REGULATORY AMLYSIS AND BIVIRDNMENTAL ^PACT OF
FIfJAL EMISSION REGULATIONS FDR m AND
LATER MODEL YEAR HEAVY DUTY ENGIIB
UNITED STATES ENVIRQtf-ENTAL PROTECTION AGENCY
OFFICE OF MOBILE SOURCE AIR POLLUTION COTfTROL
DECEMBER 1979
-------
REGULATORY ANALYSIS AND ENVIRONMENTAL IMPACT OF
FIHAL EMISSION REGULATIONS FOR 1984 AMD
LATER MODEL YEA& HEAVY DUTY ENGINES
PSEPASED BY
OFFICE OF MOBILE SOURCE AIR POLLUTION CONTROL
Michael P. Walsh, Depucy Assijcanc Adainiscracor for
Mobile Source Air Pollution ConcroL
APPROVED BY
Daca- December 11.1979
-------
NOTE
This document has been prepared in satisfaction of the Regu-
latory Analysis required by Executive Order 12044 and the Economic
Impact Assessment required by Section 317 of the amended Clean Air
Act. This document also contains an Environmental Impact Statement
for the Final Rulemaking Action.
-------
Table of Concents
A. Overview of Final Rulemaking
3. Industry Description
C. Impact on the Environment
D. Coats
E. Alternatives
?. Cost Effectiveness
II. Introduction
A. Heavy-Duty Engine Exhaust Emission
Regulation Background
3. Description of Statutory Heavy-fluty Engine
EC and CO Emission Control
1. (few Emission Test Procedures
2. New Definition of "Useful Life"
3. Revised Certification Requirements
Regarding Ourability
4. Emission Standards
5. Parameter Adjustment
6. SEA, PCA, and HC?
C. Organization of the Regulatory Analysis
III. Description of the Product and the Induscry
A. Heavy-Duty Vehicles
3. Heavy-Duty Vehicles Engines
-------
C. Manufacturers 26
1. Engine Manufacturers 27
2. Vehicle Manufacturers 31
D. Users of Heavy-Outy Vehicles 35
E. The Future of Eeavy-Outy Vehicles 39
IV. Environmental Impact 49
A.. Sackground 49
3. Primary Impact 56
1. Heavy-Outy Engine Emission Rates 56
2. Seduction in Urban Emissions from Heavy- 60
Duty Vehicles
3. Ambient Air Quality Impact o£ Regulation 64
C. Potential Secondary Environmental Impacts 63
1. Sulfuric Acid Emissions 68
2. Lead 70
3. Water Pollution, Noise Control, Energy 70
Consumption
D. Irreversible and Irretrievable Commitment 70
of Resources
E. Relationship of Short-lens Uses of che Environment 71
to Maintenance and Enhancement of Long-Term
Productivity
V. Economic Impact 73
A. Coat co Engine Manufacturers 73
1. Emission Control System Costs: Gasoline— 73
Fueled Engines
2. Emission Control System Costs: Diesel 77
Engines
-------
3. Certification Coats 33
4. Test Facilities Modification 38
a. Oynamomecers and Control Systems 38
b. Constant Volume Samp Ling Systems 94
c. Analytical Syscema 94
d. Mew Structures and Remodeling of 94
Existing Structures
a. Software and Computer Hook-up 94
3. Selective Enforcement Auditing Costs (SEA) 101
a. Test Facilities and equipment for 101
formal SEA
b. SEA Testing Costs 103
c. 10 ?ercent Acceptable Quality Level 109
(AQL) Costs
6. tocal Costs to Manufacturers 112
3. Costs to Users of Heavy-Oucy Vehicles 112
1. Overhead and Profit 112
a. Gasoline-Fueled Engines 119
b. Diesel Engines 120
2. Increase in First Costs 121
3. Maintenance Costs 121
4. Fuel Economy and Fuel Costs 123
3. Tocal Cose to Users 127
C. Aggregate Coses: 1984-1988 129
0. Socio-economic Impact 130
1. Impact on Heavy-Outy Engine and Vehicle 130
Producers
2. Impact on Users of Seavy-Outy Vehicles 140
-------
3. Impact on FueL Cost3 co.Users of Ocher 140
7ehicles
4. Balance of Trade 141
71. Alternate Actions 144
A. Introduceion 144
3. Alternatives co Specific Elements of che 144
Rulemaking
1. Te3t Procedures 144
2. Ocher Elemencs 145
C. Alternative Timing for Implementation of 146
:he Rulemaking
0. Alternative Levels of Stringency for che 146
Emission Standards
1. Lifetime Emission deductions and Cost 147
2. Cost Effectiveness 149
3. Changes in Mobile Source Emissions 150
4. Change in Air Quality 150
5. Conclusions 150
711. Cost Effectiveness 152
A. Methodology 152
3. Background 152
0 • S t^DA 153
D. Gasoline-Fueled Engines 154
1. Transient Test 154
a. 90 Percent on che 9-ttode 163
b. Transient Test Procedure 167
-------
c. Costs L 63
2. Redefinition of "Useful Life" 168
a. Benefits 169
b. Costa 170
3. Paraaecer Adjustment 170
4. Allowable Maintenance Restrictions 172
5. Selective Enforcement Auditing (SEA) 173
a. Sane fit3 173
b. Coats 177
6. Inspection and Maintenance (I/M) 178
7. Idle Test 180
E. Diesel Engines 180
1. Overall Rulemaking 180
2. Transient Test 182
3. Redefinition of "Useful Life" 183
i. Paraaecer Adjuscaenc 183
5. Allowable Maintenance 184
6. Selective Enforcement Audit (SEA) 184
Appendix A - Revised Emission Factors tor tteavy-Ducy 188
Gaseous
Regulatory Analysis
-------
CHAPTER 1
SUMMARY
A. Overview of iLulemaking
Aa cbe cotal amount of urban emissions from Light-duty vehi-
cles and trucks is reduced, the portion which heavy-duty vehicles
contribute to those emissions becomes an increasingly significant
factor- For example, ic is expected chat che fraction of total
sobile source urban non-methane hydrocarbon (HC) emissions arising
from heavy-duty vehicle operation vill climb from 122 in L976 to
36Z in L999. Similarly, heavy-duty carbon monoxide (CO) emissions
say expand from 15X ca 43Z in ch'e same time frame. Ic id in light
of these expectations chat Congress has mandated stricter controls
on the gaseous eaissions from heavy-duty engines.
This rulemaking follows from che Congressional requirement
chat EPA propose regulations to reduce by at least 90% the emitted
Levels of pollutants from heavy-duty vehicles — both HC and CO,
relative to a baseline of uncontrolled (pre-1970) emissions.
(Oxides of nitrogen (NOx) reduction vill be addressed in separate
regulations, though che present regulations do include a NOx
standard such chat no further control 3hould be required under che
new cest procedures.) the purpose of this specific document is Co
present che results of S?A analyses of che environmental and
economic impacts and che cost effectiveness of the regulations.
The reader vill find chapters devoted as veil to che make-up of che
heavy-duty industry and co alternative actions considered by che
agency.
The accompanying regulations define LeveLs for che Congre9-
sionally mandated gaseous emission standards for heavy-duty
engines. Also introduced here are requirements that the standards
be sat on a cast procedure which prescribes transient engine
operation. The shift from che steady-scate procedures completes an
EPA development program which yielded test cycles derived from
actual in-use vehicle operation; such transient testing, ic is
reasoned, more accurately assesses on-the-road emissions Chan do
che previous procedures. The procedure is designed around an
engine test and requires chat emission dumbers be determined on a
useful-vork-produced basis (i.e., grams per brake horsepower-hour
ar g/BHP-hr).
The BC and CO emissions standards which appear in chis rule-
making represent a 90 percent reduction from che average measured
emissions of twenty-three 1969 gasoline heavy-duty engines.
Numerically, the standards are 1.3 g/3K?-hr for SC and L5.5
g/BHP-hr for CO.
The NOx standard being promulgated in this regulation was-
-------
derived from data developed in a separate case program in which
1979 engines were cesced. The average of chese baseline results
was adjusted 30 chat in 1984 most engines should require no further
NOx control than chey currently have.
Several other aspects of this rule distinguish it from the
current heavy-duty engine regulations. First, a new definition of
"useful life" and revised durability requirements will alter the
certification procedure. An engine's "useful life" is reached
whenever 1) che average lifetime of ics engine family is reached or
2) che mechanical integrity has deteriorated co che poinc of
requiring a rebuild. However, in no case can chis useful life
period be Less chan 50,000 miles. The manufacturers will determine
their ovn deterioration factors based on procedures which each
develops. This procedure will be used uncil che statutory heavy-
ducy NOx emissions standards are finalized. Ac chat cime, a new
durability procedure is expected to be be promulgated.
Another change in EPA's past course involves a provision
affecting parameter adjustment. In it, che Administrator will be
allowed to require chat certain adjustable emi33ion-affeccing
parameters be set at other than recommended settings for certifi-
cation; the purpose is to encourage design of engines vich emission
characteristics chat are less 3uscepcible co in-use maladjuscaenc.
Additionally, E?A introduces in these regulations an idle cest
procedure co facilitate che promulgation of emissions' warranty
regulations under 3207(b) of che Clean Air Act. Idle operacion
occurs in situations that involve fairly direcc exposure of people
co CO and comprises a significant portion of heavy—duty operation.
Also being promulgated are regulations to control heavy-duty
diesel crankcase amissions. The current HC emission regulations
place controls only on crankcase emissions from gasoline-fueled
engines. Under the implemented changes, no crankcase emissions
from naturally-aspirated heavy-duty diesel engines will be per-
mitted.
Finally, an assembly-line emissions testing program known as
che Selective Enforcement Audit (SEA) will be implemented.
This program will aid in ensuring chac actual production engines
meet che emission levels co which chey are certified. SEA's are
initiated by a cesc order from EPA and cover only one engine
configuration per cest order. The number of SEA'3 a manufacturer
must undergo each year is based primarily, buc not exclusively, on
projected annual sales. The goal in an SEA is to ascertain whether
or not che engines cested meet a 102 Acceptable Qualicy Level
(AQL). This AQL would require virtually all engines Co aeec
applicable standards after adjustment for deterioration wich only
10Z allowed co exceed standards to provide for cesc variability and
isolated instances of nonconformity.
-2-
-------
Failure of an SEA may lead cc suspension or revocacion of che
engine's certificate of conformity. The manufacturer would then be
permitted to make running changes or quality control changes co che
engine and Chen undergo another SEA. Another possibility will be
for che manufacturer to request a Production Compliance Audit (PCA)
to determine che compliance level of the engine configuration, and
then pay a Soncomformance Penalty (MC?) based on che "marginal
cost" of compliance becveen engines in compliance and che noncon-
forming engines. The provisions for'PCA's and HCP's which were La
che proposed rulemaking are not being finalized at this cime.
However, EPA intends Co finalize cham in cime for che 1984 modeL
year.
3. Industry Description
The "heavy-duty industry" discussed here refers co chat
collection of companies which manufactures che cruck3, buses, and
engines found in on-che-road applications whose gross vehicle
weights (GVW) exceed 3,500 pounds. The rather complex picture
presented by che numerous manufacturers and cheir diverse produce
lines is simplified somewhat by che realization chat only a few of
chese companies are responsible for che bulk of che industry's
produceion.
General Mocors, Ford, Chrysler, and International Harvester
(IHC) share over 99Z of che heavy-duty gasoline engine market;
Cummins Engine, Detroit Diesel, Mack, Caterpillar and IHC are che
primary diesel engine producers. Only CM (including Detroit
Diesel) and ISC make both types of engines in significant quan-
tities .
Vehicles in che industry are produced in many configurations
(single unit or tractor, gasoline or diesel, various axle arrange-
ments and load capacities, etc.)by a number of manufacturers,
but, as with che engines, most vehicles are buile by che largest
producers. ®! (Chevrolet and QIC), Ford, Chrysler (Dodge), and IHC
make over four fifths of ail U.S.-built trucks.
The applications of trucks Co real vorld tasks vary widely
depending on load capacity, ranging from personal transportation
and agriculture, to construction, trade, and "for hire" uses. The
companies and individuals who purchase trucks and buses cake
advantage of che diversity of available products and choose
vehicle-engine combinations which economically fulfill cheir
needs.
C. Impact on che Environment
As previously noted, che amount of emissions produced by
heavy-duty vehicles are becoming an increasingly significant
portion of che cotal amount of urban amissions. In fact, it is
-------
expected that che fracnoa of cotal EC and CO emissions tram
heavy-duty engines will increase threefold in che cine period 1976
co 1999. In view of chese expectations, Congress has mandated
stricter concroLs on the gaseous emissions from heavy-duty engines.
As a result of the 3cricter controls, heavy-due/ gasoline
vehicles will exhibit life tuna improvement of 1 con in EC and 28.6
cons in CO relative co che scenario of a continuation of 1979
standards. Lifetime diesel EC emissions will be reduced by 0.77
cons. The effect which chis reduction would have on cocal mobile
source urban emissions translates co a 17 percent improvement in EC
and 30 percent for CO by 1999, again compared co che case of no new
heavy-duty regulations.
On che basis of further calculations £?A estimates chac as a
result of the rulemaking, the ambient levels of ozone and CO will
be reduced in 1999 by 2 percent ana 7 percent, res pec c ive 1/.
Secondary emission effects; water, noise and energy consump-
tion effects; and commitment of scarce resources are all expected
co be negligible as a result of promulgation of the regulations.
D. Coses
The increased costs which che heavy-ducy engine manufacturers,
and ul.cima.ceiy che consumer, will have co bear as a result of chis
regulation consists of costs for purchase and installation of new
test facilities, for development and installation of new amission
control systems, and for certification and SEA testing. For che
diesel manufacturers, che new cest facilities (mainly dynamometers
and emission 3yscem development and hardware) will be che primary
costs. For gasoline manufacturers, che emission controls will be
highest. There are additional coses falling upon che operators of
gasoline engine-equipped vehicles which are addressed below.
An increase of approximately $294 can be expected in che price
of a gasoline engine, $253 of which is attributable co che manufac-
turing cosC3 of che catalyse system. The remaining 3141 is primar-
ily attributable co profit, overhead, and equipment acquisition and
modification costs, amortized over 5 years. For diesel engines, a
first price increase of $195 is expected. Of chis $195 increase,
$56 is due co emission control development and hardware, and the
remainder is found primarily in overhead, profit and aew facilities
ralaced Co the transient cesc procedure and SEA. It is because
catalyst control will not be required or che diesel manufacturers
chac their production costs will be less ehan for che gasoline
producers. Certification cosC3 for both engine cypes are aoc
expected co rise appreciably.
No increased operating coses are expected co fall upon che.
users of diese l-equicped vehicles. However, gasoline vehicle'
-4-
-------
operators will incur the additional coses of unleaded fuel.
Offsetting these coses is che reduced frequency of replacemenc of
spark plugs and che exhaust system. In addition, SPA expects that
che manufacturers will be able co achieve ac Lease a 4 percent fuel
economy gain in gasoline-fueled engines. This could lead co a
discounted fuel savings of $788 per vehicle over Lcs lifecime.
The anticipated nee increase in operational costs for che gasoline
vehicle user amounts co about $259 (present worth on January 1,
1934, assuming a 10 percent interest rate), ignoring che fuel
economy benefit.
The aggregate cocal coses for all heavy-duty engines produced
in che five-year period beginning in model year 1984, discounted co
che effective dace of che regulations (January 1, 1984), is found
co be $705 million for gasoline engines and $243 million for
diesels. This aggregate cosc includes che increased! firsc cose
for each engine plus increased operating cosc3, but does no*,
include che fueL economy benefit for gasoline-fueled engines.
3ecause EPA expects che 1984 heavy-duty regulations to have
only slight impacts on industry-vide sales, the industry's amploy-
tnent and production should noc suffer. Also, user3 of heavy-duty
vehicles and of ocher vehicles should expect no burden as a result
of che rulemaking.
E. Alternatives
As EPA has proceeded with the development of a final rule-
making based upon analysis of comments received in response to che
February tlPSM, alternatives and options in essentially all aspects
of the rulemaking have been evaluated. These alternatives fail
into three broad areas: 1) alternatives to specific elements of
the rulemaking, 2) alternative timing for implementation of che
rulemaking, and 3) alternative levels of stringency for che emis-
sion standards. Each of chese will be summarized separately.
1. Alternatives to Specific Elements of the, Rulemaking
The test procedure was one of che oust controversial aspects
of the proposal. The alternative is to promulgate a regulation
using either the 9- or 13-mode steady-stace tests. The fundamental
question relating co che test procedure relaces to che ability of
Che sceady-scace procedure Co adequately characterize in-use
emissions of heavy-ducy engines. Available daea indicacas that
steady-state procedures are deficient in ehis regard. Therefore,
che eraasiene procedure will be used.
Alternatives relating eo the redefinition of "useful Life,"
durability testing, parameter adjustment, allowable maintenance,
assembly-Line testing with 10 percent AQL, and diesel crankcase
control are treated in detail in the Summary and Analysis of
Comments.
-3-
-------
Alcernacive Timing for Implemencacion of che Rulemaking
The NPRM called for implemencacion of che regulacion in 1983
in accordance vich che 1977 Clean Air Act. Several comments vere
submit tad by che manufacturers. EPA 3 caff has analyzed che many
commencs of che manufacturers, and has concluded chac gasoline-
fueled engines could possibly comply by 1983 ac a high risk. Some
families of diesel engines could aiso comply, buc chose requiring
significant reductions could tioc. Therefore, EPA has chosen co
delay implemencacion of che regulation until 1984.
3. Alternative Levels of Stringency of che Standards
The Clean Air Act requires chac EPA "conduce a continuing
pollucanc specific study concerning the effects of each air pol-
lutant emitted from heavy-ducy vehicles or engines and from other
sources of mobile source related pollucants on the public health
and welfare." The intenc of requiring these reports wa3 co provide
some of the framework needed co evaluate che scacutory standards
for heavy-ducy vehicles.
The scatuce provides for emission standards for boch gasoline-
fueled and diesel engines representing a 90 percent reduction from
a 1969 gasoline-fueled engine baseline. Two alternatives are
considered in this regulatory analysis. One considers an 35
percent reduction from baseline and che other considers a 95
percenc reduction from baseline.
The effect of changing che stringency of che standard is
significant over che average life of a heavy-ducy vehicle. Anal-
ysis indicates that relaxing che standard co che 35 percent level
would increase HC emissions by a factor of 1.4 for gasoline en-
gines. A similar change occurs for CO. On the ocher hand, in-
creasing the stringency would reduce HC by a factor of. 1.7 for
gasoline eaignes and by Z.O for diesel engines. Diesel CO emis-
sions are unaffected by a change in che scandard because they are
already lover than che 95 percent reduction standard.
In terms of expenditure, the 35 percenc standard would reduce
che cost per engine from 5477 co 1^26 for gasoline engines. For
diesel engines, the 35 percenc standard would reduce che per engine
cost from 3195 to $178. So gasoline or diesel engine cost dif-
ferences were estistaced for the 95 percent standard because the
targec CO level for gasoline engines and che target HC level for
diesel engines are so low that the feasibility of this opcion is
questionable.
F. Cost Effectiveness
Cost effectiveness as applied co pollution controls is the
cost of control per con of reduction Ln pollucanc. EPA's calcula-
—6—
-------
eioas yield cose effectiveness numbers for gasoline engines of $238
per con of HC and $8 per con of CO reduction. Diesel coses are
expected Co be applied coward HC control, hence the entire cose is
allocaced co HC. The estimated cose effectiveness for this diesel
HC concrol is 9253 per ton of reduction. Also evaluated in this
report are incremental cost-effectiveness values for che various
components of the overall rulemaking.
It is EPA's position, especially in Lighc of che benefies of
Che cransienc procedures, that che 1984 heavy-dut7 regulations are
indeed cost effective.
-7-
-------
CHAPTER II
urraoDucTioH
A. Beavy-Pucy Engine Exhaust Emission Regulation Background
Eeavy-duty engine exnaust emissions vere first ceguiaced by
che Stace of California beginning in 1969 (see Table II-A for a
susnnary of actual standards). The 1969 California emission stan-
dards vere expressed only in terms of exhausc gas concentration,
applied only co gasoline-fueled engines, and covered only HC and CO
emissions. SPA adopced che California standards and cesc proce-
dures for gasoline-fueled engines beginning in 1970, and imposed
exhaust and smoke emission standards for diesel engines.
The next improvement in heavy-duty engine emission measurement
cechaiques occurred with che introduction of revised heavy-duty
engine cesc procedures (for both gasoline-fueled and diesel
engines) by California for 1973. These regulations called for mass
measurement of che pollutants, and extended che standards to NOx
emissions by including a standard for ac * NQx. These procedures
were adapted by EPA for 1974. These procedures, while different
for gasoline-fueled and diesel engines, basically required che
operation of engines on an engine dynamometer ac several steady-
state speeds. Samples of engine exhaust vere collected during che
various stages of gasoline "9 mode" and diesel "13 mode" casc3, and
quantities of RC, CO and NOx pollutants vere determined. Emissions
were measured as a function of che useful vork performed by the
engine, and expressed in grams of pollutant emitted per engine
brake horsepower-hour (g/BHP-hr). The result was chat an engine
with high horsepower was allowed to pollute more chan one with less
horsepower, since it performs more useful vork. These procedures
remained in effect chrough 1973 with only minor technical improve-
ments.
As part of an EPA heavy-duty test procedure development and
technology assessment program begun in 1972, EPA evaluaced che
1974-73 steady-state heavy-duty engine test procedures in an
attempt to relate emissions measured on che cest procedure co
actual on-the-road HDV exhaust emissions. The data evaluated
indicated chat ac emission levels below the 1974-78 standards, che
results of emission casts using che 1974-78 cesc procedures were
inadequate predictors of on-che-road CO and NOx emissions, i.e., a
given reduction in emissions measured on che current cesc procedure
results in a much smaller reduction in actual on-che-road emis-
sions . iy
In 1977, EPA therefore adopced modifications co che 1974-78
cesc procedures which improved che accuracy of che cesc procedures
enough co allow che promulgation of more stringent standards. The
combination of the revised cesc procedures and new standards were
-3-
-------
Table II-A
BeavvOuty Engine Exhaust Emission Standards
Federal
California
Year
Opcion
ac
CO
NOx
hc+nox
Ootion
ac
CO
NOx
HC+NOx
1969
MR
sr
MR
MR
275a
1.5a
MR
MR
1970-71
275*
l.5a
MR
MR
275
1.5
MR
MR
1972
275
1.5
MR
MR
180
l*8
MR
1973
275
1.5
MR
—
40
—
16
L974
—
40°
—
lo
—
40
—
16
1975-76
—
40
—
16
—
30
—
10
1977-78
—
40
—
16
A
—
25
—
5
3
1.0
25
7.5
—
1979
A
1.5C
25
—
L0C
A
1.5C
25
7.5
—
3
—
25
—
5
3
—
25
—
5
1980-83
A
1.5C
25
—
10C
A
1.0
25
—
6
3
j
25
—
5
3
—
25
—
5
1984
15.1,
10.7
—
0.5
25
—
4.5
1985
1.3d
15.5d
75Ze
—
HC ¦ parts per million; CO * Z mole volume. Used for Federal
Standards 1970-73 and California Standards 1969-72.
3 Grams per brake horsepower-hour.
Measured on 1979 cesc procedure (HFTD for HC). Reduced 0.5
g/BHP-hr when 1978 procedure is used (HOIS, for EC). NDIR is
allowed in 1979 for all manufacturers, beyond 1980 only for
low volume manufacturers seeking Federal certification.
d As measured on transient test procedure.
g *
Reduction from 1972/73 baseline for gasoline engines.
MR ¦ Mo requirement.
9-
-------
applicable beginning in 1979. They were referred co as che "inter-
im regulations," because EPA intended co adopc more fundamentally
revised ceac procedures and more stringent standards later. The
interim regulations allowed manufacturers che oocion of cwo sees of
emission standards, one emphasizing 2C control and che other NOx
control. All manufacturers ware allowed co poscpone use of che
interim (modified) test procedures co 1980. Small-volume tnanufac-
curers were allowed co recain che old cesc procedures indefinitely
under EPA regulations, buc not under California regulations.
California has established progressively more scringeac standards
using che interim cesc procedures. Currenc EPA scandards do noc
change beyond 1979.2/
Since 19 77, EPA has continued its development of a fundamen-
tally new, heavy-ducy engine case procedure co make emission
reductions measured in che laboracory more represencative of
percent reductions one would expect co achieve in-use. Also, che.
1977 Amendments co che Clean Air Act directed EPA co promulgate new
HC and CO emission standards applicable in 1983 which would require
a 90 percent reduction in each pollutant from a baseline of 1969
heavy-duty gasoline engines, and a new NOx standard applicable in
1985 which would require a 75 percent reduction from a baseline of
1973 heavy-duty gasoiine engines. Based on its evaluation of che
1974-78 test procedures, EPA considers the current, interim cesc
procedures co be incapable of ensuring reduccions of chese magni-
tudes in in-use amissions. Therefore, this action consists
of che promulgation of che statutory HC and CO standards as mea-
sured on a new cransienc engine cest procedure. The new cest
procedure is the culmination of EPA's development work begun in
1972. Promulgation of the statutory 1985 NOx standard will be
proposed at a later date.
B. Description of Statutory Heaw-Puty Engine BC and CO Emission
Control
1. Mew Emission Test Procedures
EPA is establishing new test procedures for determining
gaseous exhaust emissions (including NOx) from heavy-duty engines.
Key features of che new cesx procedures, especially che engine
operating cycle, will likely be used for measuring diesel exhaust
particulates starting in a model year yec co be proposed. When che
diesel particulate regulations are proposed, che need for smoke
scandards will be addressed. Is the meantime, the current diesel
smoke measurement procedures will continue co be used afcer che new
gaseous emission cesc procedure is in effect. EPA also requires
Chat all manufacturers of heavy-duty engines use che new cesc
procedures for certification testing, i.e., chat che current
optional use of che 1974-1978 test procedures by low volume manu-
facturers be ended after 1983. Heavy-duty diesel engine aanufac-
-10-
-------
carers are. allowed che oocion of certifying their engine families
using che 1979 procedure for 1984 only. All heavy-ducy manufac-
turers atusc use che new cesc procedure for che 1985 model year.
Like che currenc cesc procedures, che new procedures measure
emissions from engines while mounted and operacing on an engine
dynamometer. However, che new procedures differ from che currenc
cesc procedures in several areas. The chree fundamencal poincs of
difference are che engine operacing cycles over which amissions are
measured, che sampling mechod used co collecc emissions during
engine operacion, and che requirement for boch cold and hoc scare
cesc segmencs. These chree differences in cum necessitate several
related changes involving engine mapping, inscrianencac ion, and
aquipmenc calibration. The new cesc procedures closely resemble
che current light-duty vehicle and light-ducy truck cesc procedures
(Subpart 3 of CFR Title ^0 Part 86) in che areas of emission
sampling, instrumentation, and equipment calibration. 3/
The new cesc procedures contain cvo cransienc engine operacing
cycles, one for gasoline-fueled engines and che other for diesel
engines. The cvo cycles were developed by EPA from daca on
the operacing characteristics of in-use heavy-duty engines of each
cype. t*f £ach cycle is specified by a second-by-second listing of
pairs or normalized engine speed and power values. Unnormalizing
che cycle inco an accual speed-power cycle requires chac che
curve of maximum engine power vs. engine speed be known. Deter-
mining chis curve experimencally is one of che earliest seeps in
che cesc sequence. After chis engine mapping is done and che
results are used to compile an actual speed-torque cest cycle, che
cesc engine is allowed a long soak or is subjected co a forced cool
down. It is chen scarced from che cold condicion, operated over
che cesc cycle, shut off for a brief soak, restarted in che hoc
condition, and operaced again over che cesc cycle. Tolerances on
how closely che engine muse follow che cesc cycle are specified in
che procedures.
Mass emissions for each pollutant and useful work output are
measured separately for che cold start and hot start segments of
che cesc. This allows emissions from in-use cold stare crips
(i.e., crips which begin with the engine at ambient cemperacure) co
be escimaced separately from emissions from in-use hot stare trips.
The cvo are then weighted with che racio of che frequencies of che
cvo types of in-use crips and divided by che similarly weighted
useful work oucpuc co gee che brake-specific emissions from an
"average" in-use crip.
Mass emissions from each cese segment are measured by diluting
che hoc exhaust gas stream with cooler air and collecting a small,
proportional sample of chis dilute mixture in a bag. The concen-
trations of pollutants in chis bag are measured using analytical
instruments suited to such measurements (a flame ionization detec-
-11-
-------
cor for HC, a non-dispersive infrared analyzer for CO and CO,,
and a chemiluminescence analyzer for NOx), che cocal volume 5c
diluce mixture is calculated from other measurements made during
engine operation, and from these the mass of each pollutant emitted
during the cast segment is calculated. SC emissions, and' at che
discretion of che manufacturer, NOx emissions, from diesel engines
are an exception: chase are not bagged but are continuously sampled
and analyzed during the test segment using heated sample lines and
a heated flame ionization deteccor and che chemiluminescence
analyzer. The cast procedure allows che use of cvo types of
constant volume sampling (CVS) systems known co be suitable cor
this cype of emissions sampling, plus ocher systems if approved in
advance.
Useful woric output is measured via che measuring systems which
are integral part3 of che dynamometer controls.
Procedures are specified for periodic equipment calibrations,
as necessary to ensure accurate cest results.
The contrasts between the new and current cest procedures
highlight che important features of che new procedures. The
operacing cycles in the current procedures consist of sequences of
specified steady-state modes (9 modes for gasoline-cue led engines,
13 modes for diesel engines) rather than of second-by-second
listings of speed-power pairs. The current procedures therefore
test engines at fewer points in their operating ranges chan will
che new cycles. The current procedures do noc allow measurement of
emissions during transient conditions representative of in-use
operation. The new procedures will. Since under che currenc
procedures emissions are not measured during periods when exhaust
gas composition and volume are changing, dilution with air and
proportional sampling into a collection bag are noc used. Instead,
paLlutanc concentrations in the exhaust gases are measured directly
over a small portion of each mode and combined with ocher measure-
ments co calculate mass emission resulcs. This measurement of
undiluted exhaust gases requires somewhat different analytical
systems. Separate cold-start and hot-3tart segments are not
performed. Calibration procedures for equipment, and tolerances on
che operating cycles, are correspondingly different.
EPA is also establishing new cest procedures co be used to
determine emissions of CO under idle conditions. The cest pro-
cedures are simple, and can be performed immediately after the
transient cest procedure. The idle cesc procedure will be used
for only gasoline-fueled engines.
2. Hew Definition of "Useful Life"
SPA is amending che current definition of "useful life" for
heavy-duty engines. The amendment vill bring che periods of use-
-12-
-------
specified in che definition inco closer agreement wich che periods
of use actually seen by heavy-ducy engines before retirement or
major refurbishment (e.g., rebuilding or major overhaul).
The amended definition will apply Co che assembly-line
testing, warranty, recall, and certification provisions of the
Clean Air Act. That Is, manufacturers will be required to furnish
owners with Section 207(a) and 207(b) warranties covering the
period of use specified ia the amended definition. A manufacturer
wiil also be liable for recall of a cacegory of its engines if the
EPA Administrator determines that a substantial number of the
cacegory does noc conform to the emission standards during, that
period. And che longer useful life definition will be incorporated
inco the certification and assembly-line testing procedures via
decerioracion factors, as described in che next subsection.
3. Revised Certification Requirements Regarding Durability-
EPA had proposed a substantially revised durability cast
procedure in the MPSH, and had intended to finalize the procedure
with this rulemaking. ¦ However, SPA is delaying the finalization of
the in-use durability testing requirements, in order to improve
this proposed procedure and to optimize all components of the
program. A revised durability test procedure is expected to be
implemented in conjunction with the statutory heavy-duty NOx
emission standard.
Beginning in 1984, and until finalization of a revised dura-
bility test procedure, the burden of durability testing wilL be on
che manufacturers. the manufacturers will determine their deter-
ioration factors in programs which they design and submit these
deterioration factors to EPA as part of the certification process.
4. Emission Standards
The HC and CO emission standards being established by EPA are
applicable to 1984 and later model year heavy-duty engines. These
standards require 90 percent reduce ions from a baseline of 1969
gasoline-fueled engines, as measured vich the new cesc procedures.
These reductions are chose mandated by che Clean Air Act as amend-
ed. The new BC and CO standards will apply to boch gasoline-fueled
and diesel heavy-duty engines.
For most engines, EPA is noc requiring more NOx control in
1984 than was required by the 1979-83 standards. It is not pos-
sible to simply keep che 1979-33 NOx standard in 1984, however,
since there was no NOx-only standard for 1979-83. Further, che
cesc procedures being promulgated for 1984 are different than chose
used in 1979-83. A NOx baseline of 1979 engines cesced vich the
new procedures was used to derive a 1984 NOx standard that is based
-13-
-------
on a statistical analysis of che NOx Levels from che engines in
this sample.
A separate idle standard is also included in chis package for
CO emissions from gasoline-fueled engines. Idle operation repre-
sents che largest single mode or heavy-duty truck operation (ap-
proximately 25 percent of cbe cime in che CAP£-21 program), and
cimes of prolonged idle can also be occasions of high-population
exposure such as at crowded intersections, Loading docks, or pickup
and discharge of bus passengers.
EPA has measured emissions from a series of L969 and 1979
engines using che new cesc procedure, co establish numerical
standards for 3C, CO, and HOx. The final standards being promul-
gated as a resulc of chis cescin'g are L.3 g/BHP-hr (BC) , 15.5
g/3HP-hr (CO) and 10.7 g/BHP-hr (NOx) in 1984. In L985, che HC and
CO standards are che same, but che NOx standard will change co a
level representing a 75 percent reduction from che 1972-73 baseline
cesc program. An exact level vill be proposed at a Lacer dace.
The idle CO scandard for gasoline-fuled engines is 0.5 percent (by
volume).
5. Parameter Adjustment
EPA is amending the certification and cesc procedures co
permit che Administrator co adjust or require manufacturers co
adjust engine parameters co physically accessible settings ocher
chan cheir recommended settings prior co emission cescs of esis-
sion-daca engines. this will encourage manufacturers co design
engines co be Less suscepcible co in-use maladjuscmenc. Such
maladjustment is capable of causing in-use emissions co be sub-
stantially higher chan allowed by standards. The paramecer
adjustment provision vill help ensure that che 90% reductions in HC
and CO mandated by statute are actually achieved by in-use engines.
The specifics of che parameter adjustment rule are essen-
tially che same as chose of che recent final rule on parameter
adjustment for light-duty vehicles and lighc-ducy trucks. Four
types of parameters on gasoline-fueled engines may be Liable to SPA
adjustment in 1984: Ldie mixture, idle speed, initial spark
ciming, and choke valve action paramecer3. tfevLy introduced
parameters on either type of engine may also be liable co adjust-
ment in che year they are introduced. In addition, exising para-
meters on either type of engine beyond the four mentioned above may
become liable co adjustment if EPA nocifies manufacturers and gives
sufficient lead cime for compliance. Parameters vill be adjusted
only if EPA determines that they pose or are reasonably Likely co
pose significant maladjustment problems in use. Procedures are
included for making and appealing chese determinations.
-14-
-------
6. The Selective Enforcement Audit Program (SEA), Produc-
tion Compliance Auditing (PCA) and Nonconformance Penal-
ties (NC7)
The SEA program is aa assembly-Line emissions testing program
used to aid in ensuring chat the engines produced meec the emis-
3ions Level to which chey are certified. SEAs are initiated by a
cesc order from EPA and cover only one engine configuration per
test order. The number of SEAs a manufacturer oust undergo each
year is based primarily, but aoc exclusively, on projected annual
sales. The goal in an SEA is to ascertain whether or not the
production engines tested meet a 10 percent Acceptable Quality
Level (AQL). A 10 percent AQL would require virtually all engines
Co meet applicable standards .after adjustment for deterioration
with only L0 percent allowed to exceed standards to provide for
test variability and isolated instances of nonconformity.
Failure of an SEA may Lead co suspension or revocation of the
engine's certificate of conformity. The manufacturer would then be
permitted to make running changes or quality control changes to the
engine configuration and Chen undergo another SEA. Another possi-
bility will be for the manufacturer to request a Production Compli-
ance Audit (PCA) co determine che compliance Level of the engine
configuration,, and then pay a Nonconformance Penalty (NCP) based on
the "marginal cost" of compliance beewe en engines in compliance and
the nonconforming engines. Ihe provisions for PCA's and for NCP's
which were in the proposed rulemaking are not being finalized at
this time. However, E?A expects chey will be finalized in time for
che 1984 regulation.
C. Organization of che Regulatory Analysis
This analysis presents an assessment of the environmental and
economic impacts of the heavy-duty engine regulations EPA is
promulgating. It provides a description of the information and
analyses used co review all reasonable alternative actions before
implementing che final rule.
The remainder of this statement is divided into five major
sections. Chapter III presents a general description of heavy-duty
vehicles and engines, a brief description of che manufacturers of
this equipment, and the market in which chey compete. It also will
discuss che uses co which heavy-duty vehicles are puc, and describe
the primary user groups.
An assessment of the primary and secondary environmental
impacts attributed to ehe heavy-duty engine regulations is given in
Chapter IV. The degree of control reflected by standards is
described and a projection of air poLLutant emissions for che
national heavy-duty vehicle population, with the standards in place
-IS-
-------
chrough L999, is presented. The impacts of chese regulations on
urban emissions and che expected air quality benefice are con-
sidered. Secondary effects on other air pollutant emissions, water
poliucion and noise are also discussed in chis section.
An examination of the cose of complying with che regulations
is presented in Chapter 7. these costs include chose incurred co
install emission control equipment on heavy-ducy engines, costs
required co purchase new emission cesting calls, che costs co
certify, che cost3 associated with che SEA program, and any La-
creased vehicle operating costs which might occur. Analysis is Bade
co decermine aggregate cost for che 1934-38 cimeframe. Finally,
che impact chat chis regulation vill have on industry and consumers
will be reviewed.
Chapter 71 vill identify and discuss che alternatives co chis
rulemaking action, cheir expected environmental impacts, and che
reasons none have been promulgated.
Chapter 711 vill present a cost effectiveness analysis of chis
rulemaking action and compare che results of chis analysis with
chose conducted on other mobile source concrol scracegies.
-IS-
-------
References
ij "An Examinacion of lacerim Emission Control. Strategies cor
Heavy-Duty Vehicles (A Regulatory Support Document)", EPA
OMSAPC, October 3, 1975.
I! Current Federal heavy-duty engine emission standards and test
procedures are contained Ln the Code of Federal Regulations,
Title 40, Pare 36.
y The new test procedure, together with a li3t of supporting
references, is contained in "Draft Recommended Practice for
Determining- Eahaust Emissions from Heavy-Duty Engines tJhder
Transient Conditions," Chester J. France and William 3.
Clemmens, SDV 78-07, June 1978, available from the Emission
Control Technology Division, EPA, Ann Arbor, Michigan.
j*j Descriptions of the surveillance project and subsequent cycle
development can be found in the references listed Ln the
report cited in the previous foocnote.
17-
-------
CHAPTER III
DESCRIPTION OF THE PRODUCT AWD THE INDUSTRY
A. Beavy-Oucv Vehicles
A heavy-duty vehicle (HDV) as defined by EPA i3 a vehicle
whose gross vehicle weight raciag (GVWR) exceeds 3500 pounds.
This differs from che definition in Che amended Clean Air Acc
which specified 6000 pounds GVW as che Lover limic of HDVs. The
reason for chis difference Ls chat a 1 chough SPA is required co
regulate all vehicles heavier chan 6000 pounds GVWR co at lease
che levels dictated by che Actl/, Light-ducy crucks in 6000-8500
pounds GVWR range are dealc vich under seoarace reguLacions.
These regulations are aimed ac che greacer chan 3500 pound CWR
population only.
The industry as well uses GVWR as a basis for reporting
produccion and sales data. Their traditional categories are:
Class Weight (Pounds ~ GVWR)
I 0-6,000
II 6,001 - 10,000
III 10,001 - 14,000
17 14,001 - 16,000
7 16,001 - 19,500
71 19,501 - 26,000
HI 26,001 - 33,000
Till 33,001 and over
SPA's definition of light-duty cruck3 sees che division
between che LOT class and heavy-duty vehicle class ac 3,500 pounds
GVWR. Thus, some o£ che Class II crucks will be ineLuded vich all
of chose in Classes III chrough VIII in che heavy-ducy vehicle
class. For purposes of che regulacory analysis Class ILA vill cover
GVWR's from 6,001 co 3,500 lbs. and Class IIB will cover 3,501 co
10,000 lbs. GVWR. In 1973, EPA estimated chat only about 5 percent
of chose crucks in weighc Classes I and II have gross vehicle
weights in excess of 3,500 pounds.21 The percentage appears co be
aomevhac higher today based on 1977 GM, Ford and Chrysler produc-
tion daca. A value of 5.5Z will be used in chis discussion. Table
III-A gives che U.S. domestic factory sales plus imports from
Canada of all crucks and buses for years 1972 chru 1978.
E£eavy-ducy crucks represent a heterogeneous class of vehicles,
in cerms of use and functional characteristics. While light-duty
crucks are used by-and-Large for personal transportation and
agriculture, heavy-ducy crucks are almosc exclusively used for
commercial purposes. The 1972 Census of Transportation conducted
¦13-
-------
Table lll-A
U.S. Trucicu and Buueu by CVUK (poundu)
(U.S. Domestic factory Sale* plus importu from Canada)
Year
»-*
8,500
8,501-
to.ouo
10,001-
14,000
14,001-
16,000
16,001-
19,500
19,501-
26,000
26,001-
33,000
33,000
and over
Yearly
Tot aid
1978
3,2111,772
Itt7,336
34,014
5,959
3,982
157.I6B
41,516
163,836
3,812,583
1977
2,972,752
173,017
30.064
3,231
4,9a9
160,396
32,249
148,728
3,525,426
1976
2,525,755
147,002
43,411
67
a,920
149,293
22,91fi
103,098
3,000,466
1975
1,790,355
104.201
19.497
6,50a
13,916
152,070
24,69a
74,896
2,186,141
1974
2,088.200
121,535
a.916
a, 120
24,366
215,221
32,3b4
160,465
2,659.187
1973
2,370,208
137,949
52,558
a, 744
37,043
199,481
40,816
155,814
3,002,613
1972
1,929,aa3
112,321
57,1103
10,353
37,492
177,723
40,150
130,328
2,496,054
*
6,001
The HVMA does
-10,000 pound
not aplit nalea
clasaea. The it
dt a, 500 pounds CVWK, but raclier
plit in the table repreaenta EPA*
publiahea
a eatiwate.
aalda for
the 0-6,000
and the
Total Vehicles Subject
to III) lieu
nlat iona
1978 593,an
1977 552.674
197b 474,7u9
1975 395,786
1974 570,987
19 73 632,405
19 72 566.170
Source: fr'ii-3, HVMA data.
-------
by Che Department of Commerce indicates that trucks are used La
agriculture, construction, mining, wholesale and retail trade,
manufacturing, and Lumbering and forestry, as veil as by the
utility, service and "for hire" industries. Most functional
applications of tIDVs are ooc readily transferable to other trans-
portacion node3 such as air, raiL, water, or pipeline.
As Table III-B shows, the uses of heavy-duty vehicles vary
with gross vehicle weight. For the lighter trucks, chose in the
8,500-20,000 pound GVWR range, we find chat the primary applica-
tions are in che agriculture, construction, services, and wholesale
and retail trade markets, where the trucks are generally used for
pickup and delivery. Personal use of trucks in this cacegory,
while limited, consists primarily of operation of motor homes built
on cruck chassis. Some people also use "heavy" pickup trucks for
personal transportation.
SDVa in the 20,001 - 26,000 pound GVWR range find uses in che
agriculture, construction,and wholesale and retail trade aarkets.
Forestry, Lumbering,and manufacturing account for most of che other
applications.
The heavier trucks (26,001 pounds GWR and over) are primarily
found in che construction, wholesale and retail trade, and "for
hire" markets. Hhile the auaber of trucks used for mining and
manufacturing is aot Large, chese markets use che heavy crucks
extensively. Trucks in this category are used only Co a Limited
extent in che other market sectors.
Since che ultimate goal of che various commercial enterprises
chat use heavy trucks is to make a profit, trucks operated by chese
businesses are designed specifically Co meet particular functional
needs in an economical manner. Thus, the heavy-duty vehicles
produced for the U.S. market are often "custom" builc to satisfy
requirements of che operational environment faced by che ultimate
user. This operational environment might be defined in terms of
economic variables (i.e., operating costs of alternative means of
transport, value of products to be transported, operating costs of
alternative types of trucks) or operational variables (i.e.,
distances co be travelled, qualities of che Load co be transported,
types of shipping procedures co be ucilized, state and federal
regulations on truck use, safety, operation).
Buses equipped with heavy-duty engines are usually in che
19,501 - 26,000 pound GWR (Class 71) cacegory. (Jses of buses
include school transportation as veLl as intercity and transit
passenger service. Most school-type buses are gasoline-fueled,
the remainder being diesels.
By defining their operating environment, users of heavy-duty
vehicles can cell vehicle manufacturers exactly what characteris-
-20-
-------
Table III-B
Trucks: Percent Distribution of Size
Classes by Vehicle and Operational Characteristic: 1972
Characteristic
MAJOR USE
Number
(Thousands)
Percent
10,000
Or Less
Lbs. GVW
10,001-
20,000
Lbs. GVW
20,001-
26,000
Lbs. GVW
26,001
Or More
Lbs. GVW
Agriculture
4,258
21.62
20.12
32.12
33.22
10.32
Forestry and Lumbering
187
1.0
.5
1.4
2.3
3.6
Mining
77
.4
.2
.6
.7
1.9
Construction
1,693
8.6
6.9
10.2
14.0
19.1
Manufacturing
443
2.3
1.3
3.3
4.4
8.5
Wholesale and Retail Trade
1,875
9.5
6.1
18.9
23.0
13.3
For ttire
770
3.9
.6
6.0
7.2
30.6
Personal Transportation
8,122
41.2
53.4
11.0
2.1
1.0
Utilities
505
2.6
. 2.5
3.1
3.8
1.9
Services
1,409
7.6
7.7
10.5
6.0
2.5
411 Other
327
1.7
1.2
3.5
3.4
2.8
BODY TYPE
Pickup, Panel,
14,464
73.32
92.62
31.32
4.42
2.17.
Multi-Stop,or Walk-m
Platform
1,645
a.4
2.2
27.4
28.9
21.0
Platform v/ Added Device
336
1.8
.4
5.6
7.0
4.4
Cattlerack
479
2.5
1.4
6.7
6.7
2.4
Insulated Nonrefrigerated Van
96
.5
.1
1.2
1.2
3.1
Insulated Refrigerated Van
. 178
1.0
.1
2.4
2.3
5.3
Furniture Van
192
1.0
3.7
2.8
3.2
Open Top Van
58
.3
.1
.6
.4
1.9
All Other Vans
610
3.1
.7
6.3
7.2
13.6
Beverage Truck
87
.5
.1
1.4
3.0
1.5
Utility Truck
370
1.9
1.7
3.4
2.0
.9
Garbage and Refuse Collector
69
.4
.1
1.3
1.4
1.2
Winch or Crane
83
.5
.1
.8
3.5
1.3
Wrecker
115
.6
.3
2.3
.6
.2
Pole and Logging
53
.3
.1
.3
1.4
2.4
Auco Transport
30
.2
.1
.2
.1
1.4
Dump Truck
468
2.4
.3
3.1
17.3
14.0
Tank Truck for Liquids
287
1.5
.1
2.3
9.7
9.1
Tank Truck for Dry Bulk
29
.2
.1
.6
1.5
Concrete Mixer
66
.4
.1
.2
.1
4.1
All Other
33
.2
.1
.6
. 6
ANNUAL MILES
< 5,000
4,621
23.51
22.02
33.22
35.82
12.72
5 - 9,999
5,540
28.1
30.2
25.6
25.2
13.8
10-19,999
6,598
33.5
36.2
27.8
24.0
22.4
20-29,999
1,647
8.4
8.1
3.1
3.3
11.5
30-49,999
772
4.0
2.9
4.1
4.9
13.4
50-74,999
270
1.4
.5
.9
1.5
11.5
> 75,000
300
1.5
.4
.6
• *
15.1
Total Percent
100.OZ
100.02
100.02
100.02
100.02
Tocal Trucks
19,745
14,598
2,822
828
1,500
Source: 1972 Census of Transportation,
U.S. Department of
Commerce.
¦21-
-------
cics cheir cruck should have when ic is completed. Examples of die
design parameters which may be specified include engine cype
(diesel or gasoline-fueled), horsepower, number of cylinders,
displacement, natural aspiration vs. curbocharging, cransmission,
body cype (single unic, or combination), gross vehicle weight,
ma-ritimtw Load weight, vehicle length, number of axles, axle ar-
rangement, distance between tandem axles, and tire size.
3. 3eavy-Duty Vehicle Engines
One of the basic parameters chat heavy-duty vehicle users oiust
consider in determining what vehicle chey need is che cype of power
plane chey will use. . Soth diesel and gasoline-fueled engines are
used co power heavy crucks and buses. • Some tradeoffs chat vehicle
purchasers consider in che selection of diesel or gasoline-fueled
engines for cheir vehicles follow:
Diesel Engines
1. Diesel fuel coses less chan gasoline.
2. Diesels get up co cwice che fuel mileage of comparable
gasoline-fueled engines.
3. Diesels require less maintenance.
4. Diesels are generally more durable chan gasoline-fueled
engines and are often rebuilc.
5. The. diesel rebuild interval (250,000-300,000) miles is
up co chree cimes longer chan gasoline-fueled engines
(100,000 - 125,000).
6. Diesels have higher resale values.
Gasoline-Fueled Engines
1. Gasoline is more readily available in mosc areas.
2. Gasoline-fueled engines generally start more easily in
cold weather and give better overall performance.
3. Gasoline-fueled engine service and part3 are more readily
available.
4. Gasoline-fueled engines weigh less.
5. Gasoline-fueled engines cost considerably less chan
comparable diesel engines (about one-third as much).
The lighter "rucks, classes LIB - VI are usually equipped wich
gasoline-fueled engines, as shown in Table LII-C.
The heavier trucks (Classes VII and VIII) are equipped wich
diesel engines, as shown in Tables III-D and III-E. However, as
fuel economy and fuel costs become more important co cruck opera-
tors, diesel engines will become more popular in 3ome of che
lighter cruck classes. Diesel engines are more fuel efficienc
and diesel fuel is cheaper chan gasoline.
-22-
-------
Table III-C
Gasoline Engine Usage in geavy-Oucy Vehiclea
8,501 10,001- 14,001- 16,001- 19,501- 26,001- 33,001- Yearly
Year 10,000 14,000 16,000 19,500 26,000 33,000 and over Tocala
1978 187,336 34,014 5,959 3,982 144,923 15,597 7,160 398*971
1977 173,017 30,064 3,231 4,989 149,254 13,526 6,005 380,080
1976 147,002 43,411 67 8,920 143,077 11,597 5,561 359,635
1975 104,201 19,497 6,508 13,757 147,267 13,509 8,74a 313,487
1974 121,535 8,916 8,120 24,325 211,861 19,382 19,138 413,277
1973 137,949 52,558 8,443 37,037 195,741 22,587 17,473 471,793
1972 112,321 57,803 10,138 37,467 174,019 27,482 13,855 433,105
Source: Table IIIA and Table IIID.
-23-
-------
Table 111-0
Ojeael Daage ia Heavy-Ducy Vehicles
8,501 10,001- 14,001- 15,001- 19,501- 26,001- 33,001- fearly
Year 10,000 14,000 16,000 19,300 26,000 33,000 and over Tucala
1978 — — — — 12,245 25,919- 156,676 194,340
1977 — — —¦ — 11,142 1.8,723 142,723 172,538
1976 — — — — 6,216 11,321 97,537 114,394
1975 — — — 159 4,503 11,139 66,143 32,299
1974 — — — 41 3,360 12,982 141,327 157,710
1973 — — 296 6 3,740 18,229 138,341 160,612
1972 — — 215 5 3,704 12,663 116,473 133,065
Sourca: FS-5, S1VMA data.
-24
-------
Table III-E
Diesels Factory Sales as a ?ercencage of
All Seavy-Oucy Vehicle Faccory Sales
8,501- 10,001- 14,001- 16,001- 19,501- 26,001- 33,001
Year 10,000 14,000 16,000 19,500 26,000 33,000 and over
1978 — — — — 82 621 962
1977 — — — — 72 582 962
1976 — — — — kZ 492 942
1975 — — — 12 32 452 382
1974 — — — — 22 402 382
1973 _ — 32 — 22 452 392
1972 — — 22 — 22 322 392
Source: tables III-A and III-0.
aii an
Vehicles
322
312
242
212
282
262
242
¦25'
-------
Manufacturers can. effectively boose che power of boch gaso-
line-fueled and diesel engines through curbocharging, chough che
first coat of che engine suffers somewhat. Secauae che avail-
ability of Curbocharged engines is a further consideration of che
prospective buyer/user, we have included a brief description.
A curbocharger combines a turbine, driven by engine exhaust
gases, with a compressor which increases che air flow into che
engine combustion chambers. Increasing che amount of air entering
Che combustion chambers permits more fuel co be injected and
cherefore mora power is generated per piston stroke. In addition
Co generacing sore power chan a naturally aspiracad engine, curbo-
chargers' improve che fuel economy and emission characteristics of
che engine. the increased air flow into che combuscion chamber
increases the inlet air pressure and inlet air temperacure. This
results in more complece combustion of che air/fuel mixture,
particularly under cruise conditions. Fuel economy is improved and
emissions of EC and 00 are reduced. DOT and EPA have estimated
chac a 0 co 7 percenc fuel economy improvement can be gained
by che use of curbochargers on heavy-duty diesel engines.3/ Though
curbocharger units can cost from 3300 co 31,200, this higher first
cost is soon paid back chrough lower operating costs.4/
C. Manufacturers
Unlike che automobile induscry in which che vehicle manufac-
turer and che engine manufacturer are one.and the same, heavy-duty
vehicles and che engines used in chem are often manufactured by
independent companies. A. single vehicle manufacturer may, in face,
use engines produced by several different companies. Even vehicle
manufacturers chat produce cheir own engines may use anocher
company's engine in che vehicles chey produce.
To simplify chis discussion of producers of domestically sold
heavy-duty crucks, vehicle manufacturers and engine manufacturers
will be considered' here separately. As an aid co che reader, che
list beiow is provided giving, che names of most of che manufac-
turers in che heavy-ducy vehicle induscry and cheir product(s}.
Summary financial and non-finaneiai information on many of these
companies can be found in Table
Manufacturers
Engines
(G-Gasoliae)
(D-Oiesel)
Vehicles
Chrysler
Ford
General Motors
IHC
Mack Trucks
Mercedes-3enz
C
c
C,D
G,D
~
0
(Chrysler, Dodge)
X
(Chevrolec-GJlC)
X
(Mack-3rockway)
1
-2S-
-------
Manufacturers
Engines
Vehicles
VoLvo
White Engines, lac
Caterpillar
0
D
0
D
D
0
0
D
0
0
D
D
X
Cummins
Qeutz
Nissan
Fiat
Hino
Isuzu
Mitsubishi
Scania Vahis
Perkins
Freightliner
Peterbilc
Kanworch
FWD
Oshkosh
Whice Mocors
X
X
X
X
X
X
1. Engine Manufacturers
Manufacturers of engines used in heavy-ducy trucks typically
fall into one of two categories, those chat produce gasoline-fueled
engines and chose that produce diesel engines. Two companies;
General Mocors and International Harvester, produce bcch gasoline-
cue Led and dieseL engines for use in on-road heavy-duty vehicles.
The manufacturers of gasoLine-fueled engines and the engines
chey certified in 1979 are Listed ia Table Ill-e. All these
manufacturers are domestically based and all produce their own Line
of heavy-duty true it a or buses. General Motors (Gil), Ford, and
Chrysler are perhaps most widely known as producers of Light-
duty passenger cars since it is from chat Line of business chat
chey derive most or cheir revenue. However, all produce light-duty
trucks and heavy-duty vehicles in addition co gasoline-fueled
heavy-duty engines.
A. fourth company producing gasoline-fueled engines for trucks
sold in the U.S. is the International Harvester Company (IHC).
Like the other three, IHC produces complete vehicles as well as
gasoLine engines — but ao passenger cars. Their concentration is
in Che heavy-duty truck market wieh some emphasis on light-duty
trucks. IHC also makes off-che—road vehicles for construction and
industry and faza equipment.
Using past sales data supplied- by the manufacturers EPA has
estimated each manufacturers market share. Assuming ao gasoliae-
fueled engine sales are made to other manufacturers the following
marke t share s r esu11:
-27-
-------
Table III-F
Manufacturers of Gasoline-Fueled Engines
for Dae i.a Eeavy-Qucv Vehicles
Manufacturer Engine Families Diaplacemeaca AvailabLe (CID)
Chrysler 2 360, 440
Ford 6 300, 351, 370, 400
429, 460, 475, 534
GM 4 292, 350, 366, 427, 454
IHC 4 345, 391, 400, 447, 53 7
Federal Register Vol. 44, No. 140, Pare III, July 19, 1979.
-28-
-------
GSi
Ford
Chrysler
IHC
432
292
162
122
These market shares are only an escimate since some engines
are indeed sold to other manufacturers and also co recreational
vehicle manufacturers.
As we curn co a discussion or diesels, it Ls important co
realize chat cheir manufacture and sale is accomplished by a
different sec of manufacturers than chose involved in che produc-
tion of gasoline engines. Only GM, via its subsidiary Decroic
Diesel Allison, and International Harvester manufacture boch engine
types in significant quantities. Together cheir production of
diesels accounts for something less chan 352 of che coca! produced.
The leading producer of dieseI engines used in che U.S. trucks is.
che Cummins Engine Company, followed by Decroic Diesel (GM),
Caterpillar, Mack True its, and IHC. A lisc of che engines, made by
these companies and several ochers i.3 given in Table III-G. Table
III-R presents a distribution, by manufacturer of diesel engines
used in U.S.-made trucks.
Like gasoline engines, most diesels are produced by domestic
companies. Detroit Diesel Allison and Perkins engine are subsid-
iaries of GM and Massey-Ferguson, LTD, respectively. Detroit
Diesel 3ells both dieseL engines and aircraft engines. In addition
to Perkins' sales of diesel engines, Ma3sey-Ferguson also produces
agricultural, industrial and construction machinery, and recreation
products.
Several of che other manufacturers of diesel engines, like
Massey-Ferguson, make off-Che-road vehicles. Caterpillar13 product
line includes construction, warehouse, agricultural, logging and
petroleum equipment, accounting for 902 of total sales.
Mack Trucks produces diesel engines and the on-road cruck3
Chat use them. Mack is a subsidiary of Signal Companies, Inc.,
¦whose business includes aerospace and industrial equipment, petro-
leum and pecrachemical products, and construction and fabricated
products.
Cummins Engine Company is the leading producer of heavy-duty
diesel engines with about 302 of che market. Cummins is unique in
Chat it does not manufacture any vehicles, either on-road or
off-road; nearly all of its sales are engines. Cummins also
produces and markecs crankshafts, curoocharger3, and related
components.
-29-
-------
table EII-G
Manufacturers af Diesel-Fueled Engines
cor Ose ia Heavy-Out7 Vehicles
Manufacturer Engine Families Displacements Available (CIS)
Cacerpillar
11
636,
638,
893,
Cumins
10
555,
855,
903,
Deucz
1
2SS
G4 (ODA)
9
212,
426,
552,
Hino
1
393
IHC
3
466,
549
Isuzu
2
235,
353
Iveco-Fiat
2
494,
534
Mack
4
672,
998
Mercedes 3eaz
3
346,
589
Mitsubishi
1
243
Scania Vabis
1
475
Volvo
3
334,
409,
526
White Engines
I
473
Source: Federal Register Vol. 44, No. 140, Part III., July 19,
1979.
-30-
-------
2. Vehicle Manufacturers
Ic i3 che vehicle manufacturers who combine cheir owa or
someone else's engine with a chassis co fabricate che final product
aeedad by che heavy-duty vehicle user. This final product is a
bus, a single unit truck, or a tractor for pulling crailer units.
tables III-I and III-J show che domestic factory sales numbers
for crucks and buses respectively during 1978. It is clear chac
some firms concentrate on producing crocks of a certain weight
class while others produce che entire spectrum. In 1978 trucks
builc by Ford, <31 (Chevrolet and QIC),' Chrysler (Dodge), and LHC
appeared in nearly every class. GM and Ford dominated che market
in almosc every cacegory and accounted for 39Z and 37Z respectively
of cotal heavy-duty sales. Along with Chrysler (Dodge) and LHC,
chey produced all but a few of che vehicles with GTVSs below 26,000
pounds. Hose of each of chese manufacturer's crucks are gasoline-
powered, using cheir own engines.
International larveseer is che largest producer of Class VII
and VIII (26,000 pounds GVW and above) vehicles, and overall is
fourth (behind Ql, Ford, and Chrysler) in che production o£ heavy-
duty trucks. As aoced earlier, IHC also produces both gasoline and
diesel engines.
The rest of che heavy-duty vehicle manufacturing industry
consists of firms which account for less chan five percent of cocal
cruck production. These firms concentrate on che production of che
"heavy heavies", che Group 7III crucks (33,000 pounds GVW and over)
chat are used primarily for long haul work. FWD is a privately-
owned company specializing in che production of custom buiit trucks
used primarily by owner-operators. FWD produces an expensive
cruck package chac is custom builc co che buyer's specifications
and produced in limited quantities. Mack, as mentioned in che
heavy-duty engine manufacturers description, produces heavy-duty
engines and Class VIII heavy-ducy trucks. The White Mocor Corpor-
ation is represented by "White", "Autocar" and "White Western Scar"
while also producing agricultural and construction equipaenc.
Other manufacturers of Groups 711 and VIII heavy trucks include
Kenvorth (PACCAA) and Pecerbilt (PACCAfi). These manufacturers
specialize in custom-built HDs which are primarily procured by
individual owner-operators (as opposed co fleecs).
In contrast co che production of heavy-duty trucks, bus
manufacturing is limited co only larger companies in che transpor-
tation manufacturing industry. As one can see from Table III-J,
IHC, <31 and Ford are che primary producers of intercity, transit,
and school bus chassis Ln chis country. For none of chese com-
panies is che sale of buses critical co che financial success of
che firm.
-31-
-------
Table lll-U
Oieiel engines Used In Trucks
Frou U.S. - 1970
Diesel Engine Manufacturer
Veli ic le
Scaaii
Hfr.
Cat.
Cusnial
CH (DIM)
IUC
Mack
Hitaubiwhi
NiftflAQ
Old*
Perkina
Vafaia
Tuttti
Chevrolet
1,236
699
4,129
--
24,004
30,3411
Cliryaler
—
446
12 .
—
—
1,404
—
4,685
ford
22.148
11.805
1,861
—
—
—
—
—
—
41,814
friiigtit 1 tottr
709
9,086
2,078
—
—
—
—
—
11,873
CHC
1,4 30
3,016
16,112
—
—
—
6,172
—
—
30,930
iuc
1,912
23,086
8,190
12,916
—
990
—
—
—
49,114
Keuwortli
2,160
9.196
2,810
—
--
—
—
—
—
14,426
Hack
211
2,412
426
—
30.865
—
—
—
—
596
34,596
Purcerbi It
2,166
3,119
1,603
—
—
—
—
—
—
—
9,550
Mi ice
913
9,143
1,115
—
—
—
—
—
—
—
14,415
Others
411
148
916
—
—
—
—
—
—
—
2,097
Tut al
11,586
80,018
48,194
12.916
30,065
2,763
990
32,256
1,404
396
243,848
Source: 1978 HVHA Pata.
-------
Table IIl-l
U.S. Truck
Salea by
Make and
CVW Claua,
1978
6,001-
10,001-
14,001-
16,001-
19,501-
26,001-
33,001
10,000
14.000
16,000
19,500
26,000
33,000
& over
Total
Ford
946,934
69B
__
137
55,899
15,531
22,293
1,041,492
Chevrolet
801,950
—
—
1,854
34,340
1,662
3,999
843,805
Dodge
322,658
52,643
15,250
85
207
17
—
390,860
IIIC
36,065
—
8
—
29,964
16,494
34,051
116,574
CMC
215,185
1,540
—
822
24,898
3,925
14,460
260,830
Mack
—
—
—
—
—
2 7,390
27,390
Kenworth
—
—
—
—
—
—
14,345
14,345
freight 1 iner
—
—
—
—
—
—
II ,725
11,725
PeterbiIt
—
—
—
—
—
—
9,454
9,454
White
—
—
—
—
—
437
10,1B9
10,626
Jeep
78,326
—
—
—
—
—
—
78,326
Brockway
—
—
—
—
—
—
37
37
Weatern Star
—
—
—
—
—
—
786
786
Autocar
—
—
—
—
—
— .
1,880
1,880
fc'WD
—
—
—
—
—
—
145
146
Miac.
6
93
—
35
1,705
929
2,087
4,855
Mine, includes Imports, Diamond Reo, Uivco, llendrickaon, Gdlikoali, etc.
Source; Automotive Newa Market Data Book, April 25, 197B.
-------
Table III-J
1978 U.S. Sua Sales
(Including School Bua Chaaaia)
8,500- 10,001- 14,001- 16,001- 19,501- 26,001- 33,001
10,000 14,000 16,000 19,500 26,000 33,000 & Over local
Chevrolet — — — — 4,430 — — 4,430
QIC — — — — 2,397 218 1,049 3,664
Ford — — — — 7,007 — — 7,007
IHC — — — — 13,968 62 — 14,030
A£l/G«nerai — — — — — — 1,036 1,036
Others — — — — 173 235 303 1,211
TOTAL .— — — — 27,975 515 2,888 31,378
Source: FS-3, 1978 MVMA daca.
-34'
-------
A brief Look at che employment picture in che industry shows
chat 763 manufacturers of truck and bus bodies (including light-
ducy crocks) employed 40,796 people in 1976, and che 292 firms
building cruck crailers employed 20,697. (Table III-Q gives che
number of employees in che major vehicle and engine manufacturers.)
D. Users of geaw-Outv Vehicles
As Section A of chi3 chapter notes, most heavy-duty vehicles
are used for commercial puposes. The eypes of trucks used co meet
che transportation needs of various enterprises are as diverse as
che needs chems elves. Basically, however, chese crucks move some
commodity from one point co anocher.
TabLe III-K lists some of che cypes of products moved by
crucks and other means of cransport and che percentage (by weight)
chat each means of cransport carries. Though che data L3 somewhat
outdated, it is interesting co see che fractional distribution of
freight and how it is transported. As of 1972 nearLy half of che
commodities lisced were shipped ay cruck, and In 1977, crucks
carried almost 25Z of all intercicy freight.5/
Trucking can be divided into cwo cypes of carriers, local and
intercity. The rule of thumb is chat local carriers are those who
conduct 50Z or more of cheir business in a metropolitan area. The
intercity (line haul or over-che-road) carriers conduce Local
pickup and delivery between metropolitan areas. Local carriers
accounted for S67.5 billion in freighc transportation expenses and
intercity carriers 567.3 billion Ln 1978.5/
Another way of examining the trucking industry is co distin-
guish between private ownership and "for hire" crucking. The
crucks in "private" fleets are under che control of each particular
company far the shipment af their own goods, crucking not being
cheir principle business. Sxamples of "private" cruck owners are
che various ucilicy companies (e.g., 3eLl Telephone Syscem)
or retail stores that own cheir own delivery trucks, and manufac-
turers of consumer products who make deliveries co retail concerns.
In contrast, "for hire" crucks are used by companies or
individual owner/operators whose business it is to transport
someone else's freighc.5/ Examples of firms in chi3 Latter cate-
gory are United Parcel Service, Roadway Express, Consolidated
Freightvays, and che various movers of household goods (United Van
Lines, North American Van Lines, Allied Van Lines). Some com-
panies, Like Hertz and Ryder, are in che business of renting crucks
for use by others.
"Tor hire" crucks accounted for about 4Z of the crucks in use
in 1975. Over fifty percent of chese trucks were combinations
(tractor-trailer) most with five or more axles (see Table III-L).67
-35-
-------
Tabic 1I1-K
Commodities Shipped by Mode of Transport
Tuna Tons/Milea
Motor
Private
Tot al
Motor
Pr ivate
Total
Croup
Carrier
Truck
Truck
Kai 1
Other
Carrier
Truck
Truck
Hai 1
Oilier
Meat & Dairy Products
41.7Z
39. U
80. ax
18.82
.4*
54. 3*
17 . 2t
71,5Z
27.82
.63
Canned, frozen & Other
20. 3
23.0
43.3
50.7
6.0
18. 3
9.5
27.8
66.8
5.4
Food Product a
Candy, Cook it; a, beverages
25.7
58.4
84.1
15.4
.4
28.8
25.8
54.6
43.1
2.2
Tobacco Products
basic Textiles & Leather
61.4
27.7
89.1
9.7
1.2
61.0
21.0
82.0
16.1
1.8
ProducL a
Apparul & Related Products
69.4
15.6
85.0
8.5
6.5
67.0
9.5
76.5
13.4
10.1
Paper & Allied Products
28.0
17.9
45.9
51.7
2.3
18.9
5.6
24.5
73.8
1.5
llauic Cheaii c a la, P i aat ica,
30.1
12.1
42.2
48.6
9.2
21 .6
4.7
26.3
63.1
10.5
Synthetic Hubber & fibers
Drugs , Paint a & Other
38.6
15.7
54.3
37.8
7.9
32.0
8.4
40.4
44.3
15.2
Chemical Products
Petroleum & Cog I Products
16.0
8.4
24.4
9.7
65.8
3.4
1.6
5.0
7.9
87.1
Itubber & Plastic Products
59.1
15.2
74.3
24.4
1.2
56.8
9.3
66.1
32.1
1.8
1.umber & Wood Products,
16.2
36.3
52.5
45.8
1.6
7.6
10.7
18. 3
76.8
4.9
Except furniture
Furniture & fixtures
41.4
34.7
76.1
22.0
1.9
39.9
20.5
60.4
37.1
2.5
Stone, Clay & Claaa
47.2
23.7
70.9
21.9
7.2
36.6
11.3
47.9
45.3
6.7
Product a
Primary Iron & Steel
44.4
6.7
51.1
43.7
5.2
35.9
4.8
40.7
51.6
7.7
Products
Primary Nonfcrrous Hetal
31.4
15.1
46.5
51.6
1.9
23.4
7.7
31. 1
67.2
1.6
Product s
fabricated Hetal Product a
55.3
25. 1
80.4
17.3
2.3
60.1
13.0
73.1
23. 3
3.6
Mutal Cans & Miuc. Metal
44. 1
17.8
61.9
36.8
1.3
40. 3
7.1
47.4
50.5
2. 1
Produc t a
Industrial Machinery,
59.4
18.9
78. 3
19.6
2.0
75.7
a.y
84.6
12.3
3.0
Except Electrical
Machinery, fcixcepL Elec-
53.4
17.7
71.1
26.5
2.3
49. 7
8.9
58.6
37.7
3.6
trical and Industrial
Coiiiiiiun icul ton Product a
64.5
12.4
76.9
13.0
10.0
59.9
5.6
65.5
18.0
16.5
& Parts
-------
Table 11 IK (Cont'd)
Coininodi t ien Shipped by Hit tit; of Traimport
Tons Tuna/Hi leu
Motor
Private
Total
Motor
Pr ivate
Total
Group
Carrier
Truck
Truck
Hail
Other
Carrier
Truck
Truck
Hail
other
Electrical Products
49.3
14.1
63.4
35.0
1.3
46.0
8.4
54.4
43.2
2.6
& Supplied
Motor Vehicle* &
37.3
3.0
40.3
59.3
.4
17.4
1.0
18.4
80.9
.8
Equipment
TransporUtimi Equip-
23.9
54. a
78. 7
19.5
1.8
30.3
43.1
73.4
24.0
2.7
ment Except Vehicles
Instrument!!, Photo
63.8
10.9
74.7
20.9
4.4
53.9
5.7
59.6
34.4
6.0
Equipment Wutcheu &
Clocku
TOTAL ALL SNIPPER CROUPS
31.12
18.32
49.42
31.72
18.82
20.92
6.82
27.72
42.02
30.33
Total all Shipper Croupu
Except Petroleum and Coal
35.72
21.32
57.02
3d.42
4.52
28.62
9.12
37.72
5b. 92
5.42
Source; Hot or Vehicle Facta and Figures, 1976
Data from 1972 Couuodity Transportation Survey - U.S. bureau of Cenuu*.
-------
Table III-L
"Tor Hire" Trucks In Use ( 1975)
Siagle-Unit Trucks
2 Axles 378,345 39.4
3 Axles 43,276 4.6
Subcocal 4^2,121 44.0
Combinacion Trucks
3
Axles
70,131
7.3
4
Axles
145,399
15.2
5
or snore
321,499
33.5
Subcocal
537,579
56.0
Total Trucks for 'dire
959,700
100.0
Tocal Trucks la Use
22,643,008
Z Trucks Used for Sire
4.062
Source: Transportation Energy Conservation Data Book, Edition 3,
February 1979, Oak Ridge tfacicnal Laboratory, Table 1.26.
-38-
-------
To remain competitive with alternative means of transport,
intercity carriers woric on a small margin over coses. Coses for
drivers are about 30Z of che Cotal. Coses of equipment account for
another 9.02-of che eoeal and operacing costs (fuel/maincanance)
about 112. In 1974 chere were approximately 2,300 Class I and II
motor carriers. Finally, employment in the trucking industry
amounted co 9,052,000 people in 1973 (ATA estimate).
Heavy-ducy engine exhaust emission regulations will, of
course, also apply co buses. As an example of how chis segment of
che vehicle population is made up, in 1977 chere were abouc 20,000
buses being operated in che U.S. by 1030 intercity cransit bus
companies, employing abouc 44,000 people. There were also 48,700
buses being operated by local cransit companies. Most of chese
cransit buses are equipped with diesel engines. School buses,
however, accounc for che overwhelming number of buses on che
roads. Ia 1977 over 298,300 publicly- and privately-owned school
buses were ia operation. They accounted for over 30 percent of all
buses on che road and were nearly alL gasoline-powered.5/
E. The Future of geaw-Outy Vehicles
The next decade is sure co bring changes in che heavy-ducy
vehicle Lnduscry. Changes in GN? and weight and length restric-
tions may cend co slow che rate of growth of che heavy-ducy vehicle
fleet. Increasing real fuel coses will certainly lead co further
development and utilization of che efficienc diesel engines.
Alchough precise predictions are impossible, che discussion which
follows addresses some of che major factors which will affect che
size and composition of che heavy-ducy vehicle fleec in che next
decade.
The GNP growth race is expected co slow in che next decade as
compared to che 1970's in which ic slowed as compared co che
I960's. The main reason is che energy problem. A corollary
of a declining race of growth in GUP is a declining race of growth
in commercial freight and therefore, a lesser growth rate in sales
of heavy-ducy vehicles co move chat freight.
Another area of change which will affecc ehe sales of heavy-
ducy vehicles in che next decade is deregulacion of che trucking
industry. Spurred by che trucking industry, the Federal Govern-
ment, and the fuel crisis, states should continue co move cowards
uniform. weighc and length liaicaeions. This will decrease che
auaiber of miles that trucks have to crave I since many unnecessary
miles are due co che differences in scaee regulations.^/ Trucks
coday go around states where regulations are more restrictive since
chat is cheaper Chan making two crips through che scats co meet
weighc restrictions or having co reload into a shorter trailer co
oeec length restrictions.
39-
-------
Along with uniform regulations, leas strict weight and length
limits may be implemented. Double and criple-crailer rigs can
substantially reduce che gallons of fuel used per con-mile cra-
ve lad. Ic is estimated chac doubling of gross combination weight
results in more than a doubling in fuel efficiency as measured by
ton-miles per gallon of fuel. Of course these weighc and length
restriction changes will continue to be debated in view of safety
and environmental concerns.
Restrictions on return crip loads 3hould be eased. This will
reduce che number of empty backhauling crips and cherefore increase
fuel efficiency.
All of che above regulation changes will tend to decrease che
rata- or growth of che heavy-duty vehicle fleet, trucks will carry
more freight per crip from both a weighc and a volume viewpoint.
Also, che number of miles per crip should decrease due co less
bypassing of overly restrictive states. On che other side of che
future heavy-duty vehicle sales equation is che fact chat heavy-
duty vehicle lifetimes may tend co diminish somewhat since chey
will be doing more work per hour and per mile. This will place
more 3cress on engines and drivetrains resulting in increased wear
and tear. Durability will become increasingly importanc.
The fuel crisis, while being an underlying cause of all of che
above changes, will be a direct cause of che shift from gasoline-
fueled engines co diesel engines. As mentioned previously in chis
chapcer, diesel engines are more fuel efficient chart gasoline-
fueled engines. Coupled with che greater durability of diesels,
che fuel efficiency advantage should continue co increase che
diesel's market share.
The switch co diesels will not be as fast as the mechanical
advantages of diesels would predict. Environmental, social, and
economic concerns will prevent che extremely rapid race of diesel-
ization predicted in some studies.4/,8/ Concern over future
particulate and NOx regulations will prevent manufacturers from
fully commitcng to diesel production until they are confident that
such regulations can be met without adversely affecting che eco-
nomic advantage of che dieseL. As more diesels are put into
use, the diesel fuel shortages may increase co a greater degree
than gasoline fuel shortages. This was demonstrated wich che fuel
shortages in the spring of 1979. The specter of diesel fuel
shortages, may dampen demand. Basic economics predicts chat as
diesel fuel demand increases, its price will increase, which will
also remove some of the diesel advantage. Finally, lack of confi-
dence in diesel cold-start capability and maintenance availability
is still a concern with many prospective owners.
SPA is projecting chac che current growth in heavy-duty
vehicle sales will decrease slightly in the mid-eighties. The
-40-
-------
major change expected is a. shift co dieseis in che heavier weight
classes.
To project total heavy-duty vehicle sales by weight class EPA
used several seeps.
First, the cotal heavy-duty vehicle 3ales by domestic manu-
facturers for the years 1967-1973 were determined from MVMA data.
A. Linear regression through this data gave a sales growth of 10,903
per /ear.
The next seep in this process was to account for imports,
primarily Canadian. The data available to EPA indicated that on a
year co year basis, Canadian imports were mathematically about 10
percent of domestic sales by U.S. manufacturers. So the growth
race was increased by 10 percent to about 12,000 per year.
To apportion che sales across che weight classes, historical
percentages from che period 1974-1978 were used. These percen-
tages, as shown below, are percentages of cotal sales in each
weight class and have been assumed not co change in che mid-19801s.
Table III-M contains che coca! projected haevy-duty vehicle sales
for the period 1984-1988.
To determine how the total heavy-ducy sales in each weighc
class will be divided between gasoLine-fueled engines and diesel
engines, SPA used several Lnpuc sources and in a few cases best
judgment. This methodology will be discussed on a per class basis
in che following paragraphs.
Classes IIB, III, IV. V - Based on data submitted in che
light-duty diasel particulate rulemaking action it appears chat
dieselixation in lighter gross vehicle weight will sot be as great
as in the heavier weighc classes. This slower dieselizacion race
will be caused in part by the larger initial purchase price of che
diesel, slightly poorer performance of dieseis, and less avail-
ability of maintenance for dieseis. Based on the Light-duty diesel
aunmary and analysis of coaoents, our projections wiLL allow 20
percent of the sales in each of these classes to be diesel by 1990
9/ and the percentage of diesel sales to grow at a steady 2 percent
per year for the period 1984-1988 (see Table III-N).
Class VI
Class VII
Class mi
Class 113
Class III
Class IV
Class V
8,501 - 10,000# 28.32
10,001 - 14,000# 5.32
14,001 - 16,000# 0.92
16,001 - 19,500# 2.23
19,501 - 26,000# 32.22
26,001 - 33,000# 5.92
33,001 and over 22.22
-41-
-------
Table III-M
Esciaaced 3DV Sales for
1984 ehrough 1988 by GVWR (pounds)
3,501- 10,001- 14,001- 16,001- 19,501- 26,001- 33,001 All HD
Year 10,000 14,000 16,000 19,500 26,000 33,000 and over Vehicles
1988 192,760 36,100 6,130 14,985 219,324 40,187 171,645 631,131
1987 189,366 35,464 6,022 14,721. 215,462 39,479 168,623 669,137
1986 185,971 34,829 5,915 14,457 211,600 38,771 165,600 657,142
1985 182,577 34,193 5,306 14,193 207,738 38,064 162,5 78 645,149
1984 179,183 33,557 5,698 13,929 203,876 37,356 159,555 633,154
-42'
-------
Table IIIrN
Estimated Diesel Sales as a Percancage of Heavy-Qucy
Sales Eor 1984 through 1988 by GVWR (pounds)
8,501- 10,001- 14,001- 16,QUI- 19,501- 26,001- 33,001 All HD
Year 10,000 14,000 16,000 19,500 26,000 33,000 and over Vehicles
1988 162 162 16Z 162 412 1002 1002 502
1987 142 142 142 142 382 962 1002 482
1986 122 122 122 122 342 922 1002 462
1985 102 102 102 102 312 392 1002 442
1984 82 82 32 82 282 852 1002 422
-43-
-------
Claaa 71 - This class is comprised primarily of "medium-duty"
cracks and school buses. Historically, school buses are abouc IS
percent of che 3ales each year. la an article published in "Fleet
Specialist" magazine, several Manufacturers estimated Class 71
diesel/truck sales in the mid-eighties. The manufacturers esti-
mated chat in 1983 becween 35 and 50 percent of Class 71 truck
sales would be diesel.10/ Currently only about 10 percent are
diesel. Our analysis will conversatively use Che 33 percent
figure. Dieseliaation in school buses is difficult to estimate.
Significant growth ia school bus sales is not expected due co
declining school enrollments and the almost complete implementation
of "court ordered busing." Most school buses do not accumulate
enough mileage on a daily basis, and thus enough fuel savings, co
fully justify the increased initial cose of a diesel engine.
Lacking a more specific estimate, besc judgment dictates chat by
1990 about 10 percent of all school bus sales will have dieael
engines.
Clasa 711 - la 1978 Class 711 sales were more Chan 61 percent
diesel with dieselisation in this clasa increasing rapidly over the
past five years. Sales in chis weight class are expected co become
mostly diesel in the mid-eighties. Based on historical ratios,
this clasa should be almost all diesel by 1988.
Clasa Till - Sales in Class Till were over 96 percent diaael
in 1978. 3ased on che recant history in chis class, chese sales
will all be diesel by 1984 or earlier. It is' reasonable chat by
1984 all Class Till sales will be diesel.
Using che data in TabLe II1-M, and che criteria in che discus-
sions above, Tables 111-0 and 111-? concain che estimated sales
split by weight class between gasoline and diesel engines for che
period 1984-1988.
Based on chis analysis, the major changes expected give an
overall heavy-duty saies growth of abouc 1.8Z per year over che 5
year period (1984-1988). The increased diese 1 izacion in all
classes will actually lead co a decrease of about 2 percent per
year in gasoline-fueled engine sales and an increase of about
6.4 percent per year in diesel engine sales over the 5-year period.
-44.
-------
Table III-O
Eaciaaced Diesel Usage in Heavy-Oucy Vehicles
for 1984 through 1988 by GVWR (pounds)
3,501- 10,001- 14,001- 16,001- 19,501- 26,001- 33,001 ail SD
Year 10,000 14,000 16,000 19,500 26,000 33,000 and over Vehicles
1988 30,841 5,776 981 2,398 89,755 40,187 171,645 341,583
1987 26,511 4,965 843 2,061 80,984 37,979 168,623 321,966
1986 22,316 4,179 710 1,735 72,490 35,824 165,600 302,854
1985 18,258 3,419 581 1,419 64,275 33,725 162,578 284,255
'984 14,335 2,685 456 1,114 56,338 31,678 159,555 266,161
-45-
-------
Table III-P
Saciaaced Gasoline-Fueled Usage in Beavy-Oucy Vehicles
for 1984 through 1988 by GVWR (pounds)
8,501- 10,001-
Year 10.000 14,000
1988 161,918 30,324
1987 162,355 30,499
1986 163,654 30,650
1985 164,313 30,774
1984 164,848 30,872
14,001- 16,001- 19,501-
16,000 19,500 26,000
5,149 12,587 129,569
5,179 12,660 134,478
5,204 12,722 139,110
5,225 12,774 143,463
5,242 12,815 147,536
26,001- 33,001 All HD
33,000 and over Vehicles
0 0 339,547
1,500 0 347,171
2,947 0 354,287
4,339 0 360,888
5,678 0 366,991
-46—
-------
Table III-Q
1978 Vehicle and Engine Manufacturer Information
Company
Total Sales (3)
Met Income (3)
No. of Emolovees
American Motors
2,533,430,000
36,690,000
27,517
Caterpillar
7,219,200,000
566,300,000
34,004
Chrysler
16,340,700,000
-204,600,000
157,958
Cmnmina Engine
1,520,750,000
64,400,000
23,298
Ford Motor
42,784,100,000
1,588,900,000
506,531
General Motors
63,221,100,000
3,508,000,000
839,000
International
Harvester
6,664,350,000
136,680,000
95,450
Mack Trucks
1,640,010,000
63,300,000
17,100
White Motor
1,095,710,000
330,000
9,232
-47-
-------
References
y Clean Air Act as Amended, August 1977; 202(b)(3)(C).
11 Based on 1973 CM and .Ford production daca.
2J Panel Report Number 7, Truck and 3us Panel Report, "Scudy
of Pocencial for Motor Vehicle Fuel Economy Improvement," (J.S.
DOT and U.S. EPA, January 10, 1973.
£/ Interagency Scudy of Posc-1980 Goals cor Commercial Motor
Vehicles; U.S. Department of Transportation, Draft Report,
June, 1976.
5/ Motor Vehicle Facta and Figures, 1973 MVMA data.
6/ Transportation Energy Conservation Data Book, Edition 3,
February 1979, Oak Ridge National Laboratory, Table 1.26.
TJ Trucking in 1995, Motor Vehicle Manufacturers Association,
Contract No. LADU 7502-C5.12, June 1975.
8/ The Impact of Future Diesel Emissions on Che Air Quality of
Large Cities; U.S. EPA, Contract Mo. 63-02-2585, February,
1979. Also available as EPA 450/5-79-005.
9/ Summary and Analysis of Comments on the Notice of Proposed
Rulemaking for Light-Duty Diesel Particulate for 1981 and
Later Model Year Vehicles, August 1979, U.S. EPA, CMSAPC,
ECTD.
10/ Fleet Specialist May, June 1979, pp. 31-39.
-48-
-------
CHAPTER 17
ENTISQNMEHTAL IMPACT
A. Background
The Clean Air Acc as amended in L9 70 contained many provi-
sions aimed ac removing harmful pollucancs from che air we breache.
Among ocher things, :he 1970 Acc called for che. establishment of
National Ambient Air Quality Standards. These levels were co be sec
such chac there would be no danger co public health and welfare.
To dace, ambient air quality standards have been sec for five
pollucancs: particulate matter, sulfur dioxide (SO2), carbon
monoxide (CO), nitrogen dioxide (NO^) and ozone (of which hydro*
carbons (HC) is che main precursor). Of these five pollucancs,
mobile sources are major contributors Co che total pollutants
emicted cor three: EC, CO, and MOx. This regulation package
concerns reduction in HC and CO from heavy-duty vehicles. There-
core, che environmental impacc analysis vill not involve NOx.
3och HC and CO emission have been related co adverse health
effects. Oecailed infomax ion on che health affects of HC and CO
will not be discussed in depeh in this Regulatory Analysis since
such information is well documented elsewhere .^1/ Briefly, HC
emissions react with sunlight co form ozone and ocher photochemical
oxidants. Ozone is a pulmonary irritant chac affects the respira-
tory mucous membranes, other lung tissues, and respiratory func-
tions. CO when inhaled replaces oxygen in the blood. The presence
of CO adversely affects che carrying and delivering capacity of
oxygen by che blood.
Although significant improvements have been made ia air
quality since 1970, a review of air quality monitoring data makes
it clear chac additional reductions in 3C and CO emissions will be
necessary if ambient air quality goals set by Congress in che Clean
Air Act are to be achieved. On March 3, 1978, EPA published in che
Federal Register a listing on a State-oy-Scace, pollucanc-by-
pollucant basis, of che accainmenc stacus of every area of che
Nation (^3 FU 3962). This information, compiled by the respective
States and reviewed by SPA, was the most accurate picture available
of the nation's air quality status as of che adoption of the Clean
Air Acc Amendments. These data indicated that of 3215 counties or
councy equivalents covered by chose designations, 607 (19 percent)
were classified as nonactaiumenc foe photochemical oxidant, and 190
(6 percent) were classified as nonaecaiamenc for carbon monoxide.
Nonactainmenc seatus indicates thac the given area fails co meet
che primary national ambient air quality standard (MAAQS) for che
poLlucant under consideration based upon either direct air qualicy
monitoring or indirect estimates for areas lacking monitoring
data. Current non-attainment data Is available co indicate che
-49-
-------
changes which have occured since 1977. As o£ July, 1979, che
non-actainmeac designations include 536 (13 percent) councies for
ozone and 164 (5 percenc) for carbon monoxide.
Since Che U.S. population is not uniformly distributed, buc
racher is concencraced in urbanized areas, che above geographically
based figures are noc represencacive of che proportions of popu-
lacion actually exposed Co excessive ambienc pollutant concentra-
tions. Indeed, ic is che very fact of urbanization which has led
co many of our air pollution problems. For example, che nonactain-
nent areas for ozone include 103 ouc of a cocal of 105 urban
areas in che U.S. wich populations greater chan 200,000 (che
exceptions being Honolulu, Hawaii, and Spokane, Washington). The
103 areas represent an exposure of over 100 million people.
Clearly, there is a great need co reduce pollutant (or pollu-
cant precursor in che case of ozone) emissions in che urban areas
of che U.S. So long as large numbers of people continue co be
exposed co concentrations in excess of che MAAQS, further emission
reductions must be sought.
Mobile sources have been recognized for some cime ¦ as major
sources of hydrocarbons (ozone precursors) and carbon monoxide.
Lighc-duty vehicles in particular have been che focus of consider-
able control work since che lace 1960*3. However, as light-ducy
vehicle emissions grow smaller, other source categories such as
heavy-duty vehicles grow in proportional significance. The wisdom
of controlling heavy duty vehicle emissions is evident when chese
emissions are placed in che context' of other sources of chese 3ame
pollucancs.
In order co properly assess mobile source emissions and cheir
control, ic is best co look ac urban areas where historically che
HAAQS contraventions have occurred. In this way a cruer perspec-
tive of che air quality impact of mobile sources can be obtained.
Ic is in chese urban areas chac improvements are most needed.
The selection of the areas, co analyze will be discussed in decail
below in Section 2. The EC analysis will be done on an A.ir
Quality Control Region (AQCH.)- basis. CO, on che other hand, will
be analyzed on a county basis. This is due co che more localized
nature of CO problems. Fifty seven AQCRs have been selected for
HC, and 52 counties for CO. Hydrocarbons analyzed include only
non-methane hydrocarbons since the methane fraction is non-reac-
tive.
Figures IV-A and IV-B present breakdowns of non-methane
hydrocarbon (NMHC) and CO emissions into various source categories
for che selected regions. These figures give che 1976 emission
levels along wich projected levels out co 1999. The data presented
in chese figures represents what is considered che base case. That
is, ic projects future heavy-duty amissions as if no new regula-
-50-
-------
I
un
M
I
12,000-
10.000-
Non-Methane
Hydrocarbons
(1000 Tons)
8,000-
6.000.
4,000
2,000
13,267
Mobile
Sol ventd
tndustrla
Petroleum
Other
33%
42%
6%
12
ax
1976
\
Figure IV-A
Annual Non-Methane llyrocarbon
Etulsalons for 57 Urban l(£(jlons
LM
28%
44%
VI
I 1%
9%
1980
\\
6,746
N.
252
45%
10%
15%
1985
6,2'16
162
472
12%
~5T
17%
1990
6.29ft
17%
46%
14%
6%
16%
1995
6,511
18%
44%
16%
1%
16%
1999
YEAH
-------
10,130
i
Ul
M
I
ia.000-
\
15,000
Carbon
Monoxide
(1000 Tons)
12,000
9,000-
,,000-J
Mobile
J,000-
ndu&tr la I
Other
842
7%
10%
16,381
B2%
11%
i 976
I9HO
1V-B
Annual Carbon Monoxide
Emissions for 52 Urban CountleB
9,784
\
7,624
7,190
7.202
70%
62%
61%
60%
12%
1U%
19115
15%
23%
1990
16%
23%
1995
16%
24%
1999
VI.- h it
-------
cions beyond chose already ia existence (che 1979 heavy-duty
standards) were promulgated. For ocher source categories, known
fucure concroi programs are included. For example, light-duty
trucks are projecced based upon che 1984 implementation of che
regulations proposed for light-ducy trucks in che July 12, 1979
Federal Register (44FR 40784).2/
For non-methane hydrocarbons, mobile sources currently repre-
sent approximately 30 percent or che urban emissions (Fig. I7-&).
With current regulations chis percentage is expecced co decline co
17 percent by 1995. A.fter chat cioe a gradual increase begins co
sec in.
Mobile source carbon monoxide emissions currently represent
over 30 percent of che urban emissions (Figure 17-3). This amount
is expecced co decline co 60 percent by 1999. So significant
change in stationary source emissions is expecced for CO. However,
since CO problems are often attributed co high localized concentra-
tions during periods of high craffic densicy, stationary sources
have minimal impact on CO air quality problems.
Lighc-duty vehicles (passenger cars) contribute che major
portion of mobile source NMHC and CO emissions. The 1976 emission
levels from light-duty vehicle and other mobile sources, and
projections of che fucure urban amissions are given in Figures I7-C
and 17-0. Again, chese projeccions are for che base case of no aew
heavy-duty regulacions. The figures give a general overview of che
contribution co air pollution chac each class of vehicles is
expecced co make chrough 1995, and of che distribucion of che
burden of control of emissions from all mobile sources. From chese
figures it can be seen chat emissions from heavy-ducy vehicles will
grow in proportion co emissions from lighc-duty crucks and light-
duty vehicles. This apparent inequitable distribution of che
burden for reducing mobile source emissions can be in part account-
ed for by the past need to concentrate control efforts on che
primary sources of mobile source pollution where potential gains
ware the highest.
It is evident fTom che figures chat for both NMHC and CO,
heavy-duty vehicles represent a growing proportion of emissions.
For hydrocarbons, heavy-ducy vehicles go from 12 percent of the
tocal in 1976 to 36 percent in 1999. The increasing share of chese
emissions going Co diesels is also apparent. For carbon monoxide
the figures are 15 percent in 1976 and 43 percent in 1999. Thus,
control of heavy-duty engines is extranely important in any overall
strategy for reducing emissions sufficient to meet ambienc air
quality standards. The remainder of chis chapcer will address che
environmental impact which would result from imposition of heavy-
ducy engine emission control strategies considered as part of chis
rulemaking.
-53-
-------
4,000.
3,000,
:ron-^echane
Kydrocarbons
(1000 Tons
2,000°
L,000'
4,398
HD Diesel
HD Gas
LD Truck
LD Vehiclt; 74Z
10Z
ft
I V
143
1976
Figure IV-C
Annual Mobile Source
Non-Mechane Hydrocarbon
Enissious for 57 (Jrban Regions
.3,314
1,691
1,140
1,088
TEAS.
-54"-
-------
16,000-
16,223
HD Diesel
HD Gas
14,000-
LD Truck
12,000
Carbon
Monoxide
(1000 Tons
10,000-
3,000-
6,000-
4,000-
2,000-
Figure IV-D
Annual Mobile Source
Carbon Monoxide Emissions
cor 52 Urban Counties
ID Vehicle
142
13%
722
1976
1 2
\
\
162 V
14J
sax
1980
/?.m
292
17?
521
2%
V\
w
\\
,\
1985
YEAR
382
122
462
1990
1995
4,364
4,317
b'/.
32
382
352
7T
492
'¦
5 IS
1999
-55—
-------
3. Primarv Imnact
1. geavy-Outy Engine Emission Races
As was noced Ln Chapter III, heavy-duty vehicles may be
equipped with either gasoline-fueled engines or diesel engines
depending on the needs of che user. The use of a diesel engine, as
opposed Co a gasoline engine, in heavy-duty vehicles is aLso
important from an emissions point of view because che emissions
characteristics or che two engines differ. Basically dieseL
engines have very low levels of HC and CO emissions, below che
levels of che current Federal 'emission standards for heavy-ducy
engines. NOx emissions on the ocher hand, for unconcrolled diesel
engines are high relative to gasoline engine MOx emissions.
Diesels also emit smoke consiscing primarily of unburned carbon
presenc in small parcicles. Gasoline engines do noc. 3uc gaso-
line engines do have higher HC and CO amissions than do diesels.
The primary reason for che different emission characteristics
of diesel and gasoline engines is explained by che way each cype of
engine funccions. Wich gasoline engines, che fuel and air are
mixed in che carburecor prior to passing into che engine cylinder.
The more or less homogenous air/fuel mixture is admicced inco che
cylinders via a throccle place, which is varied in position by che
operator co control engine power, before ic passes through che
intake manifold co che individual cylinders. The air/fuel ratio of
che mixture which enters che cylinder cends co vary ac different
power conditions, wich excess fuel under some conditions and excess
air under ochers. In Che engine cylinder an ignition source•(spark
plug) must be provided co get the combustion started, since gaso-
line air mixtures have high minimum ignition temperatures. The
compression ratio must be low enough to avoid deconacion (or
random auco ignicion), which is another basic characteristic of
gasoline-air mixtures. The effects of these constraints on pollu-
tant emissions is that carbon monoxide and hydrocarbons cend co be
relatively high, being primarily associated with engine operating
modes at which che mixtures are somewhat on che excess fuel side.
Hydrocarbons also result from "quenching" of che combustion reac-
tions because of contact becween the gasoLine-air mixture and
relatively cool surfaces of che combuscion chamber. Nitrogen
oxides are relatively high coo, because of che high peak combustion
chamber cemperacures inherent in che relatively rapid combustion
process or premixed gasoline and air.
Diesel engine ooeracion differs in many ways from that of che
gasoline engine. Fuel and air are not mixed prior to entering che
engine cylinder, and chere is no spark plug since che cype of fuel
used has ignition characteristics such chat ic can be igniced by
Che heac of compression as Long as che compression racio is high
enough. Therefore, unchroccLed air alone is indueced inco che
engine through the intake valve. Engine power i3 controlled by
-56-
-------
varying fuel flow only, wich the fuel injected under pressure
directly into che combustion chamber at Che proper time for igni-
tion co begin. Fuel continues co be injected and burned concur-
rently under highly stratified Local air/fuel, mixtures. The
overall fuel/air mixture, however, is always on the excess air side
to assure chat enough oxygen is available near che fuel spray to
support combustion. Compression racios Co achieve spontaneous
ignition tend co be much higher Chan for gasoline engines, roughly
16 to 21. Because of chese high compression ratios, diesel engines
have higher thermal efficiencies than gasoline engines which,
combined with che fact chat chere are no pressure losses associated
with having a chroctle valve in che inlec syscem, give them supe-
rior fuel consumption characteristics. As co emissions, che excess
air conditions result inherently in relatively low carbon monoxide
and hydrocarbon emissions, buc che high compression ratio cands co
cause diesel engines co have nitrogen oxide emission characteris--
cics of roughly che same magnitude as gasoLine engines, the smoke
from diesel engines is caused by che initially unmixed nature or
che fuel and air in che diesel combuscion process. Ihi3 may also
resulc in objectionable odors in diesel exhaust thac are ooc found
in gasoline engine exbausc.
Considerable work has been done within SPA in an attempt co
deceraine accurate emission factors for mobile sources. This work
depends heavily on in-use vehicle testing under EPA's Emission
Faccor Program. To answer the question of how well vehicles
perform in actual use, EPA has administered a series of exhaust
emission surveillance programs. Test fleets of consumer-owned
vehicles within various major cities are selected by model year,
make, engine size, transmission, and carburetor in such proportion
as to be representative of boch the normal production of each model
year and the contribcuion of chat model year co cocal vehicle miles
traveled. These programs have focused principally on Light-duty
vehicles and light-duty trucks.
The data collected in these programs are analyzed to provide
mean emissions by model-year vehicle in each calendar year,
change in emissions wich che accumulation of mileage, change
in emissions with the accumulation of age, percentage of vehicles
complying with standards, and effect on emissions of vehicle
parameters (engine displacment, vehicle weight, etc.). These
surveillance data, along with prototype vehicle test data, assembly
line test data, and technical judgment, form the basis for che
existing and projected mobile source emission factors. 2/
For this regulatory analysis, changes have been made Co che
emission factors for heavy-duty and light-duty trucks. The emis-
sion factors found in the mobile source emission factors document
for heavy-duty vehicles are based upon steady-state data gathered
on the 9-mode and 13-mode test procedures. In the course of
developing chese current regulations, EPA has accumulated substan-
-37-
-------
cial data on che cransienc emissions of heavy-ducy engines. 3och
che CAPE-21 data gathering program and resultant cransienc cesc
procedure were designed co accurately characterize in-use operacion
and therefore La-use emission. Therefore, che available cransienc
cesc data has been used co revise che heavv-ducy cruck emission
factors which are currently being used. The emission factors for
fucure heavy-ducy engines as well as cor future light-duty trucks
have also been revised co reflect accurately che final standards
and che implementation of Selective Enforcement Auditing vich a 10
percent acceptable quaiicy Level. Refer co Appendix A for details
of the mechodology ana che calculations.
The general form of all the' emission factors for mobile
sources i3 an equation vich some starting new vehicle emission rate
pLus a mileage dependent deterioration rate (see Tables i and 2 of
Appendix A). This means chat co determine che emissions from a
given vehicle one muse know che accumulated mileage. To determine
che average emission race for che fleet made up by a given class of
vehicles (for example, heavy-ducy gasoline-fueled crucks), ic is
necessary co account for che fact chat che on-che-road fleet
consists of a mix of vehicles of varying ages and model years. The
appropriate emission race is applied co each fraction at che fleec
and the fractions are summed into a composite.
When vehicles meecing a new emission standard are introduced
into che on-che-road fleet, they at first represent only a small
fraccion of che vhole fleet.- As cme passes, che newer technology
vehicles come co represent a larger and larger share of che entire
fleec.. This means chac che composite emission rate for che entire
fleec will show a gradual change in response Co new standards,
racher than a sudden change. As an LLluscration, che compos ice
emission race for heavy-duty gasoline-fueled vehicles for CO
changes as follows:^/
Year : 1980 1985 1990 1995 1999
Composite CO (g/mi): 256 222 106 37 40
Although Che new standard is introduced in 1984, composite
rates do noc show substantial drops until 1990 and beyond.
One way co examine the effect of che rulemaking accion is co
compare che emissions of engines built co meec che requirements of
che rulemaking vich che emissions of earlier engines. Using che
emission factor equation of Tabie 1,2 and 4, of Appendix A, che
cotal lifetime emissions of a given model year engine may be
estimated. This will be done cor 1969 (che "baseline" model year
for deriviation of che standard), 1979 (representing engines buile
co current standards), and L984 (year of implementation for chis
rulemaking) model year vehicles. The calculations will use average
vehicle lifetimes of 114,000 miles for heavy-ducy gasoline-fueled
vehicles, and 475,000 mile3 for heavy-duty diesel vehicles^/
Lifetime per-vehicle average emissions are given Ln TabLe TV-A.
-58-
-------
Table IV-A
Lifetime Emissions for 5eavy-0ucy Vehicles (Tons)
Model Year
Glass * Pollucanc 1969 1979 1984
Gasoline fueled
SC 2.71 L.17 0.17
CO 31 31 2.4
Diesel
5c 2.18 2.18 1.41
CO 5.9 5.9 5.9
-59-
-------
The impact of che new standards on vehicles produced for 1984.
(or lacer) is clearly evident in this daca. Compared Co emissions
from 1979 engines, 1984 engines are reduced 35 percenc for 3C and
92 percenc for CO in che case of gasoline-fueled engines. For
diesel engines, HC is reduced 35 percenc while CO remains un-
changed.
2. Reduction in Urban Emissions From 3eavy-0ucy Vehicles
We have seen chac as new heavy-ducy vehicles are puc inco use
and older ones retired, che emissions of che average heavy-ducy
vehicle on che road will decrease. The resulting composite emis-
sion factors can be used' to project changes in annual emissions
from che entire fleec. The same can be done for ocher mobile
3ource categories as well. To make che projections, che changes in
composite amission races are used along with escimaced growth races
in cocal vehicle miles craveled to modify che baseline amission
invencory for fucure years. Projections are also made of changes
in stationary source emission rates depending on present and
ancicipaced scacionary source concrol programs.2/
For hydrocarbons, che exhausc emissions chemselves are an
indirecc racher than a direct problem. That is, the principal
harmful effecc of HC emissions stems from che photo-chemical
reactions leading co ozone formation. The reaction process can
cake several hours, by which time che pollucants involved are
cransported and dispersed over broad areas. Therefore, che hydro-
carbon emissions have been analyzed on an Air Quality Concrol
3-egion (AQC3.) basis. The AQCS.1 s selected were chose non-Califor-
nia, non-high-alcitude regions violating che ozone standard (or
escimaced co be violacing where actual sampling data is missing) in
a 1975-1977 base period. California regions were excluded since
California has its own emission standards. Sigh altitude regions
were excluded because che emissions data used in the analysis is
not considered representative of high altitude conditions. A
separate decailed analysis would have co be done co assess che
inpacc of these regulations on high-aleicude areas. This selection
process led co a sec consisting of 57 AQCRs to be analyzed for
hydrocarbons. In addition, because methane emissions are non-
reactive and do rtoc contribute co ozone formation, che emission
invencories compiled for analysis will be based upon aon-mechane
hydrocarbons (NMHC).
Carbon monoxide emissions, in concrasc co hydrocarbons,
frequencly create localized problems of high concencracions. These
are often associated wich urban core areas experiencing high
traffic densities. It is desirable, therefore, co analyze CO on
a more localized basis Chan AQCRs. This has been done by rising a
councy based invencory. As for HC, only non-California non-high-
alcicude areas were selected. The result is 52 counties exceeding
che CO 3candard for a 1975-1977 base period.
-60-
-------
Following che selection of areas Co be analyzed, an emission
inventory for each region vas compiled. The mosc recent year for
which complete information could be obtained was 1976. This daca
chen forms che basis for future projections. Compilation of che
baseline and projection for future years is an Involved process
entailing a number of assumptions. These are discussed in detail
in supporting documents.6/7/ Two assumptions are important co
highlighc here. The firTt is che assumption chat light-duty
vehicle and light-duty truck I/M programs will be implemented in
ail che areas analyzed by 1982. Since all che areas chosen are
areas exceeding che HC and CO standards, such programs are expec-
ced.
The second assumption concerns projecced growth rates for
various source categories in fucure years. For non-methane hydro-
carbons, rollback projections were made for a range of growth
rates. The high and low end of chese ranges differ by one or cwo
percent. For chis analysis we will use che growth rates of che Low
growth opcion. For mobile sources chese races appear mosc consis-
cent with what appears likely because of energy coses and relaced
matters. The high growth assumptions would increase che absolute
Levels of emissions and decrease che absoluce levels of air qualicy
benefics projecced by che models somewhat. They would, however,
make little difference in che relative change from the base case co
che control case. The maximum air qualicy benefics wouLd peak in
1995 rather than in 1999 if che high growth case were chosen. For
heavy-duty vehicles, ocher specific adjustments in growth races are
also required. Based upon che results found in Section D of Ap-
pendix A, annual vehicle miles traveled (VNT) are expected co
decline for gasoline-fueled engines by abouc 2 percent per year,
while diesel VMXs vili increase by abouc 5 percenc per year. These
rates reflect increased use of diesel engines in the heavy-duty
industry, largely because of energy considerations.
Projections for both emission daca and air qualicy data are
made on a A.QCH. by AQCR basis (or county by county for CO).
However, the underlying assumptions on emission factors are not
region specific. &acher, chey represent typical nationwide values.
Because of this, only average results for all regions will be used
for analysis.
Figure I7-S and IV-F provide a comparison of che projecced
mobile source emissions for che base case presented in Section A of
no new heavy-duty regulations with che projected emissions for che
final regulations. They cover che years from 1990-1999. Withouc
further controls, heavy-duty emissions become a major fraction of
mobile source emissions. By 1993, heavy-duty emissions would
account for 35 percent of mobile source NMHC and 44 percent of
mobile source CO emissions. The substantial reductions in heavy-
duty emissions .expected are clearly indicated. For HC, in L999 the
-€1-
-------
k
1.200*
110 Diesel
1.00Q.
Non-Methane
Hydrocarbons
(1000 Tons)
U00
600.
400.
1,140
110 Gas
1.0 Trucks
200.
.U Vehicles
in
\t>l
12Z
.1,012
32
9X
13/
57%
64/
(a) (1,)
I «!«»(>
Figure TV—IS
Annual Mon-Methane Hydrocarbon
ttuluslons for Baseline Case
and Control Case
1,155
1,088
nz
1 2Z
ax
915
187.
5%
10"/
57%
68/
(i>) (b)
10'-> S
6X
58Z
7%
70%
Note; liar (a) = Baseline
Case of no new heavy-duty
regulations.
Uar (h) - Control Case with
new heavy-duty regulations.
(a) (u)
I')'»<»
-------
Figure IV-P
Annual Carbon Monoxide
Emissions for Baseline Case
and Control Case
5,000'
4,000-
Carbon
Monoxide
(1000 Tons)
3,000-
2,000 -
1,000-
110 Gas
LD Truck
4,732
110 Dleael 5X
JO Vehicle
382
122
48
\^6
6*
2 ax
15Z
59Z
(a) "(b)
1990
^4
62
382
71
49/.
w
\ \
\\3.072
>72
in
in
69%
(a) (b)
1995
YKAft
4,317
&%
Note: Bar (a) = Baseline
Case of no new heavy-duty
regulations.
Bar (b) = Control Caae with
new heavy-duty regulations.
35%
6%
\ vus
11%
8%
9%
51/,
7JX
1999 (l>^
-------
reduction reaches 75 percent for gasoline-fueled engines and 36
percent cor dieseis. For CO, in 1999 che reduccion foe gasoline-
fueled engines is 34 percenc. Diesel CO is unaffected by the new
regulations. These percentages are measured in comparison to the
base-case emissions for the same year, 1999.
Expressed as a percentage of all mobile source emissions, the
impact of the final rulemaking is as follows. Hydrocarbons are
reduced 17 percenc in L995 and 1999. Carbon monoxide is reduced 30
percenc in L995 and L999.
3. Ambient Air quality Impact of Regulation
(Jsing the emission races previously discussed, an anal-
ysis was done of che air quality impact in each of the selected
regions.7/ The Modified tailback method was used for oxidant and
CO co project future air quality improvements for each region. In
addition, che Empirical Kinetic Modeling Approach (EX21A) was also
used for oxidanc. The EXMA procedure has been developed by SPA. in
an attempt Co provide an improved analysis of the relationship
between oxidanc and precursor emissions while avoiding che com-
plexity of photochemical dispersion models. 7J There is uncertainty
over che applicability of SX21A, so that both EXJ1A and rollback were
used to provide a range of possible air quality impacts.
In preparing the air quality projections, baseline amission
races for various source categories were caken from che National
Emissions Data System (NESS), and projections for future control
strategies plus growth rates were made. In combination with che
mobile 3ource projections, chis data allowed an evalutation of air
quality improvements co be expected. With boch Modified Rollback
and che EXMA approach, che relative changes from scracegy co
strategy are more reliable than predictions of absolute levels of
air quality. Therefore, che results will be expressed as percen-
tage gains over baseline between various scracegies. In addition,
although che individual regions used in che analysis can be identi-
fied, che results are noc considered accurate enough Co be used for
a region by region review of che regulations. Rather, averages
over all areas analyzed will be used. The average air quality
improvements are given in Table IV-3.
The modified linear rollback and EXJ1A models differ by a
factor of nearly 2 Co 1 for ozone reductions. Sovever, they each
indicate nearly the same percencage gain from implementing che new
standards. For example,. boch methods indicate an improvemenc of 1
percenc in ozone ia 1395 when che 1984 regulations are imple-
mented. This reduction becomes 2 percent in 1999.
Table 17-3 indicates that carbon monoxide will be improved 5
percent in 1990, and 7 percenc in boch 1995 and 1999.
-€4-
-------
Table IV-B
Average Air Quality Percent Reductions
From 1976 Base Year
Ozone
(Modified Linear Rollbacic/ESMA)
Scracesy
1980
1985
1990
1995
1999
Base Case
13/7
49/25
54/30
54/31
52/29
lap lamenc HD
13/7
49/25
55/32
55/32
54/31
aegs
Carbon
Monoxide
Scracesy
1980
1985
1990
1995
1999
Base Case
16
53
65
67
67
Imp lestenc HD
16
54
70
74
74
-65-
-------
The significance of a percentage gain in air quality in terms
of progress coward attainment of standards depends upon the origi-
nal levels. For example, a 2 percent improvement in air quality
may be sufficient to bring a region chat i.3 already close to the
standard into compliance, whereas in a region experiencing very
high levels (relative to the standard) chat 2 percent would re-
present a totally inadequate reduction. In a region already
meeting the standards, such a further gain vould increase the
margin for compliance. The question could then be posed: "How many
areas originally exceeding air quality standards are brought into
compliance by implementing the new emission standards?" In Table
IV-C the air quality improvements are analyzed in this fashion.
Considering the ozone results first, the marked difference in
absolute reductions predicted by modified rollback versus SKUA
noted Ln Table 17-3 are again readily apparent. While modified
rollback indicates chac 96-98 percent of the regions originally
violating che ozone standard will come into compliance in the
1990's, QQiA puts chat percentage at 63-72 percent. Therefore, a3
noced earlier, caution oust be used in interpreting results from
either model in absolute terms. For example, the indication from
modified rollback that nearly all violating regions will seec che
ambient ozone standard by 1999 should not be considered reliable,
lather, che relative change attributable to implementation of che
new regulations is the item of maximum accuracy. The cable indi-
cates chac implementation of che heavy-duty regulations vill result
in approximately a 2 percent (rollback) to h, percent (E3MA) reduc-
tion in che number of violating regions.
The' cautions noted for ozone are equally important in inter-
preting che CO results in Table IV-C. Only rollback applies co
-his case, and chat model indicates chac with either strategy, all
regions analyzed will atcain the CO standard by 1990. However, it
has already been noced chat it is not within che abilicy of chi3
model co accurately predict absolute air quality levels. There-
fore, che indication of all regions meeting the standard is incon-
clusive. k3 an illustration of che accuracy required to accept the
absolute projections, in the finai rollback projections for 1999
only 37 percent of the regions are in compliance with the standard
by a margin of greater than 20 percent for che base case. For che
control ca3e, Chac resule changes by 5 percent co a value of 92
percent. Inaccuracies on che order of 20 percent or greater are
more than possible in che present air qualicy analysis, and would
markedly change che absoluca levels of predictions. However, such
inaccuracy would probably be relatively constant from strategy to
strategy and lead to consistent relative effects. Unfortunately,
since changes in air qualicy produced by the new regulations do not
become significant prior co 1990, no clear conclusions can be drawn
about the effect these regulations will have on attainment status.
ks noted, baaed upon che number of regions within 20 percent of che
standard, implementing the heavy-duty regulations produces a 5
percent improvement.
-66-
-------
Table IV-C
Percentage of Regions Originally Violating
Air Quality Standards Brought Into Compliance
Ozone
(Modified Linear Rollbaek/EXMA)
Strategy
1980'
1985
1990
1995
1999
Base Case
35/14
96/56
98/68
98/72
96/68
Implement
ED
35/14
96/56
98/72
98/72
98/72
Regs
Carbon
Monoxide
Strategy
1980
1985
1990
1995
1999
Base Case
2
38
100
100
100
Implement
hd
2
S8
100
100
100
-€7-
-------
C. Potential Secondary Environmental Impacts
1. Sulfuric Acid Emissions
A recenc EPA report (3/) provides an in-depch review of
the currenc status of sulfate emissions from mobile sources. On a
nationwide basis, mobile sources represent Less than 2% oc the
total man-made sulfur oxides. However, with the introduction of
the catalyse/air pump technology to control EC and CO emissions
from mobile sources, there exists the potential for a significant
source of mobile related sulfate emissions in the form of sulfuric
acid aerosol. While of negligible magnitude on a regional basis,
mobile source sulfuric acid emissions could produce a significant
Localized urban sulfate concentration in urban street canyons, or
congested urban freeway situations. Moreover, mobile source
sulfates differ from stationary source sulfates in that they are
emitted in che form of a fine sulfuric acid mist and the particles
tend to remain near ground Level.
The increase in sulfate emissions due to the use of oxidation
catalyst/air pump concroL systems on passenger cars and lighc-ducy
trucks has been of considerable concern co EPA. In ore-model year
1975 non-catalyst systems, mosc of che fuel sulfur leaves che
vehicle after combustion as SO^- In oxidation catalyst/air
pump systems used on recenc model year automobiles and Light-duty
trucks, a small amount (Less than LO percent) (8/) of che sulfur is
converted by the catalyst Co SO^ ¦ The SO, combines with vater
in che exhaust to fona sulfuric acid aerosol. Heavy-duty cacalyst
technology has not yec been used on in-use vehicles, and so Little
is known about sulface emissions from these systems.
Extensive efforts have been made within government and indus-
try to improve the information about mobile source sulface emission
factors, sulface air quality modeling techniques and sulfate health
effects as a function of exposure Level. In addition, technology
assessment work is proceeding co identify how sulfates are formed
in cacalysc/air pimp syscems, and co develop other Low sulfate
producing catalytic control systems such as che three-way catalyst.
According to currenc data, the- extent of sulface emissions is much
less than early concerns had anticipated. Major adverse health and
welfare effects from mobile source sulfates are unlikely.3/ TabLe
17-0 indicates sulfuric acid emission rates for several mobile
source cacegories.
The use of catalysts on heavy-duty gasoline engines resulting
from implementing che nev gaseous standards is not expected co
increase present mobile source 3ulface emissions significantly or
to present a future problem. Considering che much larger sulfate
emissions already associated with diesel trucks, plus che fact chat
equipping &D gasoline vehicles vich cacalysts would increase che
number of all catalyse equipped vehicles by only approximately 2
¦o3'
-------
Table IV-0
Approximate Mobile Source
Sulfuric Acid Emission Races 3/
32S04 Conversion H2S04
Source Category Race (2) (mg/aile)
Mbn-caealyae car 1 L
Oxidacion cacalyst car 10 10-15
3-vay catalyse car 5 4
Light-duty dieael car 2 9
Seavy-duty dieael truck 2 50
Aircraft gas turbine 0.03 N/A
-69-
-------
percent:, chere appears Co be no reason co expect a significant
change in roadside sulface Levels.
2. Lead
The introduction of catalyst technology for heavy-duty engines
will require use of unleaded gasoLine co replace che leaded gaso-
line used in current heavy-duty gasoline fueled engines. This
change will have a positive environmental effect as regards Che
emission of Lead particulate in engine exhausts. Reduction of
mobile source Lead emissions is an important means of reducing U.S.
urban population exposures co high ambienc air Lead concentrations.
Emission data for light-ducy vehicles shows chat approximately
80 percent of Che lead content of leaded gasoline is eventually
emitted from che cailpipe. Applying chis co a cypical lead content
for leaded gasoline of 1.5 grams per gallon gives an emission of
1.2 grams per gallon. Wich an average heavy-duty mileage of 9.9
miles per gallon, and a 114,000 mile life, cotal lead emission over
che life of a heavy—duty gasoline-fueled vehicle would be approxi-
mately 30 pounds. That is, converting co unleaded fuel will result
in a reduction of lead emissions for gasoline-fueled heavy-duty
vehicles of approximately 30 pounds per vehicle over che vehicle
life.
3. Eater Pollution, Noise Control, Energy Consumption
Complying wich che heavy-ducy engine regulations is
expected co have negligible impact on water poLlueion, or on che
ability of che heavy-ducy vehicle manufacturers co meet present and
fucure noise emission regulations. Implementing cacalysc tech-
nology can be done wich no fuel economy penalcy. In face, che
analysis of fuel economy impact done Ln che Summary and Analysis of
Comments indicates che possibility of up co a 9 percent fuel
economy benefit.
D. Irreversible and Irretrievable Commicment of Resources
Assuming chac cacalyt.ic converters are used co meet che 1984
standards, an additional commie onenc of platinum and palladium
would be required over and above chat needed for lighc-duty vehi-
cles and lighc-duty trucks which already employ catalysts. The
incremental demand in 1983 from equipping all HD gasoline vehicles
wich catalytic converters would be approximately 7 2,423 troy ounces
of plantinum and 36,212 croy ounces of palladium. These figures
are based upon vehicle 3aies, catalyst loadings and catalyst sizes
developed in Chapter V.
-70-
-------
E. Relationship of Short-Term Uses of cfae Environment co Mainte-
nance and Enhancement of Long-Tern Productivity
More 3tringent control of heavy-duty engine emissions than
that currently imposed will result in substantial decreases in
hydrocarbon and carbon monoxide emissions from this source. This
reduction will be beneficial and aid in che long-term attainment
and maintenance of acceptable air quality.
-71-
-------
Footnotes
U For a current review of this data, as well as citations co
ocher repor'3 on health effects of HC and CO, see "Health
Effects of Exposure co Low Levels of Regulated Air Pollucancs
- A Critical. Review," Benjamin A. Ferris, Jr., M.D., Journal
of che Air Pollution Control Association, Vol. 28, Mo. 5, May
1973.
It For dacails on assumed future scracagies for ocher source
categories see "Data Assumptions and Methodology for Assessing
che Air Quality Impact of Proposed Emission Standards for
Heavy-Outy Vehicles," EPA Air Management 'Technology 3rancii,
Office of Air Quality Planning and Standards, November 1979.
2/ A complete presentation of mobile source emission factors,
including future use projections, can be found in EPA-400/9-
78-005, "Mobile Source Emission Factors - Final Document,"
March 1978.
£/ "Air Quality Impacc of Proposed HDV Emission Standards -
Summary of Resuics," EPA Report TA£B 80-07, December 1979.
5J "Average Lifetime Periods for Light-Duty Trucks and Heavy-Oucy
Vehicles," EPA Report SDSB 79-24, G. Passavant, November 1979.
6/ "Data Assumptions and Methodology for Assessing che Air
Quality Impact of Proposed Emission Standards for Heavy-Ducy
Vehicles," EPA Air Management Technology Branch, Office of
Air Qualicy Planning and Standards, November 1979.
I! "Uses, Limitations and Technical Basis of Procedures for
Quantifying Relationships Becveen Photochemical Oxidants and
Precursors," EPA-450/2-77-021a, tfS EPA, Research Triangle
Park, NC, November 1977.
8/ "Emissions of Sulfur-Bearing Compounds From Motor Vehicles and
Aircraft Engines," Report co che United States Congress,
EPA-600/9-78-028, August 1978.
-72-
-------
Chapcar V
ECONOMIC IMPACT
This chapcer vill examine che coses of complying with che
gaseous emission standards and concrol strategy for L984 and Lacer
model /ear heavy-duty engines. Coses of compliance are in four
main areas: 1) purchase, installation, and check-cue of new test
facilicias designed for use vich the transient test procedures; 2)
development, production, and installation of redesigned engines ana
new or redesigned emission control systems; 3) certification
testing to assure che emission standards are being mec; and, 4) che
purchase and inscallacdoa of new SEA facilities, seli audits, and
annual SEA casting. For gasoline-fueled engines, che primary cost
Is chat for emission coacrol systems. Far diesel engines, che
primary coses are equally shared by necessary test facility modifi-
cations and engine development and redesign aimed at Lowering
emissions. Also, there are coats associated vith che anticipated
need for unleaded fuel for use in catalyst-equipped heavy-duty
gasoline-fueled engines.
The engine manufacturer oust bear che initial costs for engine
modifications, emission control hardware, certification, cesc
facility modifications, and SEA facilities and cests. These costs
vill be added ca the price of che engines it sell3 to vehicle
manufacturers or uses in its own trucks L£ it also produces vehi-
cles. These costs in turn vill be passed on co ics customers, che
cruck and bus owners and operators.' These operators oust' also bear
che coses, if any, of increased operating (fuel/maintenance) costs
which may be caused by the emission control technology.
A. Cost to Engine Manufacturers
The amission control system cost estimates discussed in the
following paragraphs inherently include costs to cover the allow-
able maintenance, provisions, che lower target emission levels with
che 102 AQL, the useful life redefinition, and the idle emission
standard for gasoline-fueled engines in addition co che revised
emission standards.
1. Emission Control System Costs: Gasoline-Fueled Engines
The actual cost of complying vith the emission standards can
be divided into two main areas: pre-production cacalyst develop-
ment and testing and the actual emission control hardware installed
on each engine produced.
Development and testing costs are difficult co estimate when
considering che magnitude of che cask. Catalyst technology is veLl
developed, which would cend co make chis cost minimal, but che
•73'
-------
application of catalytic converter technology co heavy-ducy engines
presents 3one unique problems. Prolonged operation ac high speed
and high load, as well as high speed mocoring, presents cacalyst
durability problems which oust be closely studied. EPA has esti-
mated a development and testing cost of $5.00 per catalyst sold.
This estimace is based on a rough equivalence with the research and
development cost per catalyst applied in light-duty vehicle and
light-duty crack applications.
EPA expects chat manufacturers of gasoline-fueled engines will
adopt oxidation catalyst/EGS. based emission control systems co
comply with the 1984 standards. The components. EPA expects will be
used are shown in Table V-A, and are discussed separately below:
a. Two Monolithic Oxidation Cacalyscs
This analysis assumes chat manufacturers will use one mono-
lichic oxidation cacalyst for each exhaust bank. A total catalyst
volume co engine CID ratio of 1:1 i.3 used co increase catalyse
durability and efficiency. This volume is ac least 202 larger chat
chat u3ed in lighc-ducy applications. A loading of 45 g/cubic foot
of platinum and palladium in a 2:1 racio is used co allow maximum
catalyst durability and efficiency. The catalyst manufacturing
cost was computed using che data and methodology in a cost estima-
tion report prepared under contract for EPA._1/ The methodology was
altered by using 1979 noble mecal prices and allowing for Che
effects of inflation on other cost3.2/ The use of chese catalyses
will inherently yield compliance with che idle emission standards
at no extra cost.
b. Chassis Heac Shields
Chassis heac shields may be required to protect che vehicle
chassis from heat damage by che cacalycic converter. These will
likely be simple stamped steel parts weighing cvo co chree pounds
a piece. The use of brush shields is aoc expected because of che
need co maximize converter cooling and che inherencly higher
running clearance of heavy-duty vehicles. Chassis heac shield
costs were estimated using che Rath 4 Strong report ciced earlier
and allowing for che effects of inflation.
c. Stainless Steel Exhauac Pipes
Stainless steel exhaust pipes between the exhaust manifold and
che cacalyst will be necessary co insure che proper functioning of
the cacalyst for che full useful life. The actual cost of che
pipes is dependent upon che distance between che exhaust manifold
and che catalytic converter. SPA has estimated che cost of chese
pipes using che Bath & Strong report and caking a credit for che
standard steel exhaust pipes being replaced.
-74-
-------
Table V-A
Emission Control. Componenc Manufacturing
Coses for Gasoline-Fueled Eftgines
Componenc EPA Estimate
2 Monolithic Oxidation Catalysts $175
Chassis Heac Shields 6
Stainless Sceel Exhaust 14
Air ?ump Upgrade 26
Air Modulation System 7
Deceleration Fuel Shut-off 11
Parameter Adjustment 4
Unleaded Fuel Engine Modifications 10
Sigh Energy Ignition 0
S253
1/ Assumes total catalyst volume equals expected average engine
CID (360), noble metal loading of &5 g/cu.ft. using platinum and
palladium in a 2:-l ratio. Sized and loaded Cor ful 1 useful life.
-75—
-------
d. Air Pump Upgrade
Alchough all heavy-duty gasoLine-fueled engines currently use
air pumps, chey do not provide che- air volume necessary co maximize
oxidation in che exhausc manifold, che exhausc pipe, and che
catalytic converter. The cost of chis upgrade has been estimated
by assuming chat current air pumps will be modified co provide
approximately a factor of cvo co chree increase in che air volume.
These coses were estimated using che manufacturing coac estimaces
presented in che Rath & Scrong report as an indication of che
increased material costs and labor necessary co upgrade che air
pump. The cost estimate is in close agreement with a GM estimate
submitted in cheir comments.
e. Air Modulation
Air modulation is necessary co provide che opcimum air/exhausc
gas ratio at different operating loads, chus maximizing HC and CO.
oxidation. The manufacturing cost used is a GM escimate.
f. Deceleration Fuel Shut-off
To aid in prolonging catalyst durability during high 3peed
mocoring, some form of fuel shut-off may be required. The manu-
facturing costs used here are caken from a GM estimate.
g. Unleaded Fuel Restriccor, Decal and Engine Label
These minor components will be necessary co decer tnisfueling
and comply wich useful Life regulations.
h. Parameter Adjustment
These modifications will be necessary due co che parameter
adjustment regulations. Since mosc heavy-duty engines are also
sold in Light-duty cracks, and light-duty cruck parameter adjust-
ment provisions will be implemented prior co 1934, che cost of
applying the parameter adjustment provisions co heavy-duey engines
should not exceed minor hardware modifications. 3ased on manufac-
turer comment, a manufacturing cost of S4.00 per engine i3 assumed.
i. Unleaded Fuel Engine Modifications
Unleaded fuel will require exhaust seat inserts, valve guides
and valve seals to maximize engine lifetime. The cost per engine
estimate is caken from (21 data.
j. High Energy Ignition
This item is already standard equipment on heavy-duty
gasoline-fueled engines.
-76-
-------
the component manufacturing (vendor) coses shown in Tabla V-A
ate in close agreement with confidential figures received by EPA in
ocher waiver and rulemaking act ions.3y Corporate and dealer
overhead and profit will be considered in a separate section of
this chapter.4/
2. Passion Concrol Costa: Diesel Engines
a. Gaseous amissions
Even with the increased stringency of the H.C and CO emission
standards, few heavy-duty diesel engine manufacturers are axpecced
co employ add-on devices solely to reduce emissions. Diesel engine
CO emissioas are inherently so low that no engine family will
require further reductions Co meet the cargec emission levels.
However, this is not the case for SC emissions. An SPA
analysis of diesel transient test emission data currently avail-
ableS/ shows chat 14 of the current domestic engine families (36.3
percent of sales) already meet Che cargec emission levels for HC
and CO.
Another 14 of Che current engine families (38 percent of
sales) are within easy range of meeting the cargec emission Levels
with relatively manor changes co injector ciaing, injector rede-
sign, or, in a few cases, combustion chamber redesign. EPA esti-
mates injector timing changes to cost no.t more than $5 per engine,
injector redesign $20 per engine, and combustion chamber rede-
sign $50-9300 depending on production volume. Of chese 14 fam-
ilies, cvo will not be certified in 1984 but will be replaced by
ocher families which are aoc yet certified.
The remaining 10 engine families, about 26 percenc of current
sales, are aot within range of meecing che cargec SC levels, wichout
major design changes or add-on hardware. Eased on che daca avail-
able in the analysis mentioned previously, the remaining engine
families seem co have one or more of che following characteristics:
-high raced speed, low raced BEp
-naturally aspirated
-curbocharged, but aoc incercooled or aftercooled
-cvo stroke engine
-high surface to volume (S/V) ratio
-larger Chan average sac volume
SPA expects the manufacturers of these engine families will cake
several steps to reduce che HC emissions.
The first seeps taken will be similar to chose used co bring
ocher engine families into compliance. These include changes co.
che fuel injection cising and race as well aa modifications co che
-77'
-------
fuel injector design aimed at decreasing sac volume or optimizing
che injector spray pact era.
Other mora costly and sophisticated AC control techniques
which ma/ be used include combustion chamber redesign co reduce che
3/V racio, pre-chamber injection, variable injeccioa cimiag,
curbocharging, aftercooling, and £G3.. Cose escimates for each of
these possible strategies are found in Table V-B.
Of che 10 families which will need major work co meet che HC
cargec Levels, EPA has learned chat ac lease five will noc be
certified in 1984. Of chese five families, one is simply being
dropped, and four are being redesigned primarily' due • co che 1982
noise standards, wich emissions and fuel economy as secondary
considerations. Based on che diesel cransienc cesc data analysis
mentioned previously, SPA. has used engineering judgment In deter-
mining how che manufacturers of che remaining engine families nay
reduce HC emissions co che cargec Levels required. This Ls shown
in Table V-C, cogecher wich E?A's manufacturing cost estimate for
each engine family.
It should be noted chat che injector and combustion chamber
redesign work discussed above gives che manufacturers an excellent
opportunity to perform design research and development aimed at
Lowering NOx and particulate emissions, while at che same cime
improving fuel economy.
b. Crankcaae Emissions: Naturally Aspirated Engines
The percentage of engines which ''ill be naturally aspiraced
(aon-turbocharged) in 1984 is difficult co estimate. Current
certification data shows that about 25 percent of all new diesel
engines sold are naturally aspiraced. These 25 percent are engines
primarily produced for use in either buses or inner-city class VI
trucks (also known as medium-duty).
It is well known that turbochargers are not as effective in
inner-city and suburban applications as chey are in rural and
Linehaul applications. Rowever, as diesel fuel prices continue co
increase, the fuel saved by turbocharging for these applications
will begin co offsec che increased initial purchase price and
maintenance cost of turbochargers.
With the expected influx of diaseLs into che predominately
gasoline-fueled engine markec, classes I IB-VI (3,500-26,000 Lb.
GVWR), it is inevitable chat a substantial portion of these diesels
will remain naturally aspirated.6/ Chapter III 3hows chac in 1986,
the aid year of chis analysis, classes 113-VI will account for 34
percent of tocal diesel sales. For purposes of chis analysis, E?A
shall assume chat half of these engines remain non-curbocharged.
78'
-------
table V-B
Manufacturing Cose for Diesel UC
Control Techniques 1/
Component/Technique
1. Optimize Injection timing 2/
1. Injector Sadesigu 1/^J
3. Combustion Chamber Redesign 2/4/
4. Pre-Chamber Injection 2/3/
5. Variable Injection Timing
6. Turbocharging 4/
7. Afeereooling 4/
3. Exhaust Gas Recirculation 3/
EPA Estimate
Coat per Engine
S 5
5 20
S 50-300
? 100-300
$120-450
?300-l,200
5200-400
5 40-160
U Source: Interagency Study of Pose-1980 Goals for Commercial
tlocor Vehicles, June 1976, and SPA estimates.
2/ Estimates include short-term research, development and cooling
coses, but do not include additional hardware which is not inher-
ently necessary.
3/ Slight fuel economy penalty.
4/ Slight fuel economy increase.
-79-
-------
Table V-C
Emission ConcroL Techniques and
Co3C3 for Heavy-Oucy Diesel Engines
Engine Estimated Z Control Estimated Manufacturing
Family of Total Sales Technique Cose Per Engine 1/
1 .0442 Injector redesign, $320
combustion chamber
redesign
1 6.6Z Aftercool $200
3 0.662 Optimize injection 3305
timing, combustion
chamber redesign
W 0.22Z Injector redesign, $420
aftercooi.
5 LIS' Redesign combustion $225
chamber, add 2G2.
1/ Manufacturing cose estimate i.3 based on production volume
affected.
-80-
-------
A per-engine manufacturing cose of 910.00 will be used as che
hardware coac co close che crankcase. this ascmace was provided
by Cacerpillar Tractor Co.
c. Smwwj-ry of Diesel Control CoaC3
The actual coses incurred co reduce gaseous emissions can be
divided into three main areas: research and development, cooling
changes, and hardware.
Research and Development (R&D) costs will be incurred in 1981
and 1982 as modifications and redesigns are oade co engines and
engine components. Tooling costs will occur primarily in 1983 as
changes are oade co prepare Cor 1934 production. hardware coses
vill be incurred Ln the actual /ear of production.
The major design changes vhich are expected: oocimized
injection timing, redesigned injeccors, and redesigned combustion
chambers, are primarily short term R&D and cooling costs. The
redesigned or modified hardware should aoc be iaherencly more
expensive than that which was previously used. It i3 reasonable
chat che costs accribucable co R&D and cooling would be divided
evenly over che period 1981-1983.
The actual manufacturing cost Co make any hardware modifica-
tions or add any additional hardware co an engine would be primar-
ily dependent on che production volume affected. For example, the
cost per engine for combustion chamber, redesign would be much
larger for 500 engines Chan for 10,000 engines. For chis coac
analysis, small production volume vill be considered as 2,000 or
less per year, medium production volume 2,000-10,000 per year and
large production volume greater chan 10,000 per year.
The final emission control cost will be discussed according
co che small, medium, and large volume criteria described above and
estimating coses based on production volumes.
Of the 63.7 percent of current sales vhich will require some
work co meet the standard, 2.8- percent comes from families with a
small sales volume, 6.6 percent comes from medium sales volume
families, and 46.6 percent comes from families with large sales
volumes. The remaining 7.7 percent of the 3ales will noc be
produced in cheir currenc families in 1984.
Table 7-0 shows che control strategies which may be used in
che small, medium, and large volume sales families. The concrol
strategies used are based an EPA's analysis of che diesel cransienc
test data mentioned previously. The cost estimates are taken from
Table 7-3. However, the final cost estimate used is based, in
addition, on che sales volume affecced. Small volume camiliea
would incur manufacturing costs near che high end of che range
-81-
-------
TabLe V-D
Engine Family CoticroL Cose Estimates 1/
Volume
Control
Percent of
EPA Cose
Class
Tectmiaue
Total Sales
Estimate
Small
Opciaize Injection
Timing
0.38
? 5
Injector Redesign
2.1
$ 20
Combustion Chamber
Redesign
1.45
$300
AftercooLing
0.22
S400
Medium
Optimize Injection
Timing
2.0
S 5
Injector Redesign
3.0
3 20
Large
Optimize Injection
Timing
20
S 5
Injector Redesign
13.4
5 20
Combustion Chamber
Redesign
26.6
S 50
AftercooLing
6.6
S200
Exhaust Gas Recirculation
11.0
5160 2/
U Cose estimates are Linked co production volume expected.
2/ Based on engine family specific daca.
¦82-
-------
ahown in Table 7-3, medium volume sales £amilies near che aid paint
of che range, and high volume sales families near che low end of
che manufacturing cost escimaces.
The final average cosc per engine can now be conpuced by using
che sales percentage and cosc escimace data shown in chese cables
and che daca presented above on sales already in compliance,
disconciaued engine families, and diesel crankcase control cosc.
This final computation is shown below:
Category Fraction of Tocal Manufacturing Cosc Estimate
SAD Hardware
In Compliance
.363
0
0
Discontinued
.113
0
0
Small Volume Sales
.028
4.30
.38
Medium Volume Sales
.030
.70
0
Large Volume Sales
.466
17.00
30.80
Diesel Crankcase
.170
0
1.70
Average Cost Per Engine
$22.50
333.38
Corporate, vehicle manufacturer, and dealer overhead and
profic on chis iavescmenc and hardware will be discussed in a
separate section in chis chapter.V
3. Certification Costs
Certification is che process in which EPA determines whether a
manufacturer's engines conform co applicable regulations. The
engine manufacturer must prove co EPA chat ic3 engines are designed
and will be built such chat chey are capable of complying wich che
emission standards over cheir full useful life. The certification
process begins by a manufacturer submitting co EPA an application
for certification. Subsequently, two seeps occur.
The first step involves the determination of preliminary
deterioration factors for che regulated pollutants. These deteri-
oration factors must be multiplicative in nature for both gasoline-
fueled and diesel heavy-duty engines. The engine manufacturer may
determine chese preliminary deterioration factors in any manner it
deems aecessary Co insure chat che preliminary deterioration
factors it submits co EPA for certification purposes are accurate
for che full useful life. Manufacturers must state chat cheir
procedures follow sound engineering practices and specifically
accounc for che deterioration of E6S, air pumps, and cacalysts as
well as other critical deterioration processes which che manufac-
turer may idencify. In addicion, when applicable, che manufac-
turers must state that che allowable maintenance intervals were
followed in determining che preliminary deterioration factors. The
manufacturers would submit preliminary deterioration factors, based
on che definition of useful life, in each case where current
-83-
-------
care if ication procedures require testing of a durability-data
engine. Beyond chese requirements, EPA. would cioc approve or
disapprove che durability cesc procedures used by che manufac-
curers.
Seep cwo involves emission daca engines. One co four engines
will be chosen for each engine family. These engines would be
operated for 125 hours in a procedure designed by che manufacturers
before che emission ce3C. The preliminary deterioration factor
will be multiplied by che 125 hour emission test results co predict
wnecher che emission daca engines would meet che standards for
their full useful life. If che emission daca engines are predicted
co pass che standards over che full useful life, chen che engine
family is granted certification.
For che purpose of chis cost analysis, che following assump-
cions are reasonable based on past practice. In 1984, manufac-
turers will certify one emission concrol ayscem per engine family
resulting in che need for one 3ec of preliminary deterioration
factors per family. EPA will select chree emission data engines
for each gasoline-fueled family and cwo for each diesel family7/,
since each manufacturer will develop its own preliminary deterior-
ation factors. As a base estimate, EPA has assumed chac a manu-
facturer will follow che current procedures established by EPA.
For a gasoline-fueled engine, chis is a 1500 hour period with a
cesc each 125 hours plus cests associated with scheduled main-
tenance. For a diesel engine, chis covers 1,000 hour3 with a cesc
each 125 hours plus cests associated with scheduled maintenance.
In order co estimate certification costs, unit costs must be
known for each of che following: an emission daca engine cesc
including che required 125 hours of service accumulation plus che
protocype engine and, preliminary deterioration factor assessment,
which EPA believes will be conducted in a manner similar co che
current pre-production durabilicy testing procedure. All certifi-
cation cesc costs include cransienc and idle emissions for
gasoline-fueled engines and transient and smoke emissions for
diesel engines.
Table 7-S gives EPA estimates of these costs for both types of
engines.3/ Estimates are in 1979 dollars.
Tables v-F and V-G show the calculation of initial certifica-
tion coses for each manufacturer. The number of cast operations of
each kind required by each manufacturer depends on che a umber of
engine families it will certify in 1984. Some manufacturers have
provided EPA escimates of how many families chey will certify in
the mid 1980's. These estimates have been used when possible. In
all other cases, che number of engine families certified in 1979
has been used.9/ It should be tioced chat che preliminary deterio-
ration factor assessment costs will only be incurred for chos&.
-94-
-------
Table V-£
Unit Coaca of Certification Tests
Test GaaoLine-rueled Diesel U
1. Preliminary deterioration
factor asaeasmenc. 2/
2. 125-hour emission data
engine. 3/
S122K S106K
513K S20K
U Includes transient, idle, and 3mo*e emissions.
2/ Assumes manufacturers follow current EPA procedures, but chis
is not mandatory.
3j The manner in which che 123-hour break-in period is carried
ouc is ac che manufacturers' discretion.
-35-
-------
fab la V-F
Cauoline-fuelad Engine Cerl ificat tun Coats
for Hodal Viiar >984 and faHowing
Hit.
i
ix>
(a)
Number of
Engine Families
(b)
Total Preliminary
Deterioration Factor
Assessment CosLa
) = (a) x ($ l22K/engine family).
-------
Table V-C
Diesel Engine Certification Costs
for Hodiil Year 19B^ and Following
(a)
(b)
(c>
(d)
(«)
Total Preliminary
Ntuuber
Total
Total Initial
Number of
Deterioration Factor
of Emission
Emission Data
Cert ificat ion
Hfr. Eng
ine Families
Assessment Costs
Data Engines
Engine Costa
Costs
CM
9
$ 954K
ia
$360K
$13I4K
Cummins
19
1I78K 1/
38
16 OK
I938K
Caterpi i tar
9
954K
18
360K
13I4K
Hack
4
424K
S
160K
584K
IIIC
5
5 3 OK
10
200K
730K
Deutz
2
-
2
40K
4 OK
I auzu
2
-
2
40K
40K
White Engines
1
-
1
2 OK
20K
Fiat
2
-
2
It OK
40K
Mercedes
3
3I8K
6
120K
438K
Mitsubishi
I
-
1
20K
2 OK
Scanta Vabia
1
-
I
20K
20K
Volvo
3
-
3
60K
60K
11 i no
1
-
1
20K
20K
Totul
$4 358K
$2220K
$6578K
(b) "= (a) x ($l06K/engine family).
(J) = (c) k ($20K/emisaion data engine), asuuout two per engine family for large volume manufacturers
and one for small volume manufacturers.
(c) = (b) ~ (d).
1/ Cummins stated their prel iminary deterioration factor assessment couiu at $62 ,000/ f ami ly.
-------
engine families certified by Large volume manufacturers. For small
volume manufacturers this case La virtually zero.
4. Teac Facilities Modification
These regulations will require chat manufacturers remodel
and/or purchase aew engine dynamomacers, dynamometer controls,
constant volume sampling systems, and analytical systems. some
manufacturers will have to remodel or build new cesc cells, as
well, to accommodate che new test equipment. There will also be
costs cor developing tascing software and compucar hook-ups. These
requirements are che result of the revisions in the test procedures
and che increased certification Load anticipated for the mid-
19801 s. Of all the manufacturers which are affected by these test
facility modifications, only GM and Chrysler commented in suffi-
cient detail for analysis.
EPA's analysis of these comments, given in the Summary and
Analysis of ComaencslO/, shows chat GM's modifications for
gasoline-cue led engines facilities would cioc exceed 31,573,000 and
Chrysler's costs would aoc exceed 53,334,000, La eicher case, EPA
believes these amounts co be che absolute maximum chat eicher
manufacturer would have co spend on new facilities and equipment co
comply with these regulations.
Based on SPA'3 analysis of GM's comments, heavy-duty diesel
facility and equipment modifications for GM should aoc exceed, at
most, S6,015,000.10/
Since GM and Chrysler cast facility and equipment coses have
been mentioned above and discussed in more decail in other support-
ing documentation, GM and Chrysler costs will aoc be discussed in
any further decail here but will be directly carried through to che
3ummary cables which follow.
a. Dynamomecers and Control Systems.
Manufacturers will need engine dynamometers for two purposes:
pre-production testing and emissions testing. Dynamometers used
for .emission testing using che new cransienc case procedure will
likely have co be DC-electric dynamometers with sophisticated
control systems. Dynamometers used for pre-production testing can
be substantially simpler in terms of their control systems, and can
be either DC-electric or eddy-currenc modified by the addition of a
motor co permic constant-speed motoring. It will be possible to
use an emission test dynamometer to accumulate service, but not
vice versa. This difference in required capabilities and cast
leads EPA to anticipate chat manufacturers will dedicate dyna-
mometers for each purpose, rather than perform boch operations on
-08-
-------
che same sec of dynamomecers.
The a umber of dynamometers of each cype aeeded by each aianu-
faccurer baa beea escimaced by scarring vich che number of engiae
£ ami Lies estimated for Che mid 1930*3 or, lacking a non-
confidential estimate, vich che aumber of 1979 diesel families and
gasoline-cue led families. Based on conversations vich manufac-
turers or on historical racios becveen aumber of families and
number of development-plus-certification dynamometers, che cocal
number of dynamometers aeeded for 1984 vas estimated. This cocal
was splic becveen emission cast and pre-production cescing dynamo-
aiecers by allowing one emission cesc dynamometer per engine family
plus one, unless a manufacturer indicated ic planned co make do
vich fever. The remaining dynamomecers were caken co be pre-
production cescing dynamomecers. E?A believes chis aechod is
conservative in chac ic over escimaces che aumber of che more
expensive emission cesc dynamometers.
EPA chen inventoried che dynamomecers now owned by che major
manufacturers, co identify vhere modifications or additions will be
required co meec cheir 1984 needs. Small volume manufaccurers were
creaced by assuming vorsc-case needs for new equipment. Unic costs
for modifications and additions were also estimated based on EPA
experience, manufacturers' comment,' and vendor estimates received
by EPA.
Tables V-fi and V-1 show che resulcing pre-produccion cescing
dynamomecers coses by manufacturer. Mo manufacturer vill need co
buy completely new pre-produccion cesting dynamometers, but mosc
vill have co make modifications.
Tables V-J and 7-X presenc EPA's escimaces of che need for and
cost of new and modified dynamomecers for emission cescing.
Manufacturers now using DC-eleccric dynamomecers vill have co
cemodeL chase in either of cvo ways, depending on che dynamometer
models. EPA assumes that all manufacturers vich a shortage of
DC-eleccric dynamomecers vill purchase new OC-electric dynamomecers
and aev control systems co fill che shortage. However, it is
likely chac some manufacturers' vill be able co avoid the cost of
new DC-eleccric dynamomecers- by finding a vay co remodel old
eddy-current dynamomecers. Two types of electric dynamomecers
which are suitable for emissions testing vill be considered:
motor-generator based and chyriscor based. There is a considerable
cost difference beeween chese cvo cypes of dynamomecers, but che
more expensive motor generator baaed system is much more estab-
lished. For the sake of this analysis, EPA shall assume chac half
of che manufacturers purchase che more expensive oocor-generacor
based dynamomecer and half purchase che never, less expensive
chyriscor based dynamomecers.
•89-
-------
Table V-ll
Pre-Production Dynamometer tioata, liana 1 ine-Fue I ed fclnginea
Manu f securer
No. of Engine
Fauilieu in 19B4
No. of Dynou
Needed
No. Available
Without Heuiode 1 ing
No. Heuiode led 1/
Total
Coat
1IIC
Ford
Chryaler
cm
4
a
3
4
0
ft
2
0
$1 7UK
U
\_j Coat each - $B5K. Included new control ayateui and motoring capability oil all eddy-current dynawowetera,
EPA eat iuiate.
'lj CM & Chryaler coata are included in the total coat eatiuiatea cited previoualy.
-------
Table V-l
Pre-Production Dynamometer Costa, Diesel Engines
Ho.
of Engine
No. of Dynos
Ho. Available
Total
Manufacturer
Families in
I9b4
Needed
Without Ueinodeling
Mo. Kewodeled 1/ Cost
CM 2/
9
_
_
_ _
Cuuiuin a
19
10
0
10 H50K
Caierpi1lar
9
5
0
6 51UK
Mack
k
2
0
2 17UK
IIIC
5
3
0
3 255K
beucz
2
1
0
1 U5K
lduiu
2
1
0
1 fl5K
White Engines
1
1
0
1 U5K
Fiat
2
1
0
1 U5K
Mercedes
3
2
0
2 1 /UK
Mitsubishi
1
1
0
1 H5K
Scania-Vabis
1
1
0
1 05K
Volvo
3
2
0
2 17UK
II i no
1
1
0
1 bSK
1/ Coat each "
$85K.
Inc ludea
new control syateu>
and motoring capability on
old eddy-current dynamometers
EPA eatimate.
2/ GM coata are
included in
the
total cost estimate
cited earlier.
-------
Table V-J
Emiuuion 'I'eat Dynaiuometera and Control Syateai Coat, Caaul i ne-Ftie led Eitginea
I
•D
to
I
Manufacturer
IIIC
Ford
Chryaler Jl/
CM 3/
No. of Engine
yami liea in I9U4
A
ti
3
A
No. of
Eioiaaion Teat
Uynoa Needed
5
9
No. Keuiodeled \J
11 y Adding Computer
Control Syatew
2
a
No. Keiuodeled If
by Adding Computer
Control and New
Control Cabinet
Total
Coat
$5bOK
450K
\J Coat each = $35K.
2/ Coat each E $170K.
3/ CH & Cliryaler coata are included in tlie totul coata cited previous iy-
-------
Table V-K
Eialaaiou 'feat Dynamometera and Cuntnil Syatein Coma, Dieael tnninea
No. ttewode led 2/
No. of No. ttenodeled I/ Uy Adding Cuiopuur
Mo. of Engine Emiaaion Teat By Adding .Computer Control and Mew Total
Manufacturer Families in 1984 Pynoa Needed Control System Control Cabinet Ma. New 3/ Couc
CM 4/
9
-
-
-
-
-
Cuwioinu
19
20
0
0
20
$5B00K
Caterpi1lar
9
10
0
0
10
2900K
Mack
4
2
0
0
2
i>ttOK
IIIC
5
5
0
0
5
1450K
1 Uvutz
<_j Nuzu
' White Engines
2
2
0
0
2
5H0K
2
2
0
0
2
580K
1
1
0
0
1
290K
f i at
2
2
0
0
2
5B0K
Me rcede a
3
4
0
0
4
1160K
Mitaubiahl
1
1
0
0
1
290K
Scania-Vabia
1
1
0
0
1
290K
Vo 1vo
3
4
0
0
4
116UK
II in a
1
i
0
0
1
290K
\J Coat each c §U5K
2J Coat each ~ $170K
3/ Coat each " $290K, with a range of §175K - $405K.
CM com a are included in ttie total dieael cota for CM cited previuualy.
-------
b. Constant Volume Sampling Systems
No engine manufacturer presently owns any constant volume
sampling (CVS) systems suitable for use in testing heavy-duty
engines with che cransienc case procedure. CVS systems will be
used only for emission testing. Since one CVS unic can serve more
chan one emission cesc dynamomecer, manufacturers will need fewer
CVS unics chan chey do emission ce3C dynamomecer3. Tables V-L and
V-M gives EPA'3 estimates of che number and cose of CVS syscems'
chac will be required by each manufacturer. Thi3 analysis vill
assume a 2:1 dynamomater-co-CVS racio and will use cosc esciaaces
consistent wich manufacturers' comment and EPA experience.
c. Analytical Syscems
Current analytical syscems used by engine manufacturers are
designed for measuring pollucanc concencracions Ln hoc, raw ex-
hausc. Under che cransienc cesc procedure, CO, MOx, and gasoline-
fueled engine HC will be measured afcer being diluted vich cool air
and collected Ln a sample bag. EPA anticipates chac an axiscing
system can be converted co che new requirements at less chan che
cosc of a new syscem. the cosc depends on che cype or NOx analyzer
in che existing 3ystem. Also, because of che idle emission stan-
dard, manufacturers will need an addicional raw exhaust CO2
analyzer for use in che idle emission cesc. One new raw CO2
analyzer will be required wich each analytical syscem. Tables V-N
and V-0 give che number of syscems requiring conversion and che
cocal cosc for each manufacturer. Coses are EPA estimaces.
d. Sew Structures and Remodeling of Existing Scruccures
Some manufacturers indicaced during conversations with E?A
that aew dynamometers and CVS syscems could noc be accommodated
wichouc new or remodeled buildings. Table V-P and V-Q List che
cosc of new or remodeled scruccures. the costs used are EPA
estimates, which on a manufacturer-co-manufacturer basis are
conservatively high. These coses are estimated on a square foocage
basis for che CVS modifications. Dynamometer-related costs depend
upon che cype of dynamomecer purchased. For manufacturers who
purchase che more expensive OC-eleccric dynamometers wich mocor-
generator seC3, facility modifications will also be required co
accommodate che motor-generator sees. No facilicy modifications
will be required if the manufacturer purchases electric dynamome-
ters which do not require motor-generator sees. Aa before, EPA
shall assume chac one-half purchase electric dynamometers which
require facilicy modifications and one-half purchase electric
dynamometers which require no subscancial modifications.
e. Software and Computer Hoo*-up
Manufacturers will need co develop new computer software for
•94-
-------
Table 7-L
Constant Volume Sampling System Cose
Gasoline-Fueled Engines
Manufacturer
ESC
Ford
Chrysler 2/
<21 II
dumber of
Emission Test
Dynos
5
9
4
5
Slumber of 1/
CVS Systems
Required
3
5
Total
CVS
Cost
3450K
750K
1/ Cose each a S150K, includes installation.
2/ <21 and Chrysler costs are included in che total costs cited
previously.
¦95-
-------
Tab Le V-M
Constant Volume Sampling System Coses
Diesel Engines
Manufacturer
Number of
Emission Tesc
Dynos
Number of y
CVS Systems
Required
tocal
CVS
Cose
<21 2/
-
Cummins
20
Caterpillar
10
Mack
2
IHC
5
Deuez
2
Isuzu
2
White Engines
1
Fiat
2
Mercedes
4
Mitsubishi
1
Scania-Vabis
1
Volvo
i*
Hino
1
10
5
2
51800K
90 OK
36 OK
540K
130K
180K
1S0K
L80K
360K
18 OK
180K
36 OK
L80K
U Cose each ™ ?180K, includes inscallacion.
2/ <21 coses are included in diesel eoeal coses for GJ1 cieed
previously.
-96-
-------
Table V~N
Analytical System Costs
Gasoline-Fueled Engines
i
to
I
Manufacturer
1IIC
Ford
Chrysler 5/
GM 5/
Number of Qteini lumi-
nescence Equipped
Systems to be
Converted
5
a
Number of NUIH- 2/V
Equipped Systems
to be Converted
0
0
Nuinber of Ji/4/
New Systems
0
1
Total
Analytical System
Cost
$105
238K
T7 Coat each " $2 IK
2/ Cost each - $29K
3/ Cost each B $70K
4/ Cost includes $5K for a new raw CO2 analyzer.
5/ CM and Clirysler costs are included in the total cost estimates cited previously.
-------
Table V-0
Analytical tiyatein Converaion Coal a
liieael Knginea
Number of Cheiuiluui-
neacence Equipped 1/
Number of NDIR- 2/
Total
Syateiua to be
Equipped Sy a tenia
Number of.3/
Analyt ical
Manufac turcr
Converted
to be Converted
New SyateuTa
Coat
CM
__
_
_
_
Cimuui na
0
0
10
tilGK
CaLerpiliar
0
3
2
264K
Mack
1
0
1
I07K
1IIC
0
2
1
I49K
Deuta
0
0
1
81K
Iauzu
0
0
1
81K
UliiLo Enginea
0
0
1
81K.
fiat
0
0
I
81K
Mercedea
0
0
162K
Mi t a nb i all i
0
0
I
81K
Scaitia-Vabia
0
0
I
81K
Vo 1 vo
0
0
2
I62K
Hi no
0
0
1
81K
\J Coat each = $26K
2/ Coat uacli = $34K
3/ Coat each = $8IK
-------
Table V-P
Mew or Remodeled Structures Coses \J
Gasoline-Fueled Engines
Manufacturer
CVS Systems 2/
Costs to
Accommodate
Coses to
Accomodate
Dynamometers
total
Cost
IHC
S240K
0
S240K
Ford
40 OK
0
400K
Chrysler 3/
Q1 3/
y Based on conversation with iqanufacturers.
2/ Assumes 400 sq. ft. at $200 per. sq.ft., per CVS.
ZJ GH and Chrysler costs are included in the total costs cited
previously.
99-
-------
TabLe V-q
(Tew or Remodeled Structures Coses 1/
Diesel Engines "~
Manufacturer
Cose3 CO
Accommodate
CVS Syscems 3/
Coses to
Accommodate Total
Dvnamometars 4/ Coat
GM 2 /
Cuannias
Caterpillar
Mack
IHC
Deucz
I3UZU
White Engines
c iat
Mercedes
Mitsubishi
Scania-Vabis
Vq Lvo
aino
300K
400K
16 OK
240K
30K
3 OK
80K
80K
160K
3 OK
SOK
160K
SOK
0
L500K
300K
750K
300K
300K
1S0K
300K
600K
150K
150K
600K
150K
800K
1900K
460K
990K
38 OK
38 OK
23 OK
3S0K
760K
230K
23 OK
76 OK
23 OK
y Based oa conversacioas with Larger-volume manufacturer, and on
vorac-case assumptions for smaller-voLume manufacturers.
11 GM coses are iacLuded in the cocal diesel cost cited previously.
2J Assumes an average of 400 3q.fr. at $200 per sq.ft.
¦*J Assumes an average of S150K per emission dynamometer, with a
range of S50K-5250K.
-100-
-------
use la unnormalizing engine operating eye Lea, validating cest runs,
and calculating cesc results. Most af the manufacturers now have
3uicab La computers ac their facilities for this purpose, and others
can arrange for connnercial time-sharing service. Ia either case,
chara are coses associaced with computer hook-up co che cesc area.
Tables V-& and V-S give EPA's estimates of chese software and
hook-up coses. Cose estimaces are In cLose agreement with comma tics
by GM.
5. SeLective Enforcement Auditing Coses (SEA)
la addition to che revised emission standards for 1984, EPA
is implementing a production line testing program for gasoline-
fueLed and diesel engines. this- program, known as SeLective
Enforcement Auditing, is designed to ensure that production-line
engines will msec ac least che applicable emission standards after
adjustment for projected useful life deterioration. The costs
associated with the SEA program can be broken into three main
categories for both gasoline-fueled and diesel engines: S£A cast
facilities, SSA cest costs, and costs associated wich a iOS Accept-
able Quality Level (AQL).
a. Test Facilities and Equipment for Formal SEA
The number of cest facilities required for formal SEA is
dictated by che audit rate prescribed in the regulations. ?res-
encLy this requirement is two cases per day. For small volume
manufacturers chis requirement is decreased co one cest per day.
Sased on a statistical analysis an average sample size of twelve
engines per audit is expected.11/ This average sample size assumes
chat 10 percent of the engines tested are aot in compliance.
Sased on the comments received, it appears chat manufacturers
wilL use Less expensive eddy-currenc dynamometers for engine
"break-in" and wilL then move the engines to another test site for
formal emissions testing on a DC-eLeccric dynamomecer. Using che
SEA regulations and EPA testing experience as guidelines, EPA has
determined chat two "break-in" sices and one emissions casting site
could provide six engines per week.12/ Thus, doubling these
facilities to four "break-in" sices and cwo emission testing sites
would provide che cwo engines per day necessary to fulfiLL the
requirements.
In calculating che costs for these facilities, ESA has assumed
all new facilities and equipment for gasoline-fueled engine manu-
facturers. An outline of these facilities and their cocal cost
is shown in Table V-T.12/
For diesel manufacturers, EPA has assumed all new facilities
and equipment with two exceptions. First, EPA believes the large
volume dieseL manufacturers will use che eddy-currenc dynamometer?.
-101-
-------
Table V-R
Software and Computer Hook-Op Costs
Gasoline-Fuelad
Manufacturer
Software and
Hook-do Costs
rac
Ford
Chrysler _!/
<21 1/
S125K
S150K
\J <21 and Chrysler cost estimates are included in totals cited
previously.
-102-
-------
Table V-S
Software and Computer Hook-dp Coses
Diesel
Software and
Manufacturer Hook-(Jp Coses
GH W S -
Cumins L 7 5K
Caterpillar 15OK
Mack 125K
IHC 125K
Deucz LOOK
Isuzu LOOK
White Engines LOOK
Fiac LOOK
Mercedes 125K
Mitsubishi 100K
Scania-Vabis LOOK
VoLvo 12SK
Hi no 100K
CM coses are included in estimates cited previous!/.
•103-
-------
tab La V-T
SEA Testing Facilities and Cost3
Gasoline-Fueled Engines
Number Case per
Number of Case per of Emission Emission
Mfr. Braak-in Sites 3reak-in Sice U Testing Sicas Testing Sice If TOTAL
<21 4 SS30K 2 S1.01SM 54.15H
ford 4 53QK 2 L.015M 4.15H
Chrysler V 4 - 2 - 1.60M
•JSC 4 530K 2 1.015M 4. 15M
S14.05M
y Includes new facility, supporting functions and equiamenc,
eddy current dynamometer, dynamometer control, transporters, and
receivers.
_2/ Includes new facility, supporting functions and equipment,
electric dynamometer, computer control, CVS, analytical system, and
computer interface.
y Chrysler estimated their costs Lower because ctiey would be
able to use existing facilities and equipment.
-104-
-------
displaced from cheir certification facilicies aa cheir "break-in"
dynamometers. Secondly, EPA does not believe chat small volume
manufacturers will purchase facilities dedicated co SEA, but
instead, vould use cheir certification facilities foe SLA casting.
In the final coat analysis, EFA shall assume one-half of the
certification facility costs are attributable Co SEA. Table VHJ
contains an outline of the facilitiea and coat per manufacturer.
The facilities for SEA will, in moat caaes, be built aear che
engine assembly point co decrease casting cost3 and allow che
common usage of some support facilicies. The coats allowed for
facilities in chis analysis are ample co allow che construecion of
a cesc cell which has all che essencial equipment and supporting
functions necessary for an efficient and 3afe cescing program.
b. SEA Tasting Costs
The actual cesting costs incurred by a manufacturer are
dependent on same cesting decisions made by che manufacturer and oa
che number of SEAs a manufacturer undergoes. For example, che
manufacturer decides, in advance, how many cimes (1-3) he will cesc
each engine and che length of che break-in period prior co casting
(0 co 125 hours). The number of SEAa co which a manufacturer is
susceptible is primarily dependent on che annual sales volume. For
each 30,000 engines sold, che manufacturer La susceptible co one
SEA. The minimum number of SEAs to which a manufacturer is suscep-
tible is one, regardless of sales volume. The possible aumber'of
SEAs per gasoline-fueled engine manufacturer is shown in Table V-V,
and che possible number for diesel engine manufacturers is shown in
Table V-W.
The actual coat per audit is based on che formula:
Coat/Audit* (Coat/Teat)(Tests/Engine)(Engines/Audit)
This cost analysis will be based on the following sec of
assumptions.'
(1) ail audits are passed (12 engines ce9ted);
(2) each engine is tested only once;
(3) each gasoline-fueled engine SEA cest coses SI, 750 and
each diesel test costs 91,750 (these costs cover engine
selection and cranaport, a 24 hour break-in period,
emissions testing, and miscellaneous);
(4) each cest in che audit is completed in 24 hours or less
(excluding che break-in period, which is assumed co occur
on che "break-in" dynamometer);
(5) two full. SEA cests are completed each 24 hour day (chus,.
two SEA cesc cells are used for casting and four other
dynamometers are used for break-in).
-105-
-------
Tibia V-U
SEA Tescing Facilicies and Coses
Diesel Engines
Number of
Mfr. Break-in Sices
Cummins
GM
Cacer-
pillar
Mack
I EC
4
4
4
4
2
Cose per
Sreak-in Sice y
543 OK
480K
450K
480K
430K
Qchers(9)3/ 2 each
Number
of Emission
Tescing Sices
2
2
2
2
L
1 each
Cose per
Emission
Tescing Sice 2/ TOTAL
51.045M
1.043M
1.04511
1.045M
L.045M
54.0 in
4.01M
4.01M
4.0 iM
2.01M
6.78M
1/ Includes new facilicy, supporting funccions and equipmenc,
dynamomecer inscallacion, dynamomecer concrol, cransporcers and
receivers.
11 Includes new facilicy, supporting funccions and equiptnenc,
dynamomecer, CVS, compucer concrol, analycical 3yscem, and compucer
incerface.
3/ Includes Isu2u, Mic3ubishi, Slino, Mercedes, Whice Engines,
Fiac, Scanis Vabis, Volvo, and Oeucz.
4/ One half of cercificacion facilicy coses as shown in preceding
cables.
-106-
-------
Table V-V
Number of 'Possible SEAs" 1984-1988 1/ 2V
Gasoline-Fueled
local per
Manufacturer
1984
1985
1986
1987
1988
Manufacturer
GH
6
6
9
5
5
28
Ford
4
4
4
4
i
20
Chrysler
2
2
1
2
2
10
IHC
2
2
2
2
2
10
local per year
14
14
14
13
13
58
1 / Based on projected sales (see Table V-Z) and allowing one
audic for each 30,000 3old. iouadiag of "possible audic3" La
up co che aexc whole aumber. Assumes ao change in maritec shares of
each manufacturer.
2_/ The term "possible SEAs" includes only audics which are
prompted by sales volume, and does aoc include chose for failure of
an SEA or ocher reason.
-107-
-------
Table V-W
Number of "Possible SEAs" 1984-1988 y 2/
Diesel ~~
local per
Manufacturer
1984
1985
1986
1987
1988
Manufacturer
Cummins
3
3
4
4
4
13
GH
3
3
3
3
3
15
Caterpillar
2
2
2
2
3
11
Mack
2
2
2
2
2
L0 .
ehc
1
1
1
1
1
3
Ochers 3/
9
9
9
9
9
43
tocal per,year
20
20
21
21
22
104
y 3a3ed on projected sales (see table V-Z) and allowing one
audit for each 30,000 sold. Rounding of "passible audic3" is
up co :he aext whole dumber. Assumes ao change is market shares of
each manufacturer.
2_/ the term "possible SZAa" includes only audits which ate
prompced by sales volume, and does aoc include chose for failure of
an SEA or ocher reason.
y Includes one each for Isuzu, Mitsubishi, Hino, Mercedes,
White Engines, Fiac, Scania Vaois, Volvo and Deucz.
-108-
-------
Coat/Audit ¦ (SL, 750/engine test)(12 engines ceaced/audic) ¦
321,000.
Using che case per audic figure above and che coca! number of
possible SEAs per manufacturer shown in Tables 7-X and V-W, che SEA
cesting coses £or gasoline-fueled and diesel heavy-ducy engine
manufacturers are ihown in Tables V-X and V-Y.
c. IPS Acceocable Qualicy Level (AQL) Costs
The goal or passing formal SEA ac a 102 AQL and, in essence,
producing all engines co pass che emission standards ac production,
can'be achieved through ac lease three means: research and devel-
opment aimed ac reaching Lower targec emission levels, produccion
Line qualicy concrol procedures co reduce variability, and posc-
produccion emissions testing aimed ac providing the manufaccurer
with confidence in ics efforts ac passing SEA and che 10Z AQL. The
Lower target emission levels associated with a 102 AQL were con-
sidered when che emission concrol syscem coses, shown in Sections
A-L and A-2, were computed. These coses are higher Chan chose
which would be necessary for a M)Z AQL. Therefore, no further
cosC3 for research and development and hardware will be discussed
here.
EPA expects chac as a response co the implementation of SEA
and, in addition, a 1QS AQL, che manufacturers will insci-cuce a
produccion qualicy concrol program co reduce engine-co-engine
variability. Specificall/, this program will be aimed ac ensuring
chac aaission-relaced part3 are manufactured and installed cor-
rectly and thac emission-relaced calibrations are the same as chose
of che emission data engines. EPA. has escimaced a cose of $10 per
engine for this program. This assumes one-chird of a man hour.
Finally, in response co SEA and the L0Z AQL, EPA expect* chac
all manufacturers will inscicuce a manufaccurer operated produccion
Line audit program Co measure che effectiveness of cheir compliance
efforts and provide themaeLves assurance of cheir ability co pas9 a
formal EPA audit.
EPA believes chac in che first year of SEA, 1984, che manu-
facturers may, on che average, audic as much as 0.6 percent of
their production. However, as chey gain greater confidence in
cheir SEA compliance efforts and build engines co achieve che same
emission standards for several years, this percentage will decline.
EPA has assumed chac by 1986 che audic fraction will have dropped
from 0.6 percent co 0.4 percent and will remain ac 0.4 percent
through 1988.
The costs of this production line casting program pocencialLy
lie in cwo main areas: - facilities and testing.
-109-
-------
Tab La V-X
Coses per Year £or SEA Tasting 1984-1988 _1/ 2/
Gasoline-Fueled ~~
Manufacturer 1984
GM S126K
Ford 34K
Chrysler 42K
IHC 42X
TOTAL FEZ TEAR S294K
1985 1986 1987
?126K S126K S105K
34
-------
Tab La V-Y
Coses per
Year for SEA Testing 1984-
Oiesel
CO
CO
-------
The Eaci.liei.e3 which che manufacturers vill purchase for
formal SEA vill provide each Large volume manufacturer che capabil-
ities to test L,000 engines per /ear, which ac a 0.6 percent audit
race would support a production volume of 166,000 engines. For
small volume manufacturers, cbe SEA facilities vould support a
production of 83,000 per /ear.
The sales projections shown in Chapcar III and repeated in
Table V-Z are EPA'3 projections for the heavy-due/ class as a
whole and are independent of the manufacturers involved. To
determine the need for additional SEA casting facilities and the
per manufacturer cost of production line audits, EPA shall assume
chat the manufacturers' market shares remain•unchanged from currenc
percentages. These percentages are shown as part of Tables V-AA
and V-BB.
Based on these market percentages and SPA1s sales projections,
aone of the diesel or gasoline-fueled manufacturers will require
additional facilities for their own production line audit program.
The cost of a production 3elf audit may vary quite substan-
tially from thac for a formal S£A audit test. Manufacturers vould
minimize the length of the "break-in" period to protect the angine
resale value. In addition, there would be no substantial cost
associated with engine and component selection and transport since
the engines will be removed from the production line ac random, and
pipes, cacaly3ts, etc., will be used many times. It is reasonable
chat manufacturers would design cheir "break-in" programs such chat
one engine per 16 hour day is "broken-in" and gaso 1 ine-rueisd
engine selection and transport costs would be cut to about the same
as for diesels. This would be possibLe in self auditing because
pipes and cacalysC3 can be used over again, and testing facilities
are Located near either the vehicle or engine assembly point.
So, the cosc per tesc vould be $1,072 for gasoline-fueled engines
and 31,274 for diesel engines.12/ These costs include labor and
fuel for "break-in" and emissions testing plus overhead and super-
visory coses. Using the cocai sales figures shown in Table V-Z,
and che 3elf-audic fractions aad audit cest costs described above,
the self-auditing coses per manufacturer are shown in Tables V-AA
and V-BB.
5. Total Cosc to Manufacturers
The four parts of che costs to manufacturers (emission control
system costs, certification costs, test facility modification
costs, and 3EA associated costs) are summarized in Tables V-CC and
V-DD. Costs are in 1979 dollars.
B. Costs to Users of 3eavy-fluty Vehicles
1. Overhead and Profit
-112-
-------
table V-Z
Heavy-flucy Engine Sales
1984-1988
U
Year
Gasoline-Fueled
Diesel
1984
366,991
266,161
1985
360,388
284,255
1986
354,287
302,854
1987
347,171
321,966
1988
339,547
341,383
Tocal
1,768,384
L,516,319
1/ Domestic sales plus inports.
¦113-
-------
Table V-AA
Self Audit Testing Coses
Gasoline-Fueled 1/2/
Market
Total
Mfr.
Share Z 3/
1984
1985
1986
1987
1988
Per Mfr.
rac
12
5283K
3232K
3132K
S179K
S175K
?1051K
Ford
29
635K
561K
44 IK
43 2K
42 2X
2541K
Chrysler
16
378K
309K
243K
23 8K
233K
140 IK
(31
43
1015K
332X
653K
640K
626K
3766K
Total per
year
$23615
S1934K
S1519K
S1439K
51456K
S8759K
1/ Assumes a per esse cose or S1072.
2/ Assumes audic percentages or: 1984, 0.6 percent; 1985, 0.5
percenc; 1986 so 1988, 0.4 percent.
V Market share is an E?A estimate based on past 3ales data.
-114-
-------
labia V-3B
Self Audic Testing Coses
Diesel 1/21
Mar Ice c
Tocal
Mfr. Share X 3/
1984
1985
1986
1987
1988
Per Mfr.
Cummins
30
$61 OK
S543K
S463K
5492K
S522K
S2630K
(31
25
509K
453K
386K
41 OK
435K
2193K
Cacerpillar
13
366K
326K
278K
295K
313K
1578K
Mack
14
285K
254K
216K
23 OK
244R
. 1229K
EHC
7
142X
127K
108K
115K
122X
614K
Deucz
0.8
16K
14K
12K
13K
14K
69K
Isuzu
0.7
14K
13K
UK
1 IK
12K
61K
Whice Engines
0.3
16K
14K
12K
13K
14K
59K
Fiac
0.3
16K
14K
12K
13K
14K
69K
Mercedes
1.3
37K
33K
28K
30K
3 IK
15 9K
Micsubishi
0.2
4K
4K
3K
3K
3K
17K
Scania Vabis
0.2
4K
4K
3K
3K
3K
17K
Volvo
0.5
10K
9K
dK
3K
9K
44R
Hino
0.2
4K
4K
3K
3K
3K
17K
Tocal per year
S2033K
51812K
S1543K
S1639K
S1739K
S8766K
U Assumes a per cesc cost of S1274.
2/ Assumes audic percentages of: 1984, 0.6 percenc; 1985, 0.5
percent; 1986 to 1988, 0.40 percenc.
3/ Market share is an EPA esciaace based on past sales daca.
-us-
-------
Tab le
V-CC
Total Cout to Manufacturers;
Casol ine-Fue 1 ed ling
ineu 1/
Average
Manufacturer Coat per
Eini ua ion
U & 0 Control Syateiu
Engine Coot per Engine
Teat Facility
Initial Modification CosLu
Certification Coat For Certification
1IIC
$10
§253
$ 64 4K
$1670K
ford
10
253
I288K
1988K
Chryuler
10
253
483K
3334K
CM
10
253
644K
1673K
Hanufaciuriir
Teul
Fac i1it ieu
for SEA
Formal SEA Self Audit
Testing (I984-I98B) Tenting (1984-19*18)
Qua 1i t y
Control Program
Coat per Engine
Total per
Manufacturer
UIC
S4150K
$21 OK
$105IK
$10
$ 772 5K
For J
4I50K
4 2 OK
254IK
10
10387K
Clir y t> 1 «:i
1600K
21 OK
I401K
10
7028K
CM
4I50K
58 8K
3766k
10
1082 IK
1/ All Cosla arc unditicuiiiHtiil .
'It Doiifi uol liicludo U & I), dLuititiion cuitLruJ uyuLt;iu, anj <|uuklLy couLrol program coat per engine.
-------
Table V-Dl)
Tota1 Coatu to Mannfucturera:
Dicael Engines
Manufacturer
CH
Ctuuiui nt
Caurpi I Iur
Muck
1UC
lieu 12
ItiUZII
White Engines
Fi at
More liilcs
Hi t uuli i all i
Scania Vabiu
Volvo
II i no
Avurage H & I)
Coat ptir Engine
$23
23
23
23
23
23
23
23
23
23
23
23
23
23
luui aaion
Control System
Coat per Engine
In itial
Curt i f icat ion
Coat
§1314K
I93BK
1 31 AtC
5B4K
730K
40K
40K
20K
40K
438K
2 OK
2 OK
6 OK
20K
Teat Facility
Modification Coats
for Certification
? 6015K
10235K
62B4K
IB02K
35QUK
. 703K
703K
4B3K
703k
I369K
4B3K
483K
I J69K
483K
-------
Table V-DU
Total Com:a to Manufacturers;
Diesel Engines (cont'd.)
Manufacturer
Test Facilit ies
For SEA 2/
Forma 1
SEA Testing
(1984-1988)
Se1f-Audi t ing
Te sting
(1984-1988)
Ijiia 1 i ty
Control Program
Cost per Engine
Total per
Manufacturer
CM
$401OK
$315K
$2193K
$10
$1384 7 K
Cummins
4010K
378K
2630K
10
19191K
CaterpiI lar
4010K
231K
1578K
10
13417 K
Mack
401 OK
210K
12 29K
10
7835K
IIIC
2005K
105K
614K
10
6963K
Ueut 2
703K
105K
69K
10
1620K
Isuzu
703K
105K
61K
10
16I2K
White Engines
483K
I05K
69K
10
1160K
Fiat
703K
I05K
69K
10
I620K
Mercedes
1 368K
I05K
159K
10
3439K
Hi tsubish i
483K
105K
I7K
10
liOBK
Scuniu Vabis
483K
1A5K
17K
10
I108K
Vo 1 vu
1368K
105K
44K
10
294 6K
lli no
483K
I05K
17K
10
II08K
\J All costs are undiscounted.
2/ Assumes small volume manufacturers would use certification
facilities for SEA, tlius, facility costs are shared equally by
certification and SEA.
3/ Does not include H & D, emission control system, and quality
control cost |>er engine.
-------
In addieion co cha direct cost3 of manufacturing discussed La
che previous sections and summarized in Tables V-CC and V-OD, the
manufacturers involved have increased general overhead coses which
muse be recovered and an average profit which 3hould be returned on
che corporate resources invested. To a Lesser degree chis i3 also
true for che vehicle dealer.
To determine what these corporate overhead and profit figures
should be, EPA turned co che 1979 edition of Moody's Industrial
Manual. In addition to general financial information on each
corporation Listed, this publication gives daca on corporate costs
as a funecioa of aec sales. Included in.these corporate cosc
figures are several categories which EPA has found useful in
estimating che overhead and profit figures discussed above. These
inciude: cost of sales, general expenses, administration expenses,
selling expenses, miscellaneous expenses, interest, ocher income,
and income before taxes. (Jsing the ocher income and expense
related categories, SPA was able co closely ascimata corporate
ovehead as a percentage of cosc of sales. Using che income before
cases figure, EPA computed che profic as a percentage of cost of
sales.
(a) Gasoline-Fueled Engines
For gasoline-fueled heavy-duty engines, EPA studied che
financial figures for General Motors, Ford, Chrysler, and Inter-
national 'Earvester for 1976, 1977, and 1978. The corporate over-
head percentages on a per manufacturer basis over che 3 year period
ranged from 6.7 percent co 19.3 percent, with an average of 12.2
percent. The corporate profic figures range from 0 percent co 14.5
percent , with an average of 7.9 percent.
The range on chese figures is coo large co establish a mean-
ingful average, so EPA instead has chosen co sake a conservative
assumption. The clear leader in heavy-duty gasoline-fueled engine
sales is General Mocors, vich more than 40 percent of che market.
In addition, General Mocor3 is also che 3ales and profic leader
motor vehicle industry wide. To be conservative Chen, EPA has
applied General Motors 1976-1978 average overhead and markup
percentages co sales for ail manufacturers. SPA considers che use
of <21 figures to be conservative because d profits are che Largest
percentage of che four and (31 overhead is second largest of che
four manufacturers considered. Using chese assumptions, overhead
as a percentage of cost of sales is 11.4 percent and profic before
cases as a percentage of cost of sales is 13.8 percent.
The close out of che portion of chis discussion related co
gasoline-fueled heavy-ducy engines dealer overhead and profic
should be addressed. Selling a vehicle with a larger AIR pump,
catalytic converter, or any ocher modificacion caused by chese
regulations would not increase dealer overhead. Mo additional
-119-
-------
personnel, or vehicLe servicing would be necessary. Dealer profit
after all taxes has been estimated at 1.3 percent of che dealer
purchase price.13/ To account for che effect of taxes, this arocic
margin will be doubled to estimate the effect of dealer's profit on
the first price increase.
Therefore, che total markup on che manufacturing cost becomes
L.29. This markup is applied on a per vehicle basis to all coats
incurred including hardware, research and development, certifica-
tion, SEA, etc.
The retail price equivalent (XPE) can ciow be computed as:
RPE ¦ (Ail Vendor/Manufacturing Cost per Vehicle)(1.2525(1.03)
(b) Diesel Engines
?or diesel engines, SPA studied che aforementioned 1976, 1977,
and 1973 financial data for five manufacturers: General Motors,
Cummins, Caterpillar, Mack, and Intamacioaal Harvester. The
corporate overhead as a percentage of cost of sales on a per
manufacturer basis over che 3 /ear period ranged from 3.5 percent
co 33.6 percent with an average or 16.7 percent. The before cax
corporate profit as a percentage of cost of sales ranged from 5.5
percent co L7.2 percent, vich an average of about 11.9 percent.
Although chere is once again a very large range on chese percen-
tages , sore reason is available co allow an anal/sis. The five
major manufacturers mentioned above build a vide variety of engines
and motor vehicles as well as engine ana stotor vehicle celaced
products. This diversification will obviously impact cne corporate
financial figures. In addition, chis gives £?A some reasons for
che vide range in corporate overhead and profie percentages. Of
che five manufacturers mentioned, two make only engines (Cummins
and Caterpillar), and three make vehicles and engines (GM, IHC, and
Mack). In terms of sales, chis is roughly a 30/50 split. There-
fore, using che industry wide average corporate overhead and markup
percentages ciced previously (16.7 percent and 11.9 percent respec-
tively) would certainly be a representative, if not a conserva-
tively high, estimate for the heavy-ducy diesel industry as a
whole.
Turning finally to dealer overhead and profit, £?A once again
sees no incremental increase in dealer or franchise overhead as
a result of these regulations. Ho additional personnel or engine
servicing will be necessary. Itost heavy-duty diesel engines sold
in che United States are not sold chrough conventional dealers as
are automobiles and light-ducy crucks, but instead, are 3old
chrough either dealer franchises which specialize in only crucks or
chrough manufacturers' representatives. The individual retail
price of a diesel cruck or bus may exceed 350,000, and muLcipLe
unic sales co cicy cransic systems, incer-cicy bus companies, or
Large crucking companies are quice common. Admiccedly, dealers
¦120-
-------
would cry co gee a. small profit on chair increased invesemenc La
che engine, but this could easily be very small or aoching after
the final price negotiations on che sale of che vehicle(s). The
amount of any profit will be very small, 30 EPA shall assume ic is
small enough co Call vithia ocher possible errors in estimating
manufacturing coses or corporate overhead and profic.
therefore, EPA shall use a profic and overhead markup figure
of 1.29 co determine che retail price equivalent. Once again chia
markup will be applied co all che per engine coses associated wich
chi3 rulemaking action.
These overhead and profic markups which, coincidentally, are
boch 29 percent, will be applied co che cocal tnanuiaccuring coses
co determine che first price increase.
2. Increases in First Coses
The added costs co engine manufacturers for emission coacrol
system research, development, hardware, certification and durabil-
icy testing, cesc facility modificationsf and 5£A. related costs
will be passed on to ehe purchasers and users of heavy-duty
vehicles. Assuming che average BL&O and hardware cose per engine,
the amount a manufacturer oust increase che price of ics engines in
order co recover ics expenses depends on che ciming of coses and
revenues from sales and on the cost of capical co che manufac-
turers. Tables V-EE and V-FF show che timing of costs which has
been used co estimate che average firsc price increase.
EPA has assumed chae over the Long run, manufacturers face a
10 percent cose of capical and chae chey price cheir engines
so as co recover their invescaent in five model years, 1984-1999.
Table presents che expected average first coat increases
for che 1984-1988 period for both cypes cogecher wich a restatement
.of che critical assumptions used.
The remainder of che coses discussed in chis Costs co Users
section are discounted at 10 percenc to January { of che taodeL year
in which che engine is produced unless stated ocherwise.
3. Maintenance Coats
The use of 1984 control technology is noc expected to increase
maintenance coats, but conversely is expected to actually decrease
maincanance coses in ac Lease cvo areas: exhaust ayscams and spark
plugs. the use of unleaded gasoline combined with material im-
provements in the exhaust system will reduce maintenance costs
associated with exhausc system replacement. EPA eseiaaces chac ac
che minimum one encire exhausc syscem replacement will be saved
over che vehicle lifeeima. Spark plug life will be lengthened-
-121-
-------
Table V-KE
Fixed Coat lu Ma nut acL ururn by Period:
Gauoliiic-t'imleil
Mfr. 1980 1/
1981 2/
1982 3/
1983 4/
1984 5/
1985 5/
1986 5/
1987 5/
1988 5/
Total 6/
IIIC § 557K
$1113K
§2536K
$223 IK
$ 325K
$ 274K
$ 224K
$ 22IK
$ 217K
$ 7725K
Kord 663K
I325K
305IK
2387K
769K
64 5K
52 5K
5I6K
506K
I0.387K
Cluyaler 111 IK
222 JK
JI66K
91 7K
42 OK
351K
28 5K
28UK
275K
7028K
CM 558K
11 15K
2563K
2 2 31K
11A IK
958K
779K
745k
73 IK
10,82 IK
Total $2B89K
§57?6K
$9343K
§7 766K
$2655K
$2228K
$1813K
$1762K
$1 729K
$35,96IK
1/ One-third of cert i f icat ion facility modi f icat iona .
2/ Two-thirda uf cert i f i cat ion facility modifications.
3/ I'rcl iminary deterioration factor assessment costa plua one-half of SI£A facility coata.
4/ Kmieaion data engine casta plus one-half of SEA facility coata.
5/ Formal SEA plus ae I f~auditing coats.
6/ Doea not include R & U, eiuiusion control hardware, or quality control program coata per engine.
-------
Table V-FF
Fined Coal Co Manufacturers by ftsriuil:
Diesel Engine a
Mfr .
1980 W
1981 W
1982 2/
1983 3/
1984 4i
1985 4/
1986 4/
1987 4/
1988 4/
Total 5/
CM
93007K
$3008K
$2959K
$2365K
$ 572K
$ 516K
$ 449K
$ 473K
$ 498K
$13847K
Cuuuui iiu
5117K
5118K
3I83K
2765K
673K
606K
54 7K
576K
606K
I9I9IK
Caterpi1lar
3I42K
3142K
2959K
2365K
408K
368K
320K
337K
376K
I34I7K
Mack
'JO IK
90 IK
2429K
2I65K
32 7K
296K
258K
272K
286K
7835K
IIIC
I754K
I755K
I533K
I202K
163K
I48K
129K
136K
I4JK
6963K
Deutz
-
703K
703k
40K
37K
35K
33K
34 K
35K
1620K
leuzu
-
703K
703K
4 OK
35K
34K
32K
32K
33K
16J2K
White Engines
483K
48 3K
20K
37K
35K
33K
34 K
35K
I160K
Mercedes
-
1369K
1686K
I20K
58K
54K
49K
5IK
52K
3439K
Mi t uubiuli i
-
4S3K
48 3K
2 OK
25K
25K
24K
24K
24K
II0&K
Be ania Vubis
-
483K
483K
20K
25K
25K
24K
24K
24K
llOflK
Volvo
-
1369K
1368K
60K
3 IK
30K
29K
29K
30K
2946K
II i no
-
4B3K
483k
20K
25K
25K
24K
24K
24K
I108K
Fiat
-
703K
703K
40K
37K
35K
33K
34K
35K
I620K
Total
$1392 IK
$20703K
$20158K
$1 I242K
$2453K
$2232K
3I984K
$2080K
$220IK
$76.974K
J/ One-half of certification facility modification cost e.
_2/ One-half of SEA facility coats plus pre I iuii nary deter iorat ion factor assessment costa.
"\J tine-half of SEA facility costs plus emission data engine cuuU.
4/ Formal SEA audits plus self-audit testing costs.
_5/ Does not include H & D, emission control hardware, and <|uality control program costs.
-------
Table V-GG
Average Increases in Firsc Cose
of I984-L988 Model Year Engines
Engine Type Increase in Firsc Cose _L_/_2_/
Gasoline-fueled 3 394
Diesel 3 195
y Assumes: equal firsc cose increases for aLL engines of
a cype produced during 1984-L988 model years; amorcizacion of all
cos'cs from Tables V-HH and V-II during 1984-1988 ac 10 percenc
inceresc; 1984-1988 sales from Table V-CC.
2/ Coac includes all overhead and profit ac all Levels.
-124-
-------
substantially over che currenc incerval3 as a resale of the use of
unleaded fuel.
To estiaace chese cost savings, exhausc system and spark plug
replacement coses, the currenc and fucure spark plug replacement
intervals, and a nileage accumulation race for heavy-duty gasoline-
fueled engines must be known.
EPA ascertained exhausc syscem and spark plug replacemenc
coses using pares plus labor replacemenc cosc estimates received
from several recail dealers. For a sec of 3 spark plugs, estimates
ranged from $9.50 co S20.00. SPA used the mid-range in chese
estimates, approximately 514 per sec. Exhaust syscem replacemenc
coses ranged from S1S0 co S200 for parts and labor. Since ac least
one complete exhaust syscem will noc have co be replaced, EPA
conservatively escimaces a cosc savings of $130 over the vehicle
lifecime.
The currenc heavy-ducy spark plug replacemenc interval lies in
che 12,000 co 16,000 miie range. The new maintenance incerval is
25,000 miles. For exhausc systems, E?A has escimaced chac withouc
che use of unleaded gasoline, only one replacement, probably lace
in che fourth year, would be required, and che second replacemenc
could be eliminated.
Finally, co compute che discounced values of chese savings,
che average mileage accumulation race for heavy-ducy gasoline-
fueled vehicles muse be known. This was taken from an EPA techni-
cal report14/ and Ls shown in Table V-HH cogecher with che exhausc
syscem and spark plug computations.
Using che daca in Table V-HH and a scandard 10 percent dis-
count race, che average spark plug and exhausc syscem savings can
be computed co be $176 per vehicle. EPA believes this figure co be
a minimum chac can be expected over che 3 year, 114,000 mile
average useful life.14/
4. Fuel Economy and Fuel Cosc3
EPA does aoc expect an increase in overall fuel consumption
for eicher gasoline-fueled or diesel heavy-ducy engines.
Some diesel engines which use EG2. co control HC may incur a
slight fuel economy penalty. However, diesel engines which add
aftercoolers, or redesign injectors or combustion chambers will
have a fuel economy increase which, in che aggregate, will more
Chan offset any possible penalty.
For gasoline-fueled engines, the addition of a larger AIR pump
will cause a 4-8 percent fuel economy penalty. However, che
addicion of a cacalyst will allow engine tuning for fuel economy-
-125-
-------
Table V-Uli
Enltauat
System and Spark Plug
Sav iuga
KkIiuusc Syateui Htiplaceinent
Number of
b'|>ark Plug
Hi tli and Uitliout
Average Annual
Cuiuulac i v
-------
which could yie Ld. a 12 percent co 17 percenc fueL economy improve-
ment. Thus, considering che AIR pump Losses co be 3 percenc, che
worse case assumpcion, che fuel economy improvemenc expected ranges
from 4 percenc - 9 percenc.15/
Osing a L979 fuel economy of 5.4 miles per gallonlS/ and a
recail unleaded fuel cosc of 31.10 per gallon chis could yield
savings per vehicle over a useful life of 3 years and 114,000 miles
in che amouncs shown in table 7-11. table V-II gives discounced
fuel economy benefics and fuel economy increases.
Cacalysc equipped gasoline-fueled engines will require un-
leaded gasoline which can be expected co cosc more than leaded
gasoline. The current average nationwide differencial between
Leaded and unleaded gasoline is roughly 4 ceacs per gallon, 3 cants
of which is che currenc refinery gate case differencial.17/
In the Long run, as increased quantities of unleaded fuel are
produced, chis differencial is predicted to shrink co abouc 2.5
cencs per gallon ac che reiinery gace and 3 caacj per gallon
recail.18/ This 3 cencs would include about 0.5 cencs per gallon
co cover che need for modified or increased unleaded gasoLine
pumping facilities.
Using an average fueL economy in che 1984-1988 cime frame of
9.9 miles per gallonl9/ a lifecime of 114,000 miles and a differen-
cial of 3 cencs per gallon lifecime fuel coses will increase by
$259 (discounced). Including chis cosc ignores che real Likelihood
thac by L934 lead tolerant catalyses will have been developed co
che point chac they will be commercially available on motor vehi-
cles.
5. Total Coses co Qsars
To summarize, users of heavy-duty vehicles equipped victi
diesel engines can, as a result of che scatucory emission standards
and accompanying procedure changes, expect co pay abouc 5195 more
for L984 model year vehicles chan for comparable models boughc in
1983, in 1979 dollars. Ho increase in operacing coses for chese
vehicles will occur. Assuming a 475,000 mile cocal Life, che first
cosc increase translates co an increase in operacing coses of 0.041
cencs per mile (undiscouncad).
For vehicles powered by gasoline engines, che firsc cosc will
increase by $394. Added unleaded fuel coses, less savings on spark
plug and exhausc system replacement, will cocal $83 (discounted)
over che useful Life. The sum of first coses plus operacing coses
is equivaleae co abouc 0.42 cencs per mile (undiscauaced). This
does noc include a fuel economy benefit which would actually Lead
to decreased operacing coses.
-127-
-------
Fuel
Economy Increase
I Percent
3 Percent
'* Percenc
5 Percent
7 Percenc
9 Percenc
Tabla V-II
?ual Economy Benerica
Gallons Saved
209
627
336
1045
1463
1331
Discounted Savings
at 31.10 per gallon
at 5Z discount race JL_/
S 197
S 591
S 733
5 985
S1379
SI 773
y Use ot a 3 percent discount rate is a more realistic reflec-
tion of the actual benefits because it accounts for fuel prices
which are expected to rise at a greater r3te chan a char prices.
This discount race vas suggested by the Council on Wage and Price
Stability in their Comments an the Proposed Light-Duty Diesel
Particulate Regulations. See Summary and Analysis of Comments,
Section 7 for Che . Light-duty Diesel Particulate Regulations for
further discussion.
-128-
-------
C. Aggregate Coses - 1984-1988
The aggregate cost co che aaeion of complying with che 1984
Federal heavy—duty engine emission regulations and che aew SEA
program consist of che sum of increased coses for aew emission
control technology and hardware; aew and modified cesc equipmenc
and facilities;, additional certification costs; SEA facilities and
testing; and unleaded fuel, for gasoline-fueled engines. These
costs will be calculated for a five year period (1984-1988) of
compliance, but will include operating costs incurred after 1988 by
engines produced in the five year period.
Ic atusc be noced Chae calculating aggregace coses based on a
five-year period distorts the cost impact of the regulations. Ac
the end of chat five year period, part of the initial investment in
new engine designs, aew certification, and new or remodeled test
facilities will still exist and be productive. As just one ex-
ample, the new dynamometers required by manufacturers have an
expected life of about 20 years. The "salvage value" of che 1984
investment could reasonably be subtracted from che five-year
aggregate coses. It will aoc be, since che exacc value of che
still-intact investment at the end of 1988 will depend on che
manufacturers' product plans at that time, which are uncertain now.
It must also be noted chat.aggregace cost3 will be calculated
without considering che more stringent NOx and particulate: stan-
dards expected to be promulgated for 198S. These standards will
likely force redesign and recerciricaeion of at least some engines.
The cost of this has aot been considered here. When che 1983 NOx
and particulate standards are proposed, cheir cost impacc will be
caken to be the cost increase they cause over che costs calculated
here..
The five year costs of compliance are dependent, of course, on
the number of heavy-duey vehicles sold during that period wich
eicher gasoline or diesel engines, and also on che mix of sales
between gasoline engine or diesel engine-equipped vehicles. The
accuracy and validity of projecting vehicle sales as far into che
future as 1988 is problematical, so cost estimates based on such
projections are also subject to some qualification. The engine
sales scenario . which EPA used to make these: cost calculations is
discussed in detail in Chapter III of chis analysis. They are also
shown in Table V-Z and reflect roughly a 6.4 percent growth in
diesel sales, a 2 percent decrease in gasoline-fueled engine sales,
and a 1.85 percent overall growth race in heavy-duty engine sales
over che 5 year period 1984-1988.
The various costs associated with the regulations will occur
in different periods. In order co make all costs comparable, che
present value at che start of 1984 of the aggregate costs has been
calculated, based on a discount rate of 10 percent.
129-
-------
Use of a discounc race emphasizes :hac because or che cime
value of money, a cosc Incurred coday is worth more co che nation
chan a cosc incurred in che future.
The calculation of che present vaLue in 1984 of che aggregate
coses i3 shown in Table V-JJ. The ciming assumptions used in
Tables V-ES and V-r? were used in computing che aggregate coses.
It is estimated chat che aggregate cosc of complying with che new
regulations cor che 5 years period is che equivalent of a lump-sum
investment of about $983 million dollars (1979 dollars) made at che
start of 1984. Expressed in .other cerms, che aggregate cosc of
compliance is equivalent co investments of 5L95 per diesel engine
made at che 3tarn of che year che engine is produced and S477 per
gasoline-fueled engine also made ac che 3tarc of che year che
engine is produced. Overall che aggregate cosc i3 equivalent co
5238 per heavy-duty engine.
For ease of reference che components of che cosc of compliance
and che different ways of expressing ic are shown in Tables V-XX,
V-LL, and 7-MM.
The effect of che fuel economy benefit on che aggregaca cosc
of chis package is subseanciai.
The daca from Table 7-11 cogecher wich che cocal heavy-ducy
gasoline-fueled engine production estimated for che 5 year period
(1,763,384) can be used co escimace che affecc of che fuel economy
benefic on che aggregate cosc. This i3 shown in Figure 7-a.
It is very obvious chat che minimum fuel economy improvement
determined in EPA's analysis (4 percent) would reduce the aggregate
cosc of chese regulations Co below zero. The fuel savings would be
so large chac any increase in che purchase price of heavy-ducy
gasoline-fueled vehicles would be more Chan offset by decreased
operating costs.
0. Socio - Economic Impact
I. Impact on Heavy-Duty Engine and 7ehicle Producers
The promulgation of che 1984 heavy-ducy engine emission
regulations will cause che manufacturers of chese engines co spend
about 571 million dollars for test facilities modifications and
engine certification, an additional average 583 million a year
for production of emission control systems and 590 million for che
SEA program over and above those required co meet current stan-
dards. These costs will be initially paid by individuals or
companies chac buy heavy-ducy vehicles and ulcimacely by che
consumers of che produces carried by chose vehicles. SEA and
self-audit testing costs are an exception, but most of che com-
pliance coses are incurred by engine manufacturers before chey*
-130-
-------
Table V-JJ
Presenc Value in 1984 of Aggregate Cose of Compliance
For 1984-1988 Model Years
Year Coac Presenc Value in 1984
Diesel
(1,516,819 Engines)
1980
513.921K
5 20,382K
1981
32.054K
42.664S
1982
24.732K
29.925K
1983
29,37 IK
32.308K
1984
25.669K
25.669K
1985
26.933K
24.485K
1986
28.277K
23.369K
1987
30.068K
22.591K
1988
31,996K
21.854K
3243, 247K
Gasoline-Fueled (1,768,884)
1980
2,889K
4.230K
1981
14.605K
19.439K
1982
18.172K
21.988K
1983
7.766K
8.543K
1984
142.243K
142.243K
1985
155.113K
141.012K
1986
167.135K
138.128K
1987
U6.756K
87.721K
1988
131.855K
90.059K
1989
U.593K
7,198K
1990
11.072K
6.250K
1991
-119K
-61K
1992
42.639K
19,892K
1993
25.926K
10.995K
1994
15.704K
6.055K
.1995
2,5515
894K
?704,536k
1984 Pesenc Value of Aggregace Cose: 9947,833K
-131-
-------
Table V--OC
Undiscounted Costs of Compliance Per Engine
for Engines Produced 1984.-1988
Cose Affecting Selling Price
Certification Facilities
Research and Development
SEA Facilities
Certification Testing
SEA Testing
Self Audits
Manufacturing
Quality Control
Overhead and Profit
Operating Cost
Unleaded Fuel
Exhaust System and Spark Plugs
Undiscounted Cost per Engine
Potential Fuel Savings
(low estimate)
Gasoline-Fueled Diesel
? 4.91 522.88
10.00 22.50
7.96 16.40
1.73 4.35
.81 1.44
4.95 5.78
253.00 33.38
10.00 10.00
88.61 43.34
$345.40
-237.00 -
$490.37 S160.57
-$920 0
-132-
-------
Table V-LL
Discounted Costs of Compliance Per Engine
for Engines Produced 1984-1988 1/
Cost Affecting Selling Price Gasoline-Fueled Diesel
Certification Facilities $ 3.06 $38.43
Research and Development 15.13 33.13
SEA Facilities 10.98 22.69
Certification Testing 2.45 6.19
SEA Testing .31 1.45
Self Audits 5.06 5.90
Manufacturing 253.00 33.38
Quality Control 10.00 10.00
Overhead and Profit 88.61 43.34
Operating Cost
Unleaded Fuel $258.72
Exhaust System and Soark Plugs -176.13 -
Cost of Engine at Start of
Production Necessary to Pay $476.74 $195.01
Cost of Compliance
Potential Fuel Savings -$788 0
(low estimate)2/
T7 10 Percent Discount: to January 1 of the model year.
7/ See Table.7-11.
-133-
-------
Table V-MM
Aggregate Cose of Compliance
for Engines Produced L984-L988
(102 Discount eo January 1, 1984)
Gasoline-Fueled
Diesel
Certification Facilities
5 U.918K
? 47,937K
Research and Development
22,435K
41.330K
SEA Facilities
16.228K
28,297K
Certification Testing
3,o20K
7,715K
SEA Testing
1.196K
1.813K
Self Audits
7.475K
7.363K
Manufacturing
373,924K
41.635K
Quality Control
14,780K
t—¦
n
Overhead and Profit
130,957K
54,633K
Unleaded Fuel
382.373K
-
Exhaust System and Spark Plugs
Total
-260.319K
?704,537K
3243,246K
Grand local: 3947.333 Million Dollars
-134-
-------
AG&flEGrfvre coar vs. fuel. rccAjani increase
ZooH
oh
37«
Pfetc«-nV Fuel CtoirxC»w\y lncft«3
7?.
-------
realize any revenue front sales of engines for which chis money
is expended. This regulation, therefore, will require manufactur-
ers co generate additional capital beeveen promulgation of the
final rule and 1984 either internally or from the capital markets,
sufficient to meet each year'3 coats.
Tables V-CC and V-OD show the costs of test facility modifica-
tions certification and SEA by company. (GM and IHC are che only
companies appearing in both tables.) Tables V-E2 and V-rF show
their total cost3 including control system production costs, by
year. Costs are first incurred La 1980 33 facility, modifications
begin on a large scale. The first opportunity to recover costs via
price increases will be in 1984, assuming that competitive pres-
sures keep manufacturers from raising prices on earlier engines
beyond what chey would do without these regulations.
Most engine manufacturers should have little difficulty
financing the required investment, barring a post 1930 recession.
Cummins, Chrysler and White Engines can be expected to have che
most difficulty. For Cummins che 1980-1983 investment represents
about 40 percent of its 1978 corporate profits. Chrysler will have
severe trouble financing any capital Investment without government
loan guarantees. White Engines will require several years profits
to finance their investment and may choose to seek another means of
certification to reduce their initial capital investments. Tor
Mack, Caterpillar and IHC the investment ranges from 3 percent to
16. percent. cat GM and Ford Che investment required is. less chan
one percent of 1978 profits. The foreign manufacturers should
also have little problem financing che required investment which
will likely be below the worst case estimate presented here.
Payments on recurring costs (emission control system produc-
tion, SEA related testing) will occur closer to the time revenues
are received (via sales of controlled engines) than do che payments
required before 1934 production begins. Assuming manufacturers can
pass on the largest fraction of their costs, then chey should be
able to finance most of production and installation of control
equipment from current revenues.
Changing che prices and operating costs of heavy-duty engines
may, of course, impact the sales of engine manufacturers. 3oth
tacal sales and sales mix between diesel and gasoline-fueled
engines may be affected. EPA. knows of no estimate of the cross-
alasticicias of demand20/ for gasoline-fueled and diesel engines.
When considering the change in sales mix, the cost of ownership, as
well as the increased first cost, nay cause a demand shift. Based
on che average first cost, increase (5394 gasoline-fueled, S195
diesel) and che increased costs of ownership C$83 gasoline-
fueled, $0 diesel) the demand shift would appear to be toward
diesel engines. However, fuel- economy improvement expected
for heavy-duty gasoline-fueled vehicles will offset the increased
-136-
-------
operating cost associated with unleaded fuel. Therefore, in che
long run the demand shift would be caused, only by che first price
increase. This could occur when Che larger firsc price increase
for gasoline—fueled engines, coupled wich cheir inherently larger
operacing and maintenance costs offsecs che greater initial pur-
chase price of a diesel engine for a given vehicle. Bovever, the
existing price difference beeveen comparable diesel and gasoline-
fueled engines (as much as a factor of chree) will allow chis
demand shift co occur for only a small fraction of cotal heavy-duty
sales. In addition to first, cost3 and operating costs, other
factors are considered' by purchasers of heavy-duty engines.
Gasoline-fueled engines usually give beeter overall performance and
are better suited co multi-stop applications. Diesel engines have
lover maintenance and fuel coses, a longer useful life, and give
oectar fuel economy especially in over che road applications.
The pending heavy-ducy particulate and MOx regulations may
have an effect on che gasoLine-fueled and diesel sales mix. In che
Long run, fuel economy considerations wilL be che primary reason
for che sales mix change which is generally expected by both
industry and EPA. The impact of these regulations on the selling
price and operating costs of each type of engine should not cause
any further increase in the changes in the heavy-duty market split.
EPA's Office of Noise Abatement Control has. escimaced che
overall price elasticity, of demand for oev trucks co be in che
range of -0.9 co -0.5.21/, Assuming a mid-range elascicicy of
-0.7, and a range of 315,000 co. $50,000 for che selling price of
heavy-ducy. vehicles, the added cost of compliance with che L984
regulations may reduce sales, by 0.3 percenc co 1.3 percent.22/
Manufacturers of heavy-ducy engines and vehicles withstood a much
Larger drop in sales around 1975 due co general economic condi-
cions, buc sales are now recovering well. The small decrease in
cotal industry sales from che regulations will be more than over-
come by normal sales growth, and thus can be expecced co have no
nociceable effect on any manufacturer1s growth.
EPA does not expect heavy-duty vehicle sales or the trucking
induscry in general co suffer because of a shift in che mode of
freight cransportation used. Rail and air are not reasonable
alternatives for intracity freight movement. The vast, majority of
"over-che-road" freight movement is done by heavy-duty diesel
trucks. The purchase price and operating coses of heavy-duty
diesels are noc affected by a sufficient amount to warrant anything
but a slight increase in intracity freight hauling costs.
Total bus sales and the bus transportation industry as a whole
should suffer no loss in sales or ridership. The co'sc increases
due co chese regulations will not offset the fact chac buses are
the best option for the cransporcacion of schooL children and
intracity transport. The intercity bus ridership should show ho
137-
-------
decrease because Che per passenger cose af chese regulations is a.
negligible amount when compared co ocher factors in che coca!
cicket price.
Sales by some individual manufacturers of heavy-duty diesel
engines may decline more chan predicted by overall demand price
elasticity. This could result from small volume manufacturers
having to spread their costs for test facilicy modifications
and certification over cheir smaller sales. These costs depend
primarily on che number of engine families certified, noc on che
sales of engines within chose families. Thus smaller volume,
primarily foreign manufacturers like Deucz, Volvo, Hino, Fiat,
Scania 7abis, and others will have larger price rises chan larger
volume, domestic manufacturers like Mack, Deeroic Diesel (d),
Cummins, and Caterpillar. Smaller volume diesel engine manufac-
curers may find che diesel engine martcec less proficabla aS a
resulc.
EPA cannot present manufacturer-specific estimates of how
serious chis reduction in proficabilicy will be. Such estimates
would require accurate projections of each manufaccuers's sales
chrougn L988. SPA does have available manufacturer's own sales for
recent model years and these are shown as fractions of che cocal
mar ice c in Table V-BB. EPA has used chese market shares and che
cost figures from Table V-F7 co estimace che increase in engine
price needed co recover each diesel manufaccurer1s costs. These
estimates are shown in Table V—tTN, in scrambled order and vichouc
manufacturer idencificacion. Generally, che higher increases are
for manufaccurers with low U.S. sales. Ic should be emphasized
chat chis cost analysis has assumed che worst case (i.e., higher
costs) for small volume manufacturers. Therefore, che cost figures
in Table 7-tIN will probably exceed che actual cost increases for
chese manufaccurers. The spread in che ascimaces is considerable,
and in a few cases represents a sizable fraccion of cocal engine
cost. It should be noced, however, chac che manufacturers with che
highest increases produce heavy-duty diesel engines for use in
motor vehicles sold in the U.S. as only one small part of cheir
large, and often multinational, operations. Some presendy enjoy a
price advantage over che larger manufacturers, which will offset
part of all of che differential in price increases. 3ased on
corporate size, product diversification, assecs, and total world-
vide sales, each of the manufacturers with che larger price in-
creases could absorb che cast of chese regulations wichout any
chreat co its corporate survival. Any or all could wichdraw from
che market vichouc any chreat co ics survival and vich liccle
impact on che remaining manufacturers or on competition in che
market. If any or all of che small volume manufaccurers were co
vithdraw from che heavy-duty diesel market, engine availability
would be unaffected and no significant cost increase would occur as
a result of less competition. This is crue becuase che U.S. sales
market is heavily dominated by che large volume domestic pro-
-138-
-------
Table 7-tfN
Estimated Increase La Price Needed for Individual
Diesel Engine Manufacturers to Recover Their Costs of Compliance
Manufacturer Price Increase
1
5 519
2
492
3
492
'4
492
5
243
6
221
7
221
3
213
9
177
10
142
11
123
12
115
13
108
14
108
U The order of the manufacturers has been changed from chat in
Table V-FF and the naaes omitted.
2/ Approximated by dividing the non-recurring cases of Table
7-FT by 5 tines the manufacturers share of 1986 projected sales and
adding the recurring 1934-1988 costs per engine per year. Effect
of interest and profit has noc been included as it would not
significantly affect manufacturer to manufacturer comparisions.
•139-
-------
ducers. The annual U.S. 3ales of the firsc 3 corporacioas in Table
V-NN comprise abouc 5 percenc of che cocal U.S. heavy-ducy diesel
sales per year.
Although Whice appears in che list of smaller-volume manufac-
turers, ic is noc in compecicion wich che larger-volume engine
manufacturers in che same way as che ocher3 in che lisc. In che
past several years Whice has usually certified only military
engines, which do noc compece in che civilian engine markec.
Small volume cruck and bus manufacturers should noc experience
any disadvantage, since most use engines produced by several engine
manufacturers.
Ic is noc expected chac che promulgation of chese regulations
will have any long cerm impact on employment or productivicy in che
heavy-ducy engine or vehicle industries, 3ince industry-vide sales
will be affected little.
2. Impacc on User3 of 3eaw-0ucy Vehicles
Users of heavy-ducy vehicles will be affeccad by che higher
coses for Che vehicles chey use co cransporc goods, and chis
in curn will affect che prices consumers pay for che produces
cransported by crucks.
The expected firsc cosc increases of S195 for vehicles equip-
ped wich diesel engines and S394 for chose wich gasoline engines
should noc subscancially impacc either fleet owners' or an indivi-
dual owner operator's ability Co pay for¦new heavy-ducy vehicles,
since chese costs represent at oosc 3 percenc of a vehicle's sales
price.
The regulations will add less chan one cent per mile of
gasoline-cue led vehicle operacion (undiscountad operating cost
increases divided by cocal life mileage). All operators of gaso-
line-fueled vehicles will incur these cost increases, so no sub-
group will be at a disadvantage. This does not consider che
anticipated increase in fuel economy for gasoline-fueled vehicles/
engines. To vehicle operators as a group, this cost should aoc add
significancly co cheir current vehicle operating coses, and chere-
fore should noc significantly impact either che demand for cheir
transport services or cheir profit margins.
3. Impacc on Fuel Costs co User3 of Other Vehicles
The need for unleaded fuel by gasoline-fueled heavy-ducy
vehicles will increase che demand for chac fuel. However, che
increase will be relatively small, since these vehicles presently
consume less chan 102 as much gasoline as vehicles used for per-
sonal transport ion.1ZJ Also, the increase will ccme slowly.
-140-
-------
The price difference becween Leaded and unleaded fuel should aoc be
changed significantly sinea by 1934 moat Lighc-duty vehicles and
trucks wiLL use unleaded fuel and heavy-ducy gasoline-fue Led
vehicles use only 3 percent of all gasoline consumed.24/ Con-
sequently, there will be oo significant impact an fuel coses co
users of other vehicles.
4. Balance of Trade
The implementation of these regulations will aoc have a
substantial impact on the 0.5. balance of trade.
American manufacturers '«tio sell gasoline-fueled and diesel
heavy-duty engines overseas will build these engines co comply with
che emission standards of- the importing country. Therefore, no
loss in foreign sales is expected as a result of these regulations.
As can be 3een in Table V-NW, che difference in che per engine
first price increase is aoc so great as co preclude foreign manu-
facturers from che U.S. heavy-ducy diesel market. Currently,
all heavy-duty gasoline-fueLed engines soLd in che U.S. are
manufactured domestically.
The use of oxidation catalyses on heavy-duty gasoLine-fueLed
vehicles will cause an increase in che imports of noble metals to
che U.S. The aoole metals, primarily platinum and palladium will
amount co 6.23 grams of-platinum and 3.125 grams of palladium per
engine sold. In 1979 dollars this is approximately ?63 per engine.
however, the predicted fuel economy increase (at Least 42) for
gasoLine-fueLed engines will allow a savings of at least 836
gallons per vehicle over its lifecime. Using 42 gallons per barrel
and a per barrel price of $20, chis fuel import savings ($398
undiscounted) more Chan offsets che increase noble mecal imports.
-141-
-------
References
1/ "Case Estimations for Emission Control. Related Components/
Systems and Cost Methodology Description," LeRoy H. Lindgren,
Rath & Strong, Inc., March 1978, EPA-460/3-78-002.
1J Ail Rath & Strong estimates used were adjusted for inflation.
_3/ These figures include vendor overhead and profit.
4/ See section 3-1 in :his chapter.
5_/ See Che Test Procedure issue in Che Summary and Analysis of
Comments.
6/ Class 113 represents 3,500-10,000 lb. G7WR.
7/ Larger-volume Manufacturers only. It will be assumed that
3ome smaller-volume manufacturers of diesel engines will have
only one emission-data engine selected per engine family, in
accordance with past experience.
_8/ For a more detailed description on how these costs were
computed, see the economic impact issue in che Summary and
Analysis of Comments.
9/ Based on data gathered from EPA's Certification Division.
10/ See Economic Impact issue in Summary and Analysis of Comments.
11/ Analytical Development of Sampling Plans for Selective En-
forcement Auditing, Sylvia G. Leaver, MSED, December 1978.
12/ See economic impacc issue in che Summary and Analysis of
Comments for a more in—depch discussion.
13/ Dun's Review, November 1978, Vol. 112, Mo. 5, pp. 119-121.
14/ Average Lifetime Periods for Lighc-Duty Trucks and 'deavy-Outy
Vehicles, US EPA, ECTD, SDS3 79-24, Glenn W. Passavanc,
November 1979.
15/ See che fuel economy issue in che Summary and Analysis of
Conmiencs for a more detailed discussion.
16/ 1979 Baseline Engine Data available in the public docket
supporting chis rulemaking action.
17/ Conversation with Chuck 3oehl, US. Department of Energy,
Economic Regulatory Administration, September 23, 1979.
18/ Phone communication wich Willard Smich, Economic Analysis
Division, 0PM, EPA - 3a3ed on consultant working papers from
Sobotka and Company, Inc. regarding che analysis of che retail
price differential between leaded and unleaded gasoline.
Final report is under preparation.
19/ EPA memo, "'Estimated Heavy-Duty Gasoline Fueled Engine Fuel
Economy for Che Mid Eighties." Glenn Passavant, November 12,
1979.
20/ The price cross elasticity of demand is a measure of che
proportional change in che quantity of one produce (e.g.,
ga301 ine-fueled engines) demanded resulting from a given
relative change in che price of a relaced product (e.g.,
diesel engines).
-142-
-------
21/ Background Document for Medium and Heavy Truck Noise. Emis-
sion Regulations, Appendix C, EPA. Office of Noise Abatement
Control, March 19 76, EPA-550/9-76—008.
22/ The price elasticity of demand used here considers only the
average first cost increase and does not consider the affect
of the cost of ownership of gaso1ine-fueled vehicles. SPA
knows or ao elasticity of demand taodel for cruck3 which
incorporates both increased first costs and increased costs of
ownership.
23/ Comparing Table VM-l of 1975 Highway Statistics with Table 39
of "Trucking Activity and Fuel Consumption - 19 73, 1980, 1985,
and 1990," FEA, July 1976, PB-263035."
y±] Transportation Energy Conservation Data Book, Edition 3,
February 1979, Oak Ridge National Laboratory. Table 2.3.
About 6 percent of 2 axles single unit trucks were considered
as heavy-duty gasoline-fueled engines (Table 1.26). Calcu-
lated for 1975.
-143'
-------
CHAPTER VI
ALTERNATIVE ACTIONS
A. Introduction
As EPA has proceeded with che development of a final ruie-
sulking based upon analysis of comments received La response co che
February 1979 proposal, alternatives and options in essentially all
aspects of che rulemaking have been evaluaced. Hose of che com-
ments received from manufacturers either explicitly or implicitly
involved alcernacives co icems which EPA had proposed. Thac is,
EPA was requesced co evaluate eliminating, modifying or replacing
slemencs of che rulemaking proposal in a wide variecy of ways based
upon wtiac aanufaccurer3 preceived as defeccs in che proposal, or
more desireabLe alcernacives. Some of che alcernacives raised
during che conmenc period had already been analyzed by EPA, while
some had noc.
- In che Summary and Analysis of CommenC3 decailed analysis of
all Identified alcernacives are developed. This document is
available in che public dockec (OMSAPC-73-4) and che oacerial ic
concains will noc be repeaced in chi3 chapcer beyond che Level of a
brief review of major alcernacives considered. In addicion che
Summary and Analysis of Commencs, Chapcer VII (Cose Effectiveness)
of ciiis Regulatory Analysis considers che emission beneiics and
coses associaced vich each basic alemenc of che rulemaking and
deceraines che resulting cosc effectiveness.
The alcernacives evaluaced by EPA falL Lnco chree broad
areas: 1) alcernacives co specific elements of che rulemaking,
2) alcernacive cining for implemencacion of che rulemaking, 3)
alternative Levels oc stringency for che emission scandards. Each
of chese will be reviewed separacely.
3. Alcernacives co' Specific Elements of che Rulemaking
1. Tesc Procedure
The question of cese procedure was one of che tsosc contro-
versial aspects of che proposal. The Summary and Analysis of
Commencs considers che issue of che new cesc procedure in great
depch in ehe "Tesc Procedure" section of chat document. The basic
alternative consists of promulgating emission scandards based upon
either che exiscing 9-mode and 13-mode sceady-scate cescs, or using
che cransienc cesc procedure. Between chose cwo extremes Lie a
number of variations consisting of modifications co che original
proposal, such as different approaches co deriving che cesc cycle
from che CAPE-21 daca base, field validation programs, alternative
cycLes, etc.
-144-
-------
The fundamental quescion relating co choice of cesc procedure
is che ability of the currenc procedures co adequacely characterize
ia-use emissions of heavy-duty vehicles. All available data
continues co indicate co EPA chat steady-state cests are fatally
defective in this ability, and chat che desired 90 percent emission
reductions in SC and CO could aoc be obcaiaed through cheir use.
Chapter VII estimates the additional emission reductions expected
from use of the cransienc cesc. The resulcs, summarized in Tables
VII-1 for gasoline-fueled engines and VII-2 for diesel engines,
indicate chac che incremental benefit from changing cesc procedures
is substantial and cost effective. Approximately 40 percent of che
overall benefit for gasoline-fueled engines and 60 percent of che
overall beneric for .diesel engines is attributable co implemen-
tacion of the transient test, "the re laced coses are low enough co
sake chac incremental benefit more cosc effective chan chesteady-
scate cesc opcion. Cost effectiveness numbers for chese options
are reproduced below from Tables VII-1 and VII-2.
Cost Effectiveness (S/con)
Ootion
Gasoline Fueled
Diesel
HC CO
HC
Steady-state cesc
349 13
304
Transient cesc
65 2
224
The need for che transient cesc procedure had been a funda-
mental decision included in che proposed rulemaking. Nothing
submitted during the comment period has raised substantial chal-
lenge co chac need. In fact, as che daca jusc summarized indi-
caces, che need for che cransienc cesc for gasoline-fueled and
diesel engines has become even more clearly escablished. EPA has
therefore concluded chac che transient cesc 3hould be retained for
both gasoline and diesel engines. In the case of diesel engines,
there is a possiblity chac existing eddy currenc dynamometers could
be made co perform adequately on che transient cesc. In recogni-
cion of che significant cosc savings chac could result from chis
possibility, EPA has chosen Co allow for diesels an opcional
certification procedure on che existing cast cycle for che first
year of implementation of the new regulations. This delay would
allow diesel manufacturers additional cime co explore che feasi-
bility of uaing exiscing dynamometers.
2. Other Elements
Alternatives relating to che following elements of che rule-
making in addicion co the choice of cesc procedure have been
analyzed: redefinition of useful life, in-use durability cesing,
parameter adjustment, allowable maintenance regulations, assembly.
-145-
-------
line testing with a 10 percent Acceptable Quality Level, diesel
crankcase concroi. For each, there is an appropriate portion of
che Summary and Analysis of Comment3 which can be consulted. In
addition, the coat effectiveness of each element is estimated in
Chapter 711 of this Regulatory Analysis. It is important to
realise that Chapter VII is actualLy an analysis of alternative
rulemaking packages. Each cost versus benefit ratio is derived
from a comparison of the final rulemaking with a rulemaking package,
not having the item being evaluated. Thus, each constitutes a
unique combination package, and each combination represents an
alternative approach to the rulemaking.
For the case of in-use durability testing, this review process
indicated that the proposal should not be promulgated at the
present time. for each of the remaining elements, the basic
approach originally proposed by EPA remains che best alternative.
However, modifications have been made in several of them to improve
their practicability or clarity.
• Alternative Timing for Implementatioa of the Rulemaking
The timing for introduction of new regulations can have very
important consequences. From the manufacturers point of viev it
affects the rate ac which resources muse be expended to attain
compliance, and possibly the very ability to comply. Environ-
mentally, timing defines the point at which desired emission
reductions will begin to be realized.
The proposal had called for implementation of che new regu-
lations in 1983. A great deal of commenc was received from manu-
facturers on the feasibility of that deadline. All comments
indicated that the regulations as proposed were not feasible in
1983. EPA.1 s analysis of these comments and the alternatives which
were suggested is contained in the Summary and Analysis of Comments
under che "Laadtime" issue. The conclusion of chat analysis was
chat gasoline-fueled engines might be able to meet a 1983 time-
table, out the risk of missing the deadline would be high. For
diesel engines, some families could meet a 1983 deadline, but chose
requiring significant emission reduction could not. Therefore, EPA
has chosen to delay implementation of these regulations until 1984.
D. Alternative Levels of Stringency for the Emission Standards
Section 202(a) (3) (E)(i) of the 1977 Amendments to the Clean
Air Act required that EPA "shall conduct a continuing pollutant
specific study concerning the effects of each air pollutant emitted
from heavy-duty vehicles or engines and from other sources of
mobile source related pollutancs on the public health and wel-
fare." The intent of requiring these reports was to provide a
portion of che framework needed to evaluate the statutory standards
for heavy-duty vehicles (HDV) provided in chose same amendments.
-146-
-------
Subparagraph (ii) of the above section indicates chat on the basis
of such studies, and other information available, 'EPA-may change
che statutory standards for heavy-duty engines or vehicles.
Chapter 17 or the Regulatory Analysis assesses the impact of
the statutory standards on emissions of 3C and CO aad on air
quality. Thac chapter, eonbined with the- remaining portions of
this Regulatory Analysis provides a comprehensive review of the
statutory standards. However, in order to make che judgement just
indicated concerning whether the standard should be changed,
evaluation of alternative stringency levels is required. Since
SPA's evaluation of alternative stringency levels is not reported
elsewhere, that analysis will be done here.
The statutory standard provides for emission standards for
both gasoline-fueled and diesel engines derived from a 90 percent
reduction from a 1969 gasoline-fueled engine baseline. To examine
the appropriateness of that 90 percent reduction, two alternatives
will be considered. One is an 85 percent reduction and is less
stringent than che 90 percent statutory standard. The second is a
95 percent reduction from baseline and is more striagenc chan che
statutory standard. These standards correspond to che following
numerical values (g/BSP-hr):
EC
CO
352
standard
1.9
22.2
95Z
standard
0.64
7.7
These alternatives will be evaluated in terms of lifetime
emission reductions per vehicle (and cost effectiveness), changes
in mobile source emissions, and changes in air quality.
1. Lifetime Emission Reductions and Cost
Table IV-A presents lifetime emissions for engines represen-
tative of 1969 baseline levels, 1979 engines, and che final regu-
lations. Similar results can be computed for the alternative
standards. .Emission- factors corresponding to che alternative
standards are given in Table k. of Appendix A. They were derived in
the same fashion as those for the statutory standards. The cor-
responding lifetime, emission races, compared to che statutory
standards, are as follows:
•147'
-------
Lifetime Emissions Ear Heavy-Quty Vehicles (Tons)
Engine Class
and Pollutant
Gasoline-fueled
HC
CO
Opcional
852 Standard
0.24
3.2
Statutory
90Z Standard
0.17
2.4
Opcional
952 Standard
0.10
1.5
Diesel
hc
CO
2.07
5.9
1.41
5.9
0.71
5.9
The effect of changing the stringency of the standard is
significant over the average life of a heavy-duty vehicle. For
example, relaxing the standard to the 35 percent level would
increase HC emissions from gasoline-fueled engines by a factor of
1.4. On the other hand, increasing the stringency would reduce HC
emissions by a factor of 1.7. A similar change occurs for CO.
Relaxing the standard increases diesel HC emissions by a factor of
1.5, while tightening the standard would reduce diesel HC by a
factor of 2.0. Diesel CO emissions are unaffected by a change in
the standard because they are inherently Lower than even the 95
percent standard level.
In Chapter 711, lifetime emission races are used to evaluated
the co31 effectiveness of the rulemaking. To do this for the
alternative standards necessitates assigning coses to these levels.
A prime consideration in evaluating cost variations is the change
in production target levels associated vith the standards. Target
Levels are as follows:
Production Targec Levels (g/3HP-hr)
Engine Class
and Pollutant
Gasoline-rueled
HC
CO
352 Standard
0.73
3.9
902 Standard
0.50
5.9
952 Standard
0.24
2.9
Diesel
HC
1.32
0.89
0.42
Considering gasoline-fueled engines first, the 35 percent
standard will allow some reduction in hardware costs. R&D costs
would be unchanged since mosc of the R&D effort is expected to be
directed toward system durability. Likewise, other components of
ehe per engine cost are not directly tied to the level of the
standard and' would remain unchanged. Hardware cost would include
savings on air pump requirements and catalyst loading. Air pump
-147-
-------
cosc has been estimated ac $26 in Chapcer V (before markup for
overhead and profit). This was che cosc of increasing air pump
capacity the equivalent of two additional air pumps. Ac che 35
percenc standard only one air pump would be required, and che cosc
will be reduced from <*3 g/rc co w g/cc witn a nee savings of
$11. Total cosc change, including markup, is (13 + 11) 1.29 ¦
$30.96. Applied co che cosc per engine from Chapear V of $477, che
35 percent standard would cosc $446.
1 or che casie of che 95 percent standard for gasoLine-fueled
engines, che targec level, for CO is sufficiently Low as co make che
feasibility of this option questionable. There is insufficient
daca ac chis time co determine how low-optimized heavy-duty. cata-
lyse systems will be able co operate. Therefore, no costs will be
estimated for chis case.
Turning next co diesel engines, a somewhat different situation
exists. Much of the cost of reducing engine emissions is in R&D
rather chan in add-on hardware. Hardware costs re Late largely co
chose few engine families which exceed che standards by substantial
amouncs. These are largely unaffected by che change of target
values. Most engine families can attain che desired cargecs by
design changes or calibration changes (e.g., injector design or
injection timing). These actions are included in che R&D costs.
In evaluating che cost for diesels associated with the AQL level
(Chapter Vll, Section E6), changing the target from 0.39 g/BHP-hr
co 1.05 g/3SP-hr was associated with a change of R&D cosc of
approximately 33 per engine. For che 1.32 cargec Level of che 85
percent standard given above, this change will be estimaced as
increasing to a $5 saving per engine.
A second area where diesel manufacturers would be expected co
realize savings from a relaxed standard is in self-audit costs.
The targec level for the 35 percent standard is such that most
engine famiLies expected to be offered in L984 already meet che
standard with substantial margins. Therefore, Less self-auditing
and Less stringent quality control programs would be required.
Quality control costs will be. estimated as reduced by half, and
self-audic races reduced co 0.2 percent. These changes result in a
saving of 311.92 per engine. Total cosc saving is then $5 (R&D) *
11.92 (audic plus quality control) • $16.92. Applied co che cosc
per engine from Chapter ? of S195, che 35 percent standard would
cost $173.
In the case of the 95 percent standard, as was che case with
gaso line-fueled engines, the cargec levels are sufficiently low as
co make che feasibiiicy of attainment uncertain. Therefore, no
costs will be estimated.
would be reduced $13.
Loading
-148-
-------
2. Cose effectiveness
A3 is done in Chapter VII, the benefit from implementing any
of che alternative standards is found by comparing che emissions of
engines built to those standards with che emissions of current
engines. Lifetime emissions for current engines are found in Table
IV-A. For gasoline-fueled engines they are 1.17 cons EC and 3 cons
CO. For diesel engines the current value i3 2.18 cons HC. From
these starting values, che emission reductions per vehicle from the
alternatives are:
Incremental Lifetime Emission Reductions (Tons)
Engine Class
and Pollutant
Optional
85Z Standard
Statutory
90% Standard
Optional
95Z Standard
Gasoline-fueled
ac
CO
0.93
27.3
1.0
28.6
1.07
29.5
Diesel
ac
0.11
0.77
1.47
Using che cost3 estimated above (and allocating gasoline-fueled
engine costs equally betveen SC and CO) the resulting cost effec-
tiveness is:
"Cost Effectiveness (S/ton)
Engine Class
and Pollutant
Opcional
35Z Standard
Statutory
90Z Standard
Optional
95Z Standard
Gaso1ine-fue1ed
HC
CO
240
3
239
3
N/A
H/A
Diesel
ac
1,613
253
N/A
Since it is aoc known if che 95 percent standard L3 feasible
ac this time, no costs have been estimated and, therefore, cost
effectiveness cannot be computed. The 35 percent standard for
ga30line-fueled engines is only marginally Less cost effective than
the 90 percent standard. However, for diesels, the optional
standard is shown to be ouch less cost effective than che statutory
standard. This reflects the fact that at the level represented by
the 35 percent standard very little emission reductions frcm
current levels would be required, while much of che cost would
remain constant (being associated with acquisition of equipment,
testing, etc.).
-149-
-------
3. Changes in Mobile Source Emissions
Since che feasibilicy of Che 95 percent standard is unknown ac
present, the impacts of that option will not be evaluated further.
The effect of Che 33 percent alternative on mobile source emissions
would be significant. The following table compares 1999 mobile
source emissions for the areas selected in Chapter 17 under the
base case (no new heavy-duty standard), the 35 percent option and
the statutory 90 percenc case.
Annual Honmethane Hydrocarbon and
Carbon itenoxide emission in 1999 (thousands of cons)
Optional Statutory
Base Case 85Z Standard 90Z Standard
Non-methane hydrocarbons 1155 1058 959
Carbon monoxide 4317 3082 3038
The statutory standard produces a desiraable reduction in
benefits compared both to the base case and the optional 85 percent
standard.
4. Change in Air Qualcicy
The optional standard being considered, when incorporated into
che overall emission invencory for stationary plus mobile sources,
produces some incremental changes. The average air quality im-
provement for che three cases would be as follows:
Average Percenc Reduction
from 1976 Base fear Realized in 1999
Base Case
Optional
85Z Standard
Statutory
90S Standard
Ozone (rollback/EXMA)
Carbon monoxide
52/29
67
53/30
73
54/31
74
This data indicates that an additional one percent air quality
improvement for ozone (either rollback or EdA model) and for
CO can. be associated with the 90 percent standard over che 35
percenc standard. For ozone, this is half of the tocal improve-
ment.
5. Conclusions
This analysis has concluded chat the feasibilicy of attaining
the earget emission levels associated with the more stringent 95
percent standard i3 not known for certain. Therefore, chat stan-
dard iaaoc a desirable alternative.
-150-
-------
The 85 percent standard ai.cernaci.ve resuics in a Loss of
benefits and 3ome reduce ion in cose. These changes are in such
proportions chac che cost effectiveness of the regulations becomes
prohibitive for dieseL engines. In addition, approximately half of
the ozone air quality benefit of che statutory standard would be
Lost under the 35 percenc standard.
The statutory standard represents the best of the three
choices at che present cime.
¦151-
-------
CHAPTER VII
COST EFFECTIVENESS
A. Mechodolotty
Cose effectiveness is a measure of what might be earned cite
economic efficiency of some action directed coward achieving some
goal. Expressed as cose per unit on benefit achieved, cost effec-
tiveness can be used co compare various alcernacive nechods of
achieving che same goal. In ehe concext of improving air quality,
che goal is co reduce emissions of harmful poLlucancs, and cosc
effectiveness is expressed in ceras of the dollar cosc per con of
pollutanc caacrolled.
To evaluate cosc effecciveness, cwo pieces of information on
che alternative being evaluated are aeeded. These are che cosc of
the alternative and che benefits to be gained. Costs Co be used in
this chapter vill be cotal identified costs expressed on a per
engine basis, including both costs co che manufacturer and coacs co
che operacor (all discounted co January L of che model /ear Ln
which che vehicle is produced). These costs vill be allocated
equally among che pollutants being concroLled. The benefits vill
be computed as cotal lifetime emission reductions per vehicle.
In chis chapcer, che rulemaking provisions vill be subjected
co cwo distinct analyses. The first vill be an incremental analy-
sis of each of the major components of che package. The second
vill be an analysis of the package as an integrated scrategy. The
purpose of these cvo approaches are different, and che reader is
cautioned against misinterpretations of che incremental analysis.
In che incremental approach, the effect on costs and benefits of
removing individual components vill be examined. To varying
degrees, both costs and benefits of chese components- overlap and
several components of che package may acc cogecher co oocain a
given benefit. Cn such a case, loss of any one part of che package
can result in a disproportionate loss of benefits. There are so
many overlapping interrelationships chat it would be impossible co
consider every possible combination of che various components of
che package. This analysis vill instead Look at the single set of
options produced by deleting each component one at a ciae. The
cotal Loss of benefits produced by deleting a. component vill be
associated with ehe cost of chac part of che package. Therefore, if
one were to simply sum incremental costs or incremental benefits as
an attempt at obtaining total costs or benefits, significant
amounts of double counting would occur. Such a procedure would be
invalid. The integrated cosc effectiveness analysis muse be used co
evaluate overall coses or benefits.
3. Background
In che draft Regulatory Analysis which accompanied che ?ro-
-152-
-------
posed regulations, a cose effectiveness anal/sis or che proposal
was carried auc. Thac analysis considered che overall cost effec-
tiveness of che entire proposal as an integrated compliance strat-
egy. £c was indicated ia che report that "in a multi-faceted
program such a s che proposed regulations of this heavy—ducy pack-
age, ic would be desirable co analyze separately che coses and
environmental benefits of each aspect of che package. In chis way
a decision could be made on each element as co Lcs cost effective-
ness and whecher ic should be incorporated into che final regula-
tions." However, data was Lacking with which co quantify che
benefits associated with individual elements of che proposal. In
addition, it was painted out chat che benefits are inter-re Laced
and cannot be LsoLaced easily from each other.
During che course of che comment period on che proposed
regulations, EPA has endeavored co develop more data, for example
on che comparisons between steady state and cransienc case emis-
sions, and escabLish methods cor estimating changes in emissions
which could be associated with changes in che various components of
che package. This effort has been sufficiently successful co allow
estimated cost effectiveness analysis for che main components of
che rulemaking.
It is important co bear in mind chat the benefits and costs in
chis anal/sis will overlap, and chat summing chem all would resulc
ia double councing. For example, consider che case of extended
catalyst lifetimes required under che allowable maintenance provi-
sions and che revised useful life definition. The benefit of
Increasing cacalysc lifecimes is significant. however, if che
useful life remained at 50,000 miles, che intent of che allowable
maintenance provision for catalyst change intervals would be lose,
therefore, incremental analysis of che allowable maintenance
interval and revised useful life will each separately be looking at
partly the same benefit in emission reductions.
Allowing che benefits to overlap in chis fashion may appear co
give coo much credic co individual elements of the package, this
is aoc true, since ia each case the benefit considered will be che
best estimate of what che package would actually gain or lose if
chat element were retained or removed. The purpose of an incremen-
tal analysis is co answer thac question for each element. Although
it would be desirable, ic is not che chief purpose of an incremen-
tal analysis co evaluate the benefits of che totai package. The
benefit attributed to che overall integrated package will be
determined separately.
C. Summary
Using all data now available (both chac generaced by EPA and
chat submitted co HPA during che public comment period on che
proposed regulations), an analysis of the cost effectiveness of
-153-
-------
each major element of the regulation package and of the overall
package as a unit has been done, this analysis developed benefits
expressed as cons of pollueanc removed (either HC or CO) over the
average lifetime of an individual vehicle along with cocal cases
for the same lifetime (discounted co year of sale).
Overall benefics and costs used as a starting reference the
existing regulations for 1979 and later model year heavy-duty
engines. That Is, boch overall benefics and overall costs were
developed as changes in relation co the case of the existing
regulations continuing in effect. 3enefics and costs for most of
the individual elements of the package (except for the change co
the transient test procedure), on the other hand, were evaluated in
cerss of changes to the final package. The loss in benefits chat
vould occur if each element vere removed from, the package vas
evaluated La comparison with the cost reduction that vould be
produced by chac sane change. The transient test procedure «as
evaluated in relationship to the alternative of implementing a
standard on the current test procedures corresponding co a 90
percent reduction from the 9—node gasoline-fueled engine 1969
baseline. Figures VII-L and 711-2 summarize the benefits developed
for gasoline-fueled and diesel engines, respectively. Costs,
benefits, and cost effectiveness are tabulated in Tables V1I-L and
711-2. Cose effectiveness figures for other mobile source control
strategies are provided in Table 711-3 for comparison purposes.
D. Gasoline-Fueled Engines
I. Transient Tesc
One of the key facets of the entire regulation package is the
proposed transient tesc procedure. The transient tesc is being
implemented because of che need co sake compliance testing for
heavy-ducy engines a better measure of actual ia-use emissions.
The CA?2-21 program and resulcanc transient test procedure vere
designed co accurately characterize in-use operation and therefore
ia-use emissions. Throughout this analysis, transiene cesc data
will be caken co represent in-use daca. The 9-mode steady state
procedure fails to accurately measure in-use emission races for che
Low emission engines currently being used or for advanced techno-
logies anticipated for future heavy-ducy engines. This same
failure of the 9-mode cesc co relate well co in-use emissions has
made it difficult co assess the benefit of svicching co che tran-
sient tesc. If one does not know che ia-use emissions from che
current test Chen che actual benefit or reaching a given level on
che transient cesc - cannoc be quantified. Thi3 was che dilemma
facing EPA ac checime of che proposal in February 1979. Although
there was reason to believe chac substantial benefit would accrue
as a result of implementing che cransient cesc, chere was ao way co
quantify that benefit.
-154-
-------
Figure 711-IA
Incremental Lifetime Hydrocarbon Benefits
Gasoline Fueled Engines
102 AOL £1
Allowable Maintenance
Parameter Adiuscaient
9—Mode Test
0 0.2 0.4 0.6 0.3 L.O
Lifetime EC 3enefit (tons/vehicle)
Figure 711-IB
Incremental Lifetime Carbon Monoxide
Benefits - Gasoline Fueled Engines
10% AQL
Allawable Maintenance f
Parameter Adjustment
Useful Life
Q-Mora Taqc
Transient Test
T
t-
10
—r—
15
5 1U lo 20 25
Lifetime Carbon Monoxide Benefit (tons/vehicle)
-------
Figure VI1-2
Incremental l.lfetJniu KiiiI&sIoii
lleneflta Diesel liugtnes
i
e
o\
10% AQI.(_
Allowable Maintenance
C
Useful 1.1 fe [
I 3-Motle Teat
TranHlunt Teat
—i
0.6
—I—
0.2
—i—
0.4
0.77 O.fl
l.lfetliue lie Uenefit (tons/vehicle)
-------
Table VII-I
lacremeucal Lifetime Cose Effectiveness
Gasoliae-?uelad Engines
Cose
Seaef it
(Tons)
Cose Effecciveness
(S/Ton)
Oocion
(Dollars)
ac
CO
AC
CO
90 Percent on
9-Mode
426
0.61
16
349
13
Transient Test
51
0.39
12.6
65
2
Useful Life
58
0.13
3.3
223
3
Parameter Adju3t-
aent
5
0.08
3.6
31
1
Allowable Main-
tenance
58
0.13
3.8
223
3
1Q Percenc AQL
17
0.04
0.5
213
17
Overall Package:
477
1.0
28.6
238
8
a. Transienc Test
o. Useful Life
c. Paramecer
Adjustment
d. Allowable
Maincanance
e. 102 AQL
I/M
29
0.07
2.3
20 7
V
6
* Sote: Includes all ocher aspects of che rulemaking axcspc che
transient cesc.
-157-
-------
TabLe VII-2
Incremental Liretime Case Effecciveness
Diesel engines
Coac Effectiveness:
(S/Tcra)
Seeady-Scace Tesc
0.23
304
Transient tesc
110
0.49
224
Useful Life
0.Q5
allowable Maintenance
0.05
100
10 Percent AjQL
Crankcase Control*
Overall Package
L95
0.77
253
a. Transient Test
b. Useful Life
c.. 1QZ AQL
d. Crankcase Control
* Naturally-aspirated engines only. Cose allocated over 3
pollutants. See Diesel Crankcase imissi-oo Control, Summary and
Analysis oc Co«s»nC3.
-LS8-
-------
Table VII-3
Cose Effectiveness (S/Ton) Comparison
Wich Ocher Emission ConcroL Strategies
ConcroL Program
LDV Statutory
Standards b/
LDT Interim
Standards c/
Emissions
After Control
Baseline Shmssion &! Program Initiated
I)
2)
l/M for Existing
LDV3 il
Motorcycle Standards
1978/1979 %j
1980 +
Proposed LDT
Action H
ac
CO
NOjc
ac
CO
MOx
ac
CO
NOx
ac
CO
HC
CO
ac
CO
1.5
15
3.1
2.0
20
3.1 d /
4.3
44
5.2 a/
9
34.67
3-22.5
27.4
1.7
13
ac
CO
NOx
ac
CO
JKIx
0.41
3.4
0.4
1.7
1-8
2.3
MOx » 2.3
ac
CO
ac
co
ac
CO
HOx
3-22.5 h/
27.4
a
19.3
0.3
10
2.3
Cost Effectiveness
($/Ton)
ac CO NOx
470
200
78
364
41
21
2300
73
7.7 2763
Neg.
365 Neg.
139-201 10-12
li Emission Levels in grams/mile, axeape for RD which are g/BHF-hr.
b/ Report: Interagency Task Force on Motor Vehicle Goals 3eyond 1980, March 1976.
c/ "Environmental Impact Statement - Emission Standards for Light-Duty Trucks,"
November 29, 1976.
d/ Trucks 0 - 6,000 lbs. GVWR.
e/ Trucks 6,001 - 8,500 lbs. GVWR.
£J "Cost Effectiveness Estimacad for Mobile Source Emission Control," Vector
Research, Inc. for EPA, January 1978.
£/ "Environmental and Economic Impact Statement - Exhaust and Crankcase Regulations
for the 1978 and Later Model Tear MocorcycLes."
h/ Sliding Scale Based on Engine Displacement (cubic centimecers).
U "Draft Regulatory Analysis of Proposed Emission Regulations for 1983 and Later
Model Year Light-Duty Trucks", EPA Office of Mobile Source Air Pollution Concrol,
June 28, 1979.
-159-
-------
In order co quantify cite benefit af che cransienc case, EPA
began testing engines on che cransienc case which represents
current technology and future expected technology. This testing
program measured emissions on. both che cransienc cest and che
9-mode test in order co assess the in-use levels represented by
amissions on che 9-mode test. In addition, in .preparation for
development of a 1985 SOx standard, a 1972-1973 baseline engine
cescing program is underway. Altogether, at che cime of this
analysis, there is data on five current technology and prototype
technology engines, twelve 1979 engines (representing 862 of
projected 1979 sales), 3even 1972-1973 engines (representing ^6Z of
1973 sales), and 15 of che 1969 baseline engines. Hoc all 1969
baseline engines can be included because 9-mode emission cescing
was not carried out on all engines. Emission levels, from chese
engines cover a broad range from pre-controlled Levels down co che
range approaching che new standards. Scatter diagrams of transient
emissions versus 9-mode emissions are given in figures 7II-3 and
711-4.
Three principal conclusions can be drawn from a review of chis
data. First, there is a high degree of scatter in che relationship
between che two cest procedures. For example, ac a 9-mode CO Level
of approximately 15 g/BHE-aR (che range of che new CO standard),
transient emissions vary from near 30 co abouc 120 g/3HP-hr.
Second, in spice of ¦ chis- scatter chere is a rough relationship
between transient and 9-tnode emission results. Linear regression
lines of traiisienc emissions versus 9-^node emissions are shown in
che figures. Two lines are shown in each figure. The first is an
estimate using only the pre-controL 1969 and 1972/73 baseline
results. Ihe second uses 1979 and advanced technology engine data
as well. The relatively 3mall change in the regression line
resulting from the addition of the lacter data set indicates some
stability in ehe relationship. The regression lines based upon all
che available data will be used in the remainder of this analysis
to relate given 9-mode emission levels to equivalent transient
levels. Finally, Figures VTI-3 and VTI-4 show that there is a
limiting range of transient emissions below which che 9-mode
procedure is not capable of reLiably measuring, and that this level
is well above the new transient emission standards for both 3C and
CO (1.3 g/BHP-hr HC and 15.5 g/BHP-hr CO). 9-mode emissions can be
seen co be approaching zero, while transient emissions remain at or
above the standards. This can be clearly seen from the intercepts
of the regression equations of 2.1 for SC and 62 for CO. The
impact of this fact is chat there is a iimic co the amount of
actual benefit which could be expected co result from a 9-mode
standard even if set ac an extremely low level. It is also an
expected result, since the 9-mode test measures no transient
components of che overall emissions. This effect should noc be
interpreted to mean that heavy-duty engines will not be able co
reach low transient emission levels, but only chat 9-mode testing
is incapable of identifying chose which can versus chose which cdn
-160-
-------
Trunalent
luilitJUiOllH
la.cao
ib.ooo
It.oao
12.000
10.000
a.oooo
a.oooo
4.oooo
2.00)0
0.
llljure VI1-3
Trunulciit vu 9-Mode lunlaalons
for llydrocurbonu
1972/73
Baseline /
I •
1960 baseline
All flaca
,,1969 llaaellne plus
197 2 / 7 J Uatitil 1 ite
Uaiicd oit all data
I1C (Traits liiitc) = 2.1 + 0. 72(IIC(9-inode) )
1979 Baseline plus
Currenc/l'rutoLy|iu Tcclinulo|$y
o.
t.UUOQ i.ujilo |2.004 to.000
2.0000 c.OuOH 10.000 IS.000 |a.000
9-Moilu Emissions
-------
KI (jure VI1-4
Transient
frulaslona
a\
Ki
I
no. oo
2*o.oo
210.OS
180.00
Truna lent v a 9-Mode limiaaiona
for Carbon Monoxide
1972/73
Baaellne^^^
1969 Baaellnc
90.00(1 4*
A11 Data
1969 Baseline |>lua
1972/73 baseline
JiO.Ol)
120.00
Based on all data
CO(Vranalenc) ¦= 62 + 0. 77 (CO(9-Mode) )
60.000
30.000
*/
s/
| 1979 Baueline plus
| y Current/Prototype Technology
~
o.
0.
oO.bC'j |2o.(ib lob,00 2<»Q.OO
IlO.ftbo Vu.OUJ ldi>*0U 210.00 k7U*00
9-Moila liini salons
-------
aoc. Baaed upon boch che scatter of che data and che regression
lines, chare is no 9-mode standard which could reliably attain che
desired 90 percenc reduction in actual amissions.
Using che daca of Figures VII-3 and VII-4, che emission
benefits of converting co che cransienc esse can be estimated. we
will first evaluate the average lifetime emissions of an engine
controlled Co a 9-mode standard representing 90 percenc reduction
from che 1969 9-mode baseline levels. Next, average Lifetime
emissions from an engine controlled ca che transient standard will
be evaluated. Comparison of che cvo will indicace che benefit of
the transient test. Since che only variable we desire co examine
is che change in test procedure, all other aspects of che rule-
making vill be allowed co remain intact vith eicher cest procedure.
That is, boch the 9-aode and cransienc standards vill be evaluated
based upon che inclusion of redefined useful Life, paramecer
adjustment, allowable maintenance and a 10 percent AQL. These have
che effect of maintaining in-use emissions at or near levels of
certification engines.
a. 90 Percent on che 9-mode
A 90 percent reduction from baseline standard on che 9-moae
represents levels of 1.0 g/3HP-hr HC and 12 g/BHP-hr CO. The SC
standard is being met by currenc engines, but che CO is not. In
fact, based upon che CO requirement, SPA believes chat cacaLytic
converters would be needed co achieve compliance.
Production engine carpets (certification levels) for che above
scandards can be calculated in che 3ame fashion as that developed
in Section 5 below. In that section, che produccion "argec is
estimated as (0.65) x (s tandard/!>. F. ) for a 10 Z AQL. The OF
appropriate Co heavy-duty catalyst equipped engines is derived in
che "allowable maintenance" section of che Analysis of Comments as
1.7/100,000 miles. Therefore, the production cargec levels would
become 0.65 x 1.0/1.7 - 0.38 g/BHP-HH (HC) and 0.65 x 12/1.7 - 4.6
g/BH?-H3. (CO) as measured on che 9-mode cest.
From this starting point, then, we vill let in-use emissions
increase at a race corresponding co a OF of 1.7 - so long as che
cacalyst system remains operational. Average emissions under chae
condicion can be expressed as:
BC ¦ 0.38 + 0.027 (M/10,000) 9-mode
CO • 4,5 + 0.32 (M/10,000) 9-node
where M ¦ mileage
The relationship becveen cransienc and steady state, as given
in Figures 7TI-3 and VII-4, are:
HC(Transient) -2.1+0.72 (ac(9-mode)] g/BHP-nr (VII-1)
CO(Transient) » 62 * 0.77 [C0(9-oode)] g/3HF-hr (VII-2)
-163-
-------
Equations CVII— L) and (VII-2) can be used Co estimate cran-
aient emissions relative co che 9-mode. values jusc computed. The
resulcs are:
BC - 1A * .02 (M/10,000) g/3EP-hr (VII-3)
CO » 65 + .25 10,000) g/BHP-hr (VII-4)
la using che resales of equation (VII-3) and (VII-4), ic is
specifically assumed chac engines designed co che 9-mode scandard
would produce in—use emissions according co che relationship of
equations (VIX-1) and (VII-2). What has not been considered is che
possibility chac aev engines may be designed co "beat" che cycle
and produce low 9-mode results vichout a corresponding lowering of
transient (in-use) results. This is a. real possibility, but cannot
be quantified. Insofar aa some manufacturers were to choose such a
route, che benefits here being attributed co che 9^jode procedure
would be lost.
Under che allowable maintenance restrictions of chis rule-
making, catalysts are expected to have a minimum change interval
of 100,000 miles. However, based upon probability, not all cata-
lysts will have exactly the same lifetime. Hor will all catalyses
need co be changed at che same point due co failure. Me will
creac each cacalyac as having a finite lifetime, beyond which
emission performance will begin co degrade ac a rapid rate. This
could result Erom occasional high-temperature conditions or other
operating conditions which will affect system integrity, or ran-
domly occurring factors during catalyst system manufacturer which
affect durability of che system as extended mileage accumulates. A
distribution generally found appropriate for lifetime phenomena is
che Weibull distribution. 1_/ This distribution has , che form:
F • 1 - axp [-C£bb] (VII-5)
9
If a catalyst were co fail on an in-use vehicle with extended
mileage, it is quite possible chac ic would not be replaced.
Therefore, average in-use emissions will increase somewhat near che
end of che useful life period. Ic is reasonable co assume that a
vehicle with a failed catalyse would emit at levels of 1979 en-
gines. The failed catalyst mode emissions would be:
SC - 3.0 +0.02 (M/10,000) g/BHP-iia Transienc
CO » 88 +0.22 (H/10,000) g/BHP-a£ Transient
The sera mileage levels in che above equations come from EPA'a
1979 HD baseline testing program. The mileage factors are based
upon a review of che 1979 certification OF's (which for current
engines are addicive values). All HC DFs but one were zero, and
che average CO OF was 1-1. 3ased upon che belief chat in-use DFs
for SiC would be nan-zero, a OF of 0.1 was used for SiC in combi-
-164-
-------
nation vich che 1.1 far CO. These are addicive values over 30,000
miles.
The in-use fleec average emission race can be escimaced by
combining che cacalysc and non-catalyst emissions according co che
traceion of failed catalysts (?):
EC - [2.4 + .02 (M/10,000)| [1-F] + (F] [3.0 + 0.02 (11/10,000)1
CO =¦ [65 + 0.25 (K/10,000)I [1-F] + [F] [88 + 0.22 (M/10,000)]
Combining cenns, chese can be simplified co che following:
SC ¦ 2.4 + .02 (M/10,000) + 0.6F g/3HP-HR Transienc (VII-4)
CO « 65 + 0.25 (M/10,000) + F[23 - .03 (H/10,000)]
g/BHP-HH Transienc (VII-5)
The only remaining cask is co specify che fraction of failed
cacalyscs (F) according co che W'eibull distribution of equation
VII-5. We will consider cacalysc syscem design such chac che
cacalysc change poinc of 100,000 niles corresponds co a maximum of
10 percenc railed cacalyscs. We will furcher assume a "Veibuil
3lope" of b«3. 3ased upon chese cvo faccors, che "characceriscic
value" 8"211,726 miles. A ploc of chis function is given in Figure
VII-5 and illuscrations of che effect ic has on increasing emission
races at high mileage can be found in Section 2 under che discus-
sion of useful life (Figure VII-6). The equation for F is:
F - 1 - exp [-, . M ^~31 (VII-6)
4ll, 726J
For an average Lifecime of L14,000 miles, chis equation
indicates chac about 14 percent of che cacalyscs would be expecced
co fail. Equacions (VII-4), (VII-5), and (VII-6) can be used co
evaluate lifecime emissions. The vehicle lifecime used is 114,000
miles.2/ Over chis lifetime, che average emission race i3 2.54
g/BHP-hr (HC) and 67.3 g/BHP-hr (CO). To convert chese races co
cons of emissions over che whole vehicLe Life involves che rela-
tionship between fuel consumption per BHP-hr and fuel consumption
per mile of truck travel:
% x BHP-5R x 6.1 lb. fuel x gallons x 114,000 miles x con
BHP-aR lb. fuel gallon mile Lifecime 454x2000
From che SPA 1979 baseline cesting program, che 3ales weighced
fuel consumption per 3HP-hr (inverse of che second faccor) is 0.7,
and che fuel economy is 5 mi/gal. Using chese aumbers, che conver-
sion faccor is:
6.1 x 114,000 =• .22
0.7 x 5 x 454 x 2000
-165-
-------
99.
95.
90.
80.
70.
60.
50.
40.
30.
20.
10.
UJ
a
3
< 5.0
z 4.0
'JJ
U
S 3.0
a.
2.0
1.0
I [ ] I J I
-------
The resulting lifetime emissions of a. catalyst-equipped
gasoLine-fueLed heavy-duty engine designed co meet a 90 percenc
reduction standard on the 9-mode cesc procedure are:
0.22 x 2.54 - .56 cons HC and 0.22 x 67.3 » 15 cons CO
Lifetime emissions of Che 1979 in-use fleet which constitutes
our baseline reference case have been estimated in chancer IV as
1.17 cons HC and 31 cons CO. The net benefit of che final rule-
making, if it were co be based upon a 90 percenc reduction on che
9-mode, would be 1.17 - .56 ¦ 0.61 cons HC and 31 - 15 3 16 cons
CO.
b. Transient Tesc Procedure
The impact of converting co che cransienc test procedure will
be lower certification emission Levels. As noted for che 9-mode
3Candard, che production carget Level for a 10 percent AQL has been
estimated in Appendix A at (0.65) x (standard/D?). Using che
cransienc standards of 1.3 g/BHP-hr HC and 15.5 g/3HP-hr CO,
these carget levels are:
0.65 x 1.3/1.7 =» 0.50 g/BHP-hr Transient HC
0.65 x 15.5/1.7 ¦ 5.9 g/BHP-hr Transient CO
Based upon che OF of 1.7/100,000 miles, chese can be expressed
as (so long as catalyst operates):
HC - 0.50 + 0.035 (M/10,000) g/BHP-hr
CO * 5.9 + 0.41 (M/10,000) g/BHP-hr
Adding in failed catalysts we gee:
HC - [0.50 * 0.035 CM/10,000)] [1-F] + ("1 [3.0 + 0.02 (M/10,000)]
CO - [5.9 + 0.41 (M/10,000) [1-F] + (F] [38 + 0.22 (M/10,000)]
Which simplify co:
HC « 0.50 + 0.035 (M/10,000) + F [2.5 - .015 (M/10,000)] g/BHP-hr (VII-7)
CO - 5.9 + 0.41 (M/10,000) * F [82 - 0.19 (M/10,000)] g/3HP-hr (VII-8)
Equations (VII-7) and (VII-8) yield average Lifecime emission
rates of .76 g/BHP-hr HC and 11 g/BHP-hr CO. Converting to life-
time emissions in tons ve get .22 x .76 ¦ .17 tons HC, .22 x 11 a
2-4 tons CO. The net benefit over the 90 percent on the 9-mode
scandard is .56 - .17 ¦ .39 cons HC and 15 - 2.4 ¦ 12.6 cons CO.
The emissions calculated for che transient cesc case are che
emissions for :he entire rulemaking package. Therefore, che
benefic of che entire package can be obtained by comparing chese
levels co chose previously calculated for the 1979 in-use fleet
-167-
-------
(1.17 cons 3C and 31 cons CO). The emission benefit over che
baseline Levels is chen 1.17 - .17 ¦ 1.0 cons SIC and 31 - 2.4 «»
28.6 cons CO.
c. Coses
The cost analysis of Chapcer V determined coses attributable
co various aspects of che regulation package. Oiscounced coses per
engine are given in Table V-LL. Applying che profit and overhead
factor or 1.29 developed in Section 3-1 of Chapter V, chese values
become:
Icem Discounced Cosc oer Engine
El * D
19.58
Certification Testing
3.16
Certification Facilities
10.40
SEA Facilities
14.16
SEA Testing
1.04
Self Auditing
19.43
Hardware
326.37
Unleaded Fuel
238.72
Muffler and Spark Plug Saving
-176.13
Tocal Cosc Per Engine 476.74
This cocal cost corresponds co che transient cest benefits
derived earlier and can be used co compuce che cosc effectiveness
of che overall package as given in Table VII-1.
Cosc per engine for che case of a 90 percent standard on che
9-mode cesc can be derived by che following changes. Certification
facility costs are eliminated. The SEA facility cosc becomes
$11.73 due to a savings of 9300,000 per SEA cesc sice by delecing
che need for CVS syseems (equipment plus facility). The catalyse
required co meec che 9-mode seandard will also be Less expensive.
SPA estimates a reduccion in catalyst Loading from 45 gm/ft co 38
gm/fc nd reducing che catalyst volume to engine CID ratio from
1.0 co 0.9. These changes reflect the less stringent standard, but
che continued requirement co perform over che full useful Life.
These changes reduce catalyst cost by $32.23. An additional $6
vould be saved because of air pump modifications required for che
full package but noc for che 90 percent on che 9-mode case. Tocal
cost reduction is then $10.40 (cert facilities) * $2.43 (SEA
facilities) * $38.25 (hardware) ¦ $31.08. Cost for che 9-mode case
is thus $476.74 - 31.08 ¦ $423.66, and the cost attributable co che
transient cest is $51.08. Resulting cost effectiveness is shown in
Table VII-1.
2. Redefinition of Dseful Life
In section lb above, che Lifetime emissions per vehicle using
-168'
-------
che transient case procedure were calculated co be 0.17 cons of ciC
and 2.4 cons of CO. These numbers, as has been noced, presumed
chac ail other aspects of che rulemaking were incact. The basic
assumption made ia chat regard was that che combined package would
result in in-use emissions which closely match che performance of
certification vehicles. The only exception was due co che failure
of a small percentage (14 percent of cocal) of catalyses on a
random basis near the end of che vehicle useful life.
The evaluation of the new useful life definition viil proceed
by estimating che loss of benefics and reduction in costs chat
would occur if this element were removed from che package while all
other'elements remained incacc. This method will make it possible
co evaluate che impact of aoc implementing useful life on che
overall package while at che same time escimaciag che cosc effec-
tiveness of chis element.
a. 3enefics
The extension of che useful life definicion co che average
full lifetime rather Chan someching approximating half of che full
life as is done in current practice has che efface of requiring
chac vehicles will be able co meet emission standards throughout
their average life. This will require new vehicle emission rates
to be lower so as co not exceed che standards after accounting for
emissions decerioracion over aporoxiaacaly cwice che mileage
interval of current practice. Full life useful-life will also
require the use of control syscems which are sufficiently durable
co lasc che vehicla'3 Lifetime. This makes che useful life
change a key to che effectiveness of che allowable maintenance
provisions.
These cvo aspects of full life useful-life - lower initial
emission races and more durable components - provide che basis for
estimating che benefits of chis element of che rulemaking. If full
life useful life were dropped in favor of che current 50,000-mile
useful life then both of chese areas would suffer. Emission target
levels would increase and system durability would aoc have co be
proven beyond 50,000 miles. The latter fact would have its major
emission impact in relation co catalyses. If catalyst durabilicy
need only be proven co 50,000 miles Chen a "50,000 mile catalyst"
will be used instead of a "100,000 miles catalyst".
following the procedure used in section lb, chese changes can
be quantified. We will use a cacalyst system DF of 1.3 over 50,000
miles co gee production target levels, which have been noted as
(0.65) x (scandard/DF) for 10 percent AQL. These are 0.65 x
1.3/1.3 - 0.65 g/BHP-hr HC and 0.65 x 15.5/1.3 - 7.7 g/BHP-hr CO,
These levels form che acarting point for vehicles whose catalysts
remain intact. For chose catalysts chat fail che emissions will be
as used in section lb. The resulting emission races for a 50,000-
miles useful life are:
-169-
-------
HC ¦ 0.65 * 0.039(11/10,000) + F(2.3 - 0.019(M/10,000)] g/BHP-hr (VII-9)
CO - 7.7 * 0.46(M/10,000) + F[80 - 0.245(11/10,000)] g/BHP-hr (VII-10)
Equation (VII-6) for F, when modified for a 50,000 mile useful
life, becomes:
F " I - exp [-f M > 3]
105,363
la chis case, catalyse failures will be much greater, wich up
co 70 percent having gone beyond cheir lifetime by che end of a
114,000.-mile useful life. To illuscrate chis, Figure VII-o pre-
sents che average HC amission race as a function of vehicle mileage
for che 50,000 mile useful life case and the comolece rulemaking.
The average 3C emission race rises from 0.76 g/BHP-hr for che
complete rulemaking co 1.36 g/BHP-hr wich a 50,000 mile useful
life. Over ics full life, a vehicle in che latter category would
emic 0.22 s 1.36 ¦ 0.30 cons SC and 0.22 x 28 ¦ 6.2 cons CO. The
nec loss in benefic3 from eliainacing che useful life changes is
Chen 0.30 - 0.17 * 0.13 cons HC and 5.2 - 2.4 ¦ 3.8 cons CO.
b. Coats
Costs attributed co changing useful life from 50,000 miles co
100,000 miles are basically che costs of building more durable
catalyst systems and meeting a lower emission targec. For a 50,000
mile useful life, catalyse loading is reduced co 30 g/fc3 and
cacalysc volume co engine CIO ratio is reduced co 0.9. Cast
reduction is 958.05 per engine. Cost effectiveness aopears in
Table Vil-1.
3. Parameter Adjuscment
The parameter adjuscment provisions of chis rulemaking are
designed co correct one of che largest causes of excess emissions.
EPA's restorative maintenance study3/ examined che incidence of
maladjusted parameters in 300 light-duty vehicles in 3 different
cities. For chese vehicles, which were less Chan one year aid and
had less than 15,000 miles accumulated, it was determined chat over
72 percent were maladjusted on at least one specification for
ciaing, idle CO, idle 3PM.4/ As discussed in the parameter adjust-
ment portion of che Summary and Analysis of Comments, similar rates
are expected for heavy-duty gasoline-cueled engines.
The benefits co be derived from parameter adjuscmenc are
likely co be che same for heavy-duty as chose idencified for
light-duty in che restorative maintenance study. The study indi-
cated an average reduction in HC of 32 percent and CO of 60 per-
cent, for che 300 vehicle fleec.5/ Individual vehicles experien-
cing maladjustment showed significantly greater reductions chan ch'e
-170-
-------
00:5021
mvm I
p5?r>g| |u |^»p|n/\
mam mm g hhhhii
0P3MK
CI = p*rpums
3JII tnJ3sn
" 1.1m 000*05
fflnsmj 9A P3R}| ;j||
7»3nj 3AV Pl||5|?H-Jfln1J
0
0
(J'l-Hllll/f)
5 nil
0* I
5* I
0'E
5"E
I
H
9-1 IA njiiH-u
-------
average. The average reductions, ic should also be noted, vere for
Low-mileage vehicles which had been screened co eliminate vehicles
which had been abused or extensively modified. Therefore, whole
Life impact of parameter adjustment will probably be greater than
che benefits being estimated here.
The loss of benefits chat would be experienced by elimination
of parameter adjustments is calculated as foLlows:
Lifetime emissions with parameter aduscments ¦ 0.17 tons HC
2.4 cons CO
Lifetime emissions without parameter adjustment "
0.17/(1-0.32) ¦ 0.25 cons SC
2.4/(1-0.60) - 6.0 cons CO
Lifetime benefit of parameter adjustment ¦
0.25 - 0.17 - 0.08 cons HC
6.0 - 2.4 - 3.6 cons CO
The costs of implementing parameter adjuscmenc have been
addressed in che Economic Impact section of che Summary and Analy-
sis of Comments. Parameter adjustment is expected co cost approxi-
mately 35 per engine. Cost effectiveness based upon chis amount is
given in Table VII-L.
4. Allowable Maintenance Restrictions
These regulations will affect a wide variecy of emission
related components. The overall impact will include decreasing che
amount of emission maintenance required co maintain proper vehicle
emission races. This will reduce the likelihood of excess in-use
emissions due to mal-maintenance.
Benefits attributable co many of che maintenance items are
difficult Co quantify. However, one of chese, che cacayst change
interval, exerts whac is perhaps che major emissions influence and
can be estimated. The allowable maintenance regulations will
result in a minimum catalyst change interval of 100,000 miles.
Depending upon the actual useful life co which various manufac-
turers will certify, catalyst lifetimes Longer chan 100,000 miles
may be used in some cases. However, we will use 100,000 miles cor
the analysis. Without che allowable maintenance restrictions,
catalyst change intervals corresponding co current vehicle useful
lives (50,000 miles) are expected. Based upon the position chat
catalyses on in-use. vehicles are not likely co be changed, even
though catalyst changeover would be specified in che maintenance
instruction, then the in-use fleet woud not perform as expected.
Emissions would increase afcer che 50,000 mile point and with a
shorter lifetime higher rates of catalyst failure would occur.
amission rates for this situation would be essentially the
-172-
-------
same as chose developed for a 50,000 mile useful life in Section 2a
above. Lifeciaie Loss of benefits from dropping che allowable
maintenance regulations would chus be che same as "hose developed
for useful Life:
0.13 cons HC and 3.3 cons CO.
Coses would be che same as chose escimatad for useful life
also. &esulC3 are in Table 7II-1.
5. Selective Enforcement Auditing (SEA)
The question co be evaluated with regard co SEA concerns che
acceptable quality level (AQL) co be used in chat program. This
Level identifies the maximum failure race chat can occur in audits
of production engines before chera is a significant probability of
a suspension or revocation of certification. The heavy-duty SEA
programs will use a L0 percent AQL, seaning chac 90 percent of
production engines must comply wich emission standards.
a. 3enefics
The benefits will be estimated by evaluating che change in
emissions which would resulc if che AQL were relaxed from LO
percent co 40 percent. In general, changing che AQL resulcs ia a
change in the mean production level cargec che manufacturer will
aim for. The degree of change can be calculated from statistical
considerations. There are various ways chac chese calculacions can
be approached, all of which give similar resulcs. Sere we will
follow, wich some modifications, che method used by Ford Mocors in
cheir commencs on che N?RM.
The question can be expressed ia general terms as chac of
determining an interval about the mean of a limited sampLe of
engines which will, wich a desired confidence level, contain che
desired percentage of che population:
xtlj (VII-12)
where x is che sample mean (production cargec Level).
" s is che sample standard deviation.
" K is a cabulacad statistical factor depending on sampLe
size, che desired percentage of che population (AQL), and che
desired confidence level.
Here, x must be sec so chac che upper level of che incerval s
^ Ks falls ac che level of che standard divided by che DF.
Values of & can be found in statistical cext book3, for
example, Table 3.2 of "Statistical Design and Analysis of Engineer-
ing Experiments" by Lipson & Shech.
-173-
-------
Ia aur case x represencs the cargec production mean and s the
produceion variability. The desired confidence Level will be chac
presented by Ford: 30 percenc. The desired percencage of the
population (che pass race) was determined by Ford as follows.
da lag che "power curves" for che SEA sampling plana, Ford
determined che pass race associated with che desired probabilic7
of failing che SEA (manufacturer'3 risk). These "power curves" are
presenced in Section 7, Figure 4 of che Ford commencs of 06/29/79
and display probability of failing che audit versus actual propor-
tion of engines noc in compliance. Ford used a 0 percenc manufac-
turers risk, to derive a desired pass race of 65 percenc for a 40
percenc AQL and 95 percent for a 10 percanc AQL. The use of a 0
pereenc manufacturers risk is considered overly conservative, and
the EPA calculations will be based upon a 10 percenc manufaccurers
risk. Thac is, che final resulc will be such chat there will be an
80 percenc confidence chac che manufacturer's risk will be no
greacer than 10 percenc. Independenc calculations of produccion
cargec Levels by che use of the 3candard "t" statistic yield the
same results when an overall confidence level of 90 percent is
used (e.g. 90 percenc confidence chac 90 percenc of che popula-
cion will be belov che cargec level in che case of 10 percenc AQL).
The desired pass races for a L0 percene manufaccurers risk are 57
percenc for a 40 percenc AQL and 33 percenc for a 10 percenc AQL.
rord used a sample size of 3 pre-production engines to eval-
uate emission Levels. Sample size of 3-5 engines seem to be
typical of.most manufacturers and 3 will be used for these calcula-
tions. If a manufaccurer wished to be able co raise his cargec
level vichouc increasing his risk of failure, a larger sample of
engines could be cesced.
Sased upon these faccors (802 confidence level, pass races of
57 percenc (40 percenc AQL) and 33 percenc (10 percenc AQL), sample
size of 3) a value of K can be decermined from staciscical cables.
The resulc is K ¦ 1.1 for a 40 percenc AQL and K ¦ 2.7 for a 10
percenc AQL.
Emission variability was presenced by Ford as being a funccion
of che low-mileage cargec (LtfT ¦ standard/deterioration faccor).
Thac is, a Lover cargec will be associaced wich a Lower variablicy
such chac che racio is cons can c. EPA believes chac it would be
more correcc co use che racio of variabilicy Co actual produccion
level (s/x racher Chan s/LMT) as a cons can c. This approach was
chac generally used by ocher manufaccurers. This change is che
second modification co che Ford approach made by EPA (ehe first
being Co change the "manufacturers risk" from 0 percent to 10
percenc).
The relationship between che produceion cargec level x as a
funccion of che variabilicy (s/x) can now be developed for eicher a
•174-
-------
40 percent AQL or 10 percent AQL.
sad as:
Equation CVII— L2) can be axpres-
j a s » U1I
Dividing through by x,
L +• K (s/x) - LMT/x
or
x/CIT » 1/[L * K(s/x)]
X having been determined for each AQL, this becomes:
x/urr =¦ 1/[1 * 1. l(s/x) ] 40 Z AQL
x/UlT - 1/[1 + 2.7(s/x)] 102 AQL
Figure VII-7 is a graphical presentation of these two aqua-
tions. From thac figure, once the »/x ration is known, the target
level as a fraction of the LMT can be found for either AQL.
The only piece of data now needed to estimate cargec ratio for
the tvo AQLs is the appropriate s/x ratio. Ford was the only
commenter to submit significant data on variability for gasoline-
fueled engines. In its testimony ac the May 14-15, 1979 hearings,
Ford stated that its median value of s/UiT was approximately
0.24.6/ At the hearing, Ford vas questioned on the basis for that
estimate and was asked to submit substantiating data. Ford indi-
cated that this value vas based upon "recent California audit data
of medium heavy-duty trucks.116_/ Ford's written commencs of
06/29/79 contained a table of s/Uil data (Section V, Table I, pg.
9). The highest reported median value was 0.20 for California
audits of passenger cars. The medium-duty truck data cited at the
hearings by Ford turned out to be 0.0S for HC and 0.12 for CO. (to
mention was made in the submission of the rather large drop in
3/LMT values. That the change was intended is verified by the fact
chat the markings of various 3/LMT ratios in Ford's graphical
presentation from the hearing were w'nited out on the 06/29/79
submission. Furthermore, Ford continues to use the 0.24 value in
its illustrations of the effects of the 10 percent AQL, but has
dropped any reference to its representing their median value.
Seducing variability from Ford's original 0.24 value to the
0.08-0.12 range would radically change the target values which Ford
vouid project by its own methodology. Using their un-modified
presentation, such a change would move the target level for a 10
percent AQL to approximately the level originally derived by Ford
for the 40 percent AQL. Thus, Ford would now have to acknowledge
the feasibility of the 10 percent AQL.
There is some possiblity thac as the level of emission de-
-175-
-------
I
Itatio of
I'roJuc i ton
Hiian to I.MT
I'ikiire VI1-7
ProJiictlon Target l.ev/fcls
va Variability
,40% AQl.
V
10% A<}L
I 1 1 1 1
Id .!£ .20 .2S .30
a/x
-------
dines, che s/x racio may increase 3omevhac. Therefore, chis
analysis will use che Ford California passenger car daca (for
vehicles certified co 0.41/9/1.5 g/mile) co esciaace future heavy-
ducy variablicy. The Ford daca is presented as s/LilT ratios rather
chan s/x ratios as desired. ilowever, Ford indicated in ics 06/29/
79 submission, ac pg. 6, chat current vehicles are 3uch chac
x is approximately equal co or slightly les chan LMT. Therefore,
currenc s/LJiT daca will be used co estimate che current s/x value
43 veil.
From Figure VII-4, with s/x ° 0.2, che value of x/LtfT is given
as 0.32 for a 40 percent AQL and 0.65 for a 10 percent AQL.
Emission races for che cvo AQLs can low be calculated. The 10
percenc AQL case has already been done in seccioa lb for che
transient cest procedures. Lifetime emissions were calculated co
be 0.17 cons H.C and 2.4 cons CO. For che 40 percenc AQL case,
using che same methodology as section Lb, emissions can be expres-
sed as:
ac - 0.63 + 0.04(11/1.0,000) + F[2 .4 - 0 . 024(m/ LQ , 000) ] (VII-13)
CO =• 7.5 + 0.52(M/10,0000 * Ff81 - 0.30(M/10,000)] (VII-L4)
Lifetime emissions calculated from equations (VII-13) and
(711-14) are 0.21 cons SC and 2.9 cons CO. The net Loss or bene-
fits for a 40 percenc AQL versus a 10 percent AQL is chen 0.17 -
0.21 ¦ 0.Q4 cons 'dC and 2.4 - 2.9 3 0.5 .cons CO.
b. Coses
EPA believes chac a 10 percenc AQL versus a 40 percent AQL
would mean some increase coat for hardware co meec lower engine
target Levels plus an increase in self audic and qualicy concroL
program coses. SPA believes chat engine cargec Levels are suffi-
ciently close cogecher that Che same cacalyst system would be used
for either. However, che 40 percent AQL could allow a reduction in
che size of che air pump required. Chapter 7 has estimated che
cost of increased air pump capacicy as 526 (before profit and
overhead markup) for the 102 AQL case. For che 402 AQL case, a
savings of 310 will be used ($12.90 wich markup) for che air pump.
Manufacturers designing co che 40 percenc AQL case would probably
also have somewhat more freedom in choosing how far away from che
engine to place the cacalyst, but this would noc have a subscancial
cost impact.
In addition co che $12.90 for hardware, che costs for che 10
percent AQL case include expenses for internal qualicy control
programs and self audits. Quality control program costs wilL be
che same for eicher AQL. However, ac a 102 AQL, manufacturers are
expected co increase self audic races co more accuracely dec ermine'
-177-
-------
che 90 percenc compliance Level of cbeir produceion engines. The
40Z AQL case vill use a self audic race of 0.2 percenc, which' is
typical of Lighc-duey vehicles now operating wich a &0Z AQL- For a
102 AQL, chis will be increased to 0.6 percenc che firsc year, 0.3
percenc che second year, and 0.4 percenc for che chird and subse-
quent years. The change in cose associated with these self-audic
races is 33.77. Tocal differencial cosc for che two cases is chen
S12.90 (hardware) + 3.77 (audic races) ¦ $16.67. This figure is
used in Table 7TI-1 co compute cosc effecciveness.
6. Inspection and Maincenance (I/M)
The analysis which has been done so far has contained no
specific reliance on I/M programs. In che overall rulemaking,
neicher specific benefics nor coses for I/M programs have been
included. However, chere are ways in which I/M would enhance che
effecciveness of che rulemaking and help insure full realizacion
of possible benefics. Therefore, some discussion of I/M in rela-
tion co chis rulemaking is appropriace even chough ic is not
required by che regulations being promulgated.
In che concexc of this rulemaking package, I/M can be viewed
as an "insurance policy" for many of the benefics. The presence of
an I/M progTam, which EPA expects would be implemented in chose
areas requiring maximum benefics, will insure againsc neglect or
abuse of emission related systems by che vehicle owner. The cwo
principal areas when chis might occur are misfueiing with Leaded
fuel or campering wich emission relaced hardware.
EPA has escimaced chac misfueiing in lighc-ducy vehicles
occurs in up co 8 percenc of che vehicles, wich perhaps 6 percenc
of these being persistent misfueLers (leading co catalyst poiso-
ning).^/ There are no corresponding escimaces for heavy-ducy
vehicles since catalyses have yec co be used. However, someching
similar seems possible. The incencive for misfueiing is largely
an economic one, due co che lower cost of leaded fuel compared to
unleaded fuel. In an area having an I/M program, che vehicle owner
would be faced wich a much more powerful economic incencive
againsc misfueiing. This incencive would be che cosc of replacing
Che vehicle cacalysc, which would be nearly 3500, should he fail
the I/M test. Rather than incur ehis expense, EPA believes che
owner would avoid misfueiing his vehicle. Thus, I/M insures
againsc the loss of benefits which mighc resule from misfueiing
wichoue actually generating the coses associated with catalyst
replaceaene.
A similar situaion would occur in relacion co campering.
Currene engine systems are easy co adjust, and could be adjusted
differently for an I/M cesc chan chey are for normal operacion.
For fucure engines chis will noc be the case. Engines complying
with the paramecer adjustment regulations will be difficult co-
-178-
-------
adjusc La such a way as co adversly affect amissions. The poc-
enciaL for coscly repairs from failure of an I/M case (such aa
replacing a damaged carburecor) would sake the occurrence of 3uch
maladju3Cmeac unlikely. Other fans of tampering, such as removal
of the cacalysc or ocher componenes, would also be difficult enough
to be deterred by the need co pass an annual I/M inspeccion.
The above scenario allows a rough estimate to be made of the
benefits an I/M program aighe realize. Assumptions are as follows:
o percenc of the vehicles would be aisfueled initially, without
I/M. Their catalyses would fail co the 1979 engine level used
earlier for failed cacalysc emission races. An additional 4
percenc of the cacalyscs will be esciaaced to have failed by the
end of the average useful life period due to occasional nisfueling.
This is aquivalenc co 2 percenc failed over the whole life in caras
of anissions. Tampering will be accounted for by inducing an
addicional 5 percenc. Catalyse failures would then tocal to 6
percenc + 2 percenc + 5 percenc a 13 percenc. Aitar the fashion of
the calculacions done in section lb above, there will be a 13
percenc shiic In amissions from races for operating cacalyscs to
races for failed cacalyscs. Referring co the aquations inmediacaly
preceeding aquae ion (VII-7), the change in emission can be ax-
pressed as:
HC increase - 0.131C3 +0.02(M/10,000)) - (0.5 + 0.035(11/10,000))1
CO' increase - 0.131(38 + 0.22(8/10,000)) - (5.9 * 0.41(10,000))]
Lifetime emission benefit: of Che t/M program using these
relations is 0.07 cons SC and 2.3 cons CO.
Cose for the I/M program consiscs of a S3 annual inspection
fee. On the belief chat I/M will decer the problems of nisfueling
and tampering, no ocher new coses will be incurred. The fee cosc3
over che vehicle life (8 years), discounted co /ear of 3aia, are
529.34. This is used co compuce the cosc affecciveness found in
Table VII-I.
Ic could be argued chac once an I/M program is puc in place co
decer campering ana oi3fueling, chac some of che benefics derived
from ocher componenc3 of che overall regulation (parameter adjusc-
menc, useful life, allowable naincenance) could be secured by I/M.
However, ic is che incenc of che regulaeory strategy to force the
design of durable emission control 3yscems thae are noc highly
suscepcible co mal-maincenance. Ic is less coscly for the consumer
co pay for chese feacures as part of che aew vehicle engine design
than co have co secure oaincenance or replace parts lacer on. If
che parameter adjustment, full useful life and allowable oaincen-
ance provisions of che regulacions were dropped in favor of reli-
ance of an I/M program to obcain the related benefits, the cost of
field oaincenance and replacement cacalyscs would then have to be'
-179-
-------
charged co ehe I/M program. Considering che catalyse sicuacion
alone makes chis approach much Less efficient than che approach of
retaining ail parts of che regulation package and backing ic up
vich I/M. We have previously estimated che incremental cost of a
full life versus a half life catalyst to be 358. The cost of a
replacement catalyse considering after-oarkac parts markup is aboue
§431.
7. Idle Test
The idle standard applies co CO emissions from gasoline fueled
engines. 3ased upon che idle emission daca now available co EPA,
any emission reduction brought about by che need co certify co an
idle standard would be minimal. However, che need to maincain low
idle emissions in connection vich implemencacion of Section 207(b)
of the 1977 Clean Air Act Amendments, and I/M programs, would
produce an emission reduction for in-use vehicles. This is further
discussed in the Idle Test portion of the Summary and Analysis of
Comments.
Costs associated with implemencacion of che idle test are only
the actual cose of running che addicional certification cesc. Ho
new esse equipment is required. There is also no impact on ocher
coses (e.g. control hardware). Expressed as a cose per engine, che
coses are negligible. Because this is so, a cost effectiveness
computation vould not be meaningful and will noc be attempted.
E. Diesel Engines
1. Overall Rulemaking
To estimate the benefit of the overall rulemaking for diesel
engines, we will first calculate emission races accribucable co che
new standard, and Chen compare chese co che 1979 in-use fleec
emissions. The analysis for diesel engines will focus only on HC,
because CO emissions from diesels are sufficiencly Low as co be
unaffected by che new standards.
In seccion 7 below, it is estimated ehac for a 1QZ AQL, ehe
production target level emission race would be (0.72) x (scandard
- DF). Ihe 0.72 factor accounts for che necessary margins co have
a 90 percent seaciscical confidence of passing an assembly line
audie. The OF in this case is an addieive deterioration factor
over the useful life to which an engine is certified. An addieive
DF is used because historical data is available in this form.
While ehe full lifa for diesel engines is approximately 475,000
miles, ehac lifetime includes one or more engine rebuilds.8/ The
benefits and costs to be estimated for this rulemaking will be for
the entire 475,000-mile life. Engine rebuilding is done for
performance reasons only and is not required or substantially
affected by chese regulations. Compliance wieh chis rulemaking is
-ISO-
-------
not expected to affect diesel engine operating coscj, maintenance
costs, or rebuild coses. As an engine approaches che rebuiLd
poinc, its emissions will cend to increase somewhac. After re-
build, the emission rates are expected to be returned to a level at
or slightly below thac expected from extrapolation of che Linear
DF. The nee amission impact of the rebuild operacion vill there-
fore tend to cancel the excess emissions which occurred prior to
the rebuild. For this analysis Chen, no special steps are taken to
LncLude rebuild emissions and a single OF vill be assumed adequate
on average. The useful Life for certification purposes will be che
estimated mileage to engine rebuild which is approximately 230,000
miles. From 1979 certification daca, the sales weighted OF for EC
is 0.023 g/BHP-hr for 100,000 miles. Lacking a better estimate,
che transient DF race vill be considered to be che same as che
13-mode race. For 250,000 miles chis values becomes approximately
.06 g/BHP-hr.
The production cargec emission level is thus 0.72 x (1.3 -
.06) " 0.39 g/BHP-hr. Combined with the above DF, che average
Lifetime emission race can be calculated by evaluating che value at
half of the full life, or 237,500 miles. This corresponds to .89 +
.055 3 .945 g/BHP-hr. Conversion of chis average rate co cons over
che vehicle lifetime can be done after che manner developed in
section D 1 (a) above cor gasoline-cue led engines.
Total a % brake specific dens icy average useful gm co con
Tons 3HP-hr fuel consumption of fuel a.p.g. Life conversion
Total a % SHP-hr 7.1 lb. fuel gallons
Tons 3HP-hr * .43 Lb. fuel S gai. fuel x 5.3 miles x
475,000 miles tons
Lifetime X 4j4 x 2,000 gm
which becomes
g/BHP-hr x 1.49 ¦ Lifetime Tons.
Using the conversion factor, the lifetime emission for a diesel
vehicle becomes 0.945 x 1.49 ¦ L.41 cons SC.
Lifetime EC emissions for the 1979 in-use fleet which make-up
che base case for overall benefits have been calculated in Chapter
17 as 2.13 tons. The net benefit of the complete rulemaking for
diesel engines is then 2.13 - 1.41 » 0.77 tons H.C.
The cost of the overall rulemaking for diesels has been
developed in Chapter V. Discounted coses per engine are broken
down in Table V-LL. Applying che profit and overhead factor of
1.29 developed in Section 3-1 of Chapter V, these costs become:-.
-181-
-------
Item
Cose Per Engine (discounted)
R 5, D
Certification Testing
Certification Facilities
SEA Facilities
SEA Testing
Self-Auditing
Hardware
42. 74
7.99
49.57
29.27
1.87
20.51
43.06
Total 195.01
Cose effectiveness results are given in Table VTI-2.
2. Transient Test
The implementation of the transient test procedure is a key
element of the program for diesel engines. The need for the
purchase of nev electric dynamometers to replace existing eddy-
current dynamometer results in considerable cost for dia3el
engine manufacturers. A large number of comments were received
during the conment period questioning the real need for Che trans-
ient test procedure for diesel engines. The "test procedure" issue
of the Summary and Analysis of Comments treats all these comments
at length. In that issue, the benefits of converting to the
transient test are also calculated. Those results will be used
here.
Briefly summarized, diesel engine families were examined on a
family by family basis. For each family, the additional reduction
in emissions obtained by implementing the transient test was
calculated in comparison to implementing a 13-mode standard derived
from a 90 percent reduction from the 9-mode gasoline-rueled engine
baseline. These results were then sales weighted for the overall
fleet. In order Co oalce comparison between 13-mode and transient
emission reductions, some means of estimating the transient emis-
sion race associated wich a given 13-mode emission race was needed.
This was done by using the Limited transient test data on diesel
engines to estimace a transient to 13-mode ratio. Based upon 10
available pairs of test data, this ratio is 2.40. The sales
weighted average reduction per engine is 0.49 tons HC.
The emission benefit that could be expected from implementing
the 90 percent reduction on the steady-state (90 percent from the
9-mode gasoline baseline) can be derived from the difference
between the overall rulemaking and the benefits just estimated for
the transient test. This value is 0.77 - .49 "0.28 tons.
Changes in costs related to choice of test procedure include
development costs, hardware costs, and test facility costs (certi-
fication and S£A). All of these would decline if the L3-node
-182-
-------
scandard were adopced instead oc che transient standard. Cose
items givea in seccion I chac would change, are given below:
I cam
Cose victi Transient Test Cost with 13-mode
a & d
Cere, facilities
SEA Facilities
hardware
42.74
49.57
29.27
43.06
11.65
.00
16.72
26.07
Net change in coat =* SL10.20.
The aec change in cost is che cost associated with implemen-
ting the transient test procedure and its related standards. Total
cost for che 13-rnode cest can be computed from che above as $195.01
- 5110.20 =¦ $84.81.
The decrease in R & D and Hardware costs are associated vich
che fact chac che 13-mode scandard would be easier for diesel
engines co meet. The revised values were determined in che same
fashion as che original cost in Chapter V. L979 certification data
was reviewed, and using an SC target level of 0.7 (102 AQL and a
standard of 1.0), engine family specific estimates oi 5 S D and
hardware costs were made.
Facility coses declined because with che 13-mode cest manufac-
turers will hoc need to purchase aev dynamometers or CVS sampling
systems.
3. Redefinition of Useful Life
Considering the current durability procedure as corresponding
co a 100,000-mile useful life, che difference in emissions between
chac and a 250,000-mile useful life is related co che shift of
production cargec levels due co che shorter lifetime. 3ased upon
che OF used earlier, the shift in production cargec level would be
0.0345 g/BHP-hr giving .05 cons net change in emissions over che
cocal engine life. This change in cargec emission levels is small
enough as co preclude che identification of specific associated
cost changes. It is only possible co say chac a small change in
scringency would be expected co produce some small change in costs.
3ased upon engineering judgement, chis cost will be estimated as $2
per engine.
4. Parameter Adjustment
Since at chis time no parameters have been identified co be
adjusted under chese regulations for diesel engines, these will be
no emission benefit for diesels. Likewise, chese will be no
cost.
-133-
-------
5¦ Allowable Maintenance Restrictions
These regulations principally affect curbocharger and fuel
injector service intervals. All but one of the major manufacturers
already specify turbocharger intervals consistent with the regula-
tions. The one excepcion, Mack, does not seen co have any unique
situation making the shorter interval necessary. Stacker, ic seems
zo reflect a more conservative approach to specifying maintenance.
EPA does noc believe Hack would have any difficulty meeting the
required intervals, nor would ic incur any significant cost.
Injector changes to meet the regulations, ic any, would be incor-
ported into che ouch more significant changes in injectors neces-
sitated by the emission standards themselves. Those changes have
been estimated in Chapter 7 as $20 per engine. The impact of che
allowable maintenance regulations will therefore be estimated at
0-35 per engine. By reducing 3omewhac the need for maintenance
on diesel engines, some reduction o£ in-use emissions might be
expected. Assuming a benefit similar to the useful life provi-
sions, a value of 0.05 tons will be used.
6. Selective Enforcement Auditing (SEA)
The question at issue with regard to SEA is the choice of AQL
co be implemented. Evaluation of the 102 AQL will be made in
relation co a 4QZ AQL. The impact oc the AQL on emissions is to
allow a higher production engine targec level in order co pass an
SEA audit with a 4QZ AQL instead of a 102 AQL, all other things
being the same. The amounc of change can be calculated by standard
statistical techniques. The method used here is based upon the
standard "t" statistic and is similar to that used by Caterpillar
to estimate target levels in its comments on the rulemaking.
Mended ace the desired confidence of passing an audit and produc-
tion variability (coefficient of variation). A confidence level of
90 perceac corresponds to that generally used ay manufacturers.
Concerning the coefficient of variation, the only data submitted
was by Caterpillar, and that data indicated chat a value of 0.16
was representative.9/ This value has been used in these calcula-
tions. Later data submitted after a request to Mack indicated chat
for 1979 engines the coefficient of variation was in the range of
.0B co .lfr.10/ Earlier Hack data had supported the 0.16 value.11 /
The formulas and typical calculations for estimating produc-
tion targec levels can be found in section E of Appendix A.
Results depend upon both the coefficient of variatioa and the
number of preproduction engines evaluated. For a range of these
values, the ratio of production target level (x) co the standard
minus the DF (LS£S) vould be as follows:
¦134-
-------
102 AQL 4QZ AQL
Ho. of X/LtfT X/LHT X/LMT X/LHI
engines (eovO. 16) (covQ.20) (cov»Q. 16) (cov0.2Q)
3 0.72 0.63 .35 .32
5 0.76 0.72 .90 .38
7 0.77 0.73 .92 .90
These results indicate che flexibility in target levels. A
higher coefficient of variation could lead to a lower target level,
or a manufacturer could test more engines and retain the same
target level.
3ased upon a coefficient of variation of 0.16 and a sample
size of 3 engines (which seems most characteristic oi current
praccice), targec value ratios of 0.72 (102 AQL) and 0.35 (40% AQL)
vill be used. Production target levels are found by subtracting
the appropriate OF from che standard and then applying the AQL
factor:
102 AQL
0.72 (1.3 - 0.06) - 0.39 g/BHP-hr
402 AQL
0.35 (1.3 - 0.06) - L .05 g/BttP-hr
As shown in section L, the average lifetime emission rate oust be
increased a small amount (0.055 g/BHP-hr) to account tor deteriora-
tion. When this is done, lifetime emission rates for two AQL's can
be calculated to be:
(0.39 + 0.055) 1.49 - 1.41 tons EC 102 AQL
(1.05 + .055) 1.49 - 1.65 tons BC 402 AQL
The net benefit or the 102 AQL versus the 402 AQL is then
1.65 - 1.41 ¦ 0.24 tons.
Cost differences associated with the cvo AQL values include
development cost3, SEA testing costs and 3elf auditing cost.
Item Coac ae 102 AQL Cost at M32 AQL
3. + D 42.74 39.78
SEA Testing 1.37 2.33
Self Auditing 20.51 16.19
Met change in cost *6.32
-185-
-------
The net change in cose is "he case associated with implemen-
ting che 102 rather than che 402 AQL and is used co estimate cost
effectiveness in Table T1I-2. The changes in a & D costs were
estimated on an engine family by engine family basis, as above
for che cost of che transient cesc procedure. Change in production
target levels were not great enough co affect hardware costs. SEA
eesting coses increase slightly for the 40 percent AQL, reflecting
the increased average number of engines expected to be tested to
arrive at a pass/fail decision for a 40 percent AQL versus a 10
percent AQL (15 engines versus 12 engines). Self auditing coses
reflect a change in self audic rates. Self audit costs are
based upon an audic race of 0.2 percent for a 40 percent AQL. For
a 10 percent AQL, self audit races are assumed to be 0.6 percent
che first year, 0.5 percent che second year, and 0.4 percent
thereafter. Quality control program costs (512.90) remain un-
changed .
-186-
-------
References
1/ Discussed ia many statistical "axes. See, for example,
"Statistical Design and Analysis of Engineering Experimeacs,"
Lipson & Sheth, p. 36.
2/ "Average Lifetime Periods cor Light-Duty trucks and Heavy-Outy
Vehicles," EPA. Report SDSB 79-24, G. Passavant, November
1979.
3/ An evaluation of Restorative Maintenance on Exhausc Emissions
of 1975-L976 iiodel Year In-Cse Automobiles, EPA-46Q/3-77-021,
December 1977.
4/ Ibid., labia 17-2.
5/ Ibid., Table 3-13.
6j Page 346 or Searing Transcripc and Page 1 of Ford "Supplemen-
tal Statement on Heavy-Oucy Engine Selective Enforcement
Auditing", May 14, 1979.
7/ Memorandum, "Fuel Switching," Benjamin Jackson, EPA Office of
Enforcement, August 2, 1979.
Average Lifetime Periods for Light-Duty Trucks and Heavv-Duty
Vehicles, EPA Report SDSB 79-20, G. Passavant , November
1979.
9/ See pages 10-13 of the Caterpillar submission of August 15,
1979.
10/ Letter, R.E. Kendall, Mack Trucks, Inc., co Mr. John Anderson,
October 26, 1979.
11/ Lecter, EL. E. Kendall, to Steven Turchen, November 17, 1977
and follow-on submission to Benjamin Jackson, February 14,
1978 and April 18, 1978.
-137-
-------
Appendix A.
REVISED EMISSION FACTORS FOR HEAVY-DUTY
GASEOOS REGULATORY ANALYSIS
A number of changes co the MOBILE-1 factors and the air
qualicy model inpucs are seeded for che final analysis of che
heavy-duty package. Transient daca on old and new heavy-ducy
engines can be used co updace emission faccor eseimaces. For
fucure standards, emission target Levels for manufacturing produc-
tion have been identified, and projected in-use deterioration races
escabLished cor che components. or che final package. Grovch races
cor regional heavy-ducy gasoline (HOG) and heavy-ducy diesel (HDDJ
vehicle miles traveled (VWI) need co be adjusted consiscenc vicil
projected increased dieselizacion races. Lighc-ducy cruck emission
factors appropriace co che proposed LOT package are also needed,
and dieselizacion of LDTs needs evaluation.
A. Ineorporace HP Transienc Daca Into Emission Faccor Equa-
tions
aeavy-Duty Gasoline
The DF's are based upon Light-ducy experience with various
emission conerol systems. The EPA transient casting program for KD
engines has aot provided data ac this cime co revise che DF's, so
they will remain as is.
The 1969 baseline engines were overhauled prior co testing and
will be assumed co be ac cheir new vehicle emission races. While
chis is not scriccly crue, residual decerioration associated with
basic engine wear should be quice small (witness che near-zero DFs
from certification durability engines). The "pre-1970" HDG new
vehicle emission race will therefore be revised to equal che saLes
weighted 1969 baseline values. The 1969 baseline daca represents
3L.5 percenc oc 1969 sales of HOG engines.
1972-1973 baseline daca is available ac chis cine for seven
engines representing 46.42 of 1973 sales. This daca will be used
co updace che 1970-73 factors. There is insufficienc daca co
distinguish 1970-1973 from 1974-1978 and a single faccor will be
used for boch. The 1974 HD standards were not of such stringency
as to produce any significant change in EC or CO emission controL
hardware. During the period of incaresc for our air qualicy
analysis (lace 1980*3 and beyond), vehicles of chat vintage will
have only a minor impact on overall HD emissions.
1979 baseline data is available ac chis cime on 12 engines
representing 362 of 1979 sales projection.. This daca will be used..
-138-
-------
co updace che 1979-1983 emission factors. Noce chat che analysis
will use 1984 as Che fir3C /ear of implementation for Che new
standards anc transient cast procedure.
Since the 1979 baseline data La from certification engines,
some allowance needs co be made cor che difference becveen chose
engines and productLoa line engines (since chere was ao S&A
program) This vill be done by estimating che historic ratio of
cert ificacion Levels co new engine amission raced.
In Figure 4 of che "Summary and .-kaalysis of Comments co che
NPSM: Revised 'ieavy Duty Engine S-egulations for 1979 and Later
HoceL fears", che sales weighted cercificacion Levels cor 1974
heavy-duty engines are given as 4.3 (SO/24 (CO) g/B&P-hr on che
9-mode cesc. the revised transient emission factor for chac /ear
derived from che 1972/1973 baseline program (table La) is 12.7/211
g/miie new vehicle emission race. To compare these cvo races ic is
necessary co ascimate che transient emission levels associated with
che 1974 9-mode certification levels. These are estimated at
9.6/145 g/mile. (The methodology for these ascimates can be found
in Chapcer 711. Ic iavolves escimating an approximate cran-
sienc g/BSP-hr leveL corresponding co che 9-mode steady-state
Level, and converting chac level co g/mile.)
From che above cvo data set3 che ratio of new vehicle emission
rate (g/miLe) co certification Level (g/mile) can be estimated at
12.7/9.6"l.3 for SIC and 211/145»L.45 for CO. These ratios were
applied co che 1979 baseline co derive che new vehicle amission
raca9 for 1979-1983. in Table 1A.
Heavy-Outy Diesel
Ac this time, transient cesc data on diesei engines is ex-
tremely Limited. This data will be used in cvo different manners to
estimate HDD emission factors. The first, which is probably
somewhat more accurate, but is also considerably more complicated,
will be used co estimate SC races. The second will be used for CO.
For HC emission rates, che method is based upon chat approach
developed co evaluate che cost effectiveness of che transient ce3t
procedure cor HC control. Since CO from HDD vill not be affected
by the new procedure, CO was not evaluated. Details of che cal-
culations are found in che "Test Procedure" section of che Summary
and Analysis of Commencs. Available cesc data on diesei engines
from Sw&I and Cummins was used to estimate an approximate ratio of
transient 3C emissions co L3*moda 5C emissions. This racio i3
escimaced at 2.40. Since there were relatively few engines
tested (10 used for this ratio), rather chan use the transient HC
emissions direccly, che racio was applied co 1979 certification
data co estimate a sales weighted 1979 transient emission result..
The certification daca can be found in Table a of che "Test ?ro-
-139-
-------
Table l-A
Revised Exhausc Emission Races
Heavy-Oucy Gas Vehicles
A (g/mile)
3. (g/mile)
Mew Vehicle
Decerioracion Race
Pollucanc
Model fear
Emission Race
(Per 10,000 Miles)
HC
Pre-1970
18.3
0.58
HC
1970-1978
12.7
0.53
HC
1979-1983
6.3
0.53
CO
Pre-1970
228
3.06
CO
1970-1978
211
6.15
CO
1979-1983 ¦
210
6.15
Table i-B
Revised
Exhausc Emission Races
Heavy-Ducy Diesel Vehicles
For All Areas Exceoc California and Hi?h Altitude
A (g/oile)
3 (g/mile)
Mew Vehicle
Decerioracion Race
Pollucanc
Model Year
Emission Race
(Per 10,000 Miles)
HC
Pre-1984
4.0
0.007
CO
Pre-1984
8.7
0.11
-190-
-------
cedure" wrice-up. The sales-weighced 13-mo
-------
Since these results are for new or nearly aew engines, no zero
mileage OF adjustment is necessary. HC and CO emission factors are
included in Table 1-B.
B. Modify Baseline Emission Invencory
Since the 1976 baseline emission invencory for both HDG and
HDD has been computed based upon the existing emission factors,
these values should be adjusted to account tor the change to new
factors. This should be done for each area analyzed by applying a
correction ratio. The correction ratio would be determined by
compucing 1976 composite emission factors using the old and then
the new emission factor equations. The ratio ox the new composite
factors co the old ones vould be used co correct the baseline
emission invencory.
C. Develop Estimates of Future 3D Emission Factors
Heavy-Duty Gasoline
Given the existence of an SEA program, statistical relation-
ships can be established between the standards and the design
values of emissions. Determining factors are the low mileage
target (which is equal to the standard divided by the deterioration
factor), the emission variability, the number of preproduction
engines available for testing (typically 3), and the desired level
of confidence that production engines vould be able to pass an SEA.
The details of the necessary calculations will be found in Chapter
711. F'or HDG, they are based upon a modification of the approach
used by Ford and allow for an 80Z confidence that the "manufac-
turer's risk" during an SEA would be no more than 10Z.
In addition to new vehicle emission rates, estimates of
deterioration rates are also needed. Expected deterioration rates
for the catalyst systems to be used on HDG engines have been
analyzed in the Analysis of Comments for the rulemaking (see the
"allowable maintenance" issue). That analysis indicates that a
catalyst DF of 1.7 over 100,000 miles should be reasonable.
The resulting emission races are calculated in Chapcer VII,
Section Dlb to be:
HC - 0.50 + 0.035 (M/10,000) g/BHP-hr
CO • 5.9 * 0.41 (M/10,000) g/BHP-hr
These can be converted to grams per mile, again based upon
information in Chapcer 7TI, Seccion Dla. The conversion factor is
1.74, and yields the following results:
HC - 0.37 + 0.06 (M/10,000) g/mi
CO » 10.3 ~ 0.72 (M/10,000) g/mi
-192-
-------
3ecause of such aspects of che final rulemaking as che crans-
ienc case, parameter adjustment, full life useful Life, allowable
maintenance and SEA, in—use deterioration races are expected co
correspond closely co cercificacion levels. However, a3 catalyses
approach che end of cheir useful ife, a small number of chem can be
expected co fail. Therefore, average HDG emission races will
increase somewhat near che end of che useful life. Projeccions of
cacalysc failures made in che regulacory analysis indicate that
approximately 15Z of che catalysts would fail. These are dis-
tributed according co a Veibull distribution, and are assumed co
fail co che Level of a veil maintained L979 engine.
Vehicles with failed catalyses will be assumed co emic ac
Levels corresponding co we 1l-maincained 1979 engines. The zero-
mileage emission races will be caken from che EPA 1979 HDG baseline
daca. The QFs will be chose derived in Chapter '/II, Section Dla,
converted co g/mile:
3C (failed catalyse) ¦ 4.9 * 0.035 (M/10,000) g/mi
CO (failed caeaLysc) ¦ 145 + 0.38 (M/10,000) g/mi
If F is che fraction which has failed, then che average
emission race can be developed as follows:
3C (average) » [0.87 * 0.06 {M/10,000)][L-F] * [4.9 + 0.035
(M/10,000)][F]
SC (average) - 0.37 ~ 0.06 (M/10,000) + F [4.0 - 0.025
(M/10,000)]
CO (average) - [10.3 + Q.72 (M/L0,000)][i-F] + [145 + 0.38
(M/10,000)][?]
CO (average) - 10.3 + 0.72 (M/10,000) + F [ 135-0.34 (M/
10,000)]
These final results are given in Table 2.
Heavy-Duty Diesel
Hew vehicle hydrocarbon emission rates for diesels were
estimated following the procedure used by Caterpillar. This mechod
estimates che maximum desired production mean Level based upon the
AQL and variability and Chen. escabLishes a cargee level co assure
attainment of che desired produccion level based upon a small
sample of pre-production engines (e.g. 3). The calculations are
given in Chapter 711 of che Regulatory Analysis. Expressed in
g/mile, che zero mileage emission race is 2.5 g/mi.
SDO carbon monoxide emission races fall below che final
standards and chersfore continue ac che Level of exiscing engines.
-193-
-------
Table 2
Projected Exhausc Emission Races
For All Areas Exceoc California and High Alcicude
AooLicabLe co L984 and Lacer Years
Heavy-Oucy Gasoline
aeavv-Ducv^ieael^
EC « 2.5 + 0.007 (M/L0,000)
CO - 8.7 * 0.11 (M/10,000)
Lighc-Ducy Trucks
HC » 0.39 + 0.016 (M/10,000) * F [1.5 + 0.022 (M/10,000)]
CO » 4.8 <¦ 0.17 (M/10,000) * [IS.6 + 0.30 (M/10,000)]
Where,
HC « 0.87 ~ 0.06 (M/10,000) * F [4.0-0.025 (M/10,000)]
CO » 10.3 * 0.72 (M/10,000 + FH[135-0-34 (M/10,000)]
Where,
'211,726
-194-
-------
The resulting aquae ions are found ici labia 2.
D. Growth Rates for '/TIT3
The rollback model usually assumes a common growth race for
HOC and HDD, which has been based co some extent upon the pollutant
in quescion and Che ambient problem associated vich that pollucant.
That is, GO, vhich is seen as an urban core area "hoc spot" prob-
Lest, is Limited to low growth on che assumption chat urban core
VOTs are near saturation levels. HC , an che other hand, can
increase at a greater rate because of che broader regional impact
of HC amissions. One and cvo percent growth races, respectively,
have been used for. chese polluca'nts.
The analysis for che Final HD Regulations indicates a signifi-
cant shift in diesel utilization rates. For major urban areas,
negative growth in che HDG fleec is expected, while che diesel
fleet vill increase substantially, the analysis indicates that che
usual M08IL£-1 growth races need co be adjusted co account for chis
change. The results deveLoped be Low result in growth of HDG plus
HDD 7MTs at about 22 per year, but apportion chat growth unevenly
between che cwo classes.
Fucure fleec projections used are chose of che Interagency
Scudyl/ for che L980-1990 ciae period. The cotals for HDG and HDD
are summarized below;
Projected HD Fleet Population 2,/
Vehicle Class 3/
HDG III-V 71 VII VIII Total
1973 1,595,000 2,442,000 335,000 437,000 4,309,000
1980 1,778,000 2,474,500 299,000 293,000 4,349,500
1985 2,036,000 2,123,000 217,000 184,000 4,560,000
1990 2,237,000 1,532,000 115,000 65,000 4,099,000
HDD
1973
L980
L985
1990
75,000
332,500
947,000
1,682,000
161,000
224,000
314,000
425,000
713,000
1,067,000
1,243,000
1,247,000
954,000
L,623,500
2,504,000
3,354,000
U "Interagency Study of Post-1980 Goals for Commercial Motor
Vehicles," June, 1976.
2/ Interagency Study, Fig. Ir5, 1-6.
"3/ Class III-V « 10,000-19,500 lb. G7V, Clas3 VI - 19,500-26,000
Tb. GVW, Class VII - 26,000-33,000 GVU, Class VIII « above 33,000
lb. GVW.
-195-
-------
Over the L973-199G period, ehe tocal &DG fleec is projec-
ted co decline ac an average race of -0.9X per year, while che HDD
cleec will increase by *1.1% per /ear an average. The. overall
Cleec is projected Co grow ac an average race of +1.5Z per /ear.
Of more direct interest are chose portions of che fleec
contributing most to urban VHTs (as distinguished from interstate
travel, Cor example). This generally can be done by looking ac che
lighcer weight classes. Classes III-V and VI accumulate nose of
cheir VMTs on-Local or shore hauls.' For chese classes,, che follow-
ing average growth races between 1973 and 1990 are projected (2 pec
year):
HDG
HDD
Total
III-V
+2.0
+2.0
VI
-2.2
+20
+ 1.7
III-VI
-0.2
+20
+ 1.8
The growth in the size of che fleec will not correspond
exactly co changes in VHTs. The Interagency Report also projeceed
VMTs for 1973 and 1990 for local trips, shore haul crips, and long
haul crips.
Annual SD VMT's 4/
(Billions of Miles Per Year)
19 7 3
Local Shore Long Local & Short Tocal
*dDG 35.8 10.7 3.3 46.5 49.8
HDD 10.4 18.3 30.6 28.7 59.3
Tocal 46.2 29.0 33.9 75.2 L09.1
19 9 0
Local
Short
Long
Local & Short
Total
HDG
22.8
5.6
2.0
28.4
30.4
HDD
42.3
38.3
54.3
81.1
L35.4
Total
65.1
44.4
56.3
109.5
165.8
The average annual growth rates represented by these figures
are aa follows:
Local
Short
Long,
-2.9
Local & Short
Tocal
HDG
-2.6
-3.7
-2.9
-2.9
HDD
+8.6
+4.5
+3.4
+6.3
+5.0
Total
+2.0
+2.5
+3.0
+2.2
+2.5
~kj Interagency Report, Fig. 1-7.
¦196-
-------
Comparing fleec growth races with 7MT growth races reveals
significant differences. They reflect a decline in sales of larger
(and higher 3nnual mileage) gasoline crucks and increasing sales
of lighter (and lower annual mileage) diesel trucks. The VMT
projections are che nose appropriate co use since the rollback ana
SXMA models work on VMT growch rates. On che assumption chac che
shrinkage in SDG VMTs may be somewhat overstated, we will use che
following growth projections:
Growth Race of Annual VMTs co Use for
Air Quality Projections
HDG a -2.0Z per year
HDD " +5.0Z per year
£•.. LPT Estimated 1984 Emission Factors
The- LDT amission factors for future vehicles require estimates
of "full life" decerioracion races and new vehicle production
targets.
LDT's are. expected co concinue using oxidation catalyses
similar co 'current systems. Therefore, similar DFs co current
vehicles are ertpenced.. To determine a "full life" DF, we assume
catalysts will be sufficiently durable co lasc che entire useful
life and co maintain deterioration characteristics similar co chose
observed on current 50,000 mile syscema. The full life DF can chen
be determined by linear extrapolation of the 50,000 mile OF. 1979
certification data yields che following LDV and LDT average DFs:
LDV DF(SC) - 1.20 DF(CO) - 1.18 (197 vehicles)
LDT DF(HC) * 1.13 DF(CO) - 1.11 ( 60 vehicles)
In Che interest of simplicity, a rounded off DF of 1.2/50,000
miles will be used for both 3C and CO. Assuming manufacturers
might certify for useful lives of about 100,000 miles chis becomes
1.4/100,000 miles.
Manufacturers base cheir estimates of production line mean
values upon limiced testing of pre-production vehicles (cynically
3). In order co ensure chat an SEA audit will be passed wich some
desired confidence factor (we will use 90Z), cheir carget emission
levels will, of necessity, be some point below che required level
(because of production variability and che small sample size).
This point can be estimated by standard scaeiscical cechniques,
using che "t" statistic. The following relationships will be used:
LXT ¦ low mileage cargec ¦ scandard/DF
a ¦ mtimum desired production mean ¦ LMT -1.28s
x ¦ target new vehicle emission race ¦ m - s (c//"iT)
¦197-
-------
vJhere,
9 ¦ standard deviation of emission Levels-
c ¦ "t" scaciscic for 902 confidence Level and n-L degrees
of freedom
ci ¦ saaple size
To perform che calculations, an escimace of emission vari-
ability is needed. Daca on variability expressed in. the¦ form of
s/LOT waa suomieced by Ford in* their, comments on the >, LDT MPHM.
That daca indicated a value of s/LOT of 0.20 for LDVs certified to
the 0.41/9/1.5 California standards. This value was h'igner than
thac found by Ford.on vehicles certified to higher standards, and
Ford felt thac it would be even higher at lower standards.
For our analysis, ve will assume; chacv.var:iabilicy expressed as
s/x is essentially conscant. A value of s/x can be estimated from
che cord s/LOT daca. Ford indicated in the 3ubmission..'oc its daca
thac currenc engines are such thac 1 is approximately, equal to or
somewhac less chan LOT. There"ore, currenc s/LOT dca can be. used as
an estimate of che currenc s/x racio. For engines built to meec a
102 AQL, we will assume thac s/x remains conscanc (racher than
s/LHT) as x goes down; to examine the effect of- s/x. increasing, we
will calculace results using boch 0.20 and 0.24..
We have::
x • o - s (t//a)
m ¦ LOT - I.28s
s/x ¦ 0.20 or 0.24
Combining chese we gee, depending on che s/x racio used:
x « LOT-. 20x( 1.28+c//TT) or x - LMT-. 24x( 1.28>c//TT)
x(1.256+.20c//T) « LOT or x(1.307+.24c//7) - LOT
x/LOT - 1/(1.256*. 20C//TT) or x/LOT ¦ l/(1.307+.24c//T)
For saaple size a * 2,5,7, che results are as follows:
a t x/LOT (s/x - 0.20) x/LOT (s/x - 0.24)
3 TT3I3 0753 : 04
5 1.533 0.72 0.68
7 1.440 0.73 0.70
-198-
-------
£aaad upon -what seems co be che tnosc common sample size (3)
and Che 's/x * 0.20 ratio, we will use an x/LMT ratio' of 0.63. The
above tabulations indicate chat manufacturers could raise produc-
t-ion—-target- • Levels by increasing sample size or reducing vari-
^abiiicy.---F-or-^xample, if s/x were .24 instead of .20, increasing
sise:. s^nqpiK^ifee- from 3 co 5 would allow maintenance of che 3ame
L^rgei" .ySEEir"**
New vehicle- emission rates for chese ratios are:
JL-aa (Q.r8:/J";4) '-'0.39 RC 0,68 (10/1.4) » i.3 CO
-Jr-DF--of 1.;4/;kP0-,000 miles will yield che following:
^Jgr»- 0.3930'.-0U6"- ¦ (-S^LO-i 000)
r-cd~m 4^s-(~^vI¥:Wio-r0pq).
.'ITEOSse racaf'Wtful^'.'appLy co L3Ts until such cime as a catalyst
rfailMC.'pccur?;ed;. As' catalysts approach che end of their useful
• 1 ife',;"randoB failures; will begin co occur. These failures would be
¦¦expected--to follow-a-Weibull distribution of che form:
• F-=*--l exp~£->{—j-- £
-¦#>
' One "of"" the princrple—crses' ~of "the "Weibul'l ""distribution i3 in
"chara'c'Cerlzrag~frftfcime phenomena,^/ so cha'c~i.f"is well suiced co
-our—purposes-.----To-specify che -distribution; «e will-assume che
xataiyst-changft—point of 100,000 mil.es_corresponds . co a. 5Z failure
rate (giving che manufacturers 952 confidence of catalyst surviv-
"al"T." "'~We 'wilI'"?urther assume a "Weibull slope" of b a 3. Sased
upot^tbieseparameters, % " 269,141 miles. A. plot of chis function
is given in Figure T. ~
For chose'catalysts chat fail, emission rates characteristic
of weIX-maintained, . pr.e-catalyst engines are desired. Based upon
¦review-of-emission-factors for LDT's,_6/ new vehicle emission rates
of- 1—9—HC, 23-.4 -CO-will-be- used and" combined with a DF of 1.1.
Then:
SC (failed catalyst) * 1.9 + 0.038 (M/10,000)
CO (failed catalyst) - 23.4 + 0.47 (M/10,000)
5/ Discussed in many statistical cescs. See, for example,
^Statistical Design and Analysis of Engineering Experiments,"
Lipson and Sheth, p. 36.
6/ "Mobile Source Emission Factors - Final Document," EPA-400/9-
73-005, March, 1978, Table II-l.
-199-
-------
99.
95.
90.
SO.
70.
60.
50.
40.
30.
or
/O
or
/a
%
7o
9 o a
¦« v 9
! i ! iTWiLri.
M 11 i Mill i i ¦ i i 1! ill M:- i i iiiil hH mil,II
--j •; I 1,1 I !• I ( Ll-HlT I |
1 -:i •; 111 i!: i, ( fe-,H -j
. I ! ¦ I I -i l it I till
Figure A-l
Light-Duty Truck Estimated
Catalyse Failure iaces
:• .tff. -
' 1 ' 1 '
U* .LI !¦ '• I' -i
-I-T-
7-.&r
%
%
%
20. %
i 0. %
U1
d
a
<5.0
z 4.0
Oi
U
Si 3.0
2.0
1.0
07L
/0
%
%
%
%
1111 •
Ti-r;. -!
I i I I I ! : ¦ I i t '
l i ; l I ; : : i; i . :
; i i | i i i i; i ¦ ; i
i ; ! .1 h ! i i I j j f
Mill!'!' -in
'HIM
I t
! i !
II!
nil !
! i i M t 1 M
• i 1 i t \ '• 1
' i n 11 ! i
IMS!
iiiiu
!!!ii !¦
! I
l i
j • N 11; ;!ii I
Mil ! -1 II
! S I 11 i! ill I I I i i I iM i
Mill !
mi 1 1 i
! 11 l ! !
! I I
! ! , i ! M
t I I I - i I
j j j I 1 i!;) j !
i !jlilLl
! I
i !
I !
I t
)Mi
! Ml !
i i
i 1 '8 i i ! 1 i : ' ¦' i > A I I t i I ; ill.il
t ;
I M i
I I 1 I ? I >
TT
11
1 :.! I
•4—1.
44
J-l+L
±-
tttt
4-
I ! • I , I : !!!
! ¦ - I I - /1 L ¦> I. I i II ' ! M
i i i /-i M I i i ; i i : ! : i
TT i !
1 I 1 I ! I i i I
i i ! ¦ ! ? I i ! i ! ! 1 i 11 M I
I 1 l I ; 1 j j j f j i 1 i !/| 11| j rj ¦ 1 [ llTF
1111 iMiii ill i/f nill'.!:
i I!;! I I
I I
71
¦in:..! iimim.
|¦' iih;ii » 'M'n
ill ! I j!:
Ml
-I-1 LI ¦"!")'I
11
MTliil!
¦i' i r
ii I 1
i ! rr
I ! F I
i:: liiii. i
I i
I I
TTfr
4+
-H-
H-r-H-
TTT-
tti i
T7TT
TTi
ii ! » !l Hi
MM
I i
u
kl •
¦ : I i
i t I . ! 1 .! i I
I M ; : I Hi : : i i I iT-ir f '¦ \ i MJII r
; j I I j 1 ! I j I' /i | : i I ; i I! r !
iMiii i /I ill i : 'i 1 ill
i i • i
I'll! ;
I ! ;
Li.
\Q,c
I I I ! I I
;i<
i : ¦ I I i i : ! !
1 !
= I :
i Ii
j j s r '
i S ! I ; !MJ
;;i;
tHt
I i M ; III M I i i i l!i
ill;!;
liiii i I I ;' i:"
ill ! I i ! i i
I ; ! ¦ i i ; 1 I I ';!
! I!
I !
I !
I i I
rrrri
rmrr
11 i
l-lu
TT
; !
/ >1-
!
3 -A 5 6 7 3 9 ICO.CCO 2 3 4 567391
7THICLZ MILIAGE
V/glBUU PROBABILITY PAPGS
-200-
-------
Combining che OK catalyst and failed catalyse emission race
according-.Co.che, fraction (?) of failed catalyses, we gee:
v» k
" HC*""~[Q •3'9'"-P"Dfoi6':(i!/ lOyGOO) ][1-F] * [1.9+0,038 (M/10,000) ] [F[
SC » 0.39*+-0:0116 (M/10,000) + F [1.51 + 0.022 (M/10,000)]
CO » "[4.8 -f 0-17 (M/10,000] [1-F] + [2j.4 + 0.47 (M/10,000) I[F]
CO ¦ 4.8 * 0.17 (M/10,000) * F (18.6 + 0.30 (M/10,000]
These raaulca are included in Table 2.
The above*- factors are far gasoline fueled LDTs. In future
years, some dieselizacion of LDTs is expected. Reproduced in Table
3 is data from the Summary and Analysis of Comments foe the Light
Duty Diesel 'Particulates package. This cable gives diesel frac-
tions for future JJ3V sales. These same projections will be used Co
estimate che fraction of LDT diesels.
For 23 certified LDV and LDT diesels in 1979 che average
emissions are;0.60 (HC)/1.75 (CO). The two LDT diesels in this 3et
span che average values. The following values will be used to
estimate dies.el LDT emission:
LDT Diesel Emission Rates
EC * 0.60 g/taile
CO ¦ 2.0 g/taile
These Levels are sufficiently low co be independent of AQL or
emission standard. The average DF for 11 durability vehicles
was low enough co neglect.
F. Develop Future HP Emission Standards for Optional Cases
In che course of evaluating the Clean Air Act mandated 90
percent reduction standard, opcianal reductions of 85 percent and
95 percent will be examined also. The standards corresponding to
these levels Mould be 1.9/23.3 g/BHP-hr HC/CO for 85 percent and
0.64/7.7 :g/3H?-hr SC/CO fo.r 95 percent. Corresponding emission
factors ara derived below:
aeavy-Ducy Gasoline
Fallowing the methodology developed earlier, zero mileage
emission rates for BSC would be:
85 Percent Standard
HC - 0.65 (1.9/1.7) - 0.73 g/3HP-hr
CO « 0.65 (23.3/1.7) - 3.9 g/BHP-hr
¦201-
-------
Table 3
Year-by-fear Projections of the ^0itfsai".£r action,
of Light-Duty Vehicle Salea 7_/_________
Diesel
Model Year
Fraction
L981
¦ 4-. 72
L982
1983
-«8.9S"»
1984
y.s»-
1985
¦11:4*.
1986
13.82
1987
16.52
1988
W.fifc
1989
18.72
1990 !
19.72
1991
202
1992
202-
1993
202
1994
202
1995'
'20-2
77 Source- Summary and Analysis !oc'CSniments,' EPA Light-Duty
Diesel Particulate Final Rulemaking/' Table P»St. v
-202-
-------
"95" PerSeAc Standard
ac » 0.65 (.64/1.7) - 0.24 g/BHP-hr
;GO£»:,0..5^l7r?7U?) - 2.9 g/BHP-hr
Conver'ciiig theie to grams per mile and adding in decerioracion
races corresponding to a 1.7 OF these become:
35 -Percent Standard
HC (OK Catalyst) » 1.3 + 0.09 (M/10,000) g/mile
CO (OK Ca'talyst) *• 15 .5 + 1:08 (M/10,000) g'/mile
95 Percent Standard
HC (OK Catalyse) " 0.42 + 0.03 (M/10,000) g/mile
CO (OK Catalyse) " 5.0 + 0.35 (M/10,000) g/mile
Failed catalyse emission races will be Che same as those
previously used for HSG in Section C. The average emission races
are then:
85 Percent Standard
EC (average) - 1.3 f 0.09 (M/10,000) + F [ 3.6-0-.05 (M/10,000)]
CO (average) » 15.5 * 1.1 (M/10,000) * F [129-o'.72 (M/10,000)]
95 Percenc Standard
3C (average) - 0.42 * 0.03 (M/10,000)¦* r [4.5^0.01 (M/10,000)]
CO (average) - 5.0 + 0.35 (M/10,000.) + F [ 140-0 . 03.(M/10,000)]
These results are included in Table 4.
Beavy-Oucy Diesel
The zero mileage ,amission races can-.be..estimated as vas done
in Chapter Til'of the Regulatory Analysis,: Section E. As before,
only SC is affected since diesel CO,emission .naturally fall below
even the 95 percent reduction standard.
In g/BHP-hr, the zero mileage emission rates are:
35 Percent Standard
0.72 (1,.9-0.06).- • 1.32 ,g/.3HP-hr.
95 Percent' Standard
0.72 (0.64-0.06) - 0.42;. g/3HB«-hr
-203-
-------
Converting Co-grams mile and including Che DF, Che final,
results are:
35 Percent stanaara
'•HC - 3.8 ,'*j99£MX0^f0„
95 Percent Standard
-HC-" 1:2 .007- (:Ji/10,00,0);, g/nilV
!&eaeAquations *re included in ^able,>.'
-204-
-------
Table 4
Optional Projected £^aust : Emission Rates For Ail Areas
Except Caiiro.rtiia and Hi'ghrAlticude
Applicable. ...to 1984 and: Later. - Years
A. 33 Percent Reduction Standard (1.9/23.3 g/BHB-hr-HG/GO) U
geavT-Outy Gaaoliae...
ac - 1.3 + 0.09 (mo,000) + F„ [3.6-0.05 (M/10,000).]
Q
CO - 15.5 + 1.1 (M/10\OOQ) * F„.(129t0.72.3M/10,000)1
a .
Heavy-Qucv ..Diesel
SC - 3.3 f 0.007 (M/10,000)
CO - 3.7 * 0.11 (M/10,000)
B. 95 Percent Reduction, Standard. (0.64/7 .7 %iBHf-hr gC/CO.) \l
geavr-Outy Gaaoliae
gC - 0.42 +'6.0-3 3M/10,000) *¦¦?„ [4.5-0.01 (M/10,000)]
a
CO « 5.0 + 0.35 (M/10,000) + F_ [140 ~ 0.03 (M/10,000)]
a
geavy-Puty Dieael
gC - 1.2 ~ 0.007 (M/10,000)
CO 3.7 * 0.11 (M/10,000)
Note: M and F are as defined in Table 2.
El
1/ Percent re'ductiona given indicate standards .which.are equal
the stated percent reduction -"fr'oa a 1969 .^'aaoliae-iualed baseline
-205-
-------
Tab'l e 4
Opcioaal Projected;Eshau3C;'Sis'i3si~oa.3acea Far -equal to
the seaead percent reduction fr'op a 1969 ;gaso].ine-cualed baseline.
-205-
------- |