel usage correspond
to lifetime costs of $58.13 for LHDDEs, $274.18 for MHDDEs, and
$660.55 for HHDDEs. Sales weighting of these costs gives the
average lifetime cost for a 1 percent change in fuel economy of
$365.16.

-------
3-63
Apolying these average costs, the range in the fuel
economy cost, per engine, which corresponds to the 0 to 2
percent change in fuel economy expected to be associated with
the 1987 NOx standard is $0 to $730.32. It is expected that no
fuel economy costs will be associated with the particulate
standard for 1987. In terms of the overall fuel economy costs
for the 1987 standards the range is, therefore, between $0 and
$730.32 per engine. Foe the engines produced in 1987 through
1989, the total cost for the fuel economy penalty could be up
to $804 million.
In the case of the 1990 model year standards, fuel economy
costs are expected to result from both the NOx and the
particulate standards. For the NOx standard, the range for the
fuel economy penalty is between $0 and $730.32 while the range
for the particulate standard is between $365.16 and $730.32.
The range, per engine, for the overall fuel economy effect for
the two standards is, therefore, between $365.16 and
$1,460.64. Because of the previously stated expected removal
of any fuel economy oenalty due to the NOx standard through
engine and vehicle improvements, the upper bound of this range
can be expected to decrease to $730. Expressed in terms of the
total number of engines which are expected to be produced in
1990 through 1992, the total cost is projected to be between
$452 million and $1.81 billion initially, decreasing to an
upper limit of $905 million.
c. Maintenance
Maintenance costs are not expected to be affected by
either the 1987 NOx or particulate standards for HDDEs.
In the case of the 1990 standards some increase in
maintenance costs, for some of the engines, can reasonably be
expected for both the NOx and the particulate standards. For
those engines which add a separate air-to-liquid aftercooling
system (50 percent of turbochared engines or 31 percent of all
engines), maintenance to drain, flush and refill the system can
be expected to occur twice during the lifetime of the engine.
It is also reasonable to expect that the circulation pump drive
belt will be replaced when the other maintenance is performed.
The cost at each of these maintenance points is estimated to be
$15 for labor and materials for a total lifetime cost of $30.
When expressed in terms of all HDDEs, this maintenance cost
represents a per engine cost of $9.30. Application of
air-to-liquid aftercoolers can reasonably be expected to occur
on MHDDE and HHDDEs exclusively. With lifetimes of nine years
and eight years for MHDDE and HHDDE respectively, maintenance
of the aftercooler system can be expected to occur at three

-------
3-64
years and six years into the lives of the engines. Discounting
the maintenance costs at 10 percent per year gives a total
discounted cost of approximately $20 per engine requiring the
maintenance and just over $6 per engine when expressed in terms
of all HDDEs.
In Reference 1, estimates were developed for the effects
on maintenance costs of particulate-trap systems. It is
expected that the trap-regeneration system will require
maintenance once during the engine lifetime. The cost for this
maintenance was projeted to be $35 per engine for LHDDEs and
$70 per engine for both MHDDEs and HHDDEs occuring
approximately fives years after the engine was purchased. It
is also expected that there will be a maintenance saving
attributable to the particulate standard. This saving would
come from the use of stainless steel exhaust pipe which would
last the life of the engine in place of the presently used mild
steel which could be replaced once during the engine's
lifetime. The savings in maintenance which were estimated in
Reference 1 for the three HDDE classes were $63, $98, and $156
for LHDDE, MHDDE and HHDDEs respectively. The net effects on
maintenance costs for the particulate standards are, therefore,
savings of $28 for LHDDE and MHDDE and $86 for HHDDE. When
discounted to the year of purchase, the savings in maintenance
are $17 for LHDDE and MHDDE and $53 HHDDE. The average
sales-weighted saving per engine is approximately $32. The
overall effect on maintenance costs for the 1990 NOx and
particulate standards is a saving of approximately $26 per HDDE
"($3 2-$6) .
d. Total Costs to Users
In summary, owners of vehicles which are equipped with
HDDEs can be expected to pay an average of approximately $47
more in 1987 and $715 more in 1990 for these vehicles than they
would have paid without the introduction of the NOx and
particulate standards.
In teems of fuel costs, the increased average lifetime
cost per vehicle is expected to be between $0 and $730 as a
result of the 1987 standards and to be between $365 and $1,461
for the 1990 standards.
Maintenance costs are not expected to be affected by the
1987 standards. Savings of approximately $26 per engine are
expected to occur as a result of the 1990 standards. The total
costs of the 1987 and 1990 standards to users of HDDEs are
shown in Table 3-27.

-------
3-65
Table 3-27
Heavy-Duty Diesel Engines;
User Cost per Enqine for the
1987 and 1990 Standards (discounted*
to year that standard becomes effective)
1987 Standards	1990 Standards
$47	$715
0	($26)
$0 to $730	$365 to $1,461
10 percent Der year.
First Price
Ma intenance
Fuel

-------
3-66
4. Aggregate Costs
The aggregate costs are shown in Table 3-28. por the 1987
HDDE standards the fixed and hardware costs are $49.8 million
and between $0 and $804 million for fuel. The aggregate cost
for the 1987 standards is, therefore, between approximately $50
million and $854 million. In the case of the 1990 model year
standards the fixed arid hardware costs are $814 million,
between $452 million and $1.81 billion for fuel and maintenance
savings of $32 million for an aggregate cost range of between
$1.23 billion and $2.59 billion.
The aggregate costs for the 1990 standards (HDGEs and
HDDEs) were previously presented as representing the equivalent
of a one time expenditure made at the time that the standards
become effective, i.e., 1990. The aggregate costs of the 1990
standards can also be treated as a one time expenditure made in
1987. This approach would result in all aggregate costs of the
component parts of this action being expressed as lump sum
expenditures in a common year. On this basis, the aggregate
costs in 1987 are $1.08 billion for LDTs, $16.0 million for
HDGEs, $661.4 million for HDDEs, and between $316 million and
$2.14 billion in fuel attributable to HDDEs.
IT. Socioeconomic Impact
The costs of the standards which were developed in the
first section of this chapter are used here to analyze the
effects of the standards on manufacturers, regional employment,
and the nation as a whole. The effects on manufacturers will
be analyzed with respect to sales, capital investment, and cash
flow. Nationally, the effects of the standards will be
analyzed with respect to vehicle purchasers and users, energy
usage, balance of trade, and inflation.
A. Effects on Manufacturers
1. Sales
Estimates of the effects of an increase in price on
vehicle sales can be developed from the price elasticity of
demand for the vehicle type under analysis, the price increase,
and the price of the vehicle type prior to the increase.
Jack Faucett Associates (JFA) estimates that the price
elasticity of demand for all LDTs is approximately -1.0.[6]
This value is considered by J FA to be equally applicable to
diesel and gasoline-powered LDTs. In the 1983 model year,
prices of LDTs ranged from a low of approximately $6,000 for a
small LDT with the base gasoline engine and no optional

-------
3-67
Table 3-28
Heavy-Duty Diesel Enqines,
Agqregate Cost of Compliance for
the 1987 and 1990 Standards (10 percent
discount to January of year standard is introduced)
Research, Development &
Testing
Emission Control Hardware
Fuel Economy
Ma intenance
1987 Standard
$33,660K
$16,118K
$0 to $804,082K
0
1990 Standards
$98, 360K
$715,5S9K
$452 , 068K to $1 , 808 , 272K
($32 , 188K)
Total
$49 , 778K to $8 5 3 , 860K $1,233,829K to $2 , 590,033K

-------
3-68
equipment to a high of about $15,000 for a large LDT fully
equipped with optional accessories and powered by a diesel
engine. Prices for an average gasoline LDT and an average
diesel LDT would be within these limits and can be considered
to be approximately $9,000 and $10,000, respectively. The
average price increases for gasoline LDTs and diesel LDTs
attributed to the NOx standard are $87 and $35, respectively.
These numerical price increases represent a 0.97 percent
increase in the price of a gasoline LDT and a 0.35 percent
increase in the price of a diesel LDT. With a price elasticity
of demand of -1.0, the price increases are projected to result
in a 0.97 percent reduction in the sales of gasoline LDTs and a
0.35 percent reduction in the sales of diesel LOTs.
The historical trend in LDT sales has been that of a
growth market; i.e., yearly sales increases, with the exception
of the recent past where sales of all consumer goods have
declined. During 1983, sales of LDTs have increased with the
general strengthening of the national economy. It is expected
that the return to the historical pattern will continue and
that the projected sales decrease will be reflected as a
smaller rate of growth in sales pather than an actual decrease
in sales. Values for cross-price elasticity of demand were not
available and it was not possible, therefore, to develop an
estimates of the shift in demand between engine types {either
from gasoline to diesel or from diesel to gasoline) . Tt
appears reasonable, however, that some very small shift toward
diesels may occur because of the greater price increase of
gasoline LDTs. It is expected that there will be no
significant effect on either fuel economy or maintenance as a
result of the NOx standard. Neither of these factors are
expected, therefore, to have any effect on LDT sales.
Development of estimates for the effects of the new
standards on sales of HDTs will follow the same procedure as
were used for LDTs. The price range for HDTs starts at
approximately the upper level for LDTs ($15,000), and continues
through approximately $100,000. Representative first prices
for the groups of vehicles included in the HDT class are:
$18,000 for a LHDGT, $20,000 for a LHDDT, $25,000 for a MHDGT,
$28,000 for a MHDDT, and $55 ,000 for a HHDDT. In Reference 7,
Jack Faucett Associates gave price elasticity of demand values
for four of these HDT classes as follows: -1.0 for LHDGTs,
-1.0 to -1.2 for MHDGTs, -1.0 to -1.5 for MHDDTs, and -0.35 to
-0.8 for HHDDTs. A value was not given for LHDDTs. For this
analysis, a value of -1.0 is assumed for LHDDTs which is the
same value as is used for LDDTs and LHDGTs.

-------
3-69
In the cost section of this chapter, the price increase
for HDGEs was estimated as being $2 in 1987 and $18 in 1990.
Expressing these orice increases as percentage changes of first
price and incorporating the price elasticity of demand values,
give projected percentile reductions in HDGT sales of between
0.01 and 0.02 in 1987 and between 0.06 and 0.09 in 1990.
For the 1987 standards, the average price increase for
HDDEs was estimated to be $47. This price increase, coupled
with the highest price elasticity values for each class,
results in projected sales reductions of 0.17 percent for
LHDDTs, 0.18 percent for MHDDTs, and 0.05 percent for HHDDTs.
For the 1990 model year standards, the average price
increase for HDDEs was estimated to be $715. Because of the
variation by class in the first prices and the projected price
increases of HDDTs, the average price increase will not be used
in the development of estimates of the effects on sales. For
LHDDTs, the estimated price increase of $597 ($291 from the NOx
standard and $306 from the particulate standard) represents a
2.99 percent increase in first price. For MHDDTs and HHDDTs,
the estimated price increases of $732 ($291 for NOx and $441
for particulate) and $799 ($291 for NOx and $508 for
particulate) represent 2.61 and 1.45 percent increases in first
price, respectively. Combining these values with the price
elasticity values from Reference 7 results in projected
decreases in sales of 2.99 percent for LHDDTs, between 2.51 and
3.92 percent for MHDDTs, and between 0.51 and 1.16 percent for
HHDDTs. The fractions of the projected sales decreases
attributable to the NlOx and particulate standards are
proportional to the fractions of the cost increase contributed
by each standard. For LHDDTs, the NOx standard contributes a
projected sales decrease of 1.46 percent and the particulate
standard contributes 1.53 percent of the total 2.99 percent.
For MHDDTs, the percentile sales reduction range for the NOx
standard is between 1.04 and 1.56 percent while the range for
the particulate standard is 1.57 to 2.36 percent.
Corresponding values for HHDDTs are between 0.19 and 0.42
percent for the NOx standard and betwen 0.32 and 0.74 percent
for the particulate standard.
For the 1987 standards, the effects of first price
increases on sales for both gasoline and diesel-powered HDTs
are projected to be so small as to be reasonably considered to
be negligible. For the 1990 standards, it is also reasonable
to conclude that the sales of gasoline-powered HDTs will not be
significantly impacted by the increase in first price. For
diesel-powered HDTs, the first price increases could result in
a reduction in sales on the order of 2-1/2 to 4 percent in the
light-heavy and medium-heavy duty classes. The projected

-------
3-70
reduction in sales for the heavy-heavy duty class is on the
order of 1/2 to 1 percent which is significantly less than in
the other two classes.
For HDGEs, it is not expected that the new emission
standards will have any effects on either fuel economy or
maintenance and consequently it is not expected that these
factors will have an impact on sales.
For HDDEs taken as a group, it is expected that the 1987
standards will result in a sales-weighted average cost
increases for fuel of between $0 and $730 attributable totally
to the NOx standard. For the three HDDE classes, the cost
increases for increased fuel usage are between $0 and £116 for
LHDDEs, between $0 and $548 for MHDDEs, and between $0 and
$1,321 for HHDDEs. It is expected that there will be no effect
on maintenance costs associated with the 1987 standards for
HDDEs. Combining the fuel and maintenance costs (zero) with
the operating cost elasticity of demand (-0.3 to -0.9) given in
Reference 6 results in projected percentile sales decreases for
the three classes of HDDEs as follows: 1) for LHDDEs a sales
decrease of between 0 and 0.52 percent, 2) for MHDDEs a sales
decrease of between 0 and 1.76 percent, and 3) for HHDDEs a
sales decrease of between 0 and 2.16 percent.
For the 1990 standards, the increases in fuel cost are
contributed approximately equally by the NTOx and particulate
standards. The ranges for the three HDDE classes are: 1)
between $58 and $232 for LHDDEs, 2) between $274 and $1,096 for
MHDDEs, and 3) between $661 and $2,644 for HHDDEs. For the
1990 standards, it is expected that there will be reductions in
maintenance costs for LHDDEs and MHDDEs of $11 and of $47 for
HHDDEs. The net increases in operating costs (fuel costs -
savings in maintenance) are projected to result in reduced
sales of the three HDDE classes as follows; 1) between 0.07
and 0.99 percent for LHDDEs, 2) between 0.28 and 3.49 percent
for MHDDEs, and 3) between 0.33 and 4.25 percent for HHDDEs.
The NOx and particulate standards make approximately equal
contributions to the projected reductions in sales.
The projected effects on sales of LDTs, HDGEs, and HDDEs
resulting from first price increases and changes in operatinq
costs are summarized in. Table 3-29. For the 1987 standards,
the projected effects on sales are as follows: essentially no
effect in the case of HDGEs, under 1 percent reduction in sales
of LDGTs and LDDTs stemming from first price increase and up to
approximately a 2-1/3 percent decrease in the sales of HDDE
caused mainly by increased fuel usage. Sales of HDGEs are not
expected to be significantly affected by the 1990 standards.
Sales of LHDDEs could be reduced between 3 and 4 percent as a

-------
3-71
Table 3-29
Projected Effects Of the 1987 arid 1990 Standards
On Sales Of LDTs, HDGEs, and HDDEs (percentile reductions)


1987 Standards


1990
Standards


LDGT
LDDT
HDGE
HDDE
HDGE
LHDDF,
MHDDE
HHDDE
First
0.97
0.35
0.01
0.05
0.06
2. 99
2.61
0.51
Pr ice


to
to
to

to
to
Increase


0.02
0.18
0.09

3.92
1.16
Operating
0
0
0
0
0
0.07
0.28
0.33
Costs



to

to
to
to




2.16

0.99
3.49
4.25
Total
0.97
0.35
0.01
to
0.02
0.05
to
2.34
0.06
to
0.09
3.06
to
3.98
2.89
to
7.41
0.84
to
5.41

-------
3-72
result of the 1990 standards, with first price increase playing
the major roll in the decrease. For MHDDEs, first price
increase and increased fuel usage contribute about equally to
the projected sales decrease of between 3 and 7-1/2 percent.
For HHDDEs, the projected sales decrease is from just under 1
to 5-1/2 percent caused mainly by increased fuel usage.
Because the costs of purchasing and operating diesels are
expected to be increased by the 1990 standards with little
corresponding change for gasoline engines, it is reasonable to
expect that some small shift to gasoline engines could occur.
Any shift which may occur is not expected to be significant,
however, because the inherent fuel economy advantage and
consequently operating cost advantage of diesel engines will
not be compromised.
Like the LDT industry, the HDE industry has historically
exhibited a growth trend. This trend is expected to continue.
The projected reductions in sales of HDEs are expected,
therefore, to be translated into a reduction in the rate of
growth of sales rather than an actual reduction in sales.
2. Development Costs and Cash Flow
Manufacturers of LDTs and HDEs will not have to make any
substantial investments as a result of either the 1987
standards or the 1990 standards. Preproduction costs are not
very large and will be recovered within 3 to 5 years of the
effective dates of the standards.
The bulk of the development costs associated with the
required use of trap-oxid izer s is expected to be borne by the
manufacturers of emission control equipment, such as Corning,
NGK, and Johnson-Matthey. The manufacturers of reducing
catalysts which will be employed on LDTs have already developed
the necessary technology. Even though the manufacturers of
emission control equipment will have to finance the necessary
investments, they all have indicated their willingness and
ability to supply this market. Thus, pre-production investment
costs should not be a problem for any affected entities.
The impact on cash flow will result from the inventory of
reducing catalysts and particulate traps which a manufacturer
has on-hand at any given time. In Reference 1, 90 days were
identified as the average period of time that an LDT is
normally held by a manufacturer, with a correspondina figure
for HDDVs of under four months. These turnover periods should
be short enough not to significantly affect a manufacturers
cash flow.

-------
3-73
B.	Regional Effects
Manufacturers operate about 22 plants in assembling LDTs
in the U.S. The locations for each manufacturer are shown in
Table 3-30. A total of 12 states are included. General Motors
operates nine assembly plants to produce LDTs and vans. Some
of these plants are also used to assemble passenger cars. GM
has plants located in seven states with three in Michigan.
Ford Motor Company (Ford) operates seven assembly plants spread
over seven states. Chrysler and American Motors Corporation
(AMC) currently operate two each and IHC and Nissan operate one
each. Domestic manufacturers of HDEs and HDVs currently
operate assembly plants in 17 different states, as shown in
Table 3-31.
As can be seen in Tables 3-30 and 3-31, there is a broad
dispersion across the country of manufacturing facilities which
could be impacted by the standards. Because of this dispersion
of facilities, any negative effect of the standards with
respect to employment will not be concentrated in one or more
areas of the country. As was stated previously, it is expected
that the cost increases associated with the standards will
translate into a decrease in the rate of growth of the affected
industries rather than an actual reduction in sales. The
impact on employment can reasonably be expected, therefore, to
be either a slowing in the rate at which new employees are
hired or a reduction in overtime rather than a reduction in
jobs.
C.	National Effects
1. Vehicle Purchasers and Users
Purchasers of LDTs and vehicles equipped with HDEs could
be affected to some degree by the higher vehicle first costs.
It is not expected, however, that the effects will be
significant because the price increases are small when
expressed as a percentage increase in first price (i.e., from a
low of essentially zero to a high of approximately three in the
case of LHDDTs). The expected average sales-weighted first
cost increase in 1987 of $74 for LDTs and of $715 in 1990 for
HDDE should not substantially impact the purchaser's ability to
pay for either a new LDT or a new HDV.
Users of HDVs powered by diesel engines can be expected to
be affected to some degree by the higher first cost and
operating cost of vehicles employed to transport goods. These
increased costs can reasonably be expected to be passed on as
higher freight costs which in turn will affect the prices
consumers pay for the products transported by diesel trucks.

-------
3-74
Table 3-30
Light Truck Assembly Plants
in the United States
General Motors:
Fremont, California
Lakewood, Georgia
Baltimore, Maryland
Detroit, Michigan
Flint, Michiaan
Pontiac, Michiqan
St. Louis, Missouri
Lordstown, Ohio
Janesville, Wisconsin
Ford:
San Jose, California
Louisville, Kentucky
Wayne, Michiqan
Twin Cities, Minnesota
Kansas City, Missouri
Lorain, Ohio
Norfolk, Virqinia
Chrysler:
Warren, Michigan
St. Louis, Missouri
American Motors:
Toledo, Ohio
South Bend, Indiana
International Harvester:
Fort Wayne, Indiana
Nissan:
Smyrna, Tennessee

-------
3-75
Table 3-31
States with HDE and HDV Manufacturinq Plants
HDE Plants
HDV Plants
California
111inois
Indiana
Kentucky
Michigan
Oh 10
Oklahoma
Pennsylvania
California
Indiana
Missour i
New York
Ohio
Oreqon
Pennsylvania
Tennessee
Texas
Utah
Virginia
Washington
Wisconsin

-------
3-76
Freight costs are, however, only a relatively small fraction of
the cost of a product and any small increase in freight cost
will be reflected as a proportionally smaller increase in
product cost.
2.	Energy Usage
The effects on energy usage of these standards are
expected to be limited to HDDEs. As stated previously, each 1
percent decrease in the fuel economy of a HDDE means a lifetime
operating cost increase of approximately $365. In terms of the
increased volume of fuel used, each 1 percent reduction in fuel
economy represents an increase in lifetime fuel usage of 420
gallons for an average HDDE. On an annual basis, the average
HDDE is expected to use an additional 50.8 gallons of fuel for
each 1 percent reduction in fuel economy. With an average
annual sales volume of 413,000 HDDEs (mean of 1990-92 sales
estimates) and fleet replacement every 9 years, each 1 percent
reduction in fuel economy results in increased fuel usage of 21
million gallons in the first year of the standards increasing
to 189 million gallons in the tenth year of the standards.
Combining these values with the projected fuel economy effects
of the 1987 NOx standard (i.e., between zero and two percent
reduction) gives projected increases in fuel usage of between 0
and 42 million gallons in the first year of the standard
increasing to between 0 and 378 million gallons in the tenth
year of the standard. For the 1990 NOx and particuate
standards, increased fuel usage is expected to be between 21
million and 84 million gallons in the first year of the
standard, increasing to between 189 million and 756 million
gallons in the tenth year of the standard. The yearly effects
of the HDDE standards on fuel usage are shown in Table 3-32.
3.	Balance of Trade Effects
The implementation of three-way catalyst technology over
oxidation catalyst technology on LDTs will cause an increase in
the imports of noble metals, primarily rhodium and platinum.
Imports of palladium are expected to decrease.
The incremental changes in noble metal imports discussed
below are the changes which are expected to occur as a result
of the implementation of the revised NOx standard for LDTs and
incorporate the 4-, 6-, and 8-cylinder engine sales split
expected in the late 1980's. The change in these imports is
shown below:

-------
3-77
Table 3-32
Yearly Increases in Fuel Usaqe by HDDEs
Resulting From the 1987 and 1990 Standards
Increased Annual
Fuel Usaqe
Year	(millions of gallons)
1988
0
to
42
1989
0
to
84
1990
0
to
126
1991
21
to
210
1992
42
to
294
1993
63.
to
378
1994
84
to
462
1995
105
to
546
1996
126
to
630
1997
147
to
672
1998
168
to
714
1999
189
to
756

-------
3-78
Annual Noble Metal Imports Due to LDTs (Troy oz)
1987
No Revised NOx Std
L987
Revised NOx Std
Platinum
Palladium
Rhodium
158,386
87,227
0
389,829
47,329
32,811
At prevailing market prices, these figures translate into
increases of approximately $74.8 million for platinum, $20.2
million for rhodium, and a decrease of $6.9 million due to
palladium. The sum of these three figures yields an increase
in imports of about $88.1 million per year. These figures
depend on the composition and loading of catalytic materials
used on each vehicle and will vary as a function of the
differences between the systems used by manufacturers and those
assumed in this analysis.
Another effect on the balance of trade of the standards
comes from the first price increase of imported LDTs, most of
which use 4-cylinder gasoline engines. Based on the first price
increases in Table 3-13, this comes to about $120 per
4-cylinder LDT (<($108 + $131)/2). If it is assumed that 40
percent of the 4-cylinder market is imported, this average
price increase yields a loss in the balance of trade about
$50.1 million per year based on the average of the yearly sales
shown in Table 3-1.
The third effect on the balance of trade of the new
standards comes from increased fuel usage by HDDEs. As was
previously estimated, increases in fuel usage are expected to
be between 0 and 42 million gallons in 1988 rising to between
189 million and 756 million gallons in 1999. Assuming that
each additional gallon of fuel used represents an increase in
oil imports of 1 gallon*, and using a cost of $30 per barrel (1
barrel = 42 gallons) , increased fuel usage by FTDDEs translates
into increased oil importation costs of up to $30 million in
1988 and between $135 million and $540 million in 1999.
4. Inflationary Effects - Consumer Price Index
While one gallon of imported crude oil does not produce
one gallon of diesel fuel, the other products produced
from the crude oil would tend toward a national over
supply of these products which could then be exported.
The national balance of trade would, therefore, be
approximately as assumed.

-------
3-79
The consumer price index (CPI) is one of the primary
indications for changes in the general price level. It has
been estimated that LDTs contribute about 0.5 percent to the
CPI determination and that HDEs contribute a similar
percentage. [8] Combining these percent contributions with the
average estimated price increase of about 2 percent for both
LDTs and vehicles using HDEs, will give only a .01 percent
increase in the CPI for each of the regulated classes of
vehicles. Needless to say this increase is negligible compared
to other elements of the CPI. Therefore, it is concluded that
these emission regulations will have no significant price level
impact as measured by the CPI. Further, since the public will
receive air quality and related health improvements in exchange
for the higher prices, the rise in the CPI that will occur
cannot properly be termed inflationary.

-------
3-80
References
1.	"Diesel Particulate Study," U.S. EPA, OAR, OMS,
ECTD, SDSB.
2.	"Data Resources U.S. Long-Term Review," TRENDLONG
0682, Data Resources Incorporated, Summer 1982.
3.	"The Impact of Light-Duty Diesel Particulate
Standards on the Level of Diesel Penetration in the Light-Duty
Vehicle and Light-Duty Truck Markets," Jack Faucett Associates,
EPA Contract No. 68-01-6375, January 17, 1983.
4.	"Cost Estimations for Emission Control Related
Components/Systems and Cost Methodology Description," Rath &
Strong Inc., EPA-460/3-78-002, March 1978.
5.	Oral Communication with the Bureau of Labor
Statistics.
6.	Oral communication with Jack Faucett Associates.
7.	"Estimation of Economic Elasticities In the
Heavy-Duty Vehicle Market," Jack Faucett Associates, EPA
Contract No. 68-01-6375, February 23, 19B3.
8.	"Preliminary Impact Assessment of the Non-Passenqer
Automobile Fuel Economy Standards for Model Years 1980 and
1981," Planning and Evaluation Office of Program Analysis, DOT,
NHTSA, November 29, 1977.

-------
Chapter 4
NOx ENVIRONMENTAL IMPACT
I.	Introduction
This chapter examines the environmental effects which can
be expected from implementing the proposed NOx standards for
light-duty trucks (LDTs) and heavy-duty engines (HDEs) in
non-California locations of the United States. The emission
inventory and air quality analyses presented in this chapter
are based on the implementation of split-class NOx standards
for LDTs and two-stage standards for HDEs, beginning in the
1987 model year. Class I LDTs, i.e., those up to 6,000 lbs
gross vehicle weight (GVW) or 3,999 lbs equivalent test weight,
would meet a standard of 1.2 g/mi, while LDTs from 6,001 to
8, 500 lbs GVW (Class Ila) would conform to a standard of 1.7
g/mi. Beginning in the 1987 model year, HDEs would meet a 6.0
g/BHP-hr NOx standard. For 1990 and later model year engines
this standard would be reduced to 4.0 g/BHP-hr.
The impact assessment begins with a brief review of the
most recent and significant health and welfare effects
associated with NOx emissions and provides references to more
comprehensive studies. Following this, LDT and HDE emission
rates are presented for both base-case (current levels of NOx
control) and controlled-case (new standards) scenarios. A
statutory-standard case is also included for comparison
purposes to show the reductions that would be possible if the
statutory standards were technologically practicable for all
HDEs. The effects of implementation of the above-mentioned
standards are then quantified in terms of the lifetime
per-vehicle NOx emission reductions and total area-wide NO2
inventory reductions from an analysis of Standard Metropolitan
Statistical Areas (SMSAs) which have high concentrations of
nitrogen dioxide (NO2) . The net effects of reductions in
total NOx tonnage for these designated SMSAs is then described
in terms of projected improvements in NO2 ambient air quality.
II.	Health Effects
A concise analysis of NO2 health effects is difficult to
present due to the abundance of NO2 health effects research
data and the scientific debate which results when interpreting
these studies. Two comprehensive sources of NO2 health
effects information are currently available and the concerned
reader is referred to these reports. They are the EPA
publications: "Air Quality Criteria for Oxides of Nitrogen"[1]
(the Criteria Document), and the "Review of the National
Ambient Air Quality Standards for Nitrogen Oxides: Assessment
of Scientific and Technical Information - OAOPS Staff Paper"[2]
(the OAQPS Staff Paper).

-------
4-2
An overview of the NO2 health effects may be provided as
follows. Mild symptomatic effects at NO2 concentrations
approaching short-term peak ambient exposures (0.5 ppm N02
for two hours) have been observed in asthmatics. In general,
adverse health effects, such as pulmonary function impairment,
are only observed at NOj concentrations in excess of 1.0 ppm,
while adverse health effects are inconclusive at exposure
levels below 0.5 ppm. However, the NO2 health effects
research contains convincing evidence from animal studies which
shows serious biolooical effects from lonq-term exposure to
slightly elevated ambient concentrations of NO?. These
findinqs suqgest a definite risk to human health from chronic
exposure to NO?, but such risks have not been quantified at
current ambient NO2 levels. T2] More recently, epidemioloqy
studies have suqgested concern over the health impacts of
repeated short-term exposure. These results report increased
rates of acute respiratory illness and impaired pulmonary
function, particularly among sensitive population groups, such
as children.[2]
The Criteria Document presents complete and detailed
information from animal toxicology, controlled human exposure,
and eoidemioloqy studies. A summary and analysis of selected
studies are presented in the OAQPS Staff Report, with
additional comments by the Clean Air Science Advisory
Committee, an independent oversight committee composed of
members from academia and industry which reviewed the OAOPS
Staff Report. [ 3]
III. Air Quality Analysis
The previous sections briefly summarized the health and
welfare effects of NOx in the environment and referenced more
definitive studies. This section will quantify the
contributions of LDTs and HDEs to local NOx inventories and
predict the resulting chanqes in ambient NO2 concentrations
if the NOx emission standards are changed. Specifically, this

-------
4-3
part of the analysis will present: 1) a discussion of EPA
mobile source emission factors, 2) an estimate of the per
vehicle reduction in NOx emissions, 3) base-case,
controlled-case, and statutory-case NO2 inventories for
standard metropolitan statistical areas (SMSAs) with relatively
hiqh ambient NO2 concentrations, and 4) projections of the
net effects on local ambient air duality that will result from
enacting new NOx standards for LDTs and HDEs.
The three scenarios used in this analysis are based on the
following NOx emission standards:
Effective Year	LDTs (g/mi)	HDEs (g/BHP-hr)
Base Case
2.3

10.7
Controlled Case:



1987
1.2
(Class
I) 6.0

1.7
(Class
Ila)
1990
1.2
(Class
I) 4.0

1.7
(Class
Ila)
Statutory Case:



1987
0.9

1.7
Base-case standards
represent
the
current NOx
standards. The 1.2/1.7 and 6.0 NOx standards are proposed" to
be effective for 1987 model year LDTs and HDEs, respectively,
and the 4.0 standard for HDEs would be effective with the 1990
model year. A third scenario is included representing the
statutory standards of 0.Q g/mi* for LDTs and 1.7 g/BHP-hr for
HDEs, effective with the 1987 model year.
A. Emission Factors
Emission factors are emission rates, expressed in qrams
per mile, used to calculate mass Quantities of pollutants
emitted. Mobile source emission factors are based on an
analysis of emission test results from in-service LDVs and LDTs
from the ongoing EPA Emission Factors Program, and from
baseline studies of heavy-duty engines. Emission factors
eauations take the form:
ER = ZM + DR(m)
0.9 g/mi represents a 75 percent reduction from baseline
levels. This standard is not being proposed because it
represents a stringency greater than that of the
light-duty vehicle standards. The 1.2 g/mi and 1.7 g/mi^
standards have been selected as equivalent in stringency
to light-duty vehicle standards of 1.0 g/mi and 1.5 g/mi

-------
4-4
Where ER is the emission rate, ZM represents zero mile
emissions and DR(m) is the deterioration rate per 10, 000 mile
increment multiplied by the mileaqe (in tens of thousands).
Once emission factor equations for each model year's vehicles
are established, composite emission factors for a particular
vehicle class and a specific calendar year are determined usinq
registration and mileage weighted emission factors.
Tables 4-1 and 4-2 summarize emission rate information for
four mobile source emission classes: 1) LDGTs, 2) LDDTs, 3)
HDGVs, and 4) HDDVs. The meaning of the "ZM" and "DR" columns
is described above. The "SEA" column refers to Selective
Enforcement Audit, which is EPA's assembly line testing program
performed at a manufacturer's assembly plant. The 40 percent
criterion represents the Acceptable Quality Level (AOL), or
maximum failure rate, included in SEA testing. The AOL affects
the emission factor calculation, and the results in Tables 4-1
and 4-2 reflect this adjustment. The "Useful-Life" column
desiqnates the period for which the manufacturers must certify
the emissions compliance of their vehicles/engines. These
periods are generally referred to as the half or full "useful
life" of the vehicle or engine. Terms of the emission factor
eauations for full-life periods are generally lower than for
half-life periods, since emission levels are required to be
below the standard for the full life of the vehicle or enqine,
rather than for only half the lifetime,.
As was discussed above, the mobile source emission factors
for particular vehicle/enqine types in aiven model years are
weighted according to number of registrations and mileaqe to
yield composite averaae emission rates for an emission class in
a specific calendar year. Because of this aggregation, mobile
source emission factors represent a matrix of emission results
based on time (model year) and mileage (deterioration rate).
The composite emission factors are calculated usinq a modified
version of the MOBILE2 program, an EPA computer model which
calculates composite NO2 emission factors for the following
vehicle types: 1) LDGVs, 2) 0-6,000 lb LDGTs (LDGT Is), 3)
6,001-8,500 lb LDGTs (LDGT lis), 4) HDGVs, 5) LDDVs, 6) LDDTs,
and 7) HDDVs. The MOBILE2 program is described in detail in
references [15] and [16].
B. Per-Vehicle NOx Reductions
A common method of quantifying the anticipated emission
reduction benefits of new emission standards is to determine
the difference in average per-vehicle lifetime emissions for
vehicles assumed to operate under the base- and controlled-case
emission standards. Lifetime emissions are straiqhtforward
calculations requiring only the emission rate data of Tables
4-1 and 4-2, and estimates of the lifetime mileages for the
appropriate vehicle classes. The results summarized in Tables
4-3 and 4-4 are based on the following lifetime mileages:

-------
4-5
Table 4-1
NOx Emission Rates and Assumptions
Different Than MOBILF2 '/aides for
Emission Inventory and Air OuaLitv Analysis
Low Altitude
Vehicle
Model
Emission Rate
Useful

Tvoe

Year
ZM [ 1]
DR[2]
SEA[3]
L i f e [ 4 ]
Base Case:
LDGT

1984
1985	+
1.89
1.85
0.10
0.08
40%
40%
Half
Full

LDDT

1984
1985	+
1.89
1.84
0.07
0.07
40%
40%
Half
Full

HDGV

1984
1985
193 5 +
9.45
9.45
9.45
0.09
0.07
0.07
40%
Half
Full
Full

HDDV

1984
1985
1986-95
26.72
26.04
24 .80
0.20
0.07
0.06
40%
Half
Full
Full
Controlled
Case :
LDGT
CI.
CI.
11 5]
Tla[6]
1934-86
1987 +
1987 +
1.00
1. 38
Sane
0.06
0.04
as Base
40%
40%
Case
Full
Full

LDDT
CI.
CI.
I [ 5 ]
Ha [ (51
1984-86
1987 +
1937 +
0.93
1.31
Same
0.06
0.06
as Base
40%
40%
Case
Full
Full

HDGV

1984-86
1987-89
1990 +
8.23
5.44
Same
O.Oo
0.05
as Base
40%
40%
Case
Full
Full

HDDV

1984-86
1987-89
20.94
Same
0.06
as Base
40%
Case
Full
1990-95 13.70
0.06
40%
Full

-------
4-6
Table 4-1 (cont'd)
NOx Emission Rates and Assumotions
Different Than MOBILE2 Values for
Emission Inventory and Air Quality Analysis
Low Altitude
Statutory Case:
Vehicle
Tyoe
LDGT
LDDT
HDGV
HDDV
Model
Year
1934-86
1987 +
1984-86
1987 +
1984-86
1987 +
1984-36
1987-95
Emission Rate	
7,M [ 11 PR [2] SEA [ 3 ]
Same	As Base	Case
0.76	0.07	40%
Same As Base	Case
0.71	0.05	40%
Same	as Base	Case
2.45	0.08	40%
^ame As Base	Case
6.15	0.05	40%
Useful
L i f e [ 4 ]
Full
Full
Full
[1]	Zero-mile emissions (a/ni) .
[2]	Deterioration rate (q/mi-lOK mi).
[3]	Selective ~nforce>nent Audit.
[4]	Certification to half or full useful iLfe.
[5]	Less than 6,000 lbs GVW.
[6]	6,001 - 8,500 lbs CVW.

-------
4-7
Table 4-2
NOx Emission Rates and Assumptions Different than
MOBILE2 Values for Emission Inventory and Air Quality Analysis
wiqh Altitude

Vehicle
Model
EnLss
ior.
TS.ate



Tvpe
Year
ZM

DR
SEA
Li fe
Base Case:
LDGT
1964
1985 +
1. 55
1. 51

0.09
0 .07
40%
40%
Half
Full

LDDT
1984
1985	+
1.89
1.84

0.07
0.06
40%
40%
Half
Full

HDGV
1984
1985
1986+
6.29
6.29
6.29

0.09
0.07
0.07
40%
40%
Half
Full
Full

HDDV
1984
1985
1986-95
26. 72
26.04
24 .80

0.20
0.07
0.06
40%
Full
Full
Cont rolled
Case:
LDGT
CI. I
CI. Ila
1984-86
1987 +
1987+
Same
0.90
1.13
as
Base
0.05
0.03
Case
40%
40%
Full
Full

LDDT
CL. t
ci. rra
1984-36
1987+
1987+
Same
0.93
1.	31
a s
Base
O.Qo
0.06
Case'
40%
40%
Full
Full

«DGV
1984-86
1987-89
1990 +
Same
5. 48
3.52
as
Base
0.06
0.05
Case
40%
40%
Full
Full

HDDV
1984-86
1987-89
1990-95
Same
20.<54
13 .70
as
Base
0.06
0 .06
Case
40%
40%
Full
Full
Statutory
Case:
LDGT
1984-86
1987+
Same
0. 74
As
Base
0.0 7
Case
40%
Full

LDET
1984-36
1987 +
Same
0. 71
As
¦Base
0.05
Case
40%
Full

HDGV
1984-06
1987 +
Same
2.32
As
Pase
0 .08
Case
40%
Full

HDDV
1984-86
1987-95
Same
6.15
AS
Base
0.0 5
Case
40 %
Full

-------
4-8
Table 4-3
Average Per-Vehicle
Total Lifetime NOx Emissions
and Percent deductions
Low Altitude
Vehicle Class
Liqht-Duty Gasoline
Tcucks - Class I [41
Base
Case[1]
(tons)
0.31
- Class IIa[5] 0.31
Controlled
Case[2]
(tons)
0.18
0.21
Statutory
Percent Case[3] Percent
Reduction (tons) Reduction
*1.6
30.5
0.16
0.16
<19.4
49.4
Liqht-Duty Diesel
Trucks - Class I [ 4] 0.30
- Class I la [ 5] 0.30
0.17
0.22
42.9
26.1
0.13
0.13
55 . 3
55.3
Heavv-Duty Gaso-
line (1987)
Heavy-Duty Gaso-
line (1990)
1.19
1.04
0. 69
13.0
41.9
0.35
70. 6
Heavy-Duty Diesel
(1987)
'-teavv-Dutv Diesel
(1990)
9.82
8.33
5.54
15.2
•13.6
2. 58
73.7
[1]	Base-case emissions are calculated usna 1985 model year
emission rates for LDTs and 1986 emission rates for HDVs.
[2]	Controlled-case emissions are calculated usina 1987 model
-/ear emission rates for LDTs and 1987 and 1990 model year
e.mssion rates for HDVs.
[3]	Statutory case emissions are calculated usnq 1987 -nodel
year emission rates for LDTs and HDVs.
[4]	0-6,000 lbs GVN
[5]	6,001-8,500 lbs GVW

-------
4-9
Table 4-4
Average Per-Vehicle
Total Lifetime NOx Emissions
and Percent Reductions
High altitude
Vehicle Class
Light-Duty Gasoline
Trucks-Class I[4]
- Class TIa[5]
Base
Case [ 1]
(tons)
0.26
0.26
Controlled
Case[2]
(tons)
0.16
0.17
Statutory
Percent Case[3] Percent
Reduct ion (tons) Reduction
37.8
32.1
0. IS
0.15
39.9
39.9
Light-Duty Diesel
Trucks- All
0.29
0.17
41.4
0.13
^4.1
Heavy-Duty Gaso-
line (1987)
Heavy-Dutv Gaso-
line (1990)
0.81
0.71
0.47
13.0
41.6
0.3 4
58.7
Keavv-Duty Diesel
(1987)
Ueavv-Dutv Diesel
(1990)
9.82
8.3 3
5. 54
15.2
13.6
2. 53
73.7
[1]	3ase-case emissions are calculated us na 1985 nodel year
emission rates for LDTs and 1986 emission rates for LIDEs.
[2]	Controlled-case emissions are calculated usmq 19 8*7 model
vear emission rates for LDTs and 1987 and 1990 nodel vear
emission r^tes for HDVs.
[3]	Statutory-case emissions are calculated usinq 1987 model
vear emission rates for LDTs and HDVs.
[41 0-6,000 lbs GVW
[5] 6,001-8,500 lbs GVW

-------
4-10
LDTs
120,
000
mi
HDGVs
110,
000
mi
HDDVs
350,
000
mi
The lifetime periods are based on analyses conducted as
part of the recently promulgated full-life useful-life
regulations for LDTs and HDVs. The LDT and HDGV lifetime
mileages are the typical usaqe periods for most vehicles in the
respective classes, while the HDDV lifetime mileage represents
a weiqhted composite for all HDDVs based on usaqe and rebuild
characteristics of a wide variety of vehicles and engines. [17]
Tables 4-3 and 4-4 show significant NOx reductions from
all truck classes as a result of implementation of the proposed
standards. HDVs show the greatest reductions in terms of
absolute tons of pollutant per vehicle, however, it is
important to remember that many more LDTs than HDVs are sold
each year, so LDTs do indeed contribute a substantial portion
of the total NOx inventory. Per-vehicle lifetime emissions
would be reduced by 2fi-43 percent for LDTs, while HDV emissions
would be reduced by 13-15 percent under the 1987 standards.
The NOx reduction for HDVs increases to 42-44 percent over
current levels with the implementation of the second staqe of
the HDE standards in 1990.
C. NOx Emission Inventories
1. NOx Problem Areas
NOx emission inventory projections and the resulting air
quality impacts estimates were performed for designated high
N02 SMSAs , where hiqh N02 areas are defined as
non-California SMSAs which exceeded N02 concentrations of
0.04 ppm (annual arithmetic mean) in the base year in
Question. This value ensures the inclusion of areas at or near
(within 25 percent) of the NAAOS of 0.053 ppm. The annual
average ambient N02 concentrations are termed "design
values," and are used to specify the base year condition for
calculating NOx emissions inventories and ambient air quality
Drojections. Values for 1980 and 1981 (the two latest
available) were used for purposes of this analysis. Use of a
ranqe in desiqn values provides a range of emission inventory
and air quality projections. This, in turn, removes the
sensitivity of the results to short-term changes in ambient
N02 concentration. For example, some recent reductions in
ambient N02 concentrations for the 1981 calendar year may

-------
4-11
reflect a reduction in industrial activity rather than a true
and stable improvement in air quality. The high NO2 areas,
which include six low-altitude and two hiqh-altitude areas, and
their 198D and 1981 design values, are presented in Table 4-5.
2. The EPA Rollback Model
Tables 4-1, 4-2, and 4-6 display the input data to the
modified MOBILE2 model, the results of which are then used in
the EPA Rollback Model to calculate the NOx emission inventory
and air quality projections for future years.[18] The Rollback
Model is used to model the relationship between chanqes in
source emissions and changes in air quality concentrations.
The model is linear, so chanqes in source emissions rates
result in proportional chanqes in emission inventories and air
quality concentrations. The information presented in Tables
4-1 and 4-2 has previously been described in the Emission
Factors discussion. The information presented in Table 4-6
requires some interpretation, however. The first column,
labeled "Designation," refers to the particular emission source
class. The LDT class includes information pertinent to both
LDGTs and LDDTs. The "Off-Highway" class includes mobile
sources such as aircraft, railroad, marine, construction, and
agricultural equipment. The "Combustion" source designation is
typically associated with NOx produced by furnaces and boilers
for commerical and residential heating. "Stationary Area"
sources include all other combustion sources and otherwise
unaccounted NOx sources in a particular area.
The mobile source arowth rate information presented in the
second column represents the estimated percent change, relative
to a base condition, of the numbers of vehicle miles travelled
which are expected for EPA's best estimate of a "medium" level
growth rate. The area stationary source qrowth rate
information for other sources is based on chanqes in the number
of sources.
The third column represents the replacement of existing
NOx sources with new ones. Because mobile source replacement
rates are a function of the source type and model year, they
are numerous, and therefore they are presented in the M0BILE2
documentation.[15,16]
The discount factors shown in Column 4 are used to adjust
emission sources for varying air quality impacts by inserting a
multiplicative factor in the M0BILE2 program. This factor may
vary from 0 to 1 and is used to adjust the emission inventory
values to more accurately model the given source category.
Discount factors other than 1.0 are often used to account for
the fact that stationary point sources with high effective.

-------
4-12
Table 4-5
Standard Metropolitan Statistical
Areas with High Ambient NO 2
NO2 Annual Average Concentration (ppm)
SMSAs	 1980 Design Values 1981 Design Values
Boston, MA
0.050
0.038
Chicago, IL
0.060
0.050
Cleveland, OH
0.048
0.039
Nashville, TN
0 . 047
0.049
Philadelphia, PA
0.046
0.046
Steubenville, OH
0.040
0.039
Denver-Boulder, CO[l]
0.046
0.044
Reno, NV[1]
0.048
0.043
TTj High-altitude areas.

-------
4-13
Table 4-6
Annual Growth and Replacement Rates, and Discount Factors
Growth Rate
1977-95	Replacement	Discount
Designation Medium (percent)	Rates (percent)	Factor
LDGV +1.7	As In M0BILE2	1.0
LDT +4.7	AS In MOBILE2	1.0
HDGV -0.3	As In MOBILE2	1.0
LDDV +1.7	As In MOBILE2	1.0
WDDV +6.4	As In MOBILE2	1.0
Stationary Point +3.5	+4.3	0.0
Off-Highway +2.5	0.0	1.0
Combustion +0.8	0.0	1.0
Stationary Area 0.0	0.0	1.0

-------
4-14
stack heights do not significantly affect the pollutant
concentration in the immediate local area. In this analysis,
stationary point source contributions are deducted from the
emission inventory and air quality calculations by using a
value of zero for the discount factor. This is because mobile
and area sources, rather than point sources, are likely to be
the principal cause of high annual average NO2 concentrations
at urban monitoring sites.
Figures 4-1, 4-2, and 4-3 summarize the discounted NOx
emission inventory results using the base-case scenario (no new
NOx standards), the controlled-case scenario (the proposed
standards), and the statutory-standard scenario, respectively,
for low-altitude areas. Each set of emission projections
includes total NOx emissions from six source types for each of
the calendar years 1980, 1983, 1987, 1988, 1990, and 1995,
along with the percentages of total emissions represented by
each class. Figures 4-4, 4-5, and 4-6 provide the same
information for high-altitude areas. Figures 4-7 and 4-8 show
the low- and high-altitude trends for emissions from individual
vehicle classes. The following sections discuss the results
for each of the three control scenarios separately.
3. Low-Altitude Emission Inventories
a. Base Case
The discounted results for the base case (i.e., no new
standards) presented in Figure 4-1 can most easily be
interpreted by examining the data trends for a particular
emission source as a function of time. For examnle, the
stationary area source contribution in Figure 4-1 is projected
to remain nearly constant through 1995, while the contributions
from off-highway sources is shown to increase in both total
number of tons and the percent of total NOx emissions. The
total increase for these two sources is 50,200 tons which
represents a 23 percent increase. The mobile source total
increases by some 71,600 tons, or an increase of 17 percent,
despite decreases in both the LDV and HDGV portions of the
inventory. The contributions from LDVs indicate a steady
decline in the mass quantities of NOx as vehicles designed to
meet a lower NOx standard replace older, higher emittinq
vehicles. Between 1983 and 1995 the LDT NOx contribution is
projected to increase by some 13,900 tons (an increase of 42
percent), despite the replacement of trucks designed to meet
3.1 g/mi NOx standards before 1979 by trucks certified to meet
the 2.3 g/mi standard starting with the 1979 model year. This
increase is due to the projected increase in LDT usage in
future years. HDGVs on the other hand show a slight decrease,
on the order of 1,500 tons or 5 percent, in total NOx emissions,
even without imposition of new NOx standards. This decline is
due to the decrease in the projected numbers of

-------
4-15
800-1
Figure 4-1
NOX EMISSIONS INVENTORY
HIGH N02 URBAN AREAS
Base Case Projections
Low Altitude
799.6
Legend
~ moo
ai hm
a LOT
S3 LDV
¦i orr-*oAO
Q STATIONARY S0UKX3


-------
4-16
800-)
Figure 4-2
NOX EMISSIONS INVENTORY
HIGH N02 URBAN AREAS
Controlled Case Projections
Low Altitude
Ltgcnd
~	hoo
a hoo
~	LOT
a LDV
¦ orr-ffOAO
Z2 stationary souaccs
^ n
-------
4-17
800-1
Figure 4-3
NOX EMISSIONS INVENTORY
HIGH N02 URBAN AREAS
Statutory Case Projections
Low Altitude
Legend
cz HOC *
a HOG
~ LOT
S3 LDV
¦ orr-aoAQ
ezi STATIONARY SOUPCES


-------
4-13
Figure 4-4
NOX EMISSIONS INVENTORY
HIGH N02 URBAN AREAS
Boa* Com Pro|«c1lona
High Mtltuda
Legend
o hoo
m mm
a LOT
a unv
morr-WAo
za ST*nONA*Y SOURCES
1	~f(* VtfA \/A	f

-------
4-19
O
©
1?
c
o
¦*-
o
o
o
c
o
'(A
V>
LJ
X
O
Figure 4-5
NOX EMISSIONS INVENTORY
HIGH N02 URBAN AREAS
Controlled Cqi« Projections
High Altitude
102.9
ft0  .otf
Legend
OS HDO
n hoc
~ LOT
S3 LDV
¦I Off-ROAD
ZZJ STATION ART SOURCES

-------
1
V)
C
o
o
o
o
s
Figure 4-6
NOX EMISSIONS INVENTORY
HIGH N02 URBAN AREAS
Statutory Cat* Pro|«etiena
Aitltud*
Lagerx!
C3 HOO
a HOG
~ LOT
S3 LDV
¦I OfT-OOAO
Z3 STATIONARY SOUttCCS
\9«>1 \^°

-------
4-21
:
1 HDG
e
o
o
it
C »•-
c
*
X
o
-a-
*

*^E1
1900
1983
1987
1988
1990
1995
? Mi
1993
Legend
A BASE CASE
X CONTROLLED CASE
~ STATUTORY CASE

-------
4-22
"I
Figure 4-8
CHANGES IN CONTRIBUTION TO NOx INVENTORY
o to-
o
o
o
C V
•
c
o
HDG
• i 8	W	38—	^
1980	1983	1987	1988	1990	1995
z 1980
1983
1987	1988
Calandar Year
1990
1995
Legend
A BASE CASE
X CONTROLLED CASE
~ STATUTORY CASE

-------
4-23
these vehicles as reflected in the negative growth rate value
in Table 4-6. HDDVs show the most significant increases in NOx
contribution, both in absolute terms and on a percent basis by
1995, relative to current (1983) numbers. The increase is
expected to be some 106,300 tons, or about a 66 percent
growth. This increase is attributed to a projected substantial
increase in the usage of diesel engines in HDVs. By way of
additional illustration, Figure 4-7 also shows these inventory
trends for the individual vehicle classes affected in this
rulemaking.
b.	Controlled Case
Figure 4-2 summarizes the total NOx inventory for the
controlled-case projections in the six low-altitude areas, and
Figure 4-7 shows the trends for individual vehicle classes.
Because the new NOx standards are effective beginning with the
1987 model year LDTs and HDEs, with a further reduction for
HDEs in 1990, only the 1990 and 1995 calendar years show
significant changes relative to the base-case projections. As
can be seen from comparing Figures 4-1 and 4-2, the sha'rp
increase found in the uncontrolled inventory is almost
completely eliminated under the controlled-case scenario. The
total inventory increases by only about three percent between
1983 and 1995 for the controlled-case scenario, as compared to
over 19 percent for the base-case scenario. The mobile source
contribution to the total inventory decreases by 29,900 tons,
or 7 percent, helping to offset the increases in the stationary
source and off-road contributions, which are the same as in the
base-case scenario. The 1995 LDT contribution to the NOx
inventory declines by 29 percent, or 13,700 tons in absolute
terms. About 20 percent of this reduction can be attributed to
LDDTs, the remainder to LDGTs. The HDGV contribution decreases
by 8,800 tons or 30 percent by 1995, although the percentage of
total emissions is relatively small. The effect of
introducing new NOx standards is most apparent for HDDVs, where
the HDDV inventory is reduced by 79,000 tons in the 1995
calendar year, or almost 30 percent relative to the 1995 base
case. Figure 4-2 also shows that by 1995 the HDDV class will
constitute 29 percent of the total NOx inventory compared to 35
percent for the base-case inventory. The total controlled
HDDV, HDGV, and LDT NOx contribution is about 37 percent of the
total, versus 45 percent in the base case.
c.	Statutory Case
Figure 4-3 represents the inventory reduction that could
be achieved in the six low-altitude areas if the statutory
standards were feasible from a practical standpoint for all
vehicle classes. As can be seen, the total inventory would be.

-------
4-24
reduced by 73,200 tons, or 11.5 percent. The mobile source
inventory would be reduced by 123,400 tons (a 41.5 percent
.reduction) compared to the base-case, more than offsetting the
abovementioned increase in the stationary source and off-road
total. By 1995 the HDDV portion of the inventory would be
reduced by 60 percent over the base case, a reduction of nearly
160,000 tons of pollutant on an absolute basis. HDGVs would
show a 63 percent reduction in NOx production, or 19,000 tons
on an absolute basis, while the LDT contribution would be
reduced by 38 percent or 17,000 tons. Figure 4-7 illustrates
these trends for each vehicle class. The total NOx inventory
would be reduced not only from the 1995 base case (26 percent),
but also from the 1980 base case (12 percent).
4.	High-Altitude Emission Inventories
Figures 4-4, 4-5, and 4-6 present the base-case,
controlled-case, and statutory-case NOx projections for
non-California high-altitude SMSAs, and Figure 4-8 illustrates
the trends for each vehicle class. Although the base- and
controlled-case light- and heavy-duty gasoline vehicle NOx
emission factors are lower than the corresponding low-altitude
emission factors, the same trends are observed for the
high-altitude emission inventories as were observed for the
l'ow-altitude results. The significant observations are a
decrease in the growth of the total NOx inventory relative to
the current (1983) level (from a 27 percent increase in the
base-case scenario to less than a 14 percent increase under the
controlled-case scenario) , and also a reduction in the absolute
and percentage contributions from LDVs, LDTs, and HDVs in the
controlled-case emissions inventory relative to the base-case
emissions inventory. The majority of the abovementioned
increases in the total inventory under the controlled scenario
comes from stationary source and off-road contributions, since
the mobile source contribution would be only 1,300 tons, or
about three percent over 1983 levels. A comparison of
base-case and controlled-case combined LDT and HDV emissions
shows that the 1995 controlled-case inventory contribution is
about 30 percent lower than it would be under the base-case
scenario for the comparable year. As can be seen from Figures
4-6 and 4-8, the statutory standards, if feasible, would result
in a considerably greater reduction (52 percent) from
uncontrolled levels, which would very nearly offset the
increase in the stationary-source/off-road contribution.
5.	Reductions in Total Emissions Resulting From the
Proposed Standards
The environmental impact of imposing lower NOx standards
on LDTs and HDVs has thus far been presented by examining the.

-------
4-25
relative differences and trends between emission inventory
projections as. a function of time. It is also of value to
examine in more detail the emission inventory impact of the new
NOx standards at a point in time after the implementation of
the new standards.
Tables 4-7 and 4-8 summarize the differences between the
1995 base-case and the 1995 controlled-case projections for
low- and high-altitude areas, respectively. These results were
calculated from the emission inventory data used to prepare
Figures 4-1 through 4-6. Tables 4-7 and 4-8 express the NOx
reduction in tons, as a percent reduction of the respective
vehicle class inventory, and as a percent reduction of the
total NOx inventory.
These tables clearly demonstrate the relative significance
of the LDT, HDGV and HDDV contributions to the total NOx
inventory. For example, at low altitude HDDV emissions
comprise 79,000 tons out of a total NOx reduction of 103,600
tons, or a reduction of 10.4 percent of the total inventory,
and at high altitude, 8,200 of 12,600 tons total reduction, "or
7.1 percent of the total NOx inventory. The total inventory
decreases by 13,700 tons or 1.8 percent at low altitude due to
the LDT NOx reduction, and by 3,400 tons or 3 percent at high
altitude. The HDGV reduction is 8,800 tons at low altitude and
700 tons at high altitude, which are 1.2 and 0.6 percent
respectively of the total motor vehicle inventories.
6. Sensitivity to Alternative Design Values
Emission inventory and air auality projections are based
on many assumptions and judgments concerning the input
parameters used for analysis. The average annual NO2
concentrations, or design values, mentioned earlier constitute
one of the basic inputs to the analysis. Recent EPA guidelines
suggest that alternative design values be included in such
analyses to insure the results would not be unduly biased by
unrepresentative high or low concentrations for a given year.
Therefore, inventories and air quality calculations for this
analysis were performed using both 1980 and 1981 design value
data.
The results presented in Figures 4-1 through 4-8, which
are calculated using 1980 as the base year, were reanalyzed
usinq 1981 base year projections. The net results were very
slight decreases in the emission inventories for particular
emission classes and in the total inventory. For example,
using 1981 design values results in projections of total NO2
emissions that are 1.5 and 1.3 percent lower than those using
1980 design values, for the 1995 base- and controlled-cases,.

-------
4-26
Table 4-7
1995 Incremental Reduction in NC>2
Emission Inventories for High NO2 SMSAs--Low Altitude
Reduction from 3ase Case
Vehicle Class
Heavy-Duty Diesel
Heavy-Duty Gas
All Light-Duty
Trucks
Total LDT, HDE
Heavy-Duty Diesel
Heavy-Duty Gas
All Light-duty
Trucks
Reduction	Percent of
(xlOOO Respective Vehicle
tons )	Class Emissions
Controlled Case
79.0
8.8
13.7
29.5
30.1
29.3
103.6	30.1
Statutory Case
158.0	59.0
17.0	58.2
20.0	42.7
Percent of
Total Emissions
10.4
1.2
1.8
13.6
20. 8
2.2
2.6
Total LDT, HDE
195.0
56.7
25.7
[1] 1980 base year inventories.

-------
4-27
Table 4-8
1995 Incremental Reduction in NO2
Emission Inventories for High NO2 SMSAs--Hiqh Altitude
Reduction from Base Case
Vehicle Class
Heavy-Duty Diesel
Heavy-Duty Gas
All Light-Duty
Trucks
Total LOT, HDE
Heavy-Duty Diesel
Heavy-Duty Gas
All Light-Duty
Trucks
Reduction	Percent of
(xlOOO Respective Vehicle
tons)	Class Emissions
Controlled Case
8.2
0.7
3.4
29.3
31.8
28.6
12.<5	29.9
Statutory Case
11.5	59.8
1.1	50.0
4.4	37.0
Percent of
Total Emissions
7.1
0.6
3.0
10.9
10.0
1.0
3.8
Total LOT, HDE
17.0
40.4
14.8
[1) 1980 base year
inventories.

-------
4-28
respectively. Similar changes were observed when the
high-altitude projections were calculated using 1981 as the
base year: 1995 base- and controlled-case inventory
projections decreased 2.3 and 2.1 percent, respectively. In
all cases, the differences between the 1981 and 1980 design
value analyses inventories were within 3 percent, which is well
within the uncertainty of the rollback model's projections.
Because the NOx emission reductions were calculated using
a 1980 base year, the incremental tonnage reductions in Tables
4-7 and 4-8 will also change slightly when the analysis is made
using 1981 as the base year. However, the magnitude of these
changes are again within the uncertainty of the rollback
model's assumptions. EPA, therefore, feels confident that the
benefits attributed to the new standards are reasonable
estimates and are not sensitive to small changes in the design
values.
D. Air Quality Effects
The air quality effects of imposing new NOx standards may
be determined by examining the results obtained from the EPA
Rollback Model. Tables 4-9 through 4-12 present projections of
the number of National Ambient Air Quality Standard (NAAQS)
exceedances for NO2 and the average percent change in ambient
NO2 concentration relative to a base year. The tables are
organized according to location (either low altitude or high
altitude), control scenarios, and design values. Base-case,
controlled-case, and statutory-case projections are presented
for both 1980 and 1981 design values.
Tables 4-9 and 4-10 quantify the effects of enacting new
LDT and HDE NOx standards in terms of the number of high NO2
urban areas which are predicted to exceed the NAAQS for NO2.
In the base case, the effect of the steady increase in total
NOx emissions over time is shown to adversely affect NAAQS
attainment, especially in the 1990s. The number of NO2
exceedances in this time frame is significantly decreased by
enacting new LDT and HDE NOx standards. Using the 1980 design
value results, the number of low-altitude exceedances is
reduced from three to one in 1995. Within the accuracy of the
model, this reduction appears to hold true whether the proposed
or the statutory standards are implemented. At high altitude,
on the other hand, where the 1995 exceedances are reduced from
two to one by the proposed standards, the statutory standard
results in no exceedances in 1995.*
It should be pointed out that although only three of the
six low-altitude SMSA's are projected to exceed the NAAQS,
in 1995 without new truck standards, two of the three
remaining areas studied which are currently below the -
standard are projected to be at the standard by 1995.

-------
4-29
Table 4-9
Number of SMSAs Exceeding the NO2
Ambient Air Quality Standard
Low Altitude
1983
1987
1988
1990
1995
Base Case:
1980	Design Value
1981	Design Value
1
0
1
0
1
0
1
1
3
2
Controlled Case:
1980	Design Value
1981	Design Value
1
0
1
0
1
0
1
0
1
0
Statutory Case:
1980	Design Value
1981	Design Value
1
0
1
0
1
0
1
0
1
0

-------
4-30
Table 4-10
Number of SMSAs Exceeding the NO2
Ambient Air Quality Standard
High Altitude
1983
1987
1988
1990
1995
Base Case:
1980	Design Value
1981	Desian Value
0
0
1
0
1
0
1
0
2
2
Controlled Case:
1980	Design Value
1981	Design Value
0
0
1
0
1
0
1
0
1
0
Statutory Case:
1980	Design Value
1981	Design Value
0
0
1
0
0
0
0
0
0
0

-------
4-31
Table 4-11
Average Percent Change in the Ambient
NO2 Concentration from the Base Year[l]
Low Altitude
Base Case:
1983
1987
1988
1990
1995
1980	Design Value
1981	Design Value
Controlled Case:
L980 Design Value
1981 Design Value
Statutory Case:
1980	Design Value
1981	Design Value
2
1
2
1
2
1
4
2
2
1
-3
-3
7
5
2
2
-7
-7
19
17
3
1
-12
-12
TT] Negative value means a decrease in ambient concentration.
Positive value means an increase in ambient concentration.

-------
4-32
Table 4-12
Average Percent Change in the Ambient
NO2 Concentration from the Base Year[l]
High Altitude
1983 1987 1988
Base Case:
1980	Design Value
1981	Design Value
Controlled Case:
1980	Design Value
1981	Design Value
Statutory Case:
1980	Design Value
1981	Design Value
5
3
5
3
5
3
14
12
14
12
12
12
16
14
14
12
8
8
1990
21
17
16
14
7
6
1995
37
35
20
17
5
3
[1] Negative value means a decrease in ambient concentration*.
Positive value means an increase in ambient concentration.

-------
4-33
The EPA rollback model also calculates the average ambient
air quality change expected for a particular set of NOx
standards. Tables 4-11 and 4-12 present the average percent
change in NOx air quality. Again, the low and high altitude
results are presented for the base-case, controlled-case, and
statutory-case scenarios and as a function of the desian value
year.
An analysis of Tables 4-11 and 4-12 produces the same
general observations as described for Tables 4-9 and 4-10.
Significant improvements in average ambient air quality are
apparent when comparing the base- and controlled-case
concentration changes. For example, the low-altitude average
increase in NO2 concentrations from 1983-1995 is 20 percent
(from -1 percent to +19 percent of the 1980 value) , calculated
using the base-case/1980 design values. The corresponding
1983-1995 controlled-case projection shows a 4 percent increase
in average NO2 concentration or a difference of 16 percent
due to the implementation of the proposed standards. The
statutory case would yield an 11 percent decrease in ambient
NO2 concentrations if it could be implemented. In high
altitude areas, the average increase in NO2 concentrations
from 1983-95 drops from 32 percent under the base-case scenario
to 15 percent under the controlled case scenario. This
represents an improvement of 17 percent due to the proposed
standards. The statutory standards would result in no increase
in NO2 concentrations in 1995 relative to 1983.
Breaking down t^e differences in NO2 concentrations
resulting from the proposed standards by vehicle class, 13
percent of the abovementioned 16 percent differences in NO2
concentrations at low altitude is due to the reduction in HDDV
NOx emissions. In the high altitude areas, 13 percent of the
17 percent total difference in NO2 concentrations is
attributable to the HDDV NOx reduction. LDGTs, LDDTs, and
HDGVs each account for about one percent of the difference.
As in the case of emission inventory calculations, air
quality projections are also affected by changes in the design
values. Using 1981 design values results in slightly different
air quality projections from those listed above, which can be
used to provide a sensitivity analysis for the calculations.
In general, usinq 1981 design values results in fewer projected
NAAOS exceedances and slightly lower projected NO2
concentrations than result from using 1980 desiqn values.
However, the relative differences between the base-case and
controlled-case projections remain almost the same. For
example, implementation of the proposed standards results in a
reduction in the number of low-altitude areas exceeding the
NO2 NAAOS from three to one in 1995 using the 1980 desigr;

-------
4-34
values. Using the 1981 design values results in a reduction
due to the proposed standards from two exceedances to zero in
-1995. Thus, the effect of implementing the proposed standards
is the same in either instance-1995 NO2 violations are
reduced by two. In high altitude areas, the number of
exceedances is reduced by two under the 1980 design values and
by one if the 1981 values are used.
Similar air Quality improvements are observed between 1983
and 1995 when examining the low-altitude base- and
controlled-case ambient NO2 concentration projections usinq
1981 design values. The average percent increases in NO2
concentrations during the period 1983-95 are projected to be 18
percent and 2 percent for the base-case and controlled-case
scenarios, respectively. Thus the average decrease in low
altitude NO2 concentrations under the proposed standards
would be the same when calculated using 1981 design values as
it would using 1980 values, (i.e., 16 percent). In high
altitude areas, the average increase in NO2 concentrations
between 1983 and 1995 under the proposed standards would be -15
percent using 1981 design values, as opposed to 17 percent
using the 1980 values. The base case increase would be 32
percent for either design value. Under the statutory
standards, the high altitude NO2 concentrations would peak in
1987 and decline to 1983 levels by 1995, using either design
values. Low-altitude statutory-case concentrations would
decrease 11 percent under either set of design values.

-------
4-35
References
1.	"Air Quality Criteria for Oxides of Nitrogen," U.S.
EPA, Research Triangle Park, North Carolina, EPA-600/8-82-006,
September 1982.
2.	"Review of the National Ambient Air Ouality
Standards of Nitrogen Oxides: Assessment of Scientific and
Technical Information - OAOPS Staff Paper," U.S. EPA, Research
Triangle Park, North Carolina, EPA-450/5-82-002.
3.	EPA Memo from Sheldon K. Friedlander, Chairman Clean
Aj.r Scientific Advisory Committee to Anne M. Gorsuch,
Administrator of EPA, U.S. EPA, Research Triangle Park, North
Carolina, EPA-450/5-82-002, July 6, 1982.
4.	See for example Environmental Forum, August 1983, p.
30; Science Vol. 222, p. 8
5.
Chemical & Engineering News, June 20, 1983, p. 27. -
6.
Chemical Engineering Journal, July 11,
1983, p. 31.
7.
Environmental Forum, April 1983, p. 16.

8.
Detroit Free Press, April 2, 1983, pp.
1C and 2C.
9.
Washington Post, May 24, 1983, p. A9.

10.
New York Times, June 28, 1983, p. 8.

11.
"Briefing Document" prepared for EPA
Administrator
William Ruckelshaus by EPA's Acid Deposition Task Force, August
1, 1983, excemted in the Environmental Reporter, September 2,
1983, pp. 754 and 756.
12.	Public Review Draft, "The Acidic Deposition
Phenomenon and Its Effects", U.S. EPA, EPA fiOO/8-83-0168, 1983,
pp. 4-31-4-32.
13.	"Briefing Document", p. 775. See also Environmental
Science £ Technology, Vol. 17, No.10, 1983, p. 477A
14.	"Briefing Document", p. 783.
15.	"User's Guide to MOBILE 2," U.S. EPA,
EPA-460/3-81-006, 1981.
16.	"Compilation of Air Pollution Emission Factors:
Highway Mobile Sources," U.S. EPA, EPA-460/3-81-005, March 1981-

-------
4-36
References (cont'd)
17.	EPA Memo, Derivation of HDDE Average Usage Period,
from R. Johnson, U.S. EPA, Ann Arbor, Michigan to Public Docket
No. A-81-11, July 31, 1983.
18.	"Rollback Modeling: Basic and Modified," Journal of
the Air Pollution Control Association, DeNevers, N. and J.
Morris, Vol. 25, No. 9, 1975.

-------
CHAPTER 5
PARTICULATE ENVIRONMENTAL IMPACT
I.	Introduction
This chapter will briefly examine the air quality impact
of the proposed particulate standards for heavy-duty diesel
trucks. In addition, four classes of health and welfare
effects will be considered: 1) non-cancer health effects, 2)
carcinogenic health effects, 3) visibility effects, and 4)
soiling effects. For a more in-depth presentation of the
environmental effects associated with diesel particulate
control, the reader is referred to EPA's Diesel Particulate
Study (DPS).[1]
It should be noted that the two control scenarios
considered in this chapter differ somewhat from those presented
in the DPS report. The effects described in that study were
adjusted for the difference in standards for inclusion in this
chapter, using the ratio of the respective emission inventories
as described in the sensitivity analysis chapter of the DPS
report. Since the effects models are essentially linear, -no
significant inaccuracy should result from this adjustment.
II.	Particulate Control Scenarios
Environmental effects will be presented for two basic
control scenarios. In the base-case scenario heavy-duty diesel
engine (HDDE) particulate emissions would remain at the
uncontrolled level, approximately 0.7 g/BHP-hr. The
controlled-case scenario assumes the implementation of a
particulate standard of 0.60 g/BHP-hr effective for the 1987-89
model years, followed by a more stringent standard of 0.25
g/BHP-hr effective for the 1990 and later model years.
Proportional standards of 0.72 and 0.30 g/BHP-hr are also being
proposed, effective in the 1987 and 1990 model years,
respectively, for HDDE particulate emissions under
high-altitude conditions.
III.	Air Quality Impact of Diesel Particulate Control
This section will assess the effects of the proposed
standards on localized, urban, and nationwide air quality in
order to provide a basis for assessing health and welfare
benefits. The section will include: 1) a discussion of the
relationship of diesel particulate to total suspended
particulate emissions and a discussion of the applicable air
quality standards, 2) a discussion of particulate emission
factors, or average vehicle emission rates, for the above HDDE

-------
5-2
control scenarios and the resultant lifetime emissions per
vehicle, 3) a discussion of HDDE particulate emissions relative
to those from other sources, 4) the resulting effect of the
proposed standards on localized, urban, and national
atmospheric concentrations, and 5) a discussion of particulate
emissions in high-altitude areas.
A. Relationship of Diesel Particulate to Total
Suspended Particulate and the National Ambient Air
Quality Standards
Total suspended particulate (TSP) matter includes all
particles entrained in the air. As such, it includes larqe
particles like dust and dirt, as well as small particles like
soot, and chemical elements such as lead, antimony, arsenic,
etc. TSP can be subdivided according to size into three broad
categories: coarse, inhalable, and fine. Fine particulate has
diameters of less than 2.5 micrometers (um) and, because of its
small size, almost all diesel particles fall into this
category. At the same time, inhalable particulate, generally
referred to as PM^q, has diameters of less than 10 um.
Again, because of its small size, diesel particles are all
classifiable as inhalable particulate.
There have been National Ambient Air Quality Standards
(NAAQS) for TSP since 1971. A primary NAAOS for TSP of 75
micrograms per cubic meter (ug/m^) on an annual geometric
mean basis, or 260 ug/m^ within any 24-hour period (which may
not be exceeded more than once each year) , has been set to
protect public health. A more restrictive secondary standard
of 150 ug/m^ (24 hour short term) was set to protect public
welfare (i.e., other items of value such as materials, crops,
etc.) against any known or anticipated adverse effects.[2]
EPA will soon propose a primary NAAQS for PM^g in the
range of 50-65 ug/m^ (annual arithmetic mean basis) in place
of the current primary TSP standard. Inhalable particulate is
estimated to comprise 46-55 percent or more of TSP on a
nationwide average.[3] Some 71-246 counties are currently
projected to violate the PM^g NAAOS, depending on the level
of the standard. By 1995, numerous violations of the suggested
NAAQS are still projected to occur, despite best efforts at
controlling non-mobile sources of inhalable particulate.
Because diesel particulate is all inhalable, the control of
HDDVs is germane to attaining and maintaining any PM^g
NAAQS. Thus, comparisons of the effect of diesel particulate
controls will be made in the following sections with respect to
the PM^g standards being considered.

-------
5-3
EPA has not as yet determined a need for a secondary
PMio standard, though a secondary fine particulate standard
has been considered for visibility protection. If a secondary
NAAQS is established, a comparison of the effect of diesel
particulate controls would also be directly relevant to such a
standard because almost all diesel particulate is fine.
B. HDDV Emission Factors and Resultant Lifetime
Emissions Per Vehicle
Emission factors represent the average emission rates,
expressed in grams per mile, for each pollutant under various
s.tandards. They are calculated for various classes and
subclasses of vehicles based on data accumulated through
emissions testing. Emission factors for future model years are
estimates developed from historical emissions data using
standard modeling techniques.
The derivation of the HDDV particulate emission factors
presented in this chapter is described in the DPS report and
will not be repeated here.[l] It should be noted, however,
that the rates shown in the table below differ from those
presented in the DPS report in three ways. First, the
standards differ from the scenarios in the DPS report,
resulting in slightly different emission rates. Second, in the
DPS report, emission factors changed slightly for each model
year, reflecting projected improvements in engine/vehicle
operating efficiencies. In this chapter, a single emission
rate was selected to be reasonably representative of the model
years included in each scenario. This action greatly
simplified the analysis with no significant effect on the
results. Third, the emission factors are given as a composite
factor for all HDDVs, while those in the DPS report were
presented for three HDDV subcategories. This change was made
to simplify the analysis in this chapter and to focus the
analysis on the entire HDDV class, which is the object of
control. The composite factor was derived by weighting the
individual emission based on the lifetime mileage for each
respective vehicle category, (i.e., 110,000 for LHDDVs, 268,000
for MHDDVs, and 529,000 for HHDDVs) and their respective sales
fractions (33 percent for LHDDVs, 25 percent for MHDDVs, and 42
percent for HHDVs). The following table presents average
emission rates for the base-case and the controlled-case
scenarios for the entire class of HDDVs:

-------
5-4
Average Emission Rates (q/mi)
Scenar ios
HDDV
Uncontrolled
2.1765
Controlled
0.60 g/BHP-hr (1987
1.8656
0.25 g/BHP-hr (1990
0.9325
It should also be noted that these emission rates do not
explicitly include any adjustment for the higher emissions of
HDDVs at high altitude. As explained later, very little data
are available with which to evaluate this emissions increase.
Also, because the above rates are used to characterize
emissions on a national basis and HDDV VMT at higher altitude
is very small compared to that at low altitude, the effect of
not specifically including an adjustment for vehicles at higher
elevations is negligible. Therefore, not explicitly
considering the higher emission rates for high-altitude HDDVs
has an insignificant effect on the results of this analysis.
The average emission rates can now be used to determine
total lifetime emissions per vehicle for comparison of
base-case and controlled-case levels. The average emission
rate is multiplied by the average estimated lifetime mileage
for HDDVs. The following table presents the base-case and
controlled-case lifetime HDDV particulate emission levels
resulting from the proposed 1987 and 1990 standards. The
estimated lifetime mileage used in the calculation is also
shown. This mileage was derived from a sales-weiqhting of the
lifetime mileages mentioned above.
Standard
Lifetime Emissions
(tons) Per Vehicle
HDDV
Uncontrolled
0.78
Controlled
0.60 g/BHP-hr (1987)
0.67
0.25 g/BHP-hr (1990)	0.34
(Estimated Lifetime Mileage) (325,480)

-------
5-5
C. HDDV Particulate Emissions Inventory and Comparison
to Other Sources
Now that the emission factors and lifetime emissions
per-vehicle have been described, total HDDV particulate
emissions can be determined for various future years. This
shows how the HDDV particulate emissions inventory changes with
time, including how the various standards affect emissions, and
allows for comparisons with other sources.
The methodology used in determining total particulate
emissions is contained in the DPS report. This methodology
will not be repeated here. Note, however, that the scenarios
in this chapter are slightly different from those in that
study, so the inventory contributions were recalculated using
the emission factors developed for this analysis. The figures
presented here reflect EPA's best estimate of diesel sales
through 1995. This sales estimate is also described in the DPS
report.
Table 5-1 presents the projected HDDV, LDDV, and LDDT
contributions to the 1995 mobile source urban emissions
inventory. Urban areas are of primary interest since diesels
have the greatest effect on these locations, and since the
great majority of people in PM^g non-attainment areas reside
in urban locations. The table includes the HDDV inventory
contribution for both the proposed standards and a continuation
of the 1987 standard of 0.6 g/BHP-hr to 1990 and beyond for
comparison purposes. The light-duty motor vehicle
contributions are also included for comparison purposes.[1]
The LDDV and LDDT contributions are the same under the
base-case and controlled-case scenarios and assume standards
already in place (i.e., 0.2 and 0.26 g/mi, respectively,
beginning in the 1987 model year) . The 1995 LDDV and LDDT
emissions increase substantially over 1983 levels because there
were relatively few diesel vehicles in-use in 1983 in
comparison to projected future numbers.
Table 5-1 shows that uncontrolled HDDVs were by far the
major source of urban diesel particulate emissions in the past
and that, if left uncontrolled, this relationship will
continue. It is also apparent that without the proposed
standards, the 1995 HDDV inventory will grow to nearly 90% more
than the 1983 level.
This information is important with regard to the inhalable
particulate NAAQS, which is currently being considered by EPA.
It shows that HDDVs are by far the single most important mobile
source contributor to urban violations of the suggested PM^g
NAAQS. Moreover, it shows that any substantial reduction in'

-------
5-6
Table 5-1
1983 and 1995 Urban Diesel Particulate
Emissions (tons)	
1983
Levels
1995 HDDV Scenarios
Controlled
Base
LDDV	6,400 (11%)* 21,700	(16%)
LDDT	1,500 (2%)	12,900	(10%)
HDDV	51,200 (87%)	97,000	(74%)
TOtal	59,100	131,600
(0.6)
Controlled
0.6/0.25
21,700 (18%) 21,700 (26%)
12,900 (11%) 12,900 (15%)
84,600 (71%) 48,600 (59%)
119,200
83,200
Figures in parentheses indicate percent of total,

-------
5-7
urban diesel particulate emissions would have to come from
controlling HDDVs, since the LDDV and LDDT values in this table
already reflect a significant degree of control. The proposed
standards do provide a significant level of control, since with
the standards the 1995 HDDV inventory would remain about 5
percent below the 1983 level (Table 5-1) .
Looking specifically at the effects of the proposed 0.60
and 0.25 g/BHP-hr HDDV standards, Table 5-1 shows that the 0.6
standard in 1987 would not, by itself, prevent the substantial
increase in HDDV particulate emissions from 1983 to 1995. With
the addition of the 0.25 standard beginning in 1990, however,
HDDV particulate emissions are essentially held constant from
1983 to 1995. Viewed another way, imposition of the 0.25
standard significantly offsets the projected increase in LDDV
and LDDT particulate emissions. Therefore, the proposed 0.6
and 0.25 HDDV standards, in combination, are very effective in
reducing the rate at which urban diesel particulate emissions
are expected to grow.
D. Air Quality Projections for Urban Areas
The air quality and exposure estimates developed for the
two control scenarios described above are presented in Table
5-2. The 1980 levels are included for comparison purposes.
The 1995 estimates were adjusted as described above from the
scenarios in the DPS report. The three major indicators of
impact are: 1) ambient urban concentrations, which include a
broad spectrum of city sizes and meterological conditions, 2)
annual average urban exposures, which include a variety of
individual activity pattern effects, based on exposures in four
U.S. cities, and 3) microscale concentrations, which represent
short-term (not over an hour) localized exposures.
The ambient urban concentrations of diesel particulate
under the base-case scenario are projected to increase by about
160 percent from 1980 to 1995. The ambient concentrations
resulting from this sharp increase can be placed in perspective
by comparing them to levels of the PM^g NAAQS, which is
currently being considered by EPA. In 1995 without the
proposed HDDV standards, ambient urban particulate
concentrations will rise to 1.6 to 7.8 ug/m^, representing up
to about 15 percent of the suggested NAAQS in large cities.
This clearly can represent a significant portion of the NAAQS,
especially in areas that are projected to be in continued
violation of the PM^q standard. Under the controlled-case
scenario, urban ambient diesel particulate concentrations will
also rise from 1980 to 1995, but the increases will be
significantly attenuated from those associated with the

-------
5-8
Table 5-2
Effect of the Proposed Standards on Urban Air Quality
Total Diesel Particulate
Concentration (ug/m3)--
	LDDV, LDDTy HDDV
1995
1980
Base
Ambient Urban Concentrations*
City Population;
Greater than 1,000,000
500,000 - 1,000,000
250,000 - 500,000
100,000 - 250,000
1.3-3.0
1.0-2.1
1.0-1.7
0.6-1.8
3.4-7.8
2.6-5.5
2.6-4.4
1.6-4.7
Controlled
2.2-4.9
1.7-3.4
1.7-2.8
1.0-3.0
Annual Average Exposure to U.S. Urban Dwellers
TOTAL
2.4
5.3
3.2
Microscale Concentrations
Roadway Tunnel:
Typical	57	106	65
Severe	145	270	166
Street Canyon:
Typical	2	4	2
Severe	14	26	17
On Expressway:
Typical	6	11	7
Severe	26	48	29
Beside Expressway	5	10	6
Ranges are average values plus and minus one standard
deviation.

-------
5-9
base-case scenario. The proposed HDDV standards will reduce
the 1995 levels by about 40 percent to 1.0-4.9 ug/m3.
Therefore, the proposed standards can be important in reducing
urban ambient particulate concentrations, which in turn, will
help attain and maintain the suggested PMio NAAQS.
The estimates of annual average exposure for dwellers in
urban areas increase from current levels of 2.4 ug/m^ to 5.3
ug/m-* by 1995 under the base-case scenario, or an increase of
about 120 percent. The 1995 average drops to 3.2 ug/m^ under
the controlled-case scenario, representing an increase of about
55 percent over the 1980 average. Thus, the proposed standards
would result in about a 40 percent improvement in average
exposure to diesel particulates.
The third indicator of air quality, the microscale
impacts, indicates the same trend. While these levels are well
above the more regional, long-term figures, they are probably
of less concern, due to the short-term periods involved and the
lack of acute health effects associated with diesel
particulate. However, these levels would be of concern for
anyone spending long periods of time in such areas.
The three above indicators of air quality show decreases
in air quality in future years regardless of the implementation
of diesel particulate controls. However, the proposed
standards will aid in mitigating this degradation and directly
help in attaining and maintaining the PM^g NAAQS which is
currently under consideration.
E. High-Altitude Particulate Emissions
There are no available high-altitude particulate emission
data for HDDVs, and the data for LDDVs tested in high-altitude
areas are extremely limited. Such data as are available
suggest that particulate emissions at high altitude may be as
much as 50 percent higher than those from comparable vehicles
operated at low altitude.[4] However, these data were from
naturally aspirated light-duty engines. In contrast, most
heavy-duty engines used in high-altitude areas are
turbocharged. Since the boost pressure sensors, usually
present on turbocharged engines, adjust for varying barometric
pressures (i.e., altitude), EPA estimates that the current
high-altitude increase in HDDV particulate levels should be
lower than for light-duty vehicles, on the order of 20
percent. EPA has, therefore, set the high-altitude standards
at 120 percent of the low-altitude figures.
The need for controls at high altitude should be
apparent. Higher emission levels in high-altitude areas pose

-------
5-10
an even greater threat to the environment from a given number
of vehicles than is found in low-altitude locations. As a case
in point, Denver's "brown cloud" is an example of the type of
visibility problem that could worsen considerably in the
absence of control of diesel particulates. Denver is currently
a non-attainment area for TSP, with a 1982 annual concentration
as high as 173 ug/m^ (annual geometric mean) and a short-term
concentration of 510 ug/m^ on a 24-hour basis (second highest
reading). As stated earlier, some 46-55 percent of this TSP
would likely be PMio, placing this area well above the PMio
NAAQS currently under consideration. Recent studies have
concluded that mobile sources are one of the three major
contributors to the haze problem, and that they could become
the major contributor if the current high rate of diesel sales
continues.[5]
Given the present particulate emission levels, the Denver
haze level could double by the year 2010. The more stringent
1987 LDDV and LDDT particulate standards will cut the increase
in half, but this still results in a 50 percent deterioration
in the visibility index. No quantitative study has been done
at this time to gauge the effect of the proposed HDDV
particulate standards in this particular problem. However,
from the preceding discussion of mobile source inventory
contributions, it appears that the HDDV contribution will be
significantly greater than the LDD contribution in the coming
years. The fact that the proposed standards are projected to
reduce the HDDV lifetime particulate levels by about 50 percent
by 1995 will therefore likely have a significant, beneficial
effect on the problem. Denver is not the only high-altitude
city experiencing air quality problems of this nature and the
same conclusion can likely be drawn for other areas with
similar problems.
IV. Health and Welfare Benefits of Controlling Diesel
Particulate
As mentioned previously, the DPS report examined four
aspects of the environmental impacts of diesel particulate: 1)
non-cancer health effects, 2} carcinogenic health effects, 3)
visibility, and 4) soiling.
A. Non-Cancer Health Effects
Particulate matter in general has long been regarded as
hazardous to human health. EPA recognized this danger and
established an NAAQS for total suspended particulate (TSP) as
early as 1971. Currently, EPA is planning on revising the
primary standard to focus on inhalable particles (i.e., those
with diameters of 10 um or less (PM^q))/ because this.

-------
5-11
fraction of total suspended particulate matter appears to be
responsible for most of the human health effects associated
with TSP.
As mentioned earlier, diesel particulates are all
inhalable particulate, and nearly all of it can be classified
as fine particulate (less than 2.5 um in diameter). Although a
large body of data has been developed regarding the health
effects of inhalable particulate matter, research limited
specifically to diesel particulate is relatively new and
somewhat inconclusive. An analysis of the available data
indicates that, until more is known, diesel particulate
generally should be regarded as being equivalent to other forms
of inhalable particulate matter in terms of the hazard it
presents to human health, although there is a possibility it
may be somewhat more hazardous.[1] It should be pointed out,
however, that even if regarded as posing the same hazard,
diesel particulate is emitted directly into the breathing zone,
rather than from tall stacks that would promote dispersion.
Thus, the potential for human exposure is maximized.
Two basic concerns exist with respect to the health risk
posed by inhalable particulate in general.[1] First, inhalable
particulates are small enough so that they are not as readily
prevented by the natural body defenses from reaching the lower
respiratory tract, as would coarser particles. Fine
particulate matter can penetrate to the alveoli, or deepest
recesses of the lungs, where the oxygen-carbon dioxide exchange
takes place with the circulatory system.f6] The body requires
months or years to clear foreign matter from the alveolar
region, as opposed to hours or days to clear the upper
respiratory system. The second concern is that inhalable
particulate may be composed of toxic materials or may have
hazardous materials adsorbed onto its surface.
The most obvious non-cancer health effect of an inhalable
particulate, such as diesel particulate, is injury to the
surfaces of the respiratory system, which could result in
reduced lung function, bronchitis or chronic respiratory
symptoms. The hazardous chemicals that may be associated with
particulate matter (e.g., organic compounds, lead, antimony,
etc.) can either react with lung tissue or be transported to
other parts of the body by the circulatory system. Particulate
matter may also weaken the resistance of the body to infection
and there are indications that it reacts adversely in
conjunction with other atmospheric pollutants. For example,
studies in London, New York, Buffalo, and Nashville have found
an increase in the mortality rate, especially among older
persons, when high particulate levels were accompanied by high
sulfur dioxide levels.[7]

-------
5-1?
From the above discussion, it is clear that inhalable
particulate matter (PM^g) has been linked directly with a
myriad of adverse non-cancer health effects, and it is based on
this information that EPA is proposing an NAAQS for PM^q.
Also, diesel particles are all inhalable particulate and,
therefore, can potentially cause the same health effects of
concern. This relationship can be used to assess the overall
benefits of controlling HDDV diesel particulate.
As stated previously, 71-246 counties are projected to be
currently in violation of the range of primary PM]_q standards
being considered. Even after reasonable non-mobile source
emission controls are implemented, numerous violations of the
NAAQS are still projected to occur in 1995. Without the
proposed HDDV standards, diesel particulate from these vehicles
will account for about 1.6-7.8 ug/m^ of the ambient
particulate loading in urban areas, or up to about 15 percent
of the suggested PMio NAAQS. This clearly shows that HDDVs
can be a significant source of inhalable particulate in urban
areas. Imposition of the proposed HDDV standards would reduce
this source's contribution to 1.0-4.9 ug/m^, a reduction "of
up to about 40 percent. Therefore, the proposed HDDV standards
can play an important part in in reducing urban PMj-q
concentrations. Furthermore, the resulting reduction in diesel
particulate emissions within urban areas that continue to
violate the suggested PM^q NAAOS, will directly reduce the
non-cancer health effects associated with inhalable
particulates in general.
B. Carcinogenic Health Effects
A number of studies have concluded that exposure to diesel
particulate probably poses an additional risk of acquiring lung
cancer. EPA surveyed these studies and developed
scenario-specific risk factors for lung cancer incidence. [1]
Table 5-3 shows the cancer risk estimates for both the
base-case and the controlled-case scenarios presuming the
compounds producinq the cancer-risk are reduced in the same
proportion as total diesel particulate. These estimates were
again adjusted for differences in scenarios from those in the
DPS, as described above. The estimated cancer risks from other
selected causes are also shown for purposes of comparison.
The data indicate that while the risk of contracting lung
cancer is greatest from smoking, exposure to diesel particulate
may represent a significant portion of all non-smoking-related
lung cancer. The upper limit of the uncontrolled scenario
would represent six individuals in a million, or eight percent
of all non-smokinq-related lunq cancer in the U.S. The lower
limit still represents one in a million individuals, which has.

-------
5-13
Table 5-3
Comparison of Risks from Various Sources
Estimated Annual Risk	Exposed
Sources of Risk	 	(risk/per son-year)	Population
Commonplace Risks of Death
Motor Vehicle Accident
222.0
X
10-6

Entire U.S.
Drowning
26.0
X
10"6

Entire U.S.
Burns
21.0
X
10"6

Entire U.S.
Tornados, Floods, Light-
2.0
X
10-6

Entire U.S.
ning, Hurricanes, etc.





Risks of Cancer Incidence





Diesel Particulate (1995):




Urban U.S.
Base Scenario 1.1
x 10"6
-
6.1 x
lO"6

Controlled Scenario 0.8
x 10"6
-
5.3 x
10-6

Natural Background Radi-
20.0
X
10"6

Entire U.S.
ation (sea level)





Average Diagnostic Medical
20.0
X
10"6

Widespread
X-Rays in the U.S.





Frequent Airline Passenger
10.0
X
10-6

Limited
(4 hours per week





flying)





Four Tablespoons Peanut
8.0
X
10"6

Fairly
Butter Per Day (due to




widespread
presence of aflatoxin)





Ethylene Dibromide
4.2
X
10-6

Widespread
One 12-Ounce Diet
2.6
X
10-6

Widespread
Drink Per Day





Arsenic
1.7
X
lO-6

1% of U.S.
Miami or New Orleans
1.0
X
10-6

Southern
Drinking Water (due




U.S., Urban
to presence of chloroform)





Lung Cancers:




Entire U.S.
For Smokers Due to
419.0
X
10*6


Smoking





For General Population
73.9
X
lO"6


Due to Causes Other
Than Smoking

-------
5-14
been used in the past by regulatory agencies as a cut-off point
for determining the need to regulate. Thus, Table 5-3 shows
that a relatively small but significant cancer risk may be
attributable to diesel particulate exposures. The proposed
HDDV standards would reduce this risk in 1995.
C.	Visibility Effects
Reduced visibility is one of the more readily apparent
effects of diesel particulate, as evidenced, for example, by
Denver's "brown cloud" mentioned above. Diesel particulate
reduces visibility by scattering and absorption of light.
Diesel particles are of a diameter most effective in scattering
light and their 65-80 percent carbon content produces a high
degree of light absorption. The DPS developed a method for
quantifying changes in visibility caused by changes in diesel
particulate levels. The reader is referred to that study for a
discussion of the methodology involved.[1]
Table 5-4 presents the estimated visibility impacts of the
base- and controlled-case scenarios in terms of the average
percent reduction due to diesel particulates in 1995 urban
visibility from early 1970's baseline levels. The visibility
figures were adjusted from those in the DPS report using the
ratios of emission inventories as described above. Although
the relationships between inventories and visibility is not
exactly linear, the resulting error is neglible due to the low
concentration of particles involved.
Table 5-4 shows that, in the absence of controls,
increases in diesel particulate levels will result in decreases
in visibility, ranging from a 21 percent decrease in largest
cities, to 3-8 percent decreases for less populous urban
areas. The proposed standards will cut the decrease to 14
percent in the largest cities, and to 2-5 percent in smaller
urban areas. The controlled-case scenario thus offers a 1 to 7
percent improvement in visibility over the base-case scenario,
depending on the size of the city. The lower limit of this
impact (i.e., the effect for smaller cities), may not be
perceptible. However, the effect for large cities would show a
noticable improvement in visibility. The proposed standards,
therefore, provide an overall benefit that would be most
apparent in the areas where it was most needed.
D.	Soiling Impacts
In a review of the scientific literature, the DPS found
some evidence suggesting that because of its black color and
oily nature, diesel particulate may have a disproportionate
effect on soiling compared to the effect of other types of'

-------
5-15
Table 5-4
Average Reduction in Visibility
Due to Diesel Particulate in 1995
(percent reductions from base-year visibility)
City Size (population)	Base	Controlled
More than 1,000,000	21.1	14.1
500,000-1,000,000	7.7	4.9
250,000-500,000	5.5	3.2
100,00-250,000	3.1	1.9

-------
5-16
particulate (i.e., diesel particulate would produce more
soiling than TSP on a gram-for-gram basis). The black color
may make the soiling more apparent to the observer and the oily
nature may make it more difficult to clean. The net effect
would be to increase costs to the general public for more
frequent and more thorough cleaning events. However, because
of the paucity of scientific data on the physical soiling
effects of diesel particulate and TSP, no definitive statement
of these relationships can be made at this time.
There is a somewhat larger body of literature available
regarding the costs associated with various levels of soiling.
A good summary of this economic literature can be found in an
EPA report regarding the benefits associated with the proposed
HDDV standards. [8] This report concludes that there are
significant economic benefits to be gained from control of
diesel particulates with respect to soiling.
E. Conclusions
Heavy-duty vehicles currently are the major source of
diesel particulate emissions from mobile sources in urban
areas. Unless controlled, the particulate emissions from this
source are expected to nearly double by 1995. Such high levels
can have a significant effect on ambient air quality in urban
areas. Implementing the proposed standards will essentially
hold the level of emissions from this source in 1995 at the
1983 level. The resulting reductions in ambient particulate
concentrations will be beneficial in helping numerous areas
meet the PM^q NAAQS currently under consideration by EPA.
They will also produce benefits in terms of health effects,
visibility, and soiling.

-------
5-17
References
1.	"Diesel Particulate Study," U.S. EPA, OAR, OMS,
ECTD, SDSB.
2.	Controlling Airborne Particles, Committee on
Particulate Control Technology, National Academy of Sciences,
Washington, D.C., 1980, p. 8.
3.	Letter from J. Padgett, Director, Strategies and Air
Standards Division, OAQPS to C. Gray, Director, Emission
Control Technology Division, OMS, May 1983.
4.	"Controlling Emissions from Light-Duty Motor
Vehicles at Higher Elevations," EPA 460/3-83-001, U.S. EPA, Ann
Arbor, MI, February 1983.
5.	"Assessing the Future of Denver's Haze; With
Attention to the Contribution of the Diesel Automobile,"
Dennis, R. L., National Center for Atmospheric Research,
Boulder, CO, August 1983.
6.	See Reference 2, p. 7.
7.	"Health Effects of Air Pollutants," U.S. EPA,
Washington, D.C., June 1976.
8.	"Health, Soiling, and Visibility Benefits of
Alternative Mobile Source Diesel Particulate Standards," Final
Report, EPA Contract No. 68-01-6596, Mathtech, Inc., Princeton,
NJ, December 1983.

-------
CHAPTER 6
COST EFFECTIVENESS
I. Introduction
Cost effectiveness is a measure of the relative economic
efficiency of an action toward achieving a specified goal. It
is primarily useful in comparing alternative means of achieving
that goal. The goal of the proposed regulations is to reduce
particulate and NOx emissions, or perhaps more importantly, to
reduce ambient levels of particulate and NOx in urban areas.
In this chapter, the cost effectiveness of attaining this goal
is expressed in terms of the dollar cost per ton of particulate
or NOx emissions controlled.
More specifically, the primary purpose of this chapter is
twofold: 1) to estimate the cost effectiveness of the proposed
particulate standards for heavy-duty diesel vehicles (HDDV) and
the proposed NOx standards for both light-duty trucks (LDT) and
heavy-duty vehicles (HDV),* and 2) to compare these values to
other mobile and non-mobile source control strategies. The
cost effectiveness of the proposed HDDV particulate standards
will be compared to the 1987 light-duty diesel motor vehicle
(LDD) particulate standards and to selected stationary source
controls. The cost effectiveness of the proposed NOx emission
standards will be compared to light-duty gasoline and diesel
vehicle (LDV) NOx controls and also to selected stationary
source controls.
Another purpose of this chapter is to estimate the cost
effectiveness of a particulate control scenario where separate,
more stringent standards are proposed for urban buses. The
analysis of this control scenario will be examined in Appendix
I.
To determine cost effectiveness, two pieces of information
are necessary: the costs and emissions reductions of the
strategies to be examined. The measure of costs will be either
the net present value of the purchase, operating, and
maintenance costs, which accrue throughout the life of the
vehicle, or an annualization of this lifetime cost. All the
Standards are: HDDV particulate of 0.60 g/BHP-hr at low
altitude and 0.72 g/BHP-hr at high altitude in 1987 and
0.25 g/BHP-hr at low altitude and 0.30 g/BHP-hr at high
altitude in 1990; HDV NOx of 6.0 g/BHP-hr in 1987 and 4.0
g/BHP-hr in 1990; Class I LDT NOx of 1.2 g/mi in 1987; and
Class IIA LDT NOx of 1.7 g/mi in 1987. The NOx standards
apply to both diesel and gasoline vehicles.

-------
6-2
cost figures presented in this analysis are expressed in terms
of 1984 dollars. The measure of emissions reductions will be
either the lifetime emission reduction or the annual emission
reduction. (Reasons for determining costs and benefits on
either an annual or lifetime basis will be explained later.)
Furthermore, for particulate, emission reductions will be
determined on a total, inhalable, and fine basis.* The three
classes of suspended particulate are examined in order to focus
the analysis on the most important particulate matter with
respect to public health and welfare. As discussed in Chapter
5, fine and inhalable particulate are of the greatest concern
for public health because this material is deposited in the
deepest, most sensitive portions of the lung. They also are
very important because of their disproportionate effect on
visibility. Total suspended particulate (TSP) is of concern
because it includes not only fine and inhalable particles, but
also coarser material that affects public welfare (i.e.,
soiling).
In addition, for particulate an attempt will be made to
incorporate the relative air quality impact (per unit rate of
emission) into the cost-effectiveness comparison.
The remainder of this chapter is divided into three major
sections. The first section discusses the methodology for
determining cost-effectiveness values for the proposed NOx and
particulate standards. The second section estimates the cost
effectiveness of the proposed particulate controls and compares
them to the cost effectiveness of various mobile and stationary
source particulate controls. The third section estimates the
cost effectiveness of proposed NOx controls and compares them
to the cost effectiveness of NOx control for other mobile and
stationary sources.
II. Methodology
This section presents the two methods which will be used
to estimate the cost effectiveness of the proposed NOx and
particulate standards: the annual approach and the lifetime
approach. To simplify the analysis somewhat, both methods
evaluate the costs and emission reductions associated with an
Total particulate is all suspended particulate matter
regardless of diameter, inhalable particulate is
considered to be all particulate matter less than 10
micrometers in diameter, and fine particulate is
considered to be all particulate matter less than 2.5
micrometers in diameter.

-------
6-3
average vehicle on a per vehicle basis, rather than the total
costs and benefits for the entire fleet.
With the annual approach, costs are allocated: 1) to each
year in which emission reductions were produced, and 2) in
proportion to the size of these annual reductions. The result
is a cost-effectiveness value which is applicable at any point
in the life of a vehicle, as well as over the vehicle's entire
lifetime. This approach will be used to evaluate both the
proposed diesel particulate and NOx standards. This will allow
comparisons on a consistent basis with recent EPA
cost-effectiveness estimates for LDDV and light-duty diesel
truck (LDDT) particulate controls, which were also computed
using this method. [1] It will also allow comparisons of the
proposed diesel particulate and NOx standards with stationary
source controls, where cost-effectiveness values are typically
determined on an annual basis.
With the lifetime approach, the lifetime costs are
discounted to the year of vehicle purchase and then divided by
the undiscounted lifetime emission reductions. The lifetime
approach generally yields lower cost-effectiveness values
relative to those produced by the annual approach. This -is
because the lifetime approach uses two different accounting
techniques to assess costs and emission reductions. Costs ace
disoounted, while benefits are undiscounted. The effect of
this is that lifetime emissions reductions are higher than if
they were discounted. On the other hand, the annual approach
assesses both costs and emission reductions using a similar
accounting technique (i.e., annualization) and, hence, are
comparable. The lifetime approach will be used only in
conjunction with the proposed NOx standards to allow
comparisons with past LDV cost-effectiveness studies, where
only this method was used.
In the remainder of this section, the derivation of cost
effectiveness for both the annual and lifetime approaches will
be presented. The discussion for each approach will progress
from a general overview of the method, to a more specific
discussion of the individual factors that are used in the
cost-effectiveness equations.
A. Annual Approach
1. Derivation of Annual Cost Effectiveness
As described above, the annual approach allocates the cost
of emission control: 1) to each year in which benefits were
produced, and 2) in proportion to the size of these annual
benefits.

-------
6-4
The first step in this approach is to determine the net
present value of the costs associated with the applicable
standard which accrue throughout the vehicle's lifetime. More
specifically, the net present value of the cost is the sum of
the incremental purchase price and annual costs for changes in
maintenance and operating expenses, discounted to the year in
which the vehicle was purchased. In equation form, and
assuming a conventional 10 percent discount rate evaluated
during the mid-year,* this is expressed as:
NPV = Xj/d.l)0,5 + X2/(l.l)1,5 + X3/(l.l)2,5
+ . . .X./U.l) (1 " 0,5)	(1)
Where:
NPV = net present value of vehicle's lifetime costs
(including purchase, operating, and maintenance costs).
Xi = cost in year i (includes allocation of purchase,
operating, and maintenance costs).
i = age (in years) of vehicle.
Equation 1 alone is not sufficient to determine the annual
costs, however. The next step is to allocate the costs in
proportion to the size of the annual emission reduction. For
purposes of this study, this means that the annual costs are
allocated in proportion to annual emission reductions, so that
the cost effectiveness (annual costs divided by annual emission
reductions) is the same for each year. In mathematical terms,
this means:
CEa = Xj/Ei = X2/E2 = Xi/Ei	(2)
Where:
CEA = annualized cost effectiveness.
X^ = annual cost in year i.
E^ = annual emission reduction in year i.
The use of this discount rate has been discussed in
Chapter 3. Evaluation is typically at the mid-year, which
represents the average point in time of annual cash flow.

-------
6-5
The annual emission reduction, E^, is the average
per-mile emission reduction multiplied by the average number of
vehicle miles traveled (VMT) each year. This results in the
following equation:
Ei = ER X VMTi	(3)
Where:
Ei = annual emission reduction in year i.
ER = average per-mile emission reduction.
VMTi = VMT in year i.
Combining the above three equations leads to an equation
where annualized cost-effectiveness values can be calculated.
This equation is:
CE. = NPV/[ER V VMT./(1.1)(1 " 0,5)1 = NPV/[ER(SUM)] (4)
A	i = 1 1
The numerator in Equation 4 is the net present value "of
the lifetime costs associated with each proposed standard as
derived in Chapter 3. These net present value costs will be
presented later in this chapter when the particulate and NOx
cost-effectiveness values are actually calculated.
The per-mile emission reduction for the proposed
standards, which appear in the denominator of Equation 4, are
based on emission rates that occur at the vehicle's half life
(i.e., the average lifetime emission reduction). The use of
average values assumes that the emission rate remains constant
throughout the vehicle's lifetime. In reality, emissions tend
to increase gradually with mileage. The use of averaqe values
slightly underestimates emission reductions early in life, and
slightly overestimates them late in life. The overall effect
on cost effectivenes is negligible, however, since these two
competing emission trends essentially cancel one another and
the deterioration rates for diesel particulate and NOx are
small. Hence, average emission reductions can be used to
simplify the analysis, without any significant effect on the
results.
Also, using these average values allows direct use of the
particulate and NOx emission results contained in Chapters 4
and 5, respectively. These emission results will be presented
later when the cost-effectiveness values are calculated.

-------
6-6
Unlike costs and emission reductions, the VMT summation
(SUM) portion of the denominator in Equation 4 was not
presented in an earlier chapter, and will be derived below.
2. Description of Annual VMT Estimates
The sum portion of Equation 4 requires that VMT be
estimated for each year throughout the lifetime of the average
vehicle. This study will attempt to estimate the annual VMT as
actually incurred by motor vehicles. This differs slightly
from EPA's cost-effectiveness analysis on diesel particulate
emission control in the Diesel Particulate Study {DPS)-[11
There, annual VMT was assumed to be constant throughout the
lifetime of the vehicle. It was determined by dividing the
estimated vehicle lifetime, in years, into the estimated
lifetime VMT. However, the fact that most vehicles travel more
during the first years and less during the latter years is
ignored when using constant annual VMTs. The analysis in this
chapter reflects the actual mileage accumulation for each year
of operation. This should increase the accuracy of the
estimates, since this is more representative of in-use
exper ience.
Estimates of annual VMT by aqe are shown in Table 6-1 for
LDVs,* LDTs, and the different weight classes of HDVs. These
vmt values are derived by multiplying: 1) the average annual
mileage accumulated for a vehicle with a lifetime of 20 years,
and 2) the annual survival rate for a particular model year
fleet. This manipulation accounts for the fact that as the
model year fleet ages, some of the vehicles will be scrapped
(i.e., annual VMT = 0). As will become clearer later, it is
important to include this phenomenon in the VMT estimates
because they are used to calculate the emission reductions for
the fleet average vehicle. Unless the vehicles leaving the
fleet are accounted for, the VMT for the fleet average vehicle
will be overestimated, resulting in an accompanying
overestimate of the emissions benefit. Therefore, the
corrected VMT values are used to determine cost effectiveness
rather than those for a surviving vehicle with a 20 year
lifetime. The annual VMTs and survival rates were taken from a
report by Energy and Environmental Analysis.[2]
Table 6-1 differs from that found in the EEA report[2] in
two minor ways. First, no VMT are shown beyond 20 years of
age, while the EEA report does present such values for LDTs and
Although LDVs are not affected by the rulemaking, their
annual VMT estimates are included here. As already
mentioned, the DPS assumed annual VMT was constant for
this class of vehicle. To provide better comparison, the
cost effectiveness from that study will be recalculated
using varying annual VMT.

-------
6-7
Table 6-1
Annual and Lifetime Average Per Vehicle VMT
	(thousands of miles)	
Year


Annual
VMT


LDVs*
LDTs*
HDGVs
LHDDVs**
MHDDVs
HHDDVs
1
14.69
14.78
18.10
23.09
26.42
61.50
2
13.30
13.37
16.30
21.95
25.72
60.64
3
10.95
10.42
14.67
21.02
25.26
59.72
4
9.90
10.54
12.77
19.42
23.86
52.14
5
8.50
9.19
11.14
17.68
22.04
45.32
6
7.69
8.13
9.53
15.66
19.74
38.96
7
7.12
6.57
8.20
13.73
17.41
33.55
8
6.02
6.27
7.03
11.75
14.90
28.81
9
4.85
6.03
6.00
9.59
11.99
24.62
10
4.14
5.53
5.03
7.53
9.19
18.54
11
3.52
5.00
4.15
5.81
6.92
16.95
12
2.80
4.32
3.44
4.56
5.30
14.17
13
1.87
3.83
2.80
3.56
4.87
11.53
14
1.41
3.80
2.27
2.82
3.19
9.38
15
1.09
2.08
1.84
2.25
2.53
7.58
16
0.73
2.51
1.45
1.74
1.94
6.00
17
0.58
2. 51
1.14
1.28
1.37
4.72
18
0.69
0.95
0.87
0.98
1.06
3.60
19
0.55
0.93
0.60
0. 74
0.84
2.75
20
0.34
0.84
0.40
0.40
0.53
1.97
Total
100.74
117.78
127.73
179.80
215.16
502.65
Li fetime
VMT
Gasoline-fueled and diesel-f ueled LDVs and LDTs ha\*e
identical VMT distributions.
Estimated by sales weighting Class lib, III, and IV VMT-
using estimates from the EEA study.[2]

-------
6-8
HDVs. The VMT beyond 20 years was discarded because it is
negligible and simplifies the subsequent calculations. Second,
EEA labels their HDV classes differently from EPA. The
differences are shown below.
Heavy-Duty Groups as Classified by EPA and EEA
Class IIB Class III-V Class VI
EEA Trk P.5-10
EPA	LHDV
MDV
LHDV
LHDV
MHDV
Classes
VII & VIII
HHDV
HHDV
As can be seen, EEA and EPA similarly group
even though the descriptions differ for each
exception is that EEA separates
EPA combines them as LHDVs.
LHDVs, EEA'S
weighted according to the sales projection contained in Chapter
3.*
In
Classes
order to
vehicle classes,
group. The one
IIB and III-V, while
derive annual VMT for
EEA's VMT estimates for Classes IIB and III-V were
The lifetime VMT values in Table 6-1, which come from the
EEA report[2], differ from independent EPA estimates.[1] TKe
differences are highlighted below.
Lifetime VMT Estimates (in miles)
LDVs
EPA 100,000
EEA 100,700
Ratio EPA/EEA:
0. 99
LDTs
120,000
117,800
1.02
HDGVs
110,000
127,700
0.88
LHDDV
110,000
179,800
0.61
MHDDV
268,000
215,600
1.25
HHDDV
529,000
502,600
1.05
Note that, for LHDDVs and MHDDVs, the EPA and EEA lifetime
VMT estimates differ significantly. In examining the EEA
document, it appears that the VMT data for Class III-V vehicles
(or LHDDVs under EPA labeling) are identical to VMT data for
Class VI vehicles (MHDDVs under EPA labeling). This implies
that EEA used an average VMT for both classes. Typically, the
lifetime for Class III-V vehicles is much shorter than that for
Class VI vehicles. The net effect of using an average VMT for
It should also be mentioned that the HDGVs shown in Table
6-1 are a sales weighted average of EEA data for all
heavy-duty gasoline vehicles. In the EEA report, the data
were segregated into their respective vehicle classes.

-------
6-9
both classes is to overestimate the lifetime VMT relative to
EPA's LHDDV category, and to underestimate lifetime VMT for
EPA's MHDDV category. Therefore, the difference between the
EPA and EEA estimates for these two HDV categories seems to be
readily explainable, and EPA's values appear to be better
estimates.
In order to compensate for these differences, the EEA VMT
values in Table 6-1 must be adjusted to reflect EPA's lifetime
estimates. This is accomplished by simply multiplying each
value in Table 6-1 by the ratio of lifetime VMT estimates for
EPA and EEA. These adjusted values are shown in Table 6-1A and
will be used in the subsequent cost-effectiveness calculations
with one additional change. Rather than calculate individual
cost-effectiveness estimates for LHDDVs, MHDDVs, and HHDDVs,
the analysis can be greatly simplified by combining the VMTs
for each of these classes into an average HDDV value for each
year and for the vehicle's lifetime. These composite VMT
values are found by sales-weighting the individual values. The
HDDV sales fractions used to perform this conversion ar.e:
33 percent for LHDDVs, 25 percent for MHDDVs, and 42 percent
for HHDDVs. These values were previously estimated in
Chapter 3.
Now that annual VMT values for LDDVs, LDDTs, HDGVs, and
HDDVs have been described, the subsequent cost-effectiveness
calculations can be simplified further by actually calculating
the SUM portion of Equation 4 for each mobile source category.
The resulting value for each vehicle category is shown below.
Summation of VMTi/fl.l)1 ~ ®.5 by Mobile Source Category
Category
LDVs
LDTs
HDGVs
HDDVs
Summation Value
(thousands of miles)
57.6
72.8
71. 5
211.2
These values will be used in the subsequent annual
cost-effectiveness calculations.
3. Urban Cost Effectiveness (particulate only)
Thus far, the annual cost-effectiveness methodology
considers nationwide emission reductions, without regard to
geographic location. Since the great majority of people who
are exposed to NAAQS violations for particulate matter live in
urban areas, the control of diesel particulate in these areas

-------
6-10
Table 6-1A
Annual and Lifetime Per Average Vehicle
	VMT (thousands of miles)	
Year
LDVs*
LDTs*
HDGVs
LHDDVs**
MHDDVs
HHDDVs
1
14.58
15.06
15.59
13.79
31.46
64.72
2
13.20
13.72
14.04
13.01
30.62
62.82
3
10.87
10.62
12.63
12.46
30.08
62.85
4
9.83
10.74
11.00
11.61
28.41
54.87
. 5
8.44
9.36
9.69
10.48
26.24
47.70
6
7.63
8.28
8.21
9.28
23.50
41.00
7
7.07
6.69
7.06
8.14
20.73
25.31
8
5.98
6.39
6.05
6.97
17.74
20.32
9
4.81
6.14
5.17
5.79
14.28
2 5.'91
10
4.11
5.63
4.33
4.46
10.94
19.51
11
3.49
5.09
3.57
3.44
8.24
17.84
12
2.78
4.40
2.96
2.70
6.31
14.91
13
1.86
3.90
2.41
2.11
5.80
12.13
14
1.40
3.87
1.96
1.78
3.80
9.87
15
1.08
2.12
1.68
1.33
3.01
7.98
15
0.72
2.56
1.25
1.03
2.31
5.31
17
0.58
2.56
0.98
0.76
1.63
4.97
18
0.68
0.97
0.75
0.68
1.26
3.79
19
0.55
0.95
0.52
0.44
1.00
2.89
20
0.34
0.85
0.34
0.24
0.63
2.07
Total
100.00
120.00
110.00
110.00
268.00
529.00
Avg. Life-
time Miles
* Gasoline-fueled and diesel-fueled LDVs and LDTs have
identical VMT distributions.
** Estimated by sales weighting Class lib, III, and IV VMT
using estimates from the EEA study.[2]

-------
6-11
should receive the greatest emphasis. This will be
accomplished by allocating the costs of the proposed standards
to the associated particulate reductions which occur only in
urban areas.
The urban emission reductions are found by using annual
urban VMT values in Equation 4, rather than annual nationwide
VMT. Annual urban VMT for each mobile source category is
simply the product of: 1) nationwide VMT for each year, and 2)
the fraction of nationwide VMT that occurs in urban areas. The
urban fraction for each mobile source category is: LDVs, 59.4
percent; LDTs, 48.8; and HDVs, 33.9 percent. [1] Other than
this change, the annual approach is used as described earlier.
B. Lifetime Approach
Cost effectiveness using the lifetime approach is found
simply by dividing the net present value of all costs that
accrue during the life of the vehicle by the total lifetime
emission reduction for the vehicle. That is, discounted costs
divided by undiscounted emissions. Expressed mathematically,
the lifetime cost-effectiveness value for each standard is
determined by the following equation:
CEl = NPV/(ER X VMTT)
Where:
CE^ = lifetime cost effectiveness.
NPV = net present value of lifetime costs.
ER = average per-mile emission reduction.
VMTf = total lifetime VMT.
As described in Section II.A.1. of this chapter, the net
present value of the lifetime costs used in the above equation
is found by Equation 1. The lifetime costs were derived in
Chapter 3 and will be presented later in this chapter when the
lifetime cost-effectiveness values for NOx are actually
calculated. The average per-mile emission reductions were
discussed in Chapter 4 and will also be presented later when

-------
6-12
the NOx cost-effectiveness values are calculated. The lifetime
VMT values were previously described in Section II.A.2. of this
chapter. Specifically, they are: LDVs, 100,000 miles; LDTs,
120,000 miles; HDGVs, 110,000 miles: and HDDVs, 326,000 miles.
Ill. HDDV Particulate Cost Effectiveness and Comparisons to
Other Sources
A. HDDV Particulate Control
This section will present the cost effectiveness of the
0.60 and 0.25 g/BHP-hr HDDV particulate standards, which are
proposed to begin in the 1987 and 1990 model year,
respectively. (Hereafter referred to as the 0.60 and 0.25 HDDV
standards.) As explained before, a single cost-effectiveness
value will be calculated using only the annual approach.
Table 6-2 summarizes the net present cost values and
average per-mile emission reductions that are used in
conjunction with Equation 4 to determine the annual cost
effectiveness of the proposed HDDV standards. As discussed in
Section II of this chapter, the cost information was taken from
Chapter 3. The emission factor information was taken from
Chapter 5. In addition, it should be noted that the costs and
emission reductions for the 0.60 HDDV standard are the result
of controlling an otherwise uncontrolled vehicle. For the 0.25
s-tandard, however, these values are incremental to the costs
and benefits that would result from supplementing the 0.60 HDDV.
The nationwide and urban cost-effectiveness values for
each standard are presented in Table 6-3. Two interesting
observations can be made based on the values in this table.
First, the 0.60 standard is more cost effective than the 0.25
standard. This is not surprising since the benefits of each
succeeding increment of pollution control normally is more
costly to attain. In this case, the small reduction in
particulate emissions under the 0.60 standard results from
using inexpensive control technology. The larger reduction
associated with the 0.25 standard is attained with relatively
more expensive control techniques. Second, comparison of the
nationwide to the urban cost-effectiveness values shows that,
on an urban basis, the cost effectiveness of the standards is
reduced (i.e., the cost per ton is higher). This is expected
because the costs of a nationwide control program are allocated
to a much smaller emission reduction (i.e., emission reductions
outside urban areas are ignored).

-------
6-13
Table 6-2
Costs and Emission Reductions for the
Proposed HDDV Particulate Standards
0.60 g/BHP-hr (1987) Standard
Averaqe Discounted Costs*
($/vehicle):
Purchase Price
Maintenance
Fuel Economy
Total
31
0
_0
31
Averaqe Emission Rates
(g/mi)
Without Standard 2.1765
With Standard 1.8656
Reduction 0.3109
0.25 g/BHP-hr (1990) Standard
Averaqe Discounted Costs*
($/vehicle):
Purchase Price
Maintenance
Fuel Economy**
Total
424
(32)
365-730
757-1,122
Average Emission Rates
(q/mi)
Without Standard 1.8656
With Standard 0.9323
Reduction 0.9333
1984 dollars discounted at 10 percent per annum to the
year of vehicle purchase.
One to 2 percent fuel penalty assumed.

-------
6-14
Table 6-3
Cost Effectiveness of HDDV Particulate Standards
	($/ton) *	
HDDV
Nationwide Cost Effectiveness:
0.60 (1987) Standard	430
0.25 (1990) Standard**	3,500-5,200
Urban Cost Effectiveness:
0.60 (1987) Standard	1,300
0.25 (1990) Standard**	10,000-15,000
* One ton equals about 907,000 grams.
** Incremental cost effectiveness.

-------
6-15
B. Comparison to Other Particulate Control Strategies
1.	LDP Particulate Control Strategies
To gain an understanding of the relative efficiency of the
proposed 0.60 and 0.25 HDDV particulate standards, their
cost-effectiveness values can be compared to those for other
mobile source particulate control strategies. This will be
done by comparing the HDDV standards to the trap-oxidizer based
LDD standards which are scheduled to become effective in the
1987 model year. These standards are 0.2 g/mi and 0.26 g/mi
for light-duty diesel vehicles (LDDVs) and light-duty diesel
trucks (LDDTs), respectively.
Cost-effectiveness values have already been estimated for
the LDD particulate standards in the DPS report.fl] However,
the cost-effectiveness values ( in that study assumed constant
VMT for each year of the vehicle's life, and costs were
alllocated simply by forming an annuity out of the net present
coat values over the life to the vehicle. Therefore, to be
comparable, these cost-effectiveness values must be revised
using the methodology described in Section II of this chapter.
Table 6-4 contains the net present cost values and the
per-mile emission reductions that are needed to complete the
cost-effectiveness calculations. These values were derived
using the methodology from the DPS report. [1] Table 6-4 also
shows the revised LDD cost-effectiveness values.
Comparing Tables 6-3 and 6-4 shows that the proposed 0.60
HDDV standard is more cost effective than the LDD particulate
standards. However, the reductions in particulate emissions
under the 0.60 standard are not significant, as shown in
Chapter 5. The proposed 0.25 HDDV standard appears to be
somewhat more cost effective than the LDDV controls and about
the same as LDDT controls on an urban basis. On a nationwide
basis, the cost effectiveness of the 0.25 HDDV standard as it
compares to that for LDDVs and LDDTs is just the opposite of
that described for the urban basis.
2.	Stationary Source Particulate Control Strategies
Another way of gauging the relative economic efficiency of
controlling particulate emissons is to compare the cost
effectiveness of the proposed HDDV particulate standards to the
cost effectiveness of reducing particulate emissions from
stationary sources. This comparson is presented in Table 6-5
on the basis of total, inhalable, and fine particulate. The
eight stationary source values shown in this table were taken
directly from the DPS report.[1]

-------
6-16
Table 6-4
Cost Effectiveness Calculation
for LDD Particulate Standards*
Average Discounted Costs ($/vehicle)**
LDDV	LDDT
1987 Standard 80-85	55-60
Average Emission Rates (g/mi)***
Without 1987 Standard 0.420	0.403
With 1987 Standard 0.213	0.266
Difference 0.207	0.137
Nationwide Cost Effectiveness ($/ton)
$6,100-6,500	4,200-4,600
Urban Cost Effectiveness ($/ton)
$10,000-11,000	8,600-9,400
*" Based on the annual approach.
** 1983 dollars discounted at 10 percent per annum to the
year of purchase.
*** 0.6 g/mi standard in effect for 1982-86; 0.2/0.26 g/mi
standard in effect for LDDVs/LDDTs, respectively, for 1987
and beyond. NOx standards are assumed to be 1.0 g/mi for
LDDVs and 1.2/1.7 g/mi for LDDTs.

-------
6-17
Table 6-5
Cost Effectiveness Comparison
for Particulate Control of HDDVs
and Stationary Sources ($/ton)[1,2,3]
Particulate Size Basis
Source [41		Total	Inhalable	Fine
HDDV, 1987 Standard	430	430	430
Cement Kiln	1	840	1,800
HDDV, 1987 Standard	1,300	1,300	1,300
(urban basis)
HDDV, 1990 Standard	4,400	4,400	4,400
HDDV, 1990 Standard	12,000	12,000	12,000
(urban basis)
Kraft Smelt Tank	12,000	14,300	21,700
Electric Arc Furnace	9,500	14,900	15,000
Borax Fusinq Furnace	13,300	17,500	19,700
Industrial Boiler	29,200	41,000	123,000
Kraft Recovery Furnace	32,400	41,400	59,000
Lime Kiln (baghouse)	47,200	58,100	90,800
Electric Utility	47,400	68,300	155,000
Lime Kiln (ESP)	75,700	94,400	152,000
[1]	Stationary sources are discounted to reflect their
relative ground-level effect.
[2]	1983 dollars.
[3]	For simplification, the midpoint of the ranges were used,
where applicable
[4]	Ranking based on inhalable particulate values. HDDVs are
shown as a single source.

-------
6-18
Before comparing the cost-effectiveness values in Table
6-5, two important points must be made. First, as discussed
earlier, evaluating particulate emissions based on their size
allows a better comparison of mobile and stationary source
controls from a health and welfare perspective. Diesel
particles are so small that they are almost all classified as
fine particulate. Stationary source particulate is much more
variable and included particles in all three size categories:
total, inhalable, and fine. (Note that all fine particles are
included in the inhalable fraction and that all fine and
inhalable particles are included in the total.)
The control of inhalable and fine particulate is most
important with respect, to its effect on public health and fine
particulate is solely important with respect to visibility.
Coarser particles, which are included in the total category,
are less important with respect to those two effects, but do
play a role in soiling. Therefore, evaluating emission
reductions on the basis of their size provides a better
comparison among sources with regard to the most important
health and welfare issues.
The second important point is that the stationary source
cost-effectiveness values shown in Table 6-5 have been adjusted
to reflect the relative air quality impact of those emissions
(i.e., ground level impact) compared to that of diesel
emissions. This provides a more relevant measure of
environmental impact relative to population exposure, than if
all emission reductions from stationary sources were compared
directly to those for diesels.
However, it must be added that although this adjustment
should increase the accuracy of the comparisons, it is still
rough and the result is not an ideal comparison. Also, it
still ignores the location of these ground level
concentrations, particularly with respect to the number of
people exposed and the local need for control (i.e., is the
area in or out of compliance with the NAAOS) . This may be a
significant limitation since stationary sources can often be
controlled on an individual basis (i.e., where the air quality
problems occur), while mobile sources cannot. The use of urban
cost-effectiveness values attempts to correct for this. Both
urban and nationwide figures will be used here in comparison to
stationary source figures as in some cases stationary sources
are controlled on a nationwide basis and in some cases on a
local basis.
The figures in Table 6-5 suggest that HDDV controls are
quite favorable when compared to stationary source controls,
regardless of the size of particulate examined or basis of

-------
6-19
control. Only the control of wet cement kilns appears to be
more cost effective than any of the proposed HDDV standards.
Thus, while this cost-effectiveness analysis is not able to
take all pertinent factors into account, it can be concluded
that there is no evidence that the proposed HDDV particulate
standards are not cost effective with respect to stationary
source controls, and in fact compare quite favorably.
IV. LPT and HDV NOx Cost Effectiveness and Comparisons to
Other Control Strategies
A. LPT and HDV NOx Cost Effectiveness
This section will present the cost effectiveness of the
proposed LDT and HDV NOx standards. The specific levels of
these standards are 1.2 and 1.7 g/mi for light and heavy LDTs,
respectively, beginning in the 1987 model year, 6.0 g/mi for
HDVs beginning in the 1987 model year, and 4.0 g/mi for HDVs
beginning in the 1990 model years. (Hereafter referred to as
the 1.2/1.7 LDT, 6.0 HPV, and 4.0 HPV standards.) The cost
effectiveness of proposed standards will be determined using
the lifetime and annual methodologies, which were discussed in
Section II.B. of this chapter.
A summary of the various costs and emission reductions
associated with the proposed NOx controls is shown in Table
6--6. Separate cost-effectiveness values will be presented for
the different fuel types within each vehicle class: LDGTs,
LPDTs, HDGVs, and HPDVs. The cost figures in this table were
previously developed in Chapter 3. The necessary emission
factors were taken from Chapter 4, with two important changes.
First, the NOx emission factors that are used in this analysis
are a weighted average of the low- and high-altitude factors
for each mobile source category. It is estimated that 97
percent and 3 percent of vehicles are driven at low and high
altitudes, respectively. (NOx emission factors at low and high
altitude differ only for gasoline vehicles.)
Second, the emission factors for LDGTs and LDDTs are
sales-weiqhted averages for the light and heavy classes within
each fuel type category. These composite emission factors were
used because the costs from Chapter 3 were presented as an
average of the two weight classes. The sales weiqhtina here is
also based on data contained in Chapter 3, which shows that the
light and heavy fractions of the LDGT class account for about
51 and 49 percent of the sales, respectively. For LDDTs, the
light and heavy fractions account for about 40 and 60 percent
of the sales, respectively. Moreover, because any projected
change in the sales mix is not expected to change the emission

-------
6-20
Table 6-6
Costs and Emission Reductions
for LDT and HDV NOx Standards
1.2/1.7/6.0 (1987 ^ Standard
Average
Discounted Costs*
($/vehicle)	
Purchase Price
.Maintenance
Fuel Economy**
Total Net Present
Value
Average Emission
Rates (g/mi)***
Without Standard
With Standard
Difference
4.0 (1990) Standard
LDGTs
44-87
0
0
44-87
LDDTs
35
0
0
35
2.29
1.25
1.04
2.42
1.63
0.79
HDGVs
2
0
0
9.740
8.477
1.263
HDDVs
16
0
0-730
16-746
26.715
22.385
4.330
Average
Discounted Costs*
($/vehicle)	
Purchase Price	N/A****
Maintenance
Fuel Economy**
Total Net Present
Value
Average Emission
Rates (g/mi)
Without Standard N/A
With Standard
N/A
N/A
18
0
_0
18
291
6
0-730
297-1,027
8.477
5.660
22.385
14.950
Difference
2.817
7.435
* 1984 dollars discounted at 10 percent per annum to the
year of vehicle purchase.
** Assumes a 0-2 percent fuel economy penalty for HHDVs.
*** These emission rates differ somewhat from those in Chapter
4. An explanation of the changes is given in Chapter 7.
**** Not applicable.

-------
6-21
factor significantly, it also is assumed to adequately
characterize future sales.
Table 6-7 presents the resulting cost-effectiveness values
for the proposed LDT and HDV NOx standards. Three interesting
observations can be made based on the values in this table.
First, the annual approach yields larger values than the
lifetime approach, as would be expected. Second, the overall
ranking of the proposed NOx controls from least to most cost
effective is HDGV, HDDV, LDDT, and LDGT. Third, within the
HDGV category and the HDDV catetory the cost effectiveness of
the proposed 6.0 standard compared to the proposed 4.0
standards can be considered as being very similar, again given
the uncertainty in the analysis.
B. Comparison to NOx Control Strategies for Other
Mobile and Stationary Sources
As for the proposed particulate standards, the relative
efficiency of the proposed LDT and HDV NOx standards can -be
determined by comparing their cost-effectiveness values to
those for other mobile source control strategies. Information
concerning the cost effectiveness for other mobile source
strategies is limited to: 1) inspection and maintenance (I/M)
for LDVs, and_ 2) the reduction in LDV NOx emissions from 1.0
g/mi to 0.4 g/mi.[61
The cost effectiveness of these two strategies is shown in
Table 6-8, along with the cost effectiveness of the proposed
NOx standards. As can be seen, the proposed NOx standards are
more cost effective, in all cases, than that shown for the
other LDV controls.
Turning next to stationary source control,
cost-effectiveness values were examined for four different
emission sources utilizing different control technologies.
These cost-effectiveness values were taken from a report by The
Aerospace Corporation for the California Air Resources Board.[7]
The four sources examined by Aerospace were refinery
heaters, industrial boilers, a CO boiler, and a glass furnace.
Aerospace analyzed costs for various sizes and load factors in
an attempt to cover the range of costs and benefits that would
be representative of the four selected sources.

-------
6-22
Table 6-7
Cost Effectiveness of
LPT and HDV NOx Standards ($/ton)
LDGTs
LDDTs HDGVs
HDDVs*
Cost Effectiveness
1.2/1.7/6.0 (1987) Standard
Lifetime Approach
Annual Approach
320-630
530-1,000
340
550
15
20
10-480
15-740
4.0 (1990) Standard**
Lifetime Approach
Annual Approach
N/A***	N/A
55
80
100-380
170-590
* The VMT factors used to calculate cost effectiveness are
weighted averages for LHDDVs, MHDDVs, and HHDDVs, where
the sales fraction is estimated at 0.30, 0.24, and 0.40,
respectively of total HDD sales.
*¦* Incremental cost effectiveness.
*** Not applicable.

-------
6-23
Table 6-8
Lifetime Effectiveness Comparison
of NOx Control for Mobile Sources
Cost Effectiveness
	Source			($/ton)
HDGVs, 1987 Standard	15
HDGVs, 1990 Standard	55
HDDVs, 1990 Standard*	100-380
HDDVs, 1987 Standard*	10-480
LDDTs, 1987 Standard	340
LDGTs, 1987 Standard*	320-630
LDVs (1.0 to 0.4 g/mi)	2,400
LDVs (I/M)	2,500
Ranked according to midpoint of range.

-------
6-24
Aerospace examined three different types of control
technology for each of the four sources. The control
technologies investigated include: 1) low NOx burners, 2) the
selective noncatalytic reduction (SNCR) system, and 3) the
selective catalytic reduction (SCR) system. Low NOx burners
are widely used on utility and industrial boilers. The
resulting NOx reductions are influenced by the burner
configuration, type of fuel burned, and the type of combustion
modification implemented prior to the use of the low NOx
burner. The SNCR system can be applied primarily to industrial
and utility boilers, CO boilers, and crude oil heaters. In
SNCR control processes, ammonia is injected and reacts
selectively with NO* at about 1,000°C, forming nitrogen and
water. The SCR system operates in nearly the same fashion as
the SNCR system, except that a base metal catalyst is used to
lower combustion temperature. Aerospace also considered
combinations of these three control technologies.
The cost effectiveness for each of the four stationary
sources is shown in Table 6-9. The range in costs is due to
different plant sizes, different operating load factors, and
different control technologies applied to achieve a certain
reduction in NOx emissions. These cost-effectiveness values
were inflated from 1981 to 1983 dollars by 13.7 percent, based
on the producer price index (PPI) for all industrial uses. [8]
Also shown in Table 6-9 are the cost-effectiveness values
of both the proposed 1.2/1.7 LDT, and 4.0 and 6.0 HDV NOx
standards. A comparison of these values with those for
controlling stationary sources shows that the 1.2/1.7 LDDT
standards and the 6.0 and 4.0 standards for HDGVs are more cost
effective than any of the stationary source controls. For
HDDVs and LDGTs, the range of cost-effectiveness values for the
proposed standards and stationary source controls overlap in
most cases, making specific comparisons more difficult.
Generally, however, in most cases the cost effectiveness for
HDDVs is better than that for the various stationary sources.
For LDGTs, the cost effectiveness of the proposed NOx standards
compares quite favorably to values for stationary sources.
Virtually all NOx emitted from stationary sources is in
the form of nitric oxide (NO).

-------
6-25
Table 6-9
Annual Cost-Effectiveness Comparisons for NOx Control
of LDTs and HDVs and Stationary and Mobile Sources
Cost Effectiveness
	Source*			($/ton)	
HDGVs, 1987 Standard	20
HDGVs, 1990 Standard	80
HDDVs, 1987 Standard	15-740
HDDVs, 1990 Standard	170-590
LDDTs, 1987 Standard	550
Glass Furnace	430-700
LDGTs, 1987 Standard	530-1,000
CO Boiler	310-1,700
Refinery Heater	610-2,700
Industrial Boiler	660-11,000
Values ranked according to midpoint of range.

-------
6-26
Overall, the cost effectiveness of the proposed NOx standards
for LDTs and HDVs appears to be very good relative to that for
stationary source controls. If the stationary source NOx
emissions were discounted to reflect their relative ground
level effect, as was done for particulate, the cost
effectiveness of the proposed LDT and HDV NOx standard would
compare even more favorably.
Of course, the same limitations that applied to comparison
of particulate cost effectiveness has to be considered here as
well. That is, the cost effectiveness comparisons do not
account for the fact that stationary sources can often be
controlled on an individual basis (i.e., where the air quality
problems exist), while mobile sources cannot. Nevertheless,
the available evidence indicates that controlling NOx from LDTs
and HDVs is cost effective relative to controlling these
pollutants from stationary sources.
V. Conclusion
The cost effectiveness of the proposed particulate and NOx
standards appears to be favorable when compared with other
mobile source control strategies. Also, despite some inhererft
difficulties in the comparisons with, stationary sources, the
cost effectiveness of the proposed particulate and NOx
standards appears to be favorable. Therefore, based on the
available information, the proposed standards appear to be a
cost effective means of reducing particulate and NOx emissions
compared to controlling these pollutants from other sources.

-------
6-27
References
1.	"Diesel Particulate Study," U.S. EPA, OAR, OMS,
ECTD, SDSB.
2.	"The Highway Fuel Consumption Model: Tenth
Quarterly Report," Undated Draft, Energy and Environmental
Analysis, Inc., Arlington, VA., 1983.
3.	"Particulate Size Variation in Diesel Car Exhaust,"
Groblicki, P., and C. Begeman, SAE Paper No. 790421.
4.	"Characterization of Diesel Exhaust Particulate
Under Different Engine Load Conditions," Presented at the 71st
Annual Meeting of APCA, Schrec, R., et al., June 25-30, 1978.
5.	"Characterization of Particulate and Gaseous
Emissions from Two Diesel Automobiles as Functions of Fuel and
Driving Cycle," Hare, C. and T. Baines, SAE Paper No. 799424.
6.	"Cost Effectiveness of Large Aircraft Engine
Emission Controls--Final Report," U.S. EPA, OAR, OMS, ECTD,
December 1979.
7.	"Assessment of Simultaneous Use of NOx Control
Systems on Stationary Sources in California," Vol. 1, The
Aerospace Corporation, El Segundo, CA, February 1982.
8.	"Economic Report of the President," (data gathered
by the Bureau of Labor Statistics), February 1983.

-------
APPENDIX A to Chapter 6
I.	Introduction
This appendix contains the cost-effectiveness analysis of
the proposed 0.10 g/BHP-hr diesel particulate standard for
urban buses beginning in 1990. This standard is assumed to
replace an existing standard of 0.60 g/BHP-hr, which is
proposed to begin in 1987. Therefore, the cost effectiveness
of the proposed 1990 standard is determined on an incremental
basis.
II.	Overall Methodology
Like HDDVs, the urban bus cost-effectiveness value will be
determined on an annual basis. However, somewhat less data is
available with which to characterize the urban bus fleet, so
that the calculation of annual cost effectiveness is simplified.
The cost effectiveness determination will be made on a per
vehicle basis by simply dividing the annual cost by the annual
emission reduction for the average bus. Each of these factors
is discussed separately in the following sections.
III.	Annual Cost
The annual cost of the proposed particulate standard is
found by annualizing the net present value (NPV) of the
incremental purchase price, in addition to the incremental
lifetime costs associated with maintenance and fuel economy.
The net present value of these costs are estimated by EPA as
follows.
Purchase price increment = $ 441
Maintenance increment = $ -12
Fuel economy increment = $1,750
Total NPV	= $2,179
The purchase price and maintenance increments are assumed
to be equivalent to those for MDDVs, as presented in Chapter
3. The fuel economy increment is based on an assumed two
percent fuel economy penalty, 4.30 mpg base fuel economy,
$1.20/gal fuel cost, 50,000 mi/yr travel, 10 year useful life,
and a 10 percent discount rate.
The annual cost of the total NPV, over the 10 year life of
the vehicle and at a 10 percent discount rate, is about $355.

-------
6A-2
IV. Annual Emission Reduction
The emission reduction is simply the difference in the
average lifetime emission rates under the 0.6 and 0.1 g/BHP-hr
standards, multiplied by the annual VMT. For urban buses
annual VMT is assumed to be constant over the life of the
average bus, i.e., 50,000 miles a year. The difference in the
average lifetime emission rates, i.e., per-mile rates, for the
two standards is found as follows.
Emission Reduction* = (7,1 lb/gal fuel) (0.6-0.1 g/BHP-hr)
Emission reduction (4>3 mi/gal fuel)(0.42 lb fuel/BHP-hr)
= 1.966 g/mi
Combining this average per-mile emission reduction and the
yearly VMT, as described above, yields an annual emission
benefit of 98,300 grams, or 0.108 tons per vehicle.
Note that unlike HDDVs, determining the emission benefit
in this manner assumes that none of the requisite
trap-oxidizers will fail. This assumption is based on the fact
that urban buses undergo routine maintenance that will tend to
remedy problems before failures result. Also, the operating
regime of these vehicles is fairly narrow compared to HDDVs.
This will allow manufacturers to better optimize the design
and, hence, durability of trap-oxidizer systems. If the
assumed "zero failure rate" proves to be overly optimistic, the
cost effectiveness of the proposed standard would be reduced.
V. Cost Effectiveness
Now that the annual cost and emission benefit are known,
the annual cost effectiveness of the 0.1 g/BHP-hr urban bus
standard can be determined. This is simply the annualized cost
(i.e., $355) divided by the annual emission reduction (i.e.,
0.108 tons), or $3,300/ton.
The use of this equation to convert g/BHP-hr to g/mi is
fully explained in the Diesel Particulate Study (DPS)-[1]

-------
6A-3
References for Appendix A to Chapter 6
1. "Diesel Particulate Study," U.S. EPA, OAR, OMS,
ECTD, SDSB.

-------
CHAPTER 7
ALTERNATIVE ACTIONS
I.	Introduction
In preparing a proposal for new standards, a wide variety
of alternatives have been considered by EPA. The evaluation of
alternatives is intended to identify the best approach
available to EPA, and is an essential element of a Regulatory
Impacts Analysis performed under Executive Order 12291.
This chapter will first of all briefly review the relevant
statutory requirements facing the Agency under the Clean Air
Act and the environmental needs being addressed by the
proposal. Past actions by EPA which are directly relevant to
the development and selection of options will then be
identified. Following these introductory items, individual
alternatives will be identified and analyzed.
II.	Statutory Requirements
The Clean Air Act (the Act) provides specific direction
concerning the control of both particulate and NOx emissions
from mobile sources. Considering the control of particulate
emissions first, Section 202 (a) (3) (A) (iii) mandates the
establishment of technology-forcing particulate emission
standards for heavy-duty diesels beginning in 1981. These
standards are to represent the greatest emission reduction
achievable in the leadtime available while considering costs
and other impacts. EPA had originally interpreted this
provision as applying to both light- and heavy-duty diesels and
initiated its efforts in the light-duty area. In subsequent
litigation, the Court of Appeals restricted the application of
Section 202(a) (3) (A) (iii) to heavy-duty diesels only, but
upheld EPA's light-duty diesel standards based upon EPA's
general regulatory authority contained in Section 202 (a)(1).
As for NOx control for heavy-duty engines, Section
202 (a) (3) requires NOx standards to be implemented in 1985
which provide at least a 75 percent reduction from the baseline
NOx level of gasoline-fueled heavy-duty engines. Recognizing
the possibility of excessive cost or fuel economy impacts, the
Act further provides for temporary revisions in these
heavy-duty standards, for periods not to exceed three years,
with subsequent standards always requiring additional emission
reductions. EPA announced in 1981 (through its NOx ANPRM) its
determination that the statutory NOx standard appeared
infeasible for diesel engines and indicated that EPA planned to
employ the revision provisions in its NOx standard-setting
process.

-------
7-2
The Act also provides the option to permanently change the
level of the statutory (or final) NOx standard, either up or
down, based upon study of the environmental effects of NOx
emissions from heavy-duty vehicles. However, such permanent
relaxation would have to be based upon a finding that there was
no environmental need for further NOx control from heavy-duty
engines, which is not feasible at this time.
Light-duty trucks are also included in this analysis.
This is because some light-duty trucks are included in the
Act's definition of heavy-duty vehicles. The Act includes as
"heavy-duty" all vehicles having a gross vehicle weight greater
than 6,000 lbs. On the other hand, EPA's light-duty truck
class includes all trucks up to 8,500 lbs gross vehicle
weight. Those light-duty trucks with gross vehicle weights
between 6,001 and 8,500 lbs are considered heavy-duty for
purposes of the Act, and thus are subject to the 75 percent
reduction requirements discussed above. Light-duty trucks with
gross vehicle weights under 6,000 lbs are being regulated under
the general authority of Section 202(a)(1).
III. Environmental Need
To provide perspective on the alternatives selection and
analysis process, the following paragraphs will briefly review
some of the key results of the NOx and particulate
environmental assessments from Chapters 4 (NOx) and 5
(particulate).
The analysis of Chapter 4 has shown that under current
regulations, motor vehicle emissions constitute nearly
two-thirds of the inventory in both 1980 and 1995, and they
produce half of the net emissions growth in that period.
However, NOx emissions from heavy-duty diesel engines, while
already sizable in 1980, double by 1995 and overwhelm any
expected reductions in other categories. Without the benefit
of new regulation, overall NOx emissions are expected to grow
about 20 percent between 1980 and 1995.
Currently, only one non-California region is in
nonattainment of the NAAQS for NO2, but this region has
petitioned to have its status revised and attainment is
expected to be granted soon. However, with the above growth in
NOx emissions, five non-California regions are projected to be
in non-compliance by 1995. In addition, increased NOx levels
have a deleterious effect on ambient ozone concentrations,
further aggravating the ozone attainment problem. Also, while
NOx emissions are not the main focus with respect to the
control of acid precipitation, they do contribute to the
problem and any increases are undesirable.

-------
7-3
In conclusion, further NOx controls	are needed	if
attainment of the NAAQS for NO2 is to be	maintained.	The
goal of this control should at least be	to maintain	NOx
emissions at their current levels.
Turning to particulate, the data in Chapter 5 have shown
that without further control diesel particulate emissions will
increase under any diesel growth scenario (i.e., diesel
fraction of sales). However, the light-duty portion of the
inventory is very dependent on sales and, with no increase in
current diesel market penetration, would only be 10 percent of
the total diesel particulate inventory.[1]
To put the air quality impact of these emissions into
perspective, 71-246 counties are projected to be currently in
violation of the range of possible standards for particulate
matter under 10 micrometers in diameter (PMiq) which the
Agency has proposed (49 FR 10408, March 20, 1984). The air
quality impact of current diesel particulate emissions is about
1-3 ug/m3 in urban areas, or roughly 2-5 percent of the
potential PM^g standards. Under best estimate diesel growth,
this impact will increase to about 2-8 ug/m3, or a net
increase of 1-5 ug/m3 (2-10 percent of the potential PM"io
s.tandards) , assuming £he current 1987 LDD particulate standards
are retained. However, even without this growth, numerous
violations of the NAAQS are still projected to occur in 1995
even after the application of reasonable non-mobile source
controls. Thus, additional particulate control from the mobile
source area is desirable if PM^q attainment is to be a
realistic goal.
Restricting growth in diesel particulate emissions is also
important for a number of other reasons. One, diesel
particulate is all inhalable and nearly all fine particulate,
remaining suspended for long periods of time. Two, it is
emitted at street level, directly into the breathing zone,
maximizing its impact on ambient air quality. Three, there is
a potential cancer risk associated with diesel particulate; the
average annual risk to an urban dweller being 1 to 6 per
million in 1995 with no further control. Four, due to its size
and high carbon content, diesel particulate is very effective
in reducing visibility; by 1995, reductions of 21 percent are
projected in the largest U.S. cities and 3 to 8 percent in
smaller cities. Five, due to its black color and oily nature,
diesel particulate has the potential for a greater than average
impact on soiling.
Overall, diesel particulate emissions will increase
regardless of the growth scenario chosen and will aggravate an
already serious ambient particulate problem. Cost-effective
diesel controls will need to be applied to minimize this impact.

-------
7-4
IV.	Prior Related Actions
A proposal for heavy-duty diesel particulate control was
published on January 7, 1981 (46 FR 1910). That notice
proposed that a particulate emission standard of 0.25 grams per
brake-horsepower hour (g/BHP-hr) be implemented for heavy-duty
diesels in the 1986 model year. This standard was expected to
require the application of particulate trap technology to
heavy-duty diesels, similar in concept to traps being developed
for light-duty diesels. The proposal also included appropriate
test procedures for particulate sampling based upon the new EPA
transient test requirement.
At about the same time# EPA published an ANPRM announcing
its intent to promulgate revised NOx emission standards for
heavy-duty engines and light-duty trucks {46 FR 5838, January
19, 1981). The standards discussed in the Advance Notice were
1.2 g/mi for light-duty trucks and 4.0 g/BHP-hr for heavy-duty
engines. Strictly speaking, neither of these standards
corresponded to the 75 percent reduction requirement of Section
202 (a) (3) (A) (ii). The light-duty truck standard at that level
would have been 0.9 g/mi, while the heavy-duty engine standard
would have been 1.7 g/BHP-hr. As discussed in the ANPRM, t-he
1.2 g/mi standard was chosen to provide a standard of equal
stringency to the existing 1.0 g/mi light-duty vehicle NOx
standard, while the 4.0 g/BHP-hr standard was identified as the
lowest standard applicable for heavy-duty diesels under the
revision provisions of Section 202(a)(3)(B).
Today's proposal contemplates the use of emissions
averaging to measure compliance with standards. EPA has
previously evaluated averaging for particulate emissions from
light-duty diesels, and has concluded that averaging can be
used there to reduce the cost of complying with emission
standards without significant environmental penalty. (For a
full discussion of these issues, see 48 FR 33456, July 21,
1983) . EPA has also previously announced that it was
considering the adoption of an averaging program for NOx
emissions from both heavy-duty engines and light-duty trucks
(45 FR 79382, November 28, 1980).
V.	Alternatives
Before examining specific alternatives, there are two
general areas needing consideration. These are leadtime for
new standards and emissions averaging.
It has already been noted that the Act specified the
adoption of particulate standards for 1981 and NOx standards
for 1985. In the case of particulates, the Administrator was
to consider "the period necessary for compliance" in

-------
7-5
establishing what technology would be available for the model
year to which the standards would apply. For NOx, leadtime was
also to be considered, but in this case there were specific
provisions calling for four years of leadtime.
Clearly, neither the 1981 date for particulate nor the
1985 date for NOx can be met at this time. For particulate,
the Agency sees its task as that of implementing standards as
soon as possible, while allowing necessary leadtime. For NOx,
in light of the inability to meet the 1985 deadline, the Agency
believes it appropriate to apply standards with less than four
years of leadtime, in the interest of having them take effect
as close to 1985 as possible. The criterion here is compliance
with what EPA believes the intent of the leadtime provision to
be, namely the provision of adequate time for compliance by
manufacturers. If a standard were developed which was clearly
feasible in less than four years, it would be inappropriate in
the light of the 1985 requirement to delay its implementation
beyond the period needed.
For heavy-duty engines, the analysis of technology
indicates that some modest emission reductions are possible in
the short term, but considerable time and effort will "be
required to obtain major reductions in emissions, particularly
for heavy-duty diesel engines. - Therefore, separate near-term
and long-term possibilities have been examined. The goal of
the near-term options is to implement a feasible degree of
control as soon as possible, while the longer term standards
are oriented toward attaining meaningful and needed emissions
reductions. Considering the time required to complete the
rulemaking process and technology considerations for
manufacturers to respond to an initial standard, 1987 has been
identified as the earliest year for a standard. Beyond that,
1990 has been identified as allowing time for implementation of
more advanced technology.
Light-duty trucks present a different picture than
heavy-duty engines in terms of leadtime. The technology for
NOx control applicable to light-duty trucks has already been
developed for passenger cars and will be easily adaptable to
light-duty trucks. In fact, all the techniques expected to be
used to meet new NOx standards are already being used on at
least a portion of the light-duty truck fleet. In this
situation, there is no need to look at short-term versus
long-term standards. All of the options which have been
considered are feasible in the short-term and are viewed here
as options for 1987.
The topic of emissions averaging has been thoroughly
examined through the earlier cited rulemaking to establish a
light-duty diesel particulate averaging program. In brief, the

-------
7-6
flexibility of averaging is one way to improve the
manufacturer's ability to comply with stringent standards. It
increases the economic efficiency of the standards without
decreasing the overall emission reductions they produce.
Averaging for heavy-duty rather than light-duty involves the
added complexity of two unique factors. These are the varying
useful life periods to which different subcategories of
heavy-duty diesel engines are certified and the difference
steming from the use of g/BHP-hr as the units for heavy-duty
engine emission standards, as opposed to the g/mi units
applicable in the light-duty case. Both of these factors mean
that different heavy-duty engines will have markedly different
lifetime particulate emissions even though certified to the
same standard. These two factors can, however, be dealt with
through proper design of the averaging program. Therefore,
averaging is considered below as an available option in
connection with the more stringent standards (the 1990
standards).
The following sections will describe first the heavy-duty
diesel particulate options, then the heavy-duty engine NOx
options and lastly the light-duty truck options. The emphasis
here will be on presentation of environmental and economic
impacts of the options. For further discussion of the
development proc.ess which led to the individual options
selected for analysis refer to the Diesel Particulate
Studyfl]. The technology analysis of Chapter 2 of the
Regulatory Impact Analysis also deals at length with the
feasibility of the final 1987 and 1990 options.
A. Heavy-Duty Diesel Particulate
Eight options are presented for particulate control for
HDDEs. These options and their key environmental and economic
impacts are shown in Table 7-1. Options 1-5 progress in
stringency, from least to greatest. Options 4A-4C are
variations of Option 4; Options 4B and 4C being attempts to
focus control in urban areas, where it is needed most. They
split the HDDE class into two categories, imposing less
stringent (and less costly) control on line-haul HDDEs which
spend less than a quarter of their time in urban areas. The
intended result is a more cost-effective program.
The first three options shown are based upon application
of non-trap technology to heavy-duty diesel engines. Option 1,
no further control, illustrates 1995 conditions if no standards
are adopted. Given the increase in emissions between 1983 and
1995 which it represents, coupled with statutory requirements,
this is not a viable option. Option 2 applies a 0.60 g/BHP-hr
standard beginning in 1987. This represents a level which EPA
believes feasible in the short term, without a large economic

-------
Table 7-1
Chart of Key Facts for HDDE Particulate Control Options[1]
Option
(g/BHP-hr)
1.	No Control
2.	1987 - 0.60
3.	1987 - 0.60;
1990 - 0.40
4.	1987 - 0.60;
1990 - 0.25
4A. 1987 - 0.60;
1990 - 0.25
(w/Averaging (A))
1995 HDDE
Emissions
(tons/year)[2)
97,000 (90%)[3]
84,600 (65%)
63,100 (23%)
48,600 (-5%)
48,600 (-5%)
4B. 1987 - 0.60; 63,600 (24%)
1990 - 0.25(A)
for Urban HDDEs;
1990 - 0.60
for line-haul HDDEs
4C. 1987 - 0.60; 54,600 (7%)
1990 - 0.25(A)
for Urban HDDEs;
1990 - 0.40
for line-haul HDDEs
Reduction
from Un-
controlled
Emissions
0%
13%
35%
50%
50%
34%
39%
1995 Total Mobile
Source Emissions
(tons/year)12]
131,600 (123%)[3]
119,200 (102%)
97,700 (65%)
83,200 (41%)
83,200 (41%)
98,200 (66%)
94,200 (59%)
Type of Control Technical Cost Per
System Required Difficulty Vehicle
5. 1987 - 0.60;
1990 - 0.10
34,200 (-33%)
65%
68,800 (16%)
None
Non-Trap
Non-Trap
100% Trap
70% Trap
Urban: 70% Trap
Other; Non-Trap
Urban: 70% Trap
Other: Non-Trap
100% Trap
None
Low
High
High
High
$ 31
195-390
917-1,282
757-1,122
Moderately 288-376
High
Low
Moderately 387-573
High
Moderately
High
Urban
Cost Effec-
tiveness
($/ton)
1,300
4,000-8,000
13,000-18,000
10,000-15,000
12,000-16,000
7,000-10,000
Very High 1,051-1,535 10,000-15,000
[1]	See Appendix A for description of procedures used to develop estimates of emissions, costs, and cost
effectivenesses. Costs and cost-effectiveness values are«incremental over Option 2 which is incremental over
Option 1.
[2]	Best estimate diesel sales assumed. Current 1987 PM standards assumed for LDCVs and LDDTs.
[3]	Figures in parentheses indicate change from 1983 levels. 1983 Diesel PM Emissions: HDDE - 51,200, LDD - 7,900,
Total - 59,100.

-------
7-8
impact. While it deals with the need for control in 1987, it
foregoes the opportunity to make use of available leadtime for
a technology forcing 1990 standard. Thus, this option does not
appear to satisfy the requirements of the Clean Air Act.
Option 3 adds a technology forcing 0.40 g/BHP-hr standard in
1990. A significant reduction relative to uncontrolled levels
would result from this option, while the need to develop trap
technology would be avoided. On the other hand, emissions
under this option would still increase from 1983 to 1995 and
the 1990 standard could well turn out to eventually require the
use of some traps.
Option 4 is a trap oriented option, with a 0.25 g/BHP-hr
standard for 1990. Option 4A maintains the same standards as
option 4, but applies particulate averaging for the 1990
standard. Both of these options would maintain 1995 heavy-duty
diesel particulate emissions at or even slightly below 1983
levels. Option 4A would accomplish this result at
substantially less cost and improved cost effectiveness than
would option 4 because traps would only be needed on
aproximately 70 percent of the fleet.
Options 4B and 4C have less stringent 1990 standards for
line-haul engines, in recognition of the fact that these
engines will spend much of their time in non-urban areas where
the need for control is less. This avoids the need to deal
with some of the most difficult trap applications although it
should be noted that line-haul vehicles can most easily bear
the cost since the cost of premium duty engines and line-haul
vehicles is already high. Option 4C, with its more stringent
0.40 g/BHP-hr standard for line-haul applications, will
maintain a focus on improved particulate control for line-haul
diesel engines, which is important since extensive changes in
engine technology are expected over the next decade.
Option 5 is the most stringent option considered. Traps
would be required on all engines, increasing the benefits but
also the technological risk. The potential for tampering with
emission controls (i.e., removal of trap oxidizers) could also
become a serious concern with this option since significant
fuel economy penalties might be associated with some of the
more difficult applications.
EPA's analysis of the pros and cons of all of the above
options has led to the decision to propose Option 4A. An
extended discussion of EPA's evaluation and the basis for the
final choices has been prepared for the preamble to accompany
the proposed rules, and will not be repeated here. EPA has
chosen to put that presentation in the preamble rather than in
this Regulatory Impact Analysis in order to make it available
to the widest possible audience. For the convenience of the
readers of this document, the preamble discussion is included-
as Appendix B to this chapter.

-------
7-9
Briefly, Option 4A was selected because EPA believes traps
to be feasible for heavy-duty diesel engines at a cost which
yields cost effective control compared to other particulate
control strategies. The line-haul option of Option 4C may also
be desirable but, if so, it is not clear at this time. The
0.10 q/BHP-hr standard of Option 5 was not selected both
because of the increased technological risk and because,
although indicated in Table 7-1 to be of equal cost
effectiveness to Option 4A, it is likely that Option 5 will
actually have a higher (worse) cost effectiveness than will
Option 4A.
B. Heavy-Duty Engine NOx
Four options exist for HDE NOx control, which are shown in
Table 7-2 along with their key environmental and economic
impacts. Urban control options are not presented due to the
wider geographical nature of NO2, ozone and acid rain
problems.
Option 1 would adopt no new standards for heavy-duty
engine NOx emissions. This option allows heavy-duty engine
emissions to increase some 80 percent over 1980 levels without
obtaining the .cost-effective NOx control which is available.
Option 1 also fails to satisfy the requirements of the Clean
Air Act since it does not represent the maximum degree of
reduction available for 1987.
Option 2 applies a 6.0 g/BHP-hr NOx standard beginning in
1987. This standard represents a technologically feasible and
cost effective level for that year, but not for the later
years. It would fail to make use of the available leadtime to
produce technology-forcinq standards for 1990.
Option 3 adds a 4.0 q/BHP-hr 1990 NOx standard to the 1987
standard of Option 2. Under this option, total 1995 NOx
emissions will be held to within seven percent of 1980 levels.
Ths most significant challenge in this option will be that of
feasibility for heavy-duty diesel engines where substantial
engine improvements will be required.
Option 3A is the same as Option 3 in all respects except
for the addition of a NOx averaging program for heavy-duty
engines. While it is likely that averaging would allow some
cost savings and reduce the technical challenge of the 1990 4.0
g/BHP-hr standard somewhat, the benefits do not appear to be
substantial and EPA has been unable to quantify any beneficial
effects. This is unlike the case for particulates, where, for
example, the ability to selectively aDply traps is a clear
benefit of averaging.

-------
Table 7-2
Chart of Key Facts fot HDE NOx Options[1]
Option
(q/BHP-hr)
1.
10.7 (No
further
control)
1995
HDE Emissions
(tons/year)
327,100 (80%)[4]
2. 1987: 6.0 286,200 (56%)
3.
3A.
1987: 6.0
1990: 4.0
1987: 6.0
1990: 4.0
(Averaging)
230,400 (26%)
230,400 (26%)
Reduction
from 1995
Base Case
HDE
Emissions
13%
30%
30%
1995
Total NOx
Omissions[2}
(tons/year)
874,800 (20%)[4]
833,900 (14%)
778,100 (7%)
778,100 (7%)
Technical
Difficulty
None
HDGE: Low
HDDE: Low
Cost
Cost Per Effectiveness
Vehicle ($/ton)[31
$16-746
HDGE: Moderate 18
HDDE: High 297-1,027
Improved
over 3
Improved
over 3
15
10-480
55
100-380
Improved
over 3
[1]	See Appendix a for descriptions of procedures used to develop estimates of emissions, costs,
and cost effectivenesses. Cost and cost-effectiveness values are incremental relative to the
previous option.
[2]	Assumes 1995 non-HDE NOx emissions of 547,700 (i.e., no further control).
[3]	Based upon lifetime cost approach
[4]	Figures in parentheses indicate increase over 1980 levels. 1980 NOx Emissions: HDE -
183,000, Total - 729,900

-------
7-11
EPA has chosen to propose Option 3 for heavy-duty engine
NOx standards. Analogous to the approach described above for
particulate options, EPA has formulated its discussion of the
analysis and selection process leading to the form of the
proposal for inclusion in the preamble to the NPRM. That
discussion is included as part of Appendix B to this chapter.
In summary, Option 3 provides significant cost effective
reductions in heavy-duty engine NOx emissions not obtained
through Options 1 or 2. Averaging, on the other hand, does not
have a clear enough benefit to argue for Option 3A at this time.
C. Light-Duty Truck NOx
Six 1987 control options are presented for controlling LDT
NOx emissions. These are shown in Table 7-3. Options 1 and 2
represent traditional LDT standards, applying to the entire
class. Options 2A and 2B relax the 1.2 g/mi standard to 1.7
g/mi for either all (Option 2A) or just diesel-powered (Option
2B) heavier LDTs (those above 3,999 lbs equivalent test weight
(ETW) or above 6,000 lbs GVW; referred to as LDT2S). (These
weight cut-offs are the same as those used by California and/or
the Act in determining vehicle classes.) Option 2C is the same
as Option 2 for all LDTs except heavier diesel-powered LD^s
where the existing standard (2.3 g/mi) would apply. Finally,
Option 3 is simply labeled "averaging" and applies to any of
the previous three options.
Option 1 represents a continuation of current standards
for all light-duty trucks and would require a finding, for the
heavier trucks in this class, that more stringent standards are
environmentally unnecessary. As noted in Chapter 4, although
their contribution is growing, NOx emissions from light-duty
trucks are a relatively small portion of the total urban NOx
inventrory. With the tighter standards in place for light-duty
vehicles and the proposed heavy-duty standards, more stringent
NOx standards for light-duty vehicles may not be needed to help
achieve and maintain attainment of the NO2 National Ambient
Air Quality Standard (NAAQS) until the middle or late 1990s.
This raises the possibility that more stringent LDT NOx
standards could be delayed and implemented sometime after
1987. It may also be possible to achieve similar reductions
from other sources at a lower cost. Nevertheless, there may be
environmental benefits associated with LDT NOx emission
reductions beyond the issue of NAAQS attainment (e.g., acid
rain).
Option 2 would establish a 1.2 g/mi standard for all
light-duty trucks. In the Advance Notice for NOx standards (46
FR 5838, January 19, 1981) a 1.2 g/mi light-duty truck standard
was derived by EPA as equivalent in stringency to the 1.0 g/mi

-------
Table 7-3
Chart of Key Facts for LPT NOx Options[l]
Option (g/mi)
LDDT2:2.3
3. Averaging
1995 LOT
NOx Emissions
Reduction From
1995 Base Case
LOT Emissions
1995 Total
NOx Emissions Technical
Cost
Cost
Effectiveness
1. 2.3
58,700
(42%)[4]
—
747,200
(2%)[4]
Low
—
—
2. 1.2
39,200
(-5%)
33%
727,700
(0%)
Moderate
Low
LDOT:$50-78[7]
IXJGT: 66-109[ 7 ]
350-550
460-760
2A. LDTL: 1.2[5]
LOT2: 1.7
41,600
(0%)
29%
730,100
(0%)
Low
LDOT: 35
LDGT: 44-87
340
320-630
2B. LDTl: 1.2
LDGT2: 1.2
LDOT2: 1.7
39,300
(-5%)[6]
33%
727,800
(0%)
Low
LDOT: 35
LDGT: 66-109
340
460-760
2C. LOTi: 1.2
LDGT2:1.2
40,000
(-3%)[6]
32%
728,500
(0%)
Low
LDOT: 10
LDOT: 66-109[7]
210
460-760
Same
Same
Same
Same Improved
Improved
[1]	See Appendix A for descriptions of procedures used to develop estimates of emissions, costs, and cost
effectivenesses.
[2]	Assumes 1995 non-LOT NOx emissions remain constant at 1980 levels (688,500).
[3]	Based upon lifetime cost approach
[4]	Figures in parentheses indicate increase over 1980 levels. 1980 NOx Emissions: LOT - 41,400, Total -
729,900.
[5]	LOTLs All LOTs at or below 3,999 lbs EIW or 6,000 lbs GVW.
LOT2: All LOTs above 4,000 lbs EIW and 6,000 lbs GVW.
[6]	This inventory is likely to be higher than indicated here. The model used to calculate the inventory figure
assumes a lower percentage of LDOT2S for 1995 than is currently anticipated.
[7]	Includes the effects of a one percent penalty in fuel economy for LOT^s at 1.2 g/mi.

-------
7-13
light-duty vehicle NOx standard. EPA still believes this to be
the case. This standard thus represents a measure of equity
between NOx control for light-duty vehicles and light-duty
trucks. As discussed in the Advance Notice, EPA also views a
1.2 g/mi standard as satisfying the intent of the 75 percent
reduction requirement of the Clean Air Act as it applies to the
heavier light-duty trucks (those above 6,000 lbs gross vehicle
weight). Light-duty truck NOx emissions under this option
would be controlled to the point where their overall level in
1995 would actually be somewhat below that in 1980, by an
estimated 5 percent.
Gasoline-fueled light-duty trucks would meet a 1.2 q/mi
standard principally through the application of three-way
catalyst systems, although a few of the smallest engines may
choose increased EGR rates or retarded ignition timing rather
than the expense of a three-way catalyst. Compliance would be
relatively straightforward, with about 40 percent of the fleet
already equipped with three-way systems. Diesel powered
light-duty trucks would be expected to use some combination of
EGR and injection timing retard to lower NOx emissions. The
only area of difficulty which arises in meeting this option
concerns the largest light-duty diesel trucks, such as tlje
General Motors 6.2-liter engine. It is possible that this
engine would experience performance and fuel economy penalties
in meeting a 1.2 g/mi NOx standard along with the 1987
liqht-duty diesel particulate standard. Application of an
electronic fuel injection timing system may reduce these
penalties to some degree, i.e., limit the fuel economy penalty
to no more than one percent.
Option 2A establishes a standard of 1.2 g/mi for light-
duty trucks up to 6,000 lbs gross vehicle weight or 3,999 lbs
equivalent test weight, and a standard of 1.7 g/mi above both
of those points. One impact of a tighter NOx standard for
diesel light-duty trucks is an increase in engine-out
particulate levels. This, in turn, means that compliance with
the 0.26 g/mi particulate standard will be more difficult and
require a greater reliance on trap oxidizers. Because both the
highest NOx emissions and highest particulate emissions are
associated with the heavier light-duty trucks, the potential
savings associated with maintaining a somewhat less strinqent
NOx standard for those vehicles has been evaluated. This
option also eases the difficulty noted above which the largest
diesel light-duty truck engines might have with meeting a 1.2
g/mi NOx standard.
The cutpoint chosen for subdividing the light-duty truck
class is 6,000 lbs gross vehicle weight, or 4,000 lbs
equivalent test weight. These values define two fairly

-------
7-14
distinct categories of light-duty trucks, and are parameters in
fairly common use. The 6,000 lbs gross vehicle weight criteria
is the same as that used in the Clean Air Act to separate
heavy-duty engines from light-duty trucks, while the equivalent
test weight distinction is used in California standards for
light-duty trucks. Use of a test weight distinction in
addition to gross vehicle weight is intended to discourage
artificial "migration" of vehicles from the lighter group to
the heavier group, with its attendant, less stringent standard.
In a fashion analogous to that which led to definition of
1.2 g/mi for light-duty trucks as corresponding to 1.0 g/mi for
light-duty vehicles, 1.7 g/mi can be identified as
corresponding to a 1.5 g/mi light-duty vehicle standard. Use
of a 1.7 g/mi standard for the heavier light-duty diesel trucks
will have the same effect as the current 1.5 g/mi light-duty
vehicle NOx waiver has for diesel passenger cars. The
percentage of light-duty diesel trucks requiring traps to meet
the 0.26 g/mi particulate standard will drop, from about 50 to
60 percent to about 20 to 30 percent, while the urban
particulate cost effectiveness will improve by a factor of two
to three. In exchanqe for this, 1995 light-duty truck NOx
emissions will increase slightly (about 6 percent), but still
will not exceed 1980 levels.
Option 2B is similar to Option 2A, except that the relaxed
1.7 g/mi standard would be applied only to heavier diesel
light-duty trucks. The principal reason for the split in
standards of Option 2A was that of impacts on particulate for
diesel engines. Thus an option based only on a relaxed
light-duty truck diesel standard deserves consideration. It
can be seen from Table 7-3 that this option is distinguishable
from Option 2A in terms of gasoline-fueled light-duty truck
cost and cost effectiveness, and overall emissions rates.
Option 2C is similar to Option 2B, except that the 2.3
g/mi standard would be retained for the heavier light-duty
diesel trucks. Effectively this means that the existing
standard would be retained for approximately 60 percent of
light-duty diesel truck sales.
Option 3 should be considered as a possible modification
to any of the other light-duty truck options through the
inclusion of averaging for NOx emissions. At this time,
significant benefits attributable to the adoption of such a
program have not been identified. On the other hand, there is
some potential in this option for slight increases in in-use
emissions, as there is in any mobile source averaging program.

-------
7-15
Overall, EPA believes that Option 2A is superior to the
other options. While Option 1 can not be completely dismissed,
Option 2A is preferable because it provides greater
environmental benefits and avoids the potential of being
inconsistent with the statutory requirement for heavier
light-duty trucks. Option 2A is superior to Option 2 because
it resolves a potential compliance problem for the heaviest
diesel light-duty trucks, substantially reduces the number of
traps needed to meet the light-duty diesel particulate
standard, and involves only a small penalty in NOx emissions.
While the principal motivation for this option concerns diesel
light-duty trucks, defining the option as applicable to both
gasoline-fueled and diesel light-duty trucks is desirable. The
overall impact of a 1.7 g/mi standard for all light-duty trucks
above 6,000 lbs gross vehicle weight and lbs equivalent test
weight on NOx emissions is small. It has the benefits of
regulatory simplicity and equity of standards between the two
engine types.

-------
7-16
Table 7-4
Impact of Alternative Light-Duty Truck Standards
on NO2 Attainment and Total N0X Emissions

Total Projected SMSAs
Exceeding
NCH NAAOS



(Expected and High
Growth Rates)1



1990
199 5
2000

LOT Standards
Exp Hiqh
Exp
Hiqh
EX£
Hioh
1.2/1.7
0 2
0
3
3
6
2.3
1 2
02
5
5
6
Total
NOx Emissions in 11 Urban Areas
-- All Sources


(Thousands
of Tons)




1990
1995
2000

LDT Standards
Exp High
Exp
Hiqh
Exp
Hiqh
1.2/1.7
1059 1152
1058
1201
1182
1411
2.3
1066 1159
1076
1222
1210
1444
(% Increase)
(.66) (.61)
(1.7)
(1.8)
(2.4)
(2.3)
1. Revised analysis based on 1981 design value and Mobile 3 mobile
source projections.
2. Under the expected growth and a 2.3 standard, there would be one
projected nonattainment area in 1996 increasing to five in the
year 2000.

-------
7-17
References
1. "Diesel Particulate Study," U.S. EPA, OAR, OMS,
ECTD, SDSB.

-------
Appendix A to Chapter 7
I. Derivation of 1995 Emission Inventories
The emission inventory figures for the control options
listed in this chapter were derived using standard EPA
emissions models. NOx emissions were calculated using a linear
rollback model and emission factors (emission rates) from EPA,
M0BILE2 program (modified). Particulate emissions were
calculated using a methodology developed for the EPA Diesel
Particulate Study (DPS).[1]	Inventories for the
uncontrolled-case and the controlled-case scenario for the
proposed NOx and particulate standards were discussed in
Chapters 4 and 5. Inventories for the alternative control
scenarios discussed in this chapter were calculated using the
same models or were derived from the various control scenarios
as described below. The derivation of the inventory for each
option is given below, organized by vehicle class.
A.	Particulate Control for HDDEs (Table 7-1)
Option 1 (uncontrolled case), Option 2 (0.60 g/BHP-hr 1987
and later), Option 4 (0.60 g/BHP-hr 1987-89; 0.25 g/BHP-hr 19-ftO
and later) and Option 4A (Option 4 with the addition of
averaging) are all discussed in Chapter 5. They are based on
the model employed in the DPS. These outputs, dated November
9> 1983, have been placed in the docket for this action.
Option 3 (0.60 g/BHP-hr 1987-89; 0.40 g/BHP-hr 1990 and later)
and Option 5 (0.60 g/BHP-hr 1987-89; 0.10 g/BHP-hr 1990 and
later) inventories were not discussed in Chapter 5, but are
alternative scenarios included in the computer model outputs
dated November 9, 1983. The inventories for Option 4B (same as
Option 4A, except that line-haul HDDEs, i.e., equivalent to the
upper half of class VIII, would meet a standard of 0.60
g/BHP-hr for 1990 and later model years) and Option 4C (same as
4B except line-haul HDDEs would meet a standard of 0.40
g/BHP-hr for 1990 and later model years) were derived from the
output of the computer run by combininq the HHDDE (line-haul)
inventory fraction from options 2 and 3, respectively, with the
LHDDE/MHDDE inventory fractions from Option 4A. The emission
factors data and methodology for inventory calculation may be
found in the DPS.
B.	NOx Control for HDDEs (Table 7-2)
Option 1 (no new standards), Option 3 (6.0 g/BHP-hr
1987-89; 4.0 g/BHP-hr 1990 and later) and Option 3A (Option 3
with averaging) inventories are discussed in Chapter 4 (also
see Figures 4-1 through 4-8). Together with Option 2 (6.0
g/BHP-hr for 1987 and later model years) these inventories were

-------
7A-2
calculated using the EPA rollback model (November 8, 1983).
The outputs may be found in the public docket.
C. NOx Control For LDTs (Table 7-3)
Option 1 (no new standards) and Option 2A (1.2 g/mi for
LDTs, and 1.7 g/mi for LDT2S) * inventories are discussed in
Chapter 4 and were based on November 8, 1983 outputs from the
rollback model. Inventories for Option 2 (1.2 g/mi for all
LDTs) were calculated from outputs of the rollback model
performed on October 8, 1983. Option 2B (1.2 g/mi for LDT^s
and LDGT2S, 1.7 g/mi for LDDT2S)* inventories were
calculated in rollback model outputs dated November 7, 1983.
Option 2C (1.2 g/mi for LDT^s and LDGT2S, 2.3 g/mi for
LDDT2S)* inventories were calculated in rollback model
outputs dated May 21, 1984. Copies of these outputs may be
found in the public docket.
II. Costs of Control Options
The vehicle/engine lifetime cost estimates for the control
options which were considered in developing this proposed rule
for HDDE particulate, HDE NOx and LDT NOx emission standards
were summarized in Tables 7-1 through 7-3. With the exception
of the first option (do nothing) for each combination of
regulated pollutant and vehicle/engine class shown in the
tables, the costs were derived from the cost estimates for the
proposed standards as developed in Chapter 3. The procedures
used in developing the costs of the options are described below
for each combination of regulated pollutant and vehicle/engine
class.
A. HDDE Particulate
Option 2 contained only the first part (1987 standard) of
the option which is proposed (4A). The cost was, therefore,
derived directly from Chapter 3 and is $31.
The cost estimate for Option 3 was developed on the basis
that a fuel economy oenalty would not occur as a result of a
0.4 g/BHP-hr standard and that the hardware and development
costs involved would be between one-third and two-thirds of
those estimated for trap-oxidizers, i.e., between $195 and $390.
The cost estimate for Option 4 was based on 100 percent
trap-oxidizer usage because averaging was excluded. The
effects on fuel economy were projected to be essentially equal
LDTi = 0-6,000 lbs GVW or 0-3,999 lbs ETW; LDT2 =
6,001-8,500 lbs GVW and 4,000 or more lbs ETW.

-------
7A-3
to those for the option proposed (4A). The projected cost is,
therefore, between $917 and $1,282.
Option 4B differs from the proposed option in that those
engines which are employed predominantly in line-haul
operations would continue to meet the 1987 standard in 1990
instead of meeting the more stringent 0.25 g/BHP-hr standard.
The cost for this option is, therefore, estimated as
attributable to the effects on fuel economy and the cost of
trap-oxidizers for LHDDEs and MHDDEs. The estimated cost is
between $288 and $376.
In Ootion 4Cr line-haul enqines would be required to meet
the same standard as in-Option 3 while other HDDEs would be
required to meet the same standards as in Options 4A and 4B.
The cost for this option would, therefore, be estimated as the
same as that for Option 4B plus the hardware and development
costs for HHDDE as were developed in Option 3. The estimated
cost is, therefore, between $387 and $573.
In Option 5, the costs for the trap-oxidizers and the fuel
economy penalty were assumed to be between 15 percent and 20
percent higher than those estimated for Option 4. The
resulting cost estimate is, therefore, between $1,051 and
$1,535.
B.	HDE NOx
The cost for Option 2 was derived directly from the
estimate for the proposed standard. It would consist only of
the cost of complying with the 6.0 g/BHP-hr standard which was
estimated in Chapter 3 to be $2 for HDGEs and $16-746 for HDDEs.
Option 3 is the approach which is proposed. The costs
are, therefore, those which were developed in Chapter 3 for the
1990 standards, i.e., $18 for HDGEs and between $297 and $1,027
for HDDEs.
Specific values for the cost of Option 3A were not
developed because specific tradeoffs which could be used in an
averaging program could not be identified.
C.	LPT NOx
In Option 2, all LDTs would be required to comply with a
1.2 g/mi standard. In the case of LDDTs which are above an
emission test weight of 3,999 lbs and a gross vehicle weight of
6,000 lbs (LDDT2) it is expected that additional injection
timing retard would have to be employed for compliance with the
1.2 g/mi standard. As a result of the timing retard, it is

-------
7A-4
anticipated that fuel economy could be degraded by between one
percent and two percent. The use of electronic injection
timing control could limit the fuel economy penalty to no more
than one percent and its cost is included. In estimating the
cost of this option it was assumed that essentially 100 percent
of the large diesel engines used in LDDTs will continue to be
used in LDDT2 as was the case in the 1983 model year. With
60 percent of LDDTs using large diesel engines, Option 2 would
result in a LDDT fleet fuel economy penalty of between $25 and
$50 per LDDT using the fuel economy cost estimate as developed
in Chapter 3. The cost of Option 2 for LDDTs is, therefore,
between $50 and $78. In the case of gasoline fueled LDTs, it
was assumed that Option 2 would necessitate the use of
closed-loop three-way-plus-oxidation catalysts on all
8-cylinder gasoline engines used in LDGTs, i.e., 8-cylinder
gasoline engines would be used totally in LDGT2 and about 20
percent of LDGTs would be LDGT2 (22 percent were LDGT2 in
the 1983 model year). Using the costs developed in Chapter 3,
the cost of Option 2 for gasoline fueled LDTs was estimated to
be between $66 and $109 including up to a one percent fuel
economy penalty for LDGT2S.
Option 2A is the proposed approach, and the costs a&e
those of Chapter 3. The range given in Table 7-3 for cost per
vehicle depends on the percentage of LDGTs already equipped
with three-way catalysts in model year 1986. The higher cost
(and cost effectiveness) assume that three-way catalyst use in
1986 is the same as that in 1984. The lower cost (and cost
effectiveness) assume that three-way catalyst installation on
LDGTs continues to increase between the 1984 and 1986 model
years at the same rate at which it increased between the 1982
and 1984 model years.
In Options 2B and 2C, the effect on gasoline fueled LDTs
with respect to technology application would be the same as
that of Option 2. The cost would, therefore, be the same,
i.e., between $66 and $109 per LDGT.
As in the case of Option 3A for the HDE NOx standard, the
development of a cost estimate for the effects of averaging
(Option 3 for LDT NOx) was not able to be performed.
Ill. Derivation of Cost Effectiveness
The cost effectiveness for each particulate and NOx option
was based on the per-vehicle costs as described above, the
per-vehicle emission factor data presented in the DPS report [1]
for particulate, and modified versions of the emission factors
given in Chapter 4 for NOx. The emission factors used to
calculate the lifetime NOx emissions values were modified as.

-------
7A-5
follows from those shown in Chapter 4: 1) The effects of
misfueling were removed from the calculation, under the
assumption that misfuelinq would be largely eliminated through
future reductions in the availability of leaded fuel. 2) LDDTs
between 6,001-8,500 lbs GVW currently maintain a considerable
"cushion" between the standard and the actual certification
levels. When the new, more stringent particulate standard is
implemented for the 1987 model year, it is assumed that these
trucks will be certified much closer to the actual NOx
standard, due to the NOx/particulate tradeoff. Retaining the
2.3 g/mi NOx standard with the 0.26 g/mi particulate standard
would therefore result in a net increase in NOx emissions over
the current standards. It is assumed that NOx certification
levels at 1.2 or 1.7 g/mi standards will be correspondingly
close to the standards. 3) The emission factors were corrected
to more accurately reflect the transition to three-way catalyst
systems which has already occurred for LDTs.
While none of these changes will significantly alter the
overall emission projections of Chapter 4, they do have an
impact on the LDT cost effectiveness values. Therefore, this
information was used to calculate the cost effectiveness for
each option using the methodology discussed in Chapter 6. For
particulate, the cost effectiveness was determined on an annual
basis. For NOx, it was determined on a lifetime basis. The
different methods were used for each pollutant to be consistent
with past mobile source cost-effectiveness analyses.

-------
Appendix B to Chapter 7
Discussion of Alternatives
for Heavy-Duty Engines,
from the Preamble to the Proposed Rule

-------
D.
Alternatives
As will be seen, control of either particulate or NOx to
the degree needed to adequately deal with the environmental
needs outlined above involves sianificant technological
challenae and has substantial economic implications.
Therefore, EPA has considered a wide ranqe of alternatives in
arriving at the final form of today's oroposal. These
alternatives, and EPA's evaluation of each, are reviewed
below. Interested readers are referred to the Alternatives
chapter of the Reaulatory Impact Analysis for details bevond
those presented here.
In general, available alternatives deal with combinations
of two factors: stringency of standards and leadtime. In terms
of stringency, EPA has considered options varyinq from no new
standards up to . the limits of foreseeable technoloqy. The
practicality of anv of those technological options varies,
however, depending on the time allowed for implementation.
Therefore, varying amounts of leadtime have also been
analyzed. Within the spectrum of standards and leadtime, there
is in fact a near continuum of possible options. EPA has
attempted to focus specific options at points which are
reasonably distinctive in one or more of three areas:
technology, benefits, or required leadtime.
1. Heavy-duty Diesel Particulate
Identification of Options
Clean Air Act reouirements and the environmental situation
both argue strongly for the early implementation of standards
to control heavy-duty diesel particulate emissions. On the
other hand, the technological barriers to major reductions are
of the sort which may require considerable time and effort to
overcome. Therefore, EPA has examined separate near- and
lona-term possibilities. The goal of the near-term options is
to implement the qreatest degree of control feasible without
undue adverse impacts as soon as possible, while the longer
term standards are oriented toward attaining more substantial
and needed emissions reductions. Considerinq the time required
to complete the rulemakinq process and technoloqy
considerations for manufacturers to respond to an initial
standard, EPA has identified 1987 as the earliest year for a
standard. Beyond that, 1990 has been identified as allowinq
time for implementation of more advanced technoloqy. For these
two years, the specific options are identified below alonq with
some of the key considerations for each.
Before turninq to individual options, it is important to
note that, although the particulate and NOx options are

-------
7B-2
considered separately here, they do in fact interact. In
developing and analyzing the following options, EPA has fully
considered those interactions when assessing feasibility,
technology and cost. The process was carried out in a fashion
which would insure the compatibility of the final selected
options. For example, the assessment of the number of traps
needed to meet the 1990 trap-based particulate options was done
on the basis of the existence of the more stringent NOx
standard also expected for that year. The two pollutants are
discussed separately for clarity of presentation, but readers
should bear in mind that they are interrelated and are in the
end to be viewed as pairs of particulate and NOx standards.
For 1987:
No standard. This option represents the situation if EPA
were to take no action to control heavy-duty diesel particulate
emissions. In light of both the Clean Air Act mandate and the
significant growth in particulate emissions which will occur
without particulate control, this is not a viable option.
0.60 g/BHP-hr. This option represents what EPA believe's
to be the lowest feasible level in the near term. Current
engine emissions are in the 0.4 to 0.8 g/BHP-hr range. The
target loW mileage emi'ssion level associated with a 0.60
g/BHP-hr standard is about 0.46 g/BHP-hr. Thus, some engines
already meet the required target, while reductions would be
required of others. EPA believes such reductions are feasible
with technology and engine calibration changes available in the
short term. Such things as improvements in fuel injection
systems, increased use of turbocharging, and improvements in
engine efficiencies should be able to provide the reductions
needed.
For 1990:
Since the 1990 options represent fairly stringent
standards, the incorporation of an emissions averaging program
in 1990 has been included for all but the "no further control"
case. As noted earlier, averaging reduces the risk of
non-compliance as well as the cost of stringent standards. It
will find its greatest usefulness with the trap-based options.
No control beyond 0.60 g/BHP-hr. While 0.60 g/BHP-hr is a
feasible limit for tKi near term standard, it produces
relatively small per-vehicle reductions in particulate
emissions. Overall, fleetwide heavy-duty diesel engine
particulate emissions would increase by 65 percent between 1983
and 1995 under this option. In addition, such a standard for-
1990 or later fails to take advantage of the technological

-------
7B-3
progress that should be made in the three years after 1987.
Therefore, 0.60 g/BHP-hr in 1987 without follow-on reductions
could not be considered an adequate alternative for longer-term
particulate control.
Q.40 g/BHP-hr. In moving beyond the 0.60 g/BHP-hr level,
EPA believes that a reduction to at least 0.40 g/BHP-hr will be
essential to produce further meaningful progress in reducing
diesel particulate emissions. This level represents both a
sizeable reduction from the 0.60 g/BHP-hr level and EPA's
determination of the approximate technological limit for
reductions which could be reached without the application of
particulate trap technology to heavy-duty diesel engines. It
involves substantial improvement over present engines. Some of
this improvement will come from techniques already being
developed for improved performance and fuel economy (the full
cost therefore not being attributable to new standards).
However, this level also presumes the further development and
application of advanced and costly technology, to a degree
which can not be precisely quantified at this time. Future
controls will include such things as the anticipated
application of electronic controls to engine operations lik~e
fuel injection and exhaust gas recirculation, the development
of low heat rejection techniques and ceramic materials,
increased fuel' injection pressures, and other efficiency
improving engine changes. Because of uncertainty regarding the
mix of control technology which will eventually be used to meet
the standard, costs are likewise uncertain. EPA's best
judgment estimates are that the discounted lifetime costs for a
0.40 g/BHP-hr standard would be from $195 to $390 per vehicle,
depending on fuel economy impact. Particulate emissions in
1995 would be reduced by 35 percent compared to the no-control
case, but net heavy-duty diesel particulate emissions would
still increase more than 20 percent from 1983 to 1995.
0.2 5 g/BHP-hr. Reducing particulate emissions to below
0.40 g/BHP-hr is expected to require the application of traps.
Use of this technology is still in the early stages of
development for heavy-duty diesels, and admittedly much more
work, is needed before traps will be feasible for production
engines. Traps on heavy-duty diesel engines must perform
successfully in an environment which is in some ways more
challenging than that for light-duty diesel traps (for example,
successful regeneration must be possible over a wider range of
sustained operating temperatures), and must do so for a
generally much longer useful life period. However, at this
time EPA sees no insurmountable obstacles to successful
application, especially given the substantial amount of
leadtime remaining before such a standard would go into effect'

-------
7B-4
and the already advanced stage of particulate trap development
for liqht-duty diesels.
EPA believes that it will not be possible to move below
the 0.40 a/BHP-hr option without the introduction of trap
technolooy. As the standard is lowered below that level, there
would be an increasing use of traps within the fleet, with
manufacturers making use of averaging to minimize the total
number of traps reauired. Because EPA foresees that some
applications will be technically more difficult than others,
there is some advantaqe in a standard which does not require
traps on all enqines, so that the more technically difficult
applications can be avoided. With averaging and a 0.25
q/BHP-hr standard, traps will be needed on an estimated 70
percent of the fleet. At this level, heavy-duty diesel engine
emissions would be controlled to the point where emissions in
1995 would be at or even slightly below those of 1983 . These
emissions represent about a 50 percent reduction from
uncontrolled levels, at a per-vehicle cost of about $760 to
$1,100, again deoendinq on fuel economy impact.
Urban-oriented variations on 0.25 q/BHP-hr. There is wide
variation between different heavy-duty diesel enqine
applications in the amount of total mileaqe accumulated within
urban areas. Liqhter heavy-duty diesels and such vehicles as
transit buses, trash compactors, and cement mixers spend the
predominant amount of their time in urban operation. Premium
heavy-dutv diesels used in over-the- road line-haul travel, on
the other hand, accumulate most of their mileaqe in rural
inter-citv operation. Since EPA's main concern with diesel
particulate emissions is their impact on urban air Quality, two
urban-oriented variations on the 0.25 g/BH^-hr trap option have
been evaluated which attempt to focus control primarily on
urban vehicles. Both of these variations maintain a non-trap
standard for line-haul enqines. (Line-haul enqines in this
case would be defined as those enqines used in Class VIII
trucks havinq gross vehicle weiqht ratings above 60,000 lbs.)
Elimination of the need for traps on these vehicles would avoid
the most difficult applications, in trade for a moderate loss
in overall urban emission reduction. On the other hand, it is
also true that line-haul vehicles would be best able to absorb
the high cost of traps because of their already high initial
cost compared to other heavy-duty diesel engine applications.
The line-haul standards considered are 0.60 and 0.40
q/BHP-hr. The 0.60 q/BHP-hr level is, of course, a much easier
standard to meet than the 0.40 q/BHP-hr level. Unfortunately,
even the small fraction of line-haul mileage which is urban in
nature is sufficient to raise overall heavy-duty diesel
particulate emissions in urban areas under this ootion to above'

-------
7B-5
those under a uniform 0.40 g/BHP-hr non-trap standard.
Moreover, 0.60 g/BHP-hr could not be considered a
technology-forcing standard for line-haul engines in the 1990
timeframe. Thus, a 0.40 g/BHP-hr line-haul standard is the
most loaical choice to combine with the 0.25 g/BHP-hr
trap-based standard. with this combination, there would be a
loss of control of about 12 percent in urban areas as compared
to the uniform 0.25 g/BHP-hr standard, and 1995 urban
heavy-duty diesel engine emissions would be about 7 percent
greater than 1983 levels. As can be seen from these figures,
even predominantly over-the-road line-haul operation has a
significant urban impact.
One difficulty which EPA has had with this approach has
been that of developing a regulatory procedure for successfully
separating urban from non-urban applications. Heavy-duty
diesel engines are currently certified independently from
vehicles, and a given engine may be used for a wide variety of
applications.	An urban designation would reouire
identification of intended vehicle application, and since some
engines would be used for both urban and non-urban
applications, the manufacturer would be required to develop
both trap and non-trap versions of the same engine to be able
to take advantage of the optional standard. Such a situation
might offer little advantage to the manufacturer.
Additionally, an urban option might introduce unwanted
competitive effects, for example by favoring manufacturers who
made only those engines falling into the line-haul exclusion
and who would not have to undertake the resource investments
required of other manufacturers to develop trap systems.
0.10 g/BHP-hr. It was noted under the discussion of the
previous option that a 0.25 g/BHP-hr standard can be met with
traps on about 70 percent of the fleet if averaging is
allowed. If the standard were further reduced until
essentially 100 percent trap usage were required, the resulting
level would be approximately 0.10 g/BHP-hr. This level thus
represents the maximum degree of reduction of heavy-duty diesel
particulate emissions that can be achieved with the use of trap
technology.
Requiring traps on all engine families, however, would
increase both the cost and the risk associated with the
standard. The technical challenge of successfully applying
traps to all engines is much greater than that of applying them
to most engines, since in the latter case the manufacturers
will be in a position to focus resources on predominant uses
and avoid especially difficult cases. At the same time, a 0.10
g/BHP-hr standard would maximize the available emissions
reduction, bringing about a 65 percent reduction from

-------
7B-6
uncontrolled levels and reducinq total 1995 heavy-duty diesel
oarticulate emissions to a level 33 percent below those in 1983.
There is a modified application for the 0.10 q/BHP-hr
option which relates to the control of urban bus emissions. As
has been noted earlier in the discussion of averaging, EPA
believes it is very important to obtain the maximum degree of
control for urban buses. Therefore, in addition to excluding
such engines from the averaging program, EPA is considering the
establishment of a standard of 0.10 q/BHP-hr for bus enqines to
insure the fullest use of the emission reduction potential of
traps.
Evaluation of Particulate Options
In its evaluation of all of the above options, EPA has
considered a number of factors. These included the statutory
requirements, anticipated costs and emission reductions, cost
effectiveness, and technological reauirements for compliance.
Some of these factors have alreadv been touched upon in
discussion of the individual options. The following material
provides further key information for comparinq the impacts of
the options.
The emissions'impacts of the various options are shown in
Fiqure 3, where the options are arranqed in order of decreasinq
emissions. For comparison purposes, the level of 1983
emissions is also shown.
As can be seen, meaningful reductions from uncontrolled
levels reauire a standard of 0.40 q/BHP-hr or lower. Even at
that level, heavy-duty diesel particulate emissions in 1995
will be 23 percent hiqher than in 1983. To further regulate
emissions so they do not increase above 1983 levels reauires
some form of a trap-based standard. Essentially meeting that
qoal are either the 0.25 or 0.10 q/BHP-hr standards, or the
0.25 q/BHP-hr standard with a 0.40 g/BHP-hr line-haul option.
However, it is noteworthy that the 0.60 g/BHP-hr line-haul
option does not perform especially well, having emissions
actually somewhat higher than the across-the-board 0.40
g/BHPjhr standard. The 0.10 g/BHP-hr standard attains the
maximum overall reduction, beinq 65 percent below the
uncontrolled level and 33 percent below 1983 levels.
Figures 4a and 4b present other information related to the
economic impacts of the options, in particular, cost
effectiveness, which measures the economic efficiency of each
option, and midpoint cost per engine (including operating
costs). For clarity of presentation, the cost per engine
values plotted are the midpoints of the ranges (based on.
possible fuel economy impacts) of estimated costs.

-------
Figure 3
1995 Urban Heavy— Duty Diesel
Particulate Emissions
Level of Standard (g/BHP-hr)

-------
7B-8
F.yure 4a
Midpoint Cost Effectiveness of
Heavy-Duty Diesel Engine Particulate Control
Level of Standard (g/BHP-hr)
Figure 4b
Midpoint Cost Per Engine of
Heavy-Duty Engine Particulate Control
Level of Standard (g/BHP-hr)

-------
7B-9
Cost effectiveness values commonly increase (worsen)
as the standards get lower. However, there are two
exceptions to this trend in Figure 4a. These are the 0.25
g/BHP-hr standard combined with the 0.60 g/BHP-hr standard
for line-haul trucks, which is the least cost effective of
the options shown, and the 0.10 g/BHP-hr full-trap
standard, where the cost effectiveness is shown as eaual to
that of the 0.25 g/BHP-hr standard.
The standard of 0.25 g/BHP-hr combined with 0.60
g/BHP-hr for line-haul trucks is the least cost effective
of the options presented in Fiaure 4a. As was shown in
Fiqure 3, urban heavy-duty diesel particulate emissions are
actually slightly greater under this option than under an
across-the-board standard of 0.40 g/BHP-hr. At the same
time, 0.40 g/BHP-hr represents a non-trap option, while
0.25 q/BHP-hr with 0.60 g/BHP-hr for line-haul trucks would
require about 70 percent of urban heavy-duty diesels to be
equipped with traps. This results in substantially higher
costs, relative to the 0.40 g/BHP-hr standard, while the
total emissions reductions are slightly less. The
resultinq cost effectiveness value of about $15,000 per ton
of urban particulate emission reduction makes this an
unattractive option.
Turning to the 0.10 g/BHP-hr standard, EPA believes
that the cost effectiveness of this standard will actually
be somewhat worse than that of the 0.25 g/BHP-hr standard,
but is not at this time able to quantify the difference.
The maximum benefit and least cost applications will have
already been used to meet the 0.25 g/BHP-hr standard (with
averaging) , so that subsequent use of traps on additional
enqines might be somewhat less cost effective. Amonq the
factors arguinq for hiaher cost at the 0.10 q/BHP-hr level
are greater development costs, the need to design to lower
low-mileage-target emission levels, the use of higher
quality components, the probable need for more frequent
trap regeneration, and the increased risks associated with
in-use compliance. Accordingly, although emissions would
be lower under a 0.10 a/BHP-hr standard, it appears that
this standard would be less cost effective than is shown in
Figure 4a.
On the other hand, Fiqure 4b indicates that even when
these extra cost factors are not considered, these two
options differ markedly in cost per engine, reflecting the
increased use of traps at the 0.10 g/BHP-hr level.
Althouqh not shown in Fiqures 3 or 4, EPA has also
examined the impacts of a 0.25 q/BHP-hr standard

-------
7B-10
implemented without the benefit of an averaging proqram.
Such an approach recruires 100 percent trap use, and
therefore incurs essentially the same cost as the 0.10
q/BHP-hr standard does, although the use of some less
efficient systems would be expected to reduce costs
somewhat. The emission reduction benefits, however, remain
virtually unchanqed. The result is a clearly less
efficient regulation, having a cost effectiveness of
approximately $16,000 per ton.
Based upon the available information, overall
conclusions about some of the options are readily
apparent. For 1987, the proposal of a 0.60 q/BHP-hr
standard is the clear choice as the only option which
generates feasible emission reductions in the near term.
For 1990, on the other hand, continuing with 0.60 g/BHP-hr
is not acceptable both because it produces insufficient
emissions benefit and because it fails to satisfy the
requirements of the Act. The beneficial effects of
averaginq, both on technological difficulty and on overall
costs, also support the inclusion of this program for the
long-term standards. This is true regardless of the level
of the final standards, although there would be more
benefit if a trap-based standard were promulgated.
Beyond these decisions, the choices" for long-term
standards become more difficult. Both the non-trap and
trap-based standards will demand significant technological
advances. Indeed, it appears that the technological
difficulty of a 0.40 q/BHP-hr standard usinq non-trap
techniques is at least as great for some engines as that of
a 0.25 q/BHP-hr standard usinq traps. One of the risks of
the 0.40 q/BHP-hr level is that it may in fact turn out to
require traps to meet such a standard. Given this
situation, EPA believes that a trap-based standard is
preferable. Not only would it provide needed emission
reductions at reasonable cost, but it would insure
continued progress in the development of traps for
heavy-duty diesel engines. Trap technology is being
successfully applied to light-duty diesel engines and is
expected to be available to meet 1987 standards for those
vehicles. Failure by EPA to require the development of
traps for heavy-dutv diesels would not seem reasonable
considering the statutory mandate and the significant urban
impact of heavy-duty diesel particulate. Traps are
unlikely to be applied to heavy-duty diesels unless
standards require them. Finally, as has already been
noted, even the 0.25 g/BHP-hr trap standard will succeed
only in holding heavy-duty diesel emissions at current
levels, making no actual reductions beyond that point.

-------
7B-11
Within the ranqe of trap-based standards, there are
four options. These are the 0.25 g/BHP-hr standard, the
0.25 g/BHP-hr standard with a line-haul option of either
0.60 or 0.40 q/BHP-hr, and the 0.10 g/BHP-hr standard. EPA
believes that the maximum stringency 0.10 g/BHP-hr standard
is inappropriate at the present time. Its requirement for
essentially 100 percent trap use greatly increases the risk
that some engines might be unable to meet the standard by
1990. The 0.25 q/BHP-hr standard will allow a more orderly
development and application of technology because some of
the more intractable applications can be avoided. At the
same time it meets the requirements of the statute and
promises to maintain overall heavy-duty diesel particulate
emissions at essentially current levels. As for the
line-haul options, the 0.60 g/BHP-hr standard has little to
recommend it. It requires no further control beyond 1987
for the affected engines, significantly increases overall
fleet emissions, and is the least cost effective of the
options considered here. The 0.40 q/BHP-hr line-haul
standard, on the other hand, involves a trade of some
emission benefit for enhanced overall cost effectiveness.
For this reason, EPA currently favors inclusion of the 0.40
q/BHP-hr line-haul option in the Final Rule. However,
before such an approach could be finalized, requlatory
means of identifyino urban and non-urban applications would
have to be developed, and the unwanted competitive effects
would have to be mitigated. The Aqency thus solicits
comment on this approach and its possible usefulness, and
is specifically interested in suggestions as to how the
workability of this concept might be improved.
While EPA does not believe that a standard of 0.10
g/BHP-hr should be proposed for all engines, that standard
may be appropriate for urban bus engines. The possibility
of implementing a 0.10 q/BHP-hr standard for urban buses
was mentioned earlier as a means to insure maximum control
of these engines. As a group, these engines have less
diversity in both enqine characteristics and in operatinq
patterns than do heavy-duty diesels as a whole. EPA
believes that, given the use of trap technology, these
engines may be able to meet a 0.10 g/BHP-hr standard for
little extra cost beyond a 0.25 g/BHP-hr standard, since
the trap technology currently envisioned appears to provide
the required efficiency to meet the lower standard. At
this level the urban cost effectiveness appears very qood,
primarily because these buses accumulate all of their
mileage in urban areas rather than having a portion of
emissions benefits discounted as rural. The cost
effectiveness of the 0.10 q/BHP-hr standard for urban buses
is estimated to be about $3,300 per ton.

-------
7B-12
Since EPA has not identified significant cost
differences between these two standards, the 0.10 g/BHP-hr
level appears preferable. At the same time, recognizing
that such costs may be identified by public comment on the
proposal, the Agency remains open to eventually setting a
0.25 g/BHP-hr standard for urban buses and is proposing
both options for comment.
EPA believes strongly that trap technology is feasible
and cost effective for most heavy-duty diesels for the 1990
model year, yet it also recognizes the possibility that its
judgment could be changed by information developed during
the public comment period. Therefore, it is also prudent
to indicate what course EPA would choose for non-trap
standards in that eventuality. EPA's plan at this time
would be to implement the 0.40 g/BHP-hr level should a
non-trap standard prove to be necessary.
In summary, EPA believes that some degree of immediate
control should be implemented, and that a standard of 0.60
g/BHP-hr is feasible in the 1987 timeframe. Beyond that,
substantial further reductions are called for and appear to
be feasible, for 1990. It is EPA's further judgment that
trap technology can be successfully applied to most
heavy-duty diesel engines in that timeframe, and that the
added emission reductions possible with traps justify the
increased costs of a 0.25 g/BHP-hr standard, and perhaps a
0.10 g/BHP-hr standard for urban buses. If the problems
associated with implementing a 0.40 g/BHP-hr line-haul
standard in conjunction with the 0.25 g/BHP-hr standard can
be resolved, EPA also favors including this option in the
Final Rule.
2. Heavy-duty Engine NOx
EPA continues to believe, as originally indicated in
the NOx advance notice, that the statutory 75 percent
reduction standard is not feasible for heavy-duty diesel
engines. A NOx standard will have to be developed either
as a temporary revision, based on such things as cost and
fuel economy impacts, or as a permanent change based upon
lack of environmental need. From the earlier discussion of
the environmental need for NOx control, there may or may
not be a need for controlling future motor vehicle NOx
emi ssions.
In developing NOx control options, EPA has decided to
use common standards which are appropriate for both
gasoline-fueled and diesel heavy-duty engines. While it

-------
7B-13
would be possible to have separate standards for
gasoline-fueled and diesel engines, there would be little
environmental benefit from doing so in this case. As was
noted in the discussion of environmental need, NOx
emissions from heavy-duty gasoline engines are expected to
decline in future years even without additional control.
Therefore, so long as a reasonable degree of control is
established for gasoline engines, they will not pose a
significant concern relative to future NOx emissions.
In a manner similar to the heavy-duty diesel
particulate options, EPA has examined NOx control options
in two timeframes, near-term (1987) standards and
longer-term (1990) standards. For NOx there is an added
leadtime dimension because of the statutory leadtime
provisions of Section 202(a)(3)(B) of the Clean Air Act.
The choice of 1987 does not satisfy the four year leadtime
provision of that section, but EPA has explicitly
considered the amount of available leadtime in developing
1987 options and, as indicated in the earlier discussion of
statutory provisions, believes the now conflicting
requirements of the Act necessitate such an approach. In
addition, manufacturers have been aware since the January
1981 advance notice that EPA planned to require substantial
NOx emission reductions. This fact has in essence given
them three additional years to begin to prepare for new NOx
standards. While this fact does not in any way excuse EPA
from the need to allow adequate leadtime for a new standard
after its promulgation, it has allowed progress to be made
in heavy-duty engine NOx control and reduced the required
leadtime for an early standard.
For 198 7:
No new standard. Under this option, the 1986 NOx
standard would remain unchanged. This approach would have
to be implemented under Section 202(a)(3)(E) after the
Administrator has studied the impact of motor vehicle NOx
emissions on public health and welfare. However, current
engines are actually operating below the level of the
standard (generally in the 4 to 8 g/BHP-hr range), so that
future emission levels could increase even more than
projected if manufacturers were to take advantage of
potential increases in trade for improved fuel economy or
easier compliance with particulate emission standards.
6.0 g/BHP-hr. EPA's evaluation of current emission
levels and the potential for near-term reductions from
heavy-duty diesel engines indicates that a standard of 6.0
g/BHP-hr is feasible for 1987. This standard represents a
modest decrease from current engine emissions (about 15

-------
7B-14
percent below current levels) which should be attainable in
the short term with techniques compatible with the proposed
0.60 g/BHP-hr particulate standard. Techniques expected to
be used include improvements to fuel injection systems,
injection timing retard, increased use of turbocharging and
aftercooling, minor engine modifications and improvements
to engine efficiencies.
This option is considerably less stringent than the
4.0 g/BHP-hr level envisioned in the advance notice, in
spite of the time which has elapsed since then, and in
spite of the fact that manufacturers face a standard for
their California models of 5.1 g/BHP-hr NOx on the EPA
transient test (or 4.5 g/BHP-hr HC plus NOx on the 13-mode
steady state test) for 1984. However, a Federal standard
below 6.0 q/BHP-hr would not be feasible for 1987,
principally because of leadtime constraints. The 6.0
g/BHP-hr level itself will present a challenging task in
that timeframe. Standards would need to be delayed several
years (see the discussion of 1990 standards below) to move
significantly below that level.
Gasoline-fueled heavy-duty .engines will be able to
readily meet a 6.0 g/BHP-hr standard in 1987. About one
third of current engines are already capable of meeting
such a standard, and those remaininq have well established
techniaues at hand for compliance. Recalibrations of
air/fuel ratios, enhanced exhaust qas recirculation (which
will already be in use on most heavy-duty gasoline
engines), and ignition timing retard will be able to
satisfy a 6.0 g/BHP-hr standard with no significant impacts
on performance or fuel economv.
EPA views the 6.0 g/BHP-hr 1987 option as essentially
a stop-gap measure to obtain a feasible level of control in
the short term. The Agency's main concern is directed at
the longer-term 1990 standard. Although the proposal as
structured provides a three-year period for the interim
standard, the Agency does not view it as essential to
maintain such a period should it, for example, turn out
that the 1987 proposed level has to be delayed. In such an
event, the Agency might choose to abandon the interim
standard or to implement it for a one or two year period,
depending on the situation. Any decision to change or
delay the 1990 standard will be an independent judgment
based only on facts pertinent to the feasibility of that
standard for that year. (A decision about 1990 would, of
course, consider as relevant such facts as the impact of a
delayed interim standard on the availability of
manufacturers' resources for meeting a long-term
standard.) Commenters on the proposal should treat the

-------
7B-15
near-term and long-term standards as separate and distinct
issues, and not assume that a delay in the 1987 standard
would mean an automatic delay in the 1990 standard.
For 1990:
No control beyond 6.0 g/BHP-hr. If no additional
control Ts adopted for 1990, then Reavy-duty engine NOx
emissions will continue to increase much as projected under
current standards. A 6.0 g/BHP-hr standard would result in
heavy-duty engine NOx emissions in 1990 being over 50
percent greater than in 1980. Such a standard also would
fail to meet the statutory requirement for a temporarily
revised standard based on technological, cost, or fuel
economy considerations. The Administrator would have to
issue the standard under Section 202(a)(3)(E) after
studying the effects of motor vehicle NOx emissions on the
public health and welfare.
4.0 g/BHP-hr. In order to deal successfully with the
problem 57 future growth in NOx emissions, a substantial
reduction beyond the 6.0*g/BHP-hr level is necessary. EPA
believes that further reductions are available, and that
projected improvements in technology argue for lower future
standards. Unfortunately, heavy-duty diesel engines, which
form the bulk of the problem which must be addressed, are
limited in their capability for reductions in NOx. Based
upon current knowledge of diesel NOx control technology,
EPA believes that a 4.0 g/BHP-hr standard represents the
approximate limit of available control without unacceptable
impacts on fuel economy, engine durability or engine-out
particulate levels. Primary sources of NOx control lie in
the areas of electronically programmed exhaust gas
recirculation, electronic management of fuel injection,
charge air cooling, and engine modifications to improve
efficiency and enhance combustion. These will be
implemented in concert with changes needed for particulate
control. As noted in the discussion of the 0.40 g/BHP-hr
particulate option, much of this technology is already
targeted for introduction on heavy-duty diesel engines in
the late 1980s for reasons apart from emissions control
(fuel economy improvements). The full cost, therefore,
should not be attributed to emissions control
requirements. There is also uncertainty as to the final
complement of technologies which will be used to meet a 4.0
g/BHP-hr standard, introducing further variability into
actual costs. Discounted lifetime cost would be about $300
to $1,000 per vehicle, depending on fuel economy effect.
Total discounted NOx emissions in 1995 would only be about
7 percent greater than 1980 levels under this option.

-------
7B-16
In establishing the 4.0 g/BHP-hr option, full
consideration has been given to the interactions which will
occur between NOx and particulate standards for heavy-duty
diesel engines. The most difficult combination of
standards would pair the 4.0 g/BHP-hr NOx standard with the
0.40 g/BHP-hr non-trap particulate standard. With this
combination, great care would have to be exercised in
balancing tradeoffs between NOx and particulate emissions
to bring engine-out levels of both pollutants below the
standards. The difficulty of the 4.0 g/BHP-hr NOx standard
would be eased somewhat when paired with the trap-based
0.25 g/BHP-hr particulate standard. The use of traps for
particulate control will allow some increase of engine-out
particulate in favor of lower NOx levels. In any event,
EPA recognizes that the combination of a 4.0 g/BHP-hr NOx
standard with either a 0.40 g/BHP-hr or a 0.25 g/BHP-hr
particulate standard will represent a very difficult
challenge for conventional diesel engine technology.
Heavy-duty gasoline engines will also be able to meet
a 4.0 g/BHP-hr option for 1990. Enhanced use of exhaust
gas recirculation will be the primary means of compliance,
along with engine modifications to improve tolerance for
exhaust gas recirculation without adverse fuel economy or
performance effects. These modifications might include
improvements to combustion chamber efficiency, fast burn
combustion techniques, some use of electronic controls, and
general engine efficiency improvements. Similar
non-catalytic reductions of NOx have been attained for
light-duty vehicles, and considering the amount of leadtime
remaining before 1990, EPA believes that the necessary
improvements will be attainable for gasoline-fueled
heavy-duty engines.
The cost-effectiveness of the 4.0 g/BHP-hr option is
under $400 per ton for both gasoline-fueled and diesel
engines. Cost effectiveness turns out not to have been a
significant discriminator for any of the NOx control
options, for either heavy-duty engines or light-duty
trucks, because it has been relatively low for all of the
options considered. Therefore, cost effectiveness will not
be discussed further in the context of NOx control options.
In summary, the 4.0 g/BHP-hr standard in 1990 combined
with a 6.0 g/BHP-hr standard in 1987 represents what EPA
believes to be the proper level of control for heavy-duty
engine NOx emissions, given the technological limitations
of current diesel engines.

-------
Chapter 8
Benefit-Cost Analysis
I. Introduction
This chapter provides a detailed analysis of the benefits
of controlling heavy-duty diesel vehicle (HDDV) particulate
emissions and compares these to the societal costs of emissions
control. This analysis augments the cost-effectiveness analysis
of the previous chapter in two ways. First, while cost-ef fe
-------
8-2
from other sections of this report, associated EPA reports, and
contractor reports. The benefits data and analysis are derived
from a report by Mathtech Inc. This analysis is highly detailed,
having utilized metropolitan counties as the basic geographic
unit.
II. Benefits and Costs of HDV Particulate Controls
Regulatory options for control of mobile source particles are
being considered for heavy-duty vehicles. Particulate control for
light-duty vehicles has recently been postponed to the 1987 model
year at the .2 g/rai level for automobile and .26 g/mi for light-
duty trucks. In the first subsection a detailed description of
the benefits of two HDDV standards — at .25 g/BHP hr. and at .10
g/BHP-hr. — is presented. The second subsection describes the
computation of the control costs in a format comparable to the
benefits. The last subsection presents a B-C analysis of some of
the regulatory options presented elsewhere in this RIA. These
include: (1) .6 g/BHP-hr. for 1987 and .4 for 1990; (2) .6 for
1987 and .25 for 1990; (3) .6 for 1987 and .25 with averaging
for 1990; (4) .6 for 1987 and .25 with averaging for 1990 except
line-haul trucks which remain at .6; (5) .6 for 1987 and .25 with
averaging for 1990 except line-haul trucks which are to meet .4;
and (6) .6 for 1987 and .10 for 1990.
A. Particulate Benefit Analysis for HDDVs
1. Overview of Benefits Estimation Methodology
Estimation of the benefits of reducing diesel particles is
based directly (or by extrapolation) on the geographically dis-
aggregated estimates presented in a recent Mathtech report [2]. The
Mathtech analysis covers both health and welfare effects. The
health benefits are subdivided into noncancer mortality, chronic
and acute noncancer morbidity, and cancer mortality categories.
Welfare effects include visibility.and household soiling categories.
Though diesel odor is potentially an important aesthetic effect,
it is not valued in the Mathtech report.* The scope of coverage
of effects is broad, but not entirely comprehensive, as will be
illustrated in the following two sections. No fundamentally new
concentration-response or damage effects research is presented in
the Mathtech study. Ideally, benefit estimation would be accomp-
lished using data developed specifically for diesel emissions
*The odor problem is relevant to particulate standards, since some
particulate control technologies contribute to odor reduction. On-
going work to value reduced exposure to diesel odor is expected to
be completed in mid 1984.

-------
8-3
and associated undesirable effects. This is unfortunately pre-
cluded by the absence of scientific data of this sort. Thus,
Mathtech's contribution is one of critically reviewing the relevant
particulate related studies and developing a range of estimates
that best represent prevailing knowledge.
The benefit estimation procedure follows the four steps out-
lined in Figure 8-1. In the first step, the change in air quality
is projected annually over the 1987-2000 period of analysis for
each metropolitan county with a population over 100,000 people.
This computation measures the improvement achieved after implemen-
tation of a selected diesel particulate emissions standard, and
specifically accounts for the number of operating diesel vehicles,
miles traveled., and emissions factors. The second step involves
estimation of the various health and welfare improvements that
are predicted given a lower ambient concentration of diesel
particles. This step employs the research findings extracted
from published studies. These generally take the form of concen-
tration-reponse functions from which county estimates of effects
are generated for each year.
The third step imputes a monetary value to the various changes
in health and welfare measures. These changes are valued in a
variety of ways. Reduced mortality is valued by first deter-
mining the reduction in annual mortality associated with lower
emissions. The value individuals place on this reduced risk
(determined from a separate group of studies) is then used to
value the benefit to the population at risk. The flavor of
willingness-to-pay to avoid premature mortality risk is intrinsic
to this approach. In contrast, reduced instances of morbidity are
valued by reduced costs, either in medical expenditures or in work
loss days, the latter based on average regional wage rates. The
cost to the individual in terms of pain and suffering is omitted,
so that a downward bias results from valuing morbidity in
this way. Welfare effects are valued, in the case of soiling and
materials effects, on the basis of changes in consumer surplus, which
is the common practical method of measuring changes in willingness-
to-pay; and in the case of visibility, on the basis of willingness-
to-pay data gathered from surveys designed to simulate market
bidding processes.
The final step shown in Figure 8-1 involves time aggregation
(using the OMB 10% discount rate guideline) and spatial aggregation.
The period of analysis covers 1987-2000. The heavy-duty particulate
standard is assumed to take effect in 1988. The choice of 1987 as
the initial year was dictated by a companion Mathtech study of
light-duty diesel particles. The final year was selected because
it is as about as far into the future as one could practically
predict the various growth factors. In any case, after 14 years,
the process of discounting reduces both costs and benefits to only

-------
8-4
Figure 8-1
Benefits Estimation Procedure
roughly a third of their original magnitude in the last year.
Thus, truncation at this point does not greatly affect the outcome
of the benefit-cost comparisons.
An extrapolation procedure was required to develop national
benefit estimates. The basic analysis covers only 41 metropolitan
counties (roughly 25% of total metropolitan population). These ,
are the counties which have population exposure monitors (in NAM'S
network) that measure CO concentrations. CO concentration is used
in a proportional formula to ascertain particulate concentration.
The national estimate of benefits (covering all SMSAs with a popu-
lation greater than 100,000 people) is derived via an estimated
relationship between particulate emissions and vehicle miles
traveled, which is necessary for counties where the monitored CO
proxy for particulates is unavailable.
Steps two and three will be discussed in greater detail in the
following two sections that focus on health and welfare effects
respectively. The remainder of this section concentrates on the
air quality modeling and aggregation procedures contained in steps
one and four.

-------
8-5
Figure 8-2 illustrates graphically the connection between air
quality and the measure of benefits over time. The curve AA' in
Figure 8-2 depicts the uncontrolled (baseline) diesel particulate
emissions, which increase over time given expected increases in
diesel vehicle sales. The curve BB' represents the reduction in
diesel emissions associated with a selected diesel particulate
standard. Benefits are computed annually on the basis of the
difference in concentrations between these two curves. Projecting
diesel particle loadings is not straightforward, however, since no
ambient monitoring network exists for fine particulate concentra-
tions.
Mathtech utilizes an approach similar to that in the National
Research Council study [3]. That approach takes advantage of similar
relationships between CO emissions and CO ambient concentrations
and DP emissions and DP ambient concentrations. The comparability
between emissions and ambient concentrations for CO and DP is due
to the fact both are predominately mobile source emissions with
similar dispersion patterns. Assuming proportionality, ambient DP
concentrations can then be projected from CO and DP emissions
factors and monitored CO ambient concentrations. The emissions
factors (and changes in emissions factors) account for the number
and age of vehicle by type, deterioration in emissions rate with
age, and VMT by vehicle type and age. Projections of new vehicle
sales and a scrappage model are thus an integral part of the air
quality component of the benefit model.
Figure 8-2
Time Stream of Benefits and Air Quality Change
Antoient
Concentration
A'
A
B
1987
2000
Time

-------
8-6
In the baseline case, some controls have already been
engineered into existing vehicles. By 1986 cars are projected to
be emitting at about .32 g/mi (assuming they meet a NOx standard
of 1.0 g/mi and the 5.7 liter GM engine line is dropped from pro-
duction); light-duty trucks at .56 g/mi (assuming NOx standard of
1.2 g/mi); and all heavy-duty trucks at .7 g/BHP-hr., until 1988
when the latter is assumed to fall to .6 g/BHP-hr.
Given a projected air quality change in a county, incremental
benefits can be computed for control options for any study whose
results are presented in concentration-reponse form. Choosing
the set of studies that best represents real benefits is a difficult
task. The potential for undercounting or double counting in the
process of aggregation is substantial. Further, studies of similar
effects areas may differ in terms of credibility and accuracy of
results as well as the breadth of coverage. Partly for these
reasons, and partly due to institutional constraints, Mathtech
presented four aggregation alternatives.
Alternative A is designed to be consistent with the Clean Air
Scientific Advisory Committee [CASAC] review of the health litera-
ture [4]. This review was very conservative in distinguishing
between highly reliable and less reliable quantitative evidence.
Welfare effects are excluded, since CASAC has not determined that
studies of welfare effects provide highly reliable quantitative
evidence. The incomplete coverage of both health and welfare
benefits under Alternative A implies this aggregation scheme almost
certainly substantially underestimates true benefits.
Alternative B adds, conservatively, studies CASAC judged less
reliable but still with scientific merit. These provide the
only available evidence of the potential magnitude of benefits
in some categories (e.g., noncancer mortality). Alternative B
includes minimum estimates of soiling, but assumes visibility
benefits are zero. The predominant conservative bias probably
means that Procedure B also underestimates true benefits.
Alternative C includes estimates for all categories with the
exception of soiling and materials damage in the manufacturing
sector which is considered highly speculative. Alternative C pro-
vides the best point estimate of true benefits.
Alternative D utilizes studies that give higher estimates than
C, filling some gaps in C, but also probably double-counting in
some instances. While benefits could be as high as indicated by
Alternative D, the prevailing bias is in an upward direction.
The four aggregation alternatives provide quite disparate
results. For example, for the .25 g/BHP-hr. standard, the benefits
estimated by Procedure A are $209 million (present value in 1987
for 1987-2000 interval measured in 1983 dollars); Procedure B, $752
million; Alternative C, $6638 million; and Alternative D, $12,017
million. One should not conclude from these that the range of

-------
8-7
uncertainty is almost two orders of magnitude, since the range is
a function of study selection. Rather, the reader should conclude
that selecting the most credible study set is important. It is
encouraging that the estimate generated by Alternative C is rein-
forced by ancillary evidence. Property value and wage-based
studies, that directly relate ambient particulate concentrations
to dollar value (effectively by-passing the usual damage function
chain), are consistent with the aggregated health and welfare
estimates from Alternative C. Cross-checking the individual
health and welfare benefit estimates is not possible, however',
since the property and wage studies cannot be disaggregated. The
principle underlying property and wage studies is rather simple.
Differences in air quality are expected to be capitalized into
property values, and may show up in wage differences geographically
as well. One would expect, however, that these methods underesti-
mate actual benefits, since people will not adjust for environmental
effects they do not know about. Many of the health effects of
particles are not common knowledge. Since we would expect Alterna-
tive C to be higher by some portion of the health component, the
relationship between Alternative C and the property value and wage
studies shown in Table 8-1 tends to reinforce the credibility of
Alternative C. The point estimates in the remainder of this chapter
are based on Alternative C.
2. Estimation of Health Benefits
Though at times controversial, placing an economic value on
reducing the risks of morbidity and mortality provides a way to
lend some perspective to the many different health impacts attributed
to air pollution. For example, the major emphasis on health risks
associated with diesel particulate emissions has been on cancer
[5]. Yet, because the cancer risk is less proximate in time (given
a fairly long latency period) relative to noncancer health effects,
and because effects do not pose that large a risk unless exposures
are over a lifetime, noncancer risks turn out to be far more important
in monetary terms.
As alluded to earlier, valuing changes in mortality is
accomplished by computing the change in the probability of
environmental health symptoms, which is then monetized using
separate wage premium and other data that indicate willingness-to-
pay to avoid increased occupational or other kinds of risk. To
noneconomists, the attempt to quantify the value of life often
seems presumptuous. Indeed, for most people no finite payment
would be acceptable in exchange for loss of life. But an
individual is generally not confronted with certain death as a
result of pollution. Rather, he may face the possibility that
there is one chance in a million, for example, that an environ-
mentally induced cancer will occur. What must be valued, then,
isn't the metaphysical question of the value of a particular

-------
8-8
Table 8-1
Benefit Estimate Comparisons Among Valuation Techniques
(present value in 1987 in millions of 1983 dollars
	for 1987-2000 period)	
Estimate Alternative C Property Studies Wage Studies
min.
1243.
1730
2877.
point
6635.
3461.
5812.
max.
21,105.
5768.
6619.
person's reduced lifespan (which is the sort of issue courts deal
with routinely), but instead the more tractable question of what
is the value of reducing environmental risk.
One way to do this is observe what premium individuals must
be paid to accept employment that entails greater risks of fatal
accidents. Suppose, for example, the estimated wage premium is
$500/yr. to compensate for an increased annual job-related risk of
one death in one thousand. Then, if the risk of death to 1,000
people is reduced by .001, one life is saved on the average, and
since each of 1,000 people is willing to accept $500 to take this
risk, one statistical life is valued at $500,000. This amount is
not a measure of compensation appropriate for a particular death,
since no one knows which one of the 1,000 people will die from the
hazard. In 1980 dollars, Mathtech's review of the risk literature
indicates that a one unit reduction in annual mortality risk would
have a statistical value of between $360,000 and $2,800,000 with a
point estimate of $1,580,000.
Morbidity valuation accounts for three factors: the loss
in work time due to illness; the reduction in activity levels
on nonwork days; and direct medical expenditures. The first
is valued using the average daily wage, while the second, somewhat
arbitrarily, is valued at one-half the average wage. The specific
wage rates vary geographically by county.
Health effects are divided into seven subcategories:
subcategories (1) and (2) cover noncancer mortality effects due
to short- and long-term exposure; subcategories (3) and (4) cover
noncancer acute morbidity effects due to short- and long-term
exposure; subcategories (5) and (6) cover noncancer chronic
morbidity effects due to short- and long-term exposure; and sub-
category (7) is for cancer mortality and assoicated morbidity.
Research in some of these areas is very scanty, especially with
respect to particulate effects, let alone work that specifically
addresses diesel particulate effects. In fact, with the excep-
tion of the cancer area, the studies are based on measures of total
suspended particles (TSP) rather then diesel particles. One might

-------
8-9
suspect that the fine diesel particles are potentially more
damaging, given inhalation deeper into the lungs, but in the
absence of direct evidence, the Mathtech study conservatively
assumes TSP and DP exposures result in equivalent health effects.
The Mazumdar, Schimmel, and Higgins study [6] is selected
(for aggregation Alternative C) as the best representative of
short-term noncancer mortality effects. The long-term exposure
mortality studies (e.g. Lave and Seskin [7]) are judged as too
speculative to be included in the estimate used here. The
Mazumdar et al. work is a longitudinal study of daily mortality
during 14 LonHon winters. Using both exposure to British Smoke
and SO2, short-term exposure dose-mortality functions are estimated
using excess daily winter mortality as the dependent variable.
Both lagged and nonlinear models are tested. The main limitation
to the work is that both pollutants are not present in one regres-
sion estimate, so that omitted variable bias is possible. While
this and other difficulties are apparent in Mazumdar et al, their
results are credible, and there are no alternative estimates of
short-term mortality effects.
In the short-term chronic noncancer mortality area, the Ferris
studies with various coauthors [8] constitute the best choice.
These examine chronic respiratory disease at one location over
three time periods during which TSP declined. From the three
studies, it can be concluded that respiratory health effects are
observable when TSP is in the range of 180 mg/m^ down to 130 mg/m3,
but that either no beneficial effect occurs as concentrations are
further lowered to 80 ug/m3 or the effect is obscured, perhaps by
changes in SO2 concentrations. The Ferris studies are valuable
because disease incidence is classfied by age and sex, and cigarette
smoking is taken into account. Though Ferris ej: a_l. did not derive
concentration-response functions, health effects can be tied to
appropriate aerometry data.
The Ostro [9] study is selected for the long-term acute
morbidity category. It uses the 1976 National Center for Health
Statistics Health Interview Survey to estimate the relationship
between long-term exposure to TSP and sulfates and acute illness,
measured in work-loss and reduced-activity days. The data covers
both worker and nonworker respondents in 8 4 SMSAs. The author
attempts to statistically control for a broad array of socio-
demographic differences in respondents in addition to pollution
and weather related variables in his concentration-response
functions. Both linear and logit models are tested. The Ostro
study covers a broad range of respiratory effects and is very
thorough in its statistical analysis, providing a sound basis
for the concentration-response functions in the acute morbidity
category.
The cancer mortality and associated morbidity estimates are
based on the results from several studies, reviewed in the OMS
Diesel Particulate Study [10], and a study by Harris [11]. These

-------
8-10
studies examine the incidence of respiratory cancer due to
occupational exposure to coke oven emissions and roofing tar.
These results are coupled to laboratory studies on the potency
of diesel particles' relative to these substances to predict the
incidence of respiratory cancer from diesel particles. It is
important to emphasize that the underlying studies do not relate
occupational cancer to diesel particles, and that use of surro-
gates introduces considerable uncertainty. There is simply no
alternative in the absence of a direct and credible epidemiolog-
ical study.
The valuation of cancer risks parallels the procedures used
for noncancer health risks. While this ensures consistency and
comparabilityr the results can differ substantially relative to
those constructed on a steady-state basis. Cancers that are
predicted t^occur-arfter the year 2000 are included in the
Mathtech analysis. However, three independent factors reduce
the significance and value associated with the potential loss of
life. First, some portion of the people exposed to carcinogenic
diesel substances die from other causes before the onset of
respiratory cancer symptoms. Second, the reduced exposure to
potential diesel carcinogins is assumed to occur during the four-
teen year analysis period, which covers only part of the usual
lifetime risk. Finally, the long disease latency period means
actual incidence is 20 to 40 years beyond the initial year of
analysis, and everything is discounted back to that year. At a
10% rate of discount, even large numbers are substantially reduced
over this long a period. Since the cancer benefit category turns
out to be quite small, sensitivity analysis to these procedures
was undertaken. The results of these are reported in the next
section.
In sum, while the quality of information from the studies
included in aggregation Alternative C is mixed, there will be some
natural offsetting of errors. In light of the fact that there are
gaps in the coverage (ensuring the probability of double-counting
is minimal), the overall health benefit estimate is probably on
the conservative side.
3. Estimation of Welfare Benefits
Recent economic research suggests that the general welfare
category of benefits (i.e. nonhealth benefits) may be an important
part of the total benefits derived from air quality improvement
programs (See Freeman [12]). Diesel particles degrade visibility
and contribute to soiling and other material damages. Though
these effects are sometimes accorded secondary importance, the
dollar a consumer is willing-to-spend to mitigate these impacts
is no different than a dollar he might be willing-to-pay to reduce
risks to health.

-------
8-11
Reduction in visual range is associated with the concentration
of fine particles, of which diesel particles are a part. Thus,
control of diesel particulate emissions may generate benefits via
improved visibility. Consumers experience the benefit in conjunc-
tion with outdoor activities or simply because they value the preser-
vation of scenic views. Economic studies of visibility rely on
contingent valuation techniques which make use of specially designed
surveys.
The valuation of visibility improvement is based on the
results of two studies: Tolley and Randall [13] and Brookshire,
d'Arge, Schulze, and Thayer [14]. The Brookshire j5t al. study,
which was conducted in the Los Angeles area, is used to develop
benefit estimates for urban counties in the western United States.
The Tolley and Randall study, conducted in Chicago, is used to
value urban counties in the eastern United States. These studies
do not project changes in visual range — Mathtech does this
utilizing Koschmieder's Law.
While over the past decade economists have developed and
tested many types of contingent valuation instruments to value
aesthetic changes, the technique has not yet fully gained pro-
fessional acceptance. Different attempts to value similar goods
have yielded substantially different values. Two points are worth
emphasizing. One, when available, the use of market data will
provide a more accurate valuation. Two, while contingent methods
are less accurate, there often are no alternatives for valuing
aesthetic changes with public good characteristics, and a less
accurate valuation is preferable to leaving the resource unvalued
(which is sometimes incorrectly perceived as equivalent to a
valuation of zero).
Briefly, the contingent valuation approach relies on a highly
structured questionnnaire that places respondents in a hypothetical
market situation. The questions are designed to elicit responses
that either directly or indirectly reveal their preferences about
changes in environmental quality--in this case in visibility as
measured by visual range. This information is conveyed through the
use of a set of photographs that depict a familiar scene under
different visibility conditions. In one variant of the procedure,
a specific method of paying for improved environmental quality
is suggested (e.g. an increase in utility bills or taxes). This
is designed to convince respondents that their dollar "votes"
involve real resources and to discourage "free-rider" responses.
Bids for improved environmental quality are solicited directly,
though sometimes an iterative procedure is employed. In a second
variant of the procedure, respondents are asked to rank outcomes —
consisting of a hypothetical payment and a corresponding level of
environmental quality — from most preferred to least preferred.
Two factors of this methodology have led to concerns about
the results. First, the public good nature of environmental
quality may lead respondents to purposefully misstate their true
preferences. Repeated testing for the occurrence of strategic

-------
8-12
response has not found this to be a serious concern. Second,
there are many inherent potential biases in the design and
administration o£ surveys. The wording of questions can influence
responses, sometimes significantly. These built in survey biases
continue to be a concern, and undoubtedly are the ultimate limit
on the accuracy of the contingent technique.
The Brookshire et al. study used data from six paired areas
in the southern California region. Good, fair, and poor visi-
bility conditions (2, 12, or 28 miles of visual range) were used
in the choice of the paired census tracts; no pair had identical
visibility conditions. Health effects information was distributed
to some of the respondents. Using different question formats,
Brookshire et al_. were able to identify the non-visibility com-
ponent of the total bid for visibility improvement. On average,
the aesthetic component accounted for about 35 percent of the
total bid. An analysis of the area-by-area bids indicates that
the pure visibility value is approximately $6.35 (1980 dollars)
per household per year for each mile improvement in visual range.
Using six separate contingent value question formats, Tolley
and Randall valued changes in visual range over three distances
in the city of Chicago: 9 to 4 miles, 9 to 18 miles, and 9 to 30
miles. Based on the latter two visibility scenarios, their
analysis indicates that the value of a one mile change in
visibility averages $18.50 (1981 dollars) annually per
household. This bid may include non-visibility components.
If the Brookshire et al. 35 percent factor is applied, the
mean bid would be $6.60, which is remarkably similar to the
Brookshire result in the West.
Soiling damages are estimated using a study by Watson and
Jaksch [15]. It covers the household, but not the commercial or
industrial sectors, where there may be material damages associated
with diesel particle deposition. Watson and Jaksch utilize a
simple supply-demand framework, where the supply of cleanliness
shifts downward (reflecting a decrease in the cost of obtaining a
given level of cleanliness) as emissions are reduced. This
generates a gain in consumer surplus, which constitutes their
estimate of soiling benefits.
There are a number of important assumptions embedded in their
calculations. First, using the data from the Booz, Allen, and
Hamilton [16] survey of cleaning expenditures, frequencies, and
time durations for multiple household cleaning activities to
calibrate the demand curve of their model, Watson and Jaksch
reach the conclusion that actual cleaning expenditures do not
change as pollution changes. That is, they assume a unit elastic
demand curve for cleanliness. If, as seems at least as equally
plausible, the demand for cleanliness is inelastic, the surplus
estimates would be lower. Likewise, the functional form assump-
tions for the supply curve and the way it shifts could influence

-------
8-13
the benefits estimates/ but the direction of bias cannot be
determined a priori. Second, the underlying soiling survey data
used TSP units. It has been asserted, by Wallin [17] and Sawyer
[18], that diesel particles, because they are stickier, more oily,
and darker, have 4 to 5 times the soiling potential of TSP.
Mathtech treats this claim cautiously. It is quite unlikely that
people spend four to five times as much for cleanup as CARB recently
assumed in a soiling estimate for California. Mathtech does however,
double their upper-bound soiling estimate to account for this effect.
In sum, there are offsetting biases that influence the Watson-Jasch
estimate of soiling. Since a different study provided similar,
though somewhat higher values, the Watson-Jaskch benefit estimate
is the conservative choice.
4. Summary of Results and Sensitivity Analysis
The basic results are summarized in three tables. Table 8-2
shows the lower-bound, point, and upper-bound benefit estimates
disaggregated by category for the .25 g/BHP-hr. standard. Table
8-4 illustrates the geographic distribution of these benefits.
Table 8-5 provides a category-by-category benefit breakdown for
two heavy-duty truck standards: .25 g/BHP-hr. and .10 g/BHP-hr.
Table 8-2 is designed to provide a sense of the uncertainty
associated with the point benefit estimates. This uncertainty
arises from numerous sources. Predicting the magnitude of air
quality improvements involves use of CO as a proxy for DP and
projections of the size and composition of the vehicle fleet
until the year 2000. In addition to the uncertainties asso-
ciated with these factors, the potential air quality improvement
in counties adjacent to those included in the analysis is not
counted. The evidence from the health and welfare studies that
form the basis for valuation is at times contradictory. Even
within studies, considerable variation is often apparent as a
result of using different functional forms for concentration-
response relationships. Sampling variation is sometimes large.
Finally, the assumed equivalence of TSP and DP effects could be
in error. The actual value assigned to response changes involves
additional uncertainty. Overall, there is the potential to under
or over count, given the fuzziness of boundaries between the health
categories. Some benefit categories, for example diesel odor, are
not included, and others are not fully valued, which could be the
case for morbidity, where health cost and wage data are used in
lieu of actual willingness-to-pay data.
Many of these uncertainties can be collapsed into two observed
sources of error. There is variation in the results in
different studies covering similar effects, and there is variation
within a given study that results from sampling error and the use
of sensitivity analysis to crucial assumptions. Mathtech provides
information on both, by their use of different aggregation schemes
and by provision of bounded estimates for the individual studies.

-------
8-14
As indicated earlierf it would be a mistake to judge uncertainty
on either o£ these alone, since the sources of uncertainty are
associated with both. The uppers and lower-bound estimates
attempt to account for both, but it is important to point out
that without statistical information on the distribution of
errors, no overall formal accounting of uncertainty is possible.
Thus, the bounded estimates in Table 8-2 are subjective.* The
following guidelines are employed. If the results from studies
selected as credible in a given category are substantially different
from the best estimate, the estimates from these other studies are
incorporated into the lower- and upper-bound estimates. If, on the
other hand, the studies are similar, the error bounds are then
derived from the study used in Alternative C, taking account of
statistical variation and variation in critical assumptions.
Since these bounds are subjective, they cannot be interpreted as
confidence intervals. They only indicate the full range of
uncertainty that is present.
Table 8-3 summarizes the sources for the estimates in Table
8-2. In the noncancer mortality area, thfe lower-bound estimate is
based on Mazumdar et al. utilizing assumptions on functional form,
the value of risk, and confidence intervals for estimated parameters
that give low estimates. The upper-bound estimate is based on the
Lave and Seskin [7] study. It is used because it produced estimates
quite a bit higher, though in general less reliable, than those
from Mazumdar et al. Since the results of studies in the noncancer
acute morbidity area differ substantially, studies other than
Ostro are used for the upper- and lower-bounds. The lower-bound
estimate is based on a study by Samet, Spiezer, Bishop, Spengler,
and Ferris 119] and the upper-bound by Crocker et a^. [20]. There
is likewise considerable range in the benefit estimates derived
from studies in the noncancer chronic morbidity area. With a
point estimate of $21 million, the lower-bound estimate cannot
realistically be distinguished from zero. The upper-bound estimate
is based on the chronic effects research of Crocker et al. In the
cancer area, the lower-bound estimate is zero, based on the studies
EPA reviewed [10], while the upper-bound estimate is based on
Harris [11] .
In the visibility welfare area, the underlying contingent
market studies are very similar. However, standard errors of the
estimated average bids are not reported. Thus, the upper-bound
estimate is based on the SRI [21] study, which has the highest
bid for a given change in visibility. The lower-bound estimate
would be considerally larger than the $100 million shown in Table
*Mathtech did not follow such a procedure. The upper- and lower-
bound estimates reported in Table 8-2 are based on extrapolation
and reorganization of results reported in the Mathtech analysis.

-------
8-15
Table 8-2
Benefits of Heavy-Diesel Particulate
Standard of .25 g/BHP-hr.*
Benefit Cateoqry	Lower-Bound
Health
Noncancer Mortality	6.
Noncancer Acute Morbidity	351.
Noncancer Chronic Morbidity	0.
Cancer Morbidity & Mortality	0.
Health Subtotal	357.
Wblfare
Visibility
Household Soiling
Welfare Subtotal
100.
786.
Point
187.
3035.
21.
	1;_
3244.
886.
1987.
1404.
3391.
Upper-Bound
1802.
8418.
2250.
18.
12,488.
3001.
5616.
8617.
Tbtal Benefits
1243.
6635.
21,105.
* Present value over 1987-2000 interval in 1987 measured in millions of 1983 dollars.
Annialized values for 14 year interval at 10% discount rate can be obtained by dividing
entries by 7.367.

-------
8-16
Table 8-3
Source Summary of Lower-Bound
Upper-Bound and Point Estimates
Benefit Cateoqry
Realth
Noncancer Mortality
—Nbncancer_Acute Morbidity
Noncancer Chronic Morbidity
Lower-Bound
Mazundar
Samet
Ferris
Cancer Morbidity and Mortality EPA*
Welfare
Visibility
Household Soiling
Tblley & Randall,
Brookshire et al.
W&taon & Jakscb
Point
Mazundar
Ostro
Ferris
EPA*
Tbllery & Randall
Brookshire et al.
Watson & Jaksch
Upper-Bound
Lave & Seslcin
Crocker
Crocker
Harris
SHE
Watson & Jaks
*Enoompasses several studies reviewed in EPA Diesel Particulate Study [10]

-------
8-17
8-2, if it were based on the low estimates from the contingent
market literature. However, as indicated in chapter 5 of this RIA,
the increases in visibility attributed to control of diesel
particles range from 1.2 to 7.0 percent (See Table 5-4). Only one
city in the Mathtech's forty-one county coverage has a visibility
increase exceeding one mile. Improvements much less than one
mile are probably not perceptible, and since the contingent market
evidence is predicated on questions designed on 5 or 10 mile
changes in visual range, whether or not consumers value a change
that they can barely perceive is conjectural.* The $100 million
estimate is conservatively derived assuming only 1 in 40 counties
have enough improvement in visibility to be noticeable and valued.
As with the visibility studies, the household soiling studies are
similar. Thus, the lower- and upper-bound estimates are based on
the Watson and Jaksch analysis, by varying a shift parameter in
the supply function.
As apparent from Table 8-2, the health benefits are roughly
equal to the welfare benefits. Noncancer acute morbidity, visibility,
and soiling account for almost all the total benefits. Noncancer
mortality and noncancer chronic morbidity are relatively unimportant,
though there is some potential for significant effects in these areas
as indicated by higher upper-bound estimates. The best information
available points to a rather small cancer benefit associated with
reduced diesel emissions. It is important to note that diesel odor
is an important omitted welfare category.
Generally, the uncertainty is greatest in the health relative
to the welfare area. The uncertainty is decidedly nonsymmetrical,
with greater uncertainty on the lower end. The net result of these
two factors is that the lower-bound total benefit estimate is
influenced most by the greater certainty of welfare effects,
particularly soiling; in contrast the upper-bound total benefit
estimate is influenced more strongly by the health categories.
Table 8-4 shows the regional distribution of benefits. The
benefit-cost criterion establishes whether or not a regulation is
efficient. This means that those that benefit gain more than those
who bear the costs. Distributional consequences, being separate
from efficiency concerns, are usually judged on noneconomic grounds.
*It should be pointed out that programs to reduce particles from
other sources in conjunction with diesel vehicles could very well
improve visibility substantially. The multiprogram improvement,
of which diesel control is a part, would be highly valued.

-------
8-18
Sometimes, when the groups bearing the costs are noticeably
different from those deriving the benefits, perhaps separated
geographically as in the case when pollutants are transported long
distances, the question of property rights becomes the paramount
policy concern. This does not seem to be the case with the diesel
particulate emission issue. While some areas have larger diesel
loadings than others, as reflected in the results of Table 8-4,
the costs of control are diffused among the owners of heavy-duty
trucks and the customers of transported goods and services.
Table 8-5 compares the benefits of a .25 g/BHP-hr. standard
to a .10 g/BHP-hr. standard. The results are exactly what one would
expect.
It is desirable, to the extent possible, to test the
sensitivity of results to key assumptions. This has been undertaken
in three areas: (1) with respect to the assumed diesel sales
penetration rate; (2) with respect to the real discount rate; and
(3) with respect to the use of discounting and accounting for
prior causes of death in the cancer area.
Forecasting diesel sales as a fraction of total sales out to
the year 2000 clearly involves a significant amount of guesswork.
Thus, additional analysis assumed that heavy-duty trucks under
26,000 lbs. gross vehicle weight have a market penetration rate 25
percent higher than in the basic analysis. The heavier trucks,
over 26,000 lbs. gross vehicle weight are left unchanged, since
most trucks of this weight are already diesel. The result is that
total benefits increase by less than 5 percent. Thus, errors in
predicting heavy-duty diesel penetration rates are not very impor-
tant.
The second sensitivity check was with respect to the use of a
real rate of discount of 10 percent. This choice of rates is in
line with OMB guidelines for the preparation of RIAs. Much debate
surrounds the choice of the correct rate, and no consensus has
formed. Most economists believe, however, that 10 percent is too
high. TO test sensitivity to the use of the 10 percent real rate,
the benefits were re-calculated using a discount rate of 5 percent.
With this change, benefits increased approximately 93%. This will
not significantly influence net benefits (total benefits minus
total costs), however, since total costs will also approximately
double, given the relatively smooth stream of equipment and main-
tenance costs. Details of the cost calculation are discussed in
the following section.
Finally, recall that the calculations for cancer benefits,
though comparable to other health areas, intrinsically tend to
give low values. Two sensitivity checks were run. First, given
the latency period, most of the lives lost to respiratory cancer

-------
8-19
Table 8-4
Geographic Distribution of .25 g/BHP-hr.
Beavy-Duty Diesel Particulate Standard Benefits*
CPA Region
Tbtal Benefits
I
New England
184.
II
New York - New Jersey
612.
III
Middle Atlantic
544.
IV
South Atlantic
812.
V
East N. Central
1258.
VI
South Central
661.
VII
Midwest
173.
VIII
Mountain
242.
IX
South Pacific
1890.
X
Jbrth Pacific
260.
U.S.
Tbtal **
6635.
^Present value over 1987-2000 interval in 1987 measured in millions of 1983 dollars.
**U.S. total has been oorputed prior to rounding 90 that mount indicated differs
slightly frcm the actual sun of regional benefits.

-------
8-20
Table 8-5
Benefits of Heavy-Duty Diesel Particulate Standards*
Standard
Benefit Cateogry	.25 q/BHP-hr.	.10 q/BHP-hr.
Health
Nbncancer Mortality 187.	262.
ttoncancer Acute Morbidity 3035.	4272.
Moncancer Chronic Morbidity 21.	30.
Cancer Morbidity and Mortality 1.	1.
Health Subtotal 3244.	4565.
Vfelfare
Visibility 1987.	2909.
Household Soiling 1404.	1981.
Welfare Subtotal 3391.	4890.
Tbtal Benefits	6635.	9455.
•Present value over 1987-2000 interval in 1987 measured in millions of 1983 dollars.

-------
8-21
occur 20 to 30 years after 1987. The value of reducing the risk
of death this far out in tine is substantially reduced by the
10 percent rate of discount. To find out how much this matters,
the discounting is removed. Thus, people are assumed to value a
reduction in risking immediate death the same as death much later
in their lifetime. The point estimate increased by a factor of 6,
and the upper-bound estimate increased by a factor 4. This is
somewhat perplexing. Since the time period covered over 50 years,
it was believed a priori that discounting would have a larger
impact. Thus, a second sensitivity check was run. Some individuals
exposed to diesel particles in the period 1987-2000 will die of
other causes prior to the manifestation of cancer. When this
assumption is removed, along with the use of a 0 percent discount
rate, the point cancer benefit estimate approximately increases by
a factor of 30, and the upper-bound estimate by a factor of 14.
Of course, one might question whether one should attribute risk
to cancer, when the actual cause of death is expected to be
unrelated to cancer. In any case, while under different assump-
tions cancer benefits would be larger, they would still be small
relative to noncancer health effects.
B. Particulate Cost Analysis for HDDVs
1. System and Operating Trap-Oxidizer Costs for HDDVs
In this subsection, the component•and operating costs for
equipping a HDDV with a trap-oxidizer will be estimated. The
primary estimate is based on EPA analysis from the Diesel
Particulate Study [10] . Since this estimate is in point form,
other analysis, in particular from a study by Energy Resource
Consultants (ERC) that was summarized in a recent SAE paper [21],
is used to develop a range estimate. As was the situation in
the benefit analysis, statistical control of estimated errors is
not possible with the available data. Thus, no formal accounting
of uncertainty is possible. Rather, to parallel the benefit
analysis, lower-bound, point, and upper-bound estimates are provided.
These are subjective and are only meant to provide a sense of the
measurement uncertainty on the cost side. In the following section
the component and operating cost estimates are used to calculate
the total discounted control costs associated with the .25 g/BHP-hr.
and .10 g/BHP-hr standards.
Table 8-6 summarizes the EPA and ERC cost analyses. Where
comparisons are possible, the estimates differ by 30 to 80 percent.
Some of the difference is explained by the nature of the approxi-
mations. Accurately estimating the fuel penalty is particularly
tricky, since there is little operational data available. It
should also be noted that EPA and ERC use somewhat different

-------
8-22
definitions of truck classes.* By and large, however, the
difference in estimates is due to different assumptions concerning
dealer margins on pollution control equipment, and on different
cost estimates for equipment components.
Both analyses follow the costing procedures developed by
Lindgren [22J to obtain the "Retail Price Equivalent" (RPE) which
is used as the equipment cost estimate shown in Table 8-6. The
RPE is found by estimating the cost of the components in a trap-
oxidizer system to the manufacturer, and then adding markups to
account for the cost of assembly and the manufacturer's and
dealer's profit and overhead. Component costs are based on data
for light-duty trap-oxidizers, data from manufacturers, and
engineering judgment. A simplified version of the formula is:
RPE ¦ [Variable and Fixed Manufacturing Costs] x [1 + Corporate
Level Overhead + Corporate Profit + Dealer Overhead and
Profit]
EPA and ERC follow the Lindgren guidelines faithfully at the
manufacturer level, but differ substantially in estimates of the
markups in the second bracket. ERC uses the Lindgren estimates of
.2, .2, *4 for corporate overhead, corporate profit, and dealer
overhead and profit, respectively. EPA substitutes lower estimates
at the dealer level, which results in the overall bracketed factor
being equal to 1.2 in lieu of 1.8.
EPA's reasoning for adopting lower markup factors is summarized
in a 1980 report [23]. EPA found for the 1976-1978 period that
corporate overhead as a percentage of sales was about 10 percent,
and corporate profit (for GM) 14 percent. These are smaller than
the 20 percent estimates for both categories used by Lindgren.
Using the argument that the incremental cost of sales at the dealer
level is the appropriate markup (that is, unless the control device
itself adds to overhead, fixed costs should not be attributed to
the cost), EPA found that dealer profit as a percent of sales
before taxes amounted to 3 percent in 1977, and used this in lieu
of Lindgren's 40 percent estimate.
The perspective reflected in the Lindgren analysis and ERC's
estimates is different. EPA, in effect, assumes that any cost
producing change, whether for safety purposes, to improve per-
formance, or to reduce emissions, does not add much incrementally
to cost at the dealer level. Overhead in this view is independent
*EPA divides the classes into: 8500-19,000 lbs., 19,000-26,000,
and 26,000 and up; ERC into 8,500-14,000, 14,000-40,000 and
40,000 and up.

-------
8-23
of vehicle sticker prices and gross sales. In contrast, the
Lindgren view is that all costs are variable, so that cost pro-
ducing change should be included in dealer margins. While EPA's
procedure probably more closely approximates actual market condi-
tions, this area has not been carefully analyzed, in particular as
it applies to the retail market for trucks. Thus, a comparison of
the two estimates provides a gauge of a portion of the cost uncer-
tainly .
There are differences in the control cost estimates as well
as in the markup factors. The use of stainless steel piping
reduces maintenance on the exhaust system. EPA accounts for this,
the ERC estimates do not. The result is that the EPA cost esti-
mates for maintenance indicate net savings (i.e., they are negative).
However, even without the maintenance savings the ERC maintenance
estimates are larger than the EPA figures. Thus, with che exception
of the fuel penalty estimates which are quite similar, the ERC
estimates for equipment and maintenance are larger than EPA esti-
mates. Using EPA markup factors, ERC's estimated equipment costs
are 22, 2, and 44 percent higher than for comparable EPA equipment
for the three truck weight categories respectively.
In sum, the ERC estimates are uniformly higher than EPA's
estimates. Since EPA addresses the markup issue more realistically
and is somewhat more detailed, it serves as the best, or point
estimate. Even so, it is clear that some uncertainty exists on
the appropriate markup and on equipment costs, that is reflected
in differences between the EPA and ERC estimates. Since the ERC
estimates are the larger of the pair, they are utilized to compute
an upper-bound estimate. The lower-bound estimate is based on the
EPA calculations; no credible evidence exists that the costs could
be substantially lower. In the following section these vehicle
cost estimates will be used to derive lower-bound, point, and
upper-bound total cost estimates for the 1987-2000 period.
2. Total Discounted Consumer Costs
The basic calculation of discounted total cost is straight-
forward. For each model year, starting in 1988 and running through
the year 2000, the equipment and operating costs detailed in the
previous section are allocated over the typical truck's lifetime.
This can be done in one of two ways. The vehicle costs can be
annualized so that the costs are spread uniformly by year over the
vehicle's lifetime, or they can be allocated in the year in which
they occur. Normally, the latter procedure is most appropriate,
since it reflects the timing of actual resource costs of achieving
emissions reductions. In this application, however, the annualized
allocation of a model year's cost makes more sense. The reason
for this is best understood by examining the benefits and costs of
control for a vehicle manufactured toward the end of the analysis
period, say in 1999. With control devices installed, this car
will continue to emit fewer particles in 1999 and 2000 and well

-------
8-24
Table 8-6
Estimated Lifetime Vehicle Costs for Trap-Oxidizer
	Systems in HDDVs (1983 (toliars)	
Class/Technology/Source
Disoounted Costs
Equipment Maintenance* Fuel Penalty natal
I. Light-Heavy Trucks
Burner Monolith — EPA
— ERC
$363.
624.
($17.)
47.
$126.
116.
$472.
787.
Catalytic
—	EPA
—	ERC
754.
740.
13.
87.
840.
II. Medium-Heavy Trucks
Burner Monolith — EPA
— ERC
$556.
800.
($17.)
60.
$386.
369.
$925.
1229.
Catalytic
—	EPA
—	ERC
1207.
1140.
17.
276.
1433.
III. Line-Haul Trucks
Burner Monolith
Catalytic
EPA
ERC
EPA
ERC
$652.
1320.
1444.
2140.
($53.)
295.
1052.**
$917.
1107.
738.
$1516.
2722.
3930.
^Parentheses indicate negative maintenance expense. See text for explanation.
**Includes discounted cost of replacing catalyst at 250,000 miles.

-------
8-25
beyond. The benefits of control are not counted after 2000.
Thus, if the costs of control were allocated as they actually
accrue, substantial equipment costs are recorded in the year of
purchase, with only operating costs, thereafter. This front-
loading will mean will that most of the costs are counted, in
contrast to only two years of the benefits. If, however, the
costs are annualized, two years of costs will be compared with
two years of benefits.
Using the latter procedure,* each of the model year's costs
are recorded on a spreadsheet, each entry computed by multiplying
per vehicle costs by projected sales. The costs are summed by
year, then discounted, and finally summed over the years in the
1987-2000 interval to give the final result. The worksheet of the
computations based on ERC estimates for the burner-monolith trap-
oxidizer technology is shown in Table 8-7 as an example of this
procedure.
As was the case with the benefit estimates, emissions control
costs are not estimated with certainty. Total cost estimates in
range form are derived from Table 8-7 and a similar spreadsheet
analysis of EPA cost estimates. The lower-bound and point estimates
are based on EPA vehicle cost estimates, while the upper-bound of
the range is based on ERC cost estimates. Total cost estimates
for the .25 g/BHP-hr. and .10 g/BHP-hr. standards are shown in
Table 8-8. Table 8-7 was derived assuming 100% trap usage, which
corresponds to the .10 g/BHP-hr. standard. Total control costs
for the .25 g/BHP-hr. are computed by applying the 70% trap usage
factor calculated in the Diesel Particulate Study [10] to the
results obtained from the spreadsheet analysis. The 70% trap
usage factor assumes that emissions averaging is permissible.
*As indicated, this is not the conceptually correct procedure.
If the information to implement it were available (it is not), the
following procedure would be preferable (though the overall result
in all probability would remain the same in this application).
On the cost side, control costs should be allocated as they occur.
On the benefit side, the benefits accruing from emissions reduc-
tions beyond 2000 from vehicles manufactured prior to 2000 should
be calculated, discounted, and added to discounted benefits in the
1987-2000 interval. In this way all the benefits and costs asso-
ciated with vehicles manufactured in the 1987-2000 interval are
counted at the time they actually occur.

-------
Table 8-7
Spreadsheet Ibtal Cost Analysis for
	HDDV Particulate Control
Calendar Year
Model Year

87
88
89
90
91
92
93
94
95
96
97
98
99
2000
87
88


92.5
92.5
92.5
92.5
92.5
92.5
92.5
92.5
76.8
76.8
76.8
14.8

89



97.0
97.0
97.0
97.0
97.0
97.0
97.0
97.0
80.0
80.0
80.0
15.9
90




101.9
101.9
101.9
101.9
101.9
101.9
101.9
101.9
83.6
83.6
83.6
91





106.4
106.4
016.4
106.4
106.4
106.4
106.4
106.4
86.8
86.8
92






111.5
111.5
111.5
111.5
111.5
111.5
111.5
111.5
90.5
93







116.0
116.0
116.0
116.0
116.0
116.0
116.0
116.0
94








121.1
121.1
121.1
121.1
121.1
121.1
121.1
95









125.8
125.8
125.8
125.8
125.8
125.8
96










130.9
130.9
130.9
130.9
130.9
97











135.4
135.4
135.4
135.4
98












140.5
140.5
140.5
99











-

145.0
145.0
2000














150.1
CD limn Ttotal

0
92.5
189.5
291.4
397.8
509.3
625.3
746.4
872.2
987.4
1105.8
1228.0
1291.4
1341.6
Discounted
*It>tal

0
i-H
•
00
156.6
218.9
271.7
316.1
352.9
383.0
406.8
418.7
426.3
430.4
411.5
388.6
Nates on following page

-------
8-27
	for Table 8-7
1.	Entries in table are in undiscounted millions of 1983
dollars. It is assumed that all vehicles manufactured
have a burner-monolith trap installed. Policies that
require fewer traps can easily be evaluated on a pro-
portional basis.
2.	Per vehicle costs are based on ERC estimates in Table
7-6, which have been annualized as shown:
Vehicle Class
light-heavy medium-heavy heavy-heavy
Average
lifetime
Annualized
cost per
vehicle
8 year	12 year	11 year
147.52	180.37	419.09
3. The ERC estimates provide the upper-bound total cost
estimate.- A similar spreadsheet was constructed for
the EPA estimates, which provided the basis for the
lower-bound and point total cost estimates. (see text.

-------
8-28
Table 8-8
Particulate Control Costs for HDDVs*
Standard
Lower-bound
Point
Upper-bound
25 g/BHP-hr.**
1788
1788
2986.
10 g/BHP-hr
2554
2554
4266
*Present value over 1987-2000 interval in 1987 measured in
millions of 1983 dollars. Annualized values for 14 year interval
at 10% discount rate can be obtained by dividing entries by 7.367.
**Assumes 70% trap usage factor under emissions averaging. Without
averaging, trap usage under the .25 g/BHP-hr. standard would be
near 100%.

-------
8-29
C. Benefit-Cost Comparisons for Particulate HDDV Options
There are six HDDV particulate control options under
consideration. (Refer to Table 7-1 for key information pertaining
to these options.) In all cases, significant control is proposed
in 1990, with an interim standard of .6 g/BHP-hr. in effect in
1987. Baseline emissions are assumed to be .7 g/BHP-hr. The 1990
standards range from .4 g/BHP-hr. down to .1 g/BHP-hr. Averaging
of emissions is permitted in three options, and line-haul trucks
are exempt from the 199 0 standard in two variants. See Table 8-9
for a list of options (option numbers correspond to those in Table
7-1).	The benefit and cost estimates of achieving each option are
summarized in this table along with a tabulation of net benefits
(benefit - cost). The information in Table 8-9 is based on
analysis from earlier sections of this chapter, though the results
are presented in annualized rather than present value form. The
annualized dollar basis is necessitated by the need to have benefit
and cost estimates independent of the policy initiation date, which
differs between the options in this section and those examined in
previous sections.
Options 4A and 5, .25 and .1 g/BHP-hr., respectively, with
averaging permitted on the former, are identical to those considered
previously, except for implementation dates. Thus, the cost and
benefit estimates are derived directly from Tables 8-2, 8-5, and
8-8	for these options. The point estimate for net benefits is
computed as the difference between point estimates for benefits and
costs. The lower-bound net benefit estimate is calculated by sub-
stracting the upper-bound cost from the lower-bound benefit estimate.
Similarly, the upper-bound net benefit estimate is calculated by
subtracting the lower-bound cost from the upper-bound benefit
estimate.
Option 4 is similar to option 4A, except emissions averaging
is not permitted on the former. The benefit of option 4 is equal to
that of option 4A, given the same level of reduced emissions. The
cost of option 4 is higher — essentially equivalent to the cost
required to meet the .1 g/BHP-hr. standard — since trap usage is
100%.
Option 3 is less stringent than options 4 or 5. The benefit
estimate for option 3 is extrapolated linearly on the basis of
relative emissions reductions using the .25 and .1 g/BHP-hr.
estimates from the previous section. (See Table 7-1 for urban
emissions reductions for relevant options). The .4 g/BHP-hr.
standard can be met without usage of traps for most vehicles. The
cost estimate for option 3 is derived using EPA estimates of
requisite engine modifications as a proportion of the costs of
achieving option 4A.
Options 4B and 4C are the split standards. They are proposed
because line-haul trucks operate between cities and only contribute
marginally to urban loadings. Yet, the cost of particulate control

-------
8-30
on a vehicle basis is quite high for line-haul trucks which fall
into the heavy-heavy class. (See note to Table 8-7 for cost
comparison.) The cost of meeting options 4B and 4C are computed
on the basis of relative cost reductions similar to the calcula-
tion required in option 3. The benefits are extrapolated on
the basis of relative reduced emission loadings, again similar to
the calculation required in option 3.
In most important respects, the benefits and costs reported
in Table 8-9 are comparable. Each is based on similar air quality
information. Each is valued in 1983 dollars, and reflects a common
time horizon and common discount rate. This does not mean, however,
that the accuracy of estimating benefits and costs is always high
enough so that differences between benefits and costs are necessarily
significant. The many caveats throughout this chapter should make
it clear that this may not be the case. Thus, the results reported
in Table 8-9 should be viewed in conjunction with the uncertainty
analysis reflected in the lower-bound and upper-bound estimates.
With this perspective, several results come through rather
clearly. First, the benefits outweigh the costs across all options,
and for the most part by an amount that is large. That is, the
difference between benefits and costs are typically larger than
the magnitude of the costs. Thus, it seems reasonable to conclude
that some degree of further control of diesel particulate HDDV
emissions is efficient from a societal perspective. Second,
obtaining a ranking of options on an efficiency basis involves far
more uncertainty. The differences in the point estimates of net
benefits are smaller than the apparent magnitude of estimation
errors. Thus, while the best estimate of net benefits for option
4A indicates that it is better than options 3 and 4, for example,
there is a real possibility option 3 or 4 may in fact be better.
This is evident since the upper-bound net benefit estimates for
options 3 and 4 are at least a factor of three higher than the
point estimate of option 4A.
While the lower-bound estimates of net benefits are uniformly
negative, we do not conclude that the "no control" possibility
should be considered seriously. In the first place, the lower-
bound net benefit estimates are based on the difference between
the lowest conceivable benefit and highest conceivable cost, and
thus do not represent a very probable outcome. However, they
provide useful information, especially to a risk-adverse decision
maker. Given no significant basis for choosing between options
4 and 4B, for example, a risk-adverse individual would probably
prefer option 4B, with a much smaller lower-bound negative net
benefit. The same could be said for option 3.
In sura, option 4A is the preferred choice on an economic
efficiency basis. Option 5 has higher net benefits, but not
significantly so, and it entails some risk of a policy error given

-------
8-31
the large negative lower-bound net benefit estimate. Options 3
and 4 do provide positive net benefits in all probability. However,
option 4A would always be preferable to option 4, given they are
identical except that option 4 is necessarily more costly given
averaging is not permitted. In addition, option 4 has the dis-
advantage of a high lower-bound negative net benefit estimate.
The first of the split options, 4A, is almost identical to option
3 while the more stringent split standard, 4B, has marginally
higher net benefits.

-------
Table 8-9
Benefit
-penefit-Cost Summary of HDCV Particulate Oontrol Options
(annualized millions of 1983 dollars)	
Cost
Net Benefits
lower-bound point upper-bound lower bound point upper-bound lower bound point upper-bound
87 -	.6
90 -	.4
87 -	.6
90 -	.25
119.
169.
87 - .6
90 - .25(A) 169.
87 - .6
90 - .25(A) 117.
90 - .6 line
haul
HDDVs
87 - .6	148.
90 - .25(A)
90 - .4 line
haul
HDEVs
87 - .6
90 - .1
240.
632.
901.
901.
622.
790.
1283.
2010.
2865.
2865.
1977.
2513.
4082.
113.
347.
243.
114.
113.
347.
243.
114.
189.
579.
405.
191.
177.
177.
295.
347.
347.
579.
-70.
-410.
-236.
-74.
-147.
-339.
519.
554.
658.
508.
613.
936.
1897.
2518.
2622.
1863.
2336.
3735.
* Baseline is assumed to be .7 g/BHP-hr.; (A) denotes emissions averaging permitted. See text for explanation of
entries in table, in particular for options not examined in previous sections of this chapter.

-------
8-33
REFERENCES
1.	Executive Order 12291 of February 17, 19 81, Federal
Register 46: 13193-13198.
2.	Health, Soiling, and Visiblity Benefits of Alternatives
Mobile Source Diesel Particulate Standards, Mathtech Inc., Report
for U.S. EPA, Contract No. 68-01-6596, December, 1983.
3.	Diesel Cars, Benefits, Risks, and Public Policy, National
Research Council, National Academy Press, Washington, D.C., 1982.
4.	CASAC review is summarized in Air Quality Criteria for
Particulate Matter and Sulfur Oxides, U.S. EPA, EPA-600/8-82-029c,
December 1982.
5.	Health Effects of Exposure to Diesel Exhaust, National
Research Council, National Academy Press, Washington, D.C., 1981.
6.	Relation of Air Pollution to Mortality; An Exploration
Using Daily Data for 14 London Winters, 1958-1972. Electric Power
Research Institute, Palo Alto, 1980.
7.	Air Pollution and Human Health, Johns Hopkins University
Press, Baltimore, 1977.
8.	The most recent of which is: "Chronic Nonspecific
Respiratory Disease in Berlin, New Hampshire, 1967-1973: A Further
Follow-up Study," American Review of Respiratory Disease 113:475-485,
1976.
9.	"The Effects of Air Pollution of Work Loss and Morbidity,"
J. of Environmental Economics and Management, 10:371-382, 1983.
10.	Diesel Particulate Study, Draft, U.S. EPA, OAR, OMS, ECTD,
October, 1983 .
11.	Potential Risk of Lung Cancer from Diesel Engine Emissions,
National Academy Press, Washington, D.C., 1981.
12.	Air and Water Pollution Control: A Benefit-Cost Assessment,
John Wiley and Sons, New York, 1982.
13.	Establishing and Valuing the Effects of Improved Visibility
in Eastern United States, Draft Report, U.S. EPA 1983.
14.	Methods Development for Assessing Tradeoffs in Environmental
Management, Vol. II, U.S. EPA, EPA-600/6-79-001b, 1979.
15.	"Air Pollution: Household Soiling and Consumer Welfare
Losses," J. of Environmental Economics and Management, 9:248-262,
1982.

-------
8-34
16.	Study to Determine Residential Soiling Costs of
Particulate Air Pollution, National Air Pollution Control
Administration, October 1970.
17.	"Calibration of the D.S.I.R. Standard Smoke Filter
for Diesel Smoke," International Journal of Air and Water Pollution,
9:351-356, 1965.
18.	"Assessment of the Impact of Light Duty Diesel Vehicles
on Soiling in California," Report for California Air Resources
Board, 1983.
19.	"Relationship Between Air Pollution and Emergency Room
Visits in an Industrial Community," J. of the Air Pollution Control
Association, 31:236-240, 1981.
20.	Methods Development for Assessing Air Pollution Control
Benefits Vol 1: Experiments in the Economics of Air Pollution
Epidemiology, EPA Report, 19 79.
21.	Weaver, C. "Particulate Control Technology and Particulate
Standards for Heavy-Duty Diesel Engines," in Diesel Particulate
Traps P-140, Paper No 840174, Society of Automotive Engineers,
Warrendale, Penn., 1984.
22.	Cost Estimations for Emissions Control Related Components/
Systems and Costs Methodology Description, U.S. EPA Report: EPA-
460/3-78-002, 1978.
23.	Summary and Analysis of Comments on the Notice of Proposed
Rulemaking for the Control of Light-Duty Diesel Particulate Emissions
from 1981 and Later Year Vehicles, U.S. EPA, OMS, ECTD, 1980.

-------