Regulatory Impact Analysis
       Control of  Sulfur and
  Aromatics Contents of On-Highway
            Diesel Fuel
              June  1990
U.S. Environmental Protection Agency
     Office of  Air  and Radiation
      Office of Mobile Sources

-------
EPA^20-R-90-103
   June 1990
                         TABLE OF CONTENTS
                                                            Page
  Chapter 1:  Introduction	   1-1
       I.    Background	   1-1
       II.   Diesel Fuel Control Options	  .   1-4
       III.   Structure of Report	   1-6

  Chapter 2:  Refinery Impacts of Controlling Sulfur and
             Aromatics in Diesel Fuel	•.  .  .   2-1
       I.    Trends in Commercial Diesel Fuel Properties  .   2-1
             A.    Current Fuel Quality Standards	2-1
             B.    Commercial Diesel Fuel Survey Results  .   2-2
       II.   Refinery Cost Impacts	2-5
             A.    Background	2-5
             B.    Energy Resource Consultants/Sobotka
                   Study	2-9
             C.    Bonner and Moore's Study	2-12
             D.    National Petroleum Refiners Association
                   Survey	2-19
             E.    Best Estimate Refinery Cost Estimates  .   2-23
       III.   Effect of Hydrodesulfurization on Fuel
                   Aromatics	2-25
       IV.   Small Refinery Impact 	   2-29
       V.    Leadtime Requirements 	   2-31

  Chapter 3:  Engine Control Technology and Cost Prior to
             Fuel Control	3-1
       I.    Engine-Out Particulate Emissions - Current
             Fuel	3-1
             A.    Heavy-Duty Diesel Engines  	   3-1
             B.    Light-Duty Diesel  Emissions	3-27
       II.   Exhaust Aftertreatment Technology 	   3-29
             A.    Exhaust Aftertreatment Types.  	   3-31
             B.    Exhaust Aftertreatment Efficiencies .  .   3-32
             C.    Exhaust Aftertreatment Device Costs .  .   3-41
             D.    Exhaust Aftertreatment Device
                   Deterioration 	   3-48
       III.   Aftertreatment Technology Mix for Compliance
             with Standards	3-49
             A.    Methodology	3-49
             B.    Results	3-50
                   Appendix 3-A: Projected 1991 and 1994
                                 HDDE Particulate Emission
                                 Distribution	3-A

  Chapter 4:  Effect of Fuel Quality on Emissions and
             Engine Cost	4-1
       I.    Effect of Fuel Properties on Diesel Emissions  4-1
             A.    Effect of Fuel Sulfur on Highway
                   Diesel Emissions	4-2
             B.    Effect of Fuel Aromatics on Highway
                   Diesel Emissions	4-6
       II.   Effect of Fuel Quality on Off-Highway Diesel
             and No. 2 Fuel Oil Emissions	4-14

-------
                  TABLE OF CONTENTS  (Cont'd)
     III.  Diesel Aftertreatment Technology With Sulfur
           Control	4-15
           A.    Light-Duty Diesels	4-15
     IV.   Diesel Aftertreatment Technology Fuel
           Aroraatics Control 	  .   4-18
     V.    Sensitivity Analysis	4-21

Chapter 5: Effect of Fuel Modification on Engine Wear.  .   5-1
     I.    Introduction	5-1
     II.   Effect of Fuel Sulfur on Engine Wear	5-1
           A.    Background	5-1
           B.    Experimental Results	5-2
           C.    Used Oil Analyses	5-3
     III.  Effect of Reduced Wear on Operating Costs .  .   5-7
           A.    Oil Change Cost Reduction	5-10
           B.    Extension in Engine Rebuild Interval/
                 Vehicle Life	5-16

Chapter 6: Effect of Fuel Modifications on Air Quality,
           Public Health, and Welfare	6-1
     I.    Emissions Inventories 	   6-1
           A.    Emission Sources	6-2
           B.    General Inventory Estimation Methodology 6-2
           C.    Fuel Consumption	6-3
           D.    Emission Factors	6-5
           E.    Emission Inventory Results	6-15
     II.   Effects of Emission Changes on Pollutant
           Ambient Concentrations	6-22
           A.    Indirect Sulfate Analysis 	   6-23
           B.    Air Quality Impacts	6-30
     III.  Visibility Assessment 	   6-40
           A.    Methodology	6-40
           Appendix 6-A: Diesel Fleet VMT, Fuel	6-A
                         Consumption, Registrations Fuel
                         Economy, and Annual VMT/Veh
                         Calculations
           Appendix 6-B: Nationwide Particulate, HC       6-B
                         and CO Emissions
Chapter 7: Cost Effectiveness of Fuel Controls
     A.    Overview
     B.    Methodology
7-1
7-1
7-3

-------
                           Chapter 1

                         Introduction

I.   Background

     The analyses in this  document  were originally  prepared  in
1987  and  revised in  early 1988.   However,  since then,  due  to
more  recent  technical  information  and  the   July  19,  1988
agreement  by  the  oil  and  engine  industries,  some  of  the
assumptions  about  the  use  of  control  technology  are  now
outdated.    These   assumptions   relate  to   both  when  add-on
technology will be needed and the types of technology likely to
be used.   This most significantly  affects  the  analysis  of  the
cancer  impacts  of  sulfur  control.   The   Agency  no  longer
considers  this analysis  valid.  Therefore,  this  analysis  has
been  deleted  from  the regulatory  impact  analysis.   The other
analyses  in  this  document  are much  less  sensitive  to  the
revised  assumptions.    Economically,  these  newer  assumptions
would  lead  to  slightly  better  cost  effectiveness  than  is
reported here.   This  is  not considered  significant  since,  as
will be shown  later,  sulfur control is already considered to be
very cost effective.   Given the  relative  insensitivity of these
analyses to  the changing  assumptions,  and  the  need to provide
sufficient  leadtime for the 1994 model year, this  document is
being published without revision.

     Diesel   engines    contribute   a   significant   amount   of
particulate   emissions  to  the  nation's   total  particulate
emission  inventory.   Much of  these emissions  occur  in urban
areas where population density,  and thus, personal  exposure to
diesel  particulate  emissions  are  higher  than  in  non-urban
areas.  The particulate emissions  from diesel  engines  are  all
less  than  10  microns   in  diameter,  which means  that when they
are  inhaled,  the particulates  are  deposited deep  in  the lung
tissue.  This  can  cause aggravation of respiratory conditions,
reductions  in  lung   functions,   increased   susceptibility  to
infections,  structural  changes  in  lung tissue and  alveolar
macrophage damage.  Diesel  particulate  has  also  been identified
as   having  probable   carcinogenic   effects.    Finally,  diesel
particulate reduces visibility  (primarily in urban  areas),  and
can  cause soiling of materials.

     Diesel  particulate is made up of three basic components,
or fractions  — the carbonaceous fraction, the  soluble organic
fraction  (SOF), and the sulfate fraction.   Carbon particulates
are  formed  from incomplete combustion  of  hydrocarbon molecules
which make  up diesel  fuel.  The soluble organic  fraction is not
really  a   separate  particulate,  but   consists  of  unburned
hydrocarbons  which  are absorbed onto  the  carbon  particulate
(and thereby add to the weight  of the carbon particulate).  The
sulfate  particulate  is predominately  condensed  sulfuric  acid

-------
                              1-2


gas, which  is derived  from sulfur• in. diesel fuel.   Sulfur in
the  fuel  oxidizes  to  form  S02«  "Some  of  the  SC>2  further
oxidizes to  form 803, which quickly reacts with  water to form
sulfuric acid,   Some  of  the sulfuric  acid can react  with wear
products  (iron,  chrominum),  forming  metal sulfates.   Gaseous
products  emitted by  diesel engines  are  hydrocarbons,  carbon
monoxide (CO), nitrous oxides  (NOx), and (as mentioned earlier)
sulfur dioxide (802).

     Due to  statutory requirements  in the Clean  Air  Act (CAA)
mandating reductions  in  NOx and  particulate emissions, and the
projected growth in  the  diesel vehicle population, in  March of
1985 EPA promulgated  emission standards  for  heavy-duty trucks
and  buses.[1]    These  standards  were  expected  to bring  about
technological  changes in  diesel  engines  and  exhaust treatment
that would lead  to a  reduction in particulate emissions.  These
technologies  are  concentrated  in  two primary   areas —  the
development  of   strategies  to   reduce   engine-out   emissions
(combustion    chamber    modifications,    turbocharging    and
aftercooling,  electronic controls,  etc.),  and the development
of  technology to  "trap"  and   burn  diesel  particulate  in  the
exhaust   system.    These   technologies   affect    the   various
particulate   types   in   different    manners.     For   example,
combustion  chamber  modifications  which  reduce   "dead  space"
result  in a  greater reduction  in  SOF  than carbon  or sulfate
particulate  by  reducing  the quantity of   unburned hydrocarbon
emissions  which  are  available  to  adsorb  onto  the  carbon
particulate.

     Devices used  to reduce particulate  in the  exhaust system
are  called  aftertreatment devices,  and there are three primary
types.   Uncatalyzed  traps are  devices which filter   or  trap
diesel  particulate until the  trap becomes  loaded  to  a certain
point.   Then the trapped particulate is burned off periodically
by   a   regeneration   system,  which  is  needed  because  diesel
exhaust  is  not  usually  hot   enough  by  itself   to   burn  the
particulate.   Next  there are  catalyzed traps,  which  function
quite  similarly to  uncatalyzed  traps.   The  catalyst  simply
reduces  the  ignition temperature  of the  carbon particulate,
thereby reducing  or  eliminating  the need  for  a   separate
regeneration system.   The catalyst can also reduce HC  emissions
and  thus,   SOF.   Last,   there  are  flow-through    oxidation
catalysts,  which  are used  extensively in  light-duty  gasoline
vehicles and  trucks.   Oxidation  catalysts  are   effective  in
reducing gaseous HC  emissions,  which reduces SOF.  Generally,
initial  and  maintenance  costs  are highest  in  the trap  systems,
and  much lower  for catalysts.   Also,  most trap systems  increase
exhaust backpressure,  and  thereby  increase fuel consumption
slightly.   There  is  little  or  no  impact  on  fuel consumption
from oxidation catalysts.

-------
                              1-3
     Although, in-cylinder and aftertreatment  controls  are known
to have  some effect on  carbon  and SOF  emissions,  neither kind
of control  has  been demonstrated  to be  effective at  reducing
sulfate  particulate.   Furthermore,  some  catalyzed  traps  and
oxidation catalysts  are  known to  promote  further  oxidation  of
S02  to  sulfate  particulate.   This   situation  can  lead  to
sulfate  concentrations  being  higher   in   the  aftertreatment
outlet   than   the    inlet,   thereby   reducing   the   overall
effectiveness of  the  aftertreatment device.   So far,  the  only
demonstrated  way  to  remove  sulfate  particulate  is  through
reducing the sulfur  content of diesel fuel.

     At  the  time  the  particulate  emission  standards  were
proposed, several engine manufacturers  raised  the  issue  of the
impact of diesel fuel sulfur  levels on  emissions.   According to
their  comments,   sulfur  from  fuel could  be  a problem  either
through  trap plugging  from  engine-out sulfate  emissions,  or
through  the generation  of  significant measurable  particulate
sulfate  emissions which  would make  it   impossible  to meet the
stringent  emission  standards.   In  the preamble  to the final
rule on particulate standards, EPA's response  was  that  it would
continue to investigate this  issue  and resolve  any  continuing
difficulties.  EPA further stated,  "while it  does  not  appear at
present  that regulating  sulfur content of  diesel  fuel  is  a
prerequisite to the  feasibility  of traps,  if it is shown to be
necessary based  on  ...  further  analysis,  EPA  will investigate
potential action under Section 211(c) of the Act."[l]

     Shortly  after  the  particulate  standards  were  finalized,
EPA  initiated a  preliminary  study  to  examine the  impacts  of
diesel  fuel  controls on  a  number  of  areas. [2]   The  study
explored some of  the  costs and benefits of sulfur and aromatics
controls.    Included   were  estimates  of   potential   emission
reductions,  the  costs of  control  to refiners,  and engine wear
benefits.  Major conclusions of the study were:

     0     Reducing  sulfur in diesel  fuel  would  significantly
           reduce sulfate particulate and S02 emissions.

     0     The  sulfur content of  diesel fuel  could  be  reduced
           from  0.27  wt.  percent  to  0.05  wt.  percent  and
           aromatics  from  about  30 percent  to 21  percent for a
           cost of  about 1.2 cents  per gallon.   However,  this
           would  require  extensive  segregation  of  distillate
           and burner fuels.

     0     Reducing  the  sulfur  content  of  diesel fuel  would
           reduce  corrosive  wear   in  engines,  bringing  about
            lower  cost oil changes  and improved   engine  life,
           leading to  a  savings  of approximately four times the
           above  refinery cost.

-------
                              1-4


     In order  to obtain  comments on  the technical  aspects  of
this  study,  EPA  published  a  Federal  Register  notice  of  the
availability of  the  report and  opened  a  docket  (A-86-03)  to
receive comments.[3]  Many comments were  received from industry
and  other  interested  parties.   These   comments  identified  a
number of  areas  where  the  analysis  could  be  improved.   Also,
due  to the  fact that  this  was a  preliminary  study,  certain
items were not investigated  in the  analysis.   For  example,  the
study did  not  address the effect that fuel controls  could have
on the cost to engine manufacturers  to comply with  the 1991  and
1994  particulate standards.   Some  manufacturers  have  stated
that the availability of  low  sulfur  fuel could  enable  the  use
of  oxidation  catalysts  on  diesel  vehicles,  presumably  at  a
lower cost than trap  systems.   Also,  the study did not  address
the  impacts  fuel controls  could have  on air  quality,  health,
and welfare.

     EPA therefore  initiated  a  second  and more comprehensive
study  of   diesel  fuel  controls,  which  is  contained  in  this
Regulatory Impact Analysis  (RIA).   Like  the  preliminary study,
this  RIA  examines   sulfur   and  aromatics  controls.    However,
additional  refinery  modeling  has  been  conducted  to  address
concerns  raised  in  the  preliminary  study.    A more thorough
analysis of  the effect  of reduced sulfur levels  on engine wear
and  its  effects  on oil  change  cost and engine life has been
performed.   The  analysis  also  includes  an  assessment of  the
impacts  that   fuel  controls  would  have   on  the   types  of
aftertreatment  technology needed  to meet  the  1991  and  1994
standards, and resultant costs.   And finally,  the  effects  of
fuel  controls  on  air  quality,  health   and  welfare   have  been
determined.

II.  Diesel Fuel Control Options

     Two basic diesel fuel  control  options  are explored in this
RIA.  The first one is  sulfur  control.   The current diesel fuel
sulfur content  is about  0.25  wt. percent.   This study examines
sulfur controls down to as low as 0.05 wt.  percent.   The second
control  option is  the  addition of  aromatics  control to  the
above  sulfur control.  Aromatics control alone  is not treated
as  a separate  option,  because  it  is more expensive  and less
cost effective than sulfur control, and  therefore would not be
viewed  as   an  alternative  to  sulfur   control.   The  current
national  average aromatics  level  of highway  diesel  fuel  is
about  35  volume percent.  This  RIA examines  control  levels in
the vicinity of 20 volume percent.

     Section 211(c)  of  the  CAA gives the EPA administrator the
authority  to regulate fuels  used in motor vehicles.   Therefore,
the  diesel fuel controls  discussed in  this  study would apply
only  to  on-highway diesel  fuel.  However,  a significant  amount
of   diesel    fuel    is   consumed    by   off-highway   sources
(agricultural,  construction  equipment,  etc.).   Also,  fuel oil,
which  is  used  as  a  burner  fuel  in  residential  furnaces and

-------
                              1-5
commercial  and  industrial   boilers,   is  very   similar   (and
identical  in  many applications)  to  diesel  fuel.   Therefore,
many  refiners  sell  diesel  fuel  and  fuel oil  from  a  common
middle distillate  pool.    Refiners  would  have  a choice of  how
much fuel to treat,  in the  range of their highway  diesel  fuel
alone up to_their  entire common distillate pool.   The amount of
fuel treated might vary  considerably from refinery  to  refinery
depending on many factors,  which  are  discussed in  Chapter  2.
However,  it  is likely  that,  in many refineries, more than  just
on-highway diesel  fuel will be treated.   The approach used  in
this analysis  is to  include the costs of  treating  this  "extra"
fuel  (above  on-highway  fuel demand), and also to  account  for
any benefits  that  accrue,  such  as  reduced  S02 emissions  from
residential furnaces and off-highway vehicles, etc.

     Sulfur  Control  - Removing  sulfur  from  diesel  fuel  would
require the  refinery industry to  construct a certain amount  of
new sulfur  refining  capacity.  Thip must  be  amortized  over  the
fuel  thus  treated,   thereby  raising  its  price.    Thus,   the
primary  cost  of  sulfur  controls   would be observed  in  the
increase  in  the price  of  middle  distillate products  (diesel
fuel and fuel oil).

     Many of  the major  diesel  engine manufacturers have stated
that if the  sulfur is removed  from diesel fuel,  their  engines
will  be  able  to  meet the  1991  particulate standards  without
aftertreatment,  and  could  also meet  the  1994  standards  with
oxidation catalysts  instead  of with traps.   This  would  result
in  a  lower  cost  per  vehicle  to  meet  the  1991  and  1994
standards,  an  obvious  economic  benefit  of  sulfur  control.
However,   sulfur   controls,  if  promulgated,   could  not   be
implemented  until   around  1993  because  the  refinery  industry
would need leadtime  to construct  new  capacity.   Therefore,  one
action  also  being considered  is  whether to lower  the  sulfur
content of certification fuel  (in  1991)  prior to implementation
of in-use sulfur control.

     Removing  sulfur  from  diesel   fuel  (or  middle distillate
fuels) would have a number  of  emission  benefits.  First,  it
would reduce sulfate particulate from all highway, off-highway
and  stationary  sources  that  use  the  treated  fuel.   Second,
SC>2 emissions  from these  sources  would also be  reduced.   Much
of  the  SO2  reacts in  the  atmosphere  to form  further  sulfate
particulate,  and this  would also  be reduced,  leading  to lower
ambient particulate  levels and improved  visibility.   If  sulfur
controls  lead  to   significant use of  oxidation  catalysts  or
catalyzed  traps  on  diesel  trucks  these   technologies  would
reduce  HC  emissions,  leading  to  lower particulate  SOF  and
reduced cancer incidences.

-------
                              1-6
     Finally,  a  potentially  sizable  benefit  of  diesel  fuel
sulfur control is reduced engine wear.   There is  a  growing body
of  evidence  that  indicates  that  reduced  sulfur  levels  could
lead  to  lower  corrosive  wear   in   diesel  engines,   thereby
possibly  extending  engine  and  vehicle life.   Or,  the  reduced
sulfur levels could enable  lubrication oil  producers  to  produce
oils  at  a lower  cost,  or  allow truck owners and  operators  to
extend oil change intervals, thereby reducing oil  change  costs.

     Sulfur  and Aromatics  Control -  Removing  aromatics  from
diesel fuel would also  require the refinery industry to  install
additional   refinery  capacity.    However,  the  removal   of
aromatics  is  more  difficult  than  removing  sulfur,   so  the
refining costs are higher.  These costs would have  an impact  on
diesel fuel prices.

     Reducing  the  aromatics  content  of   diesel  fuel  reduces
carbon  and  SOF  particulate   emissions   from  diesel   engines
somewhat.  This  has a number  of  implications.  First,  reducing
in-use fuel aromatics levels would  reduce  emissions from trucks
currently on the  road,  thereby improving  ambient  SOF and carbon
concentrations.   The reduction in ambient SOF  could  further
lead to a reduction in cancer incidences.

     The  reduction  in certification fuel  aromatics  levels would
reduce  engine-out emissions  of engines  designed  to meet  the
1994 particulate  standards.   With  lower  engine-out  levels, less
aftertreatment  control  would be needed to meet  the  standards,
thereby  yielding a vehicle  cost  benefit.   In  the  long term,
ambient  levels of  carbon  and  SOF would  be essentially egual
with  or without  aromatics  control since under either  scenario
engines must meet the  same exhaust standard in 1994.  The cost
of  aromatics  control,  therefore,  must be  evaluated against the
potential  technology  savings  in   obtaining  the  net  cost  of
expected  emission reductions from aromatics control.

Ill. Structure of Report

     The  primary analysis  contained  in this report  focuses on
the  cost effectiveness  of diesel  fuel  sulfur  and  sulfur  and
aromatics  control.    (Cost  effectiveness  analysis   relates  the
control  cost  to  the  amount  of  emission  control   achieved  in
terms of  dollars per ton  of pollutant reduced.)

     Chapter  2 discusses the  implications   of  fuel controls on
the  refinery  industry.   First,   fuel composition  trends  are
discussed.   Next,  costs  are  reported  from recent  contractor
studies  and  compared with other studies.    The  impacts  on small
refiners   are   also  discussed.   Last,   the  effects  of  fuel
controls  on other  fuel  parameters  such  as  cetane  number  and
pour point are addressed.

-------
                              1-7
     Chapter 3  discusses diesel  engine  technology,  cost,  and
emissions  without  fuel  controls.    Engine-out  emissions  are
developed for  four  heavy-duty vehicle  classes.   Aftertreatment
technologies are  then  discussed,  with the  focus being  on  the
efficiencies  of  each  technology  on  the  various  particulate
fractions,  and  the  mix of technologies needed to  meet  the 1991
and 1994 standards.   Finally,  the  per vehicle  technology costs
to meet  the  1991  and 1994 standards on current diesel  fuel  are
estimated.

     Chapter 4  presents the effect  of  fuel controls on  diesel
engine  technology,   cost and  emissions.   The effect  of  fuel
controls on  engine-out emissions is  examined  first.    Then  the
information on  aftertreatment  efficiencies and costs  developed
in  the  previous  chapter  is  used  to   develop  per   vehicle
emissions  and  control  technology  costs  with  the  two  fuel
control cases.

     Chapter 5 presents an analysis of  how reducing diesel fuel
sulfur  levels  can reduce engine wear,  thereby extending engine
and vehicle life  or  reducing  lubricant oil costs  and  improving
oil change  intervals.  The first part  of the  chapter  discusses
a  contractor  analysis  of experimental  and  in-use wear  data.
The  second part  of  the chapter  develops potential  operating
cost reductions which could be experienced by truck owners  and
operators with low sulfur fuels.

     The effect  of  fuel modifications on air quality,  public
health  and welfare  are discussed in Chapter 6.   The first part
presents inputs  used in  inventory  modeling of both mobile  and
stationary sources.   The second  part estimates the  effect fuel
controls would have on  urban  and  rural  particulate  and  SO2
concentrations.   This   is  followed  by  a  discussion  of  SO2
conversion  to  sulfate  particulate  in  urban areas,  and  the
resultant effect  on  urban particulate concentrations.   The last
two  sections   estimate  welfare  benefits   such  as   improved
visibility and reduced national SC-2 inventories.

     The  analysis   of   diesel  fuel  control   alternatives  is
presented  in  Chapter 7.   The effect  of sulfur  and   aromatics
control on the costs  for heavy duty trucks  to  meet the 1991 and
1994 particulate  standards  are  examined.   Also examined is the
net  cost  effectiveness of   sulfur  and  aromatics   controls,
estimated in total dollars spent per ton  of pollutant reduced.

     A  complete discussion  of  how the Agency used this study in
selecting  specific   options   for  proposal   is  found  in  the
preamble to the associated Notice of Proposed Rulemaking  itself.

-------
                              1-8
                    References  (Chapter  1)

     1.     "Control of  Air Pollution  From New  Motor  Vehicles
and New Motor Vehicle Engines;  Gaseous Emission and Particulate
Emission Regulations,"  Federal Register,  10606, Friday,  March
15, 1985.

     2.     "Diesel   Fuel   Quality    Effects    on   Emissions,
Durability Performance, and Costs,"  ERG, Sobotka,  EPA Contract
868-01-6543.   Available in Docket ttA-86-03.

     3.     "Diesel   Fuel   Quality    Effects    on   Emissions,
Durability, Performance and Costs;   Availability of Preliminary
Study," Federal Register,  23437, Friday, June 27, 1986.

-------
                            Chapter  2

                 Refinery Impacts  of Controlling
               Sulfur and Aromatic  in Diesel Fuel


     As discussed in Chapter  1, two diesel  fuel properties have
been identified  as  affecting diesel particulate emissions.  The
control of these two  properties,  sulfur  and  aromatics content,
shall be  investigated  in this chapter.   The chapter begins with
a summary of  the results  of two  commercial  fuel  surveys which
track  and identify trends  for various  diesel  fuel  parameters
(Section  I).    Section   II  begins  with  background  on  the
production and distribution of diesel fuel as well as a general
discussion of the refinery methods  which can be used  to reduce
sulfur  and  aromatic  content.  This  section  also  presents  and
evaluates  a  number  of   studies   and  surveys  concerning  the
refinery  cost   estimates   of  controlling  these  diesel  fuel
parameters.   These  studies will be  discussed in detail  in this
section  and  the  results   of  the studies  will  be  used  to
determine  an estimate  of   the  total U.S.   refinery  cost  for
on-highway  diesel fuel  modifications.   Section III  deals with
the  effects  of   hydrotreating on  the types of  aromatic species
in  diesel  fuel.   Section  IV  discusses  the  results  of  an
analysis  addressing the  cost impact of diesel  fuel regulations
on  small  refineries in the U.S.  Section V,  the final section,
presents  an  analysis   of   the   leadtime   requirements   for
implementation of fuel controls.

I.   Trends in Commercial  Diesel Fuel Properties

     In   considering   diesel  fuel   quality  standards,  it  is
important  to identify existing  diesel   fuel  regulations,  and
also  to examine any trends in diesel  fuel  quality which can be
determined  from fuel survey data.   A  synopsis of  the current
diesel  fuel  quality standards and  the available survey data are
presented below.

     A.    Current Fuel Quality Standards

     Several  states  have  established  legal  requirements  for
diesel  fuel.    In  the  northeast,   regulations on the  sulfur
content  of  heating oil  force average distillate  sulfur  levels
down.   California has mandated that  no  diesel fuel sold in the
South Coast Air  Basin shall have  a  sulfur content which  exceeds
0.05  percent by weight.   Recently,  California has promulgated
regulations   limiting  highway   diesel   sulfur  and   aromatics
content  throughout  the  entire state.    These  regulations will
take   effect  in   1993.    In  most  states,   however,   strict
.requirements for diesel  fuel quality  have not been legislated.

-------
                               2-2
     ASTM  document   D-975   lists   recommended   fuel   quality
standards for commercial No. 2 distillate  fuel  oils.   Table 2-1
shows some of the  specifications  for No.  2 diesel fuel and also
for  No.  2  fuel  oil.   It  should  be  noted   that  these  ASTM
specifications  are  recommended  levels  and  are  not  legally
binding  for  refiners.   Table  2-1  also   shows   all   of  the
specifications which  diesel  test  fuels are currently required
to meet  as  part of the Federal Test Procedure (FTP) when new
diesel vehicles  or engines are officially  tested for  exhaust
emissions.

     In general,  the EPA test fuel standards and  the  ranges for
the  various   specifications are more  stringent,  to  limit the
variability in test results due to variations  in  fuel  quality.
For  example,  the ASTM  cetane  specification  for diesel  fuel  is
40 (minimum) while FTP fuels require  42-50.  There is no cetane
specification for  No.  2 fuel oil.   ASTM  recommends that diesel
fuels  and  fuel  oil  do  not  contain greater  than 0.50  weight
percent sulfur,  compared to a range for the FTP fuel of 0.2-0.5
weight  percent.   ASTM has  no  recommendation with  regard  to
aromatic content while an  EPA test  fuel must  have a  minimum of
27 volume percent aromatics.

     B.    Commercial Diesel Fuel Survey Results

     This section will  focus on the trends of  diesel  fuel• with
respect  to  three  fuel parameters; sulfur  content,  aromatic
content, and cetane  number.  Sulfur and  aromatic content are
reported here due  to the  correlation  of these  two  parameters
with  particulate emissions, as discussed  in Chapters  1  and 4.
Cetane  number will also  be shown  since  it  is an indicator of
the  fuel's  ignition quality and tends to vary  inversely with
aromatic content.  Results  of surveys which have  been conducted
by  two organizations  will   be  reported.   One  is  the  National
Institute  for  Petroleum and Energy  Research (NIPER)  survey,
which  is  conducted  annually.   The second  is   a  semi-annual
survey conducted by the Motor Vehicle Manufacturers Association
(MVMA).

     Table 2-2  lists  minimum, average,  and  maximum levels from
the  NIPER  survey as  well   as the MVMA  survey.  As can be  seen,
cetane  number has  been steadily  declining  over  the  past few
decades,  while  sulfur content,  for the  most part,  has been
relatively constant, with only a very slight increase.

     Table  2-2  also  shows  trends  in  aromatic  content.   MVMA
measures  aromatic  content   as  part  of their  survey,  and the
results  indicate that aromatic content may  have   increased only
slightly during  the  1980's.   The  NIPER survey does not measure
aromatics,  although  the  reported  decline  in  cetane  over the
last 25 years   suggests  that  aromatics   levels   may  have been
increasing.[1]

-------
                              2-3

                           Table 2-1


   Specifications  for Highway Diesel No.2  and No.2  Fuel Oil
                             No.2  Diesel
No.2 Fuel Oil
Cetane Number
Distillation Range:
IBP, °F
10% point, °F
50% point, °F
90% point, °F
EPA, °F
Gravity, °API
Total Sulfur, weight. %
(max. )
Aromatics, Min. Vol. \
Flashpoint, °F, Min..
Viscosity, GST. (40°C)
ASTM*
40 (min.)

540-640
0.50 (max.)
-
125
1.9-4.1
EPA**
42-50
340-400
400-460
470-540
550-610
580-660
33-37
0.2-0.5
27
130
2.0-3.2
ASTM*
-

540-640
30 (min)
0.50
-
100
1 . 9-3 . 4
     ASTM D 975.
**   CFR  Title  40,  Part  86.   Applies  only  to  emission  test
     fuels.

-------
                               2-4
                           Table  2-2
                 Trends  in  Diesel  Fuel Quality
                         NIPER  Survey*
Year
             Sulfur Content
               (weight  %)
Cetane Number
1986
1985
1984
1983
1982
1981
1980**
1979**
1978**
1977**
1970***
I960***

Year/
Season
1987/W
1986/S
1986/W
1985/S
1985/W
1984/S
1984/W
1983/S
1983/W
1982/S
1982/W
1981/S
1981/W
1980/S
.000-.260-.960
.001-. 256-. 930
.010-. 267-. 940
.040-. 275-. 960
.020-. 272-. 970
.010-. 283-. 950
.010-. 232-. 980
.000-.230-.890
.010-. 230-1. 100
.004-.220-.900
.22
.23
MVMA
Sulfur Content
(weight %)
.030-.218-.297
.030-. 238-. 352
.047-. 238-. 361
.034-. 237-. 389
.082-. 228-. 341
.140-. 277-. 403
.119-.220-.338
.168-. 344-. 498
.100-. 253-. 349
.151-. 299-. 462
.103-.253-.363
.133-. 276-. 465
.082-.197-.353
.078-.217-.370
36.8-45.1-54.
38.0-44.7-52.
39.0-45.3-53.
37.0-45.0-55.
39.0-44.4-53.
29.0-45.6-52.
39.0-44.9-54.
39.0-46.3-55.
39.0-46.3-61.
40.0-47.3-66.
48.7
50.0
Survey****

Cetane Number
42-44-46
41-44-47
42-44-47
41-45-49
42-45-49
43-45-48
37-44-50
42-45-53
42-44-49
42-46-53
43-47-49
42-47-51
44-46-53
44-47-54
6
1
2
4
0
4
0
9
5
3



















                                                    Aromatics
                                                    (volume %)

                                                  25.9-31.1-38.3
                                                  32.4-39.6-43.1
                                                  26.6-32.3-36.6
                                                  27.9-31.7-38.4
                                                  26.5-31.6-38.8
                                                  27.8-31.7-36.1
                                                  23.8-31.9-37.7
                                                    (not given)
                                                  24.6-30.4-34.5
                                                  26.7-32.4-39.3
                                                  29.4-31.8-35.5
                                                  23.3-29.9-36.3
                                                  23.9-31.3-35.8
                                                  23.2-31.6-38.1
      "Diesel  Fuel Oils," National  Institute for  Petroleum  and
      Energy Research.
      Classification  was  truck-tractor   fuel   in  these  years.
      Classification  changed  to diesel No. 2  in  1980.
      Approximated from  relationship between  clear  cetane number.
 ****  "MVMA National  Diesel Fuel  Survey,"  MVMA,  Inc.
* *
***

-------
                               2-5
     There are  a number  of  reasons  to expect  distillate  fuel
quality  to  deteriorate  in  the  future.    The  processing  of
heavier  and   poorer  quality  (with  regard  to   sulfur  content)
crude oil will likely result  in  higher  distillate sulfur  levels
due to  the direct blending of  more sulfur  into the distillate
fraction.    Likewise   the  increased   conversion   of   heavier
components to  lighter  components should result  in lower  cetane
and higher aromatics levels in  the  future.  As  fluid catalytic
cracking  has increased in  severity,  in  order  to  produce  more
gasoline,  the   aromatic   content  of   light   cycle  oil   has
increased.   Blending of light cycle oil,  generally the  poorest
quality  diesel  blending  component,  into  diesel fuel has  also
been increasing  in recent years  resulting  in increased aromatic
content and  corresponding decreases  in cetane  number.   Without
regulations  it   is  expected  that  a gradual  decline of  diesel
fuel quality will occur.[1,2]

II.  Refinery Cost Impacts

     Several studies have been  conducted recently investigating
the  refining  cost  impact   of   sulfur  and  aromatic  content
restrictions of  highway diesel  fuel.   An  initial  study  of the
issue was  performed  under contract  for EPA by  Energy Resource
Consultants  and  Sobotka  and Company  (ERC  &  SCI),  and  was
released for public comment  in  June,  1986.[2]   SCI responded to
many of  the  comments which were raised in  a  follow-up  study,
while Bonner &  Moore Management Science also prepared  a study
under  contract   for   EPA.[3,4]    In   addition,  the  National
Petroleum  Refiners Association   (NPRA)  surveyed  member  refiners
and  published a report  estimating  the  cost   of  diesel  fuel
control  in 1986. [5]   Other studies have also  been performed on
the cost of diesel fuel control  in limited geographical regions.

     In  this section,   a  discussion  of  the  background  and key
issues   relating  to   the   refining  cost   of   diesel  fuel
modifications will be  presented, as well  as a  synopsis  of the
aforementioned  studies.  Following  the discussion of  each of
the studies, a final analysis will  incorporate  information from
each in deriving  a "best  estimate" of refinery costs.

     A.    Background

     Two  major  issues   exist  which make estimation of the total
refinery  cost of highway  diesel fuel control  difficult.   These
issues   were  handled  differently   in   each   study,   and  an
understanding of  these  two  issues  is  required before discussion
of  the  study results.  The first major issue affecting  highway
diesel  fuel  control  costs concerns  how much  fuel will actually
need  to be  treated  if on-highway  diesel fuel  regulations are
implemented.  The light fuel  oils distilled  during the refinery
process  are  known  as   the  distillate  fuel  oils.  Included are
products  known   as  No.l, No.2,  and No.4  fuel  oils  and No.l,
No.2,  and No.4  diesel  fuels, conforming to ASTM  Specifications

-------
                               2-6


D396  and D975,  respectively.   On-highway diesel  fuel is  one
product within  the No.2 distillate  pool,  though,  as  indicated
in Table  2-1,  ASTM specifications  show No. 2 diesel  and  No.  2
fuel oil to  be  essentially the  same product.   Table  2-3  shows
Department   of   Energy  (DOE)   statistics  on  the   end   user
categories for the No. 2 distillate  pool,  including  both diesel
and  fuel  oil.[6]  On-highway diesel  represents  41  percent  of
the No. 2 distillate  pool,  and 55 percent  of  the No.  2 diesel
pool.   Residential heating  oil  represents 18  percent of  the
total No. 2  distillate pool,  but 69 percent of  the No. 2 fuel
oil pool.

     Comparison  of  the recommended  ASTM  specification for No.2
diesel  fuel   and No.2  heating  oil  (Table 2-1)  reveals  some
differences between the  two  distillate pools.   Diesel fuel must
meet  a minimum  cetane number  of  40  whereas  fuel  oil has  no
cetane  number  specification.   Since  fuels which satisfy  the
ASTM No.  2 diesel  requirements generally satisfy  the  No.2 fuel
oil  specification  as  well, some refiners  choose  to  produce one
product which is sold  to  both markets,  rather  than  incur  the
cost  of segregating  the products  and blend  stocks.   The same
storage  tanks,  loading facilities,  and pipelines are  used for
the common product.   However, if on-highway diesel fuel quality
specifications are mandated,  there would  be a  greater  incentive
to  separate  the products  to  avoid  having  to  reduce the  sulfur
content  of  the  unregulated  No.   2   fuel oil,   as  well  as
distillate   used  for   other  off-highway  applications.   If
segregation  is  not feasible  for some or  many refineries, some
fuel  oil and off-highway  distillate  may also  be  reduced  in
sulfur  or  aromatic content  if such regulations  are  imposed on
on-highway diesel, thus increasing the cost of both products.

      In  its  written   comments  on the  ERG study,  the American
Petroleum  Institute  (API)  stated  that refiners  will  have  to
control  the  quality  of  all  No.2 diesel  products  listed  in the
left   column  of  Table   2-3,  except  for   railroad,   vessel
bunkering,    and   military   use,    which    are     currently
segregated.[7]    However,    Sobotka   and   Co.   believes   that
segregating  on-highway  diesel  fuel  through  the distribution
system would be  possible, but  at some additional cost where the
ability does not  currently  exist.[8]   The ability of  pipelines
to  segregate  distillate  products  would  be  limited to  the
availability  of tankage  at  on-line tank  farms,  with  pipelines
currently  shipping  products  on  a  fungible  basis   apparently
having the most difficulty.   Additional  costs  for  new tankage
and piping would likely  be incurred.  The ability of  distillate
to  be  segregated  at   bulk  terminals would be  dependent on the
mix of tanks  currently used.

      In  summary,  Sobotka   and  Co.  stated  that  most  major
pipelines  would have relatively  little  difficulty in  handling
an  additional grade  of  diesel fuel.   Bulk terminals would,  in
most   cases  be  able   to  handle  another  grade  of  diesel with
relatively  low cost as well.   If  carrying additional  grades  of
distillate  fuel proved  to be too  costly,  bulk terminals  would

-------
                              2-7
                           Table 2-3
          .  End-Use of No.  2 Distillate Products[6]
                            Percent of Total No.2 Distillate
Residential
Commercial
Industrial
Oil Company
Farm
Electric Utility
Railroad
Vessel Bunkering
On-Highway Diesel
Military
Off-Highway Diesel
All Other

Total
No.2 Diesel*
     0
     5
     4
     2
     7
     0
     7
     5
    40
     2
     4
     0

    76
No.2 Fuel Oil
     16
      4
      2
      0
      0
      1
      0
      0
      0
      0
      0
      1

     24
     Includes  small  amounts  of Mo,
     Distillate.
            1  Distillate  and  No

-------
                               2-8


have  the  option  of  marketing  only  one  of  the  distillate
products, depending on the number of  terminals  nearby and their
response to the added grade.

     The segregation options  available  to  the refiner to reduce
sulfur or  aromatic content will  vary greatly depending  on  his
raw  materials,  product  slate  and  the  configuration  of  the
refinery.  Where segregation  of  final products is  possible  the
refinery  may  change  blending practice  to  direct  low  sulfur
blend  stocks  to  the  on-highway  pool.    Directionally,  this
practice may have  the  undesirable side effect of increasing the
sulfur content of  the non-highway  distillate fuels,  although
significant  desulfurization   will  still be required  to  reduce
the on-highway diesel  to  0.05 weight  percent.   The  costs  for
fuel control presented in this chapter were generated under the
assumption  that the.  sulfur  content  of  the  off-highway  pool
would  not  increase.   To  the  extent the sulfur  content  of
off-highway  distillate  does   increase    in   reality,   actual
refining costs  and many of the resultant environmental benefits
will be commensurately lower.

     Since it  is  impossible  to predict at  this time  the degree
to which segregation  would take  place  if  fuel regulations were
promulgated, the  effects of  fuel  segregation on  refining cost
has  been  bracketed for  this analysis.  As  will  be presented
later in this  Chapter, refining costs were developed under' two
segregation  scenarios.   In  the  first scenario,  it was assumed
that no  increase  in fuel segregation would take  place.   In the
second  scenario,   complete segregation of  highway  diesel from
all other  distillates  was assumed.   The  actual  degree  of fuel
segregation  employed and the actual cost   of  fuel  control will
lie somewhere within this range.

     The  second major issue  concerns  the refinery  facilities
that  are currently  available to  reduce sulfur in diesel  fuels.
The refining unit  used to desulfurize  fuel is  referred  to as  a
hydrotreater.   The  hydrotreater  removes  sulfur  by  introducing
hydrogen to the fuel and  passing  it over  a  catalyst at elevated
pressures  and temperatures.   Public  information  on  current
industry hydrotreating capacity  is  difficult  to interpret since
the  related severity  of  the process and how the unit is being
used is not reported.  Also,  little is known  publicly about the
capability of  existing units,  such as the maximum severity an
individual unit can withstand.

     Most   mid-distillate  hydrotreating   units   currently  in
operation  may  technically  be  capable of reducing  sulfur to
level s   as  low  as   0.05  weight   percent,   but  not  without
significant  changes in  their operation.[9]   Running the units
at  much  lower liquid hourly  space velocities  (a  measure of
residence  time)  and  higher  temperatures   will  increase  sulfur
removal,  but  at a reduction  in  fuel throughout.  A few units
may  be  able  to keep  current typical  space  velocities if  they

-------
                               2-9


can  reach  the  required  operating  temperature.   Either  way,
operating costs  will  be higher.  If a  significant  reduction in
the space velocity is required, or if a unit  is  unable  to reach
the desired operating conditions, then  additional  units will be
required, and refiners will have to make  capital investments in
order to meet the nationwide diesel  fuel demand.

     Hydrogen  required  for  hydrotreating is  generated  from the
refinery's•reformer  (a  producer of high  octane gasoline)  or  a
hydrogen  manufacturing  plant.   The  hydrogen  reacts  with  the
sulfur to form hydrogen sulfide (HoS)  which  can be removed in
a  subsequent  process.    The  H2S   is   eventually  converted  to
elemental   sulfur  and   sold.    Thus,   increased  hydrotreating
requires  increases  in  hydrogen generation,  gas treatment,  and
sulfur recovery processing.

     Hydrodearomatization  is  similar  to  hydrodesulfurization
(i.e.,  hydrotreating),  differing  mainly  by the  more  severe
operating    conditions    present    in    the    hydrotreater.
Hydrodearomatization    requires   greater    temperatures   and
pressures,  and   also   requires  the  use  of   a   noble  metal
catalyst.   Large  reductions  in aromatic  content  would force
refiners  to   invest  in  new   units   or   significantly  revamp
existing  units,  since  most  units  today  cannot  operate  at
significantly    increased   pressures.[9]     Other   potential
processing   options   for   aromatics   removal   also   exist.
Hydrocracking  can be used to  mildly reduce  aromatic  levels in
distillate  fuels.   Solvent extraction  can  also  be   used  for
aromatics removal but  may not  be  practical  in many  refining
situations.

     In   addition  to  these   processing  technologies,   other
options  may be   available  to   refiners who  face   reduction of
either  sulfur  or  aromatic' content  of  on-highway  diesel fuel.
One  option  would be  product sharing.   This  practice  currently
exists  in  the  gasoline market where  a refiner in one part of
the  country will produce  and  sell  gasoline  for a  refiner who
produces  an equal  volume of  gasoline for exchange  in  another
part of the country.  It may be possible  for  a similar practice
to exist  for the distillate market.  For example,  if a refinery
has  access   to  low  sulfur  crude  or  has  excess  hydrotreating
facilities, he may be able to  produce  low  sulfur diesel  fuel in
exchange  for  production  of non-regulated  fuel  oil at  another
refinery.   Should   certain  refiners   find  it  impossible  to
finance  additional  processing  equipment,  this  type of approach
may become  attractive.

     B.     Energy Resource Consultants/Sobotka Study

   • Energy Resource  Consultants,  Inc.   (ERG) prepared a report
under contract to EPA entitled,  "Diesel Fuel  Quality Effects on
Emissions,   Durability,   and    Performance,"   dated   September
30,  1985.[2]   Sobotka and Company,  Inc.  (SCI) consultants  were

-------
                              2-10


subcontracted to estimate the refining  impact  support  of sulfur
and aromatic content  reductions.   Sobotka's  approach is similar
to  Bonner  and  Moore's   (discussed  below)  in   that  linear
programming  computer  models  were  developed  to  represent  the
nation's  refinery  capabilities.    The  model  used   for   the
analysis was  a version of  the  Department of  Energy's  Refinery
Evaluation Modeling System (REMS).

     Two models  were  developed,  one to  represent  the Eastern
U.S. (PADDs 1  through 4),  and one to represent the Western U.S.
(PADD 5).  Projections were made  of  refinery supply and demand,
product  qualities,  and processing capacities for  the  modeling
year which was 1987.  No new  investments,  beyond  the facilities
currently in place  or projected to be in place,  were allowed in
the modeling.  After  the  base  cases  were run, two  diesel  fuel
control cases  were  run.   The  first considered sulfur reductions
from  current   levels  of  0.27  weight  percent to  0.05  weight
percent.   The  second  case  controlled  sulfur to   0.05  weight
percent  with  simultaneous  control  of  aromatic  content to  17
volume percent.

     Sobotka used  a diesel fuel  demand of 1.25  million barrels
per day (42.8  percent of  middle distillate  demand),  assuming
complete segregation  of on-highway diesel from the rest of the
distillate  pool.   Total  middle  distillate  demand  was  2.92
million  barrels  per  day.   Crude  oil price  was  assumed  to be
roughly $29 per  barrel.   The  Number 2 heating oil  pool was not
allowed  to serve  as  a sink for the  sulfur removed  from the
diesel  fuel  pool.   The heating oil pool  was  allowed,  however,
to  serve as  a sink  for  aromatics  which  were removed  from the
diesel fuel.

     The  results from  the ERC/SCI  report  are  shown  in  Table
2-4.   The  results  show the nationwide average cost to control
sulfur  content  to  0.05  weight  percent  to  be  1.2 cents  per
gallon.   The  average cost  to   control  sulfur  to  0.05 weight
percent  and  control  aromatic content to   17  volume percent was
1.5 cents per gallon.

     Due to the  assumption of complete segregation  of  products,
the model was  able to alter the refinery's blending practice to
a great  degree to  help achieve the diesel quality  levels.  This
blending  flexibility accounts  for  the reductions  in  aromatics
that  was  noted when  sulfur  was   the  parameter  controlled.
Aromatics  decreased  from  28.7   volume  percent  to  20.3 volume
percent  when  sulfur  was  controlled to  0.05 weight   percent.
This was  a result  of the model  directing higher  aromatic blend
stocks  to  the No.  2 fuel  oil  and not to  the diesel  pool.  In
late'r  work (described below),  Sobotka  and  Co.  agreed  that the
modelling  overstated the  ability  of  refiners  to selectively
blend   distillate   streams,  and  thus   the  degree   to  which
aromatics  would be  reduced when  sulfur   is  controlled would
likely be  less than  that presented  here.

-------
                                 2-11


                              Table  2-4


                  Results  of Refining Analysis From
             the Energy and Resource Consultants Study[2]


                                                       0.05% Sulfur
                               	     0.05% Sulfur   17% Aromatic
Costs
Change in Cost (tf/gal)          	        1.2              1.5
Change in Total Cost(106$/yr)   	         214              277

Fuel Properties
Sulfur (wt.%)                   0.274       0.048            0.048
Aromatics (Vol.%)               28.7        20.3             17.0
API Gravity                     34.0        37.5             37.9
Calculated Cetane Index         46.0        48.2             51.3

-------
                              2-12


     Another important modeling  assumption made by  Sobotka  and
Co.  was  that  existing  refinery  capacity  was  sufficient  to
reduce the sulfur content of the  on-highway  diesel  pool.   Thus,
the  costs  in  Table  2-4  reflect  only  changes  in  refinery
operating and  raw material costs.   This  was partly  the  result
of the  assumptions  that  the industry could segregate the diesel
fuel from the other distillate products and  thus  only treat  the
on-highway diesel  product.   The  analysis  did  not determine  the
minor cost of making the piping changes within  the  refinery  nor
the  potentially  more   significant  cost   of   segregating  the
product distribution system.

     Having recognized the  sensitivity of modelling  results  of
these assumptions,  EPA  asked  Sobotka and Co. to prepare another
study of the issue. [3]  Sulfur control to  0.05  percent with  and
without  aromatics  control to 20  percent  was analyzed  for both
segregated and  combined production  of  highway diesel  fuel  and
other distillates.   Additional  constraints  were  also placed on
the ability to  selectively  blend refinery streams,   in  order  to
better  reflect  actual  refinery  operation.   Another  important
aspect  of  the  study was  that  it  was  performed assuming that
aromatics   reduction  could   only   take   place   via  solvent
extraction  (i.e.,   hydrodearomatization  technology   was  not
available).

     Two   assumptions   concerning   existing   desulfurization
capacity  were used in  Sobotka's  updated  analysis.   The  first
case  assumed  that  the  1992 base  case  desulfurization capacity
was  utilized  at 69  percent.   The  second  case was  analyzed
assuming   no   excess   desulfurization,   hydrogen,   or  sulfur
recovery capacity was available.

     Results  showed that  sulfur  control to  0.05  weight percent
would  cost  between  1.4  and 2.3  cents  per gallon of controlled
fuel  depending on  the  segregation  and  investment  assumptions.
Cost  for controlling both  sulfur and aromatics ranged from 1.9
to  3.5  cents per  gallon of controlled  fuel.   Detailed results
are shown in Table  2*5.

     C.    Bonner and Moore's Study

     EPA  contracted  with  Bonner   and   Moore  to  use  their
proprietary  linear  programming  (LP)  computer  model, designated
the  Refinery  and   Petrochemical  Modeling  System   (RPMS),  to
estimate separately the  refinery costs of  reducing  the  sulfur
and  aromatic  content of  diesel  fuel.[4]   The  cost  of  sulfur
content control to 0.05 weight percent and  sulfur  control with
aromatic content control  to 20  vol. percent  was   investigated
for' each  of   four  geographic  refining   regions.   Region   1
represents  the  East Coast or the Petroleum Administration for
Defense District (PADD)l,  Region 2  represents  the  mid-continent
area,  that  is, PADD 2, 4 and 5  (excluding California), Region  3
represents  the  Gulf Coast,  PADD  3,  and  Region  4  represents

-------
                                        2-13


                                     Table  2-5


              Updated Sobotka Costs for Diesel Fuel Quality Controim


                                   Excess HDS Capacity*    No Excess  HDSCapacity

                        Controlled
                           Fual
                          Volume         Refining    Cost**     Refining     Cost**
	Case	 (1000 bbl/day)  (SMillion/yr)  U/qal) ($Million/yr)   U/gal)

Segregated-Low Sulfur      1071            232         1.4        340         2.1
Segregated-Low Sulfur/
  Aromatics                1071            310         1.9        430         2.6
Combined-Low Sulfur        2351            654         1.8        321         2.3
Combined-Low Sulfur/
  Aromatic*                2351            1110        3.1        1275        3.5
*    Costs  were developed  assuming  existing  HDS  capacity was
     utilized at 69 percent.
**   Cost shown are on a per gallon of controlled fuel basis.

-------
                              2-14
California.  Costs were developed  based on an average crude oil
acquisition cost of $22 per barrel.

     The  national  average   costs  for  sulfur   and  aromatic
reductions  were  estimated  by production  weighting  the  PADD
specific costs obtained from the  RPMS.   The estimated  cost  for
each  region  was  developed by the RPMS using  a  single  "super
refinery" to represent all of  the  refining  capabilities  of  that
region  (i.e.,  the average refinery).   This super  refinery  was
required  to  produce  all  of  the  products  projected  to  be
produced by  all  of the individual refineries in  that region in
the  time  frame   of  the   study,  which  was  1990.   Individual
refineries would  be expected to  experience costs both above and
below that  estimated  by  the RPMS, but  on  average,  the  actual
costs  should be  close  to  that  projected by  the model.   The
complexities  involved  with modeling individual  refineries  made
such  an  approach economically infeasible and make  the  use  of a
single refinery  for  each  PADD a  necessary limitation of  this,
study.

     The   costs   estimated   by   the   RPMS  are,  by   design,
incremental  in nature  and do not  attempt to represent  the  full
costs of  refining diesel  fuel.  This avoids a number of complex
issues  associated with  valuing  capital equipment  already  in
place.   As  the  desired   output  is  the effect  of  sulfur  and
aromatic  control  on  refining  costs,   the  difference  in  -cost
between  an   uncontrolled  and  controlled  scenario  is  fully
satisfactory for  this study.

     A  base  1990 case  was  run   to  determine  optimal  process
requirements  and refinery costs   associated with producing the
1990  product  slate, considering  process capacities known to be
available  in 1986 along  with capacities announced to  be built
and   available  by  1990,   and,   thus,   not  requiring  capital
investment.   The controlled  case  was  run  in   an  analogous
fashion  (i.e., a  fresh optimization  from 1986 capacities),  only
with  a  diesel  product   of  lower sulfur   or  lower  sulfur  and
aromatic content.

      It  should be noted  that  the  RPMS runs  tend  to  project
sizable  capital   investments  between 1986  and  1990 for the base
cases even though the refinery industry as a whole is expected
to  invest little for refining  capacity.   This  occurs because
the   current  capacity   of  many  peripheral  processes  (e.g.,
cooling  towers)   is not   known  and was  presumed  to  be  zero in
1986  for  modeling purposes.  Thus,   the  required  1990  base
capacity  for  these  processes  is  considered  to be  entirely
incremental,  though in all  likelihood, the vast  majority of it
is  currently in  place.   These sizable capital investments have
no  direct effect  on  the  estimated  sulfur or  aromatic content
control  costs,  nor  the  estimated capital investment required
for control,  since the investments are present  in both the base
case  and  the  control case  and do  not affect  the  incremental

-------
                              2-15
control  cost.    It  simply  means  that  the  capital  investment
shown  for 'either the  base or  controlled cases  should not  be
used to  estimate the total capital  investment required by  the
refining industry between 1986  and 1990.  However,  the current
capacities  of  some  of  these peripheral  processes  may  be  in
excess of that needed in 1990.  If the model  is  predicting that
additional  capacity  is  needed  in the  control case  it may  be
overestimating the additional capacity needed  for  control.   The
degree to which  this may be occurring is unknown and not easily
estimated.

     Each regional refining model  was  required to  meet product
demand for  1990  using crude  supplies projected  to be available
in that  same timeframe.  Five  types of distillate  fuels  were
modeled;   naphtha jet   fuel,  kerosene  jet  fuel,   heating  oil,
highway diesel fuel and non-highway  diesel fuels.   Three  types
of residual fuels were modeled;   low-sulfur,  high  sulfur,  and
bunker fuels,  as well  as  three  grades  of  gasoline;  unleaded
premium,  unleaded regular,  and leaded  regular.   The distillate
fuel demand for each region is shown in Table 2-6.

     Bonner  and  Moore  evaluated  two  degrees  of  distillate
segregation  in their  study.  The  first  was  labeled 100 percent
segregation, where all heating oil was segregated  from the rest
of  the diesel  fuel pool.   The  second  scenario  was evaluated
assuming that the current  degree  of  segregation of heating oil
from  diesel  fuel,  as   documented by  the  NPRA  survey,  would
continue.  Bonner and Moore did not, however,  evaluate the cost
of  completely segregating  diesel used  in  on-highway vehicles
from  all other  distillates  (heating  oil  and  diesel used  in
off-highway  applications).   This  will  be discussed  further in
Section E below.

     The base case diesel  fuel  used in  the  model  had a sulfur
content  of  0.25  weight  percent and  an  average aromatic content
of  34 volume  percent.   Bonner  and  Moore  did not   allow  the
quality  of  other products  to  degrade   in order  to  accommodate
the control imposed on  highway  diesel  fuel.   More  specifically,
heating oil  was  not  allowed to be a sink for sulfur which would
no longer be allowed in  the highway diesel fuel.

     The  results from the  Bonner  and Moore modeling are shown
in  Table 2-7.   Bonner  and Moore performed modelling runs for
each  of  the four regions  described  above,  and  aggregated the
results  to  yield  national  cost  estimates.   After  issuing   a
draft  report,  a  follow  up  to the original  study  was  performed
to  account  for  several changes.  These included  reducing the
estimated efficiency  of the  aromatics  removal process  used by
the   model,   adjusting   capital   investments    required  for
incremental  process  capacity,  and  restricting   the internal
segregation capability  of  the  refinery to  a more  realistic
level.  Cases were run  that  investigated the sensitivity of the
model  results to two  parameters:   distillate segregation, and

-------
                              2-16
                           Table 2-6
               Bonnet & Moore Refinery Distillate
           Production Retirements  (1000 bbl/day)[4]
Kerosene
Heating Oil
Total
Kerosene
Heating Oil
Subtotal:  No. 2 pool

Total
100 Percent Segregation*

Fuel
c Fuel

sel Fuel
1 Fuel
tfo. 2 pool
Fuel
t Fuel
sel Fuel
1 Fuel

1
34.0
4.0
37.0

1
90.8
155.6
12.6

NPRA

1
34.0
4.0
37.0
31.81
214.6
12.6
Region
2 3
54.0 61.0
41.0 76.0
253.0 595.0
Region
	 2__ __3 —
136.8 141.6
724.0 975.3
126.2 312.1

Segregation*
Region
2 3
54.0 61.0
41.0 76.0
253.0 595.0
20.5 2.8
840.3 1114.1
126.2 312.1

4
33.0
1.0
222.0

4
4.6
251.0
43.4


4
33.0
1.0
222.0
0.0
255.6
43.4
Totals

 182.0
 122.0
1107.0
Totals

 373.8
2105.9
 494.3

2974.0

4385.0
Totals

 182.0
 122.0
1107.0

   55.1
2424.6
 494.3

2974.0

4385.0
     Bonner  and Moore designations

-------
                              2-17


                           Table  2-7
     Bonnet and Moore Costs of Diesel Fuel Quality Control
     	(g/gal of Controlled Fuel)[4]	
                                                   Added Capital
                                      National     ($ million)
100% Segregation with investment
Sulfur to .05 Wt.%                       2.57         1065.9
Sulfur & Aromatic                        5.20         3023.2

NPRA Segregation with investment

Sulfur to .05 wt.%                       2.69         2023.1
Sulfur & Aromatic                        5.16         4877.7

Volume of fuel controlled (MBPD)

100% Segregation                      2105.9
NPRA Segregation                      2424.6

-------
                              2-18
capital investment.  Bonner  and  Moore's  final adjusted national
costs ace also shown in Table 2-7.

     The assumption that the refiners are  able  to  segregate all
heating  oil  from  other  distillates  and  then  only treat  the
highway fuel is  referred  to  as the 100% segregation  case.   The
costs  under  this  scenario  include only refinery costs  and do
not  include  any  cost  that  may  be  required  to  change  the
distribution-  system  downstream   from  the  refinery   to  further
segregate  fuels.   The   volume   of  fuel  treated   under   the
assumption  is  shown  in  Table 2-6 and  also at the bottom of
Table 2-7.  The other sensitivity case that was  run  with regard
to segregation  is  referred  to  as NPRA  segregation.  This  case
involved  using  NPRA's  survey  results  to  determine  current
segregation  practices.    This  involved  including  some of  the
home heating oil into the on-highway diesel pool and treating a
common  fungible  product.   Since, the  NPRA  segregation  data
reflects  current  segregation  practices,  no  changes  in  the
distribution system would be required.

     It should be  noted that this segregation analysis tried to
bracket  the  cost   by modeling   the  extremes  of  segregation.
Bonner and Moore estimated  total diesel fuel demand to be 2.106
million  barrels  per  day,  a volume, which  included  all  diesel
fuels  required  to meet  a  minimum  cetane  specification,  not
merely  that fuel  which  is  used  in  highway   vehicles.   This
volume  is just  slightly  lower  than  the  volume of  fuel  which
NPRA  projected  would  require  treatment  (2.46   million barrels
per day).  The  actual projected 1990 demand for diesel fuel for
highway vehicles is only  1.29  million barrels  per  day, however
(Table 6-1).  Because Bonner and Moore's estimate of on-highway
diesel fuel demand  included  all  fuel  for applications  requiring
a  minimum cetane specification,  the  control pool size did not
vary   to   a   significant   degree  under   the  two   segregation
extremes.   Therefore,  Bonner  and Moore  did not  find  a strong
cost  sensitivity to segregation  due  to  the assumed  volume of
on-highway diesel fuel.

     The  amount  of investment required  by refiners  to meet the
diesel  quality  specifications   is  also  an   important  factor
affecting   cost.     Base   case  capacity   for   distillate
hydrotreating includes  current industry capacity plus  announced
capacity.   The  maximum throughput for  this  process depends on
the  operating severity,  but  the  publicly  available  data are
published without  defining  the  related severities, thus  there
is uncertainty concerning the  real capacity limits.   Bonner and
Moore   only  ran  cases  under   the  assumption  that  current
operations utilize hydrotreating capacity to  their  limits, and
that   any desulfurization   or   dearomatization beyond current
practice  would  require  new capacity  to  be  built.   Thus, the
refinery  costs  shown  include capital  recovery  costs for new
desulfurization  equipment.    Should  additional desulfurization

-------
                              2-19
equipment  currently  be  available,  costs  would  be  somewhat
lower.   This will be discussed further in Section E.

     Assuming  NPRA  segregation,   and   that   any   additional
desulfucization   beyond   the  base  case   requires   capital
investment,  Bonner  & Moore  estimated  the  nationwide  average
cost of control sulfur content to 0.05 weight percent  to  be 2.7
cents   per  gallon   of   controlled  fuel.   For  the  maximum
segregation-  case,  with  investment  required, Bonner   and  Moore
estimated  the  sulfur  control  cost to be 2..6 cents per gallon of
controlled  fuel.   The  total  cost  to control  aromatics  to  20
volume   percent   and  sulfur   to  0.05  weight   percent,   was
approximately  5.2  cents  per  gallon for  the  both  the  NPRA
segregation  and  maximum segregation  case.  All  of the costs in
Table  2-7  are expressed on  a  cents  per  gallon  of  controlled
fuel basis.

     The added capital  requirements for  treating diesel is also
shown in Table 2-7.  Capital requirements  range  from  l.l  to 2.0
billion dollars  for  sulfur control, and from 3.0 to 4.9 billion
dollars  for sulfur  and aromatics  control,  assuming  no  excess
desulfurization capacity currently exists.

     D.    National Petroleum Refiners Association Survey

     During  the  final quarter  of  1986, the  National Petroleum
Refiners Association (NPRA) conducted a survey of  its  members
in  order  to assess  the  capability of U.S.  refiners to produce
diesel fuels of lower sulfur and  aromatic  content.   As noted on
the transmittal  letter  from the NPRA to its members, the survey
was  conducted  "...to  ensure  the   complete  awareness  of  the
associated costs  and logistical  implications  to the industry
are  made  known  to  the  government."[ 5 ]   The  survey was  also
performed  in order to gather data  which could  be used to check
Bonner  & Moore's  industry model.    Responses  were gathered from
139  refineries  which represent  98  percent  of the  total U.S.
operating  capacity.

     The   NPRA  survey  consisted   of  five  parts.    Part   I
determined current  and  future  practices  under present product
quality  specifications.   Part  II determined  the minimal sulfur
content  achievable with  the use  of  existing  facilities only.
Parts   III  thru   V  investigated  the  costs  associated  with
manufacturing    diesel    fuel    under   various   diesel   fuel
specifications.

     The   NPRA  survey   recognized   the   importance   of  the
segregation of distillate  fuel  issue and asked the refiners to
list  their  projected (1991) product volumes  in the following
categories:   base  diesel  fuel,  common  diesel/distillate No.2,
distillate No.2  (fuel oil),  and other diesel.  Base diesel  fuel
was  defined  as  that fuel  intended  specifically for highway,
off-highway  construction,  farm,  industrial,   and   commercial

-------
                              2-20


uses.   The  common   diesel/distillate  No.2  pool   refers   to
finished stocks  which may be marketed as  either  diesel  fuel  or
fuel  oil.   Other  diesel  fuel  includes  railroad,  marine,  and
military fuel.   The  refiners  projected product volumes for each
of  these categories  for  1991  are  shown  in Table  2-8.   The
average sulfur content is also shown for each product pool.

     Table 2-8  also  shows findings by NPRA concerning how much
additional - segregation  is possible by the refiners.   The NPRA
asked refiners how much  fuel  could be segregated, above current
practices, "...to the limit of  perceived capability to maintain
segregation   through  distribution  and  marketing  systems."
Although  30  percent  of  the  refiners  indicated  they  would
initiate  segregation of  final  products, the  actual  volumes  of
segregated products  changed very  little as indicated  in Table
2-8.

     It  should be  noted  that  ideally the volumes in each column
of Table 2-8  should  be equal  but in  fact  they are  not  due  to
inaccuracies  involved in conducting  a survey of  many parties.
It  is  also  interesting to  note the trends in Table  2-8.  One
would expect  the volumes  of highway diesel and distillate No. 2
to increase as the specifications  become more stringent and the
industry  attempted  to  break  up  the  common  pool  into  two
distinct  pools.   The  table  shows  the  highway  diesel  pool
increases,  but  the  distillate No.  2  pool  actually  decreases.
Also,  in the  base  case  there  is  a  small  amount  of   "other"
diesel  fuel  which  is high in  sulfur content.   In the  control
cases, this volume was included with other products.   The table
shows  441,000  barrels per day  of segregated highway diesel and
2,022,000  barrels  per  day  of   common  diesel/distillate  fuel
would have to be controlled in order to lower sulfur content of
the  on-highway  fuel.   Examination of  the detailed  results  of
the  NPRA survey, however,  revealed  that 59,000  barrels of fuel
in  the  highway  diesel pool was  fuel  which was  designated for
either  railroads,  military, or vessel bunkering use.  In their
comments on the  ERC  report, API stated that fuel designated for
these  uses  was  currently  segregated  from  on-highway   diesel
pools   and   should  not  be   included  in  the  control  pool
volume.[3]  Subtracting  this  fuel from the total yields  a total
control  pool  volume  of 2,404,000  barrels per  day.   This  volume
represents the refining  industry's estimate of their capability
to  separate the distillate products,  as determined by the NPRA
survey.  However,  as shown in Table 6-1, on-highway  diesel fuel
consumption is expected to be only 1.30  million barrels  per day
in  1991.

      NPRA's   estimate  of  the  nationwide  average  increase   in
refining  costs   to   meet   lower    on-highway   diesel  fuel
specifications   are  shown  in,  Table  2-9.    The  total  cost   to
reduce  sulfur to 0.05 weight  percent is  3.11 cents  per  gallon
(on a dollar per  gallon  of controlled fuel basis).  The Bonner
and Moore result  which  corresponds  to  this  cost is  the NPRA
segregation case,  which was 2.69  cents  per gallon.   The  primary

-------
                                   2-21
                                Table 2-8
      Distillate Product Volumes From NPRA Survey  (1000 bbl/dav)[5]
                        Base
Base Diesel
Common Distillate
Distillate No.2
Other Diesel
  Total
                       363(.23)*
                     2,053(.28)
                       240(.29)
                        59(.55)
                     2,715
                                   Sulfur  Control
                                   to  0.05 wt.  %
  441(.05)*
2,022(.05)
  219(.12)
  N.R.***
2,682
                   Sulfur Control
                   to 0.05 wt.  % &
               Aromatics to 20  Vol %
  406 ( .05,19.6)**
2,029 (.04,20.0)
  221 < .12,27.0)
  N.R.
2,656
**
     Sulfur content,  weight percent,  shown in parentheses.
     Sulfur content  (wt.  percent),   aromatics  content  (volume
     percent),  shown in parentheses.
***  Not reported,  volume included with other products.

-------
                                         2-22
                                      Table 2-9
        NPRA Survey Results of Nationwide  Diesel  Fuel  Quality Control CostsTSI
                                                       Sulfur  Control
Additional Manufacturing Costs (Ł/gal)*




   Operating Expense




   Capital Cost




   Total Added Cost




Total Capital Investment (S million)
.15 wt. %
0.48
1.33
1.31
2028
.05 wt .%
0.92
.920
3.11
3313
.05 wt. S
Aromatics Control
to 20 vol. S
1.90
4.45
5.35
6651

     On a per gallon of controlled fuel basis.

-------
                              2-23
difference  between the  two  studies  is  in  the capital  cost.
NPRA projects capital expenditures of $3.3  billion,  versus $1.0
to  $2.0  billion  as  calculated by  Bonner  and Moore.   The NPRA
survey did find that a small amount of distillate  fuel  could be
produced at  0.05  weight percent  sulfur  without  further capital
investment.  If existing facilities were used to  their fullest
potential,   it  was  estimated  that  316,000 barrels  per day  of
0.05 weight  percent  sulfur  diesel fuel  could  be  produced  by
increasing-operating severity only.

     NPRA's  estimated costs  of  controlling  on-highway  diesel
fuels  sulfur  content  to  0.05  weight  percent  and  aromatic
content to  20  volume percent  is  also shown  in  Table 2-9.   The
total cost of control is estimated at 6.35  cents per  gallon,  20
percent higher  than  Bonner  Moore's  estimate of 5.2 cents  per
gallon.  Total  capital  investment was  estimated  to  be  $6.65
billion, versus $3.0  to  $4.9 billion as estimated by Bonner and
Moore.

     Twenty  seven  percent  of the  refiners,  representing  17
percent  of   distillate  production,   stated  they  could  not
implement the low-sulfur  diesel specifications  within  3  years,
due  to  a  lack  of  capital  or  to  environmental  permitting
constraints.   It  should  be  noted  the  NPRA  survey  did  not
address  an  implementation  schedule  longer  than   three  years.
Despite  the  fact  that  these  refiners  could  not  meet  the
specifications, their cost  estimates  are included in all of the
cost data presented above.

     The survey also  instructed the responding  refiners to keep  .
their  product  slate constant  from the base case to the control
case.   This  implies  that  a  refiner  who  has  facilities  to
hydrotreat  75  percent  of  his  current  product  must  make  an
investment  to  treat the remaining 25 percent,  while a refiner
with  excess  hydrotreating  facilities   could not  produce  any
additional  low sulfur  fuel.   The production   volume  lost  by
those  refiners  who cannot produce  low  sulfur diesel fuel could
presumedly  be made  up  by  the  refiners   who   do  have  excess
hydrotreating  facilities,  thus .decreasing  the total  capital
investment required by the industry.

     E.    "Best Estimate" Refinery Cost

     This  section will  outline how  a   "best  estimate" of  the
refinery  cost  impact was  determined  for  sulfur  and  aromatic
control  of diesel  fuel.   While  the  analysis  used  information
which  was  provided by each  of the studies presented above, the
control cost  estimates were  derived  primarily  from  the  Bonner
and  Moore  study.   The Bonner  and Moore study  is  an  objective
analysis  which  was  conducted  after   the  ERC/SCI  study  was
complete,  thus  being  able to address  and incorporate relevant
comments  on the  ERC/SCI study.   It  is worth  noting  that the
updated SCI study's  sulfur  control costs  (Table  2-5)  are very
close  to those  of the Bonner  and  Moore  study,  once  corrections
are made by EPA to adjust capital  cost assumptions.

-------
                              2-24


     The  costs  presented  in  the  NPRA  survey  were  not  used
directly, 'since the  respondents  to  the NPRA  survey were  the
parties which would be  regulated  and the survey  was  conducted
to provide  information specifically  to  regulatory authorities.
Although  the  information  is  valuable,  and  some  of  it  is
incorporated below, the  cost  estimates  can not be used directly
since  it is  difficult  to confirm the  accuracy of  the  data
provided by each respondent.

     The  "best  estimate"  control  costs  were  determined  as
follows.  Per gallon control costs, as  determined  by Bonner  and
Moore, were first  adjusted to reflect a lower after-tax rate of
return and were  then applied to  the  volume of  fuel controlled
in 1991  under two  extreme distillate  segregation scenarios.   In
the  first scenario,  100  percent  segregation,  it  was  assumed
that  only diesel fuel used by  on-highway diesel vehicles would
require  treatment.   In the  second scenario, NPRA segregation,
it  was. assumed  that  current   segregation  practice  would  be
maintained.   The  amount  of  existing  excess  desulfurization
capacity  was  also  accounted for in the  cost'estimates,  as will
be described below.

     An  upper  and  lower  bound  to the  per  gallon  control cost
was  determined   for  both  the  sulfur and subsequent  aromatics
control  scenarios.   The range  in  cost   is dependent largely on
the  segregation  capability of the industry/  as this affects the
volume  of fuel  to  which the control costs  are to  be applied.
If   on-highway  diesel  fuel  specifications   are  mandated,  the
economic  incentive to  segregate  the products  will  be strong.
The  minimum amount of fuel that would need to be  treated  is the
on-highway  diesel  pool  itself,  which is  estimated to  be 1.30
million  barrels per day in 1991.   This  volume  of fuel was used
in estimating  a lower  bound  for  the  refining cost.   The NPRA
control  volume  estimate (2,404,000 barrels per day) was used to
determine the upper bound.

      The  costs  developed  by  Bonner and Moore  were modified as
follows.  First, the Bonner and Moore costs were  computed using
a  capital  recovery   factor  of   0.226  (Table  2-7  in  their
report).   In other  words, supporting one dollar of  investment
would  cost   0.226  dollars  per   year.   When  allowance  for
maintenance,  local taxes,  insurance,  and overhead  is included,
this capital  recovery  factor   becomes  0.286.    These  capital
recovery factors  are  based  on  a 15 percent  after-tax cost of
capital  (in real,  not  nominal,   dollars).   However,  a  recent
report  prepared by Jack Faucett Associates under contract with
EPA  indicates  that  a  10  percent  after-tax  cost  of capital is
more appropriate  for  the refining  industry.   [10]  When  this
figure  is used  the capital  recovery .factor,  as  calculated by
Bonner  and Moore's methodology,  becomes 0.171  (0.231  including
maintenance, local taxes,  etc.).

-------
                              2-25


     Using  these  adjustments,   sulfur  control  costs  for  the
maximum  segregation  case  were  determined  to  be  0.357  dollars
per barrel  of  controlled  fuel  (2.04  cents  per gallon)  assuming
HDS investment is  required,  and 0.482 dollars per  barrel  (1.15
cents per gallon) for the  "no investment" case.

     An  additional adjustment  was  made  to Bonner  and  Moore's
100 percent segregation  case results.   In the Bonner and  Moore
report,  control  costs  for  desulfurization were  determined  by
comparing   investment  and  operating  costs  for   two   cases:
maximum  segregation with  0.25 percent sulfur fuel (baseline)and
maximum segregation with  0.05 percent  sulfur fuel.   However,  to
accurately  estimate  the  cost of  processing changes as  actually
perceived by  the  industry,  investment  and operating costs  for
the  "maximum  segregation with  0.05 percent sulfur" case should
be compared with a baseline case representing current  industry
practices  (i.e.,  the  NPRA  segregation  case).   Differential
processing  investment  and  operating  costs were  recalculated
accordingly.  Bonner  and  Moore's  fuel control costs for the 100
percent  segregation  and  NPRA segregation cases were  adjusted,
and are shown in Table 2-10.

     These  costs were generated  under  the assumption  that  no
excess  desulfurization  equipment was available.   However,  the
NPRA  survey found that  if existing facilities  were  used  to
their fullest potential,  316,000  barrels  per day of 0.05 weight
percent  sulfur  diesel  fuel could  be produced.   Thus,  316,000
barrels per day of fuel could be  treated  at a lower cost.   This
cost  was  determined  by  subtracting  the  per  gallon  capital
recovery  cost  associated  with  new  desulfurization  equipment
from  the per gallon "HDS  Investment Required" costs  shown  in
Table 2-10.  These "No  HDS Investment Required" costs  are also
shown in Table 2-10.

     As  stated   previously,  the  volume of  fuel which  will  be
controlled  ranges  from 1.30 to  2.404  million barrels per day in
1991.   A total  of  316,000 barrels per  day can be controlled at
the  "no  HDS  investment"   cost  shown  in  Table  2-10.    The
remainder of the pool will be  controlled at the "HDS Investment
Required Cost."  The  average national control costs for control
of   sulfur  and  subsequent  control   of   aromatics  were  thus
calculated  and  are shown in Table 2-11.   Sulfur  control  costs
range from  1.8  to  2.3 cents per  gallon of  controlled fuel ($360
to   $830 million   per   year),    while   subsequent  control  of
aromatics costs  from 2.1  to 2.4  cents per gallon of controlled
fuel ($470  to $770 million per  year).

III. Effect of Hydrodesulfurization on  Fuel Aromatics

     The refinery  processing required to remove the sulfur from
diesel   fuels  involves  introducing  hydrogen to  the  fuel   and
passing it  at  elevated  temperature  and  pressures  over   a
catalyst.   This process  is  similar   to  that used to   saturate

-------
                              2-26

                           Table 2-10

     Adjusted Bonnet and Moore Costs of Diesel Fuel Control
     	(Cents per Gallon of Controlled Fuel)	
                                                 Cost (^/gallon)
100% Seareg-ation (v/HDS investment repaired)
Sulfur to .05 wt.%                                     2.06
0.05% Sulfur & Aromatic to 20 Vol %                    4.34
NPRA Segregation (w/HDS investments regruired)
Sulfur to .05 wt.%                                     2.39
0.05% Sulfur & Aromatic to 20 Vol %                    4.44
100% Segregation (no HDS investment required
Sulfur to .05 wt.%                                     1.04
0.05% Sulfur & Aromatic to 20 Vol %                    3.58
NPRA Seo;reqation( no HDS investment recruired)
Sulfur to .05 wt.%                                     1.32
0.05% Sulfur & Aromatic to 20 Vol %                    3.69

-------
                                           2-27
                                        Table  2-11
          "Best Estimate" Refinery Cost Cor Controlling On-Hiqhwav Diesel Fuel
                          Sulfur Control to 0.05 Weight Percent
   Lower Bound (100 Percent Segregation)
   Volume      Control Cost    Control Cost
(1000 bbl/dav)   (cVgal)
    316
    982
   1298
1.04
2.06
1.81
                	Upper Bound (MPRA  Segregation)	
                   Volume     Control Cost   Control Cost
(8 million/yr)  (1000 bbl/day)    (*/gal)      (S million/yr)
      50.4          316           1.32          63.9
     310.1         2088           2.39         765.0
     360.5         2404           2.25         329.0
                       Sulfur Control to 0.05 Weight Percent With
                         Aromatic Control  to  20 Volume Percent
                Lover Bound
                                            Upper Bound
    Volume     Control Cost
(1000 bbl/dav)   (*/gal)
    316
    982
   1298
3.58
4.34
4.16
 Control Cost       Volume    Control  Cost
(t Million/yr)  (1000 bbl/day)
                    316
                   2088
                   2404
                                 3.69
                                 4.44
                                 4.34
Control-'Cost
($ Million/yr)
    178.8
  1/421.2
  1,600.0
Incremental Aromatics
 Control:         2.35
                466.2
                                 2.09
                                                 771.0

-------
                              2-28


aromatic compounds in  the  fuel,  although larger-scale aromatics
saturation  requires   more  severe   hydrotreating   conditions.
However,  although  desulfurization  units  serve  primarily  to
remove  sulfur  from the  fuel,  some  mild  ring saturation  does
take place,  and consequently, the amount  of  aromatic carbon is
reduced.

     Typically   desulfurization   does  not   result   in   the
saturation- of  monocyclic   aromatic   structures.   Dicyclic  and
polycyclic aromatic compounds  present in  diesel  fuel,  however,
are often  partially  saturated.  This  is because  the second and
third aromatic  rings  are less  stable than the first  ring,  and
are therefore saturated more easily during hydrodesulfurization.
The result of  this  partial  saturation  is that  the  amount  of
aromatic carbon in the fuel  is  reduced,  while aromatics levels
measured  by  Fluorescent  Indicator   Adsorption   (FIA)  analysis
(which  counts  mono-,  di-,  and  tricyclic  compounds  equally)
remain  the same.   Fuel  analysis for  aromatic content  by  mass
spectroscopy,  however,  shows  the   shift  from  polycyclic  to
monocyclic structures.

     Hydrodesulfurizing experiments were  carried  out on various
fuels by Jack  R.  Yoes and Mehmet Y. Asim of Akzo Chemie America
with  the  purpose  of  determining  hydrotreater performance and
aromatic  saturation.[11]   Two  cracked feeds   and a  high sulfur
virgin stock were hydrotreated  at  hydrogen partial  pressures of
400, 800,  and  1200 psig, at temperatures designed to produce of
0.05  wt.   percent  sulfur   product.    Both   cobalt-molybdenum
(KF-742-1.3Q)   and  nickel  molybdenum (KF-843-1.3Q)  catalysts
were  evaluated to determine  differences  in sulfur  removal and
aromatics  saturation.   Although  conditions  were severe enough
to produce reductions  in aromatics as measured by  FIA  as well
as  sulfur  content, what is interesting to note  is  the shift in
aromatic distribution  upon  hydrotreating.   For the  three fuels,
using the  nickel-molybdenum catalyst, the fraction of aromatics
falling  in the monocyclic and  dicyclic  categories  shifted on
average from  about 70 and 18,  respectively, to about 77 and 13,    i
respectively upon hydrotreating.   In  other words, approximately    I
30  to  40  percent of  the dicyclic aromatics  were partially
saturated.                                                          i

     Another   evidence  of  this   shift  took   place   in  the
hydrotreating   of   fuels   for   the   CRC VE-l   project.[12]
Hydrotreating  was used to  produce   a  low  sulfur   fuel   (0.06
weight  percent) from  a  high  sulfur  feed  (0.32 weight percent).
Although    FIA   aromatics    levels    remained   constant   at
approximately  40 volume  percent, the distribution  of  aromatic
species changed drastically.   The mass  percent  of  mono-,   di-,
and'  tricyclic  aromatic   carbon  was  8.4,   14.2,   and   1.6,
respectively,   in   the   feed,   and   13.7,    7.9,    and    0.8,
respectively,  in the  hydrotreated product.   This  represents  a
reduction  of   roughly  50  percent in dicyclic  and tricyclic
compounds.

-------
                              2-29


     The  increase  in  monocyclic   structures   indicates  that
nearly all  those di- and  tricyclic  molecules  became monocyclic
aromatic  structures.    It  also  appears   that   little  or  no
monocyclic  aromatic   saturation  took  place.    Treatment  of
another fuel  used in the  CRC test  program  to  reduce  both FIA
aromatics   (to   10  percent)   and  sulfur   levels  resulted  in
eliminating nearly all dicyclic  (92  percent) and tricyclic (100
percent) structures.

     Another  study of  the  effect   of  hydrotreating  on  diesel
fuel  aromatics  was funded by the National Research  Council  of
Canada.[13]    A   coal-derived   middle  distillate   fuel   was
subjected  to  three  levels  of   severity of  hydrogenation.   A
reduction in  total aromatics  of approximately  30 volume percent
showed  a  corresponding  decrease  of  70  and  85  percent  in
dicyclic  and  tricyclic   aromatic   structures.    More  severe
treatment showed  reductions  in excess of  80 and  95  percent  of
dicyclic and tricyclic structures, respectively.

     What  the  first  two  examples   suggest  is  that  even when
aromatics    levels    measured   by    FIA   stay    constant,
hydrodesulfurization  may  result  in  a  reduction  in aromatic
carbon, due to a  shift  the distribution of aromatic compounds
greatly  in the direction  of  monocyclic  compounds.   The more
severe  hydrodearomatization  process may result in  the partial
saturation  of nearly  all  di-  and  tricyclic  aromatic species.
As will be  discussed in Chapter 4,  emissions  may be dependant
on   the  type  of  aromatics  present.    Further  experimental
quantification of the shift in  aromatic  structure of fuels with
hydrodesulfurization may be  warranted as emission models become
more sophisticated and are able to utilize this  information.

      In this  report, it will  be estimated  that:   1) a reduction
in fuel  sulfur  levels to 0.05  weight percent will result  in the
partial  saturation of approximately  50  percent of  the di- and
tricyclic  aromatic species in  the  fuel  to monocyclic aromatics
and  2)  processing to  reduce fuel sulfur to  0.05 weight percent
and  aromatics to  20  volume  percent will  result in the partial
saturation  of  80 and  90  percent  of  the di-  and  tricyclic
aromatics respectively.

IV.  Small  Refinery Impact

     Of  particular  interest  in this  study is the  impact  on
small  refiners  of a  regulation  of diesel  fuel  quality.  To
investigate this, EPA commissioned  Sobotka and Company Inc. to
study  the issue.[14]

      In   Sobotka's  report,   small  refiners  were  categorized
according   to   the   type  and  configuration  of   processing
technology  employed,  and  then  simulated by aggregate refining
models.   A comparison  of  costs  for   fuel  control  for  small
versus large  refiners was  then  made.

-------
                              2-30


     Three  types  of   small   refineries   were  identified  for
purposes  of  the  study;  cracking,  hydroskimming,  and  topping.
Cracking refers to refineries  employing both catalytic cracking
and  reforming   technology.    Hydroskimming   refers   to   those
refineries  with   reforming   but  without   catalytic   cracking
facilities.   Refineries  with  neither  catalytic  cracking  or
reforming facilities were defined as topping facilities.

     Sobotka   found   that   74   refineries  operating   at  the
beginning  of   1986   were   classified  as  small   refineries*
representing  about  8.4  percent  of the  total  U.S.  crude  oil
distillation capacity.   Modeling runs  for  fuel  sulfur control
and  fuel  sulfur and aromatics  control were performed,  assuming
both  segregated  and combined  production  of  diesel   fuel  and
other distillates.

     The  study concluded  that  small  topping  refineries  would
have  little  ability to  alter  the  quality  of  their  product.
Instead  of making  the  necessary investment,  to produce  a  low
sulfur fuel, these facilities  would most  likely discontinue to
produce highway diesel fuel.

     For  hydroskimming  refineries,  the  study   concluded that
costs for  sulfur  control would range  from  1.5  to 2.2  cents per
gallon  of  controlled  diesel,  depending  on  what  is assumed
regarding  current   excess   desulfurization  capacity  and'  the
ability   to   segregate   on-highway   diesel   fuel.    Cost  for
producing  fuel with 0.05  wt  percent  sulfur  and  20 percent
aromatics  ranged  from  2.7 to 4.5 cents per gallon of controlled
fuel.

     Sobotka  estimated  that the  cost  for sulfur  control  at
cracking  refineries  would  range  from  1.3   to 2.5  cents  per
gallon.   Costs   for control   of  sulfur  and  aromatics  were
estimated to range from  3.1 to  5.2 cents per gallon.

     Sobotka  compared  these small  refinery  cost  estimates to
their cost estimates for the aggregate U.S.,  which  ranged  from
1.4  to  2.3   cents  per  gallon  of  controlled  fuel  for   sulfur
control,  and  from 1.9  to 3.6  cents  per gallon  for  sulfur and
aromatics  control.  As  can  be seen,   sulfur  control   costs for
small refiners  are similar to  those of the entire U.S. refining
industry.   However, as  the  Sobotka  report   states,  reducing
sulfur would require investments  of  $49 to $98 million, a level
which may be  difficult for small refiners  to  finance.  The  cost
     Crude   oil  feedstock  of  50  MBPD   or   less,  owned   or
     controlled by  a  refiner  with  total capacity  less  than
     137.5  MBPD.

-------
                              2-31


of subsequently reducing  aromatics is substantially  higher  for
small refiners than for large refiners.

V.   Leadtime Requirements

     One important  aspect of the  leadtime requirement question
is the close proximity of this regulation  and  the  regulation of
gasoline volatility.   Investment  requirements  on the  order of
$1 billion- for volatility regulations  in  conjunction  with  the
requirements placed  on refiners due  to  diesel  fuel  regulations
may make it  more  difficult  to make both changes  simultaneously
than  separately.    Consideration  of   this  issue  was  given  by
Sobotka and Company in an analysis  of leadtime requirements  for
diesel fuel control.[15]

     SCI concluded  that, if on-highway and  off-highway diesel
fuel could be  segregated, capital  requirements  would  range from
$1 to  $2  billion,  and compliance  could be achievable by late
1993,  given promulgation of  a  final  rule, by  mid-1990.   If
segregation  of diesel fuels  proved infeasible, and  investments
of  $3.3 billion,  as  estimated  by  NPRA,  were  required,  SCI
concluded  that compliance  might  not be  feasible until  1995.
SCI could  not  estimate the   leadtime  requirements  for  aromatics
control, but given  the large amount of capital investment which
would be  required,  compliance would  likely not be  feasible in
time for the 1994 diesel engine particulate regulations.

     In a  Joint industry proposal  submitted  to EPA,  members of
the  oil  refining and  engine manufacturing  industries proposed
that diesel  fuel  containing  no greater than 0.05 percent sulfur
by weight  and meeting  a  minimum cetane index  specification of
40 could  be supplied  commercially  by October  l,  1993.[16]  The
proposal   indicated   that    segregation   of   on-highway   and
off-highway  diesel  would  likely occur  to some  extent.   Thus,
both  those  organizations responsible  for the joint  industry
proposal   and  Sobotka  agree  that  with  an   increase  in   fuel
segregation,   compliance   with  regulations   requiring  sulfur
control to 0.05 percent could take  place by late 1993.

     Even  if  domestic  refiners  were  unable to  install  the
necessary  processing  equipment  by  the  effective  date of  low
sulfur  regulations,   it  is   likely  that  foreign  refiners  could
supply  some  amount   of   low sulfur  to  the  United   States to
supplement any shortage.   Indeed,  as  indicated in a preliminary
analysis   performed   by  Sobotka  and  Company,  Inc.,  foreign
refiners  may  even  be  able  to  produce  low  sulfur   fuel  at  a
competitive   advantage   over   their  U.S.   counterparts.[17]
Sobotka  predicted  a  cost   advantage  of  0.6  to  1.4  cents  per
gallon for sulfur control for  foreign refiners.  This  is due to
the  fact  that only  a small portion of  their distillate  pool
would  be  subject  to  sulfur regulations.   The  flexibility of
foreign refiners  in minimizing production costs would therefore
be  greater.    This  analysis  did not, however,  investigate the

-------
                              2-32
cost  to  foreign  refiners  of  segregating  the  production  of  a
relatively  small  portion   of  their  distillate  pool,  nor  the
costs associated with transporting  low  sulfur  distillate  to  the
U.S.   Although  these  factors  could  reduce  the  competitive
advantage of  foreign refiners,  Sobotka suggested  that  current
imports  of  distillate  fuel of  200,000 barrels  per day  could
double or even triple if diesel  fuel  sulfur control regulations
are implemented.

-------
                              2-33


                     References (Chapter 2)


     1.     "Diesel Fuel Quality—Refining  Constrictions  and  the
Environment," George  H. Unzelman,  HyOx Inc., presented  at  the
1987 NPRA Annual Meeting,  San Antonia, Texas, March 29-31, 1987.

     2.     "Diesel   Fuel   Quality    Effects    on   Emissions,
Durability, • and  Performance,"  Craig  Miller  and  Christopher
Weaver,   Energy  and Resource  Consultants,  Inc.,  and  William
Johnson and  Terry Higgins, Sobotka and Co.,  Inc.,  EPA contract
68-01-6543 report, September 30,  1985.

     3.     "Cost and Feasibility of Lowering Diesel Fuel Sulfur
and Aromatic  Content,"  Sobotka s Co.,  Inc.,  draft final report
for EPA contract »68-01-7288, November 2,  1987.

     4.     "A  Study  on Restriction   of   Sulfur  and  Aromatics
Content of Highway  Diesel  Fuel - An Estimate of Economic Impact
on  the  U.S.  Refining Industry,"  Franklin P.  Frederick,  Bonner
and  Moore Management  Science,  final  report  for  EPA Contract
68-03-3353, June 9, 1988.

     5.     "U.S.  Refining  Industry  Capability  to  Manufacture
Ultra  Low  Sulfur  Diesel  Fuels,"  National  Petroleum Refiners
Association, Washington D.C.,  1986.

     6.     "Petroleum Marketing Monthly,"  DOE/EIA-0380 (88/06),
June, 1988.

     7.     "Comments  from the American Petroleum  Institute in
Response to EPA's Federal  Register  Request (51 F.R. 23437, June
27,  1986),"  October 27,  1986,  (Available in  Public  Docket  No.
A-86-03).

     8.     "Effects of  Diesel Fuel Standards on Transportation
and Bulk  Terminal Storage of Distillate Fuels," Memorandum  from
Doug  Koplow  and Eric  Hillenbrand,  Sobotka  &   Co.,   Inc.,  to
Millard Smith, EPA, December 17, 1987.

     9,     "Higher    Severity   Diesel    Hydrotreating,"    D.c.
McCulloch, N.D.  Edgar,  and  J.T.  Pistourius,  American Cyanamid
Company,   presented  at  the  1987  NPRA  Annual  Meeting,  San
Antonio, Texas, March 29-31, 1987.

-------
                              2-34
     10.   ".Butane Suppliers:  An  Industry Profile and Analysis
of the  Impacts of Decreased Market  Prices Caused  by Gasoline
Volatility  Control,"  Jack  Faucett  Associates,  Final  report
prepared  for  Work  Assignment  1*16,   EPA  Contract  &68-03-3244,
February 1988.

     11.   "Confronting    New    Challenges    in    Distillate
Hydrotreating,"  Jack  R.   Yoes,  Mehmet  Y.  Asim,  Akzo  Chemie
America,  presented  at  1987  NPRA  Annual  Meeting,  San Antonio,
Texas, March 29-31,  1987.

     12.   CRC  Project VE-l  "Investigation of  the  Effects  of
Fuel  Composition and  Injection  and  Combustion System  Type  on
Heavy-Duty Diesel Exhaust Emissions," SWRI Project No. 08-8673.

     13.   "Production  and Analysis  of EOS Coal-Derived Middle
Distillate  Test  Fuels from  Hydrogenation  at  Three  Levels  of
Severity," J. Erwin and N.  Sefer,  Southwest Research  Institute,
and  B.  Glavincevski,  National  Research Council  of  Canada,  SAE
Paper no. 872038, 1987.

     14.   "Cost  and Feasibility of  Lowering Diesel Fuel Sulfur
and Aromatic  Content—Impact on Small Refiners,"  Sobotka & Co.,
Inc., draft report for  EPA contract 1168-01-7288.

     15.   "Evaluation of  the Leadtime to  Comply with  Gasoline
Volatility  and Diesel  Fuel Regulations,"  Sobotka  &  Co.,  Inc.
May 5,  1988.

     16.   "Recommended   Federal    On-Highway   Diesel   Fuel
Specifications  to Assist  Engine  Manufacturers  in  Meeting the
1991  and 1994  Particulate  Standards,"  submitted  by API,  NPRA,
EMA, and NCFC,  July 19,  1988.

     17.   "Assessment of the  Relative Impact of the  Proposed
Diesel  Fuel  Regulations  on  Domestic  and Foreign  Refiners,"
Sobotka and Company, Inc., draft,  April 14,  1988.

-------
                           Chapter 3

                   Engine Control Technology
                 and Cost Prioc to Fuel Control


     The purpose  of  this chapter  is  to develop  and examine  a
scenario which  represents the  response of the diesel  industry
to emissions standards set for upcoming model  years  assuming no
change  in  diesel fuel  quality.   An  analysis  of  the  likely
developments in engine-out emissions will be presented,  as  well
as a determination of  the emission target levels  expected to be
used by industry.   The  feasibility,   application,  and cost  of
the various types of  exhaust aftertreatment devices which  may
be necessary  for compliance  with standards  for  diesel  engine
particulate emissions  will also  be  examined.   A discussion  of
the influence  of fuel  sulfur  on  each type  of  aftertreatment
device, an  estimation of the costs  and the emission  reduction
efficiencies of each technology,  and  a projection of the types
of aftertreatment devices which will  be  used assuming no  fuel
control, are presented below.

I.   Engine-Out Particulate Emissions - Current Fuel

     A.    Heavy-Putv Diesel Engines

     This section will present engine-out particulate emissions
for heavy-duty  diesel engines  (HDDE)  using baseline  fuel.   It
is  convenient  to break this  analysis  into  four  timeframes
(pre-1988,  1988-1990,  1991-1993,  1994+)   due  to  progressively
more stringent  particulate standards  effective  in 1988,  1991
and  1994.   While it  is sufficient for  air  quality  modelling
purposes  to analyze average  emission levels of  all  engines
within  a  HDDE  subclass, beginning in 1991 it is  necessary to
address  the distribution of  emissions  about the  average  in
order  to  model  aftertreatment compliance  strategies  under the
emissions averaging  program.   It  is  also necessary  to  analyze
the exhaust particulate  composition for  all  timeframes  for two
reasons:    1)  to quantify the environmental  effect  of  fuel
modifications,  since  each  type  of  exhaust  particulate  has  a
different environmental  effect,  and  2)  beginning  in  1991  to
correctly apply the various  aftertreatment devices since their
control efficiency can differ dramatically between  the  various
types  of  particulate.   Thus,  this   discussion  of  engine-out
emissions will  be  structured  as  follow:   First,   an  emission
target  level  will  be established  for each  standard.    Second,
engine-out  particulate compositions  will  then  be estimated  for
each   of  the   four   timeframes.  -Third   and  fourth,   average
emission  levels  of  pre-1988  and 1988-1990  engines  will  be
presented,  respectively.   Fifth and  sixth,   the  rationale  and
derivation  of  the   1991-1993  and  1994+  projected  emission
scenarios will  be presented.   Finally, a discussion of how  the
emission  distributions   for   these  projected  scenarios  were
derived will be presented.

-------
                              3-2


     1.     Selection of an Emission Target  Level  foe  Heaw-Dutv
           Diesel Engines

     In order  to meet the particulate  standard in 1991-93  and
1994 and  beyond,  manufacturers need  to select  a target  level
below that of  the  standard to allow for emission deterioration,
variability  in  testing,  and   engine  to  engine   production
variability.   A  methodology  for selecting  design  target levels
was developed  in the Draft  Regulatory  Impact  Analysis  for  the
NOx/Particulate Rulemaking.[1]   In that document, the  equation
for determining the zero-mile target level  (TL)  was developed:

     ZMTL = MPL-DF
             AQL

where

     MPL = Maximum emission level passing the standard,
     DF = the particulate deterioration factor,  and
     AQL = Selective Enforcement Audit Adjustment Factor.

     In  the  aforementioned  document,  the  AQL  for  particulate
was estimated  to be  1.10-1.15.   For this  analysis  the higher
level of  1.15 was used, since the  lower  emission  levels  may
produce some  upward pressure on  measurement variability  (as   a
percentage of  the  standard).   Using the deterioration factors
developed below and an AQL of  1.15,  particulate emission target
levels  were  calculated  for  the  1991  and  1994  particulate
standards.  These  target  levels  were  then used  in  estimating
the cost  and  technological requirements for compliance with the
particulate standards.

     2.    Particulate Composition

     The particulate  emitted by diesel  engines is made  up  of   a
variety of  components whose  relative amounts  vary depending on
a number of factors such as  engine  size, engine type  (direct or
indirect  injection)  and  fuel  and oil  consumption level.   When
analyzing  the   effects   of   fuel  modifications   and  exhaust
aftertreatment on  pacticulate  levels it is helpful  to know the
relative  amounts of each component  in the  particulate because
fuel  modification  and  aftertreatment  devices  tend to impact
some portions of the particulate much more than others.

     For  the  purposes of this  study the standard breakdown of
the  particulate  into soluble  organic  fraction  (SOF), sulfate
and bound water,  and  residual  carbonaceous portion  (RCP)  will
be  used.   The  SOF  consists  primarily of  hydrocarbons coming
from unburned  or partially burned fuel and  lubricating oil.   A
small  amount   (about  2 percent,  as discussed  in  Chapter  4) of
the  fuel  sulfur is  converted to  sulfates  which attract water.
Finally,  the  RCP  consists  primarily of solid  carbon but  also
includes  trace  amounts  of   ash and  other  components  from the
fuel and  lubricating  oil.

-------
                               3-3


     The   standard   method   of   determining   the   particulate
composition  is   to  determine  the   SOF  through   a   solvent
extraction  using   methylene   chloride  as   the  solvent.    The
particulate'  filter  can   be   weighed  before  and   after   the
extraction or the extract  can be dried and weighed to determine
the weight  of  the  SOF.   Following the  soxhlet  extraction  the
particulate filter, which  now contains only the RCP and sulfate
and bound water  components,  is washed with water  and isopropyl
alcohol to  remove the sulfate and bound water  leaving the  RCP
on the filter.

     With   this  technique    the   three   components   of   the
particulate can be measured   fairly  accurately.   However,  this
technique  is  time-consuming   and  costly.   As  a  result,  more
testing  is  being  done   using  the  vacuum  oven  sublimation
technigue, which consists of  baking the  particulate  filter  in a
vacuum  at  200-225°C  for  16-18  hours.   This evaporates  all of
the SOF.   However,  the literature suggests that  roughly  25-40
percent of  the  sulfate  portion  is  also removed  during vacuum
sublimation.[2,3]   Also,  since vacuum sublimation  is generally
used to get  a  guide estimate  of SOF,  an additional  analysis of
the  sulfate is  generally  not  performed   and  the  results  are
reported  only  in terms  of volatile  and nonvolatile fractions.
Since  it  is desirable to  utilize as  much  data as  possible in
analyzing  particulate  composition,  data  obtained  using  the
vacuum  sublimation technique will  be  incorporated into   this
analysis after  being  adjusted to account for the inclusion  of a
portion  of the sulfate  emissions.    To do  this,   it  will be
assumed here that  30  percent  of the sulfate is baked off during
the vacuum sublimation process  based on  the range  of roughly
25-40 percent reported in the literature.

     As  was  mentioned  previously,   the level  of  sulfate  and
bound  water  can  easily  be   calculated.   It  is   primarily  a
function  of  fuel sulfur  level,  brake-specific fuel consumption
(BSFC),  sulfur   to  sulfate conversion,  and  relative humidity.
The  equation used  to  calculate  the sulfate  and  bound  water
level for a transient test is as follows:

     S04 •»• H2O - 0.316 X (BSFC x S x CONV)         (eq.3-1)

     Where:

     §04 + H20   =     Mass    of   sulfur   plus   bound   water
                       
-------
                              3-4


sulfate   level   foe   that  engine.     In   the   absence   of
engine-specific data,  average values  are assumed as  follows:
1)  a  fuel  sulfur  level  of  0.25  weight  percent,  which  is
representative   of   typical  diesel   test   fuels,   2)   the
class-average BSFC value  found  in  the MOBILE4 conversion factor
analysis  (0.54  Ib/BHP-hr  for light HDDEs,  0.43  Ib/BHP-hr  for
medium  HDDEs and  0.39  Ib/BHP-hr  for  heavy HDDEs),  and 3)  a
sulfur  to sulfate conversion of  2.0  percent  as  discussed  in
detail in the next chapter.

     Figure  3-1  shows how  the adjusted  vacuum  sublimation data
compares  to data  obtained  through soxhlet  extraction.   This
data  set  is made up of both  current technology engines  as well
as future development engines.  At low particulate levels it is
difficult  to make  a  direct comparison due to a lack of  soxhlet
extraction  data  on  low particulate engines.   However,   in  the
mid-level particulate  range the  results  from the two techniques
seem  to relate fairly well.  Therefore,  this technique  will be
used on vacuum sublimation data to be used in this analysis.

     Data  on a  number  of  current  production  and  development
engines  were collected in  order   to  analyze  the  particulate
composition  of  both  current and future  engines.   Because much
of the data submitted  by  the manufacturers  for  this  analysis
are confidential,  it  will only be presented graphically without
identification with  respect to specific  manufacturer  or engine
family.   These  data  are  shown in  Figure 3-2.   It  should be
noted  that ten of  the twenty-one  heavy-duty data  points were
obtained  through  vacuum   sublimation  while  all  ten  of  the
medium-heavy data  points  come from soxhlet extraction  and only
one  of the  four  light-heavy data points was  obtained  through
vacuum sublimation.   It should also be  noted that  some  of  the
data  presented  in this  table were  obtained using  low sulfur
fuel  and,  as will  be discussed in  the  next chapter,  the overall
particulate  levels would  be  higher  on baseline  fuel  due to an
increased sulfate  level.   However, the SOF and  RCP levels would
remain unchanged.

      As   was previously   mentioned,  sulfate   and  bound  water
levels are primarily a function of fuel  consumption,  sulfur to
sulfate  conversion and fuel sulfur  level and  do not  tend to
vary  with total particulate level when  these  other parameters
are  held  constant.   Therefore,  it  is convenient  to  treat the
subject of sulfates  separately from  that of SOF  and  RCP.  The
data  from  Figure  3-2  are replotted in  Figure  3-3  showing SOF
vs.  total non-sulfate  particulate.   It  may  also  be helpful to
examine SOF as a percentage  of non-sulfate particulate  as shown
in Figure 3-4.   Examining the medium and heavy HDDE classes in
Figure 3-3 it appears  as  though the average  SOF  level decreases
slightly  as the  non-sulfate particulate level  decreases.   In
Figure 3-4  it  appears that the  average percent SOF increases
slightly  at  lower particulate  levels.   In  both  instances the

-------
                        FIGURE 3-1
                HHDDE SOF LEVEL vs. PARTICULATE LEVEL






Ł*
1
a.
IE
CD
_J
UJ
UJ
_J
U.
8







v.c. —
0.19 -
0.18 -
0.17 -
0.16 -
0.15 -
0.14 -
0.13 -

0.12 -
0.11 -
0.1 -

0.09 -
0.08 -
0.07 -
0.06 -
0.05 -
0.04 -
0.03 -
0.02 -
0.01 -
0 -
0







a


a

D . a -_."*"
° Dp
+ ,
D
a
+

u .,+
+ +


2 ' 0.24 ' 0.28 ' 0.^2 ' 0.^6 . ' o!4 ' 0.44 ' 0.48 ' 0.^2
p   VACUUM OVEN SUB.
PARTICULATE LEVEL (g/BHP-hr)
             t   SOXHLET EXTRACTION

-------
                     FIGURE 3-2
                SOF LEVEL vs. PARTICULATE LEVEL







o
.c
1
Q.
X
m
3
u_
0
to








u.^u —
0.24 -

0.22 -

0.2 -
0.18 -
0.16 -

0.14 -

0.12 -

0.1 -

0.08 -
0.06 -

0.04

0.02 -
n


0
0
o

o




+
*
+ i-i
n _P D
° + D 9
D
D

D o3 u
D D

02
       0.3             0!4:
            PARTICULATE (g/BHP-hr)
a   HHDDE
-»-   MHDDE
                     0.5
                                        o   LHDDE
0.6

-------
             FIGURE 3-3
      SOF LEVEL vs. NON-SULFATE PARTICULATE






-^
CL
I
^
_l
111
>
Ul
_J
u.
O
CO







U.Ł.*t —

0.22 -

0.2 -
0.18 -
0.16 -

0.14 -

0.12 -


0.1 -
0.08 -
0.06 -

0.04 -


0.02 -
0 -
n

o
0
0

0



+ a

+ +
+ + +
ib Q
* +D QD s° ?D
a
a

u , , o
u u
a

1R ' n P«i ' n ^»«; ' 045 ' n
p
    NON-SULFATE PARTICULATE (g/BHP-hr)
HHDDE       +   MHDDE       o   LHDDE

-------
      FIGURE 3-4
NON-SO4 PM:SOF PERCENT VS. LEVEL
ow -


111
«j 50 -
g
cc
2 40 -
UJ
<
^
9 30 -

Q
z
u_
O
g 20 -
O
CC
UJ
Q.
u. 10 -
O
CO
0 -
n
O
O
O
+
*
*. ,- *
+

O
+ D

D •
D
D
s
D
n


_

D D
n

•IK no«; A OR n A^, n
NON-SULFATE PARTICULATE (g/BHP-hr)
a   HHDDE
        V   MHDDE
o  LHDDE

-------
                              3-9
change is very slight and a case could be made  to  represent SOF
as  a  constant level  oc  a constant  percentage  over a  range  of
particulate levels.  However, SOF levels must decrease  in order
for engines  to  reach the very  low particulate  levels  projected
for the  1994  timeframe.  Figure  3-3  demonstrates  that  some
engines  are  already  capable   of   very  low  SOF  levels  and
therefore  shows  that  this   assumption  is  not  unreasonable.
Thus,   a  constant  percentage  SOF will be used for  all  projected
particulate- levels.  Averaging the data in Figure  3-4  shows the
SOF percentage  of non-sulfate particulate to be  51  percent for
light  HDDEs, 44 percent  for  medium  HDDE's,  and 24  percent for
heavy  HDDEs.   Urban bus  engines,   generally DDA's,  generally
fall  into  the  category  of  medium  HDDEs  as defined  in  this
study.   Thus,  44   percent  will  also be  used for the urban bus
class.

     In  summary,  the particulate composition for  all   cases  in
this  study  will be determined as follows:   first,  the  level  of
sulfate  and  bound  water will  be  calculated using  appropriate
input  information for the scenario bring  considered.   Second,
this calculated sulfate  level will be  subtracted  from  the total
particulate  level  to obtain the  non-sulfate particulate level.
The appropriate SOF  percentage  for  that weight class  will then
be  applied  to  the non-sulfate  particulate to determine the
level   of SOF with the  remainder  being RCP.   With respect  to
deterioration, the deterioration  factor  (DF) will apply only to
the SOF and RCP (i.e., sulfates  will remain constant throughout
the   useful  life  of   the  engine)   and  will   be   applied
proportionally  to  the   SOF  and RCP such  that their  relative
amounts   remain   constant  throughout   engine   life.    This
assumption  is  made  due  to  the  limited  knowledge  of  how the
factors    that    affect   particulate    deterioration   impact
particulate composition.

     3.    Pre-1988 HOPE Particulate Emissions

     The data used to obtain average  HDDE  particulate emission
rates  for pre-1988 engines were taken from the Draft Regulatory
Impact Analysis for  the HDDE NOx/Particulate Rulemaking  and is
reproduced  in  Table 3-1.[1]  At that time  transient  test data
were  available  on  a number of  current engines  (representing
approximately 66  percent of HDDE  sales of  that  timeframe) and
therefore  the  data  are   fairly   representative  of  engines
produced  in the  early  1980s.   It   should  be noted that  these
engines  were   very  low  mileage -  and  could   essentially  be
considered   to   represent   zero-mile  emission    levels.    A
discussion of in-use deterioration will be presented shortly.

     Sales  weighting these  data  (using actual sales  data from
1985)  yield  average zero-mile  particulate  emission  levels  of
0.46  g/BHP-hr  for  light HDDEs (LHDDEs),   0.66  g/BHP-hr for
medium  HDDEs   (MHDDEs)   and  0.63   g/BHP-hr   for  heavy   HDDEs
(HHDDEs).   While  engines  from  the  1970's probably  emit  at

-------
                              3-10
                            Table  3-1

            HDDS Emission  Levels  -  Pre-1988 Engines
Manufacturer/
    Engine

Light HOPE:

Detroit Diesel
    78-6.2

Medium HOPE;

Caterpillar

    3208D1T
    3208D1NA
    3208P1NA
    3208P1NA

Cummins
    NH250
    NH250

Petroit Diesel
    8.2T
    8.2T
    8.2T
    8.2T
    8.2T
    8.2T
    8.2T
Particulate
 (g/BHP-hr)
      0.46
      0.59
      0.63
      0.88
      1.09
      0.72
      0.83
      0.43
      0.43
      0.44
      0.67
      0.69
      1.06
      1.42
   NOX
(g/BHP-hr)
     3.01
    10.00
     8.49
     4.30
     4.13
     8.11
     6.80
     5.99
     6.80
     6.88
     5,
     5,
  04
  60
4.17
4.73
Navistar  (International Harvester)
    PTI466B                   0.81
    PTI466C                   0.62
    PTI466C                   0.62
    DTI466C                   0.57
    DTI466C                   0.57
    PT466C                    0.56
    PT466C                    0.53
    DT466C                    0.43
                         05
                         64
                       4.62
                       5.50
                       6.95
                        5,
                        7
       50
       05
                        9.65

-------
                              3-11


                        Table 3-1 Cont'd

            HOPE Emission Levels  - Pre-1988 Engines
Manufacturer/
    Engine

Heavy HDDEt

Caterpillar
    3306DITA
    3406DITA
    3406DITA

Cummins
    NTCC240
    NTC350
    NTC350
    NTCC400
    VTB903

Detroit Diesel
    6V92TA
    6V92TA
    8V71N
    8771TA
    8V92TA
    8V92TA
    8V92TA
    8V92TA
    8V92TAS
Mack
    EC6-235
    EM6-250
    EM6-250R
    EM6-285
    Eft 6-3 00
    EN6-300R
    E6-350
Particulate
 (g/BHP-hr)
      0.73
      0.81
      0.58
      0.77
      0.70
      0.58
      0.85
      0.72
      0.83
      0.83
      0.69
      0.42
      0.46
      0.41
      0.34
      0.52
      0.44
       0.81
       0.40
       0.49
       0.61
       0.55
       0.55
       0.36
   NOx
(g/BHP-hr)
     9.02
     4.78
     8.20
     4.78
     9.00
     7.23
     5.31
     5.27
     4.92
     4.59
     6.31
     7.03
     8.68
     9.53
    11.65
     5.61
     9.24
     9.73
     9.11
     8.81
     6.97
     8.20
     7.98
     8.61

-------
                              3-12
somewhat higher levels  than  this due to higher fuel consumption
and less control  of  HC emissions  and smoke, these  levels  will
be used for'  all  pre-1988 engines  for  two  reasons.   First,  an
accurate  assessment   of   older   engine   emission   levels   is
difficult due to  the  limited  amount  of  transient  cest  data
available on  them.   Second,  due  to  the  declining  number  of
these older engines  still in  operation,  their  impact on future
air quality projections is small.   Because actual cerrification
data for  1988 LHDDEs show particulate  levels  greater than 0.46
g/BHP-hr (i.e.,  0.514 g/BHP-hr),  and because  it  is  not likely
that   particulate   emission   levels   from  LHDDE   increased
coincidentally with  the introduction of  particulate standards,
it will be  assumed that SOF and RCP emissions from pre-control
(pre-88) LHDDE1 s  are  the  same  as those  from  1988-1990  model
year engines.  Average  end of  life emissions from ore-'88 model
year vehicles are shown in Table 3-2.

     4.    1988-1990 HOPE Particulate Emissions

     The  data  used  to determine  1988-1990  model  year  HDDE
particulate   emissions   are   presented   in   Table  3-3   and
graphically in Figure 3-5.  This  data  set  is made  up entirely
of  1988 certification  data  and  represents nearly  every major-
manufacturer  and engine  line.  Using  manufacturer's projected
1988  sales   (which  cannot  be  reproduced  here  due to  their
confidential  nature),  average  1988 particulate certification
levelsv for  the  three  HDDE subclasses were  calculated  to  be
0.514 g/BHP-hr for LHDDEs, 0,479  for MHDDEs, and 0.436 g/BHP-hr
for  HHDDEs.    Similarly,   sales  weighting  the  certification
deterioration factors  (DF)  yields  DFs of  0.009  g/BHP-hr  for
light HDDEs,  0.033  g/BHP-hr  for medium HDDEs and 0.016 g/BHP-hr
heavy HDDEs.   These  DFs will  be  subtracted proportionally from
the  SOF  and RCP  as  discussed  in  the  previous   section  on
particulate   composition  to  derive  the   zero-mile   particulate
levels   and   compositions.    Using  the  BSFC  values   and  SOF
percentages  presented  in  the  composition  section,   the  end of
life particulate  compositions  were calculated.   The  particulate
levels   and   compositions  for  all  HDDE   subclasses   for  the
1988-1990 timeframe are presented  in Table 3-4.

     5.    1991-1993 HDDE  Particulate Emissions Scenarios

     EPA has  been  monitoring  HDDE  manufacturers'   progress
toward  meeting the 1991  NOx  and  particulate  standards.  Based
on  progress   to date  it is  EPA's  assessment that manufacturers
on  average will be  able  to   meet the  1991   HDDE  particulate
standard without  aftertreatment  using  low sulfur  fuel  (0.05
percent sulfur   by   weight)   and  possibly even with   current
(baseline)  fuel.   Two  future  particulate emission scenarios
were  developed  for  this  study.   The  first,   which  will  be
considered  the nominal scenario,  assumes that manufacturers on
average will  be  able to  meet the 0.25  g/BHP-hr standard with
engine-out   reductions  alone   (i.e.,  without   aftertreatment)

-------
                        3-13
                      Table  3-2

Average Pre-1988 HDDE End of Life Particulate Levels
             and Estimated Compositions
Class
Light HDDE •
Medium HDDE
Heavy HDDE
Urban Bus
Total
(q/BHP-hr)
0.5156
0.6946
.0.6444
0.6931
SOF
(q/BHP-hr)
0.2195
0.2744
0.1399
0.2737
S04+H20
(q/BHP-hr)
0.0852
0.0678
0.0615
0.0710
RCP
(q/BHP-hr)
0.2109
0.3492
0.4430
0.3484

-------
                              3-14


                           Table  3-3

                  1988 HDDE Certification Data
Manufacturer/
    Engine

Light HDDE;.

Cummins
    4BT3.9

Detroit Diesel
    V8-6.2L

Hino
    W04C-TE
    WO4C-TD

Mitsubishi
    4D31-OAT
Navistar
    7.3L

Medium HDDE:

Caterpiller
    3208T
    3208 ATAAC

Cummins
    6BT5.9
    6BTA5.9
    6CT8.3
    6CTA8.3

Daimler-Benz
    OM366-190
    OM366-170

Detroit Diesel
    8.2N
    8.2T
   NOx       Particulate
(q/BHP-hr)    (q/BHP-hr)
Ford
     7.8L-210
     7.8L-215
     7.8L-240
     7.8L-215C
     7.8L-185F
     6.6L-165F
     6.6L-170
  5.2


  3.2
  6.8
  5.4
  5.72
  5.67
  4.7
  5.2
  4.4
  5.6
  4.7
  4.6
  5.7
  5.6
  5.4
  7.8
  8.3
  5.9
  5.4
  8.5
  5.3
  7.6
  8.4
  5.2
0.50


0.537
0.54
0.51
0.54
0.57
0.50
0.60
0.60
0.47
0.55
0.58
0.50
0.41
0.44
0.513
0.406
0.376
0.40
0.34
0.42
   415
  ,597
            Particulate DF
              (q/BHP-hr)
              0.04


              0.02
              0.00
              0.00
               0.04
               0.04
               0.00
0,
0
 0.578
               0. 128
               0.128
               0.04
               0.04
               0.12
               0.04
               0.04
               0.02
               0.02
               0.00
0.00
0.00
0.00
0.00
0.042
0.042
0.042

-------
                              3-15


                        Table 3-3 Confd

                  1988 HDDS Certification Data
Manufacturer/
    Engine

Medium HOPE:
   NOx       Particulate
(g/BHP-hr)    (q/BHP-hr)
            Particulate DF
              (g/SHP-hr)
Hino
    EM100-E
    H07C-G
    H07C-F
    H06C-TJ
    H06C-TH

Isuzu
    65A1T

Navistar
    DT-360
    DTA-360
    DT466-210
    DT466-185F
    DT466-185C
    OT466-240
    OT466-245

Nissan
    FE6-M
    FE6-A
    FE6T-M
    FE6T-A
    NE6T-M
    NE6T-A

Perkins
    180TI

Volvo
    TD71-230
    TD71-227
    TD71-FCQ
  7.6
  9.4
  5.5
  5.3
  5.7
  5.6
  8.2
  5.9
  8.4
  8.6
  5.6
  5.0
  7.8
  4.9
  5.0
  5.3
  5.5
  5.0
  5.1
  9.46
 10.3
 10.7
  5.4
0.59
0.47
0.40
0.44
0.52
0.31
0.46
0.52
0.45
0.43
0.53
0.47
0.44
0.46
0.48
0.30
0.26
0.44
0.41
0.582
0.52
0.56
0.54
0.04
0.00
0.00
0.00
0.02
0.028
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.04
0.04
0.00
0.00
0.00
0.00
0.04
0.04
0.04
0.04

-------
                              3-16


                        Table 3-3 Cont'd

                  1988 HOPE Certification Data
Manufacturer/
    Engine
   NOx
(g/BHP-hr)
Particulate
(q/BHP-hr)
Particulate DF
  (q/BHP-hr)
Heavy HDDE;

Caterpillar
    3306 PCTA
    3306B
    3306B ATAAC
    3406B ATAAC

Cummins
    NHC-250
    NTC 093E
    NTC 093J
    KTA-19
    LTA10

Detroit Diesel
    6V92TA DDEC
    360-11.1L
    S60-12.7L

Mack
    E9-500
    EMS-300
    EN6-300L
    EM6-275L

Volvo
    TD102FBQ
    TD102FGQ
    TD102FCQ
    TD122FAQ
    TD122FCQ
    TD122FEQ
    TD122FGQ
  5.5
  7.3
  5.8
  6.0
  9.1
  7.9
  5.3
  7.6
  5.4
  9.74
  8.0
  7.78
  7.5
  5.6
  6.6
  5.1
  9.6
 10.2
  5.8
  6.4
  9.7
  5.8
  5.1
  0.59
  0.56
  0.46
  0.51
  0.43
  0.46
  0.38
  0.60
  0.38
  0.363
  0.42
  0.42
  0.38
  0.51
  0.39
  0.47
  0.37
  0.38
  0.40
  0.43
  0.39
  0.39
  0.47
   0.04
   0.04
   0.04
   0.00
   0.00
   0.00
   0.01
   0.04
   0.00
   0.07
   0.17
   0.17
   0.04
   0.00
   0.01
   0.01
   0.04
   0.04
   0.04
   0.00
   0.00
   0.00
   0.00

-------
    FIGURE 3-5
NOx LEVEL vs. PARTICULATE LEVEL






_J
LU
>
UJ
_J
x
0
z






II —

10 -
9 -

8 -


7



6 -
5 -

4 -

3 -
0.
+
D +

D
+ + +
+ + a u
6 + t
D

O
a
a
+ JDD + y + j
+ + ++ n + ^
i

o

25 ' 0.35 ' .. 0.45 ' 0.55
  PARTICULATE LEVEL (g/BHP-hr)
       +   MHHDE       o
LHHDE

-------
                              3-18
                           Table 3-4

       Average 1988 HDDE Certification Particulate Levels
                   and Estimated Compositions
Class
Urban Bus

:DDE .
HDDE
:DDE
IUS
Total
(q/BHP-hr)
0.5140
0.4790
0.4360
0.4790
SOF
(q/BHP-hr)
0.2195
0.1809
0.0899
0.1796
S04+H20
<
-------
                              3-19
using low sulfur  fuel.   The second scenario, the  low emissions
scenario, assumes that manufacturers  on average will be able to
meet  the  standard  with   engine-out   reductions   alone  using
current (baseline) fuel.

     In order  to  meet the  0.25  g/BHP-hr standard  manufacturers
must establish  a  target  level  somewhat below  the  standard  in
order to  account for  engine to engine  production variability,
as  discussed  earlier   in   this   Chapter.    According  to  the
methodology outlined  there, meeting the 0.25 g/BHP-hr standard
essentially means meeting  a zero-mile  target  level of  about
0.20 g/BHP-hr,  or an  end of life target  level of  about  0.22,
assuming a  DF of 0.02  g/BHP-hr,   as described  below.   For  the
low  emissions  scenario  this  target   level  will  represent  the
average  certification (end of  life)   emissions   rate  for  all
three HDDE  subclasses which would result using current sulfur
level (0.25 percent by weight)  fuel.   However,  for  the  nominal
scenario,  this  target  level  represents  the  average  emission
rate which would  result  from using low  sulfur  (0.05  percent by
weight)  fuel.   The  average  emission  level on  baseline  fuel in
the nominal emissions  case  is higher, due to the  higher amount
of  sulfates  produced.   The difference  between  sulfate  (and
bound water) levels using low sulfur  fuel  and baseline fuel was
added to  the  target  level  to yield the average emission rates
for  the  nominal  scenario on baseline  fuel   (0.29  g/BHP-hr  for
LHDDE's,  0.27  g/BHP-hr  for  MHDDE's   and  HHDDE's,  and  0.28
g/BHP-hr for urban buses).

     Due to  the  small amount  of  composition data available on
future  development  engines,  as  well  as uncertainty  as  to how
new engine  technology will   impact particulate  composition,  two
projections of  composition will be used for all  1991 and 1994
scenarios.   Instead  of  using  the   average  percent  SOF  of
non-sulfate  particulate  as calculated  in Section  I-A-2,  this
value was  bracketed by  ±10 percent.   This  was done  since the
benefits of catalysts will  depend on  the  level of SOF and it is
important  to  determine  if  the results of  this   analysis  are
sensitive to  the  level of SOF.  Thus, for light HDDEs each case
will be  run using non-sulfate SOF .levels  of  61 and 41 percent.
For medium  HDDEs  these values will be 54  and 34 percent and for
heavy HDDEs and  urban  buses 34  and  14 percent  will  be used.
The  average end  of   life  emission levels  and  compositions for
both the nominal and  low  emission scenarios  are presented in
Table 3-5.

     In  order  to meet  a  certification level  of  0.22  g/BHP-hr
the  zero-mile  level  must  be  lower  than this to  account for
deterioration.   Due  to  a  lack of durability testing  data on
1991 development  engines,  there is some uncertainty as  to what.
DFs  will look  like  in  this timeframe.   However,  they are not
expected to change  significantly  from  1988  levels.   In  looking
at  the  1988  DFs there  is  a  great variability  among the  HDDE
subclasses.   This variability  is generally due  to  the  small

-------
                              3-20
                              3-20
                           Table 3-5

            Projected  1991-1993 Average End of Life
                Emission Levels and Compositions
 Engine
Subclass
Light HDDE
Medium HDDE
Heavy HDDE
Urban Bus
Light HDDE
Medium HDDE
Heavy HDDE
Urban Bus
Light HDDE
Medium HDDE
Heavy HDDE
Urban Bus
Light HDDE
Medium HDDE
Heavy HDDE
Urban Bus
Total
(cr/BHP-hr)
SOF S04H2O
(q/BHP-hr) (q/BHP-hr)
Nominal Scenario - Low SOF Case
0.2873 0.0842 0.0820
0.2747 0.0709 0.0662
0.2709 0.0293 0.0615
0.2772 0.0707 0.0694
Nominal Scenario - Hiqh SOF Case
0.2873
0.2747
0.2709
0.2772
Low Emission
0.2217
0.2217
0.2217
0.2217
Low Emission
0.2217
0.2217
0.2217
0.2217
0.1252
0.1126
0.0712
0.1122
Scenario -
0.0573
0.0529
0.0224
0.0518
Scenario -
0.0852
0.0840
0.0545
0.0822
0.0820
0.0662
0.0615
0.0694
Low SOF Case
0.0820
0.0662
0.0615
0.0694
Hiqh SOF Case
0.0820
0.0662
0.0615
0.0694
0.1211
0.1376
0.1801
0.1371
0.0801
0.0959
0.1382
0.0956
0.0824
0.1026
0.1378
0.1005
0.0545
0.0715
0.1057
0.0701

-------
                              3-21
number of engines  tested and the fact that the  standard is not
very stringent  and manufacturers have  not yet  had  to  develop
low DF engines.   It is  uncertain whether  this  variability will
remain in the future due to  likely  changes  in  produce offerings
by some  manufacturers.   Therefore,  a sales-weighted average of
the 1988 HDDE  subclass  DFs  yields  a DF  of  0.020 which  will  be
used for  all HDDE subclasses in the 1991-1993  timeframe.

     6.    -1994 and Later HDDE Particulate Emissions

     The  same  scenarios   presented  for   1991-1993  emissions
(nominal  and low) including the two composition  cases (high and
low  SOF)  will  be  retained  for   modelling  1994  and  later
emissions.   Due  to uncertainties in  the particulate reduction
potential   of   advanced  engine   technology   which  will   be
implemented in 1994, an  accurate assessment of  1994 particulate
levels is difficult.  However,  rapid  advances  are being made in
engine-out  particulate   controls,  as   evidenced by the  large
decrease  in  emissions  expected  by  1991,  and  EPA  believes
continued  progress  beyond  1991  is  likely.    To  model  this
progress, it was  assumed for  all  emissions scenarios  and HDDE
subclasses  that  there   would  be   a  35  percent  reduction  in
non-sulfate particulate from  1991  end-of-life  levels  by 1994.
Sulfate  levels  tend  to   be   primarily  a  function  of  fuel
consumption and will  not likely be reduced by engine technology
but  only by  reductions in fuel  consumption  (conservatively
assumed  here  to  be  zero).  Thus,  a  35  percent  reduction in
non-sulfate  particulate  translates   into  a   23-30   percent
reduction  in  overall   particulate,  depending  on   the  engine
subclass and emissions   scenario.   Due  to uncertainties  in the
effect   of   future   control   technologies   on    particulate
composition,    this   35   percent   reduction    was   applied
proportionally  to  both the  SOF  and  RCP  portions  of  the
particulate  (i.e.,  a 35  percent  reduction will be  applied to
each).   The average 1994 end of life particulate levels for all
HDDE subclasses and emissions scenarios are presented  in Table
3-6.

     For  the  1991-1993  timeframe  an  additive  DF of  0.020
g/BHP-hr was used.  However,  some  advances are  expected to be
made by  1994 in the areas  generally associated with  particulate
deterioration such  as oil control.  To  model  this progress a 25
percent  reduction  in deterioration was  assumed.   Therefore,   a
DF of 0.015 g/BHP-hr was used for all 1994  and later model year
HDDE emission scenarios.

     7.    Engine-out Emissions Distributions

     In  the preceding  sections  average particulate  levels were
projected for  the   1991  and 1994  timeframes for  all three HDDE
subclasses.    However,    in  order   to   model   aftertreatment
compliance  strategies   more  accurately  under  the emissions
averaging  program, the distribution of  emissions   from  various

-------
                              3-22
                            Table  3-6

                Projected 1994 and Later Average
          End of Life Emission Levels and Compositions
 Engine
Subclass
Light HDDE
Medium HDDE
Heavy HDDE
Urban Bus
Light HDDE
Medium HDDE
Heavy HDDE
Urban Bus
Light HDDE
Medium HDDE
Heavy HDDE
Urban Bus
Light HDDE
Medium HDDE
Heavy HDDE
Urban Bus
Total
(q/BHP-hr)
SOF S04H20
( q/BHP-hr ) ( q/BHP-hr )
RCP
(q/BHP-hr)
Nominal Scenario - Low SOF Case
0.2159
E 0.2041
0.1985
0.2053
0.0562
0.0474
0.0196
0.0468
0.0788
0.0646
0.0583
0.0678
0.0809
0.0921
0.1206
0.0908
Nominal Scenario - Hiqh SOF Case
0.2159
E 0.2041
1 0.1985
0.2053
Low Emission
1 0.1727
iE 0.1692
! 0.1660
0.1691
Low Emission
! 0.1727
IE 0.1692
! 0.1660
0.1691
0.0836
0.0753
0.0477
0.0743
Scenario - Low
0.0385
0.0356
0.0151
0.0344
Scenario - Hiqh
0.0573
0.0565
0.0366
0.0547
0.0788
0.0646
0.0583
0.0678
SOF Case
0.0788
0.0646
0.0583
0.0678
SOF Case
0.0788
0.0646
0.0583
0.0678
0.0535
0.0642
0.0925
0.0633
0.0554
0.0690
0.0926
0.0669
0.0366
0.0481
0.0711
0.0466

-------
                              3-23


engines about  the  average level  must also be  projected.   This
section describes  the  method used  to arrive at  a  standardized
distribution and then discusses  its application specifically to
the 1991 and 1994 emission scenarios.

     In  the   absence  of  sufficient  data  on  the   emission
distributions of future engines  the next step is to examine the
distribution.   of   current   engine   emissions.     The   1988
certification  data presented in  the  discussion of  1988-1990
emissions are shown in Figure 3-6 in  a  sales  vs.  emission level
format.  It  can be seen  from this  figure  that  the  majority of
HDDEs are clustered around the  0.5 g/BHP-hr range with  some as
high  as  0.6 g/BHP-hr  (the  level  of the  1988 standard)  and a
larger number  spread out  below the 0.5 g/BHP-hr level  all  the
way down into  the  0.3  g/BHP-hr  range.  Analyzing the subclasses
separately,  the  heavy  HDDEs   also  follow this trend,  while  the
medium HDDEs tend  to  be  more evenly  spread out  and  light HDDEs
are all clustered  around  the 0.5  g/BHP-hr  level.  While  it is
apparent  from  this  graph   that  there  is  a definite   skewed
distribution in  1988 there are  two main factors which  lead EPA
to conclude  that this distribution is  not appropriate  foe  the
1991/1994 timeframes.   First, in 1988 all engines  are required
to meet  the 0.60  g/BHP-hr  standard whereas  in  1991  and  1994.
under the emissions  averaging program  engines can  be produced
which  do not  meet the  standards  as  long  as  all engines on
average meet the standards.   For  this  reason EPA  expects  some
manufacturers  to continue to produce some engines  lines which
emit well above  the standard based on  their  proven performance
in the marketplace.    Second, the  1991 and  1994  standards  are
significantly  more stringent than the  1988  standards  and  EPA
does  not  expect to see  a large  number of engines  emitting at
levels  substantially  below   the  level   of  the   1991  and   1994
standards.  For  these  reasons a  log-normal distribution skewed
toward higher emissions was used to  represent  the 1991 and  1994
emissions distributions.

     The first  step in arriving at the 1991 log-normal emission
distributions  was  the construction  of  a discrete  standardized
normal distribution.   The discrete  distribution  was assumed to
contain eleven intervals.  This  number  is small  enough  to  keep
the  analysis  to  a manageable  level yet large  enough to allow
sufficient  resolution  of  the distribution  to  adequately model
aftertreatment  compliance strategies.   Also, a  total range of
±3 standard deviations (+3Z), which.covers 99.74 percent of  a
normal  population was  used.   Dividing the  range  of  +3Z  into
eleven discrete  segments  yielded the twelve  segment end points
and   their   corresponding   population   fractions  (here  sales
fractions)  as  shown in Table 3-7.   Choosing the  midpoint  for
each  segment  to represent that  segment yields the  Z values  and
the  corresponding fraction  of  engines  represented by  each of
the eleven  intervals, also shown  in Table 3-8.

-------
                               FIGURE 3-6
zg
      400
                         ENGINE SALES vs. PARTICULATE LEVEL
      350 -
      300 -
      250 -
      200 -
      150 -
      100 -
       50
        0
             0.^7
            ..   0.45
PARTICULATE (g/BHP-hr)
       MHDDF
                                                      LHDDE

-------
                            3-25


                         Table  3-7

         Standardized Discrete  Normal Distribution

Standardized Normal Deviate
	(Z-value)	      Cmnmulative Sales Fraction

           -3.00                          0.0013
           -2.45                          0.007L
           -1.91                          0.0281
           -1.36                          0.0869
           -0.82                          0.2061
           -0.27                          0.3936
            0.27                          0.6064
            0.82                          0.7939
            1.36                          0.9131
            1.91                          0.9719
            2.45                          0.9929
            3.00                          0.9987

-------
                              3-26


     The_  relationship  needed to  move  from this  standardized
normal distribution to a log-normal distribution is:

     log(x) = Z log (0) + log (H)                (eq.  3-2)

     or:

     x » exp (2 log(tf)+log (U»

     where:

     log (V4)     =  mean  value  of  normal  distribution  (i.e.,
                 logarithm of emissions)
     log (°)     » standard deviation of the normal distribution
      z          - standardized normal deviate
      x          »  mean   value   of   interval   point   in   the
                 log-normal distribution (i.e.,  emissions)

     In order  to  make use of this  equation  both  the  mean value
and standard deviation of the normal distribution (log  x)  must
be known.   The mean value for all scenarios were derived in the
previous  sections  and  the log  of  these  values  can  simply be
found.    Standard  deviations   were   chosen  which   provided
reasonable  upper  and  lower  bounds  for  the  scenario  being
considered.   This  is described  below in more detail  for  the
1991, nominal  emissions case.   Since  the distribution  of  only
the   non-sulfate   portion   of    the  particulate    is   being
constructed, here  the ratio of  the maximum values for  the  low
emission  scenarios was assumed  to  be the same as the ratio of
the mean emissions of the two scenarios.

     In  light  of  progress  being made  on  particulate  control
technology,  for the  nominal,  1991  scenario, an upper bound of
0.40  g/BHP-hr  total  particulate  appears  to  give   reasonable
lower bounds  for  all  cases.   For  the lowest emission case (low
emission  scenario  for  HHDDEs)  this  yields  a  zero-mile total
particulate level of  0.1405 g/BHP-hr.  All  the  other vehicles
classes  in  the 1991  low  emission scenario  and all vehicles
classes  in the  nominal  emission scenario  have higher lower
bounds.

     First,  the calculated sulfate levels were subtracted from
this  value to  give  the   class-specific   non-sulfate   maximum
values.   Class-specific non-sulfate maximum values  for  the low
emission   scenario  were  then calculated,  as  described above.
Finally,   Equation 3-2  was  used  along  with  the  non-sulfate
maximums  and means (2 =»  2.725)  to  yield the standard  deviations
for  all vehicle  classes  of  the two  1991  emission   scenarios.
However,  because  the  urban bus  market  is  so  limited in nature
with   respect  to  the   number  of   engines   available,   the
distribution of emissions about the mean  is expected  to  be much
narrower   than  for   the   medium HDD! class.   To  model   this
difference the  standard  deviation for  the  urban bus scenarios

-------
                              3-27


was  assumed  to be  one  half  of  the  standard deviation  of  the
corresponding medium HDDE scenarios.

     These standard deviations  were used  in Equation  3-2 along
with the  non-sulfate  means and the Z values  from  Table  3-8 to
calculate   the   non-sulfate    end   of   life   1991   emission
distributions.  The DF  of  0.02 g/BHP-hr  was  then  subtracted
from each  point in each distribution  to yield  the r.on-sulfate
zero-mile   distributions   for   1991.    Each   point   in  each
distribution  was  then  broken  down  into   its"  SOF   and  RCP
components  according  to  the  compositions presented  in Section
II-A-5.  Following  this  the calculated  sulfate  level  for  each
scenario   was  added   to   each   point   in  the   non-sulfate
distributions   to   obtain  the  zero-mile   total   particulate
engine-out  emission  distributions   for   1991.   The  zero-mile
distributions   for   all   1991   emission   scenarios   and  HDDE
subclasses are presented in Appendix 3-A.

     As was discussed  in Section II-A-6  on 1994 emissions, both
emission scenarios  for  1994  assume a 35  percent  reduction in
non-sulfate particulate  from  1991.  To achieve this, the  SOF
and  RCP   values   for   each   point  in  each  1991  zero-mile
distribution will be  reduced  by 35 percent.  The 1994 zero-mile
distributions  for all  emission scenarios  and  HDDE  subclass  ace
also presented in Appendix 3-A.

     B.    Light-Duty Diesel Emissions

     Although  the  vast  majority  of highway  diesel  fuel  is
consumed by engines  classified  as  heavy-duty,  any modification
to  diesel  fuel quality will  also  have  an effect  on emissions
from light-duty diesel vehicles (LDDVs)  and trucks (LDDTs).   In
order  to determine  the magnitude of this effect, the first  step
to  be  performed is  to  establish  baseline  emission  values  for
both classes of vehicles.

     A number of  studies  of  emissions  from  light-duty  diesel
vehicles  have  been performed   in  which  the  fraction of   fuel
sulfur   converted    to   particulate    sulfate    has     been
measured.[4-7]   These   studies  show  that  typical  sulfur  to
sulfate  particulate conversions range from  1  to  2 percent of
the  sulfur  in the  fuel.   The  mean  conversion  level  of  1.5
percent  was  used  in  conjunction   with  fuel  economy data to
determine  emissions of  sulfate, bound water,  and sulfur  dioxide
for  LDDV and  LDDT.

     Baseline  total  particulate emissions  for LDDVs  sold in
model   years   prior  to  1987   were   estimated  from  pre-1987
certification test  results.   Sulfate  emissions  were calculated
based  on  the  1.5 percent  conversion of fuel  sulfur to  sulfate
and model year specific fuel  economy data.   SOF was  estimated
to  be 18  percent  of  the total  particulate,  the  average of
several  LDDV  engines  tested  in a  particulate  trap  evaluation

-------
                        3-28


                     Table  3-8

Standardized Normal Distribution Interval Parameters

 Mid-Point Z-value           Fraction of Engine Sales

       -2.725                        0.0058
       -2.180                        0.0210
       -1.635                        0.0588
       -1.090                        0.1192
       -0.545                        0.1875
        0.000                        0.2128
        0.545                        0.1875
        1.090                        0.1192
        1.635                        0.0588
        2.180                        0.0210
        2.725                        0.0058

-------
                              3-29


program.[8]   The remainder of  the particulate by definition was
assumed  to  be   RCP.    For   1987  and   later  LDDVs,   1987
certification  test  results  were  used  to estimate  non-sulfate
particulate emissions  emissions.  SOF  was estimated  to be  18
percent  of  total   particulate.    Sulfate  emissions  were  then
calculated  using  MOBILES  fuel   economy  projections.   As   a
simplifying assumption,  particulate  levels for  all  LDDVs  were
assumed to remain constant over the full life of  the vehicle  at
the  end-of-life  level.   Emissions  of   HC   were  taken  from
AP42.[9]   Baseline end-of-life emissions for  LDDVs  are  shown  in
Table 3-9.

     Baseline  emissions  for LDDTs less than 6,000  pounds  GVWR
(LDDTls)  were determined in the same way as LDDVs.   Once again,
total  particulate  end  of  life   emissions  were  estimated  from
certification test results.  For LDDT1,  it was  assumed  that the
SOF  represented  50  percent of total particulate,  based on  test
data obtained on a  light-duty  diesel truck.[10]   Sulfate, bound
water, and  sulfur  dioxide emissions  were determined  assuming
1.5  percent  sulfur to  sulfate  conversion as  explained above.
RCP  emissions  were calculated by subtracting  SOF  and sulfate
emissions  from  total  particulate  levels.    As  with  LDDVs
particulate  emissions  were   assumed   to   remain  constant  at
end-of-useful-life  levels  over   the   life  of   the  vehicle.
Emissions of  HC were  taken  from AP42.[9]   Baseline emissions
for LDDTls are shown in Table 3-9.

     Total particulate  emissions  for  LDDTs  over  6,000  pounds
GVWR  (LDDT2s)  for  model  years  prior  to  1988  were determined
from  certification  test  data.   Sulfate  and  sulfur  dioxide
emissions were determined  from fuel  economy data assuming a 1.5
percent  sulfur  to  sulfate conversion as  described  above.   The
remainder of the particulate was assumed to  be  SOF and RCP,  in
relative  proportions  the  same  as  those  of  light  heavy-duty
diesel engines because  the only engine  used in LDDT2s  is also a
LHDDE.   Total   engine-out  particulate  emissions   from engines
certifying for  model  year 1988  and beyond were estimated to be
0.295  grams   per  mile,   as   projections   by  manufacturers
indicated.   No  reduction  in  engine-out  emissions  beyond  1988
levels  is anticipated  (i.e.,  reductions  necessary to  comply
with the 1991  standard  of 0.13 gpm will be achieved by the use
of trap-oxidizers).  Baseline  emissions for  LDDT2  are  shown in
Table 3-9.

II.  Exhaust Aftertreatment Technology

     In  this  section,  the various  types  of  exhaust  treatment
devices  which  may be   employed  to  meet  the   1991   and  1994.
heavy-duty  particulate  standard,  both  trap-oxidizer  and  flow
through  catalyst type  systems,   will  be  presented.   Both the
emission reduction  potential  and the costs of such  devices will
be examined.

-------
                                  3-30
                                Table 3-9

                        Baseline Emissions - LDDV
Model
Year
LDDVs
1979
1980-81
1982-84
1985-86
1987
1988-90
1991-93
1994+
LDDT1
1979-80
1981
1982-84
1985-86
1987
1988-90
1991-93
1994+
LDDT2
1979-80
1981
1982-84
1985-86
1987
1988-90
1991-93**
1994+**
Fuel Econ
(MPG)

25.1
25.1
27.1
28.4
28.4
29.8
31.4
34.2

21.0
21.0
23.0
23.8
23.8
24.7
25.9
27.8

21.0
21.0
23.0
23.8
23.8
24.7
25.4
27.3
Emissions (a/mi)
SOP
0.046
0.046
0.046
0.046
0.024
0.024
0.024
0.024
0.135
0.135
0.135
0.135
0.108
0.108
0.108
0.108
0.188
0.188
0.188
0.188
0.180
0.138
0.043
0.044
RCP
0.178
0.178
0.178
0.178
0.079
0.079
0.079
0.079
0.099
0.099
0.099
0.099
0.072
0.072
0.072
0.072
0.167
0.167
0.167
0.167
0.159
0.122
0.012
0.013
SO*
0.035
0.035
0.032
0.031
0.031
0.029
0.028
0.025
0.041
0.041
0.038
0.036
0.036
0.035
0.033
0.031
0.041
0.041
0.038
0.036
0.036
0.035
0.034
0.032
TPM
0.260
0.259
0.256
0.255
0.134
0.132
0.131
0.128
0.275
0.275
0.272
0.270
0.216
0.215
0.213
0.211
0.396
0.396
0.393
0.391
0.375
0.295
0.089
0.089
S02
0.651
0.651
0.603
0.575
0.575
0.549
0.520
0.477
0.779
0.779
0.711
0.687
0.687
0.662
0.631
0.588
0.778
0.778
0.710
0.686
0.686
0.661
0.637
0.594
HC*
0.42
0.29
0.29
0.29
0.29
0.29
0.29
0.29
0.86
•0.43
0.43
0.43
0.43
0.43
0.43
0.43
0.86
0.43
0.43
0.43
0.43
0.43
0.15
0.15
     Zero-mile emissions.
**   Emissions levels reflect the use of  trap oxidizers

-------
                              3-31


     A.     Exhaust Aftertreatment Types

     1.     Trap-Oxidizer Systems

     A  trap-oxidizer  system  is an  exhaust  treatment  device
which collects  solid  particulate matter  in  the  exhaust  and
subsequently   oxidizes   the   collected   particulate   (i.e.,
regenerating  itself)  at  certain intervals.   Possible  devices
used  for  trapping  of  the  particulate  matter include  ceramic
wall-flow monolith,  ceramic  fiber  coil, and  wire mesh  media.
Possible  methods   of  regeneration  include  the  use  of  burner
assemblies,   electrical  heaters,  fuel  additives,  and  catalyst
formulations  for   passive regeneration.   A brief  overview  of
these different systems is presented in this section.

     a.     Ceramic Wall-Flow Monolith

     The wall-flow  monolith trap is  designed  similarly  to  the
ceramic  honeycomb  substrate  used  in  gasoline-fueled  vehicle
catalytic converters.   Alternate upstream and  downstream cells
of the  honeycomb  matrix  are  plugged,  causing  particulate-laden
exhaust gases  to  filter  through the  porous walls.   The result
is a  high efficiency capturing  of  the carbonaceous particulate
components  which  must  in turn  be  periodically  regenerated-or
burned  off.   The  system  can  be either uncatalyzed  or  may be
catalyzed with  either base  or noble metals.   A more  complete
description  of   this   device  can   be  found  in  "An  Updated
Assessment of the Feasibility of Trap Oxidizers."[11]

     b.     Johnson-Matthey Catalyzed Wire Mesh

     The  catalyzed wire  mesh trap  is  a  passive  regeneration
system  which  traps  particulate and  subsequently oxidizes  it
under certain operating modes with  the aid of  a  catalyst.   The
wire  mesh  system  has   the   attraction  of  not  requiring  an
expensive regeneration  system, but does  not  trap  particulate as
efficiently  as  the  ceramic  monolith  system.   Problems  with
increased   sulfate   emissions  due   to  the   catalyst  remain
unresolved  for  current high  sulfur fuels.   For a more detailed
description  of  the  operation of such a  system  the  reader is
referred to the Diesel  Particulate Study.[12]

     c.    Ceramic Fiber  Coil

     This type  of trap was  developed as part of a cooperative
effort  by  Daimler-Benz  and  Mann  &   Hummel.[13]   The   ceramic
fiber coil  trap  consists of  a number  of  perforated stainless
steel cylinders wound with silica fibers  (the trapping  medium)
and  arranged  in  parallel  in a  stainless steel  housing.   The
number  of windings affects the filtration  efficiency as  well as
the  back pressure  associated with  the trap.   Regeneration of
the  trap   takes  place  via   some  type  of   oxidant   injection
system. [14]   For  more  details on this  type of trap,  the reader

-------
                              3-32
is   referred  to   the   RIA  for   the   1985   NOx/Particulate
Rulemalcing. [1]

     2.     Flow-through Catalyst

     A low-cost alternative to the trap-oxidizer  type  system is
a  flow-through  oxidation catalyst  similar  to  those  used  on
gasoline  vehicles.   Work  aimed  at  developing  both  ceramic
monolith type  and pellet type catalysts  for  diesel application
is currently underway.   The  operational  requirements  are  such
that  the  catalyst  formulation  must satisfactorily  oxidize  the
organic particulate while minimizing  the oxidation of S02  to
particulate  sulfate if  high sulfur fuel  is used.   Typically,
those  catalysts  which  are   the  most  active  on  the  organic
particulate  also  tend  to oxidize  sulfur dioxide to the greatest
extent.    Determining  the   optimal  catalyst  formulation  and
loading is the subject of a considerable research effort.

     B.     Exhaust Aftertreatment Efficiencies'

     Accurately estimating the  efficiency of  the performance of
the  aftertreatment  systems  discussed  above  on  the  various
components of  diesel  emissions  is a critical link in projecting
their application on vehicles certifying with the 1991  and 1994
particulate  standards.   A great deal of test data exist  in the
literature,  much  of which is contained in several  SAE special
publications on diesel particulate  control.   Both transient and
steady  state testing  on engine  and  chassis  dynamometer  have
been  performed  on  these  systems.   A  discussion  of  system
efficiencies by type is presented below.

      1.    Trap Qxidizer Systems

      a.    Ceramic Wall-Flow  Monolith (Uncatalvzed)

     The  analysis of  the test data was  approached with  some
degree  of  selectivity;  particularly  in  the  areas  of ceramic
monolith  testing,  where  a   substantial   body  of  data  exist.
Since   emission  characteristics   of   heavy-duty  engines  are
different in steady-state and transient modes,  and because the
federal  test  procedure  for  heavy-duty  diesel  engines  is   a
transient  test,  only data generated in the  FTP transient mode
were  considered.   Data generated from light  duty vehicles were
excluded  in this analysis,  due to the  difference in trap sizes
and  particulate  composition for  these vehicles.   After  these
criteria were  applied, the data used in the  heavy-duty analysis
consisted  of  SAE papers, EPA  reports and  confidential  engine
manufacturer submissions.[13,15,16]  A summary  of  the results
of  the  tests performed  on heavy-duty  engines is shown in Table
3-10.    Though    confidential    data   received    from  engine
manufacturers  are  not  presented,  consideration  of  these data
were  included in  determining the  "Best  Estimate  Efficiencies"
shown in  the  table.   Many  ceramic monolith  traps  of varying

-------
                                                 Table 3-10
                    Heavy-Duty Uncatalyzed Ceramic Monolith Efficiency-Transient Testing
                                                                        Efficiency (Percent Reduction)
Source
EPA [15]
EPA (16]
EPA [16]
SAE [13]

Vol.
' Engine(Vehicle) Trap (Cu In) Cat
NTC400 EX47 1193 N
DDAD 6V71 EX47 1193 N
(GMC RTS II) EX47 1193 N
(MB 0 405 Bus) — — N
Best Estimate = N
SOP RCP S04 TPM*
82
37.5 93.5 66.8 61.3
87.1 93.0 15.4 92.1
70-95 — 51-82
55 90 0 71*
HC CO
8 -50
0.5 14.1
34.6 16.6
4-15 0
13 -6
Contingent on engine emission characteristics.

-------
                              3-34


size, porosity  and pitch  have been  tested.   Efficiencies  and
fuel  economy penalties  vary  with  design  parameters  of  the
monolith  system,   although   general   conclusions  about   the
efficiencies of  the ceramic monolith trap can be drawn.

     As  illustrated  in  Table  3-10  the  ceramic  particulate
filters design acts primarily  on  the solid carbonaceous portion
of  the  diesel   particulate.    Solid  carbon   (RCF)   trapping
efficiencies  reported  generally  range  from  70  to 95 percent.
The  average  efficiency  yielded  from  the  available  data  is
approximately 90  percent.   The  efficiency of  the trap  on  the
SOF fraction of  particulate is significantly  less  than this,  as
most  passes through the  trap  in the gaseous  phase.   However,
some  of the SOF  particulate is  adsorbed onto the  trapped RCP,
and  by  this  mechanism  is  controlled   to   a  certain  extent.
Although the  SOF  efficiency varies  widely from test  to test,
the data indicate  that  a  SOF reduction of 55 percent on average
can be expected.

     Many references in the literature also  report decreases in
the  sulfate portion of the particulate with  the use  of ceramic
monolith traps.   However,  no  mechanism by which the  trap might
reduce  the  sulfate to  S02  seems likely, and  theoretically no
reduction in the  amount  of  sulfate formed  in  the   engine  is
expected. It  is probable  that the apparent reduction  in sulfate
particulate  levels by  the trap  is  attributable  to  temporary
sulfate storage  in the trap,  which,  upon the  eventual release
of the  sulfate  from the trap  will  disappear.   It  is  therefore
concluded that sulfate particulate  emission will not be lowered
by using an uncatalyzed ceramic monolith trap.

      A  reduction  in gaseous HC  emission with  ceramic trap use
has  also been reported in some of the  literature [13, 15, 16],
although the actual  degree of   reduction  varies  widely.   The
preponderance  of   data  suggest  that on average  a 13  percent
reduction  in HC   emissions  will  result  from  the use   of  an
uncatalyzed  ceramic monolith  trap,  although  actual   reductions
may   range   from   0  to  50  percent  depending  on  operating
conditions.   The  data also suggest  an  increase in emissions of
CO  of 6 percent.   Actual  test  data reported  range   from  a 50
percent increase to a 17 percent  decrease.

      b.     Ceramic Monolith (Catalyzed)

      The  emission  reduction  efficiencies  described  above  are
those which can be  expected for  an  uncatalyzed ceramic monolith
trap.   Some attention has  been given to  ceramic monolith trap
designs implementing base  and/or  noble  metal  catalysts in order
to   facilitate  passive  regeneration  and/or   increase   overall
efficiency.   Based  on  data  in the   literature  as well  as
confidential  data  submitted,  the  efficiency  of  the  catalyzed
trap on the RCP fraction of  the  particulate appears to be about
the  same as  that  of  the  uncatalyzed  ceramic  trap.   The data

-------
                              3-35


show  that  the  efficiency  of  the  catalyzed  trap  on  the  SOF
fraction is much higher than the uncatalyzed version.

     As expected,  however,  the use  of high-activity  catalysts
with  the  ceramic  monolith  may greatly  increase emissions  of
sulfate, up to seven times the engine out level with the  highly
active  catalyst  formulations which  appear  to be  necessary  for
dedicated  passive  regeneration.   This  is   a   major   problem
preventing  the  use of high  activity catalysts in these  devices
to meet 1991 and 1994 particulate standards  without  fuel  sulfur
control.

     A  passive   regenerating  ceramic   monolith   system   not
requiring a supplemental active regeneration  system  as a  backup
has  not  yet been  satisfactorily demonstrated   for  heavy-duty
engines,  and  therefore  can  not  currently  be  considered  a
feasible option.   It  may, however,  in some cases be  desirable
to use  a  catalyst with  a active regeneration  ceramic trap  in
order  to  increase  the  SOF  reduction capability of the  trap.
This  type  of  device  would  be  particularly  attractive  for
1ight-heavy-duty engines  applications,  where the SOF  comprises
a substantial portion of the  particulate.   While this  type of
trap would  still  require an auxiliary regeneration  system,  the
resulting  increase in particulate  collection efficiency  above
and beyond  an uncatalyzed ceramic trap  might in some cases be
the most cost effective approach to particulate control.

     In order to estimate the efficiency of such  a  device,  the
particulate  reduction efficiency of an  uncatalyzed  particulate
trap was  combined  multiplicatively with  the efficiency of  the
flow-through oxidation catalyst devices  described later  in this
chapter.  Efficiencies for two  types of  catalyzed traps  ("high"
and "low"  cost)  were determined, analogous  to the two  types of
flow-through catalysts systems  described below.   The efficiency
of the  "high cost" catalyzed  trap was determined by combining
the  efficiency  of   an   uncatalyzed  ceramic  trap  with  the
efficiency  of  the  "high cost" or "high  activity"  flow-through
catalyst.    The efficiency of the "low coat"  catalyzed trap was
determined  likewise.  These  emission reduction efficiencies are
shown in Table 3-11.

     It is  worthy  to note that,  while some  data are available
in  the literature  on the particulate collection efficiency of
catalyzed  ceramic   monolith  traps,  .these  data   were  not  used
directly  in this  analysis.   Due  to the proprietary  nature of
the research,  in many  instances it  is  difficult to  determine
much  about  the  catalyst formulation  and loading corresponding
to any  specific  piece of test  data beyond  whether the catalyst
was a'll noble metal or a mixture of base and noble  metals.  To
assure  consistency  between  the  efficiency  and cost  of  the
catalyzed traps  used in  this  study  and  the efficiency and cost
of  the other  aftertreatment  devices used,  efficiencies  were
determined  by combining  the efficiencies  of uncatalyzed traps
and flow through catalysts as described above.

-------
                              3-36


                           Table 3-11


         Efficiency of Catalyzed Ceramic Monolith Traps


                               Efficiency (Percent Reduction)
   Type           SOF        RCP        SOA         HC       CO

Low Cost-Low       69         90         0          39       -6
  Efficiency
High Cost-High     91         91        -100        83       90
  Efficiency

-------
                              3-37
     c.     Johnson Matthey Catalyzed Wire Mesh

     Trapping  efficiencies  of  Johnson  Matthey  catalyzed  wire
mesh traps  as  found in  the  available  literature  are shown  in
Table    3-12.[15,17,18,19]      Total    particulate    trapping
efficiencies for  this  trap  type   (excluding  the  type  I  trap
tested by  Ulman [17])  range  from  46 to  86  percent.   It should
be  pointed  out,   though,  that   some  of  the   higher  total
particulate  efficiencies  reported  were  taken  from  high  SOF
engines.     Only  two   reports   in   the   literature    gave   a
compositional       breakdown       of      the       particulate
efficiencies.[17,19]   The  average  of  particulate  collection
efficiencies reported  were  87 percent on  SOF,  47 percent  on
RCP, and -136  percent  on sulfate (i.e.,  additional  136  percent
over engine-out sulfate).

     d.     Ceramic Fiber Coil

     Little  test  data  exist  in the literature on  the  ceramic
fiber  coil particulate  trap.    However,   results  of   some
development  work   and  demonstration   programs   have   been
documented   by   Hardenberg    and   others    in    the    SAE
literature.[13,14]   The transient  test results that  have been
reported  show   that these  devices'   total  particulate  and  HC
emission  reduction  efficiencies  are  similar  to  that  of  the
ceramic  monolith  system.   Soot  trapping  efficiency  of  the
ceramic fiber  coil is somewhat higher  than  that  of the ceramic
monolith,  approaching  100 percent  as the trap becomes  loaded.
A  slight  increase  in  CO emissions  (about  10 percent) occurs
with the use of the ceramic fiber coil.

     2.      Flow Through Catalyst

     Little  recent  information  exists in  the literature on the
performance  of  flow  through  catalysts  on diesel  emissions,
although  the  results  of steady-state  testing  of  an Engelhard
catalyst  on a single-cylinder  engine  are  available.[20]   The
majority of  the information on  catalyst efficiency was  obtained
through  confidential submittals  and conversations  with engine
and  catalysts manufacturers.    Because  of  the  confidential
nature of the  material,  little discussion  can be presented here.

     Of  major   concern  in developing an oxidation catalyst for
diesel  engines  is  determining  a  catalyst formulation  which
satisfactorily   oxidizes   the   organic   particulate   while
minimizing   the  oxidation   of   sulfur  dioxide   to  sulfate
particulate.   Typically,  the  formulations  which  provide  the
highest  reduction in organic  particulate also tend  to produce
the most sulfate.

     Based  on   the  data  received,  two  representative  types of
flow through catalysts  have been used in this analysis.   One is
a  "high  efficiency, high cost" system.   Theoretically,  assuming

-------
               Table 3-12
HDD Johnson Matthey Hire Mesh Efficiency
                                 Efficiency (Percent Reduction)
Source Engine Trap
[15] NTC400 JM
[17] DDAD 8V-71TAC JM I
[17] " JM II
[18] Scania DS11 15 JM
[19] Mining JM

* Contingent on engine emission
** Doe rat ion on low sulfur fuel.
Vol.
(Cu In) Catalyst SOF
__ v — —
Y(aged) 82
Y(aged) 90
— V — —
Y 72-97
Best Estimate 87
characteristics .
RCP S04 TPM* HC
— . 46 86
-100 -250 25 80
32 -136 64 85
71** 40
60-63 Stor. 65-86
47 -136 64 70
CO
-9
60
46
71
—
36

-------
                              3-39


that the  conversion of  the  catalyst  is reaction  rate  limited
(i.e.,  that the system can be designed so that  mass transfer is
not  the  limiting  factor),  it  should  be possible  to design  a
system for  a diesel vehicle  achieving  the  same conversion of
gaseous hydrocarbons  (or  SOF) as  is seen  in gasoline vehicles.
One would merely have to design volume and catalyst loading for
the different  reaction  rates  and  reaction concentrations in the
exhaust.    The  concentration  of  SOF   in   diesel   exhaust  is
generally  less than the  HC concentration in gasoline exhaust,
and temperatures are generally  lower  in diesel  exhaust,  and so
a  larger  catalyst  may  be necessary in a diesel vehicle  for the
same percent conversion.  The concentration  of  oxygen in diesel
exhaust is  typically higher  than  in gasoline exhaust, however,
which may enhance  the  performance of the  catalyst  and reduce
the size  requirement  (or  at least eliminate the need for an air
pump).

     In summary, developing a "high efficiency"  system  capable
of oxidizing most  of  the SOF fraction of the particulate should
be feasible.   Reductions  of 80 percent  in SOF (similar to the
reduction of HC  in gasoline vehicles)  should be achievable with
proper design.  The use of  such a  catalyst  should also  increase
sulfate emissions.   An  increase   in sulfate of  100  percent is
believed  to be reasonable.   A  nominal  reduction  in RCP  of 10
percent should also be seen, as evidenced by some of the data.

     A low cost, low  efficiency  oxidation  catalyst should  also
be  feasible.   Using a  less active catalyst  formulation,   lower
particulate  reduction  efficiencies would be achieved.   Based on
confidential submittals,  a  reduction  in SOF  of 30 percent with
no  effect  on  the  RCP  or  sulfate portion  of  the particulate
would  be   expected.   Efficiency  estimates  for   of  the   flow
through catalyst technologies used in this study are  summarized
in Table  3-13.

     The   effect   of   oxidation   catalyst  use   on  gaseous
hydrocarbon emissions  is not  clear.   Test results published in
the  literature as  well  as confidential submittals  indicate  that
the  HC reduction efficiency  of all flow through catalysts are
on  the order  of  60 percent.  There  is  no  reason,  however, to
expect that the effect of  the  catalyst  on SOF emissions  should
be  significantly different than  the  effect of catalysts  on HC
emissions.   Both HC and  SOF  emissions  should be in the gaseous
phase  when  they  pass  over the catalyst,  and  should be  acted
upon by  the catalyst  in a  similar manner.   Therefore,  in this
analysis,  it will  be  assumed that the  HC reduction  efficiency
of   the   catalyst   will  be  the   same  as   the  SOF reduction
efficiency.  As  shown  in  Table  3-13,  HC efficiency is estimated
to  be 30 and 80  percent  for  the low  cost  and  high cost  flow
through catalysts,  respectively.

     CO   emission   reduction  characteristics  of  an  Englehard
flow  through   catalyst   have   also   been  reported  in   the

-------
                              3-40


                           Table 3-13


         Efficiency of Flow Through Oxidation Catalysts


                               Efficiency (Percent Reduction)
   Type           SOF        RCP        SO*         HC      CO

Low Cost-Low       30          0         0          30        o
  Efficiency
High Cost-High     80         10        -100        80      90
  Efficiency

-------
                              3-41
literature.[20]  Under high power  application,  the CO reduction
efficiency was  high,  approximately  90  percent.   Tests  results
on other, lower activity, catalyst  systems  showed no effect on
CO emissions.   For  this  analysis,  therefore,  it  was estimated
that the  low  cost  and high  cost  flow through  catalysts  would
have  CO   reduction   efficiencies  of   0   and   90   percent,
respectively.

     C.     Exhaust Aftertreatment Device Costs

     Information   on  the   costs   of   the  emission   control
technologies  considered  in this  analysis were taken from  the
March   1985   Regulatory   Impact   Analysis   accompanying   the
NOx/Particulate  Rulemaking,  and  a  1978  report prepared  under
contract for  EPA by Rath and Strong.[1,2]  Costs developed from
these studies were adjusted  to 1986 dollars  according to  the
Producers Price  Index for Internal Combustion Engines.

     1.     Trap-Oxidizer Systems

     a.     Uncatalyzed Ceramic Monolith

     The  1985  NOx/Particulate  RIA(1]  included cost estimates
for trapping  devices  and  regeneration  systems for  three classes
of heavy-duty vehicles,  roughly equivalent to the  three classes
of heavy-duty vehicles  used in this study.  Cost  estimates for
a  ceramic  monolith  trap,  with   a  bypass  burner regeneration
system,  were  generated  assuming trap volumes of 11,  21,  and 39
liters  for  LHDDE,  MHDDE,  and HHDDE  respectively.   Costs for the
trap and regeneration system are included  in Table  3-14.

     The   NOx/Particulate   RIA   also    quantified   increased
maintenance   costs   to  be  born  by the  purchaser  of  a  trap
equipped vehicle.   Those costs, discounted back to  the  year of
purchase were determined to  be $68 for  LHD,  $110 for MHD, and
$136 for HHD  diesels.

     In addition  to hardware  and  maintenance costs,   it  was
estimated  in  the  NOx/Particulate RIA that  each   of the seven
largest HDDS  manufacturers  would  spend about $2.8  million ($2.9
million  in   1986   dollars)    in  general  trap  research  and
development.   For  the  purpose of this  study,  it  was  assumed
that  approximately  one  half  of  this money  has  already been
spent,  and  that only half of  this  ($1.45  million)   remains to be
spent.   Also,  in the NOx/Particulate  RIA  it  was assumed  that
specific  system designs  would   cost  an  additional   $230,000
($235,000 in 1986 dollars)  in R&D  per engine family.

     The  1987  Model  Year  Certification Test Results  show  a
total  of 101 heavy-duty diesel  engines  families certified;   9
LHDDEs, 49   MHDDEs,  and  43  HHDDE s.   Of  the 49  MHDD  engine
families, it was projected  based on indications given by  engine
manufacturers that  two  families  would be used in urban  buses,

-------
                              3-42


                           Table 3-14


                       Trap System Costs


                          LHDDE        MHDDE        HHDDE

Uncatalvzed Ceramic Monolith
Trap Cost*
Regen System*
Maintenance
R & D
103
275
68
10
176
282
110
85
299
289
136
79
           Total           456          653          303
"High-Efficiency" Catalyzed Ceramic Trap

        Trap*                 103         176         299
        Regeneration*         275         282         289
        Catalyst*              56         ill         141
        Maintenance            68         110         136
        R & D	           10          85          79

           Total              512         764         944
 "Low Efficiency" Catalyzed Ceramic Trapp

        Trap*                  103         176         299
        Regeneration*          275         282         289
        Catalyst*               9          19          24
        Maintenance            68         110         136
        R & D	           10          85          79

           Total               465         672         827
 "Low Cost" Trap Cost Estimate

        Trap Cost*             75           108          143
        Regen  System*        154           170          209
        Maintenance            40            83          141
        R & D	           10            85           79

           Total             279           446          572
      Retail price equivalent (RPE)

-------
                              3-43


while the  remaining 47 families would be used  in non-urban bus
MHDDVs,   While  uncertainties  in  the  future  bus  market  render
this projection  to  be speculative  at best,  this  breakdown of
MHDDV  and  bus  engine  families  seems  to  be  a  reasonable
projection  for  the  purpose  of apportioning  R&D  costs  between
vehicle  types.   Accordingly,  in  this analysis,  general  trap
research  and  development  costs  were divided  among the  four
classes according to  the  relative number  of engine  families in
each.  Engine  family specific R&D  expenditures were determined
according to the number  of families in each  subclass projected
to  require  traps  under a  given fuel  scenario.   R&D costs were
assumed to be spent on a schedule over four years preceding the
production  of  the  trap   (two  years  assuming   introduction in
1991).    The schedule  for  R&D  expenditures,  analogous to  that
used in  the NOx/Particulate RIA  is  shown  in  Table 3-15.   These
research  and development  costs  were  then  discounted to  their
present valued in  the year of the initial trap application, and
the cost was  assumed to be  recovered by the manufacturer over
the first three model years of sales.

     Using  this  accounting procedure,  the  pec-vehicle increase
in  cost  due  to  research  and development  expenditures  can be
calculated.   For  example,  assuming  traps  are  introduced in
1994, and  that  50  percent of vehicles would require traps, the
per  vehicle equipped  R&D  costs will  be  $10,  $85,  and  $79 for
LHDDVs, MHDDVs, and HHDDVs vehicles,  respectively.   The  sum-of
the  maintenance cost, the trap cost,  and the R&D costs  for the
burner  regeneration  ceramic  monolith trap-oxidizer  system is
shown in Table 3-14.

     b.    Catalyzed Ceramic Monolith

     As described  previously,  it  was  assumed that two different
types   of   catalyzed  ceramic   monolith  systems    would  be
available.   One  system would  be  loaded with  a catalyst similar
to  that  of the  "high efficiency"  flow through  catalyst.   The
hardware  cost of  this device  was  assumed to  be equivalent to
the  cost  of a complete uncatalyzed  trap oxidizer system  with  a
burner regeneration system plus the catalytic metal  cost of the
high efficiency flow  through  catalyst system  described below.
Maintenance  and R&D  costs were assumed to be  the same as  that
of   an  uncatalyzed  trap.   Costs   for  the  "high   efficiency"
catalyzed  ceramic trap are also shown  in Table  3-14.

     The  cost of the "low efficiency" version of the catalyzed
ceramic  monolith  trap was determined  by the  same method  as the
high  efficiency  version,  catalytic  metal  process  were   taken
from the cost estimates  of  the  "low  efficiency"  flow through
catalyst  system described later  in this  chapter.   These  costs
are also  shown  in Table 3-14.

-------
                      3-44


                   Table 3-15


R&D Cost Expenditures for Aftertreatment Devices


              Introduction in 1991


    Year                  Percent of R&D Spent

    1987                            0
    1988                            0
   .1989                           88
    1990                           12


              Introduction in 1994


    Year                  Percent of R&D Spent

    1990                           19
    1991                           47
    1992                  .         30
    1993                            4

-------
                              3-45


     c.     Johnson Matthev Catalyzed Wire Mesh

     No attempt  to determine the  cost  of this device  was  made
for this study.  While  this  system remains a viable  option  for
particulate  control,  current unresolved  problems  related  to
increased  sulfate  formation make it  difficult  to  base   the
results  this   study   on  the   feasibility  of   this   device.
Furthermore, • due to the relatively low  efficiency of this  trap
on carbonaceous  particulate  (approximately one half  of  that  of
the ceramic monolith),  it  is doubtful that this system  will  be
able to  compete  economically with the  ceramic  monolith  even  if
sulfate production problems   are  resolved.   In any  event,  the
cost effectiveness of  this  device should not  be  substantially
greater than  that  of the  ceramic monolith under  any scenario.
Since  the   purpose of  this  study  is   to determine the  cost
savings associated with  fuel  control,   and  not necessarily  to
appraise  in  detail  every   possible  type  of  aftertreatment
device,  there should  be  little   error  involved  in using  the
ceramic monolith  as  the  primary candidate  for  a  particulate
trap.

     d.    Ceramic Fiber Coil

     As with the catalyzed wire mesh  system,  it is difficult to
determine  whether  the   ceramic   design will  be   available  for
particulate  control  in  1991 and 1994.   Durability  problems
still  need to be  resolved.   The rational for not  basing  the
results of  this  study on the feasibility  of  such a  device are
the same as in the case of the wire mesh trap.

     It may,  of  course, be the case  that  a cheaper  system such
as   the   ceramic   fiber   coil   system   described   in   the
NOx/Particulate  RIA will  be available.  To  examine  the impact
of such a  development,  a  sensitivity case was run  (Chapter  4)
using  a low  cost  trap  alternative.   For  simplicity,   it  was
assumed that  the emission reductions characteristic  of  the low
cost trap  were  the same  as the  uncatalyzed ceramic monolith,
but  that the cost  would be  similar  to that  developed  for  the
ceramic  fiber coil  in  the  NOx/Particulate RIA.  As explained
therein,   costs   for   the   regeneration   system   would   be
significantly   lower,   due   to   the   different   regeneration
technique  employed.  Since  no  trap bypass   would   be  needed,
there  would  also  be  a  credit  equivalent  to the  cost  of  a
standard  exhaust  pipe  realized,  as  well   as  a  decrease  in
maintenance  costs  due  to  the replacement  of  the standard steel
exhaust pipe with  a  stainless  steel pipe.   There would likely
be  an  increase  in maintenance  costs,  however, due  to the need
to  periodically replenish the  catalyst  supply..   The  R&D  cost
were * determined  in  the  same  way  as  those of  the  ceramic
monolith trap.   The  "low-cost"   trap oxidizer  system   consumer
cost estimate is shown  in  Table 3-14.

-------
                              3-46


     2.     Flow Through Oxidation Catalyst System

     The costs  of the  flow through oxidation  catalyst  systems
were calculated in accordance with the procedure  developed  in a
cost  estimation  report  prepared under  contract  for  EPA.[21]
The methodology was  altered  by using 1986  noble rcetal  prices
and  allowing  for  the  effects  of   inflation  on  other  costs
according to  the Producer  Price  Index.   The  markup  on  vendor
costs originally  used  in the referenced  report was also changed
to more accurately reflect industry practice.[22]

     The type,  size,  geometry, and  loading characteristics  of
the flow  through catalysts  which may be applied to heavy-duty
diesels  was  not  readily  available,   as  this  technology  (for
diesel  engines)  is  still  in  the  development  stage  and  most
information is  highly proprietary.    The  direct  application  of
gasoline   catalyst  technology  to   diesel  engines  is   not
necessarily   feasible,    due  to    differences   in   exhaust
composition,  exhaust  flow  rate,  temperature,   etc.  A catalyst
formulation   designed   to  minimize   sulfate   formation   while
satisfactorily  oxidizing organic  particulate is  required.   The
low temperature (relative to  gasoline)  of  the  exhaust  and the
high  volume of  exhaust  flow need to  be designed for as well.
Also of concern is the carbonaceous  particulate in the exhaust/-
which may reduce  the efficiency of the catalyst  by occupying
some of  the active sites.  To  summarize,  the  design  of  such- a
system requires a detailed evaluation of many parameters.

     While  all  of the inputs necessary to determine the cost of
a  flow-through  catalyst  system for  heavy-duty diesels are not
known,  conversations  with members of  industry and confidential
information received  made it possible to size  and cost  out two
systems  for heavy-duty  diesels.  Using  information received on
catalyst loading  and volume  on prototype systems, the necessary
catalyst   requirements  were   determined   for   each  subclass.
Catalyst  loadings (in gram/cubic foot)  were assumed  to  be the
same  for all  subclasses.   Catalyst  volumes for LHO, MHD, and
HHD were determined according  to relative  exhaust  flow  rates.
Using   this   information,   the    "high-efficiency"  oxidation
catalyst  system cost was calculated and is presented in Table
3-16.   As shown  in the  table,  the  flow-through  catalyst device
cost  alone  was  determined  to be  $94   for  LHDDVs,  $159  for
MHDDVs,  and $194 for HHDDVs.   In addition to the hardware cost,
to insure the proper  functioning of the catalyst  for the full
useful   life  of  the  vehicle,  stainless  steel  exhaust pipes
between  the  exhaust  manifold  and  catalyst will be  necessary.
The   cost   for   this  alteration,  including  a  credit for the
standard steel  exhaust  pipe  replaced,   as estimated   in  the
NOx/Particulate RIA  will be  $34 for  LHDDVs, $43  for MHDDVs, and
$73 for  HHDDVs. [1]  Chassis  heat  shields will  also  be required,
estimated to be  $11  per vehicle  equipped. [23]   The use  of this
type   of  control  technology  is   not   expected  to  increase
maintenance  costs   for   vehicle   owners.    Conversely,    the

-------
                              3-47


                           Table 3-16


                      Catalyst  System Costs


                            LHDDE        MHDDE        HHDDE
High Efficiency Catalyst

        Catalyst              94          159          194
        SS Exhaust            34           43           73
        Maintenance          -42          -52          -83
        R8.0                    5           43           39
        Heat Shields          11           ll           11

           Total             102          204          234



Low Efficiency Catalyst
        Catalyst              48           67           77
        SS Exhaust            34           43           73
        Maintenance          -42          -52          -83
        R&D                    5           43           39
        Heat Shields          ll           11           11

           Total              56          112          117

-------
                              3-48


improvements  in  exhaust  system  materials due  to  the use  of a
catalyst system will  actually decrease maintenance  costs.   The
maintenance savings from  the  elimination  of the need to replace
part of the exhaust  system when discounted back to  the  time of
purchase were determined  to be $42 for LHD, $52 for MHD and $83
for HHDCl].

     As  in  the  case  of   particulate  traps,  expenditures  for
Research and Development  will be recovered over the first three
years of sales after  the  technology is marketed.   RSO  expenses
for catalyst  systems  are  expected  to  be  approximately one-half
of  that  required  for  traps  (i.e.,  each  of  the  seven  largest
manufacturers would  spend a  total  of  $1.43 million  in  general
research  and  development,  only  half  of  which  remains to  be
spent, plus  $117,500  per  engine  family using  a catalyst.)   The
RS.D funds  are assumed to be spent according  to  the  schedule
shown  in Table  3-15.   Table  3-16 also  shows  an example  of
estimated R&D costs for each  vehicle class (generated  under the
assumption that 50 percent of engine families require catalysts
in 1994.)

     The  consumer  cost of the  low cost/  low efficiency  flow
through  catalyst  system  was  determined by  similar methodology.
and is also shown in Table 3-16.

     D.    Exhaust Aftertreatment Device Deterioration

     The  performance  deterioration   of   the   three  types  of
systems  which have  been   indicated  as primary  strategies,  the
burner  regeneration  uncatalyzed ceramic  monolith  system,  the
burner  regeneration  catalyzed ceramic monolith,  and the  flow
through oxidation catalyst, will be addressed in this section.

     1.    Uncatalyzed Ceramic Monolith Deterioration

     Durability testing of  a Corning noncatalyzed  trap  on a
1980 Mercedes  300SD  engine was  performed by SwRI for EPA.[8]
Emissions   tests   over  80,000   km  of   operation   showed  no
deterioration in particulate  collection efficiency of the trap,
with overall  particulate  collection efficiency  ranging  from 87
to  93  percent.  Based on these data,  it was  assumed  that no
deterioration  in  trap  collection efficiency  would occur  over
the full life of the vehicle.

     2.    Flow through Catalyst Deterioration

     The   loss   of   efficiency  of   automotive  catalysts  is
primarily due to the presence of lead  in  gasoline.   Since  lead
is  present  in  much  lower forms  in  diesel  fuel,  it  will be
assumed   that  catalyst   performance   will   not  deteriorate
significantly due to lead.

-------
                              3-49
     A study of the effect of fuel sulfur on the  operation  of  a
gasoline  automotive  catalyst  compared  hydrocarbon  conversion
efficiency over 15,000 miles of  operation on fuels with  sulfur
contents  of  0, 0.03,  and 0.09  weight  percent.  The  data  show
that when  fuel containing 0.03  and  0.09 weight  percent  sulfur
is  used,  HC oxidation  efficiency is reduced  to 80  percent  of
baseline  (zero-mile)  efficiency.  The  effect  appears to  take
place during, the  first 5,000  miles,  after which no  additional
deactivation due to sulfur appears to take  place.   Since  diesel
fuel sulfur  levels under consideration  in  this study are  0.05
weight  percent or greater,   it  is  expected  that  this  same
deactivation will  take place  in diesel flow  through oxidation
catalysts.   For  calculation  of  emissions  in  this   analysis,
therefore,  the   end-of-life  efficiency  of   a  flow  through
catalyst  device will  be  assumed  to  be  80  percent of baseline
(zero-mile) oxidation efficiency.

     3.     Catalyzed Ceramic Monolith System

     with  the  catalyzed  ceramic monolith system,   one  would
expect the  deterioration in efficiency  to  be  some combination
of  the  20  percent decrease in  efficiency of  the  flow-through
catalyst  and  the  zero   percent  efficiency  reduction  of  the
uncatalyzed ceramic trap.   The  efficiency reduction factor  of'
this  system type  was  therefore calculated  according to  the
relative  reductions in particulate via  trapping and  catalysis.
This  calculation  yields  an  average  end-of-life  particulate
reduction efficiency  of 91  percent of  the zero-mile  efficiency
for  the  "high  efficiency"  catalyzed  system  and 98  percent  of
the zero-mile efficiency for the low efficiency system.

III. Aftertreatment Technology Mix for Compliance with Standards

     A.    Methodoloo^r

     l.    Light-Duty Vehicles

     The  analysis  of  exhaust  aftertreatment  requirements  for
light-duty vehicles was not  necessary for  LDDV and LDDT1, since
no  aftertreatment  devices are  needed to comply with particulate
standards  under  present  conditions.    In  the  case  of  the
Mercedes-Benz  300D,  the  recent recall of  these systems  and
discontinuation of its sale in  the U.S.  make it  difficult  to
predict much about  future use  of the system.

     Rulemaking proceedings  currently in progress  on the LDDT2
class  indicate that a  0.13  gram per mile  particulate emission
level will  be  required for these vehicles beginning  in the 1991
mode! year.  It was assumed that this standard would be  met by
applying  ceramic  monolith traps  to  100  percent of the vehicles
beginning in 1991.   Because of the similarity  of these engines
to  those  in LHDDE class, the  same trap  costs as those developed
for  LHODE  vehicles were used.    No  research  and  development

-------
                              3-50


costs were included,  as  the same  amount  will be  spent  whether
fuel control  is implemented or not.

     The  emission  level  determined by  applying  100  percent
traps to  projected 1991  engine-out  levels was  defined  as  the
emission  target  level for achieving  the  0.13 gpm  standard  for
this vehicle class.  This target  level was  subsequently  used to
determine the  fraction of LDDT2 requiring traps under any given
fuel control scenario (only 99 percent traps  would  be needed in
1994, due to expected fuel economy improvements).

     2.     Heavy-Putv Vehicles

     The  lowest  cost scenario  for meeting particulate emission
standards was determined  for each heavy-duty  vehicle  class,  for
model years  1991  and 1994,  assuming no fuel control.   Using the
engine-out emission  distributions described previously for each
vehicle  class, a  compliance  strategy model  was used  to apply
the  various  types of aftertreatment  devices  described  in this
chapter  to  the highest  emitting  engines  in  each  vehicle class
until  the  end  of  life emission target  level  was  reached.
Aftertreatment  devices  were  applied to  the highest  emitting
engines according to the  criterion that the lowest  cost  per ton.
of emission  control  was  achieved.  Modeling was performed using
both the  "nominal-high  SOF"  and the  "nominal-low  SOF" emission
distributions  discussed  earlier  in this  chapter.   The model
results   for   both  nominal  emission  distributions  were  then
averaged  together.   These averaged  results  were  used  in  the
cost  effective analysis  of  this  report.   It  should be noted
that,  according  to  this methodology,  not  every vehicle  was
required  to  comply  with the  particulate standards.    Rather,
emission  averaging was allowed over the entire vehicle class.

     In  the  modelling it was  assumed that trap-oxidizers would
have  a  two  percent  fuel economy  penalty  associated  with their
use  (one percent due to  trap  backpressure, one percent  for the
regeneration system), as estimated in the Draft NOx/Particulate
RIA.[25]   While a lower  fuel  economy penalty (1.5 percent) was
used in the  NOx/Particulate  RIA, trap volumes  currently under
development   are  significantly   smaller   than  those   assumed
therein.[1]   Since the large trap volumes  in  the Final RIA used
as  a basis for adjusting the  estimated fuel  economy  penalty for
a burner regenerated ceramic monolith system downward from 2.0
to  1.5  percent  no  longer  seem  likely/  the 2.0  percent  fuel
economy  penalty used in  the Draft RIA will  be used  here.

     B.    Results

      l.    Light  Duty Vehicles

     As  described in Section A,  100  percent  of  the  LDDT2 fleet
will be  equipped with  traps  in  1991,  at a cost of  $455 per
vehicle.   Emission  factors  and  aftertreatment  costs  for  1991
light duty diesels are  shown  in Table 3-17.

-------
                                       Table 3-17
                 Light-Duty  Diesel  After-treatment Coats and Emissions
Vehicle
Class
Percent Aftertreatment
Fuel Econ Emissions (g/mi)
Traps Cost($/vehicle) (nog) SOP RCP SO* TPN
1991-1993
LDDV
LDDT1
LDOT2
1994 and
LDDV
LDDT1
LDDT2
0 0
0 0
100 455
Later
0 0
0 0
99 450
31.4 0.024 0.079 0.028 0.131
25.9 0.108 0.072 0.033 0.213
25.4 0.043 0.012 0.034 0.089
34.2 0.024 0.079 0.025 0.128
27.8 0.108 0.072 0.031 0.211
27.3 0.044 0.013 0.032 0.089
S02 HC*
0.520 0.29
0.631 0.43
0.637 0.15
0.477 0.29
0.588 0.43
0.594 0.15
Zero mile levels.

-------
                              3-52


     The 1994  and later MY  emission  factors  and aftertreatment
costs for light-duty diesels are also shown in Table  3-17.   AS
shown,  only  99 percent  of  LDDT2s  required traps  to meet  the
emission  standard.   This   is   lower  than  the   100  percent
utilization  in   1991,  due   to  projected  increases  in  fuel
economy.  Since  less  fuel is  consumed per  mile,   less  sulfate
will  be produced,  and  overall  engine-out emissions will  be
slightly lower.   The slight decrease  in  the  number  of  traps
required in 1994 is the result.

     2.    Heavy-DutYL Vehicles

     Results for  1991-93 heavy-duty vehicles  are shown in Table
3-18.    Aftertreatment   costs   (not   including   fuel  economy
penalties)  range  from  $157  per  vehicle for LHDDVs  to $205  for
HHDDVs  to  meet the  0.25  g/BHP-hr  standard.    Costs  for  urban
buses are significantly higher  ($656) due to the more stringent
particulate standards.   Due  to  the contribution of  sulfate  to
total  particulate,  it  was not  possible  for urban  bus  engines
which  convert  2 percent of  fuel sulfur to  sulfate particulate
(see Chapter 4  for  detailed  analysis) to meet the 0.10 g/BHP-hr
standard, even  with  100  percent traps.  Certifying  to the 0.10
g/BHP-hr standard therefore  will require  the  use of  an engine
which converts  less than  1.5 percent  of fuel sulfur to  sulfate
and  still  require  the  use  of  traps  costing $656  per vehicle.
The urban bus cost estimates in Table  3-18 reflect  the use of a
"low  sulfate  producing"  engine,  the  availability  of which  is
difficult to predict.

     Table 3-18 also shows the  modeling results  for  heavy-duty
vehicles  sold  in 1994  and   beyond.   Aftertreatment  costs  per
vehicle  range   from  $462   for LHDDVs   to  $702  for  HHDDVs  (not
including fuel  economy  penalties).  Because of the  low emission
standard  and   the  large  contribution  of   sulfate  to  total
particulate, it was not possible for LHDDVs,  MHDDVs,  and urban
buses  to  meet   the  standards,  even  with  100  percent  trap
application.    Certifying  to  the  0.10 standard  with  current
sulfur fuel will  require the use  of vehicles  which  convert less
than  2.0 percent of  fuel sulfur  to  sulfate;  as low as  a 1.5
percent  conversion would  be required  for LHDDVs.   Table 3-18
reflects the use  of these  "low  sulfate producing" engines.

-------
                     Table 3-18
Heavy-Duty Diesel Aftertreatment Usage and Emissions
Veh.
Class
1991-93
LHDDE
MHDDE
HHDDE
BUS
,1994 and
LHDDE
MHDDE
HHDDE
BUS
Uncatly.
Traps
0
0
11
0
Later
0
0
43
0
Low Coat
Catalysed
Trtpf
27
21
9
100
100
99
45
100
Low Cost Average
Flow Aftertreat-
Through ment Cost
Catalysts ($/Vehicle)
50 157
39 192
29 205
0 656
0 462
0 642
0 702
0 659
SOP
0.070
0.063
0.038
0.027
0.023
0.018
0.014
0.018
RCP
0.072
0.095
0.125
0.015
0.008
0.011
0.020
0.010
End of Life
Emissions (g/BHP-hr)
-S0.4
0.082
0.066
0.061
0.052
0.063
0.064
0.059
0.066
TPM
0.224
0.224
0.224
0.094
0.093
0.093
0.093
0.093
SQ2
1.16
0.94
0.87
1.00
1.14
0.93
0.84
0.98
HC
0.66
0.82
1.04
0.44
0.36
0.37
0.67
0.36
CO
4.0
4.4
4.9
4.6
3.6
3.6
3.6
3.6

-------
                              3-54


                     References (Chapter 3)


     l.     "Regulatory  Impact  Analysis,  Oxides   of   Nitrogen
Pollutant Specific Study  and  Summary and Analysis of Comments -
Control  of Air Pollution  from New Motor Vehicles and  New Motor
Vehicle   Engines:   Gaseous  Emission  Regulations  for   1987  and
Later Model .Year  Light-Duty  Vehicles,  and  for  1988  and Later
Model   Year   Light-Duty   Trucks   and   Heavy-Duty   Engines;
Particulate Emission Regulations  for  1988  and Later Model  Year
Heavy-Duty Diesel Engines,"  EPA, OAR, OMS,  March 1985.

     2.     "Determination of   a  Reliable  and Efficient  Diesel
Particulate Hydrocarbon  Extraction  Process,"  Shirish  A  Shimpi
and Ming Li Yu, SAE Paper 811183, 1981.

     3.     "An Improved  Method for Determining  the Hydrocarbon
Fraction of Diesel Particulates by  Vacuum  Oven Sublimation," R.
Halsall, M.L.  McMillan,   B.J.  Schwartz, SAE Paper  No.   872136,
1987.

     4.     "Measurement   of   Sulfate  and   Sulfur   Dioxide  in
Automotive  Exhaust,"  Melvin  N.  Ingalls, and  Karl  J.  Springer,.
Southwest Research  Institute, prepared  for  EPA,  OAWM,   OMSAPC,
ECTD, EPA-460/3-76-015, August 1976.

     5.     "Characterization    of   Particulate   and    Gaseous
Emissions From Two Diesel Automobiles as Functions  of Fuel and
Driving  Cycle,"  Charles  T.  Hare,  Southwest  Research  Institute,
Thomas  M.  Baines, U.S.  Environmental Protection  Agency, Paper
790424   present   at   SAE  congress  and  Exposition,   Detroit,
February 26  - March  2,  1979.  (Available  in Docket  * A-80-18,
II-I-18).

     6.     "Characterization  of Sulfates,  Odor,  Smoke,  POM and
Particulates  From Light  and  Heavy  Duty  Engines  -  Part   IX,"
Springer,  Karl J.,  prepared   for U.S.  Environmental  Protection
Agency  by  Southwest  Research  Institute,  EPA-460/3-79-007,   June
1979. (Available  in Docket A-80-18, ll-A-6).

     7.     "Characterization  of  Exhaust Emissions  From Diesel
Powered Passenger Cars With Particular Reference to Unregulated
Components,"   Lies,    K.H.,    Postulka,   A.,  and   Gring,   H.,
Vokswagenwerk A.G., SAE Paper  #840361.

     8.     "Diesel   Car   Particulate  Control   Methods,"   C.M.
Urban, L.C. Landman, R.D. Wagner, SAE Paper  No.  830084, 1983.

     9.     "Compilation   of   Air  Pollutant  Emission Factors,
Volume  II:  Mobile Sources,"  AP-42  4th Edition,  September 1985.

-------
                              3-55


     10.   "Effects  of   Fuel   Properties  and   Engine  Design
Features on  the  Performance of  a Light-Duty Diesel  Truck -  A
Cooperative  Study,"  E.G.  Barry,  J.c.  Axelrod,  L,J.  McCabe,  T.
Inoue, and N. Tsuboi, SAE Paper No. 861526, 1986.

     11.   "An Update  Assessment  of  the Feasibility  of  Trap
Oxidizers,"  Regulatory  Support   Document,   J.   Alson  and  R.
Wilcox,  U.S.- EPA,  OANR,  QMS,  ECTD,  SDSB,   June  1983,  Public
Docket No. A-82-32.

     12.   "Diesel  Particulate  Study,"   ECTD,  QMS,  OAR,  EPA,
November 1983.

     13.   "Experiences  in   the  Development  of  Ceramic  Fiber
Coil Particulate Traps,"  H.O.  Hardenberg, H.L.  Daudel, and H.J.
Erdmannsdorfer, SAE Paper No. 870015, 1987.

     14.   "Particulate Trap Regeneration Induced  by  Means  of
Oxidising   Agents   Injected  Into   the   Exhaust  Gas,"   H.O.
Hardenberg,  H.L.   Daudel,  H.J.  Erdmannsdorfer,   SAE  Paper  No.
870016, 1987.

     15.   "Heavy-Duty    Engine   Exhaust    Particulate    Trap
Evaluation,"  Charles M.  Urban,  Southwest Research  Institute,''
EPA 460/3-84-008, September, 1984.   '

     16.   "Emission Characterization of  a  2-Stroke  Heavy-Duty
Diesel  Coach Engine  and  Vehicle With and  Without a Particulate
Trap,"  Terry  L.  Ullman  and Charles T. Hare, Southwest Research
Institute, EPA 460/3-84-015,  prepared for U.S.  E.P.A., OMSAPC,
ECTD, March, 1985.

     17.   "Emissions   Performance   of   Two  Catalyzed  Trap
Oxidizers  on a  Bus  Engine,"  Terry  L.  Ullman,  SAE   Paper  No.
860132, 1986.

     18.   "Chemical   Analysis   and   Biological  Testing   of
Emissions  From a  Heavy Duty Diesel  Truck With and Without Two
Different  Particulate Traps," SAE  Paper No. 860014, 1986.

     19.   "Investigation  of  the CTO  Emission  Control  System
Applied to Heavy-Duty Diesel Engines Used in Underground Mining
Equipment,"  J.P.   Mogan,  E.D.  Dainty, H.C. Vergeer,  A.  Lawson,
K.C.  West away,  J.K.  Weglo,  and  L.R.  Thomas,   SAE   Paper  No.
850151, 1985.

     20.   "Diesel Particulate SOF  Emission  Reduction Using  an
Exhaust Catalyst," G.E.  Andrews,  I.E. Iheozor-Ejiofor, and S.w.
Pang/  SAE  Paper No.  870251,  1987.

-------
                              3-56


     21.   "Cost  Estimations   foe   Emission  Control   Related
Components/Systems and  Cost  Methodology Description,"  LeRoy H.
Lindgren, Rath  & Strong,  Inc.,  March  1978,  EPA-460/3-78-002.
(Available in Docket tt A-80-18, II-A-18).

     22.   "Update  of  EPA's  Motor  Vehicle  Emission  Control
Equipment Retail  Price Equivalent  (RPE)  Calculation Formula,"
prepared for. U.S.  EPA by Jack  Faucett Associates,  September 4,
1985.

     23.   "Regulatory  Analysis  and  Environmental  Impact  of
Final Emission  Regulations  for 1984 and Later Model Year Heavy
Duty Engines,"  U.S.  EPA,  OMSAPC, December 1979.  (Available in
Docket H A-80-18, II-A-8).

     24.   "Deactivation   of   Three-way   Catalysts   by  Fuel
Contaminants:—Lead,  Phosphorus and  Sulfur," W.B.  Williamson,
H.S.  Gandhi,  M.E.   Heyde,   and  G.A.   Zawacki,   SAE Paper  (To.
790942,  1979.

     25.   "Draft   Regulatory   Impact  Analysis,    Oxides   of
Nitrogen Pollutant  Specific  Study  and Summary  and Analysis of
Comments - Control of Air Pollution from New Motor Vehicles and
New  Motor  Vehicle  Engines:   Gaseous  Emission  Regulations for'
1987 and Later Model Year Light-Duty Vehicles, and for 1988 and
Later  Model  Year  Light-Duty  Trucks   and Heavy-Duty  Engines;
Particulate  Emission  Regulations  for  1988  and Later Model  Year
Heavy-Duty Diesel Engines," EPA, OAR, QMS.

-------
              Appendix 3A
        Projected 1991 and 1994
HDDE Particulate Emission Distributions

-------
              1991  LHP  NOMINAL SCENARIO
Percent   Total
  of    Particulate     SOF
Engines (q/BHP-hr)  (q/BHP-hr)

0.5800
2.1000
5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
5.8800
2.1000
0.5800

0.1945
0.2066
0.2198
0.2343
0.2500
0.2673
0.2860
0.3065
0.3289
0.3533
0.3800
LOW SOF CA,
0.0461
0.0511
0.0565
0.0624
0.0689
0.0760
0.0837
0.0921
0.1012
0.1112
0.1222
Sulfates and
 bound water
 (q/BHP-hr)
                                  0.0820
                                  0.0820
                                  0.0820
                                  0.0820
                                  0.0820
                                  0.0820
                                  0.0820
                                  0.0820
                                  0.0820
                                  0.0820
                                  0.0820
 Residual
(q/BHP-hr)
                0.0664
                0.0735
                0.0813
                0.0898
                0.0991
                0.1093
                0.1204
                0.1325
                0.1457
                0.1601
                0.1758
                    HIGH SOF CASE
0.5800
2.1000
5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
5.8800
2.1000
0.5800
0.1945
0.2066
0.2198
0.2343
0.2500
0.2673
0.2860
0.3065
0.3289
0.3533
0.3800
0.0686
0.0760
0.0841
0,0929
0.1025
0.1130
0.1245
0.1370
0.1506
0.1655
0.1818
0.0820
0.0820
0.0820
0.0820
0.0820
0.0820
0.0820
0.0820
0.0820
0.0820
0.0820
0.0439
0.0486
0.0538
0.0594
0.0655
0.0722
0.0796
0.0876
0.0963
0.1058
0.1162

-------
              1991 MHD NOMINAL SCENARIO
 Percent   Total
   of    Particulate     SOF
 Engines (q/BHP-hr)   (q/BHP-hr)
                     Sulfates and
                      bound water
                      (q/BHP-hr)
 0.5800
 2.1000
 5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
 5.8800
 2.1000
 0.5800
0
0
  1764
  1892
0.2034
0.2189
0.2359
0.2547
0.2752
0.2979
0.3227
0.3500
0.3800
LOW SOF CASE
0.0342
0.0381
0.0425
0.0473
0.0526
0.0584
0.0648
0.0718
0.0795
0.0880
0.0973

0.0662
0.0662
0.0662
0.0662
0.0662
0.0662
0,0662
0.0662
0.0662
0.0662
0.0662
                                    Residual
                                   (q/BHP-hr)
0.0760
0.0849
0.0946
0.1053
0.1171
0.1300
0.1442
0.1598
0.1770
0.1958
0.2165
                     HIGH SOF CASE
0.5800
2.1000
5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
5.8800
2.1000
0.5800
0.1764
0.1892
0.2034
0.2189
0.2359
0.2547
0.2752
0.2979
0.3227
0.3500
0.3800
0.0562
0.0627
0.0699
0.0779
0.0866
0.0961
0.1066
0.1181
0.1308
0.1447
0.1600
0.0662
0.0662
0.0662
0.0662
0.0662
0.0662
0.0662
0.0662
0.0662
0.0662
0.0662
0.0540
0.0603
0.0672
0.0748
0.0832
0.0923
0.1024
0.1135
0.1257
0.1391
0.1538

-------
              1991 HMD NOMINAL SCENARIO
 Percent   Total
   of    Particulate     SOF
 Engines (g/BHP-hr)   (g/BHP-hr)
 0.5800
 2.1000
 5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
 5.8800
 2.1000
 0.5800

0.1710
0.1840
0.1984
0.2142
0.2317
0.2509
0.2720
0.2952
0.3208
0.3490
0.3800
LOW SOF CA
0.0153
0.0172
0.0192
0.0214
0.0238
0.0265
0.0295
0.0327
0.0363
0.0402
0.0446
Sulfates and
 bound water
 (q/BHP-hr)
   0.0615
   0.0615
   0.0615
   0.0615
   0.0615
   0.0615
   0.0615
   0.0615
   0.0615
   0.0615
   0.0615
 Residual
(q/BHP-hr)
  0.0942
  0.1054
    1177
    1314
    1463
    1628
  0.1810
  0.2010
    2230
    2472
0,
0
0
0
0
0
  0.2739
                     HIGH SOF CASE
0.5800
2.1000
5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
5.8800
2.1000
0.5800
0.1710
0.1840
0.1984
0.2142
0.2317
0.2509
0.2720
0.2952
0.3208
0.3490
0.3800
0.0372
0.0417
0.0466
0.0519
0.0579
0.0644
0.0716
0.0795
0.0882
0.0977
0.1083
0.0615
0.0615
0.0615
0.0615
0.0615
0.0615
0.0615
0.0615
0.0615
0.0615
0.0615
0.0723
0.0809
0.0904
0.1008
0.1123
0.1250
0.1389
0.1543
0.1711
0.1897
0.2102

-------
              1991 BUS NOMINAL SCENARIO
Of
Engines

0.5800
2.1000
5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
5.8800
2.1000
0.5800
Particulate
< q/BHP-hr)

0.2136
0.2215
0.2298
0.2385
0.2476
0.2572
0.2672
0.2777
0.2887
0.3002
0.3123
SOF
(q/BHP-hr)
LOW SOF CASE
0.0447
0.0471
0.0497
0.0524
0.0552
0.0582
0.0613
0.0646
0.0680
0.0715
0.0753
bound water
(q/BHP-hr)

0.0694
0.0694
0.0694
0.0694
0.0694
0.0694
0.0694
0.0694
0.0694
0.0694
0.0694
                                                0.0995
                                                0.1049
                                                0.1107
                                                0.1167
                                                0.1230
                                                  1295
                                                  1365
                                                0.1437
                                                0.1513
                                                0.1593
                                                0.1676
                                     0,
                                     0
                    HIGH SOF CASE
 0.5800
 2.1000
 5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
 5.8800
 2.1000
 0.5800
0.2136
0.2215
0.2298
0.2385
0.2476
0.2572
0.2672
0.2777
0.2887
0.3002
0.3123
0.0735
0.0776
0.0818
0.0862
0.0909
0.0958
0.1009
0.1062
0.1118
0.1177
0.1239
0.0694
0.0694
0.0694
0.0694
0.0694
0.0694
0.0694
0.0694
0.0694
0.0694
0.0694
0.0706
0.0745
0.0786
0.0829
0.0873
0.0920
0.0969
0.1021
  1074
  1131
0
0,
0.1190

-------
           1991  LHP  LOW EMISSIONS SCENARIO
Percent   Total
  of    Particulate     SOF
Engines (q/BHP-hr)  (g/BHP-hr)

0.5800
2.1000
5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
5.8800
2.1000
0.5800

0.1521
0.1604
0. 1694
0.1792
0.1899
0.2017
0.2144
0.2284
0.2436
0.2602
0.2784
LOW SOF CA,
0.0288
0.0321
0.0358
0.0399
0.0443
0.0491
0.0543
0.0600
0.0663
0.0731
0.0805
Sulfates and
 bound water
 (g/BHP-hr)
                                  0.0820
                                  0.0820
                                  0.0820
                                  0.0820
                                  0.0820
                                  0.0820
                                  0.0820
                                  0.0820
                                  0.0820
                                  0.0820
                                  0.0820
 Residual
(a/BHP-hr)
                0.0414
                0.0462
                0.0516
                0.0574
                0.0637
                0.0706
                0.0781
                0.0864
                0.0953
                0.1052
                0.1159
                    HIGH SOF CASE
0.5800
2.1000
5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
5.8800
2.1000
0.5800
0.1521
0.1604
0.1694
0.1792
0.1899
0.2017
0.2144
0.2284
0.2436
0.2602
0.2784
0.0428
0.0478
0.0533
0.0593
0.0658
0.0730
0.0808
0.0893
0.0986
0.1087
0.1198
0.0820
0.0820
0.0820
0.0820
0.0820
0.0820
0.0820
0.0820
0.0820
0.0820
0.0820
0.0274
0.0306
0.0341
0.0379
0.0421
0.0467
0.0516
0.0571
0.0630
0.0695
0.0766

-------
           1991 MHD LOW EMISSIONS SCENARIO
 Percent   Total
   of    Particulate     SOF
 Engines (q/BHP-hr)  (g/BHP-hr)
                     Sulfates and
                      bound water
                      (g/BHP-hr)
 0.5800
 2.1000
 5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
 5.8800
 2.1000
 0.5800
  1433
  1529
  1634
0.1750
0.1877
  2017
  2170
0.2339
0.2524
0.2728
0.2951
0,
0,
0
0
0
LOW SOF CASE
0.0239
0.0269
0.0301
0.0337
0.0377
0.0420
0.0467
0.0520
0.0577
0.0640
0.0710

0.0662
0.0662
0.0662
0.0662
0.0662
0.0662
0.0662
0.0662
0.0662
0.0662
0.0662
                                    Residual
                                   (3/BHP-hr)
0.0532
0.0598
0.0671
0.0750
0.0838
0.0935
0.1041
0.1157
0.1285
0.1425
0.1580
                     HIGH SOF CASE
0.5800
2.1000
5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
5.8800
2.1000
0.5800
0.1433
0.1529
0.1634
0.1750
0.1877
0.2017
0.2170
0.2339
0.2524
0.2728
0.2951
0.0393
0.0442
0.0496
0.0555
0.0620
0.0691
0.0769
0.0855
0.0950
0.1053
0.1168
0.0662
0.0662
0.0662
0.0662
0.0662
0.0662
0.0662
0.0662
0.0662
0.0662
0.0662
0.0378
0.0425
0.0476
0.0533
0.0595
0.0664
0.0739
0.0822
0.0912
0.1012
0.1122

-------
           1991 HHD LOW EMISSIONS  SCEMARIQ
Percent   Total
  of    Particulate     SOF
Engines (q/BHP-hr)  (q/BHP-hr)
Sulfates and
 bound water
 (q/BHP-hr)
LOW SOF CASE
0.5800
2.1000
5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
5.8800
2.1000
0.5800
0.1405
0.1505
0.1615
0.1736
0.1870
0.2017
0.2178
0.2356
0.2552
0.2767
0.3004
0.0111
0.0125
0.0140
0.0157
0.0176
0.0196
0.0219
0.0244
0.0271
0.0301
0.0335
0.0615
0.0615
0.0615
0.0615
0.0615
0.0615
0.0615
0.0615
0.0615
0.0615
0.0615
 Residual
(q/BHP-hr)
                                               0.0680
                                               0.0766
                                               0.0860
                                               0.0964
                                               0.1079
                                               0.1205
                                               0.1344
                                               0.1497
                                               0.1666
                                               0.1851
                                               0.2055
                    HIGH SOF CASE
0.5800
2.1000
5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
5.8800
2.1000
0.5800
0.1405
0.1505
0.1615
0.1736
0.1870
0.2017
0.2178
0.2356
0.2552
0.2767
0.3004
0.0269
0.0303
0.0340
0.0381
0.0427
0.0477
0.0531
0.0592
0.0658
0.0732
0.0812
0.0615
0.0615
0.0615
0.0615
0.0615
0.0615
0.0615
0.0615
0.0615
0.0615
0.0615
                                                0.0522
                                                0.0588
                                                0.0660
                                                0.0740
                                                0.0828
                                                0.0925
                                                0.1032
                                                  1149
                                                  1278
                                                0.1420
                                                0.1577
                0
                0

-------
           1991 BUS LOW EMISSIONS SCENARIO
 Percent   Total                 Sulfates  and
   of     Particulate     SOF     bound water
 Engines (g/BHP-hr)   (g/BHP-hr)   (g/BHP-hr)
LOW SOF CASE
0.5800
2.1000
5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
5.8800
2.1000
0.5800
0.1697
0.1755
0.1816
0.1880
0.1946
0.2017
0.2090
0.2167
0.2248
0.2332
0.2421
0.0311
0.0329
0.0348
0.0368
0.0388
0.0410
0.0433
0.0457
0.0482
0.0508
0.0535
0.0694
0.0694
0.0694
0.0694
0.0694
0.0694
0.0694
0.0694
0.0694
0.0694
0.0694
                                    Residual
                                   (q/BHP-hr)
                                                0.0692
                                                0.0732
                                                0.0774
                                                0.0818
                                                0.0864
                                                0.0913
                                                0.0963
                                                0.1016
                                                0.1072
                                                0.1130
                                                0.1191
                    HIGH SOF CASE
 0.5800
 2.1000
 5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
 5.8800
 2.1000
 0.5800
0.1697
0.1755
0.1816
0.1880
0.1946
0.2017
0.2090
0.2167
0.2248
0.2332
0.2421
0.0512
0.0541
0.0572
0.0605
0.0639
0.0674
0.0712
0.0751
0.0792
0.0835
0.0881
0.0694
0.0694
0.0694
0.0694
0.0694
0.0694
0.0694
0.0694
0.0694
0.0694
0.0694
0
0
0.0492
0.0520
  0550
  0581
0.0614
0.0648
0.0684
0.0722
0.0761
0.0803
0.0846

-------
              1994 LHP NOMINAL SCENARIO
 Percent   Total
   of    Particulate     SOF
 Engines (g/BHP-hr)  (g/BHP-hr)
 0.5800
 2.1000
 5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
 5.8800
 2.1000
 0.5800
Sulfates and
 bound water
 (g/BHP-hr)
LOW SOF CASE
0.1519
0.1598
0.1684
0.1778
0.1880
0.1992
0.2114
0.2247
0.2393
0.2552
0.2725
0.0300
0.0332
0.0367
0.0406
0.0448
0.0494
0.0544
0.0598
0.0658
0.0723
0.0794
0.0788
0.0788
0.0788
0.0788
0.0788
0.0788
0.0788
0.0788
0.0788
0.0788
0.0788
 Residual
(g/BHP-hr)
                0
                0
                0
    0431
    0478
    0529
                0.0584
                0.0644
                0.0710
                0.0782
                0.0861
                0.0947
                0.1041
                0.1143
                     HIGH  SOF CASE
0.5800
2.1000
5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
5.8800
2.1000
0.5800
0.1519
0.1598
0.1684
0.1778
0.1880
0.1992
0.2114
0.2247
0.2393
0.2552
0.2725
0.0446
0.0494
0.0547
0.0604
0.0666
0.0735
0.0809
0.0890
0.0979
0.1076
0.1182
0.0788
0.0788
0.0788
0.0788
0.0788
0.0788
0.0788
0.0788
0.0788
0.0788
0.0788
0.0285
0.0316
0.0349
0.0386
0.0426
0.0470
0.0517
0.0569
0.0626
0.0688
0.0755

-------
              1994  MHD NOMINAL  SCENARIO
Percent   Total
  of    Particulate     SOF
Engines (g/BHP-hr)  (g/BHP-hr)
Sulfates and
 bound water
 (q/BHP-hr)
LOW SOF CASE
0.5800
2.1030
5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
5.8800
2.1000
0.5800
0.1362
0.1446
0.1537
0.1638
0.1749
0.1871
0.2005
0.2152
0.2313
0.2491
0.2686
0.0222
0.0248
0.0276
0.0308
0.0342
0.0380
0.0421
0.0467
0.0517
0.0572
0.0632
0.0646
0.0646
0.0646
0.0646
0.0646
0.0646
0.0646
0.0646
0.0646
0.0646
0.0646
 Residual
(q/3HP-hr)
                                               0.0494
                                               0.0552
                                               0.0615
                                               0.0685
                                               0.0761
                                               0.0845
                                               0.0938
                                               0.1039
                                               0.1150
                                               0.1273
                                               0.1407
                    HIGH SOF CASE
0.5800
2.1000
5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
5.8800
2.1000
0.5800
0.1362
0.1446
0.1537
0.1638
0.1749
0.1871
0.2005
0.2152
0.2313
0.2491
0.2686
0.0365
0.0408
0.0455
0.0506
0.0563
0.0625
0.0693
0.0768
0.0850
0.0941
0.1040
0.0646
0.0646
0.0646
0.0646
0.0646
0.0646
0.0646
0.0646
0.0646
0.0646
0.0646
0.0351
0.0392
0.0437
0.0486
0.0541
0.0600
0.0666
0.0738
0.0817
0.0904
0.0999

-------
               1994 HMD NOMINAL SCENARIO
 Percent   Total
   of    Particulate     SOF
 Engines (q/BHP-hr)  (q/BHP-hr)
 0.5800
 2.1000
 5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
 5.8800
 2.1000
 0.5800
Sulfates and
 bound water
 (q/BHP-hr)
LOW SOF CASE
0.1295
0.1380
0.1473
0.1576
0.1689
0.1814
0.1951
0.2102
0.2269
0.2452
0.2653
0.0100
0.0112
0.0125
0.0139
0.0155
0.0172
0.0192
0.0213
0.0236
0.0262
0.0290
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
 Residual
(q/BHP-hr)
                0.0612
                0.0685
                0.0765
                0.0854
                0.0951
                0.1058
                0.1177
                0.1306
                0.1450
                0.1607
                0.1780
                    HIGH SOF CASE
0.5800
2.1000
5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
5.8800
2.1000
0.5800
0.1295
0.1380
0.1473
0.1576
0.1689
0.1814
0.1951
0.2102
0.2269
0.2452
0.2653
0.0242
0.0271
0.0303
0.0338
0.0376
0.0418
0.0465
0.0517
0.0573
0.0635
0.0704
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
0.0470
0.0526
0.0587
0.0655
0.0730
0.0812
0.0903
0. 1003
0.1112
0.1233
0.1366

-------
              1994  BUS  NOMINAL SCENARIO
Percent   Total
  of    Particulate     SOF
Engines (q/BHP-hr)  (q/BHP-hr)
Sulfates and
 bound water
 (q/BHP-hr)
LOW SOF CASE
0.5800
2.1000
5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
5.8800
2.1000
0.5800
0.1615
0.1667
0.1721
0.1777
0.1836
0.1898
0.1963
0.2032
0.2103
0.2178
0.2257
0.0291
0.0306
0.0323
0.0341
0.0359
0.0378
0.0399
0.0420
0.0442
0.0465
0.0489
0.0678
0.0678
0.0678
0.0678
0.0678
0.0678
0.0678
0.0678
0.0678
0.0678
0.0678
                                               Residual
                                              (q/BHP-hr)
                                               0.0647
                                               0.0682
                                               0.0719
                                               0.0758
                                               0.0799
                                               0.0842
                                               0.0887
                                               0.0934
                                               0.0983
                                               0.1035
                                               0.1089
                    HIGH SOF CASE
 0.5800
 2.1000
 5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
 5.8800
 2.1000
 0.5800
0.1615
0.1667
0.1721
0.1777
0.1836
0.1898
0.1963
0.2032
0.2103
0.2178
0.2257
0.0478
0.0504
0.0532
0.0561
0.0591
0.0622
0.0656
0.0690
0.0727
0.0765
0.0805
0.0678
0.0678
0.0678
0.0678
0.0678
0.0678
0.0678
0.0678
0.0678
0.0678
0.0678
                                               0.0459
                                               0.0484
                                               0.0511
                                               0.0539
                                               0.0568
                                               0.0598
                                               0.0630
                                                  0663
                                                  0698
                                                  0735
                0
                0
                0
                                                0.0774

-------
           1994 LHD LOW EMISSIONS SCENARIO
 Percent   Total
   of    Particulate     SOF
 Engines (g/BHP-hr)  (g/BHP-hr)
                     Sulfates and
                      bound water
                      (g/BHP-hr)
 0.5800
 2.1000
 5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
 5.8800
 2.1000
 0.5800
0.1244
0.1298
0.1356
0.1420
0.1490
0.1566
0.1649
0.1739
0.1838
0.1946
0.2064
LOW SOF CASE
0.0187
0.0209
0.0233
0.0259
0.0288
0.0319
0.0353
0.0390
0.0431
0.0475
0.0523

0.0788
0.0788
0.0788
0.0788
0.0788
0.0788
0.0788
0.0788
0.0788
0.0788
0.0788
 Residual
(g/BHP-hr)
  0.0269
  0.0301
  0.0335
  0.0373
  0.0414
  0.0459
  0.0508
  0.0561
  0.0620
  0.0683
  0.0753
                     HIGH SOF CASE
0.5800
2.1000
5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
5.8800
2.1000
0.5800
0.1244
0.1298
0.1356
0.1420
0.1490
0.1566
0.1649
0.1739
0.1838
0.1946
0.2064
0.0278
0.0311
0.0347
0.0385
0.0428
0.0474
0.0525
0.0580
0.0641
0.0707
0.0779
0.0788
0.0788
0.0788
0.0788
0.0788
0.0788
0.0788
0.0788
0.0788
0.0788
0.0788
0.0178
0.0199
0.0222
0.0246
0.0274
0.0303
0.0336
0.0371
0.0410
0.0452
0.0498

-------
           1994 MHD  LOW EMISSIONS  SCENARIO
Percent   Total
  of    Particulate     SOF
Engines (q/BHP-hr)  (q/BHP-hr)
Sulfates and
 bound water
 (q/BHP-hr)
LOW SOF CASE
0.5800
2.1000
5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
5.8800
2.1000
0.5800
0.1147
0.1209
0.1278
0.1353
0.1436
0.1526
0.1626
0.1736
0.1856
0.1989
0.2134
0.0155
0.0175
0.0196
0.0219
0.0245
0.0273
0.0304
0.0338
0.0375
0.0416
0.0461
0.0646
0.0646
0.0646
0.0646
0.0646
0.0646
0.0646
0.0646
0.0646
0.0646
0.0646
 Residual
(j/3HP-hr)
                                               0.0346
                                               0.0389
                                               0 . 0436
                                               0.0488
                                               0.0545
                                               0.0608
                                               0.0676
                                               0.0752
                                               0.0835
                                               0.0926
                                               0.1027
                    HIGH SOF CASE
0.5800
2.1000
5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
5.8800
2.1000
0.5800
0.1147
0.1209
0.1278
0.1353
0.1436
0.1526
0.1626
0.1736
0.1856
0.1989
0.2134
0.0256
0.0287
0.0322
0.0361
0.0403
0.0449
0.0500
0.0556
0.0617
0.0685
0.0759
0.0646
0.0646
0.0646
0.0646
0.0646
0.0646
0.0646
0.0646
0.0646
0.0646
0.0646
0.0245
0.0276
0.0310
0.0346
0.0387
0.0431
0.0480
0.0534
0.0593
0.0658
0.0729

-------
           1994 HHD LOW EMISSIONS SCENARIO
 Percent    Total
   of     Particulate     SOF
 Engines  (g/BHP-hr)   (g/BHP-hr)
                     Sulfates and
                      bound water
                      (g/BHP-hr)
LOW SOF CASE
0.5800
2.1000
5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
5.8800
2.1000
0.5800
0.1097
0.1162
0.1233
0.1312
0.1399
0.1494
0.1599
0.1715
0.1842
0.1982
0.2136
0.0072
0.0081
0.0091
0.0102
0.0114
0.0128
0.0142
0.0158
0.0176
0.0196
0.0217
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
                        Residual
                       (q/BHP-hr)
                                                0.0442
                                                0.0498
                                                0.0559
                                                0.0627
                                                0.0701
                                                0.0783
                                                0.0874
                                                0.0973
                                                0.1083
                                                0.1203
                                                0.1336
                    HIGH SQF CASE
 0.5800
 2.1000
 5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
 5.8800
 2.1000
 0.5800
0.1097
0.1162
0.1233
0.1312
0.1399
0.1494
0.1599
0.1715
0.1842
0.1982
0.2136
0.0175
0.0197
0.0221
0.0248
0.0277
0.0310
0.0345
0.0385
0.0428
0.0476
0.0528
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
0.0583
0.0339
0.0382
0.0429
0.0481
0.0538
0.0601
0.0671
  0747
  0831
  ,0923
0
0
0
0.1025

-------
           1994 BUS LOW EMISSIONS SCENARIO
Percent   Total
  of    Particulate     SOF
Engines (g/BHP-hr)  (g/BHP-hr)
Sulfates and
 bound water
 (g/BHP-hr)

0.
2.
5.
11.
18.
21.
18.
11.
5.
2.
0.

5800
1000
8800
9200
7500
2800
7500
9200
8800
1000
5800

0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.

1330
1368
1407
1449
1492
1538
1585
1635
1688
1743
1800
LOW
0
0
0
0
0
0
0
0
0
0
0
SOF CASE
.0202
.0214
.0226
.0239
.0252
.0266
.0281
.0297
.0313
.0330
.0348
0
0
0
0
0
0
0
0
0
0
0
.0678
.0678
.0678
.0678
.0678
.0678
.0678
.0678
.0678
.0678
.0678
 Residual
(g/BHP-hr)
                                               0.0450
                                               0.0476
                                               0.0503
                                               0.0532
                                               0.0562
                                               0.0593
                                               0.0626
                                               0.0661
                                               0.0697
                                               0.0735
                                               0.0774
                    HIGH SOF CASE
0.5800
2.1000
5.8800
11.9200
18.7500
21.2800
18.7500
11.9200
5.8800
2.1000
0.5800
0.1330
0.1368
0.1407
0.1449
0.1492
0.1538
0.1585
0.1635
0.1688
0.1743
0.1800
0.0333
0.0352
0.0372
0.0393
0.0415
0.0438
0.0463
0.0488
0.0515
0.0543
0.0572
0.0678
0.0678
0.0678
0.0678
0.0678
0.0678
0.0678
0.0678
0.0678
0.0678
0.0678
                                               0.0320
                                               0.0338
                                               0.0357
                                               0.0378
                                               0.0399
                                               0.0421
                                                 0445
                                                 0469
                                                 0495
                                                 ,0522
                0,
                0,
                0
                0
                                               0.0550

-------

-------
                         Chapter 4

    Effect of Fuel Quality on Emissions and Engine Cost

     One of the major purposes of this study  is  to  estimate
the impact a. modification in diesel fuel quality will  have
on  engine  technology  requirements.   In this  chapter,   che
engine-out  emission  estimates,   aftertreatment   technology
costs, and aftertreatment  device efficiencies developed  in
Chapter  3  will be  used to  estimate  the cost of  complying
with  the  particulate standards under  various fuel  control
scenarios.   The difference  between compliance  costs  with
baseline fuel quality  (Chapter 3)  and the compliance costs
with  fuel  quality  control developed here will then  be  used
in Chapter 7 to evaluate the cost effectiveness  of  the  fuel
quality regulation.

     The  two  primary  fuel  control  scenarios   for which
exhaust  aftertreatment technologies mixes  were  determined
are:   a sulfur  reduction  to 0.05  weight percent,  and  a
sulfur reduction to 0.05 weight percent  accompanied by  an
aromatics  reduction  from the current  average level  of  34.2
to  20 volume percent.  Hereafter,  these  scenarios  will  be
referred to  simply  as  sulfur control and aromatics control,
respectively.   Other  degrees  of   fuel  control   will  be
discussed in section IV of this chapter.

     The first  subject to  be discussed in this chapter  will
be the effect of fuel  sulfur  and aromatics on the emissions
of  diesel  vehicles.    Second will  be a presentation of the
aftertreatment  requirements  and costs foe compliance under
the two  fuel control scenarios.   This will be followed by a
sensitivity   analysis   of   the  aftertreatment  requirement
estimates  to  some  of  the key  assumptions  that were  made
about    emissions,     aftertreatment     efficiency,    and
aftertreatment cost  in Chapter 3.

I.    Effect  of Fuel Properties on Diesel Emissions

      Over  the  past  few  years  a  number  of studies  have
investigated the  effects  of  diesel  fuel  parameters  on
diesel   emissions,   both  for   heavy-duty   and   light-duty
vehicles.  On the  heavy-duty side,  some of the  most recent
investigation into  the  effects  of  fuel quality  on diesel
emissions  is being  performed under  contract by  Southwest
Research  Institute   under   contract  with  Coordinating
Research Council.[1]  Emission  testing is  being performed
on  various heavy-duty engines (including a Cummins  NTCC400,
and  a  DDA  Series  60)  on  fuels  of  varying  volatility,
aromaticity,  and  sulfur content.  A number of other studies
of  this  nature have  also  been  performed   and published.
Also,  many   engine  manufacturers  have  supplied  EPA  with

-------
                             4-2
emission test data, which,  although confidential  in nature,
have been helpful in confirming the  effect  of  changing fuel
quality on  emissions.   The effect  of  diesel  fuel  qualities
on light-duty .diesel emissions has  also been the  subject of
investigation.

     The  collective  knowledge gained  from these  published
studies  was. used  in  predicting   the   effect    of   fuel
modifications   on   emissions  of   both   heavy-duty   and
light-duty  diesel vehicles.   The  confidential  submictals
were  only used  to improve the  choice of  a  best  estimate
from within a range of data where the  scatter  was  such chat
a simple averaging of the data was not appropriate.

     A.    Effect of Fuel Sulfur on Highway Diesel Emissions

     1.    Heavy-Duty Diesel Engines

     The  relationship  between  fuel sulfur and  particulate
and  sulfurous emissions has  been  the subject of  study and
testing   for   the   past  several   years.    It  has   been
demonstrated  that  fuel  sulfur contributes  to  the emissions
of diesel engines, both through gaseous emissions of sulfur
dioxide  (S02>,  and  also through  the further oxidation of
SO2  to   sulfate  particulate.    The  relative  amounts  of
sulfur  dioxide   and   sulfate  particulate  matter  emitted
depend  on the engine design  and  operating conditions. Most
test  data show that between  one  and  three percent  of the
sulfur  in the fuel is emitted as particulate sulfate, while
the  rest   is  emitted   as  S02-   The  determination  of  a
representative  sulfur-to-sulfate  conversion  fraction from
the  available test  data is  necessary for  the  purpose  of
projecting  vehicle  emissions,  and  a  discussion  of  the
methodology  and  data   used  in   this   determination  are
presented below.

     Some   studies  have  also shown  a correlation between
fuel  sulfur  content  and  non-sulfate particulates.   This
effect  appears  to be, small,  and, in  fact,  has  not been
observed  in  many experiments.   If  this   correlation does
exist,   it   may  be  partially  explained  by  the  partial
saturation  of polycyclic aromatic  hydrocarbons in  the fuel
during   the  desulfurization  process.    A  more  detailed
discussion  of this  issue  was presented  in Section  III of
Chapter 2.

     Table  4-1  documents  the average  results of  transient
mode  testing   performed   on  various   heavy-duty   diesel
engines.[1-5]   The  data  indicate  that  the  conversion of
sulfur  to   sulfate  varies   significantly from  engine  to
engine.   Predicting  sulfur  to  sulfate  conversion  for  a
given engine  operating  in   a  given  mode  is a   difficult
matter  and  would require a  more  complete understanding of

-------
                             4-3


                         Table 4-1

    Sulfate Conversion in HDDS Exhaust (Transient cycle)

                                     Average
                                     Percent
Manufacturer         Engine         Conversion     Reference

    Cat    •            3406B           1.5            [21
    Cumm             L10-270           2.2*           C3]
    Cumm            NTCC-400           2.8*           [3]
    DDA               	           1.1            [4]
    Nav               7.3IDI           1.5            [5]
    Nav           '    DTA466           1.6            [5]
    Nav               DTA466           1.9            [5]
    Cumm            NTCC-400           2.7            [1]

Sales Weighted Percent Conversion »   1.9
     Excluding low sulfur fuel results.

-------
                            4-4
the mechanism  of sulfate  formation  in  the engine  than is
available.   Thus,  a   sales   fraction  weighting   of   the
conversions   seen   in  the   various   test   engines   was
performed.  Projected  sales figures were taken  from engine
manufacturers'   confidential  estimates  submitted  to  ZPA.
The results  of  this   sales  weighting  indicate  an  average
conversion  of  approximately  1.9  percent  for   heavy-duty
diesels.

     However,  if  the  results for  one engine,  the  Navistar
7.3 liter  IDI  engine, were excluded from the data  (due to
the fact  that  it  is  an IDI engine),  the  average conversion
for heavy-duty engines becomes  2.2  percent.  Therefore,  a
conversion  of   2.0   percent  was   used  hereafter.    The
sensitivity  of  results  of  the  study  to  the  conversion
assumed was also investigated in Section IV below.

     Studies have  reported that  a  significant  increase in
the   average   conversion  of   fuel   sulfur    to   sulfate
particulate  occurs   when  operating  on   a    low   sulfur
fuel.[1,3]   A  discussion  of   the   data  recording  this
phenomenon and possible explanations  for  it  are  presented
below.

     The   Coordinating   Research   Council   VE-1   project
currently underway at Southwest Research  Institute  had, at
the  time  this  analysis  was  performed,  tested  only  one
engine, a Cummins NTCC400, on a  variety  of fuels under both
steady-state and  transient conditions.[lj  The  fuel matrix
for this  test program consists of  nine fuels,   three  with
low  sulfur  content   (0.05  percent  nominal),  one  at  an
intermediate  sulfur   level (0.15  percent),  and five  high
sulfur  fuels  (0.30 percent).   At nominal fuel sulfur levels
of  0.30 wt  percent,   fuel  sulfur  conversions ranging   from
1.6 percent  to 2.7 percent had been reported,  with average
levels  of  2.0  percent.  It appears from  the  data available
that  sulfur  to  sulfate conversion increases to nearly 5.0
percent on the low sulfur  fuels.

     Transient  emission testing of two Cummins  engines,   a
1986  L10-270 and a  1986  NTCC-400,  was performed by Cummins
Engines   Company  in   cooperation   with  the  Department  of
Energy, Mines, and Resources  of Canada.[3]  Average sulfur
to  sulfate conversions  using normal engine timing and at
normal  sulfur levels  (e.g.,  0.23  percent  and 0.39 percent
by weight) were 2.2 percent for  the L10  and 2.8  percent for
the NTCC-400.   This  test data showed a dramatic increase in
sulfur  conversion (up to  18  percent) using  a.  low sulfur
(0.02 percent  by weight)  fuel.

     One   possible   reason  for  this  observation   is  the
collection of  background sulfate particulate in addition to
that  produced by the engine.   For  instance,  for  a typical
heavy-duty engine operating on  0.05  weight  percent sulfur

-------
                            4-5
fuel,  one  would   predict   sulfate  emissions   of   0.0057
g/BHP-hr  (assuming  0.42  Ib/BHP-hr  fuel  consumption,   no
bound   water,   and   2.0   percent   sulfur   to   sulfate
conversion).   Whatever  background  sulfate  emissions  are
being measured  in  addition to  the  fuel  sulfur  emissions
would distort the  percent  conversion  calculated from  the
data.  A study performed  by John C. Wall of  Cummins  Engine
Company  on  a  LTA10-300  engine  using  sulfur  free  fuel
(dodecane)   showed  measured  background  sulfate  particulate
levels  of   0.005   g/BHP-hr.[3]    In   the   example   above,
assuming this background  particulate level  was  measured  in
addition  to  the  fuel  related  sulfate  emissions   during
testing, a  total of  0.0107 g/BHP-hr  of sulfate would  be
collected,   and  the  sulfur  to  sulfate conversion calculated
would  be  3.7 percent  instead  of  the  true  2.0  percent
conversion.   If  0.25  weight percent  sulfur  fuel were being
used the relative contribution of  background sulfate would
be smaller,  the  apparent  sulfur to sulfate conversion would
be  calculated  to   be only 2.4  percent  versus  the  true
conversion  of 2.0  percent  in  the engine.   Assuming this
background  sulfate measurement  exists in   the  test  data
reported,  this  may  account  for  some  of  the  perceived
increase in conversion with  low  sulfur  fuel  apparent  in the
literature.

     Another  explanation  for  the  experimental observations
of a higher indicated conversion rate at low sulfur   levels
may  be  the  imprecision in  experimental measurement.   Since
percentage  conversion  is  proportional  to  the  ratio  of
sulfates collected  to fuel  sulfur  level,  any  error  in the
measurement  of  either value   will affect   the  calculated
conversion.   An  illustration  of  this,  similar  to  that
presented by John C. Wall,  follows.[3]

     Due to the  low fuel sulfur level  being  measured (0.05
weight  percent)  any  error   in  measurement  might  produce  a
drastic  change  in  calculated  conversion.    For  example,   a
fuel with  a measured level  of  0.05 weight  percent may have
an  error in  sulfur  measurement of  +0.02  weight  percent.
Assuming the actual percent conversion of sulfur to  sulfate
is  2.0  percent,  if the  actual level  of  fuel  sulfur were
0.03  percent (while  being measured  as 0.05 percent),  the
sulfate  (dry) collected would  be 0.0034 g/BHP-hr.  Assuming
a  ±0.003 g/BHP-hr  sulfate  measurement  error  (i.e.,   sulfate
measured  -  0.0004  to   0.0064  g/BHP-hr),   the  calculated
conversion  would range   from  0.1  percent  to  2.3 percent.
Conversely,   if  the  actual fuel   sulfur  level  were 0.07
percent  (measured  as  0.05  percent),  calculated conversion
would range from 1.7 percent to 3.9 percent.   For an  actual
sulfur  to  sulfate  conversion of 2.5 percent,  experimentally
measured conversions  could range any where from  0.1  to 3.9
percent.

-------
                            4-6
     If  the   effects  of   background  sulfate   emissions
described  earlier   (0.005   g/BHP-hr)   are   included,   the
situation gets  worse.  For  the  same   situation  as  above,
calculated  conversions   could  range   from  1.9   ~o   5,6
percent.    Of  course  these  errors can  be  minimized  by
repeated  experimental  measurement (including   background
sulfate  measurement)   in   a  variety  of   laboratories.
However,  the- possibility  for this  type of error  does  exist
and, assuming a  normal  distribution of measurement  errors,
it  is  not  surprising that  some studies  have reported  an
apparent   increase    in   calculated   sulfate   conversion.
Furthermore,  data also  exist indicating the reverse effect
(i.e.,   that  conversion  is   lower  at  low  fuel   sulfur
levels).[4]

     To  summarize,  at this  point  in  time  it has not  been
conclusively  demonstrated  that  sulfate percent  conversion
increases  at  low fuel  sulfur  levels.   EPA's  conclusion  is
that the  few reports documenting this  effect merely  point
to  the variability  in  the data  and the error  inherent  in
measurement.   Even if conversion  increases with  low sulfur
fuel in  some  engines,  it  certainly  does not appear to do so
in  all  engines,  and  therefore  this study will  assume  that
the sulfur to sulfate conversion is 2.0 percent at all fuel
sulfur levels.

     2.    Light-Duty Diesel Engines

     Although the vast  majority of highway  diesel  fuel  is
consumed   by   engines   classified  as   heavy-duty,   any
modification to  diesel  fuel  quality will have an  effect  on
emissions  from   light-duty  diesel   trucks (LDDT)  and diesel
passenger cars  (LDDV) as well.   To  determine  the magnitude
of  this effect  the  existing  base  of  emissions  test  data
from  light-duty  vehicles   operating on  fuels  of  varying
sulfur, aromatics, and volatility was examined.

     As  described in Chapter  3, light-duty diesel  engines
typically  convert  approximately 1.5 percent of  the sulfur
in  the fuel  to  sulfate  particulate.  Using the  mean level
of  1.5 percent,  and  assuming  that  the  ratio  of associated
water  to sulfate particulate is 1.32 (50 percent humidity),
the effect  of fuel  desulfurization on particulate  and  S02
emissions  can be readily calculated for LDDV and LDDT from
average  fuel economy data.   Results  of such calculations
will be shown in Section III below.

     B.    Effect  of   Fuel   Aromatics  on  Highway  Diesel
           Emissions

     1.    Heavy-Duty Diesel Engines

     Many   of   the   aforementioned   studies  were   also
instrumental  in determining  the effects of fuel aromatics

-------
                            4-7
content on  emissions.   Fuel aromatics  have been correlated
with   particulate   emissions,    particularly  the   soluble
organic  fraction   (SOF)   and  the   residual   carbonaceous
portion (RCP)>  as well  as with  gaseous  HC emissions.   A
discussion of  the  data as  well  as some  of the confounding
factors involved in interpreting it is presented below.

     First,   .a  number  of  different  methods   exist  for
measuring  fuel  aromatics.   ASTM  test  D-1319,  which  is
commonly  used  in  industry,   utilizes  what  is  known  as
Fluorescent   Indicator   Absorption   (FIA).    This   test
procedure   measures   volume   percentages   of   saturates,
olefins,  and   aromatics   in   petroleum   samples.    Those
compounds  measured  as  aromatics  include  monoeyelie  and
polycyclic  aromatics,  aromatic  olefins,  some   dienes,  and
compounds containing sulfur,  nitrogen,  and oxygen  atoms.
Unfortunately,  compounds which are entirely aromatic,  those
which  are  aromatic  with  an  olefin  sidechain,  and  any
sulfur,  nitrogen,  or  oxygen  compounds  are • all  counted
equally as  aromatic  species.   The contribution of  each  of
these  types of compounds to  emissions,  however,  is  not
likely  the  same.  Two  fuels with identical FIA aromatics
measurement  may  thus  have  radically  different  aromatic
compound  types,   and  consequently,  will   have  different
emission forming characteristics.

     Furthermore,   the   FIA   measurement   technique   is
recommended only for petroleum  fractions that  distill  below
600°F, at which temperature only about 90 percent of diesel
fuel  distills.   As   stated in  the  ASTM   test  procedure,
results  of  the  test  are erratic  on  petroleum fractions
boiling near  the 600°F limit.   Because of  these problems,
there   is   some   question  over   the   appropriateness  of
correlating particulate emissions with FIA aromatics.

     Other  test procedures  for  measuring  aromatic  species
exist.  Mass Spectroscopy,  and  Proton and Carbon-13 Nuclear
Magnetic  Resonance (NMR)  all  quantify  aromatic  levels  in
fuel.   These  methods  make distinctions  between  aromatic
species  type,  whether  monocyclic, dicyclic, or tricyclic.
The  resulting  measurement  is therefore more representative
of  the  amount  of  aromatic  carbon  atoms in  the  fuel.
Unfortunately,  these  aromatic  measurement procedures  are
not  currently used  widely, and  most available  studies  of
emissions versus  aromatic  level  do not  contain measurements
of aromatics by Mass Spectroscopy or NMR method, only FIA.
Further,  little data are available other than FIA aromatics
describing   the   speciation   of   aromatic  compounds   in
commercial  fuel.

     Because  of the  large body  of data based  on  the FIA
measurement  procedure,  an attempt  to correlate  HC  and
particulate emissions  with  FIA  aromatics  levels  will  be

-------
                            4-8
made.  However, NMR and  Mass  Spectroscopy have been used to
analyze the aromatics  content of the fuels used  in  a single
study,   the   CRC's   VE-1   heavy-duty  Emissions   testing
project.[1]  The  emission data  taken  from  the one engine
tested  there  (Cummins NTCC400)  will  be  used to  correlate
emissions  with  NMR   and/or   Mass   Spectroacopy  aromatics
levels.    A   description  of   this   correlation   will   be
presented later in this section.

     a.    Correlation of Emissions with FIA Aromatics

     As  described  above;   SOF,   RCP,  and   HC  (gaseous)
emissions  have   been  observed  to   correlate   with  fuel
aromatics.  Preliminary  data from the VE-1  project  on  the
Cummins  NTCC400  show  that   a correlation  exists  between
these  variables  and FIA  aromatics.[1]    Regressions   of
emission data  taken from this engine,  show that  a reduction
in  fuel aromatics  from  34.2 to  20  volume  percent  would
result  in  SOF  and  RCP  reductions  of  14.4 and  9.6 percent
from baseline, respectively.

     The   effect  of   aromatics   on  hydrocarbon  emissions
appears  from  this  data  to  be  similar to  the  effects  of
aromatics  on  carbonaceous particulate.  From  inspection of
the  data,  it appears  that  control of  fuel  aromatics would
result  in  HC  emission  reductions   of  9.6  percent  from
baseline,  respectively.

   .  The    Mobil/Caterpillar    cooperative    study   also
investigated   the  effects  of   aromatics  on  particulate
emissions, albeit on  an older Caterpillar 3406B engine.[2]
Regressing  SOF emissions  against FIA aromatics (transient
test  results  only),  it appears that a one percent reduction
in fuel aromatics corresponds  to  a  2.5 percent reduction in
SOF  emissions.  Similarly,  it appears  that  a  one percent
reduction  in  aromatics  will  result  in  a  2.4  percent
reduction  in RCP.  HC emissions were also  correlated with
fuel   aromatics.   On this  engine,   the  reduction  of  HC
associated with  aromatics  control   was  1.1  percent  per
percent  aromatics  reduction   in  the  fuel.    This  would
translate   into  SOF  reductions  of  35.5   percent,  RCP
reductions of  34.1  percent,   and HC reductions   of  16.1
percent  for  an  aromatics  reduction  to  20  volume percent
from  baseline   levels,   significantly   higher   than  the
reductions achieved on  the  VE-1 Cummins engine.  Since the
results  of this engine represent older  technology than the
VE-1  data,  the VE-1 data will  be used  preferentially in
this   analysis.   However,   the  results  of   the  Mobil
Caterpillar  study were  used  as  a sensitivity  analysis,   a
full  discussion of which is  reserved for Chapter 7.

-------
                             4-9
     Correlation  of  emissions with FIA aromatics  can also
be  performed with  the  data  of  Shirish  Shimpi.[3]   Only
three  fuels  were  used,  and  the  correlation  of  emissions
with  aromatics  levels  was  not  very  clear,   An  overall
downward trend  in non-sulfate  emissions with fuel aromatics
reduction  was  observed,  however.   Several  other  studies
have  been  performed  investigating  the  effects  of  fuel
aromatics  on  diesel  emissions  as  well.   However,  these
studies  were performed  predominantly  on older  technology
engines  operating under steady state  conditions.   Although
somewhat useful in  confirming the  results  of  the studies
described  above,  the data contained in them were  not used
directly in this analysis.

     This  analysis assumes that  fuel aromatics control will
result  in  straight  percentage  reductions on  the  various
emission   components    of   heavy-duty   diesel   exhaust,
regardless of  the absolute  emission  level  of  the engine.
However, not all  of  the SOF and RCP particulate, nor all HC
emissions,  can be  attributed  to   aromatics   in  the  fuel.
Some  of the  emissions have been  shown to be  derived from
other  sources  (e.g.,   lubricating   oil,  non-aromatic  fuel
components,  etc.).    For  current  engines,   the  percent
reductions in emissions  estimated   above  should implicitly
take  into  account these  other sources.  More uncertainties
arise,  however, when  this  approach is  applied  to future
engines.

     Many  manufacturers  have pointed to improved engine oil
control  in future engines as  a means  of  reducing  lube-oil
derived  emissions.    Were  oil-derived  emissions  reduced
disproportionately,  the  percentage reduction   in   emission
corresponding  to fuel  aromatics control would  be greater
since  the  fractional amount  of emissions  derived  from fuel
aromatics  would   be  greater  than  in  current   engines.
Therefore,   emission reduction estimates  based on current
engines  and  applied  to future  engines  may underestimate
actual  emission reductions.   However,  the  opposite  could
also  be true if  the reduction of  fuel-based RCP emissions
reductions   outpaced  those   of   oil-related  emissions  on
future   engines.   Unfortunately,   either  situation  could
occur.   This is another  reason  for using the  range of SOF
levels  foe future engine emissions described in Chapter 3.
Therefore,   it   appears   reasonable   to   apply   percent
reductions in emissions  to both current and future engines.

     b.    Correlation of  Emissions  with Aromatic Carbon
           (Mass Spectroscopy, NMR Analysis)

     As  mentioned in Section I-A-1  of  this chapter, some of
the  studies  investigating  fuel  effects  on  emissions have
observed that  there may  be  a  direct  correlation between

-------
                            4-10
fuel  sulfur  levels  and   the  amount  of   SOF   emissions
collected.[1,6]    Unfortunately,  the  effect  has  only  been
seen with a  few engines, while  on the majority  of  engines
tested   it   has  not   been  demonstrated.    It  has   been
hypothesized that the  effect may  be  attributable  to  the
"scrubbing"   of  hydrocarbon compounds  out  of the gas  phase
by  reaction  with sulfuric  acid  on the particulate  filter,
or that  the  cause may be improved filtration efficiency due
to the presence  of  sulfuric acid and  associated  water.[6]
It may  also  be  the  case that the  correlation  between fuel
sulfur   and  SOF   emissions   is   attributable   to   the
desulfurization  process rather  than  the fuel  sulfur  level
per  se.   Diesel  fuel  desulfurization  via  hydrotreating
often  results   in  the  partial  saturation  of  polycyclic
aromatic  compounds  (see  Chapter  2,  Section  III).    The
elimination  of   some  of these polycyclic  aromatics  in the
fuel may be  the cause of lower  SOF emission  levels.   Using
the  information  presented  in Chapter 2 on  the effect  of
hydrotreating on aromatic  species  and  by  determining the
effects  of  various  aromatic  species  on emissions  (as will
be done  this section),  it  is  possible to predict reductions
in SOF  and other carbonaceous emissions resulting from fuel
desulfurization.

     It  has  been   hypothesized  that  different  aromatic
species  may  affect  emissions  from  diesel  vehicles   to  a
different   extent.     More   specifically,   diesel   engine
emissions may correlate  more closely  with aromatic carbon
content  of  the  fuel (as  determined  by Mass Spectroscopy,
NMR)  than  with  the volume fraction  of aromatic containing
compounds  (FIA   analysis).   Unfortunately,   little  emission
and  fuel composition data exist in which fuel aromatics are
measured by methods  other  than FIA.   Further,   little  is
known   about  the  speciation  of  aromatics  compounds  in
commercial  fuel. ' Still,  the fuels  used   in  the CRC VE-1
emissions  study have  been  analyzed  thoroughly by   these
various  methods, and by making  some  assumptions  about the
distribution  of  aromatic  compounds   in commercial  diesel
fuel,  it was possible  to predict here the effect  of fuel
control  on SOF,  RCP, and HC emissions.[1]

     Fuel   aromatics  analysis  by   Mass   Spectroscopy  is
performed by first  obtaining the aromatic fraction  of the
fuel by chromatographic separation.   Mass Spectroscopy (MS)
is  then performed  to  separate  the  aromatics by class and
aromatics  composition  data are derived  in volume percent.
Determination of aromatics  content  of the  fuels used  in the
CRC VE-1 project by MS was  performed by  Chevron Research
Company and Unocal  Corporation.  Determination of aromatic
carbon  by Proton and  Carbon-13 Nuclear Magnetic Resonance
Spectrometry (NMR)  was  performed  on  the  same fuels  by
General  Motors  Research Laboratories.  Results  of  the NMR

-------
                            4-11


analysis  are  reported as  the  mole percent  aromatic  carbon
of the fuel.

     The  emission   data   from  the   VE-1   project   were
correlated with, fuel  aromatics  levels  determined using both
MS and  NMR.   A "normalized  MS  value"  was first  determined
for  each fuel  by combining the volume  percent of  mono-,
di-,   and tricyclic  structures,  weighted  according  to  the
relative  amount   of  aromatic  carbon   in  each type  of
structure.  Correlation  of  Emissions  were  then  correlated
with  this  "normalized MS value" as they  were with the mole
fraction aromatic carbon  in the fuel,  as determined  by the
carbon-13 NMR method.

     In   order  to  use   these  correlations   to  estimate
emission  reductions   resulting  from fuel control,  it  was
necessary   to  make  some  assumptions   concerning   the
speciation of  aromatics  in  commercial  fuel  both currently
and  under  the fuel  control   scenarios  being  considered.
Little  data  exist  on the  distribution  of  aromatics  in
commercial fuel.  Great variability  may exist  from refinery
to refinery  as well as  among  streams  within  a  refinery,
depending  on  the  processing  each has  been subjected to.
However, most distillate fuels  on  which data were reported
show  that the majority  of aromatic . species are monocyclic
structures.   Approximately  10  to   50   percent  of  the
aromatics  are  dicyclic   structures,  while  typically  5  or
less   percent  are   polycyclic   (tricyclic  or   greater)
structures.[1,7]

     As  a basis  for  this  analysis,  it  was  assumed that
baseline  fuel  (FIA  aromatics  of  34.2  percent)  has  an
aromatics  distribution   of  mono-,   di-,  and  tricyclic
aromatics  of  26.4,   6.8,  and  1.0 volume  percent.    While
selecting any distribution of  aromatic species  to represent
commercial fuels is  highly subjective  based on  the  limited
data  available, a  baseline assumption  had to be made.  This
distribution was selected assuming 20  percent  of aromatics
were  dicyclic,  3  percent  were  tricyclic, and  the remaining
77  percent were   monocyclic.    This  was  perceived   to  be
fairly  representative of  the  distribution  of  aromatics in
the  mid-range  volatility/ mid-range aromatic  fuel used in
the  VE-1  test  program (fuel tt5) and of the distillate fuels
tested   by  Yoes   and  Asim.[7]   While   a  baseline  fuel
containing up to  50 percent dicyclic  aromatics  could have
reasonably  been  selected,  selecting  such  a   fuel   as   a
baseline   would  have   changed  the   ensuing   particulate
emission  reduction   estimates   by  only a  few  percent.
Therefore,    the     sensitivity   of    emission   reduction
projections to the  distribution of  aromatics   in baseline
fuel  is  low and can  be ignored.

-------
                            4-12
     As described  in  Chapter 2,  a  50 percent  reduction  in
di- and tricyclic  structures with  a  corresponding increase
in monocyclic  structures  should  occur  with sulfur  control
(i.e., when the fuel  undergoes  hydrodesulfurization).   This
would yield an aromatic distribution  of 30.3,  3.4,  and 0.5
volume percent.  Additional  processing necessary  to  reduce
aromatics   to  20  volume  percent  would  further  shift  the
distribution  of  mono-,  di-,  and  tricyclic  structures  to
18.5,  1.4, and 0.1 percent, respectively.

     Using  these   distributions   of  aromatics,   and  the
correlations  of  emissions  with  the  "normalized  MS  value"
developed  from the VE-1  project  data, emission  reductions
can  be estimated.    From this  analysis,   it  appears  that
small reductions of 2.8,  1.1,  and 1.9 percent  in emissions
of SOF, RCP,  and HC would result from fuel sulfur control.
Reductions of  18.5, 6.8,  and 10.8 percent  in SOF,  RCP,  and
HC  emissions  were  predicted  to  result   from  control  of
sulfur in conjunction with aromatics control to 20 percent.

     These   emission   reduction   estimates   are   fairly
consistent  with   those   projected   according   to  the  FIA
analysis  for  the  Cummins  engine  (reductions  of  14.4,  9.6,
and 9.6 percent  in SOF, RCP, and HC  with  aromatics control
to 20  percent),  and  more  accurately reflect the  fact that
some  degree  of   partial   fuel  aromatics   saturation  takes
place  with   fuel   hydrodesulfurization.    Results  of  the
"normalized  MS value"  correlation were  therefore  used  to
predict   reductions   in   emissions   corresponding  to  fuel
control in the remainder  of  this report.   However, the data
with   which   to   construct   reliable  models   correlating
emissions  with  more  sophisticated  aromatics  measurement
procedures such as Mass Spectroscopy  are limited  and so the
analysis  presented here  should be  considered  preliminary.
As more  data  are  generated,  more  sophisticated  emissions
modeling   of   this nature  should   become  feasible,  thus
improving the ability to  precisely  predict the relationship
between fuel aromatics and emissions.

     2.    Light-Duty Diesel Engines

     Due   to   the   differences    between  IDI    (indirect
injection) and DI  (direct  injection)  diesel engines and the
different duty  cycles  of  cars  and large  trucks,  it   is
generally impossible  to  use data  generated on   heavy-duty
engines  (mostly  DI)  to predict  fuel effects  on  emissions
from  light-duty  IDI  engines.   Given  this limitation,   in
order  to  determine the effect of fuel control on  light-duty
diesel  emissions,  one must  turn to  the   data  generated  on
light-duty diesel  vehicles.

-------
                            4-13
     One of the most  recent  studies which  investigated  the
effects  of  fuel  variables  on  emissions  was  CRC  Project
CAPE-32-80,  performed  under  contract   by  the   Southwest
Research Institute.[8]  This  test program  consisted of  the
measurement of FTP  HC,  CO, NOx,  and particulate  emissions
from  four  1982 "MY  light-duty  diesel vehicles  operating on
nine  different  fuels  of   varying aromaticity,  volatility,
and sulfur content.   The  fuel  matrix was selected such that
the  effect of  aromatics,  Tgo/   and T^Q  on  HC,  CO,  and
NOX   emissions  could  be  evaluated   independently.    The
amount   of  sulfur   in   the  fuel,   however,   varied   in
conjunction with the  other three  fuel parameters and so  the
effect of these three parameters  on  particulate emission is
somewhat  confounded  with the  effects  of  fuel  sulfur  on
particulates.  By  assuming that  1.5  percent  of  the  fuel
sulfur is  converted to particulate  sulfate (see Chapter  3),
and that 1.32  grams  of water  is  associated with  each gram
of  sulfate  particulate,  one  can  normalize the  emissions
results of this study to remove  the effect of  sulfur from
the   particulate   emissions   data.    This  was  done   by
subtracting  the  assumed  weight   of  the  sulfate  and  bound
waters from the total particulate levels.

     After  doing  this,  these  adjusted  particulate levels
were  then  regressed  against  fuel  aromatics  (FIA),  Tgg/
and  TIQ.   From the  regression  it  was  determined  that  for
the  four vehicles  tested,  a  reduction  in fuel  aromatics
from  34.2  to 20  percent  would  result  in a  non-sulfate
particulate  reduction  of  approximately  7.2   percent   on
average.

      Inspection of  the CAPE-32-80  test  data  also revealed
that  a decrease  in HC emissions  of about  50  percent would
result   from  dearomatization   from  34.2  to   20   volume
percent.   Data  from this  testing program  also  showed some
possible  effect  of  aromatics  on  NOx  emissions,   but  the
supporting   evidence   is  not   sufficient   to  make  any
quantitative statement.

     A  recent  cooperative  study  by  Mobil  Research Dev.
Corp. and  Toyota Motor  Corp.  also  investigated the effects
of  fuel  properties on FTP emissions from a light-duty  (2.2
liter) Toyota  truck  engine.[9]   Eight  different  fuels were
tested;  two  containing  20   percent  aromatics,   five with
aromatics  levels  ranging  from 32  to 37  percent,  and  one
containing  49  percent aromatics.   A curvilinear regression
of  test  results  showed  an  average reduction  of  about 43
percent  in total  particulate emissions  corresponding to  a
reduction  in  aromatics  from  34.2   to  20  volume   percent.
This  corresponds to  a reduction of  Approximately 50 percent
in  RCP  and   SOF,   assuming   that   .he  sulfate  emissions
remained constant.  A reduction of   approximately 60 percent
in  HC emissions was also observed.

-------
                            4-14
     The  results   of   these  two   studies   were  combined
according to the number of vehicles tested in each  (i.e.,  a
four  to one  vehicle  weighting)  in  order  to estimate  an
average  effect  of  fuel  aromatics  control  on  particulate
emissions for this  analysis.  Combining the  results  of the
two  studies  in  this  way,  one  would   estimate   that  a
reduction in aromatics  to  20 volume  percent  would reduce HC
emissions  by   about  52   percent   from  baseline,   while
non-sulfate  particulate  would be   reduced  by  roughly  15
percent.

     Due  to  the  lack  of  data  and  the  relatively  small
contribution  of  light-duty  diesels  to  the  total  diesel
particulate emission inventory,  it  was assumed  that  fuel
sulfur  control  (and  the corresponding partial saturation of
polycyclic aromatic structures) would have no effect  on LDD
HC or SOF emissions.

II.  Effect of  Fuel Quality on Off  Highway  Diesel  and No.2
     Fuel Oil Emissions

     The  diesel fuel  sulfur and sulfur/aromatics  controls
studied  in  this  paper are  intended for  on-highway  diesel
fuel  only.   However,  as  explained  in  Chapter 2,  refiners
market  pn-highway,  off-highway and  a  significant amount of
fuel  oil  from  a   common  distillate  pool.    With  fuel
controls/  they  are  faced  with  the  decision to treat the
whole  pool  or  treat part  of the  pool  and  segregate the
products.  It is  expected that some refiners will treat the
whole pool, and thus at least some of the  fuel oil  marketed
will  have  a   low  sulfur  content,   or a  low  sulfur  and
aromatics content.  This  will  result  in emission reductions
for  those sources that use this  fuel  oil.   The  purpose of
this section is to explain the changes  in emissions.

     Total  sulfur   oxide  emissions  are   almost  entirely
dependent  on   the  sulfur   content  of  the   fuel  and  are
generally not affected  by boiler size or  burner design.  On
the  average,  more  than 95  percent  of  the   fuel sulfur  is
emitted as  SC*2  (like  internal  combustion  engines),   while
the  rest  is  emitted  as  sulfate   particulate   and  sulfur
trioxide,  which  guickly  reacts  with  moisture  to  become
additional  sulfate  particulate.   Due to   the  similarity
between the sulfur  conversion for  these applications and
diesel  engines,  identical sulfur  to sulfate  conversion will
be used for  these stationary fuel oil  sources,  and engines
burning off-highway No. 2 diesel.

     Non-sulfate   particulate  emission   from   off-highway
sources were assumed  to remain constant with  fuel  aromatics
control.   As   indicated   in the  Mobil/Caterpillar  study,
emissions   from   engines   operating  under  steady-state
conditions   show  little   sensitivity   to  fuel   aromatics

-------
                            4-15


levels.[2]  Since  engines  used in  off-highway  applications
operate primarily  in a  steady-state mode,  it  was  assumed
that  fuel  aromatics  control  would   have  no  effect  on
off-highway particulate emissions.

     While it  can  be argued  that  such an  assumption  tends
to  underestimate  the  possible  emission benefits  of  fuel
aromatics  control,   results   of   the   cost-effectiveness
analysis  (Chapter  7) indicate  that this assumption is not
critical.    Even  if  aromatics  control   (under   the   NPRA
segregation  scenario)   resulted   in   the  same   emission
reductions from off-highway as  from on-highway  diesels, the
cost-effectiveness   ($/ton   of   urban   particulate)   of
aromatics  control  estimated  under  the  NPRA   segregation
scenario  would be no  lower  than  that derived  in the 100
percent  segregation  case.   Since,  as  will  be  seen  in
Chapter  7,  aromatics  control  does  not  result  in   cost
effective   emission   control  under   either   segregation
scenario,    this   simplifying   assumption  is    somewhat
perfunctory.

III. Diesel Aftertreatment Technology With Sulfur  Control

     The  aftertreatment  technology mix for  compliance with
1991 and  1994  particulate  standards was determined assuming
fuel  sulfur  was  reduced  to  0.05  weight  percent.    The
relationships between fuel  sulfur  control and diesel engine
emissions  established  in  Section   I  of  this  chapter  were
used  to   adjust  the   engine-out   emission  distributions
developed   in   Chapter   3.    The   adjusted   engine-out
distributions  were then -used in  estimating the most  cost
effective  mix  of aftertreatment technologies necessary for
compliance with  the  particulate standards.   Results  of the
analysis are presented below.

     A.    Light-Duty Diesels

     Emissions for 1991-1993  LDDs  with  fuel sulfur control
are  shown  in  Table 4-2.    Because  vehicle technology is
unaffected, LDDV and LDDT1  sulfate and SC-2  emissions were
simply  reduced from baseline levels  (Chapter 3)  to reflect
the  degree  of  fuel  sulfur  control.    The  fuel  sulfur
reduction  also  caused   a  reduction in  sulfate particulate
emitted by LDDT2s.   This did not show up  as a  reduction in
total  particulate,  however,  but  rather  as a  reduction in
the  number  of   traps   required  to  meet   the   0.13  gpm
particulate  standard  (86.7   percent).    Table  4-2   shows
emission   levels  and   aftertreatment   costs    for    these
vehicles.   Emission  reductions   and  aftertreatment  cost
savings from baseline levels  are shown in the Table 4-3.

-------
                                     4-16
                                  Table 4-2

             1991 Light-Duty Diesel After-treatment Technologies,
               Costs and Emissions Under Various Fuel Controls
Vehicle
Class
Percent
Traps
Aftertreat
Cost
(S/veh)
Fuel
Economy
(mpq)
Emissions (q/mi)
SOF
RCP
Fuel Sulfur Control
_S04
TPM

S02

HC*

1991-93 Model Years
LDDV
LDDT1
LDDT2
1994 and
LDDV
LDDT1
LDDT2
0
0
86.7
Later Model
0
0
86.5
0
0
394
Years
0
0
393
31.4
25.9
25.5
34.2
27.8
27.3
.024
.108
.055
.024
.108
.056
.079
.072
.027
.079
.072
.027
.006
.007
.007
.005
.006
.006
.109
.187
.089
.108
.186
.089
.104
.126
.128
.095
.118
.120
.29
.43
.188
.29
.43
.19
Subsequent Fuel Aromatics Control
1991-93 Model Years
LDDV
LDDT1
LDDT2
1994 and
LDDV
LDDT1
LDDT2
0
0
79.6
Later Model
0
0
79.3
0
0
362
Years
0
0
361
31.4
25.9
25.5
34.2
27.8
27.4
.020
.092
.053
.020
.092
.053
.067
.061
.029
.067
.061
.030
.006
.007
.007
.005
.006
.006
.093
.160
.089
.092
.159
.089
.104
.118
.128
.095
.118
.119
.139
.206
.072
.139
.206
.072
Zero mile emissions.

-------
                            4-17


                         Table  4-3
              Light-Duty Savings  and Emission
          Reductions Corresponding to Fuel Control
               Aftertreat.            Total Direct
Vehicle       Cost Savings         Particulate Emissions
 Class           ($/veh)             Reduction (q/mi)

Fuel Sulfur Control

1991-93 Model Years

LDDV                 0                    0.022
LDDT1                0                    0.026
LDDT2     .          61                    0.000

1994 and Later Model Years

LDDV                 0                    0.020
LDDT1                0                    0.025
LDDT2               57                    0.000

Subsecruent Fuel Aromatics Control*

1991-93 Model Years

LDDV                 0                    0.016
LDDT1                0                    0.027
LDDT2               32                    0.000

1994 and Later Model Years

LDDV                 0     "               0.016
LDDT1                0                    0.027
LDDT2               32                    0.000
     Incremental to sulfur control.

-------
                            4-18
     Table 4-2  also shows  aftertreatment requirements  and
emissions for 1994  and  beyond.   Once again,  LDDV  and  LDDT1
emissions of  sulfate  and  S02  were  reduced from  baseline
fuel levels and trap requirements  for LDDT2  (86.5  percent)
were also  reduced.   Emission  reduction  and  aftertreatment
cost savings are shown  in Table 4-3.

     2.    Heavy-Duty Diesels

     The  technology mix,  cost, and emissions  for  1991-3
HDDE's  with sulfur  control,  based  on the average of the two
nominal  engine-out  emission   distributions  developed  in
Chapter  3,  are shown  in Table 4-4.  As  can be  seen,  with
sulfur  control it is projected that no traps  will  be needed
to  meet  the  0.25  g/BHP-hr particulate  standard.   While
traps  are projected  to  be  needed   to  meet  the  standard
without  fuel  sulfur  control,   it  is  possible  that  trap
technology will not be  available in time.  If this  were the
case,  then  the cost  savings shown  can  be  used as  a  very
rough   indication   of   the  reduction   in  non-conformance
penalties which would  occur.   Urban buses  will  require  a
mixture  of  catalyzed  traps and flow-through catalysts  to
meet  the  0.10  g/BHP-hr   standard.    Aftertreatment  cost
savings  of $156 per urban bus sold,  not  including  the  cost
of  a fuel economy  penalty  due  to traps  would  be realized.
Aftertreatment  cost   savings   and   particulate   emission
reductions for each heavy-duty class are shown in Table 4-5.

     Table   4-4    also    shows   the   projected   exhaust
aftertreatment  requirements  and  emissions  for  1994  and
later  heavy-duty  diesels  under  fuel  sulfur control.   As
shown,  aftertreatment costs  savings  range from  $268 to $335
per  vehicle  as  shown in Table 4-5.  The utilization of high
activity flow  through  catalysts  was  increased  from  the
baseline fuel  case due  to the reduction in  fuel sulfur,
with utilization  rates as  high as  55  percent  of  vehicles
sold in the  case  of  LHDDEs.   The  use  of  some  traps
(approximately  30 percent  of vehicles sold),  however,  will
still be required to meet the 1994 particulate standard.

IV.  Diesel Aftertreatment  Technology Fuel Aromatics Control

     1.    Light-Duty Diesels

     Projected   emissions    and   exhaust   aftertreatment
technology   requirements   for   light-duty   diesels  under
subsequent  fuel control are shown in Table  4-2.   As  shown
in  Table 4-3,  emissions  of  particulate  are slightly reduced
from those  with  only  fuel  sulfur   control  for  LDDVs and
LDDTls.   For  LDDT2s, with fuel aromatics  control only 79.6
and 79.3   percent   of  the  vehicles  in  1991   and   1994,
respectively,  will require traps  to  comply with  the 0.13
gpm particulate standard.

-------
Table 4-4
Technology Usage (%)

Engine Non Cat
Class Traps
Fuel Sulfur Control
1991-93 Model Years
LHDDE 0
MHDOE 0
HHDDE 0
BUS 0
1994 and Later Model
LHD 0
MHO 0
HHD 20.7
BUS 0
Low Cost
Catalyzed
Traps


0
0
0
67.8
Years
29.5
36.0
15.8
34.4
Low Cost
Flow
Through
Catalysts


0
0
0
0

0
0
0
0
High Cost
Flow
Through
Catalysts


0
0
0
20.5

54.8
30.3
29.0
50.4
Average
Af tertreat .
Cost
($/vehicle)


0
0
0
$500

193.6
307.4
367.9
343.9


SOF


.108
.085
.050
.034

.028
.029
.017
.024

End
RCP


.101
.124
.160
.044

.042
.049
.062
.050

of Life
_§04


.016
.013
.012
.016

.023
.016
.014
.019




Emissions (g/BHP-hr)
TPM


.220
.221
.222
.094

.093
.093
.093
.093
so?

*
.231
.187
.173
.195

.221
.183
.165
.191
HC


1.12
1.19
1.30
.56

.614
.745
.950
.645
CO


3.97
4.35
4.87
3.88

2.47
3.49
3.96
2.86
Subsequent Fuel Aromatics Control
1991-93 Model Years
LHDDE 0
MHDDE 0
HHDDE 0
BUS 0
1994 and Later Model
LHDDE 0
MHDDE 0
HHDDE 18.3
BUS 0

0
0
0
61.4
Years
24.7
27.0
13.7
28.6

0
0
0
0

0
0
0
0

0
0
0
19.7

50.1
38.4
26.1
39.7

0
0
0
453

167
265
327
284

.086
.071
.042
.031

.028
.024
.015
.024

.095
.116
.151
.048

.044
.052
.064
.052

.016
.013
.012
.016

.022
.017
.014
.018

.197
.200
.204
.094

.093
.093
.093
.093

.231
.187
.173
.197

".221
.182
.165
.191

1.02
1.08
1.18
.56

.61
.70
.90
.68

3.9
4.3
4.9
3.9

2.6
3.2
4.0
3.2

-------
                              4-20
                           Table 4-5
            Heavy-Duty Vehicle Savings and Emission
            Reductions Corresponding  o Fuel  Control
vehicle
 Class
 Aftertreat.
Cost Savings
   ($/veh)
   Total Direct
Particulate Emissions
Reduct ion (g/BHP-hr)
Fuel Sulfur Control

1991-93 Model Years

LHDDE
MHDDE
HHDDE
BUS
     157
     192
     - 15
     -56
1994 and Later Model Years
LHDDE
MHDDE
HHDDE
BUS
     268
     335
     334
     315
Subsequent Fuel Aromatics Control*

1991-93 Model Years
LHDDE
MHDDE
HHDDE
BUS
       0
       0
       0
      47
1994 and Later Model Years
LHDDE
MHDDE
HHDDE
BUS
      27
      42
      41
      65
       0.004
       0.003
       0.002
       0.000
       0.000
       0.000
       0.000
       0.000
       0.027
       0.024
       0.020
       0.000
       0.000
       0.000
       0.000
       0.000
      Incremental to sulfur control.

-------
                            4-21
     2.     Heavy-Putv Diesels

     Table   4-4   shows   the   projected   emissions   and
aftertreatment  requirements   for  heavy-duty  diesels  under
subsequent fuel  aromatics  control.   As seen  in Table  4-4,
and as  summarized  in Table 4-5, emission estimates  of  SOF,
RCP, and HC  for 1991 model  year  HHDDEs are  slightly lower
than with only  fuel  sulfur control,  due to  the reduction in
fuel aromatics.   Urban buses  in 1991 and all  HDDDEs  in  1994
will  still  require  the use  of trap  oxidizers under  this
fuel  control  scenario,  although  slightly  fewer  than  with
only  fuel  sulfur control.   Aftertreatment  costs  savings in
1994 range from $27  to  $65  per vehicle,  not  including the
cost of the added fuel economy penalty due to traps.

V.   Sensitivity Analysis

     The cost effectiveness  of  fuel  quality control will be
determined  in  Chapter  7.    The  emissions  and  technology
results presented  thus far will be used there  as  the "best
estimate" of  the engine manufacturing  industry  response to
the  fuel  control  scenario under  consideration and  will be
used  primarily  in  that  analysis.   However, several  of the
assumptions  used in  this  analysis may have  a  significant
impact  on  the  vehicle  savings  and  emission  reductions
associated with fuel control.  These  include:   the percent
conversion of  sulfur  to  sulfate  in  the   engine,  baseline
engine-out  emission  levels,  and  the  cost  of  the  trap
oxidizer   systems.    To  evaluate   the  impact   of  these
parameters,  two  sensitivity  cases  were   run  on 1994 HHD
engines,  a "greatest savings" and a "lowest  savings" cost.
The  impact  of   these  parameters  on  results  for  the other
vehicle class was assumed to be proportionally similar.

      In  the  "greatest  savings"  case,  the  assumptions  made
regarding   sulfur   to   sulfate   conversion,   engine-out
emissions, and  trap-oxidizer costs  were  selected  in  order
to  maximize  the per-vehicle  savings  attributable  to   fuel
control.  A  sulfur to sulfate conversion of  2.5  percent in
the   engine   was  assumed,   as  were   the  "low  emission"
engine-out  distributions developed  in Chapter  3.   It was
assumed  that  trap  costs  were  the  same   as  in  the  "best
estimate"    scenario   (i.e.,   no   low-cost   traps    were
available).    Vehicle   savings  and   emission   reductions
corresponding to fuel  control  for  the HHDDE  vehicle  class
in  the  "greatest savings"  case  are shown in Table  4-6.

      In  the  "lowest  savings"  case,   modeling assumptions
were  selected  in   order to  minimize  the  vehicle   savings
resulting  from  fuel  control.    A  sulfur   to   sulfate
conversion of 1.0  percent was used, as  were the  "low-cost"
trap   cost   estimates   (Table  3-8).   Results   for   the
"lowest-savings" scenario  are also  shown  in Table  4-6.

-------
                            4-22


                         Table  4-6
    Sensitivity of Heavy-Duty 1994 and Later Model Year
             Diesel Engine Savings and Emission
          Reductions Corresponding to Fuel Control	
                                                 Total
                                               Particulate
                             Aftertreatment     Emission
                  Vehicle     Cost Savings      Reduction
   Scenario        Class        ($/veh)         (q/BHP-hr)

Fuel Sulfur Control

Best Estimate       HMD           ?334              0
Highest Savings     HHD           ? =.52              0
Lowest Savings      HHD           $131              0

Subsequent Fuel Aromatics Control*

Best Estimate       HHD           $ 41              0
Highest Savings     HHD           $ 31              0
Lowest Savings      HHD           $ 36              0
     Incremental to fuel sulfur control.

-------
                            4-23
     Inspecting Table 4-6, one  sees  that,  depending on  the
assumptions  used,   the  impact  of  fuel control  on  engine
technology can  vary greatly.   Aftertreatment cost  savings
range from $131  to $552  per  HHDDE for  sulfur control  (not
including  fuel  economy benefits)  around the  mean of  $334
per vehicle.   Cost  savings for  subseguent  aromatics control
range from $31 to  $36  per vehicle around a mean of  $41  per
vehicle.   These   values   will   be  used   in  the  overall
sensitivity analysis in Chapter  7.

-------
                            4-24


                   Reference! (Chapter 4)
     1.    "Investigation   of    the   Effects   of    Fuel
Composition  and  Injection and  Combustion  System  Type  on
Heavy-Duty  Diesel  Exhaust  Emissions,"  Terry  L.   Ullman,
Southwest Research  Institute,  Coordinating Research Council
Contract CAPE-32-80, Project VE-1, March 1989.

     2.    "Heavy-Duty   Diesel   Engine   Fuel   Combustion
Performance   and   Emissions   -   A  Cooperative   Research
Program," E.G.  Barry  and  L.J. McCabe,  Mobil  Research  and
Development  Corp.,  D.H. Gerke and  J.M.  Perez,  Caterpillar
Tractor Co., SAE Paper 852078, 1985.

     3.    "Fuel  Sulfur Reduction  for  Control  of  Diesel
Particulate  Emissions," J.C.  Wall,  S.A.  Shimpi,  M.L.  Yu,
SAE Paper no. 872139, 1987.

     4.    "General   Motors   Comments   on   Environmental
Protection Agency Notice Federal  Register Vol. 51,  No. 124,
June  27,  1986.  Diesel  Fuel  Quality Effects  on  Emissions,
Durability,   Performance and Costs;  Availability  of a Draft
Study," Docket No.A-86-03,  December  22, 1987.

     5.    "Fuel   Sulfur   Effect   on  Diesel   Particulate
Emission,"   Navistar   International  Corporation,   Engine
Division Engineering, June 29, 1987.

     6.    "Fuel  Composition  Effects  on  Heavy-Duty  Diesel
Particulate  Emissions,"  J.C.  Wall,  and  S.   K.  Hoekman,
Chevron Research Company, SAE Paper  841364, October, 1984.

     7.    "Confronting   New   Challenges   in   Distillate
Hydrotreating,"  Jack R. Yoes,  Mehmet Y.  Asim,  Akzo Chemie
America,  presented  at  1987  NPRA Annual   Meeting,  San
Antonio, Texas, March 29-31, 1987.

     8.    "Study  of the  Effects of Fuel Composition,  and
Injection  and  Combustion  System  Type  and Adjustment,  on
Exhaust  Emissions   from  Light-Duty Diesels,"  Charles  T.
Hare,    Southwest   Research   Institute,   Prepared   for
Coordinating  Research  Council   Inc.,  Project  CAPE-32-80,
April  1985.

     9,    "Effects of  Fuel   Properties  and  Engine  Design
Features  on the  Performance  of  a Light-Duty Diesel Truck  -
A  Cooperative  Study," Barry  E.G.,  Anelrod,  J.C.,  and
McCabe,  L.J.,  Mobile  Res.  S<~   9V.   Corp.,  Inove,   T.,  and
Tsuboi N,, Toyota Motor Corp.,  AE Paper no. 861526.

-------
                           Chapter 5

           Effect of Fuel Modification on Engine Wear

I.   introduction

     This chapter explores the relationship between  diesel  fuel
sulfur level  and engine wear, and  estimates  the  reductions  in
certain  operating  costs  that could  be  experienced  by  truck
owners and  operators using  low sulfur  fuel.   The  results  are
used  in  Chapter 7  to  estimate  the net  cost  effectiveness  of
sulfur controls.

     This chapter  is divided  into  two  parts.   The  first  part
estimates the amount of reduced engine wear to be  expected  with
a low  sulfur  fuel.   Both experimental  and in-use oil  analysis
data are discussed.   The second part discusses  the implications
of this  reduced  wear  on lubricating oil composition, oil change
interval, engine rebuild interval and vehicle life.   Two likely
engine wear  benefit scenarios are  selected,  and  reductions  in
certain  operating  costs are  estimated and  compared for  these
scenarios.

II.   Effect of Fuel Sulfur on Engine Wear

     A.    Background

     The  ERC  Study  investigated  the  effect  of  diesel  fuel
sulfur  content  on  engine  wear.[l]   ERC  concluded   that  a
reduction  of  diesel  fuel  sulfur  content  from  0.27   to  0.05
weight percent would  result  in a 30 to 40 percent reduction  in
engine wear and  therefore a  30 to 40 percent  increase in engine
life  and oil  drain  interval.   These  conclusions  were  based
primarily on  the results of a study by Tennyson and Parker, [2]
which was conducted on  a two-cycle  locomotive engine.  However,
today's  engine  oils  have  additive packages which  are better
able  to  handle the corrosive  wear  from  fuel  sulfur,  and  some
uncertainty  existed  in  extrapolating  the  locomotive  engine
results to on-road diesel operations.   EPA therefore contracted
with  Southwest Research, Inc.  to  further  study the effect  of
diesel fuel sulfur level in the range of  0.30 weight percent to
0.05  weight  percent  on  wear  in  on-road  heavy-duty  diesel
engines. [3]    A   two   part   approach   was   used   in   this
investigation.   The  first  part  was  to  conduct  a  literature
search to uncover  more data  than  that  contained  in   the  ERC
analysis.  The second part  was an empirical check on the first
in which Southwest  contacted diesel fleets currently operating
on   0.05  weight  percent  sulfur  fuel   to  obtain  used  oil
analyses.   This  effort  was  concentrated  in   the   Southern
California area,  which has undergone  a legislated reduction in
diesel  fuel   sulfur  content  to   0.05  weight  percent   maximum.
Comparison  was  made  of  used  oil  analyses  prior  to sulfur

-------
                              5-2


reduction with  those  obtained after the  reduction  to  determine
the effect of fuel sulfur reduction on engine wear.

     In  addition  to contracting  the wear  study to  Southwest,
EPA received comments on the wear benefits presented in the ERG
report  from  oil  companies  and  engine  manufacturers.   Some
companies also  submitted in-house data which characterized the
wear   rates   of  current   engines  with  current  fuels   and
lubricants.   Most of these  data  and comments were also provided
by IPA to Southwest to  assist them  in  their analysis/  and are
discussed in this chapter where appropriate.

     B.    Experimental Results

     The   literature   search   done  by   Southwest  was   more
comprehensive than  any  previously undertaken  in  the   area  of
sulfur  and  engine  wear.   However,  much of  the data  relating
sulfur to engine  wear was old (1940s and 50s)  and, due to the
difference  in  oil  qualities,  past  and  present,  cannot  be
applied quantifiably to  current engines.  Another limitation of
the older literature data is that the impact on engine wear was
addressed under carefully  controlled  steady  state conditions,
and not  more representative transient conditions.   Most of the
studies  however,  showed that  operating  on   low  sulfur  fuels
resulted in less piston ring and liner wear.

     There was  one recent  comprehensive wear  study that  used
current  engine  lubricating  oils  and fuels.    This was  conducted
by  Daimler  Benz,  and   has  been  recently  reported  in  the
literature.[4]   Wear  testing  was performed  on   a  Daimler  Benz
OM422  engine (4  stroke, -DI,  naturally   aspirated,  14.6 liter)
operating  on  API  CC  20  W  20  oil.    The  oil  was  changed
frequently to retain  resolution  of the experimental measurement
technique  and therefore  there was  no  significant Total  Base
Number   (TBN-a   measure   of   the   oil's    alkalinity  value)
depletion.  The effects of  fuels with sulfur  levels of  0.26 and
0.05  weight  percent  were  investigated over  a  range  of engine
operating conditions  (load, speed,  temperature).  Reduced wear
due to low sulfur  fuel  was observed during  conditions typical
of cold  start and warm-up conditions.  For  instance, at coolant
temperatures of 122°F,  an 80 percent reduction  in  bore  wear was
observed.   At  158°F,  the  reduction  was  28 percent,  and  at
176°F,  there was  no  difference  in the  observed rates of wear
for  the  different  fuels.    At   higher   temperatures,   the  low
sulfur fuel showed a slight increase in wear.

     Several experimental studies also  demonstrated that,  for  a
given  fuel  sulfur level,  engine  wear is sensitive to  oil TEN
value   [5,6,7].   It  has  been  recognized  that,  when certain
alkaline compounds  are  added to  lubricating  oil,  these help to
neutralize  the  sulfuric acid produced   in the  engine, thereby
reducing corrosive wear.   However,  the  alkalinity of  an oil
drops  with  use  due  to  the  neutralizing   process.   At low

-------
                              5-3


alkalinity values, wear can  increase  rapidly.   Because of this,
engine  manufacturers  generally  specify   a   minimum  new  oil
alkalinity value  and  oil change  interval which  depend  on  an
expected in-use fuel sulfur level.   This is done to  ensure that
oil TBN does not  drop  to  such a level that excessive wear would
occur.   However,   oil   changes,   as   will   be   subseguently
discussed,  are expensive  for  truck  owners  and operators.  Truck
owners  and  operators  who   attempt   to  stretch   oil  change
intervals may  experience  more wear on current sulfur fuel than
they would on a lower sulfur fuel.

     Experimental data  are useful   for  identifying  the general
causes and mechanisms  of  engine wear -  that:  1) wear increases
with sulfur level,  2)  wear generally decreases with increasing
temperature  (although  the  Daimler-Benz data  seem  to  indicate
that  at  high  operating temperature wear may start  to  increase
somewhat),  and 3)  wear  increases  at  higher loads.   However,
in-use engines are  operated  under  a variety of conditions that
are difficult  if not  impossible to  simulate in  a  laboratory.
In an attempt  to  get a better  idea of  in-use wear  levels with
different  fuels,  Southwest   analyzed  oil  samples  from  fleets
operating in southern California before  and after  sulfur levels
were reduced.

     C.     Used Oil  Analyses

     On January  1,  1985,  the sulfur content of diesel  fuel was
reduced by regulation to  0.05 weight  percent  maximum throughout
the   South  Coast  Air   Basin,  namely  Los   Angeles,   Orange,
Riverside   and  San Berandino  Counties  in  California.   This
change provided  an  opportunity to  gather  fleet  data indicating
whether  or not   lower sulfur  diesel  fuel  may  result  in  a
reduction in  engine wear.  Data on diesel  fuel  and  used engine
oil  analyses   were  sought  from fleets  operating  in  the area
before   and   after   the    implementation   of   0.05   sulfur
legislation.    The oil  analyses  were performed by  the fleets at
the  time of  oil change.  Data  collected  varied  between  the
fleets,  but all four analyzed for iron content  and reported the
mileage  since  the  last oil  change of  each  sample.  Southwest
selected  the  largest  fleet,  the  Southern  California  Rapid
Transit  District  (a bus  fleet  with  mostly two-stroke Detroit
Diesel engines)/  and grouped the data by calendar  year, engine
type  and mileage interval, looking for  a  trend  in iron content
and  sample  mileage.   No  consistent  pattern of increasing oil
iron  content  with mileage interval was  noted  across all engine
types,  and  calendar years.   However, there  did  seem to  be  a
somewhat consistent trend  of  increasing  oil  iron  content vs.
mileage  in 1984  and 1985  for all engine types  in  the two lower
mileage  intervals (1-9,000 miles and 9-15,000 miles).  The lack
of  correlation at  higher  mileage intervals  could  be  due to
significantly  smaller   sample  sizes  in  those  intervals, which
were  not  reported.   Another  cause  could be that  those engines
that  go longer   intervals  between oil  changes  might  be those

-------
                              5-4


that  also  burn more  oil, and  require  more frequent  fresh oil
additions,  thereby diluting  the concentration of wear products
in the  oil.   Because of the generally  wealc correlation between
iron content and mileage, Southwest grouped  all  samples in each
fleet  by calendar  year  (1984,  1985 and  1986), and  estimated
average mileages and iron content.  Statistical  tests  were used
to note  any significant differences  in iron content  or sample
mileage from 1984 to 1985 and 1986.

     A summary of the  reliable  iron content data from the four
fleets  is  shown  in  Table 5-1.   The  data  are disaggregated by
engine manufacturer  and  calendar  year.   Seven  engine  samples
are  shown  from three  fleets.   The percent reductions  in iron
content  in  1985  and  1986 were both  estimated  from  the  1984
pre-sulfur  control   iron content  values.    The  Detroit  Diesel
engines from the SCRTD bus fleet are the largest samples in the
group.

     There  were  judged  to  be  seven  unreliable engine  samples
from  the  four  fleets  (not shown  in Table 5-1).  Two samples,
the   Chandler   Cummins  and  Rental  GMC,  were  statistically
significant but had very low numbers of oil  samples  in one year
as  compared  to  another.   The  trends in iron  content  were
dissimilar;   the  Chandler Cummins experienced  a  120  percent
increase  in iron content from 1984  to 1986,  while the Rental
GMC experienced a 72 percent decrease over the same period.

     There  were  also a  number  of  non-significantly  different
engine  samples  from  each  fleet.   The   SCRTD  fleet had  one
(Cummins  engines),   the  Chandler  fleet  had  3  (Caterpillar,
Detroit  Diesel  and  Mack)  and  the Rental  fleet had 2 (Cummins
and Duetz).   In most of these instances where  the difference in
iron  content was  insignificant, there was  either a small sample
size  in  one year  compared to  another,  or  little  difference in
oil  iron content  from year-to-year, or both.  Also,  there were
no predominant trends  in iron  content from  year-to-year, in the
insignificant  samples;  in some cases  iron content dropped in
the years with sulfur control and  in other cases it increased.

      The  iron content  data  in  Table  5-1  show a  mixture of
results.   The  SCRTD  DDA-71 two-stroke   engines  show  a   19-23
percent  reduction  in  1985  and  1986  over 1984.   The  DDA-92
engines  show  about  a  38  percent  reduction.   The MAN  866
4-stroke engines  show  a 67 percent increase in  iron  content in
1985  over 1984.   DDA engines from  the Rental fleet show a  26-34
percent  reduction,  similar to  the SCRTD fleet.   The  Rental IHC
4-stroke engines  show a  25-31  percent reduction which is  also
similar  to  the  DDA 2-stroke  engines  in  the  Rental  and  SCRTD
fleets.  The Cummins and DDA samples  in  the Laidlaw  fleet  show
increases in iron content of between 26 and 85  percent in  1986
over  1984.

-------
                 5-5
             Table  5-1
Samples with Significant Differences
        in Oil Iron Content
-


Iron

Content (ppm) Standard
Fleet Enqine Type
SCRTD DDA71 2-stroke


DDA92 2-stroke


MAN 866 4 -stroke

Rental DDA 2-stroke


IHC 4-stroke


Laidlaw Cummins 4-stroke

DDA 2-stroke

Year
1984
1985
1986
1984
1985
1986
1984
1985
1984
1985
1986
1984
1985
1986
1984
1986
1984
1986
No.
2054
5111
4836
458
4387
5551
59
67
126
162
200
90
114
108
25
60
33
63
Mean
82.1
63.6
66.2
82.1
50.1
. 51.6
78.3
130.9
85.3
62.9
56.4
72.0
53.8
49.3
32.9
41.4
31.0
57.4
Deviation
65.7
44.7
46.3
65.7
50.7
40.6
51.3
153.3
47.7
33.2
24.2
36.4
26.1
25.8
20.6
30.4
20.2
74.7
Percent
Iron
Reduction
(8x to 84)

23%
19%

39%
37%

-67%

26%
34%

25%
31%

-26%

-85%

-------
                              5-6


     Overall,  .with  the  exception  of  the  small  Laidlaw  DDA
sample, the  two-stroke engines  show a  consistent decrease  in
wear  with  low  sulfur  fuel.    The  four  stroke  engines  show
increases and  decreases  in wear,  but  the largest  sample (IHC)
shows  a  wear  decrease similar  to  the  2-stroke  engines.   The
other  two  samples  showing increases  in  wear,  especially  the
Laidlaw Cummins sample, could arguably  be considered either too
small  or  imbalanced (in number  of measurements in  1984  versus
1986) to make a valid comparison.

     It should be  pointed  out that not all operating parameters
were constant during  this  three year period.   The new oil  TBN
used in  the  SCRTD  fleet  increased from 5.4  in 1984 to  7.4  in
1985 and  6.1  in 1986  (oil viscosity remained  the same).   New
oil  TBN  values for the other  fleets  were not  available.   The
oil  iron values for DDA  engines in  the SCRTD fleet  show lower
oil/iron contents  in  1985  than  1986, which follows the trend of
oil  TBN levels.  However,  the  differences in percent reduction
in  oil  iron   content  from  1984  to either  1985  or 1986  are
minimal,  and so the 6.1  and 7.4 TBN oils  are assumed to  result
in  equivalent  wear  on a  low sulfur fuel.  Also,  the Chandler
and  Rental Truck  fleets experienced  a  significant  increase  in
the  levels  of  zinc present in  the  oil in 1985-86  compared to
1984.  Zinc  dithiophosphate (ZDTP)  is well  established  as  an
inhibitor of  abrasive wear,  and  it is therefore difficult to
determine the  relative contributions of  the  zinc  increase  and
sulfur decrease  to  the  reduction  in wear  observed.   Southwest
did  conclude  that  the   zinc   increases   did   not  contribute
significantly to the reduction  in  wear, but this conclusion was
not  well  supported and therefore the data from  the  SCRTD fleet
should be considered preferentially.

     Although  there  is  some  uncertainty  with  some  of  the
samples,  and it would be better to have more oil  analysis data
on  four-stroke  engines operating  under  line-haul  conditions on
current and low sulfur fuels, the  available   oil  analysis data
indicate  that  a   significant  decrease  in  oil  iron  content
occurred with  low  sulfur  fuel  for the majority  of the samples.
And, in  spite  of  the  problems  with the  oil  analysis data, the
primary advantages  of  this data over the  experimental  data are
that it was gathered under in-use operating conditions and that
it measures total  iron worn in the engine due to all causes.

     The  size  of  the  wear  decrease  among   samples  showing
decrease  in wear  appears to   be  in  the range  of  20   to  40
percent.  A  value  of  30  percent,  therefore,  is  reasonable to
expect from all diesel engines, and will be used  in  all further
analyses.   Adjustment  to  this value  is  necessary,  however,
because  the  pre-  and post-control  sulfur  levels  in California
were  0.35  weight  percent  and  0.03  percent,  respectively,
whereas   0.25  weight  percent   and  0.05  weight  percent  were
defined  in chapter 2 to  be representative  levels  to use for
this analysis.

-------
                              5-7


     In a preliminary wear  analysis  which EPA released to  peer
reviewers it  was assumed  that  the  difference in  above  sulfur
levels would  result in  a  proportional  reduction  in  the  wear
benefit.    In  their   comments  on   this  analysis,  the  Engine
Manufacturers  Association   and  the  Motor Vehicle  Manufactures
Association  stated   that  experimental  evidence  suggested  the
effect was  non-linear,  and  that  this   should  be  taken  into
account. [8]    EMA   and   NMVA  suggested  using   a   wear   model
developed by Daimler Benz to  do this.[4]   However, there  are  a
number of  inputs  to this  model  whose  in-use  values are  not
known.   Therefore,  this  analysis  will   continue  to  assume  a
proportional reduction  in  wear  benefit.  The size of the  wear
benefit thus  adjusted  is 18  percent and this is  used  in  all
further analyses.

III. Effect of Reduced Wear on Operating Costs

     In the  previous section it was  shown by both experimental
and  in-use  oil  analysis data  that  'lower  diesel  fuel  sulfur
levels will  likely  lead  to  less  diesel engine  wear.   Exactly
how  this will  translate  into  benefits  to truck  owners  and
operators,   however,  is  difficult  to  predict.   There  are  a
number of possible  outcomes.   First, lubrication  oil  producers
and  engine  manufacturers could  determine that  since there is
less  sulfur  in diesel fuel,  lube  oils could be produced  with
lower  concentrations  of  anti-corrosive  agents.   This  could
reduce the cost  of  lube oil,  thereby reducing oil  change costs
to  truck owners  and operators.   The  second  possible outcome is
that  engine manufacturers  could  recommend   longer  oil  change
intervals,   or  truck  owners  and  operators  could  decide  to
stretch  oil  change intervals.   According  to  a  rebuild  and
engine  maintenance  survey  performed  by  EMA,[9]  hereinafter
called the EMA Rebuild Survey, heavy heavy-duty  truck operators
schedule oil  changes at about  11,000 mile  intervals.  A truck
lasting  500,000  miles  would  therefore  experience about  45  oil
changes.   These  are expensive for  the truck owner and operator,
since both the labor and materials'   cost of the oil  change as
well  as  truck  downtime  are  involved  (this  latter  cost  is
probably minimized  by scheduling  other  maintenance  along  with
oil  changes  and  by performing oil changes during natural truck
downtime).    Nonetheless,  truck owners   and operators  have  a
significant  incentive  to increase oil  change intervals  if they
perceive it  will not adversely  affect engine wear.   The third
possible outcome is  that  reduced  engine  wear  could lead to
increased  mileage  to  rebuild  and  possibly increased  vehicle
life.

     Any of  these  outcomes seems  possible.   Furthermore/ it is
possible that there  could be  combinations  of  these outcomes.
The  ERG  report   assumed that  there would  be  an  increase in
engine rebuild interval,  engine life, and oil  change interval.
The   increase  in   engine   life,  however, was not assumed to
increase vehicle life, nor  were oil  costs assumed to be  lower.

-------
                               5-8


None of the work  done by Southwest revealed any  information  on
which combinations of these would be most likely.

     Comments  received by EPA on  the  ERG report  were  varied.
The  engine  manufacturing  industry generally  acknowledged  that
one  or  more of these outcomes was  likely.  Caterpillar  Co.,  in
their  comments  on  the  ERG  report,  stated  that  "the  most
attractive prospect could  be  that  the combination of low sulfur
fuel and reduced particulate  emissions would  result in extended
oil  change  periods."[10]   Caterpillar indicates  here that  soot
accumulation in oil  also  has  an effect on oil  change interval,
i.e., truck  owners and operators  schedule oil  changes based  on
oil  appearance  or ash content.  The "reduced  particulates"  in
Caterpillar's  comments  refers to  future engines  designed  to
meet  the  1991  and   1994  particulate   standards,  which   are
expected to  have  engine-out  particulate  levels that  are 60  to
75   percent   less   than   current   (pre-1988)   levels.    The
implication is that  an increase in oil change interval would be
possible for  future engines,   but  perhaps not  current engines.
Caterpillar,  Co.   also  submitted  some  information showing  the
difference  in  the  cost  of lube  oil at  different TBN  levels.
These costs  were  based  on different types  of oils  ordered  by
their  test  facilities,   and   not   on a  complete  analysis  of
additive components  that would be  required or  desirable with  a
low sulfur fuel.

     API  submitted  information about  the  primary  causes  of
rebuilds in  support of a position  that  any  reduced engine  wear
due  to  low  sulfur fuel  would not  affect  rebuild  interval  or
vehicle  life.[11]   Their  comments were  focused  in  two areas.
The first comment cited the 1981 EMA Rebuild  Survey which found
that  engines  are  usually   rebuilt  for  either   loss  of  oil
control, or  for mechanical failures such as  bearing, camshaft
and  fuel  injector failures.   API  concluded  that trucks in the
second  category would  not experience  any  increase  in engine
rebuild  interval  with  low sulfur  fuel.   EPA  agrees with  this
conclusion   since    these   failures   appear   unrelated   to
corrosivering  and  liner  wear.  The.second comment  was that loss
of  oil  control was the  result of  factors  also  unrelated  to
corrosive  wear,  such as  abrasive  wear  from  top  land piston
deposits,  piston  ring  scuffing and  broken  piston rings.   API
cited two  experimental  test   programs  which  tested  engines  on
several  different   fuels  (of  varying   sulfur   content)   and
lubricants  to  support  this.[12,13]   The test  programs  also
showed,  however,  that  sulfur, while it does  not  lead to the
formation  of  top  land  deposit  it  does  form deposits  in the
bottom  land  and   bottom  ring  groove.    Also,    the   programs
demonstrated  that  corrosive wear  could cause  increased  oil
consumption if oil  alkalinity values  were too  low.

     The  best way  to resolve this  issue would be  to examine
in-use  oil consumption data  on two- and four-stroke  engines on
current and low sulfur fuels.  Unfortunately,  this data was not

-------
                               5-9


available from  the  Southern California  fleets.   The fleet  oil
analysis .data,  did,  however,  show  a difference in  wear  due to
fuel sulfur level.  The iron contents of  samples with both high
and low sulfur  fuels  would have included wear products  from all
causes of wear (abrasive or corrosive),  and in  all parts  of the
engine.  Southwest did  say that 80  percent of the wear  products
in  the oil would come  from the ring  and liner.   If  this  is
true,  then  the  difference  in  wear  between high and low  sulfur
fuels  in the  fleet  oil  analysis data would be  expected to lead
to differences in oil consumption.

     Overall,  both  the experimental  and  in-use  oil  analysis
data indicate that  engine  wear is  lower with low  sulfur  fuels.
However, there( is  still some  uncertainty  as to how this will
benefit truck owners and  operators.   Therefore, two  scenarios
have  been  chosen that  are  believed to  set  upper and  lower
bounds to the size of this  benefit.   The  first  scenario  assumes
there  will  be a  reduction in oil cost  and an increase  in oil
change  interval,  with no  increase  in engine rebuild  interval,
engine  life,  or  vehicle  life.   The basic hypothesis  of this
scenario is that  wear testing and/or experience on  the part of
both engine manufactures and  lubricating oil produces  with the
lower  sulfur  fuels  will result in  the  oil  producers  reducing
the   TBN   of  lubricating  oils   and  engine   manufacturers
recommending  longer  oil change  intervals  for  their  engines.
Since  comments  from at  least one  manufacturer  indicated oil
change   interval   to   some   extent  depends   on   engine-out
particulate levels,  this benefit will be assumed  to apply only
to  cars  and trucks  with low particulate levels  (this will  be
discussed in  the  next section).  It is further assumed that oil
change  interval is  not  limited in the  range  it is  extended by
the degradation of other oil parameters such as viscosity.

     The second scenario assumes  there  will be no change  in oil
cost or oil change  interval, but there  will  be an  increase in
engine  rebuild  interval  for those engines which are rebuilt for
reasons  related  to  high  oil consumption.    Two   cases  are
examined  in this  scenario which are derived  from  assumptions
about  why  trucks  are scrapped.   The first case  assumes   trucks
are  always  scrapped due to  engine problems.   In this  case, the
extension  in  engine  rebuild  interval leads to  an  extension in
engine  and  vehicle  life.  The second case  assumes that there is
no  correlation  between  engine rebuild interval   and  vehicle
life,  i.e.,  that  some  or  many trucks are scrapped because of
reasons  unrelated  to the increase  in  rebuild  interval  (for
example, transmission or  rear  axle failure).  The benefits for
this  case,  then,  are an  extension  in  rebuild interval  and  a
decrease  in the total number  of rebuilds needed for the  in-use
fleet  of trucks.

     Operating   cost   reductions    for   these   scenarios  are
developed below.

-------
                              5-10


     A.    Oil Change Cost Reduction

     This  scenario  assumes  that   lubricating  oil  producers
reduce oil alkalinity values  and engine manufacturers recommend
longer oil change intervals to yield wear  that  is eguivalent to
the  wear  experienced with current  sulfur fuels.   The  vehicle
model  year  applicability  of  these  benefits  will  be discussed
first.   Next,  a  discussion  of  the   potential  for  oil  cost
reduction with a  reduced  alkalinity content will be presented.
Finally,  this section will  be  concluded  with a discussion of
the  procedures  for  estimating the  reduction in  the number  of
oil  changes and  a presentation  of the operating cost reductions
using these methods.

     Comments from  one  engine manufactuer have  indicated  that
increased oil change  intervals might be experienced on vehicles
with  low  engine out particulate levels.   The  task here is  to
decide  for  this  analysis what  level  is  low  enough for  each
vehicle type.  Engine out particulate  levels have  been  reduced
by   the   engine   manufactures  in  order  to  comply  with  the
particulate  emission  standards,   so   these  can  be  used  to
estimate  the model  year  applicability of  this benefit.   The
particulate emission standards  for the different vehicle types
are  shown in Table  5-2.  For  LDDVs  and LDDTs,  the  emission
standard  in  1988 remarks  a  60-70% reduction  in particulata
levels.   For  heavy-duty  vehicles,  the most  stringent  standard
is  in 1994,  but  the  1991 standard  respresents about a  60-70%
decrease  from   pre-1988   levels.   For   buses,   the  emission
standard  in 1991  represents almost  a 90%  reduction  from current
levels.   This  analysis  will  assume that  the  increase  in  oil
interval will apply to 1988 and  later LDDVs  and LDDTs,  and 1991
and  later HDVs and buses.

      It  may  actually  (for  example, 1996  or   1997)  be  several
years  beyond  the implementation  of  sulfur  controls that  engine
manufacturer  decide  they can   recommend  increased  oil  change
intervals for their new engines.  However, this recommendation,
brought  about  by lower  sulfur  fuel which older  engines  also
would be using,  would  be  expected to   influence  oil  change
intervals on  the existing low  emitting 1991 and late HDVs and
buses, and 1988 and later  LDVs and  LDTs) fleet.

      Therefore,  this  benefit will  be   assured  to  apply  to all
1988 and later  LDDVs and LDDTs,   and  1991 and  later  HDVs and
buses  in  all   calander   years  after  the   start  of  sulfur
controls.   The  oil change intervals,  lifetime mileages and oil
change costs for all vehicles  are  shown  in  Table  5-3.   Oil
change  intervals  for   LDDVs,   LDDTs  and   LHDVs  came   from
conversations  with automobile  dealers, and  the  intervals  for
medium heavy and heavy  heavy-duty trucks were  taken  from the
EMA  Rebuild Survey. [9].  The  bus oil change  interval came  from
the  in-use SCRTD bus oil analysis data  in  the Southwest Wear
Report.[3]   Lifetime  mileages are  from the  Diesel  Particulate

-------
                              5-11
                           Table  5-2
                      Particulate Emission
                      	Standards	
Vehicle Type
LDDV & LDDT

LHDV - HHDVs
Buses
MYR Group
1982-87
1988+
1988-90
1991-93
1994+
1988-90
1991+
Particulate Standard
0.6 gpm
0.2 gpm
0.6 g/BHP-hr
0.25 g/BHP-hr
0.10 g/BHP-hr
0.6 g/BHP-hr
0.1 g/BHP-hr

-------
                                      5-12

                                    Table 5-3
                              Input Parameters for
                    Estimating Reductions in Oil  Change Costs
                           LDDVS    LDDTS    LHDVs    MHDVs    dtJCSk    Buses
Oil Change Interval (Mi)   5,000    5,000    5,000     11.000    12,000   6.000
Lifetime Mileage (Mi)      100.000  124.000  128,000   268.000   529.000  540,000

Lifetime (yrs)             7.2      10.5     8.0       8.1       8.2      12.0
Oil Change Cost ($)        30.00    30.00    30.00     100.00    100.00   100.00

Oil Changes Per Year       2.78     2.36     3.20      3.0       5.4      7.5

Annual Cost                83.40    70.80    96.00     300.00    540.00   750.00

-------
                              5-13


Study.[14]   Oil  change costs  were  determined  by  contacting
several service shops in the Detroit area.

     When  lifetime  mileage  is  divided by  oil change  interval
and vehicle  life, the  result is the  number  of oil  changes  per
year,  which is  also  shown  in Table 5-3.   The  number of  oil
changes   ranges   from   2.36  for   LDDTs   to  7.5   for  buses.
Multiplying  annual  oil  changes by oil  change  cost gives  the
annual oil change cost, shown in the bottom row.

     Information  supplied  by  Caterpillar  Co.  indicated  that
there is  about  a  1.3  percent decrease in the wholesale price of
lubricating oil for every unit of reduction  in TBN value.   This
was based on the wholesale  cost  of  oil  at  two TBN  levels  - a
TBN  of  13  ($2.01  per gallon),  and  a  TBN  of   7  ($1.86  per
gallon).   This analysis  will  assume an equivalent  percentage
change  in  the  retail  price  of  oil  over  this  TBN  range.
Additional   items   needed   to   estimate   the    reduction   in
lubrication oil cost are  the current TBN level of  oils used in
trucks  and the expected  oil TBN  level  when  using  low sulfur
fuel.  The  ERG  report  stated that new oil TBN values typically
are  in  the range of  6-10.   Engine manufacturers  commenting on
the  report concurred,   and data on new  oil  used  in  the  1984
SCRTD  fleet in the wear  report indicate new  oil  TBN values of
5.4 to  7.4.  Therefore,  this analysis will  use  a  TBN  value of
7.0 for current sulfur fuel.

     It  is  difficult  to  predict  how  low  oil producers  would
reduce TBN with low sulfur  fuel.  The engine manufacturers have
widely  varying  specifications  for  new  oil  TBN.[3]   However,
Caterpillar  Co.  recommends  a new  oil TBN  level  of  about  20
times  the  sulfur  level.   A 0.05  weight  percent  sulfur  fuel
would thus  yield  an oil TBN value of  1.0.   It is  doubtful that
an  engine manufacturer would recommend  such  a low TBN level,
even with low sulfur fuel, and so for  this  analysis a TBN value
of  3.0 will be used.  As  will be be presently  demonstrated,
overall  oil  change  costs  are not at  all  sensitive  to  this TBN
level given the TBN price relationship described above.

     The  final piece  of  information needed  is  a  typical oil
crankcase capacity.   These  can  also vary   considerably  from
manufacturer to manufacturer,  and  from  engine to engine.   The
1980  edition of Motor's  Truck  and  Diesel  Repair  manual  lists
the  crankcase  capacities  of Cummins engines in   the  range of
28-36  quarts.[15]  For  Mack  engines,  the capacity ranges from
15-20  quarts and for  Detroit Diesel  engines  ranges between 20
and 30 quarts.  This analysis will  use 24 quarts  (6 gallons) as
a typical volume of oil needed  for  an oil change.

     The  savings  in new  oil cost  obtained  by using this  input
data  for a 3.0 TBN oil is  about  $1.50 per oil change.  At TBN
levels   of   2.0  and   4.0  the  savings   are   $1.00  and $2.00,
respectively.   When  these  savings  are  compared  against  the

-------
                              5-14


total  oil  and filter  change cost  of $100  (Table  5-2)  it  is
clear  that oil TEN  reduction appears to have a  small  impact  on
oil  change  price,  and  that overall  oil change  cost  is  not
sensitive to  TBN  level.   Therefore,  oil  cost savings  of  $1.50
(at  a  TBN  level  of  3.0)  will  be  used  for   all  further
calculations.

     To conclude the  oil  change cost analysis,  the extension in
oil  change  interval  with low  sulfur  fuel must be  estimated.
The  hypothesis  used  in  this   analysis   is  that  oil  change
intervals can be  increased  and oil  cost decreased  to  yield
equivalent wear  to  a current  sulfur  fuel.   The  conclusion  of
the  Southwest oil   sample   analysis  section  was   that  diesel
vehicles  would experience  18  percent  less  wear  with the  low
sulfur  fuel  (.05  weight  percent)  than  with  the current  sulfur
fuel (.25 weight  percent).   It  is therefore  assumed that trucks
and buses could travel 18 percent further between oil changes.

     The  adjusted  oil  change   intervals  and  numbers  of  oil
changes per  year are shown in  Table  5-4.   Also  shown  are  the
adjusted annual costs, the  net  savings per year and discounted
lifetime  cost  reductions.   The  net   savings  per  year  are
estimated by  subtracting adjusted  annual  cost  as  presented  in
Table  5-4 from the  annual oil change cost as presented in Table
5-3.  These are further discounted  over the vehicle life in the
last row.  The results show discounted lifetime cost reductions
of $35  to $45 for LDDVs, LDDTs  and LHDVs, and  $300 to $900  for
the  heavy duty vehicles  and buses.  On a fuel consumption basis
(taking   undiscounted   lifetime   costs    and   dividing   by
undiscounted  lifetime  miles,  then multiplying by model  year
2000 fuel economies  as listed in Table 5-4 [16,17])  this  works
out  to savings  of  between  0.92 cents per  gallon  (HHDVs)  and
1.75 cents  per gallon (buses).   These benefits are  assumed to
apply  only to 1988  and later LDDVs and LDDTs and 1991 and later
heavy-duty duty trucks and buses.

     In  this  example  the most  uncertain input parameters  are
the  oil  cost  reduction  and  increased   oil  change  interval.
However, only 6 to 8 percent  of  the benefit  (depending on truck
type)  is due. to the  oil cost   reduction.  The  remaining 92-94
percent  is  due  to  the reduction in  oil  change  frequency.
Therefore/ the size of this  benefit  is  directly related to the
extension in  oil change  interval—if oil change  intervals  can
be   extended  further  than  18  percent,   the  benefit will  be
larger.   If  oil  change  intervals  are  extended  less  than  18
percent,  the  benefit  will  be proportionally smaller.   At this
juncture  there is  no technical  basis for  either a  higher or
lower  percentage  oil  change  interval extension.

-------
                                       5-15

                                     Table  5-4
                           Discounted Lifetime Coil for
                           Oil Chance Interval Scenario
                                              LDQTs  LhpTs   hHD     HHDT    Bus
Adjusted Oil Change Interval (mi)      5.410  5.410  5,4)0  12.980  14.160  7.080
Adjusted Oil Changes Per Year          2.56   2.18   2.95   2.5     4.6     6.4

Adjusted Annual Cost ($)               76.60  65.40  88.50  246.25  453.10  630.40

Net Savings Per Year (()               6.60   5.40   7.50   53.75   86.90   119.60
   (undiscounted)
Discounted Lifetime Cost Reduction (() 37.02  36.78  45.46  326.00  526.00  895.00

Cost-Reduction Per Gallon" <*/gal)     1.56   1.18   0.75   1.32    0.92    1.75
     Fuel economies used

     LOOV:  33 mpg
     LOOT:  26 mpg
     LHOV:  16 mpg
     HHOV:  8.1 mpg
     HHOV:  6.8 mpg
     Bus:   6.6 mpg

-------
                              5-16


     B.    Extension in Engine Rebuild Interval/Vehicle Life

     This   scenario   assumes  there   is  no   change  in   oil
composition  (or  cost)  or  oil  change  interval,   but  there  is
reduced  engine wear  which  leads to  increased engine  rebuild
intervals,  engine  life and vehicle  life in those  engines  that
fail   or   are  rebuilt   for  reasons   related  to   high   oil
consumption.   In this  scenario,  it  is necessary to  analyze all
diesel vehicles, not  just those  with engines  that are rebuilt,
since reduced engine wear could affect vehicle  life in vehicles
whose engines are not rebuilt also.

     Although  LDDV,  LDDT  and  LHDV  engines  are  occasionally
rebuilt  where the  condition of  the  rest  of  the  vehicle  is
excellent,  these  situations  are  believed to  be  rare  and  this
analysis  assumes these engines are  never rebuilt.  Also,  some
MHDVs  with nonsleeved engines  are  not  rebuilt.C18]   These are
referred  to  in this  paper as MHDVls, and are  analyzed with the
LDDVs, LDDTs,  and  LHDVs.   (The  remainder of the MHDV class  will
be discussed in a later section.)

     The  question then becomes  how  increased engine  life could
affect vehicle life for  these lighter vehicle  classes that are
not  rebuilt.   Several sources were  contacted to  determine why
vehicles  are scrapped.   Little   information was  obtained which
could be  used  to quantify the effect of  engine  life on vehicle
life.   Therefore,   two  sensitivity   assumptions   are  possible.
The  first one  is that vehicle life is not  influenced  by engine
life  (i.e., all vehicles are  scrapped  for  non-engine related
reasons  such  as transmission,  body, etc.),  and  in  this  case
there  is  no  economic  benefit.   The  second  one  is that  an
extension in  engine   life  always  leads to  increased  vehicle
life.  In this case an economic  benefit exists in deferring the
purchase  of   a  replacement  vehicle.    The  actual  situation
probably  lies somewhere  between  these two  extremes,  and  since
the  engine  is  the single  most expensive  piece  of  equipment  in a
vehicle,  it  probably  lies  closer   to  the  second one,   since
vehicle   ownersare  likely  to  replace   or  repair  cheaper  car
components   (except  for  perhaps the transmission)   until the
engine fails.   However,   there  is currently no  technical  basis
for  estimating the  "actual"  situation  for  these  vehicle  types
(LDDVs, LDDTs, LHDTs,  and MHDVls),  and  so only the second case
is  estimated (i.e.,  extension in engine  life always leads  to an
extension in vehicle  life).

     The  reduction  in  engine  wear is  also expected to result in
longer  rebuild intervals for  rebuilt MHD,  HHD  and bus  engines
which  are rebuilt  for reasons related to high oil consumption.
There•is an economic  benefit, then,  in  rebuild costs which are
delayed.    Like   the   lighter   diesel  vehicles,   there  is
uncertainty  in  how   reduced  wear  will  affect   vehicle  life.
Therefore,  two cases  are examined for these vehicles also.

-------
                              5-17


     The first assumes that  an  extension in rebuild interval(s)
always extends vehicle life.  In this case an  economic  benefit
occurs not only  due  to delayed rebuilds, but also  to  a  delayed
replacement  purchase.    The  second  case   assumes  that   no
correlation exists between extended  engine  rebuild interval and
vehicle  life.   In this second  case  an economic  benefit  occurs
from  delayed  rebuilds, and  a  further  economic  benefit  occurs
from  some rebuilds  which   are  not  needed.    These  occur  in
engines  which  experience first  and  second rebuilds  at  high
odometer  values  with  a  current  sulfur  fuel,  but  only  one
rebuild  (at  an even higher  mileage)  with low  sulfur  fuel.   in
this case, the cost of a second rebuild  is  avoided.  Under this
approach, vehicle life is not extended.

     As  stated earlier, some MHDVs  have  nonsleeved engines that
are  not  rebuilt  by  truck owners   and  operators  (MHDVls).
However,  many  of  the non-sleeved   MHD  engines  are  rebuilt.
Also, many manufacturers  make  sleeved diesel  engines  for  MHDV
use,  and these  have  characteristics (mileage  accumulation vs.
age,) like HHDV  engines.   Therefore,  all three classes  of MHDV
engines  will  be  analyzed separately and weighted  together  at
the  conclusion of the analysis  to   yield  a single MHDV  wear
benefit.   For  convenience,   unsleeved  nonrebuilt  engines  are
referred to  in this  analysis as MHDVls  and unsleeved  rebuilt
engines  will  be  referred to as  MHDV2s.  The  rebuilt,  sleeved
MHDVs will be  assumed  to  have  the  same  benefits  as HHDVs,  and
thus will not receive a separate label.

     To  summarize,  the types of benefits for  which reductions
in  operating  costs   need  to   be  estimated  depend   on  the
correlation   between   engine   and   vehicle  life.    Where  an
extension  in  engine  life   also  extends   vehicle  life,  the
benefits  are  deferred replacement  vehicle cost  for all  diesel
vehicles, and  deferred rebuild  costs for rebuilt  MHDVs,  HHDVs
and  buses.  Where an extension in engine  life  does not  lead to
an  extension  in vehicle  life,  the  benefits  are  reductions  in
the  total  number of  rebuilds,  and a delay  in  when they occur.
These benefits apply  only to MHDVs,  HHDVs and  buses;  there are
no benefits for LDDVs, LDDTs  and LHDVs for this situation.

     Operating  costs  for  the situation  where  an  extension  in
engine  life   leads  to  an extension  in  vehicle  life  will  be
examined  first.   Base  case  (current   sulfur  fuel)  operating
costs  are examined  first,  followed  by  low   sulfur  operating
costs.

     l.    Reduction  in  Operating  Costs  -  Increased   Engine
           Rebuild Interval  and Vehicle Life

     Base  Case  -  Base  case  vehicle   replacement  costs  are
estimated  foe  all vehicles  first.   It  should be mentioned that
this  analysis  does  not  examine  all operating  costs  such  as
minor   maintenance,   fuel,   etc.,   but   only  those  that  are

-------
                              5-18
significantly  affected  by   an  increase   in   engine  rebuild
interval  and vehicle  life.   All  other  operating  costs  are
assumed to be the  same on  a per mile basis with current and low
sulfur fuel.

     To  estimate  the  base  vehicle   replacement   costs,   the
analysis  requires  estimates  of  new vehicle  cost   and  vehicle
lifetime (in- years and mileage).   Vehicle cost and  mileage  can
then be  discounted and  divided to yield  a  net  replacement cost
per mile for the base case.

     New vehicle cost/ lifetime  mileage,  and lifetime  in  years
for all vehicles used  in this analysis are  shown in Table 5-5.
New  vehicle   costs  were   obtained   by   contacting   several
Detroit-area  car  dealerships.   Very   few  manufacturers  are
offering diesel  LDVs for sale  in the U.S.  at  this  time  (Ford
and  Daimler-Benz),  so  this  cost  is  difficult  to  project.
Rather  than  rely  on  the  price  of  the  one  or   two  models
available,  a price  was  selected  that  may  be closer to  an
average   if   several   manufacturers   were  marketing   diesel
light-duty vehicles.   Costs  for  LDDTs,  LHDTs, MHDVs  and  HHDVs
were  determined   from   conversations   with  manufacturers  and
dealerships.   These  costs  can  vary greatly with  the  type  of
optional equipment chosen by the buyer, and  are not  meant  to be
any kind  of  industry  weighted  retail  average.  The  bus  price
was determined from conversations with transit managers.

     Lifetime mileages were taken from the  RIA  for  the 1988 and
later  heavy-duty  engine NOx and  particulate  standards.[19]
Vehicle  lifetimes  were  estimated   from  the  above  lifetime
mileages  and  the  MOBILE4  VMT vs.  age distributions  for  these
vehicle types.[20]   The ages  for  LDDVs,  LHDVs,  and  MHDVs appear
somewhat  low  in this analysis,  but  should  not have much  of an
effect on the results  of this analysis since the analysis will
be  focusing  on  the difference in lifetime and lifetime mileage
between fuel  scenarios,  rather  than on the  absolute  values of
these estimates.

     When vehicle  costs  are  discounted  over the  ages listed in
Table  5-4  with  a  factor   of   10   percent,   and   divided  by
discounted lifetime  mileage,  the result is  discounted base case
operating costs,  also shown  in  Table  5-5.  These  costs  range
from  6.57 cents  per  mile  for  LDDTs  to  18.66  cents/mile for
MHDVls.

     Next,  rebuild costs  must be developed for MHDV2s,  HHDVs
and buses.   The key  input  parameters are  rebuild  mileages and
ages, the fraction of  vehicles  receiving  rebuilds,  and rebuild
costs..   Rebuild mileages  and ages  can be  used  in discounting
mileage  and  rebuild  costs  to  determine  their   net  present
values.   The fractions  of  vehicles  that  receive  rebuilds are
used to  estimate a net fleet rebuild  cost,  since some  vehicles
may be scrapped before a particular  rebuild occurs.

-------
               5-19
            Table  5-5
Base Case Vehicle Replacement Cost
Input Parameters for All Vehicles
Parameter
Vehicle Costs (1987 $)






Lifetime Mileage (mi)






Lifetime (yrs. )






Discounted Vehicle
Replacement Cost (tf/mile)






Vehicle Type
LDDV
LDDT
LHDT
MHDV1
MHDV2
HHDV
Buses
LDDV
LDDT
LHDT
MHDV1
MHDV2
HHDV
Buses
LDDV
LDDT
LHDT
MHDV1
MHDV2
HHDV
buses

LDDV
LDDT
LHDT
MHDV1
MHDV2
HHDV
Buses
                                      Value

                                      12,000
                                      14,000
                                      20,000
                                      45,000
                                      45,000
                                      80,000
                                     150,000

                                     100,000
                                     124,000
                                     128,000
                                     200,000
                                     321,000
                                     529,000
                                     540,000

                                      7.2
                                     10.5
                                      8.0
                                      5.4
                                     10.8
                                      8.2
                                     12.0
                                       8.89
                                       6.57
                                      10.83
                                      18.66
                                       8.10
                                      10.50
                                      14.90

-------
                              5-20
     Rebuild mileages for MHDVs,  HHDVs  and buses were developed
from  survey  data  reported  by  a  publication  called  Fleet
Equipment  (formerly  Fleet Maintenance  and Specifying).[is]   A
portion  of the  data on rebuild  mileages for  sleeved  diesel
engines  is shown  in  Table 5-6.  Fleets  were requested by  the
publication to report engine mileage before and  after  the  first
major  overhaul.   Data were  disaggregated by  fleet type.   The
"Miles  Before Overhaul"   data were  used  for  the  mileage  to
first  overhaul,  and  the  "Miles After  Overhaul"  data  were  used
for  the mileage  between the first  and  second  overhaul.   The
survey  did not  presume  a  second  overhaul,   so  data  were  not
gathered on the mileage after a second  overhaul.   For  buses  the
bus  fleet  data from Table  5-6  were used  in  this  analysis;  for
HHDVs  the  non-bus weighted  averages  were  used.   The  Fleet
survey  listed  a mileage  to  first  rebuild for  MHDV2s of 175,000,
and a mileage after rebuild of 146,000  miles.

     The next  item that  needs to  be estimated to  determine  the
base  case  rebuild  costs  is  the  fraction   of  vehicles  that
receive  rebuilds.  At  this point,  estimates  have been  made of
average vehicle lifetime  mileage  and  rebuild  mileage.   However,
few  vehicles  are  actually  rebuilt  (or  scrapped)  at  these
average  mileages,   but   are   rebuilt   (or   scrapped)  over  a
distribution   of  mileages   centered   about   these   average
mileages.   Therefore,  it  is possible  and  probable  that  some
HHDVs and buses are scrapped before the  second rebuild, or that
some  MHDVs are  scrapped before  the  first rebuild.   For  these
vehicles, there is no rebuild cost to estimate.

     A  standard  Monte  Carlo  analysis  was used  to  predict  the
fraction of  trucks receiving rebuilds.[21]   In  this technique,
distributions  are  created which are centered  about  the average
rebuild  and  scrappage  mileage.    A  finite truck sample  is  then
created, and for each truck in  the  sample, the analysis picks a
scrappage  mileage and  rebuild  mileages.  This  is accomplished
with  the   average rebuild and  scrappage mileage,  the estimated
sample  coefficient of variation  (COV),  and  a  table  of random
normal deviates.   Scrappage and rebuild mileages are calculated
for  each truck by multiplying  the random normal  deviates,  the
COV  and the  sample average.  A new random deviate  is selected
each  time  a  new  rebuild  mileage  or  scrappage  mileage  is
estimated.   The scrappage  and  rebuild  mileage  for  each  truck
are  then compared to see if  the  scrappage mileage is less than
the  rebuild mileage.   If so,  then there  is no rebuild for that
truck.   If the  scrappage mileage  is  greater than  the rebuild
mileage, then  a rebuild  is assumed to have taken place.

     The  use  of this  technique  requires  estimates  of  the COV
for  the scrappage and  rebuild mileages.  The  only data found on
how  scrappage  mileages  are distributed,  were  some  data  on
rebuild interval  distributions  in  the  EMA  Rebuild  Survey.
Truck  owners  were asked to  recall the mileage  at  which the
first  rebuild was performed.   A  portion  of  the data for Class

-------
                              5-21

                            Table  5-6

              Rebuild Mileages for Sleeved Engines

Category
Private Fleet
Contract/
Common Carrier
Bus Fleet
Utility Fleet
Lease/
Rental Fleet
Government Fleet

N
176

120
18
32

4
10
Non-Bus
Fraction
.51

.35
-
.09

.02
.03
Miles Before
Overhaul
273,000

308,000
304,000
227,000

123,000
123,000
Miles After
Overhaul
224,000

268,000
236,000
163,000

224,000
97,000
Non Bus Total       342
Non Bus Weighted Avg.                  277,000         230,000

-------
                              5-22


6, 7 and 8 trucks  is  shown in Table 5-7.  The data  for  Classes
7 and  8  appear to be approximately normally  distributed,  while
Class  6   is  bimodal.   The  Class  6   bimodal  distribution  is
probably the result  of  non-sleeved diesels  being rebuilt  at  a
lower  mileage  and sleeved  engines being  rebuilt at  a  higher
mileage.   Ideally  it would  be  best to  develop separate COVs for
all  of  the  classes  for  which  rebuild  mileages   are  being
estimated, but  there  does not  appear  to be  sufficient  data to
do that.  Therefore,  a  COV will  be estimated from  all  of  the
data  in   Table 5-7  and  applied  to   all vehicles.   About  60
percent  of   the  rebuilds  for  these   vehicles   are  performed
between   200,000   and  350,000   miles.    The  coefficient   of
variation (COV),  then,  is  about  27 percent.  Therefore,  this
analysis will  use a  value of  25 percent  for the  COV of  all
average rebuild mileages.    Furthermore,  the analysis  will  also
use  this  value  to   establish the  distribution  of  scrappage
mileages.  The  percentages  of  trucks  receiving  rebuilds  using
these  values  and  the  Monte Carlo  analysis  are  shown in  Table
5-8.

     From the data in Table 5-8  it is evident that nearly all
trucks receive  a  first  rebuild,   about  one-half  of the trucks
receive  second  rebuilds,  and   a  few   trucks   receive  third
rebuilds.  At  this point  this portion of the analysis rests on
how  well the  rebuild  and scrappage  distributions  have  been
estimated. No survey data have yet been found to  compare to the
values in Table 5-8.  It  is also  now  evident that  to estimate
rebuild costs,  estimates  of vehicle mileage  between second and
third  rebuilds,  and third  rebuilds and scrappage  are  needed.
(Mileages before  and   after   first  rebuild  were presented  in
Table  5-6).  The  mileage between the  first  and  second  rebuild
as presented  in  the  previous  section will  be  used  for  these
mileages.  The  final  item  that  needs  to be  determined  is  the
cost of the various rebuilds.

     Rebuild  costs  are  a  function   of the cause  of  engine
rebuild,  whether  the  rebuild  can  be performed in the  frame or
must be  removed,  the complexity of the engine,  and  a  host of
other  factors.    Generally,    if  only   the  power   cylinder
components  are being replaced,  the rebuild  is  done  in-frame.
If  the  rebuild   requires  crankshaft  removal,   however,  the
rebuild  is  performed with the  engine  removed from the frame.
According to the EMA rebuild  survey  the components most often
replaced  or  rebuilt  during the first  rebuild are piston rings,
connecting  rod  bearings,   main  bearings,   cylinder  liners,
injectors  and  nozzles,  and  the  cylinder   head assembly  and
pistons.     Camshafts,   crankshafts,   push   rods/tubes,   roots
blowers,  aftercoolers  and vibration dampers  are  seldom rebuilt
on  the first  rebuild,  but may  need  attention  in the second
rebuild.

     Several  repair  shops  were  contacted  for  estimates  of
rebuild  costs  and most  would  only give  ranges of  costs.   From

-------
                     5-23
                  Table 5-7
Distribution of Mileages to First Overhaul by
      J3VW Class from EMA Rebuild Survey

First Overhaul
Under 100,000 Miles
100,000 - 149,999
150,000 - 199,999
200,000 - 249,999
250,000 - 299,999
300,000 - 349,999
350,000 - 399,999
400,000 - 449,999
450,000 - 499,999
500,000 - 549,999
550,000 - 599,999
600,000 - Over

Average (xl,000)
Median (xl,000)
Sample Size:

Total
5%
7
9
19
18
23
9
5
2
1
—
2
100%
259
278
423

8
5%
6
8
19
19
24
9
5
2
1
__
2
100%
262
283
271
GVW
7
2%
4
8
22
21
18
5
14
—
3
—
3
100%
280
281
100

6
6%
16
24
24
3
16
9
1
—
—
—
1
100%
203
208 .
52

-------
                  5-24


               Table 5-8
Percentage of Trucks Receiving Rebuilds
 from Monte Carlo Analysis (COV a 25%)

MHDV2S
HHDVS
Buses
1st Rebuild
92%
95%
93%
2nd Rebuild
54%
62%
53%
3rd Rebuild
8%
14%
9%

-------
                              5-25


these conversations, a first  rebuild  repair  cost for MHDV2s  of
$4,000  is  estimated,  and second  rebuild cost  for MHDV2s  are
estimated at  $6,000.   First  rebuild  costs  for HHDVs and  buses
are  estimated  at  $5,000.    Second  rebuild  costs  for  heavy
heavy-duty trucks  and buses  are  estimated at  $8,000.   Third
rebuild costs  are  assumed to be  the  same as  the  first rebuild
cost for all  vehicle types.

     An analysis of base  case rebuild  costs for  HHDVs using  the
data developed  in  previous  sections  is  shown  in of  Table 5-9,
Lifetime  mileages,  vehicle   life,  and  discounted   lifetime
mileage  were  the  same values  as  those  used  in  the  oil  cost
analysis.  Mileages  to  rebuild were  determined  from the  Monte
Carlo  analysis  of  the   fractions of  trucks  receiving  each
rebuild.  Total present rebuild cost  is estimated  by weighting
the  present   value of  each   rebuild  cost  by  the  fraction  of
vehicles receiving that   rebuild.   Lifetime rebuild  costs  are
obtained  by  dividing the   total   present  cost  by  discounted
lifetime mileage.  Lifetime base  case rebuild  costs  range from
1.29 cents/mile  for buses to 1.93 cents  per  mile  for MHDV2s.
Total base case  costs  for heavy-duty  trucks  and buses  can  now
be  estimated  by  adding   the  vehicle  replacement  costs  (Table
5-5) to the  rebuild  costs.   Total  base  case  costs  for  all
vehicle  types are  summarized  in  Table  5-10.    Replacement  and
rebuild  costs  for MHDVs have  been  weighted  together  using
weighting  factors  from  the   RIA   for  the  NOx  and  particulate
Final Rule.[19]   Total  costs range from 6.57 t/mi  (LDDTs)  to
16.19  Ł/mi  (bases).   Rebuild  costs  range  between 8 and  15
percent of the total of rebuild and replacement costs.

     Sulfur  Control  Case -  To  estimate  the  sulfur   control
vehicle  replacement  and  rebuild  costs estimates of  the number
of vehicles that  will  benefit from sulfur  control and size of
the  benefit  are needed.  These estimates  can  then be   used to
extend  rebuild  interval  and  vehicle   life  for   those  that
benefit, resulting  in  lower  fleet average operating costs.  The
first  item  that  is  estimated  is  the  fraction  of  vehicles
benefiting from  sulfur  control.    It  was  established  earlier in
this chapter  that the vehicles  that  would benefit  from sulfur
control  are  those that  are  rebuilt or  scrapped  due to  the
problems related to high oil-consumption.

     Information  on  this  item  was   available  from  the  EMA
Rebuild   Survey.    This   survey   included  both   fleets  and
owner-operators.   Participants  were   asked   to  indicate  the
criteria  they  use  to  determine   when  an  engine is  rebuilt.
Several criteria were provided on  the  data  form  (such as engine
hours,  years  in use,  oil consumption),  and participants could
also write-in alternate criteria.

     Some of the results  of this survey  are  shown in Table 5-11
in  the  column  labeled  "Percent  of Fleets  citing."   Eighty
percent  of the  fleets  and  71  percent  of  the  owner   operators

-------
                               5-26
                             Table  5-9
                  Base Case Engine Rebuild Costs
Lifetime Mileage (mi)
Vehicle Life (yrs)
Discounted Lifetime Mileage (mi)
 MHDV2S

321,000
10.8
196,927
 HHDVs

529,000
8.2
348,689
 Buses

540,000
12.0
320,290
First Rebuild

Mileage to First Rebuild
Age at First Overhaul (yrs)
Cost of Overhaul ($)
Present Value Cost of
  First Rebuild ($)
170,960
4.5
4,000


2,605
266,114
3.4
5,000

3,616
293,473
6.5
5,000

2,691
Second Rebuild
Mileage to Second Rebuild (mi)     297,665     480,123     490,157
Age at Second Rebuild (yrs)        9.4'         7.1         10.9
Cost of Second Rebuild ($)         6,000       8,000       8,000
Present Value Cost of
  Second Rebuild ($)               2,449       4,066       2,831
Third Rebuild
Mileage at Third Rebuild (mi)
Age at Third Rebuild (yrs)
Cost of Third Rebuild ($)
Present Value Cost of
  Third Rebuild ($)
Total Present Cost ($)
Lifetime Rebuild Cost
368,966
13.5
4,000

1,104



3,807



1.93^/mi
631,751
10.8
5,000

1,786



6,206
596,910
13.3
5,000


1,407



4,130

-------
                              5-27


                           Table 5-10
                      Base Case  Replacement
                 and Rebuild Costs (cents/mile)
Vehicle Type    Replacement Costs    Rebuild Costs    Total Cost

   LDDV              8.89^/mi           none            8.89^/mi
   LDDT              6.57               none            6.57
   LHDT             10.83               none           10.83
   MHDV             11.58               1.46^/mi       13.04
   HHDV             10.50               1.78           12.28
   Buses            14.90               1.29           16.19

-------
                                 5-28


                              Table 5-11

                     EMA Rebuild Criteria Survey

                             Fleet Survey


                              Percent of        Frequency of
	Criteria	       Fleets Citing      Reasons Cited %

Oil Consumption                  80                 20.2
Maj Engine Failure               65
Reduction in Oil Press           62
Blowby                           57                 14.4
Loss of Performance/
 Reduced Power                   46                  5.8*
Miles                            32
Oil Analysis                     22                  5.5
Compression Pressure             18                  2.2
Engine Mrs.                       9
Years in use                    	6                 	
Total Responses                 397                 48.1%
                     Owner  Operator  Survey


                          Percent of Owner       Frequency of
	Criteria	    Operators Citing     Reasons Cited (%)

Oil Consumption                 71                  16.3
Maj Engine Failure              69
Reduction in Oil Press          65                  12.9
Blowby                          56
Loss Of Performance/
 Reduced Power                  53
Miles                           34
Oil Analysis                    33
Compression Pressure            29
Engine Mrs.                     15
Years in use                    10
Total Responses                435
     Value includes only half of responses.

-------
                              5-29


cited oil  consumption as  the major  cause for  engine  rebuild.
However,  participants  in  many  cases  listed  more  than  one
criteria for  engine  rebuild.   Therefore,  it  cannot  necessarily
be inferred that 80 percent of engines are rebuilt for  high  oil
consumption,  since  there may be  other failures  in  conjunction
with high oil consumption  that  actually trigger a rebuild.   To
determine the likely percentages  of  those  engines  which  were
rebuilt for oil consumption alone (and that would  experience an
increase in  rebuild  interval  with  lower  oil  consumption  and  a
low  sulfur  fuel), criteria were selected  in  Table 5-10  that
relate to oil consumption,  the  total responses were  summed,  and
the  oil  consumption  criteria were  renormalized  to the  total
responses.    The  results   are   shown  in  the  column  labeled
"Frequency  of Reasons Cited  Percent."   The  criteria  selected
were  oil  consumption,   blowby  and  oil  analysis.    Excessive
blowby  can  only  be  caused by  worn  rings  and/or  liners.    A
rebuild  undertaken  because  of  oil  analysis  would  only  be
performed if  excessive wear products were observed  in  the oil,
again primarily  related  to ring  and liner wear.  Reduction in
oil  pressure  was  not selected  because this problem is usually
caused by a faulty oil  pump or looseness  in  engine  parts where
oil  is  applied.    (Low  oil   pressure  due  to  excessive  oil
consumption  can  usually  be  solved by  adding  oil,  or topping
off.)   Loss   of   performance/reduced  power   and  compression
pressure could be caused  by valves  or lack of fit between rings
and  liners,  or both.   For this  analysis, it  will  be assumed
that one-half  of  the responses  are due  to valves and one-half
due  to  ring/liner wear  for  each of  these  two  criteria.   The
resulting percentage  of  engines  rebuilt  due  to ring  and liner
wear for  fleets  is  48.1  percent,  and  for  owner operators  is
46.2  percent.   The   remainder  of  this  analysis will use  47
percent as the percent of trucks benefiting from sulfur control.

     In their comments on the draft  EPA wear  analysis EMA  and
MVMA  questioned  the  use  of  this  survey data to  predict  the
percentage  of  engines benefiting from sulfur  control. [22]   It
is   expected   that  maintenance  supervisors   for  fleets  and
owner/operators responding in the  survey have  identified some
reasonable,  experience-based   criteria  for  rebuilding engines
that seeks to minimize total  maintenance  cost.   That the survey
participants  identified  oil   consumption  is  not  surprising,
since excessive oil  consumption can lead to excessive  wear  and
a  catastrophic   engine   failure,   which  could  substantially
increase the cost of an engine rebuild or result in  that engine
being  scrapped.   Therefore,   EPA  views  this   survey  data  as
representative for actual engines.

     There  are no  data  on  the  causes  of engine  failure  in
LDDVs,- LDDTs, LHDTs,  and MHDVls.   Therefore,  this analysis will
use  the  truck fraction developed above  (47%)  for the fraction
of these vehicles benefiting from reduced sulfur levels.

-------
                              5-30


     Data from  the  in-use  oil  analysis  indicated  that  diesel
vehicles would  experience  18 percent less wear on  a low sulfur
fuel.  The  assumption  used  in  this analysis  is  that  vehicles
benefiting  from sulfur control  travel  18 percent  further  with
low  sulfur  fuel before engine  rebuild  or scrappage.   However,
some  of  these  vehicles  will  fail  for  non  oil-consumption
reasons,  or crash  in  the   extension  period.   These  vehicles
receive some, but not the full, benefit.

     The effect  of  these two phenomena can be  accounted for by
reducing the percent of  vehicles benefiting (i.e.,  47  percent)
from  low  sulfur fuel to a  lower level.  To do this,  estimates
of  the  length  in  miles   of  the  extension  period  must  be
evaluated.   Next,  the  rate  of crashes  and  non-oil  control
failures  with  mileage  must be  calculated.    Combining  these
estimates  will  give   the   percentage   of   vehicles  receiving
partial  benefits.   This   analysis  will  assume   that  those
vehicles that  crash or fail for non-oil  control  reasons in the
extension  period  receive   one-half  of  the  full  18  percent
benefit, or  9  percent.   However,  this  can be  accounted for by
assuming  that  one-half of  the  vehicles  which  fail  in  the
extension period receive the full  benefit.   Therefore, all that
is needed  is to estimate  the percent of  vehicles  crashing or
failing for non-oil control reasons in the extension period.

     Using HHDVs as  an  example,  the mileage to first rebuild is
266,114 miles  (Table 5-9).  Some  trucks  will travel  18 percent
farther  until   rebuild,  or  to  314,014 miles,  which  is  47,900
miles  further  and 0.8 years older.  Data from MVMA  show  that
0.3  percent  of  trucks  crash every  year,  so 0.24  percent of the
trucks  crash in  the extension  period.[23]   If  47 percent  of
trucks  are  rebuilt  for  oil  control  reasons at  266,114 miles,
then the remaining 53 percent of trucks are  rebuilt for non-oil
control reasons  at  266,114  miles.   Assuming the rate of non-oil
control failures  in the extension period is the  same  as in the
first  266,114  miles, the percent of trucks  failing for non-oil
control reasons  in  the  extension period of 47,900 miles  is 9.5
percent.  Therefore,  the  total  percent  of  trucks  receiving  a
half  benefit is  9.74  percent,  and if this  is  rounded to 10
percent  and halved, the crashes  and  non-oil  control  failures
can  be  accounted for by subracting 5  percent  from  47 percent,
so  that  42  percent of all  trucks receive  a  full  18 percent
benefit.  This  is  the  same as saying that 37  percent  receive  a
full benefit,  and  10 percent receive a one-half benefit, and is
easier  from  a  calculational  perspective.   Although   this  was
estimated for  HHDVs, the same percentage can  be  applied to the
rebuild  and scrappage mileages  of  the other  vehicle   classes,
since  the  extension benefit  and  the  percentages  of vehicles
crashing  and  failing  for  non-oil control  reasons   are   both
mileage-based  (i.e.,  a higher  rate of  non-oil control  failures
in a lighter vehicle class will  be  balanced  by a lower  absolute
mileage extension).

-------
                              5-31


     The results  of  the above analyses show that  42  percent  of
vehicles will  travel  18  percent further  on  low sulfur  fuel
before rebuild and scrappage.  These  factors  need to be applied
to  the  vehicle  replacement  and  rebuild costs  as developed  in
the base case.

     New fleet vehicle  replacement costs  are  estimated by first
estimating  replacement  costs  for the  42  percent of  vehicles
traveling  18  percent further,  then weighting  the two  vehicle
groups back together.   These costs are shown  in  Table  5-12  in
the  "Vehicles Benefiting," and "Fleet" costs.   The  last column
shows the  difference in cost  from the base  case, which  range
from 0.91 cents/mile for LDDVs to 1.78 cents/mile for MHDVls.

     Next,   rebuild  costs  for vehicles  benefiting from sulfur
control must be  estimated.  These  are shown for  MHDV2s,  HHDVs
and  buses  in Table  5-13.    Rebuild  costs  and  the fraction  of
vehicles receiving each  rebuild  are assumed  to be the  same  as
the  base  case.   The sulfur control  rebuild  costs  range  from
0.91  cents per  mile for  buses  to  1.52   cents  per  mile  for
MHDVs.   These  are about 20 percent  lower than the  base  case
costs.  The total  costs for the  base and  sulfur  control  cases
for  all  the vehicle types  are  summarized in  Table  5-14.   The
differences are  shown  in  both cents/mile and  cents/gallon  of
fuel,   the  latter   using   model  year   2000   current   fuel
economies.[16,17]  The benefits  in cents/gallon  range from 8.4
cents/gallon for HHDVs to 30.7 cents/gallon for LDDVs.

     2.    Reduction  in Operating Costs  - Increase  in Engine
           Rebuild Interval  and Reduction in Number of Rebuilds

     The prior methodology  assumed that  an increase  in engine
rebuild  interval  also increased  vehicle  life.   In this method,
an  increase in engine rebuild interval leads- to  a reduction  in
numbers of rebuilds,  particularly second rebuilds  for HHDVs.

     The  base  case  costs  are  exactly  the  same   as  in  the
previous example.   To predict the reduction  in  the  number  of
rebuilds,  the  Monte Carlo  analysis  was rerun without adjusting
vehicle  lifetime  mileages.    The effect   on   the  fraction  of
second and third rebuilds  is shown in Table 5-15  (the base case
fractions  are also shown for comparison).   Second rebuilds have
dropped  about  10 percent  for  each vehicle type.  The impact on
control  case operating  costs is  shown in  Table 5-16.   Benefits
are  much  lower  than the  previous example because there  is  no
deferral of vehicle  replacement costs.  The benefits range from
1.98  cents/gallon for   buses  to   2.52  cents  for  HHDVs.  Since
there  is no extension  of  vehicle life,  there  are  no  benefits
for  LDDVs, LDDTs  and LHDTs in  this method.

-------
                              5-32
                           Table 5-12
              Vehicle Replacement Cost Comparison
                                  Costs, (cents/mile)
Vehicle Type

   LDDV
   LDDT
   LHDT
   MHDV1
   MHDV2
   HHDV
   Buses
Sulfur
Vehicles
Benefiting
6.730/mi
4.16
7.38
14.41
5.45
7.58
10.76
Control
Fleet
7.98^/mi
5.56
9.38
16.88
6.99
9.27
13.16
Base
Case
8 . 89^/mi
6.57
10.83
18.66
8.10
10.50
14.90
Difference
Base-Control
0.910/mi
1.01
1.45
1.78
1.11
1.23
1.74

-------
                              5-33


                           Table 5-13
                     Sulfur Control Engine
             Rebuild Costs for Vehicles Benefitting
Parameter
Lifetime Mileage (mi)
Vehicle Life (yrs)
Discounted Lifetime Mileage (mi)
First Rebuild
Mileage to First Rebuild (mi)
Age at Rebuild (yrs)
Second Rebuild
Mileage to Second Rebuild (mi)
Age at Rebuild (yrs)
Third Rebuild
Mileage to Third Rebuild (mi)
Age at Rebuild (yrs)
MHDV2S
378,780
14.2
213,280
201,732
5.5
351,244
12.3
435,380
20
HHDVs
624,220
10.6
384,326
314,014
4.2
566,545
9.1
745,466
14.6
Buses
637,200
14.2
392,317
346,298
7.7
578,385
12.9
704,354
15.7
Lifetime Rebuild Cost (tf/mi)      1.51        1.42       0.91

-------
                              5-34


                           Table 5-14


                 Operating Costs and Reductions


                                                 Reduction
Vehicle Type         Base       Control      
-------
                    5-35



                Table  5-15
      Percentage of Trucks Receiving
Rebuilds - Monte Carlo Analysis (COV = 25%)
Case
Base


Control


Vehicle
MHDVs
HHDVs
Buses
MHDVs
HHDVs
Buses
1st Rebuild
92%
95%
93%
91%
93%
90%
2nd Rebuild
54%
62%
53%
43%
51%
41%
3rd Rebuild
8%
14%
9%
7%
7%
6%

-------
                          5-36
                       Table 5-16
             Operating Costs and Reductions
           for Changes in Numbers of Rebuilds
Vehicle Type
MHDVs
HHDVs
Buses
Base Control <Ł/mile ^/gallon*
13.04 12.73 0.313 2.51
12.28 11.91 0.37 2.52
16.19 15.89 0.30 1.98
Fuel economics used
MHDV  - 8.1 mpg
HHDV  - 6.8 mpg
Buses -6.6 mpg

-------
                           CHAPTER 6

              EFFECT OF FUEL MODIFICATIONS ON AIR
              QUALITY, PUBLIC HEALTH,  AND WELFARE

     This  chapter  develops the  emissions  impacts for  the  fuel
control  options,  and  assesses  the  resulting  changes  in  air
quality   (particulate  concentrations)   and  welfare   effects
(visibility  and  S02   reductions).    The   emission   reductions
developed  here  are  further  used  in  Chapter  7  to  develop
particulate cost effectiveness estimates.

     The  chapter  is  divided  into  three  sections.   The  first
section  explains  how  emission  inventories  are  estimated  for
on-highway mobile,  off-highway  mobile and  stationary  sources,
and presents the urban  emission  reductions for the fuel control
scenarios.   The second section  presents  the  methodology  for
predicting improvements in  particulate  and  SO2  concentrations
in  urban and  rural  areas.   Included in  this  section is  the
development  of a methodology  for estimating S02  conversion to
sulfate  particulate,  and  the  resulting  impact  on air  quality.
The   final   section  estimates   changes   in  urban   and  rural
visibility,  and discusses  some  of  the implications of  the S02
reductions.

     The    methodologies     used   to    develop    particulate
concentrations,  and  visibility  impacts  draw  extensively  on
methods  developed in  the Diesel  Particulate Study and the Draft
and    Final    RIAs    for     the    heavy-duty    particulate
standards.[1,2,3]    Some  modifications  and  improvements  have
been made, however,  and these are noted where applicable.

I.   Emissions  Inventories

     This   section   develops   gaseous    (SO2   and   HC)   and
particulate  (carbon,  SOF and sulfate) emissions inventories for
the  base  and  control  cases  outlined   in  the  introduction.
Emissions  are  estimated  on  an  annual  basis  for  on-highway
mobile,  off-highway mobile,  and  stationary diesel  and No.   2
fuel  oil emission sources.  An  overview of  these sources and  a
description  of which  sources are affected by on-highway diesel
fuel  controls  for  both  maximum  and  minimum  segregation  is
presented  first.  This  is  followed by a general  description of
the  methodology  used  to   estimate  the  emission  inventories.
Next,  since  emission  inventories depend  on how  much  fuel  is
consumed each  year,  a discussion of how future fuel consumption
was estimated  is presented.  Next,  the emission factors of each
source  are  discussed.   Finally,  the emission  inventories for
the base and fuel control  cases  are presented and discussed.

-------
                              6-2


     A.     Emission Sources

     The controls  discussed  in  this  RIA would  be  applicable
only to  on-highway diesel  vehicles.   However,  as discussed  in
Chapter 2,  refiners  may choose  to  treat a  substantial  portion
of  the  off-highway  diesel  fuel  and  No.  2  fuel  oil  also.
Sources potentially affected by diesel fuel  controls  other  than
on-highway  sources  fall  into   three  general  categories:    1)
off-highway  mobile  sources,  which  includes  construction  and
agricultural  equipment,  2)  stationary  diesel  engines used  as
generators and in  other  power equipment, and 3) No.  2 fuel  oil
applications, which  include  commercial  and industrial  boilers
and residential furnaces.   Rail  sources, vessels, and military
applications, while  they  are off-highway mobile  sources  which
use No.2 diesel fuel,  would be segregated under  these controls
as discussed  in Chapter  2,  and are therefore excluded from this
emissions  analysis.    For  the  remainder   of  this   report,
off-highway   mobile   will   include   only   agricultural   and
construction.   Stationary   diesel  .will  include  only   those
sources burning No. 2 diesel.

     B.     General Inventory Estimation Methodology

     An  overview  of  the   methods  used  to  estimate  emissions
inventories  is as  follows.   First,  for  highway mobile sources,
emission rates  in grams per  mile are estimated  for  all  diesel
vehicle types and ages. These have been  developed in  Chapters 3
and  4.   (The emission rates  for  heavy  duty  trucks   and  buses
were developed in units of  g/BHP-hr, but are converted to units
of  g/mi  in  this  chapter using  BHP-hr/mi  conversion  factors.)
The  emission rates by vehicle  type and  age are  then combined
with  registration  and  travel   data  to  obtain  fleet  emission
rates by vehicle type in a  particular  calendar  year.   These are
then multiplied  by  vehicle miles  traveled  (VMT) data  in  each
calendar year to obtain total annual emissions  for each  vehicle
class.    These   VMT   estimates   are  readily  developed  from
estimates  of total  annual fuel  consumption by  vehicle  class,
and class  specific  fuel  economy values.   The emissions by class
are  then  summed  to  obtain  total  inventories for   all  diesel
mobile sources.

     The  emission  rates   for  off-highway  mobile,   stationary
diesel and fuel  oil  sources  are available  from  the  literature
in  units  of  grams of  pollutant per  gallon of  fuel consumed.
All that is  needed to estimate pollutant  inventories  for these
sources are  estimates of fuel consumption by source category.

     The emissions  analysis  in  this  chapter estimates emission
inventories  on an annual basis for  a variety of future calendar
years.  The  purpose in doing the analysis in  this  manner is to
evaluate the changes  in emission reductions as  old  trucks are
replaced with new one with  lower emission levels.

-------
                              6-3
     The year of the most stringent emission  standards  for most
trucks  is  1994,  and  generally,  most  if  not  all  trucks  are
scrapped before they are  20  years of  age.  Therefore,  the 2015
calendar year emissions analysis  for  on-highway trucks includes
only the  lowest emitting trucks  (all  higher  emitting  pre-1994
trucks are  assumed  to  be  scrapped),  and so this year represents
a steady  state  condition in  terms of emission  standards.   The
cost  effectiveness  methodology  presented  in the next  chapter
however, utilizes a 33-year  discounted  analysis that  requires
the estimate  of emission reductions beyond 2015.  This  33 year
analysis will start  with the  1994  calendar  year.  Therefore,  to
accommodate   this   analysis,   emission   reductions  will   be
estimated through the calendar year 2027.

     Although the  33  year  analysis will  extend from  1994  to
2027,  diesel  fuel controls  could be implemented before 1994,  as
early as  1992.   EPA is conducting a  leadtime analysis  on this
issue,  however  the  results  are  not yet  available.   Therefore,
emission inventories will be  developed  for 1992 as well  as  for
the years 1994-2027.

     C.     Fuel Consumption

     Fuel consumption  estimates  are needed for the four sources
mentioned in  the previous  section:  on-highway  mobile  sources,
off highway mobile  sources,  stationary diesel sources  and fuel
oil sources.  On-highway  sources are discussed  first,  followed
by the other sources.

     Estimates  for  diesel   fuel  consumption  for   on-highway
diesel  vehicles were  taken  from the  Motor Fuel  Consumption
Model Thirteenth  Periodical Report, hereinafter  referred to as
the MFCM.[4]   Estimates of diesel fuel  consumption by  vehicle
type for  all  classes  except buses are available from the report
for calendar  years  1980-2000  (only the data  in  years  1992-2000
were  used  in this  analysis).  Total non-bus on-highway diesel
fuel consumption  ranges from  18.8 billion gallons  per  year  in
1992 to 21.9 billion gallons per year in 2000.

     One  adjustment was made  to  the  MFCM  fuel  consumption for
1992-2000.   These  estimates  do   not  include  fuel   consumed  by
buses.   The techniques used to  estimate bus  fuel  consumption
are explained in Appendix 6-A.

     A  number of techniques  and  sources  of  data were  used  to
obtain  total  annual VMT by vehicle type  for  the calendar years
1990-2000,  and  these  are  described  in  Appendix  6-A.   The
purpose of  these  calculations was to ensure  consistency between
five commonly estimated and  used  parameters - fuel  consumption,
VMT, fuel economy,  the total numbers of vehicles in each class,
and annual  VMT  per  vehicle.  Generally,  VMT  for  each  class for
1992-2000   was   determined   by  dividing   class  specific  fuel
consumption by  class  specific  fuel  economy estimates.   This

-------
                              6-4
resulted in a  2.75  percent  compound annual increase in VMT  for
all  diesels  between 1992  and  2000.   This  was  rounded  to  3
percent  and applied  in  a  compound fashion  for the  2000-2027
calendar years,  since  the  rate  of growth  was rising  between
1992  and 2000.   Fuel consumption beyond  2000 could  then  be
estimated for  calendar  years 2000-2027 by dividing VMT  by fuel
economy.  The" 3  percent  compound  VMT growth  rate resulted  in
about  a 2.5  percent compound annual  increase in  total  fuel
consumption over the  entire period of 1992-2027.  The  specific
procedures  used  in  this analysis are discussed  in detail  in
Appendix 6-A.

     The foregoing discussions developed  on-highway  diesel fuel
consumption  and  diesel  VMT   for  1992   through   2027.    Fuel
consumption   estimates   for   off-highway    mobile    sources,
stationary diesel and fuel  oil sources for the same time period
are also needed.

     As discussed in  Chapter 2,  the.minimum  (NPRA)  segregation
volume  of  fuel  controlled  includes  on-highway,  off-highway
mobile,  stationary  diesel   and  fuel  oil  sources.    With  the
on-highway  fuel  consumption developed  above,  the  sum total  of
off-highway mobile,  stationary  diesel  and fuel  oil  consumption
can be  estimated by subtracting the on-highway fuel consumption
from  the total  volume  of   fuel  controlled  under the  minimum
segregation scenario.  For  1990  the volume of  fuel  controlled
with,  minimum  segregation   is   37.2  billion  gallons.   The
on-highway  demand    (from   Appendix   6-A)   is   20.2   billion
gallons.   Thus,   for 1990, the  volume  of  fuel   consumed  by
off-highway mobile,   stationary diesel  and fuel oil  sources  is
about  17 billion gallons.   This  consumption  must be  properly
allocated  between  the  different  sources,  and projections  of
fuel consumption must be made for the years 1992-2027.

     Table  2-2 in  Chapter  2  lists the  percentages  of  middle
distillate  fuel  used   by  the   emission  sources   and  these
percentages can be  used  (after removing  military,  railroad  and
vessel  bunkering,   and   renormalizing)   to  allocate  the  non
on-highway  fuel  consumption   into  the   different  categories
(off-highway  mobile, stationary  diesel  and  fuel oil).   When
this is  done,  about  24  percent  of the off-highway  mobile plus
stationary  diesel  plus  fuel  oil  is  off-highway  mobile,  20
percent is stationary diesel, and  the  remainder  (56  percent)  is
fuel  oil,   However,  as  will  be  shown shortly  in  the  emission
rate analysis, off-highway  mobile  and  stationary diesel  engines
have  similar  emission rates that  can  be combined to form  one
source.  Thus,  this  analysis  will  combine  the estimates from
these   sources,   and  refer  to   it   as   "other  diesel"   fuel
consumption.  The resulting  percentage for  this  category of  non
on-highway  diesel  is 44 percent (24 +  20).   The 1990 projected
volume of fuel consumed by  "other  diesel"  and fuel oil  sources,
estimated  by  the  above  methodology  is  7.5  and 9.5  billion
gallons,   respectively.     The   only  item   remaining   is   to

-------
                              6-5
estimate  fuel  consumption   for   these  categories   in  other
calendar years.   The  VMT growth  rates  developed  for on-highway
mobile sources resulted  in  about  a 2.5 percent compound annual
increase in  on-highway  fuel  consumption.   This  percentage  was
applied to the other  diesel  and  fuel  oil  categories  as  well.
This  is  consistent  with  growth  rates  used  for  off-highway
emissions in the  past.[2,3]   A summary of  all  fuel  consumption
values for calendar year 1990-2027  thus  obtained for  the  NPRA
segregation case obtained is shown in Table 6-1.

     D.    Emission Factors

     This section describes in more  detail  the  emission  rates
used  for  the  different  sources  and  the  methodologies  for
estimating  these   emission   rates.    The  situation   is   most
complicated  for  on-highway  sources  since the  emission  rates
change by model  year,  and  thus,   vehicle  scrappage  and travel
must  be  accounted  for  to  properly estimate emission  rates  by
vehicle type.  For the other sources, the emission  rates do  not
change from  year-to-year, other  than with  fuel composition,  so
estimating inventories for these sources is easier.

     The emission rates  and methodologies  for on-highway mobile
sources  are  discussed  first, followed  by  the other  diesel
emissions  (off   highway  mobile   and  stationary  diesel)   and
finally, fuel oil sources.

     1.    On-Hiqhway Mobile Sources

     Emission  rates from the  six vehicle types  (LDDVs, LDDTs,
LHDTs, MHDVs, HHDVs, and buses) for base and  control fuels were
developed in  Chapters 3  and  4,  with the  exception of emission
rates  for vehicles with  trap  failures  which will  be discussed
presently.   The  LDDV  and LDDT emissions  are in  units of grams
per  mile,  and can  be  multiplied  by  class-specific  VMTs  to
obtain  inventories.  However, the other classes are presented
in units of  g/BHP-hr,  and must be converted  to  units  of g/mi.
The  conversion factors used to do  this are  those developed for
MOBILE4,   and   are  listed   in    Table   6-2.[5]    The   GVWR
class-specific conversion factors in the  referenced report were
weighted by  nationwide VMT to obtain  similar  factors  for  the
somewhat  broader  vehicle  class  categories  used  here.   For
example, this  analysis groups  classes 6-8a   in  the  above report
into the single class of MHDVs.

     The  emission  rates developed  in Chapters  3  and  4  were
end-of-life  emission   rates   for  fleets  of   vehicles,   some
including traps,  with  no trap failures.   As  developed in  the
supporting   analyses   for  the   diesel particulate  standards
[1,2,3],  trap  failures  may occur   because   of   failure  of
electronic controls or  due to unforeseen  operating conditions.
When  trap failure occurs, the vehicle  is assumed to be emitting
at  its engine-out  emission  levels.   (This   applies  to  HC  and

-------
                              6-6
                           Table 6-1
              Middle Distillate Fuel Consumption
                by Source  (10M gal per year)
Calendar Year

    1990
    1991
    1992
    1993
    1994
    1995
    1996
    1997
    1998
    1999
    2000
    2001
    2002
    2003
    2004
    2005
    2006
    2007
    2008
    2009
    2010
    2011
    2012
    2013
    2014
    2015
    2016
    2017
    2018
    2019
    2020
    2021
    2022
    2023
    2024
    2025
    2026
    2027
Highway
1.97
1.99
2.01
2.03
2.05
2.08
2.12
2.16
2.2
2.25
2.31
2.35
2.41
2.46
2.53
2.59
2.66
2.74
2.81
2.89
2.97
3.06
3.15
3.24
3.34
3.43
3.54
3.64
3.75
3.86
3.98
4.09
4.22
4.34
4.47
4.6
4.74
4.88
Other
0.768
0.784
0.799
0.815
0.832
0.848
0.865
0.882
0.900
0.918.
0.936
0.955
0.974
0.994
1.014
1.034
1.055
1.076
1.097
1.119
1.142
1.165
1.188
1.212
1.236
1.261
1.286
1.311
1.338
1.364
1.392
1.420
1.448
1.477
1.507
1.537
1.567
1.599
Fuel Oil
0.978
1.007
1.037
1.068
1.101
1.134
1.168
1.203
1.239
1.276
1.314
1.354
1.394
1.436
1.479
1.524
1.569
1.616
1.665
1.715
1.766
1.819
1.874
1.930
1.988
2.048
2.109
2.173
2.238
2.305
2.374
2.445
2.519
2.594
2.672
2.752
2.835
2.920
Total

3.716
3.781
3.847
3.914
3.983
4.062
4.153
4.246
4.339
4.445
4.561
4.659
4.779
 ,890
 ,023
 ,148
5.285
5.466
  573
  725
5.879
6.044
6.212
6.382
6.564
6.739
6.935
7.125
7.326
7.530
7.746
7.955
8.187
8.412
8.649
8.889
9.143
9.399
4,
5,
5
5
5

-------
                        6-7
                     Table 6-2

Heavy-Duty Diesel Conversion Factors (BHP-hr/mi)[5]
MYR
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994+
LHDVs
^~B
	
	
	
	
	
	
	
	
	
	
.942
.923
.922
.921
.919
.919
.919
.919
.919
.919
.919
.919
.919
.919
MHDVs
2.72
2.77
2.81
2.85
2.87
2.83
2.82
2.82
2.72
2.57
2.41
2.39
2.31
2.44
2.40
2.44
2.32
2.32
2.32
2.31
2.31
2.31
2.31
2.30
2.30
HHDVs
3.20
3.27
3.27
3.35
3.35
3.30
3.35
3.40
3.36
3.31
3.33
3.26
3.15
3.15
3.14
3.13
3.13
3.13
3.13
3.13
3.13
3.13
3.13
3.13
3.13
                                                  Buses
                                                  3
                                                  3
  07
  00
3.19
2.99
2.82
2.93
2.80
2.84
2.84
2.83
2.83
2.88
2.81
2.86
2.62
2.56
2.56
2.55
2.55
2.53
2.52
2.51
2.50
2.49
2.48

-------
                              6-8


carbon and SOF particulates  -  there is  little  or no  effect  on
S02  or  sulfate  particulate.)    To   estimate   fleet  emission
rates, the  fraction  of  trap  failures  must  be  estimated,  and
emissions of  vehicles  with operating  traps  and failed  traps
weighted together.   This  analysis uses  a  two  percent  annual
trap  failure  rate estimate, which is  the same  as the analysis
for   the   final  diesel   particulate   standards.    Engine-out
emission levels used  for  trap-failed  vehicles were developed in
Chapter 3 and 4.

     The registration and travel data used  in this  analysis are
shown in Tables 6-3,  6-4,  and  6-5.  The registration by age and
mileage accumulation  rates  by  age  data comes from  MOBILE4.[6]
The  diesel  sales fractions  were developed  from  data available
in the MFCM.[4]   Registration  fractions,  VMT, and diesel  sales
fractions in  each model  year are multiplied, and the product is
summed over  all  the  model  years  within  a  calendar year  and
normalized to obtain model  year  travel  fractions.    The  model
year  emission  factors  are   then weighted by  these  travel
fractions to  obtain  class-specific emission rates  by  vehicle
types for each calendar year.  These emission rates  can then be
multiplied by estimates  of  class specific nationwide annual VMT
to yield annual emission inventories.

     Lastly,   to  obtain  estimates  of  urban,   as  opposed  to
nationwide,  emissions,  the VMT  estimates  by class  developed in
previous  sections  are  multiplied  by  the  fractions  of  urban
travel by class.  These  are obtained from the conversion factor
analysis for MOBILE4 and are shown in Table 6-6.[4]

     2.    Off-Highway Mobile and Stationary Diesel Sources

     Other sources  which  use  diesel  engines are  agricultural
equipment,   construction   equipment,   and   stationary   diesel
engines used  in  generating electricity  and  operating machinery
(such as oil drilling rigs).

     The   emissions   of   interest   in  this   analysis   are      i
particulates,  hydrocarbons,  and  sulfur dioxide.  The analysis
of HC  and particulate emissions in Chapter 4 showed that HC and
particulate  emissions vary  with  fuel  aromatics content  under      ,
transient  conditions,   but  do  not   vary  under  steady-state
conditions.    Most of  the diesel  engines  in these  off-highway
categories    are   operated   under   steady-state   conditions,
therefore, HC emission from  these  sources  were  not estimated at
all.   S<>2  emission rates  are  a function  of sulfur  content and
sulfate  conversion  in  the  engine  and exhaust  system.   This
analysis  uses  a two percent  S02-to-sulfate conversion  rate,
which  is  the same  rate  used  for  on-highway diesel  engines in
Chapter 4.  Using a diesel  fuel density of  7.1  Ibs  per gallon,
and  current  and  low sulfur  percentages  of  .25   and   .05  wt
percent,  the SC-2 emission rate for non-highway diesel engines
on current sulfur fuel  is  15.8  g/gal,  and for  low  sulfur fuel
is 3.15 g/gal.

-------
                              6-9


                           Table 6-3

           Vehicle Registration Distributions by Age
                                          All
Vehicle                                Heavy-Duty
  Age          LDDVs        LDDTs       Trucks*      Buses

   1           0.062        0.070        0.082        .077
   2           0.082        0.092        0.165        .077
   3           0.079        0.088        0.135        .076
   4           0.075        0.083        0.111        .074
   5           0.071        0.077        0.091        .072

   6           0.067        0.072        0.075        .070
   7           0.063        0.067        0.061        .068
   8           0.060        0.063        0.050        .065
   9           0.056        0.057    .    0.041        .062
  10           0.052        0.051        0.034        .059

  11           0.048        0.047        0.028        .055
  12           0.045        0.041        0.023        .050
  13           0.041        0.036        0.019        .043
  14           0.037        0.031        0.015        .034
  15           0.033        0.026        0.013        .026

  16           0.029        0.021        0.010        .020
  17           0.026        0.016        0.009        .016
  18           0.022        0.011        0.007        .013
  19           0.018        0.007        0.006        .011
  20+          0.034        0.044        0.024        .032
     Classes 2b-8.

-------
                     6-10





                    Table  6-4





Mileage Accumulation Rates by Age (miles per year)
Vehicle
Age
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20+

LDDV
17,825
16,475
15,233
14,081
13,017
12,033
11,124
10,283
9,506
8,788
8,123
7,509
6,942
6,417
5,932
5,484
5,069
4,686
4,332
4,005

LDDT
20,140
17,572
15,434
13,639
12,133
10,863
9,788
8,877
8,103
7,444
6,883
6,405
5,999
5,655
5,365
5,123
4,924
4,763
4,637
4,543

LHDV
23,611
20,947
18,583
16,486
14,625
12,975
11,511
10,212
9,059 .
8,037
7,130
6,325
5,612
4,978
4,416
3,918
3,476
3,084
2,736
2,427

MHDV
43,946
40,504
37,332
34,408
31,713
29,229
26,939
24,829
22,885
21,092
19,440
17,918
16,514
15,221
14,029
12,930
11,917
10,984
10,123
9,331

HHDV
86,375
79,434
73,051
67,408
61,782
56,817
52,252
48,053
44,191
40,640
37,374
34,371
31,609
29,069
26,733
24,585
22,609
20,792
19,121
17,585

Buses
45,000
45,000
45,000
45,000
45,000
45,000
45,000
45,000
45,000
45,000
45,000
45,000
45,000
45,000
45,000
45,000
45,000
45,000
45,000
45,000

-------
                             6-11
                             Table 6-5

               Annual Diesel Vehicle Sales Fractions
Model
Year

1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000+
LDV
0.0
.001
.002
.002
.003
.003
.003
.003
.009
.026
.045
.060
.039
.019
.014
.008
.003
.004
.004
.004
.005
.006
.006
.007
.007
.008
.009
.009
.010
.010
.011
LDT
0.0
0.0
0.0
0.0
0.0
0.0
.001
.011
.011
.013
.056
.051
.071
.074
.037
.015
.019
.019
.021
.021
.023
.023
.023
.023
.023
.023
.023
.023
.023
.023
.023

0
0
0
0
0
0
0
0
0
0





















LHDV
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.041
.081
.122
.183
.198
.215
.232
.250
.260
.270
.280
.290
.300
.300
.300
.300
.300
.300
.300
.300
.300
MHDV
.160
.160
.145
.130
. 150
.170
.191
.249
.307
.365
.409
.454
.500
.543
.578
.633
.575
.582
.593
.605
.613
.629
.639
.648
.662
.670
.680
.689
.690
.690
.691
HHDV
.925
.923
.923
.921
.920
.920
.960
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
 Buses

 .082
 .122
 .139
 .1145
 .196
 .209
 .212
 .129
 .200
 .230
 .308
 .473
 .451
 .509
 .473
 .662
 .743
 .791
 .832
 .861
 .886
 .907
 .926
 .942
 .956
 .968
 .977
 .986
 .999
1.00
1.00

-------
                             6-12
                           Table  6-6

                    Urban Travel Fractions
                           All  Years
 LDDV       LDDT       LHDDT        MHDDT        HHDDT       Buses

0.6          0.5        0.5          0.45        0.26        0.55

-------
                             6-13


     Particulate  emission  rates  for  these  off-highway  diesel
engine  sources   are  shown  in Table 6-7.   The  source  of  the
particulate data  is  the Fourth Edition  of AP-42, Volume  l.[7]
Particulate emissions  range  from 11.8  g/gal  for  construction
equipment   to   23.3   g/gal   for   non-tractor   agricultural
equipment.   The  emission rates  for  agricultural equipment  and
stationary  diesel  engines   are  very   similar.   Due   to  the
similarity in emission  rates,  an  arithmetic  mean emission rate
was estimated for all of  the sources,  which is  shown under  the
dashed  line  in Table 6-7.   The non-highway  diesel  particulate
average  (agricultural,   construction  and  stationary)  is  17.6
g/gal,  and  this  is   used   for   all  off-highway   mobile  and
stationary   diesel    engines.     However,   these   are   total
particulate  emissions,   and  must be  split  into  the  various
particulate types.   Using   a  two  percent  sulfur   to   sulfate
conversion rate,  sulfate emissions for  a 0.25 wt percent sulfur
fuel  are  1.1  g/gal   fuel  consumed.   Data from  pre-1988  engines
in Chapter 3  shows  that of  carbon plus  SOF  emissions,  about 75
percent  is carbon   and  25  percent • is  SOF.   Applying  these
fractions  to   the   difference   between   total  and   sulfate
emissions,  the  result  is 12.3  g/gal  for  carbon emissions  and
4.1 g/gal  for SOF  emissions.   On a  low sulfur  fuel,  only  the
sulfate particulate  are affected.  A  drop of  80 percent  (from
0.25  weight percent  to 0.05 weight  percent) in sulfur  content
leads to  an  80 percent  drop in  sulfate  emissions.   Therefore,
the sulfate emission rate of non-highway  diesel engines  on low
sulfur fuel is 0.22  g/gal.

     To  obtain  urban  emissions  for  these   sources,  an  urban
fraction  for  this   category is  needed.   Data  on  urban  fuel
consumption by these sources is very hard  to  find.   However, it
is likely  that  agricultural  fuel  consumption is almost entirely
rural,  and  that  construction   and  stationary diesel   fuel
consumption  are  almost  entirely urban.    There may  be  some
overlap  in  these  categories,  however.   Assuming  that  only
construction  and stationary diesel  consumption is urban  and
using the  renormalized  consumption  percentages from Table  2-2
(that is,  factoring  out  railroad, vessel bunkering and military
use), the  urban  fraction for  these  categories  is  0.63.   Thus,
urban "other  diesel" emissions are  estimated to be  63  percent
of total nationwide other diesel emissions.

      3.     Fuel Oil Sources

      Sources which  use  fuel oil  are commercial  and industrial
boilers,   and  residential   furnaces.    Some   commercial  and
industrial  furnaces use  residual  oil,   but the emissions  of
these sources  are not affected by these middle distillate fuel
controls.  According to AP-42, the  fuel oil sources  emit only
about 2 Ibs of non-sulfate  particulate  matter  per  1000 gallons
of diesel  fuel,  and  this  is  probably the  result  of  the very
high  air to  fuel ratios used in burners. [7]  This is only about
one-tenth  of the  particulate emission  rates  of  diesel engines.

-------
                              6-14
                           Table 6-7

               Particulate Emission Factors for
       Off Highway Mobile and Stationary Diesel Sources
	Category	                     (g/gal)

Agricultural
      Tractor                                    20.8
      Non-tractor                                23.3
      Average                                    22.0
                        /
Construction
      Average of 10 types                        11.8

Stationary Diesel
      Industrial                                 15.2
      Large Bore                                 22.7
      Average                                    18.9
Non-Highway Diesel (All 3 above combined) Averages
           Total                                 17.6
           Carbon                                12.3
           SOF                                    4.1
           Sulfate                                1.1

-------
                              6-15
Also, these sources are  operated under steady-state conditions,
and so the carbon  or  SOF emission emitted by these  sources are
not expected  to  be affected  by changes in  fuel  quality.  With
respect  to  sulfate  emissions,  AP-42   reports  that  sulfur  to
sulfate  conversion  is  on the  order  of 1-5  percent,  so a value
of two  percent  will be  used  to be consistent  with the other
sources.   At  two   percent,  sulfate  emissions on 0.25  weight
percent  sulfur  fuel are 1.13  g/gal,  and 0.226 g/gal on  a low
sulfur  fuel.    S02  emission  rates  are  16.1  g/gal  on  high
sulfur fuel,  and 3.2 g/gal on low sulfur fuel.

     Urban fuel  oil consumption  is  also  difficult to  estimate
accurately.    In   the   midwest,  much  of   the   fuel  oil  for
residential furnaces is consumed  in rural areas since the urban
areas are  served by natural  gas pipelines.   In  the  northeast,
however,  fuel   oil  is  burned  even  in  the   urban  areas.
Therefore,  this  analysis  will  assume  that  urban  fuel  oil
consumption is proportional  to urban population.  According to
1980  census  data,   61  percent  of  the  U.S.   population  lives in
urban  areas   of  50,000  and  greater.[8]   Thus,   the  urban  to
nationwide fraction of  fuel  oil consumed  is  estimated  to be
61 percent.

     E.    Emission Inventory Results

     The  changes   in   total  urban,  particulate   and   gaseous
emission  inventories  due to sulfur and  aromatics control are
shown   in  Tables   6-8   and  6-9,  respectively.    Nationwide
reductions are  shown  in Tables  6-10  and 6-11.   Rural  emission
reductions can  be  obtained  by  subtracting  the  urban  from the
nationwide estimates.   The  rural estimates  are  used  in  later
analyses  to   estimate  the  effects of  fuel  controls  on  rural
visibility.   However,   the   following   discussions  will  center
only  on  urban emission  reductions (Tables  6-8  and  6-9).   The
emission  changes  in Table 6-8  and 6-9  due  to aromatics control
are incremental  to  sulfur control.  Total  emission  inventories
for the  sources addressed here are contained in Appendix 6-B.

      In  Table 6-8,  SOF,  carbon, directly  emitted sulfates, and
indirect  sulfates  are  shown separately, as well  as  the sum,
total PM.  Direct  sulfates  are those emitted at the tailpipe of
a  diesel vehicle;  indirect  sulfates  are those formed  from the
conversion   of   middle   distillate-derived   S(>2  to   sulfate
compounds,   The extent  of  conversion of   SO2  to sulfates is
developed  in  the next  section (Indirect  Sulfate  Analysis), but
the  results   are used  here.   To summarize  the   next  section,
about  12 percent of urban  S02  is converted to sulfates within
the  urban atmosphere  (mostly ammonium  sulfate).   Put another
way,  for  every  ton  of  SC>2  emitted,  conversion   of  S02  in
urban areas  to  sulfates   results  in .about   0.288  tons  of
sulfates.   The  quantity  of  these  "indirect"  sulfates  far
outweighs the directly emitted sulfates.

-------
                             6-16
Particulate

Sulfates
  (direct)
Sulfates
  (indirect)
SOF
Carbon
Total Direct
  PM
Total PM
                               Table 6-8

                   Annual Urban Particulate Emission
                       Reductions (tons per year)
Calendar
Year
1992
1995
2000
2005
2010
2015
1992
1995
2000
2005
2010
2015
1992
1995
2000
2005
2010
2015
1992
1995
2000
2005
2010
2015
1992
1995
2000
2005
2010
2015
1992
1995
2000
2005
2010
2015
Sulfur
100 %
7,500
7,630
8,230
9,322
10,734
12,440
30,203
31,856
36,506
41,920
48,439
56,172
-126
-1,033
-907
-979
-1,144
-1,289
-711
-3,101
-5,689
-7,208
-8,570
-9,969
6,663
3,496
1,634
1,135
1,020
1,182
36,866
35,352
38,140
43,055
49,459
57,354
Control
NPRA
18,344
19,336
21,544
24,461
27,951
32,028
74,189
79,324
90,425
103,203
118,132
135,473
-126
-1,032
-907
-979
-1,144
-1,289
-711
-3,101
-5,688
-7,208
-8,569
-9,969
17,507
15,203
14,949
16,274
18,238
20,770
91,696
94,526
105,373
119,479
136,369
156,244
                                                         Subsequent
                                                      Aromatics  Control
100 %
-2
-10
-32
-42
-50
-59
0
0
0
0
0
0
3,212
2,172
1,023
722
681
761
2,391
1,508
320
-42
-214
-281
5,601
3,670
1,311
612
417
421
5,601
3,670
1,311
612
417
421
NPRA
-2
-10
-32
-42
-50
-59
0
0
0
0
0
0
3,212
2,172
1,023
722
680
761 .
2,391
1,508
320
-68
-214
-281
5,601
3,670
1,311
612
417
421
5,601
3,670
1,311
612
417
421

-------
Type
S02
HC
CO
                            6-17

                           Table 6-9
            Urban S02,  HC and CO Emission Reductions
Calendar
Year
1992
1995
2000
2005
2010
2015
1992
1995
2000
2005
2010
2015
1992
1995
2000
2005
2010
2015
Sulfur
100 %
104,873
110,612
126,756
145,556
168,190
195,042
-3,189
-8,996
9,509
18,829
23,692
29,530
26,215
67,698
176,810
234,110
279,883
328,132
Control
NPRA
257,599
275,430
313,976
358,346
410,180
470,396
-3,189
-8,996
9,509
18, .829
23,692
29,530
26,215
67,698
176,810
234,110
279,883
328,132
Aromatics
100 %
0
0
0
0
0
0
18,125
14,946
8,217
6,684
7,171
8,133
24,077
23,998
26,424
28,878
32,696
37,697
Control
NPRA
0
0
0
0
0
0
18,125
14,946
8,217
6,684
7,171
8,133
24,077
23,998
26,424
28,878
32,696
37,697

-------
                             6-18
Particulate

Sulfates
  (direct)
Sulfates
  (indirect)
SOF
Carbon
Total Direct
  PM
Total PM
                              Table 6-10

              Nationwide Particulate Emission Reductions
( tons )
Calendar
Year
1992
1995
2000
2005
2010
2015
1992
1995
2000
2005
2010
2015
1992
1995
2000
2005
2010
2015
1992
1995
2000
2005
2010
2015
1992
1995
2000
2005
2010
2015
1992
1995
2000
2005
2010
2015
Sulfur
100 %
21,202
21,648
23,276
26,417
30,422
35,267
84,928
89,352
101,291
116,380
134,472
155,995
-355
-2,738
-2,268
-2,395
-2,789
-3,127
-2,112
-9,284
-16,784
-21,217
-25,213
-29,337
18,735
9,626
4,224
2,805
2,420
2,803
103,662
98,978
105,516
119,186
136,891
158,798
Control
NPRA
38,980
40,839
45,105
51,242
58,656
67,390
157,034
167,168
189,684
216,844
248,723
285,998
-355
-2,738
-2,268
-2,395
-2,789
-3,127
-2,112
-9,284
-16,784
-21,217
-25,213
-29,337
36,513
28,817
26,053
27,630
30,654
34,926
193,547
195,984
215,738
244,476
279,376
320,924
Aromatics
100 %
-2
-23
-79
-103
-122
-143
0
0
0
0
0
0
8,425
5,650
2,552
1,756
1,628
1,810
6,865
4,258
728
-403
-860
-1,097
15,288
9,885
3,201
1,274
646
570
15,288
9,885
3,201
1,250
646
570
Control
NPRA
-2
-23
-79
-103
-122
-143
0
0
0
0
0
0
8,425
5,650
2,552
1,756
1,628
1,810
6,865
4,258
728
-403
-860
-1,097
15,288
9,885
3,201
1,274
646
570
15,288
9,885
3,201
1,250
646
570

-------
                            6-19
Type
S02
HC
CO
                           Table 6-11
         Nationwide S02* HC and CO Emission Reductions
Calendar
Year
1992
1995
2000
2005
2010
2015
1992
1995
2000
2005
2010
2015
1992
1995
2000
2005
2010
2015
Sulfur
100 %
294,888
310,251
351,708
404,096
466,916
541,649
-9,422
-29,856
25,382
54,827
69,674
87,979
76,080
182,257
495,757
654,388
781,347
916,293
Control
NPRA
545,260
580,443
658,626
752,932
863,622
993,047
-9,422
-29,856
25,382
54,527
69,674
87,979
76,080
192,257
495,757
654,388
781,347
916,293
Aromatics
100 %
0
0
0
0
0
0
46,421
37,908
18,082
13,173
13,462
14,956
71,077
69,488
69,339
74,999
84,854
97,777
Control
NPRA
0
0
0
0
0
0
46,421
37,908
18,082
13,173
13,462
14,956
71,077
69,488
69,339
74,999
84,854
97,777

-------
                             6-20
     The two segregation  cases are  shown  in both  tables  - 100
percent segregation, which  represents  refinery control  of only
on-highway vehicle fuel, and NPRA  segregation,  which represents
refinery  control  of  nearly all  middle  distillate.   Emission
reductions  are  represented  by positive  values.    Increases  in
emissions are represented by negative values.

     Examination of  the  data  in  Table 6-8  reveal  two  general
items.  First,  changes  in SOF and  carbon  part icul ate emissions
are  identical   for  both  100  percent  segregation  and  NPRA
segregation  because  only  on-highway  engines  are  affected  by
changes in fuel  aromatic content.   Changes  in sulfate emissions
(both  direct  and  indirect)  with aromatics  control  only  arise
from changing control technology  of on-highway diesels.   Thus,
such   sulfate  emission  changes  only  occur  with  on-highway
diesels .

     1 .    Sulfur Control

     Emission changes due  to  sulfur  control  are  shown  in the
first  two columns.   The  reductions  in directly emitted sulfates
for the 100  percent segregation case  range from  7,500  tons  in
1992 to 12,440  tons  in  2015.   The increase in reductions is due
to  an  increase  in  fuel  consumption.   Reductions   in  sulfate
emissions for NPRA segregation range  from 18,000  tons  in 1992
to  32,000  tons  in  2015.   The  differences  between  the  100
percent and NPRA cases  represent  reductions from other diesel
and fuel oil sources.

     Reductions    in   indirect   sulfates   for   the   maximum
segregation case  range  from about 30,000 tons  in  1992 to about
56,000  tons  in  2015.   These  reductions  are  the  result  of
eliminating  much   of  the   urban   SC>2  emitted  by  on-highway
sources.   For  NPRA  segregation,   the  reductions   range  from
74,000 tons in 1992 to 135,000 tons in 2015.

     SOF  emissions   from   on-highway   vehicles  increase  with
sulfur  control.   The general upward  trend in  SOF emissions  is
due to newer vehicles (1991-93  and 1994+  heavy-duty vehicles),
which, because  of  the reduction  in direct  sulfate particulate,
can  emit more  SOF  and  carbon  emissions  in meeting  the same
particulate emissions standards.   There is  a larger  increase in
SOF  emissions  for  1991-93  trucks  than 1994+  trucks.   This  is
what  causes  SOF  to temporarily peak  in 1995 and  drop somewhat
in  2000.   For  example,  the  increase in  SOF   emissions  for
1991-93 HHDVs  is about 0.0119 b/BHP-hr,  which is  a  factor  of
4.5 times  the  1994+ HHDV SOF increase  due  to sulfur control of
0.0027 g/BHP-hr.  After  calendar year  2000,  SOF  emissions rise
again  due to the increase in fuel consumption.

     Carbon  particulate  emissions  show distinct  increases with
sulfur  control,  similar  in direction  but  greater  in magnitude
than  SOF  emissions.   The  reason   for  the  increase  in carbon

-------
                             6-21
emissions  is  the  same  as   for  SOF  emissions:   with  sulfur
control  and  emissions  averaging  (as discussed  in Chapter  3),
some HDVs do not  need traps  and others with traps  can meet the
emission standard with higher carbon emissions.

     In  using   these  results,   particularly  the   projected
increases  in  SOF  and  carbon  PM  emissions,   it  should  be
remembered that  the  analyses of  Chapter  3 and  4  assumed  that
the  0.1  g/BHP-hr  PM  standard  could  be  met  with high  sulfur
fuel.  If this were not the  case,  SOF  and carbon  PM emissions
could be much higher  without fuel sulfur control.   Then,  sulfur
control would show significant reductions in these emissions.

     Direct  PM  includes  direct   sulfates,  carbon   and  SOF.
Reductions in total direct PM for 100 percent segregation range
from  6,663  tons  in   1992  to  1,182 tons  in 2015.   They  drop
because  the  sulfate  reductions  are balanced  by  increases  in
carbon and SOF.   They do not  go to  zero,  however,  because there
are always some net PM reductions (all sulfates) from LDDVs and
LDDTs.   Reductions  in  total  direct  PM  emissions  for  NPRA
segregation  range from 17,000 tons in  1992  to  almost  21,000
tons in 2015.

     Reductions   in   total   PM  emissions   for   100   percent
segregation  range from nearly 37,000  tons  in  1992  to  57,000
tons in  2015,  and for  NPRA  segregation  range from  92,000  tons
in  1992  to   156,000   tons   in  2015.    These  reductions  are
dominated by the indirect sulfate emission reductions.

     Turning  momentarily  to  Table   6-9   to   complete   the
discussion  on  sulfur  control,  urban  emission reductions  of
SO2  for  100  percent  segregation  range  from  105,000 tons  in
1992  to  195,000  tons in 2015.   These  reductions  are  based  on
S02   inventories   for  the  base   and   control   cases   before
conversion  of  some  of  the  SC-2  in the  atmosphere  to  sulfate
compounds.    The  growth in reductions  is  due to the  growth  in
fuel  consumption  over  the same period.    For NPRA segregation,
S02  reductions  range  from  258,000  tons  in  1992  to  about
470,000  tons in  2015.   These reflect additional  reductions  in
S02 from other diesel and fuel oil sources.

     Urban HC  emissions increase in 1992 and 1995  with  sulfur
control,  but  decrease  later.   The   changes   range  from  an
increase of  8,996  tons in 1995 to  a reduction of  almost  30,000
tons  in  2015.   The near  term  increases  are  due to the absence
of  aftertreatment  on  1991-93  trucks  and  buses  with  sulfur
control.   If  trap  technology  is   not   available  until  1994,
sulfur  control will  not  affect  the  HC  emissions  of  1991-93
engines  and  this  temporary   increase  will  not  occur.   The
ability  to  use  oxidation catalysts on  some  1994+  trucks  and
buses brings significant HC reductions in  later years.

-------
                              6-22
     Urban CO  emissi-n  reductions  range  from  26,000  tons  in
1992 to  328,000  tons in  2015.   In 1992  the  emission reduction
is  entirely  due to  a  5.7  percent  decrease  in engine-out  CO
emissions from  the  in-use  fleet.   In  the later years,  sulfur
control allows  the use of oxidation catalysts,  providing  higher
percentage reductions in CO emissions.

     2.    Aromatics Control

     Table   6-8    indicates   that  directly   emitted   sulfate
emissions  increase  slightly  for the  100  percent  segregation
case   (aromatics   control   has   no   effect  on  either  S02  or
indirect sulfate emissions).  The increases  are  negligible when
compared to  the  sulfate reductions due to  sulfur  control.   The
cause  of  the increases  can be  traced to  slight  increases  in
sulfate emissions  of some  of the  heavy-duty vehicle  types due
to  shifts  from  trap technology to  oxidation catalysts  (which
oxidizes   slightly  more   exhaust  S02   to  sulfates).    The
increases  in sulfate  emissions  are  identical  for  the  NPRA
segregation case,  since they occur only  in on-highway vehicles
in either case.

     SOF emissions are  clearly  lower with aromatics  control,
especially in the  earlier years, where the  entire  in-use fleet
experiences  a  reduction  in  SOF  emissions.    The  reductions
decline because  not as many 1994+ engines will  need traps with
aromatics control,  thus  resulting  in  the same  average tailpipe
emissions.   The  reductions  do not go  to zero,  though,  because
there  is  always some benefit of  aromatics control on  LDDVs,
LDDTs, and HDVs with failed traps.

     Carbon  particulate   emissions   show  reductions   in  the
earlier years  and  increases  in  later years.   The  reasons are
the  same as  for  SOF emissions.   The net  increases  in carbon PM
emissions  are   due   to  greater  absolute   growth   in  fuel
consumption  in the HDV classes  as opposed to  the  LDDV and LDDT
classes.

     Turning again to Table  6-9,  there  are  no SO2 reductions
due  to   aromatics  control,  but  there  are  significant  HC
reductions.  They  range from about 18,000 tons  in  1992 to about
8,000  tons  in  2015.  The decrease in HC  reductions is brought
about  by newer  trucks (1994+),   in which  the  differences  in HC
emissions due to aromatics  control  is less than for in-use cars
and  trucks.   CO emission reductions due to  aromatics  control
range  from 24,000  tons in 1992 to about 38,000 tons  in 2015.

II.  Effects   of   Emission  Changes   on   Pollutant   Ambient
     Concentrations

     The  previous  section  developed   urban  and  nationwide
emission  reductions  in  tons  due  to  diesel  fuel  sulfur and
aromatics  control.  The  purpose of  this third  section  is  to

-------
                              6-23
determine the effect those emission reductions will  have on the
ambient  concentrations  of these  pollutants.  The  improvements
or  changes  in  air  quality  are  important  results  in  and  of
themselves,  but  they are further  used in a  subsequent  section
of this  chapter  to  predict  changes in visibility, and  are also
used  in a  later  chapter to  estimate  other welfare  benefits
(reductions  in  materials and crops damage)  due to  diesel fuel
controls.

     Urban  ambient   concentrations   are  developed   for   five
pollutants  -  carbon particulate,  sulfate  particulate,  total
particulate    (SOF+carbon+sulf ates),    SC>2   and   ozone.    Rural
ambient  concentrations  are  developed  for the same  pollutants
with  the exception  of   ozone.   Changes  in  concentrations  of
carbon  and  sulfate particulates due  to  fuel  controls  are used
in  Section  IV  of  this  of  this  chapter  to develop urban  and
rural   visibility   impacts   of  fuel   controls.     Changes   in
concentrations   of   total  PM,  SC>2   and  ozone   due   to   fuel
controls are used in Chapter 8 to develop welfare benefits.

     One of the  most significant  factors in  this  analysis that
affects  the  ambient concentrations   of  both  S02   and  sulfate
particulate   is   the  extent  of  the  reaction of  SO2  to  form
sulfate  particulate in  the  atmosphere.    It is known  that  862
reacts  with other  species  in  the atmosphere  to form  various
sulfate  species  like ammonium  sulfate and  sulfuric  acid.  The
reaction rate is  affected by many factors,  such  as  meteorology
(sunlight,  temperature)  and  the  presence   of  other  reactant
species.  About  98  percent  of  the   sulfur  burned  in  diesel
engines  reacts  to  form  SO2,  only two percent  react  somewhere
in  the  engine  or   exhaust  system  to  form  directly  emitted
sulfates.   Therefore,  if  much  of the  SO2  emitted by  diesel
engines  reacts   in  the  urban  atmosphere   to   form  sulfate
particulate,  urban  particulate reductions due  to fuel controls
can  be  large.   For  example,  if  only  10  percent  of  the  S02
emitted  by  diesel  engines reacts  to form sulfate particulate in
urban  areas,  the amount  of  sulfate  particulate  formed in  the
atmosphere  is  five times  the amount of particulate  directly
emitted by diesel engines (at a two percent conversion rate).

     The first part of this  section discusses the approach used
in  this study  to  determine  the extent  of the reaction  of  S02
to  indirect  sulfates  in  urban areas.   (The  results  of this
analysis  have  already  been  used  in  the previous  section  to
quantify total  sulfate  and total  particulate reductions.)  The
second  part discusses the methodologies  used to develop ambient
concentration  changes of the  various pollutants.   The  final
part presents and discusses the air quality  results.

     A.     Indirect Sulfate Analysis

     There  are  two basic  questions   to  be  addressed  in this
analysis  -  how much  of  the S02 reacts  in urban areas  to form

-------
                              6-24


sulfate particulate, and what types of sulfate  particulate are
formed?  The  answers to  these  questions  help  to  quantify the
amount of  ambient  particulate  reductions  due  to  diesel  fuel
sulfur controls.

     This section  is divided  into  three parts.   The  first part
summarizes estimates that have been made by the California Air
Resources Board (GARB)  on  SO2  conversion,  and  discusses  some
of  the  theory  of  SC>2  transformation.    The  second  section
discusses   the   modeling   approach   used   to   estimate   S02
conversion  in  several  (10)  urban  areas  based  on  ambient  804
and  SO2  measurements.   The  third and  final section  presents
the  results  of  this  analysis,  and  estimates  the  mass  of
indirect sulfate  that  would  be  expected  to be  reduced by each
unit of mass of reduction of SC>2  emissions.  This has already
been  used  in   the   previous  section  to  characterize  indirect
sulfate reductions due to diesel fuel sulfur controls.

     1.    Background

     SO2 reaction  rates have  been studied  in  the Los Angeles
Basin  by  Cass.[9]   The  rates range from  about  6% per hour in
the spring,  summer,  and fall  to about 0.5% per hour  in  the fall
and winter.   At 3  percent  per hour, it only takes a little over
16  hours  for  the  rest of  the  SO2 to  react,  so  it  has  been
estimated that the  other  50  percent of  S02  that has  not been
lost  to  surface  deposition   reacts entirely  to  form  sulfates.
GARB  estimates that most of  this  ends up  as ammonium sulfate,
with one associated  water  molecule.[10]   On a mass basis,  each
pound  reduction  in  SC-2 is   expected  to  lead to a  1.16  pound
reduction in ambient (ammonium) sulfate particulate.

     It  has   been  widely    recognized,   however,   that   the
meteorological  conditions   affecting  SO2  conversion   (such  as
sunlight,   temperature,  humidity,   wind   speed,    and   other
pollutants  present)  in the   LA  basin  are quite different from
other  parts  of  the U.S.    For  example,  the  SOo  conversion
rates  used  by  EPA  in several  air  quality models  in  the draft
RIA  for  SC<2  National  Ambient  Air  Quality  Standards  ranged
from  0 to  2  percent per hour. [11]   These are  much  lower than
the  California conversion  rates.   Thus,  the  rates or  level of
SC<2   conversion  used   in   this   analysis   should   be  more
representative  of the  average meteorological  conditions  in the
U.S.

     2.    SO2 Reactions and  Products

     S02  follows   a basic   reaction  tree   that  starts  with
oxidation and  branches  off into a number of  reactions,  one of
which  is ammoniation.   For  the  purpose  of  assessing sulfate
particulate,   the  oxidation   and   ammoniation  steps   are  most
important.

-------
                              6-25
     S02   initially  reacts   with   oxygen  to   form   sulfur
trioxide,  or  803.   803  is  very  unstable  and quickly  reacts
with  water  vapor   (H20)  to   form  sulfuric  acid   (^804).
Sulfuric  acid  is  very  hygroscopic,  and  so  depending on  the
relative  humidity,  significant  quantities  of  water vapor  can
become  attached  to the  sulfuric acid molecule  and  agglomerate
into  a   particle  that  would  not  have  existed   if  the  ^804
had  not  been  present.   Next,   if  there   is  ammonia  present
(significant  sources   of  ambient   ammonia  come  from  decaying
plant matter  and  animal  waste),  the sulfuric  acid can  react
with  ammonia   to   form   ammonium  bisulfate    (NH4HSC>4)   or
ammonium   sulfate   (^4)2804).   The  ammonium   sulfates   are
less  hygroscopic   than   sulfuric   acid,   but  the   amount   of
associated water  is again related  to relative humidity.   These
reactions are summarized in Table 6-12.

     Methods used to sample these species  in the atmosphere are
limited  to  measuring  the  concentrations of the various  ions  -
804=, H+, and  NH4.   Unfortunately,- when  H+  is  measured,  it
cannot  be determined  whether  the  H+  ion  being  measured  was
associated  with   H2SC>4  or  NH4HS04,   so  estimates  of   the
quantity   of   species   present  are   always  subject   to  some
uncertainty.   A  few   recent  studies  have   been   conducted,
however,  on  sulfate species  in urban areas,  and most  show  a
predominance  of  the   (1^4)2804  species   over  sulfuric  acid
and  ammonium  bisulfate.   For example,  one study measured  the
sulfate  species in  Philadelphia  in  the daytime and  found  it to
be    about    80    percent     (^4)2804     and    20    percent
H2SC>4.[12]   A  nighttime  study  of   the   air  around  Houston
found the percentages to be 93 and  7 percent, respectively.[13]

     In  addition  to the uncertainty in sulfate  species,  there
is also uncertainty in the amount of  associated water  with each
molecule.   In  the  Regulatory Impact  Analysis  on the  National
Ambient  Air  Quality Standards for  Sulfur  Dioxide,  a  range was
developed  for  the  ratio  of  fine  particulate   mass  (sulfate
anions,  cations, and water) to sulfate mass.[11]   The amount of
water   associated   with  ammonnia   sulfate  and  bisulfate  was
estimated  in the  0  to  50 percent range.   The  ratio  ranged from
1.2  (100 percent  ammonium bisulfate  and no water)  to  1.7 (100
percent  ammonia sulfate and  60 percent water).  A value  of 1.6
was  used in  the RIA as  a  high estimate  (1.5 was the middle and
1.4  was  the  low  estimate).   The  previous  measurements  of
Philadelphia  and  Houston indicate,  however,  the  presence  of
some  sulfuric  acid in  urban areas.   At  50  percent  relative
humidity,   it   has  been  estimated   that  7  water  molecules
associate  with each sulfuric acid molecule. [14]   The fine PM
mass to sulfate mass  for this species is about 2.3.   Therefore,
this analysis will  use the fine mass  to  sulfate ratio  of  1.6 in
estimating  indirect sulfate  inventories  from the  reaction of
S02.

-------
          6-26





       Table 6-12





Simplified SOg Reactions





        Oxidation







       SO2 + 1/2 02 -> 803



         SO3  +  H2O •» H2SO4







      Anunoniation



        H2SO4 + NH3 -*• NH4HSO4





           or





      H2S04  +  2  NH3 * (HN4)2 SO4

-------
                              6-27


     The  above  analysis  has  established  the kinds  of  sulfate
species  that are  to  be expected  from  S02  reactions  and  the
mass of sulfate species per mass of sulfate  anion.   In order to
estimate  sulfate   inventories,  however,   the  extent  of  862
conversion to sulfate species must still be estimated.

     3.     Methodology for Estimating 80^ Conversion

     In  order  to   assess the  overall  conversion  of  862  to
sulfates  in  the  urban  atmosphere,  an  approach to  model  this
transformation  is   necessary.  Two  general  methods  exist  to
determine  the  values   of   SC>2  conversion-one  by  dispersion
modeling  and the other  by  modeling based on  measured  sulfate
and 802 concentrations.

     Several   photochemical   dispersion   models   have   been
developed to  predict  short-term ground level concentrations and
deposition   fluxes   of   one  or   two   gaseous  or  particulate
pollutants.   The  Pollution  Episodic- Model  Version  2  (PEM-2),
developed  by K. Shankar Rao,  is  one  such  model.[15]   Because
this  program  accounts   for  the  effects  of  dry  deposition,
sedimentation,  and  first-order   chemical  transformation,  it
possesses many capabilities,  including  a consideration of up to
300 isolated point  sources  and 50 distributed area sources.  As
a  result,  PEM-2  is  intended  for  studies  of  the  atmospheric
transport, transformation, and  deposition of acidic, toxic, and
other  pollutants   in  urban  areas  to  assess  the  impact  of
existing  or  new sources  or  source modifications on air quality
for  regulatory  purposes  and  urban   planning.    However,  a
necessary  input to the  model is  the  chemical transformation
rate of  SO2  to  sulfate,  and this  is  not  known to  any degree
of precision for various cities under typical conditions.  As a
result,  such models could not  be used  here; and  a different,
simplified method of conversion modeling is used in this study.

     A  simpler  method,  and   the one  used here  is  to calculate
SC>2  conversion  from  urban  ambient  sulfate  and  sulfur  dioxide
concentration data.   This   data  in  Hg/m3  can  be  converted
into molar  concentrations by dividing the  mass concentrations
by  the  molecular weights of  each  substance,  and estimating the
ratio  of sulfate  moles  to  total  sulfur  moles (S02  +  sulfate
moles).   Estimating SC>2  moles is straightforward;  however, to
estimate   sulfate   moles  requires  determining  what  sulfate
compound  is  present  in the  atmosphere/  since  only  the  864
anion  concentration   is  measured.   This  step,  however,  was
accomplished in the previous section.   The molecular weight of
a compound with a fine PM mass  to sulfate  ratio  of  1.6 is 1.6 x
98, or 156.8 grams  per gram-mole.

     This   simplified  procedure  for  SO2  conversion  requires
several  assumptions which could make the  conversion rates  thus
developed somewhat  uncertain.   First,  all  of the  sulfate is
assumed  to  come  from conversion  of  SC>2»  and  none  as  direct

-------
                              6-28
sulfate emissions.  As  such,  the conversions might  be somewhat
high,  since  indirect sulfate  concentrations would  be  lower if
direct  sulfate  emissions  were  subtracted.   However,  it  was
shown  in  the  last section that  indirect sulfates  probably far
outweigh  direct  sulfates,   even at  low  conversions  of  SC>2,
therefore, the  error  due to  this  assumption  is  probably not
large.  Secondly,  this  technique  does  not  attempt  to  account
for  the  influence of sulfate  and/or S02 from  other areas, nor
does  it explicitly account for deposition of either sulfate or
S02-   In  the  "simple"  urban area where  there  is  no  influx of
sulfate  or  SO2  from  surrounding  urban  areas,  one need  only
consider  the  relative  differences  in advection  and deposition
of  sulfate  and  SO2.    If  the  deposition  of   sulfate  is  much
higher  than  SC-2/  then  measured sulfate  will be   low and SC>2
conversions based on ambient data will be low.  The converse is
also  true.   Indirect  sulfate particles  are  extremely small, on
the order of  2 microns or less,  and  there is evidence that they
disperse  like   a gas.   Therefore  the rates of  deposition and
advection between S02  and  sulfate  particles  are  probably not
materially different,  so estimates  of  SO2  conversion  based on
ambient SC>2 and  sulfate in urban areas should be reasonable.

      In  the  more complex   situations,  sulfates  and  SO2  are
coming  into   the  urban  area  from  an  adjacent   urban   area.
Assuming  equal  deposition,  there is  a  good chance the sulfate
to  SC>2  concentration  is  higher  than   the  ratio  of the  area
from  which it  came due to  the  fact that  there  has  been  more
time  for  the various  sulfate reactions  to take place.  However,
both  the  sulfate and   sulfur   dioxide  concentrations  of  the
subject  urban   area  are  increased  by  the  emissions from the
adjacent  area.   The  apparent  S02   conversion  in   the  subject
urban  area  therefore   may   be  higher   than  the  actual  S02
conversion.

      The  above discussions  illustrate some  of the assumptions
involved  in  this modeling  approach.  While there  may be some
uncertainty in the  conversion rates  developed, by  selecting  a
variety of urban areas  and  looking at  the  range of conversion
rates, the analysis will yield results which can  be compared to
the GARB  analysis.

      4.    Results

      Sulfate  and sulfur  dioxide concentrations  were available
from   ten cities.[16,17]    These   were   average  concentration
measurements over the period of 1982-1984.   The  data are  shown
in the first two columns of  Table  6-13.   Sulfate concentrations
range  from    4.4    Ug/m3    in   Tucson    to   13   Hg/m3   in
Cleveland.     S02   concentrations   range   from    14    Ug/m3
(again,   Tucson)  to   64  Ug/m3  in  New   York   City.     Molar
conversion ratios estimated  from these  concentrations are  shown
in  the third  column.   They  range  from  a  low of 10 percent of
SO2  reacted  in  New York City to 21.4  percent in  Toledo.   The

-------
                         6-29


                     Table  6-13


       City and National SO? Conversion Ratios

                                                  Molar
                  Anibient Cone.  (Mg/m3) *       Conversion
City
New York
Denver-
Boulder
Kansas City
Nashville-
Davidson
All ent own
Toledo
Tucson
Columbus
Cleveland
Buffalo
National
Average
S04
11.00
6.90
6.03
8.23
10.17
11.47
4.43
12,03
13.00
9.70
^—
S02
63.69
30.54
34.03
30.54
40.03
27.92
14.83
42.75
52.35
42.75
^^
Ratio
.103
.132
.106
.152
.156
.214
.166
.158
.170
.131
.120
Averages of concentration measurements from 1982-1984.

-------
                              6-30


national  population  -  weighted   average  using  1980  census
populations for these areas is 12 percent.

     These molar  conversions  are  all  much  lower  than the  50
percent  estimated  by GARB  for  the  LA  Basin,  but this  in not
surprising,   since   they   are   based   on   yearly   average
concentrations which reflect  much  colder weather than typically
occurs in  the  LA basin.   An  862  conversion  rate of  12 percent
will be  used  in the remainder of  the  air quality analysis.  At
a  12  percent  conversion rate  and 1.6  Ibs of  fine particulate
mass  per  pound   of  sulfate,  about   0.288   Ibs   of  sulfate
particulate mass are produced for every pound of SC>2 emitted.

     B.    Air Quality Impacts

     Two basic methods  are  used  in  this section  to  estimate
changes  in  urban  concentrations  of  particulates,   S02/  and
ozone.   The   methods   used   to  estimate  concentrations  of
particulates  and  S02   are  very  similar  to  the  methods  that
were  used  in  establishing   the  heavy-duty   engine  NOx  and
particulate standards  and rely  on atmospheric  lead  monitoring
data  as  a  surrogate for  diesel  particulate   and  S02-Cl»2,3]
The  ozone  analysis  uses changes  in  urban  HC  inventories and
results  from   the  Empirical  Kinetic  Modeling  Approach  (EKMA)
model  to  estimate  ozone  concentration  changes  due  to  diesel
fuel controls.   The lead surrogate  methodology, and resulting
diesel   particulate   and  SO2   concentrations,   are   discussed
first.

     1.    Particulate and SO
-------
                              6-31


pollutant   levels.    An  estimate   is   made  of   the  fleet's
automotive  lead emissions  which  caused the  observed  ambient
lead  levels,  and  is  compared  to  the composite  particulate or
SC>2  emissions  developed  from  the  sources  covered  in  this
analysis.   These  composite  emissions  are estimated  by summing
the  annual  emission  inventories  of the different  sources  and
dividing  by  total highway   VMT.   The  result  is  an emission
factor in g/mi  that  includes stationary source  and off-highway
emissions,  but  is directly comparable to  the  lead emission
factor developed only from on-highway sources.   Very generally
speaking,  if  the  composite particulate  or  SC>2  emissions  in
2015  are  expected  to be  three  times automobile  lead emissions
in  1975,  for  example,  then ambient  composite  particulate or
SO2  concentrations  in 2015  can be expected  to  be  three  times
the  1975  ambient  lead  concentrations.   In this  case,  1975
monitoring data was chosen  over more recent data  to  avoid,  for
the  most  part,  the  errors   associated  with  estimating  the
leaded/unleaded vehicle mix.

     The basic mathematical expression of this methodology is:

C(P)cy = (E(P)cy/E(Pb)75)*(S(P)/S(Pb))*(VMTcyr/VMT75)*C(Pb)75

Where:

C(P)Cy      =   Ambient   concentration   of   pollutant  (Mg/m3)
                in a  calendar year (cyr)

E(P)Cy      =   Fleet-average pollutant  emission  factor  in the
                calendar year of interest (g/mi)

E(Pb>75         Fleet-average emission  factor for  lead in  1975
                (g/mi)

S(P)        =   Dispersion  factor  for  diesel  particulate  or
                S02 emissions

S(Pb)       =   Dispersion factor for lead emissions

VMTCy       =   Total  vehicle  miles  traveled  in  the calendar
                year

VMT75       =   Total on-highway mobile  source VMT  in  1975

C(Pb)75     -   Ambient lead  concentrations in 1975 Hg/m3

     The  above equation  is  generic  for urban  and rural areas,
so  E(P),  E(Pb),  VMT  and C(Pb)  must  all  be either  urban or all
rural, depending on  whether an urban or rural estimate  is being
made.

     Multiplying   E(P)Cy  by  VMTCy  (and  dividing  by  908,000
grams per ton),  however,  gives an estimate of the urban tons of

-------
                              6-32
the  pollutant  in  the  particular  calendar  year,  and  these
estimates are already available from  the  previous  section.   The
work  done  for  the  final heavy-duty  particulate  standards  and
previous work  done  by  EPA  developed estimates   of  the  lead
emission factor in  1975,  and the  dispersion factors of lead and
diesel particulate  (diesel  particulate was assumed to disperse
like  a gas  because  of  its  small size,  so  S02  will use  the
same  dispersion factor).   The lead  emission  factor  used  was
0.11  g/mi.    The  lead  dispersion  factor  was  determined to  be
0.43,  and   the  diesel  particulate  dispersion   factor   was
determined  to  be  1.00.   The  1975   urban VMT  is  758  billion
miles.[1]

     The use of  the  factors mentioned   above  results  in  the
following general equation:

C(P)cy =  P(tons,urban,cy) * 1.00  * C(Pb,1975)    *      l
                0.11 g/mi      0.43   7.58xlO"miles   1/908,000

or


  C(P)cy  =  P(tons,urban,cy) * C(Pb)75 urban
                   39,500

     For rural  areas, the  lead emission  factor E(Pb) has  been
estimated to  be the  same as for urban  areas  (0.11 g/mi).   The
1975 rural VMT can be estimated by multiplying urban VMT  by the
ratio  of rural to  urban VMT  for  1975  from the  MOBILES  Fuel
Consumption Model.[18]   In  this model,  rural  VMT  is about  76
percent  of  urban VMT.  Rural VMT is  estimated therefore  at 572
billion miles.  The equation for rural areas, then,  is

     C(P)cy   =  P(tons,rural,cy)  * C(Pb)75 rural
                      52,340

     At  this  point it  is necessary to consider  one additional
factor, which was  also used  in previous  analyses,  to make the
lead  ambient  concentrations  fully  compatible with mobile source
modeling.  This  factor  is used to correct the  lead air  quality
data  for the  fact  that only  an  estimated  89 percent  of  the
total  nationwide  lead  emissions  are  due to mobile  sources.
Therefore,  the ambient  lead concentrations were  multiplied by
0.89 to account for  the possible  contribution of non-automotive
sources.   This adjustment  is probably conservative,  though to
an  unknown  degree, because  only  lead monitors  in  areas of  no
known  large  stationary  source  of lead   were  chosen  for  this
analysis and  the majority of  the non-automotive sources  of  lead
emissions  reside in  a few  identifiable  areas which  have  been
excluded from this  analysis.

-------
                              6-33


     Therefore, the final equations become:

Urban;  C(P) = P(tons,urban) *  0.89  * C(Pb)75  urban
                               39,500

Rural:  C(P) = P(tons,rural) *  0.89  * C(Pb)75  rural
                               52,340

     Tables   6-14   through   6-17   present   the   lead-based
concentrations  for  carbon, total  sulfates, total  particulates
and  S02  for  each  fuel  control  scenario  and  four  groups  of
cities.   Two  calendar  years  are  shown,  1990  and 2015.   The
first two columns  show  concentrations for 1990 and 2015 without
fuel  controls.   The  rest  of  the  columns show  concentrations
with  fuel  controls for both  100  percent  and  NPRA segregation.
The cities  used and  their  corresponding lead values  are shown
in  Table 6-18  (other  data  are  also  shown  for  these  cities,
which  will  be  explained  in  the  section  on  visibility) .   The
rural  lead  value  comes  from the rural New York Bight area. [19]
(Distributions of  concentrations by  city size category  are  not
presented here,  but  can  be  estimated from the  coefficient of
variation  of  lead  concentrations,   which is  41  percent  for
cities  over 1,000,000, 35  percent  for cities  between  400,000
and  1,000,000,  31  percent  for  cities  between  200,000  and
400,000,  and 37 percent  for  cities  under 200,000.)  Generally,
pollutant concentrations  are  lower  in the  smaller  cities, with
pollutant concentrations  being  the  lowest in rural areas, where
they  are  only  about one-tenth of the values in  cities  over  one
million in population.

     Carbon  concentrations  in  Table 6-14  for  1990  range from
4.31   Ug/m3   to   0.43  Hg/m3.    They   are   only   slightly
higher  in  2015 with no  fuel controls.   This  is  the  result of
two counterbalancing  effects.   First, the 1991  and 1994 diesel
particulate  standards  cause a decline  in carbon particulate
from  on-highway  sources.   However,   this is  balanced  by  an
increase  in fuel  consumption  in both on-highway, off  highway
mobile,  and stationary diesel  sources  of carbon  particulate,
which  act  to  cancel  out  the  reductions in  on-highway carbon
particulate.   Sulfur  control raises  carbon particulate because
on-highway  engines are able  to increase  carbon  particulate as
direct  sulfate particulate  decreases with  sulfur controls  in
meeting   the  same  particulate  standards   (1991  and   1994).
Subsequent  aromatics  control  has  virtually no effect  on carbon
particulate concentrations,  due  to  only  a  small  change  in
emissions.

      Sulfate   concentrations  are  shown   in  Table  6-15.   The
levels  in  2015 with  no  fuel  controls  are  almost  double  the
concentrations   in  1990,  owing  to  the  increase   in  fuel
consumption.   For  sulfur  control  under 100 percent segregation,
2015  sulfate  concentrations  are significantly  lower than  the
levels  in  2015 without  fuel  controls,  but  still higher than

-------
                                      Table  6-14

                                Average Ambient Carbon
                         Concentrations by Population (Mg/n>3)
      City  Size
      Grouping
Greater than
  1,000,000

400,000 - 1,000,000

200,000 - 4,000,000

< 200,000

Rural
 No Fuel Control
1990  1994  2015


4.31  3.86  4.36

2.96  2.66  3.00

2.49  2.23  2.52

2.61  2.34  2.64

0.43  0.33  0.25
	Sulfur Control	
	100%	   —NPRA	
1994  2015  1994  2015
3.96  4.76  3.96  4.76

2.72  3.27  2.72  3.27

2.29  2.75  2.29  2.75

2.40  2.88  2.40  2.88

0.35  0.32  0.35  0.32
     Subsequent
—Aromatics Control	
	100%	   —NPRA	
1994  2015  1994  2015
3.87  4.77  3.88  4.77

2.67  3.27  2.67  3.28

2.24  2.75  2.24  2.76

2.35  2.88  2.35  2.89

0.33  0.32  0.33  0.32
cr»
 I

-------
                                         Table 6-15
      City  Size
      Grouping
Greater than
  1,000,000

400,000 - 1,000,000

200,000 - 4,000,000

< 200,000

Rural
                                Average Ambient Total Sulfate
                            Concentrations by Population (Mg/m^)
No Fuel Control
1990 1994
4.
3.
2.
2.
0.
49
09
59
72
42
4.
3.
2.
2.
0.
89
37
83
96
45
2015
8
5
4
5
0
.51
.86
.92
.15
.78
—Sulfur Control—
-100%	   —NPRA-
                   1994  2015  1994  2015


                    3.31  5.74  0.98  1.74

                    2.28  3.94  0.67  1.19

                    1.91  3.31  0.57  1.00

                    2.01  3.47  0.59  1.05

                    0.21  0.37  0.90  0.16
     Subsequent
—Aroraatics Control	
	100%	—NPRA	
1994  2015  1994  2015
                         3.31  5.74  0.98  1.74

                         2.28  3.94  0.67  1.19

                         1.91  3.31  0.57  1.00

                         2.01  3.47  0.59  1.05

                         0.21  0.37  0.09  0.16
                                     o\
                                      I
                                     CO

-------
                                       Table 6-16
      City  Size
      Grouping
Greater than
  1,000,000

400,000 - 1,000,000

200,000 - 4,000,000

< 200,000

Rural
                           Average Ambient Total Particulate
                          Concentrations by Population
 No Fuel Control
1990  1994  2015


10.60  10.3  14.48

 7.29  7.07   9.96

 6.13  5.94   8.37

 6.42  6.22   8.77

 1.02  0.91   1.13
	Sulfur Control	
	100%	   --NPRA	
1994  2015  1994  2015
 8.83  12.16 6.50  8.16

 6.08  8.37  4.47  5.62

 5.10  4.03  3.76  4.72

 5.35  7.37  3.94  4.95

 0.69  0.79  0.57  0.58
     Subsequent
—Aromatics Control	
	100%	   —NPRA	
1994  2015  1994  2015
8.65  12.14 6.31  8.15

5.95  8.35  4.34  5.61

4.99  7.02  3.65  4.71

5.24  7.36  3.82  4.94

0.67  0.79  0.54  0.58
CTl
 I
CO

-------
                                      Table  6-17
                                  Average Ambient 802
                         Concentrations  by  Population (Mg/m^)
      City  Size
      Grouping
Greater than
  1,000,000

400,000 - 1,000,000

200,000 - 4,000,000

< 200,000

Rural
 No Fuel Control
1990  1994  2015


10.96  11.99 20.92

 7.54  8.24  14.39

 6.34  6.93  12.09

 6.65  7.27  12.67

 1.01  1.10   1.92
	Sulfur Control	
	100%	   —NPRA	
1994  2015  1994  2015
 8.12  13.98 2.40  4.19

 5.59  9.62  1.66  2.88

 4.68  8.08  1.38  2.42

 4.82  8.47  1.46  2.53

 0.52  0.90  0.22  0.38
     Subsequent
—Aromatics Control	
	100%	   —NPRA	
1994  2015  1994  2015
8.12  13.98 2.40  4.19

5.59  9.62  1.66  2.88

4.62  8.08  1.38  2.42

4.92  8.47  1.46  2.53

0.90  0.90  0.22  0.38
a\
I

-------
                             6-38
                          Table  6-18
        City Lead Values, Baseline Visibilities, Radii
City Category (by pop)
 Ambient
   Lead
Conc.(Mg/m3)
                 Baseline
                Visibility
                  (mi)
               City
            Radius(mi)
> 1,000,000
Houston
Los Angeles
New York
Philadelphia
 2.09
 2.68
 1.05
 1.34
                  18
                  11
                  15
                  15
              13.3
              12.2
               9.8
               6.6
400,000 to 1,000,000
Atlanta
Boston
Denver
Kansas City, MO
New Orleans
Phoenix
Pittsburgh
San Diego
St. Louis
 1.05
 0.92
 1.59
 0.80
 1.06
 2.10
 0.85
 1.13
 1.58
                  12
                  21
                  75
                  17
                   9
                  58
                  10
                  25
                  13
               6.5
               3.9
               6.4
              10.0
               8.0
              10.1
               4.2
              10.1
               3.2
200,000 to 400,000
Birmingham
Cincinnati
Jersey City
Louisville
Oklahoma City
Portland
Tucson
 1.22
 0.81
 1
 0
 1
  03
  96
  66
0.81
0.75
10
11
15
10
17
24
60
 5.6
 5.0
 2.0
 4.4
13.9
 5.7
 5.6
Less than 200,000
Mobile
New Haven
Salt Lake City
Spokane
Trenton
Waterbury
Yonkers
 0.96
 1.15
 0.98
 0.58
 0.88
 1.88
 1.16
                  10
                  17
                  80
                  36
                  17
                  17
                  15
               6.3
               2.5
               4.9
               4.1
               1.5
               3.0
               2.4
Rural
 0.13
                  30
               N/A

-------
                             6-39
1990  levels.   Sulfate  concentrations  are  lowest  for  sulfur
controls  with  NPRA   segregation.    There  is  no   effect  of
aromatics control  on sulfate concentrations.

     Total particulate concentrations are  shown in  Table  6-16.
Significant   reductions   in   particulate   concentrations   are
obtained  with sulfur  control  for  both  maximum  and  minimum
segregation cases.  The impacts of  subsequent  aromatics control
on total particulate reductions are extremely small,  however.

     SO2  concentrations   are   shown  in  Table  6-17.   The  SO2
concentrations essentially follow the same  trends as  sulfate PM.

     2.    Ozone Concentrations

     Section  I of  this  chapter developed  urban HC reductions
due to diesel fuel controls.  The purpose of this  section  is  to
estimate the impacts of those reductions on urban ozone levels.

     The  model  used  in  this  analysis  to  estimate  changes  in
ozone  concentrations  from  changes  in  HC  inventories  is  the
Emperical  Kinetic Model,  or  EKMA.   Instead  of running  EKMA,
however, the urban  HC  reductions will  be  compared  to  urban  HC
reductions   from    the   proposed   regulations   on   gasoline
volatility, which used EKMA directly in estimating  the effects
such changes would have on urban ozone levels.

     The  RIA  for  volatility  controls  included  analysis  of
nationwide inventories for  a  variety of scenarios such  as  with
and without  onboard refueling requirements, analyses that were
based  on  design-value  versus  July  average  temperatures,  and  a
few others. [20]   The  choice of scenario is not important;  what
is important is the change  in  the ozone levels  that  corresponds
to  a  specific change  in  inventories.   This analysis  will use
the  HC   emission  inventories  for  61  non-attainment  areas
(equivalent to "urban" here)  volatility controls at 9.0 and 9.5
psi   RVP,   assuming   a    prior   onboard   refueling   control
requirement, and  based on design-value  temperatures  (Table 3-25
of  the RIA).  In that  analysis,  the  inventory  with  9.5  RVP
control  in  2000  is 6.144  million  tons,  and  the  9.0  psi RVP
inventory  is  6.113  million tons,  for  a net  difference of RVP
31,000  tons.   The reduction from 1983  ozone levels  for 9.5 RVP
control was  10 percent,  and for 9  RVP  control was  11 percent.
However,  there clearly could be round-off  error involved,  since
no  change  in average  ozone  levels  was   seen  between  other
similar control levels.

     The  HC  reductions due to diesel fuel  controls  in 2010 are
at  most 30,000 urban tons,  which is about  the same  as the  HC
reduction  due to  the 9 RVP versus  9.5  RVP  volatility controls.
The HC  reductions due to  diesel  fuel controls,  therefore,  would
be  expected to bring  about at most a  one percent  decrease  in
ozone  levels.

-------
                             6-40
III .  Visibility Assessment

     Section  II  developed  estimates  of  the  changes  in  the
concentrations  of  different  types   of   particulate  (carbon,
sulfates)  for  the  fuel  control  scenarios.   Because  particles
scatter  and  absorb light,  changes are  expected  in  visibility
with  the changes  in  particulate  concentrations.  Furthermore,
different types of particulate have  different effects  on  light
scattering  and  absorption.   In  section . II,   there  was  an
increase  in  carbon concentrations,  but  a  large reduction  in
sulfate  concentrations.  The  purpose  of  this   section  is  to
estimate the net visibility changes due to diesel fuel controls.

     This  section  is   divided  into two  parts.   The  first  part
summarizes  the  visibility  methodology  used  in  setting  the
heavy-duty particulate  standards,  which  is also  used  here,  and
the  modifications  made  for  this  analysis.   The second  part
presents  the  changes  in  visibility due to  fuel  controls  using
this methodology.

     A.    Methodology

     The methodology used in  setting  the heavy-duty particulate
standards is based on  an  analysis of the visual properties of
air  and  the  Beer-Lambert  law,  in  which the reduction in light
intensity  is   a   function  of   distance  and   the   extinction
coefficient (bext)of  the media:

     I  =  I0 e~bextL

Where

     Io  =  light intensity at object being observed,  and
     I   =  light intensity at distance L from the object. [1]

     This method also used Koschmieder "s Law, which  states that
objects  become  invisible to the human eye  when  the  ratio of  I
to Io becomes 0.02.[1]   The above equation then becomes
     LV =  3.91
             bext
where

          = visual range, and
          = total extinction coefficient of the media

     The  most important  parameter of  these  equations  is  the
total  extinction coefficient  (bext) .   This coefficient  is the
sum of four components:

     1.    scattering by gas molecules
     2.    absorption by gas molecules
     3.    scattering by particles
     4.    absorption by particles

-------
                             6-41


     The impact of fuel control on items 1 and  2  is negligible,
therefore, this analysis is primarily  concerned with scattering
and  absorption  by particles,  the particles  of  interest  being
sulfates  and  carbon participates .   SOF emissions,  even  though
absorbed  onto  the  carbon  particulate   are  assumed  to   be
transparent  (as  they  are  when  in  the  gaseous  phase)  and
therefore do not affect visibility.  Also, as will  be discussed
subsequently,  the  extinction  coefficient  used  in this analysis
will be for fine elemental  carbon, which does not include SOF.

     The equations given above  assume  that the particles  which
absorb  and  scatter light are  evenly distributed in  the  visual
range.  This is not  necessarily the case  in urban  areas  where
the  visual  range  may extend beyond the area  affected by diesel
particulate.    The   methodology   used   in   the   heavy-duty
particulate   standards    analysis   assumed   that    the   PM
concentrations estimated in  Section II applied within but  not
outside the city  radius, and  derived two  new equations from the
above equations to account for  this.'  The first  equation  below
was   used   for   urban  areas  or  cities  where   the  baseline
visibility  extended  beyond  the  radius of  the  city, and  the
second  equation was  used where the city radius was  equal  to or
larger than the baseline visibility.

 LV - (18. 6xlO-4 [miles/m] - bext dp Mc La)/bo
(1)

 Lv - 18.6xlO-4[miles/m]/[18.6xlO-4/Lvo)+bext/
-------
                             6-42
     The  concentrations  of  carbon  and  sulfate  are  available
from  section  II.    The  extinction  coefficient  of  elemental
carbon  is  11.5 m^/q  and the  extinction coefficient  used  for
sulfates  is  3.5   rn^/g  (assuming  a mean  particle  radius  of
0.1-0.6 urn for the  latter).[1]   Baseline visibilities  and  city
radii are the  same  as  these used in the previous  analyses,  and
are shown in Table 6-18.

     For rural areas,  baseline visibility was not  available,  so
this  analysis  used 30 miles,  which falls  in  the range  of  the
cities less  than 200,000.  In most rural areas of  the  west,  the
baseline visibility  is  probably  significantly greater  than  30
miles,  but  in  other  areas  of  the midwest,  east,  south  and
northwest  where   relative   humidities  are  much  higher   the
baseline visibility  may be  significantly  less  than  30  miles.
The  absolute  baseline  visibility  is  not  that  significant,
however, as  the analysis will be  focusing on  percentage changes
in visibility, rather than on their absolute values.

     Also concerning the analysis  for  rural areas, the affected
radius was assumed to extend beyond  the  baseline visibility (30
miles),  as  there  is  probably a  fairly  even distribution  of  a
low concentration of particulate emissions in these areas.

     The results  of  the visibility  analysis  for  fuel  controls
are  shown  in  Table  6-19.    The  estimates  shown  are percentage
changes  in  visibility  from  the  base  case  in  1990  to  each
control  case  in   2015.   Positive  values  are   improvements  in
visibility,   negative   values   are  reductions.   Results   are
disaggregated  by  city  size  category  and by  level  of  fuel
segregation.   The  base  case  shows  percentage   changes   in
visibility  due  to  on-highway,   off-highway  and  other  diesel
sources  from  1990 to  2015   with  no fuel  controls.    All  other
fuel  control  cases  measure  percentage  changes  in  visibility
from 2015 with controls to 1990 without controls.

     In  the base  case, visibility is  projected  to  decrease
between  1.8  and  11.2  percent from  1990  to 2015.   Rural areas
show a small improvement (1.1 percent) in visibility.   There is
a reduction in particulates  from on-highway sources (because of
the particulate standards)  which improves  visibility,  but  this
is  outweighed  by the  negative  impact  of  the  growth  of  SC>2
emissions (and thus,  indirect  particulates)in  all  sources  in
the  absence  of fuel controls.  Rural  areas show an improvement
of visibility due to improvement  of  on-highway  emissions, which
are mostly (67 percent) rural.

     An  improvement in visibility is noted  with sulfur controls
with  100 percent  segregation over the base case.   This  is  due
to  lower direct and indirect  sulfate  emissions  from on-highway
vehicles.   There  is   no change  in  visibility  for  aromatics
control.

-------
                             6-43
     A significant  further  improvement  in  visibility  is  noted
if diesel  fuel  sulfur  level is lower  for off-highway and other
diesel sources as well  as  on highway vehicles, as  shown  in the
percentage changes  in  visibility for  minimum  segregation.   The
positive values  indicate  an improvement  in visibility in  2015
over 1990 levels.

-------
                              6-44
                            Table 6-19

             Average Percentage Changes in Visibility
             From 1990 to 2015 by City Size Category
    City  Size
>1 million

400,000 1 million

200,000-tOO,000

<2JO,000

Rural
       -100% Segregation-

 Base   Sulfur  Aromatics

-11.2    -7.4     -7.5

 -5.1    -3.3     -3.4

 -4.4    -2.9     -2.9

 -1.8    -1.2     -1.2

  1.1     2.1     - 2.1
-NPRA Segregation-

 Sulfur    Aromatics

  3.4        3.3

  1.6        1.5

  1.4        1.3

  0.5        0.5

  3.2        3.2

-------
                             6-45
                    References  (Chapter 6)

     1.    "Diesel Particulate  Study,"  SDSB,   ECTD,  QMS,  OAR,
EPA, October 1983. Docket A-80-18.

     2.    "Draft  Regulatory  Impact  Analysis  and  Oxides  of
Nitrogen Pollutant Specific  Study," SDSB, ECTD,  QMS,  OAR,  EPA,
October 1984.   Docket  A-80-18.

     3.    "Regulatory  Impact  Analysis,  Oxides   of   Nitrogen
Pollutant Specific Study  and Summary and  Analysis of Comments,"
ECTD, QMS,  OAR, EPA, March 1985.  Docket A-80-18.

     4.    "The   Motor   Fuel   Consumption   Model,   Thirteenth
Periodical  Report," prepared for DOE by  Martin Marietta Energy
Systems, Inc., and Energy and Environmental  Analysis,  Inc., May
26, 1987.

     5.    "Heavy-Duty Vehicle  Emission  Conversion  Factors II,
1962-2000," Paul Machiele, EPA-AA-SDSB-89-01, October 1988.

     6.    MOBILE4   Travel   Characterization   Data   Handout,
MOBILE4 Workshop, November 1987.  EPA, OAR,  QMS, ECTD.

     7.    "Compilation   of  Air  Pollutant   Emission  Factors,
Volume  I:  Stationary  Point and Area  Sources,"  4th  Edition,
AP-42, September 1985.

     8.    "State  and Metropolitan  Area Data Book  -  1982,"
Bureau of Census, Dept. of Commerce, August 1982.

     9.    "Sulfate Air   Quality  Control Strategy Design,"  by
Glen  R.  Cass,  Atmospheric  Environment,  Vol.  15,  No.   7,  pp.
1227-1249,  1981.

     10.   "Review of  Motor  Vehicle  Diesel Fuel  Modification
Emission Reduction Strategies," SSD, GARB, October 1986.

     11.   "Regulatory Impact  Analysis  on the  National Ambient
Air  Quality  Standards   for   Sulfur  Oxides  (Sulfur  Dioxide)
(Draft)," SASD, OAR,  EPA, May 1987.

     12.   "A Composite Receptor  Method Applied to Philadelphia
Aerosol," by Dzubay,  T.G.,; Stevens,  R.K.;  Gordon,  G.E.; Olnez,
I.;  Sheffield,  A.E.;  and Courtney,  W.J.,  submitted to Environ.
Sci. Techno1., Vol. 22, No. 1,  pp. 46-52,  1988.

     13.   "Visibility  and  Aerosol Composition  in  Houston,
Texas,"  by Dzubay, Thomas  G.;  Stevens,  Robert K.;  and Lewis,
Charles  W., Environ.  Sci. Techno1., Vol.  16,  No.  8,  514-524,
1982.

-------
                              6-46
     14.    "An Acid  Aerosols  Issue  Paper:   Health  Effects  and
Aerometrics," EPA, OHEA, ECAO, May 1987.

     15.    "User's Guide  for PEM-2:   Pollution  Episodic  Model
(Version 2),"  prepared for  the U.S.  EPA by  Rao,  K.  Shankar,
NOAA, Contract No. 1AG-DW13930021-01-1, 1984.

     16.    EPA  Memorandum  from Jake  Summers,   RIS,  NADS,  to
Angela Lindner, January 29, 1987.

     17.    "National  Air  Quality  and  Emission  Trends  Report,
1986," OAQPS, EPA, EPA-450/4-88-001,  February 1988.

     18.    "MOBILES  Fuel  Consumption  Model,"  Mark  A.  Wolcott,
U.S. EPA,  OAR, OMS,  and  Dennis F.  Kahlbaum,  Computer  Sciences
Corporation, February 1985, EPA-AA-TEB-EF-85-2.

     19.    "Air Quality Criteria  for Lead",  EPA 600/8-77-017,
Health Effects Research Lab, RTP, NC,. December 1977.

     20.    "Draft  Regulatory  Impact   Analysis,   Control   of
Gasoline Volatility  and Evaporative  Hydrocarbon  Emissions from
New Motor Vehicles," OMS,  OAR, EPA, July 1987.

-------
                         APPENDIX 6-A

       DIESEL FLEET VMT,  FUEL CONSUMPTION,  REGISTRATIONS
         FUEL ECONOMY,  AND ANNUAL UMT/Veh CALCULATIONS


     The calculations  involved  in  this report  required  diesel
fleet  fuel  consumption  data, diesel   fleet  annual  VMT  data,
diesel annual  VMT  per  vehicle data,  diesel fleet  fuel  economy
data,  and diesel  fleet registration data  which were  accurate,
and  consistent for  all  five  categories.   Since  the DOE  13th
Periodical   Report  appeared   to  provide  a   fairly  accurate
estimate of  total  fuel  consumption,   it  .was  used  as much  as
possible as  a basis for the following calculations.[4]

Fuel Economy

     Fuel economy values for  the years through 2020  were taken
from the MOBILES  fuel  consumption  model  with  the  exception  of
those  for  HHDDTs.[18]   For  the years  2021  through 2023,  the
fuel  economies were  assumed  to remain  constant  at the  2020
value.   Since  the  fuel  economy for  the  HHDDTs  was  available
from  the DOE  13th  report  through  the  year 2000,  it  was taken
from there to  remain  as  consistent  as possible with  the total
fuel consumption data.   In order to project it out  to 2023,  it
was assumed to increase in proportion to that for the HHDDVs  in
the  MOBILES  fuel  consumption model.   The fuel economy  values
for all  classes are shown  in Table 6-A-l.  The average  for all
buses  was drawn from the  MOBILES  information  by weighting the
bus categories by their fleet  registrations (which  is discussed
later).

Fleet VMT

     Fleet  VMT   through  the   year   2000 was  determined  by
multiplying   the   class    specific   fuel   economy   estimates
determined  above,   by  the  diesel   fuel  consumption  values for
these  classes  found   in   the  DOE  13th  report. [4]   The  13th
report,  however/  contained  no  information  on  buses.   As  a
result diesel  bus  fleet VMT  was determined by multiplying bus
fleet  registrations by  the  average annual VMT  per  bus.   The
derivation of  these factors  is discussed below.  For  the years
2001  through  2023  fleet VMT values  were  calculated by assuming
an   annual   increase   of  three  percent.   The  diesel  fuel
consumption values for all the classes are shown in Table 6-A-2.

Fleet  Registrations

     Once again,  to remain  consistent with the  assumptions  of
total  fuel  consumption,   the diesel  fleet registration  values
through  the year  2000  for all vehicle classes except buses were
taken  from the 13th periodical report.   The projections for the
2001  to  2023  time frame were  derived  by  dividing  the fleet VMT
values by the estimates of average  annual VMT per vehicle.

-------
                                6-A-2
                            Table  6-A-l

                  Diesel Fleet Fuel Economy  (mpg)
Year

1975
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
  LDDV

21.85
28.26
28.76
29.32
29.872
30.4
30.924
31.412
31.858
32.287
32.707
33.114
33.441
33.723
33.953
34.147
34.314
34.444
34.566
34.66
34.737
34.798
34.84
34.881
34.907
34.925
34.938
34.948
34.949
34.956
34.958
34.957
34.957
34.957
34.957
LDDT
20
23.57
23.76
23.957
24.134
24.277
24.438
24.644
24.842
25.079
25.305
25.543
25.738
25.901
26.041
26.153
26.25
26.333
26.405
26.464
26.518
26.562
26.598
26.627
26.654
26.673
26.693
26.702
26.712
26.717
26.717
26.722
26.722
26.722
26.722
LHDDT
8.235
14.04
14.195
14.353
14.511
14.667
14.846
15.01
15.19
15.381
15.562
15.752
15.918
16.078
16.199
16.372
16.469
16.56
16.633
16.695
16.753
16.789
16.817
16.843
16.869
16.887
16.905
16.904
16.914
16.919
16.918
16.921
16.921
16.921
16.921
MHDDT
6.069
7.285
7.37
7.463
7.544
7.628
7.707
7.783
7.858
7.933
8.003
8.07
8.126
8.167
8.196
8.222
8.238
8.254
8.262
8.272
8.281
8.286
8.291
8.294
8.296
8.301
8.303
8.304
8.308
8.309
8.31
8.31
8.31
8.31
8.31
HHDDT
4.35
5.933
6.024
6.118
6.214
6.312
6.411
6.509
6.606
6.689
6.76
6.82
6.901
6.965
7.012
7.049
7.076
7.097
7.113
7.125
7.135
7.142
7.148
7.152
7.154
7.156
7.158
7.160
7.161
7.162
7.163
7.163
7.163
7.163
7.163
  Buses
4.78
6
6
  06
  16
6.26
6.39
6.51
6.62
6.73
6.84
6.97
7
7.
7
7
7.
  08
 ,20
 ,25
7.29
7.34
7.40
7.44
7.49
  54
  59
7.63
7.67
7.72
7.76
7.81
  85
  89
  94
7.97
8.02
8.06
8.06
8.06
8.06
8.06
7,
7,
7,

-------
      6-A-3
   Table  6-A-2



Annual Fleet VMT
Year
1975
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
LDDV
4.37E+08
1.89E+10
1.75E+10
1.64E+10
1.55E+10
1.49E+10
1.47E+10
1.49E+10
1.52E+10
1.56E+10
1.61E+10
1.68E+10
1.73E+10
1.78E-I-10
1.83E+10
1.89E+10
1.95E+10
2.00E+10
2.06E+10
2.13E+10
2.19E-I-10
2.26E-I-10
2.32E+10
2.39E-I-10
2.47E+10
2.54E+10
2.62E-HO
2.69E+10
2.77E+10
2.86E+10
2.94E+10
3.03E+10
3.12E+10
3.22E+10
3.31E+10
LDDT
O.OOE+00
8.98E+09
9.39E+09
9.80E+09
1.02E+10
1.06E+10
1.09E+10
1.12E+10
1.15E+10
1.20E+10
1.24E+10
1.28E+10
1.32E+10
1.36E+10
1.40E+10
1.44E+10
1.49E+10
1.53E+10
1.58E+10
1.62E+10
1.67E+10
1.72E+10
1.77E+10
1.83E+10
1.88E+10
1.94E-HO
2.00E+10
2.06E+10
2.12E+10
2.18E+10
2.25E+10
2.32E+10
2.39E+10
2.46E+10
2.53E+10
LHDDT
O.OOE-t-00
1.45E+10
1.63E+10
1.79E+10
1.94E+10
2.09E+10
2.23E+10
2.36E+10
2.49E+10
2.61E4-10
2.73E+10
2.85E+10
2.94E+10
3.03E+10
3.12E+10
3.21E+10
3.31E+10
3.41E+10
3.51E+10
3.61E+10
3.72E+10
3.83E+10
3.95E+10
4.07E+10
4.19E+10
4.31E+10
4.44E+10
4.58E+10
4.72E+10
4.86E-I-10
5.00E+10
5.15E+10
5.31E+10
5.47E+10
5.63E+10
MHDDT
7.28E+08
4.09E+10
4.17E+10
4.28E+10
4.38E+10
4.51E+10
4.64E+10
4.80E+10
4.98E+10
5.18E+10
5.39E+10
5.63E+10
5.80E+10
5.98E+10
6.15E+10
6.34E+10
6.53E+10
6.72E+10
6.93E+10
7.13E-flO
7.35E+10
7.57E+10
7.80E+10
8.03E+10
8.27E+10
8.52E+10
8.77E+10
9.04E+10
9.31E+10
9.59E+10
9.88E-I-10
1.02E-I-11
1.05E+11
1.08E+11
1.11E+11
HHDDT
4.73E+10
6.58E+10
6.72E+10
6.85E+10
6.99E+10
7.13E+10
7.29E+10
7.46E+10
7.64E+10
7.84E+10
8.05E+10
8.27E+10
8.52E+10
8.78E+10
9.04E+10
9.31E-I-10
9.59E+10
9.88E-I-10
1.02E+11
1.05E+11
1.08E+11
1.11E+11
1.15E+11
1.18E+11
1.22E+11
1.25E+11
1.29E+11
1.33E+11
1.37E+11
1.41E+11
1.45E+11
1.49E+11
1.54E+11
1.59E+11
1.63E+11
Buses
2.91E+09
5.42E+09
5.67E+09
5.94E+09
6.17E+09
6.45E+09
6.74E+09
7.06E+09
7.35E+09
7.67E+09
8.01E+09
8.34E+09
8.59E+09
8.85E+09
9.11E+09
9.39E+09
9.67E+09
9.96E+09
1.03E+10
1.06E-I-10
1.09E+10
1.12E+10
1. 15E+10
1.19E+10
1.22E+10
1.26E+10
1.30E+10
1.34E+10
1.38E+10
1.42E+10
1.46E+10
1.51E+10
1.55E+10
1.60E+10
1.65E+10

-------
                               6-A-4


         Since  buses  were  not   included  in  the  13th  periodical
report, fleet  registration values had to  be derived  independently.
The values  in the  MOBILES fuel  consumption model for  public  buses
were assumed  to  be representative of  transit and  commercial  buses.
However,  the  estimates  for diesel school  buses appeared  to  grossly
underestimate  the  recent  trends  in  sales  of  diesel  school  buses.
Diesel  sales  data  for  school  buses  from  the  MOBILE4  conversion
factor analysis  was  used  to  estimate  the  fraction  of  school  buses
which  were  diesel for  the years  through  1990.  The  assumption  was
then made that 20 years  after all school bus sales were diesel that
all  school  bus  registrations  were  diesel (The  year  2019).   To
estimate the  fraction of  school  buses  which  were diesel  for  the
years  1991 to  2023,  a linear interpolation between the 1990  and 2019
values was performed.   Diesel school  bus  registrations for  1991  to
2023 were  then determined by multiplying these diesel  registration
fractions by  the total school bus registrations available  from  the
MOBILES  fuel  consumption  model.   The  school bus  registrations plus
the public bus registrations were then  combined to  form the  total
bus  fleet  registrations.  The diesel  fleet  registration  values  for
all classes are shown in Table 6-A-3.

Average Annual VMT per Vehicle

         For  all vehicle   classes  except   for  buses,  the  average
annual VMT per vehicle values through the  year  2000  were calculated
by  dividing  the  fleet  VMT  estimates by   the  fleet  registrations
(which were  discussed above).   In  order  to take  into  account  the
effects of changes  in diesel penetration  into  a vehicle  class,  the
estimates for  the years  2001 through  2023  were obtained by changing
the  2000  values  proportionately  to  an   independent  estimate  of
average  annual  VMT   per   vehicle  which  was  based  on MOBILE4  and
MOBILE4   conversion   factors   information.    This   estimate   was
determined   by   multiplying    the    MOBILE4   fleet   registration
distributions  by vehicle  age by  estimates  of  model  year  specific
diesel sales  fractions,  and  renormalizing  these products to  get  an
age  specific  diesel  registration  distribution.  These in turn were
multiplied by  estimates  of the age  specific annual VMT per  vehicle.
When  summed  together  across  all  vehicle   ages,  this  results  in  a
calendar  year  specific  average  annual  VMT  per  vehicle  for  each
vehicle class.

         For  buses,  average  annual  VMT  per vehicle  estimates  for
public buses,  and school  buses  were taken  from the  Federal Highway
Administration's  Highway  Statistics  Book,  and  weighted  together
based  on  projected   diesel   registrations  in  each  class.   This
resulted in  a  weighted average annual VMT per  bus  for each  calendar
year.   Average  annual  VMT  per   vehicle  values for  all  classes  of
diesel vehicles are shown  in Table 6-A-4.

-------
            6-A-5
         Table 6-A-3



     Fleet Registrations



LDDT       LHDDT      MHDDT
           HHDDT
           10000
50000
1030000
             Buses
94000
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
1739000
1647000
1557000
1475000
1403000
1347000
1313000
1292000
1279000
1273000
1278000
1304342
1337082
1400006
1450056
1497873
1563562
1625492
1693813
1760087
1812889
1867276
1923294
1980993
2040423
2101635
2164685
2229625
2296514
2365409
2436372
2509463
2584747
2662289
1021000
1085000
1147000
1207000
1265000
1320000
1372000
1422000
1471000
1518000
1564000
1641967
1700702
1769603
1830334
1885771
1963019
2039779
2121129
2204005
2270126
2338229
2408376
2480627
2555046
2631698
2710649
2791968
2875727
2961999
3050859
3142385
3236656
3333756
950000
1089000
1233000
1375000
1514000
1649000
1782000
1910000
2035000
2154000
2269000
2358385
2441790
2524688
2608796
2695733
2783637
2871590
2961702
3053785
3148289
3245032
3342383
3442655
3545934
3652312
3761882
3874738
3990980
4110710
4234031
4361052
4491883
4626640
947000
991000
1036000
1082000
1130000
1178000
1228000
1280000
1331000
1382000
1433000
1480953
1529727
1577892
1628111
1678278
1729194
1782360
1836918
1892836
1950392
2009566
2070331
2132863
2197138
2263277
2331252
2401189
2473225
2547422
2623844
2702560
2783636
2867146
1324000
1340000
1356000
1372000
1387000
1406000
1425000
1447000
1470000
1495000
1522000
1567660
1614690
1663130
1713024
1764415
1817348
1871868
1928024
1985865
2045441
2106804
2170008
2235108
2302162
2371226
2442363
2515634
2591103
2668836
2748901
2831368
2916309
3003799
296000
318000
342000
365000
391000
417000
446000
474000
504000
536000
568000
594391
622687
651687
679997
710537
742882
774353
808324
841849
877422
914193
952099
991619
1030967
1071352
1114242
1157985
1203951
1250782
1288526
1327182
1366997
1408007

-------
              6-A-6
           Table 6-A-4



       Annual VMT/Vehicle
  LDDT
N/A
  LHDDT
N/A
  MHDDT
14566
  HHDDT
45950
  Buses
30957
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
10888
10652
10545
10511
10617
10928
11340
11737
12193
12667
13137
13258
13321
13104
13031
12994
12821
12703
12556
12446
12446
12446
12446
12446
12446
12446
12446
12446
12446
12446
12446
12446
12446
12446
8795
8650
8543
8438
8348
8220
8137
8106
8149
8168
8199
8044
7999
7918
7885
7883
7800
7731
7658
7591
7591
7591
7591
7591
7591
7591
7591
7591
7591
7591
7591
7591
7591
7591
15311
14964
14551
14142
13805
13550
13258
13019
12834
12679
12572
12459
12394
12347
12307
12268
12237
12218
12201
12189
12177
12169
12169
12169
12169
12169
12169
12169
12169
12169
12169
12169
12169
12169
43233
42123
41270
40516
39875
39425
39105
38903
38902
39036
39303
39171
39060
39003
38934
38904
38891
38863
38840
38823
38808
38795
38786
38778
38773
38769
38768
38768
38768
38768
38768
38768
38768
38768
49722
50130
50546
50944
51424
51853
52355
52830
53330
53836
54367
54367
54367
54367
54367
54367
54367
54367
54367
54367
54367
54367
54367
54367
54367
54367
54367
54367
54367
54367
54367
54367
54367
54367
18311
17830
17368
16904
16496
16163
15830
15506
15218
14944
14683
14452
14232
13994
13804
13607
13405
13246
13070
12926
12774
12628
12489
12351
12236
12128
12011
11904
11793
11692
11690
11690
11690
11690

-------
                                6-A-7
Fleet Fuel Consumption

       Fleet  fuel  consumption  for all  calendar years  could now  be
determined by dividing the fleet VMT values by the  fleet  fuel economy
values.   (Since  for 1975  and  1990 through  2000  these were  based  on
the  estimates  of  fuel consumption  from the  13th  periodical  report,
with the  exception  of  buses,  the fuel consumption  for  these  years  is
identical  to  that  from  the  13th  report.)   These fuel  consumption
values are shown in Table 6-A-5.

-------
              6-A-8
           Table 6-A-5
Fleet Fuel Consumption (gal/year)
Year
1975
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
LDDV
2.00E+07
6.70E+08
6.10E+08
5.60E+08
5.19E+08
4.90E+08
4.76E+08
4.74E+08
4.76E+08
4.33E+08
4.93E+08
5.07E+08
5.17E+08
5.28E+08
5.40E+08
5.53E+08
5.67E+08
5.82E+08
5.97E+08
6.14E+08
6.31Et-08
6.48E+08
6.67E+08
6.86E+08
7.06E+08
7.27E+08
7.49E+08
7.71E+08
7.94E+08
8.18E+08
8.42E+08
8.67E+08
8.93E+08
9.20E>08
9.48E>08
LDDT
O.OOE+00
3.81E>08
3.95E4-08
4.09E+03
4.22Ei-08
4.35E+08
4.44E*08
4.53E+08
4.64E+08
4.78E+08
4.90E-I-08
5.02E*08
5.13E+08
5.25E+08
5.38E+08
5.52E+08
5.66E+08
5.81E4-08
5.97E*08
6.14E+08
6.31E+08
6.49E*08
6.67E>08
6.87E+08
7.06E+03
7.27E+08
7.48E+08
7.71E*08
7.93E>08
8.17E+08
3.42E+08
8.67E+08
8.93E+08
9.19E>08
9.47E+08
LHDDT
O.OOE+00
1.04E>09
l.lSE-i-09
1.25E*09
1.34E+09
1.43E>09
1.51E>09
1.57E+09
1.64E-I-09
1.70E+09
1.76E+09
1.81E*09
1.85E4-09
1.88E1-09
1.92E+09
1.96E+09
2.01E-I.09
2.06E+09
2.11E+09
2.16E>09
2.22Et-09
2.28E>09
2.35E+09
2.41E+09
2.48E+09
2.56E>09
2.63E+09
2.71E+09
2.79E>09
2.87E+09
2.96E*09
3.04E+09
3.14E+09
3.23E+09
3.33E+09
MHDDT
1.20E+OE
5.62E+09
5.66E+09
5.73E+09
5.81E+09
5.91E+09
6.03E+09
6.17E+09
6.34E+09
6.53EH-09
6.74E-^09
6.98E+09
7.14E+09
7.32E+09
7.51E+09
7.71E+09
7.93E+09
8.15E-t-09
8.38E+09
3.62E+09
8.87E+09
9.13E*09
9.40E*09
9.68E*09
9.97E+09
1.03E*10
1.06E*10
1.09E>10
1.12E+10
1.15E>10
1.19E+10
1.22E+10
1.26E+10
1.30E*10
1.34E+10
HHDDT
1.09E-t-10
l.HE^-10
1.12E+10
1.12E+10
1.12E*10
1.13E*10
1.14E+10
1.15E+10
1.16E4-10
1.17E+10
1.19E+10
1.21E>10
1.23E+10
1.26E>10
1.29E+10
1.32E+10
1.36E+10
1.39E+10
1.43E-t-10
1.47E>10
1.51E>10
1.56E+10
1.60E+10
1.65E+10
1.70E+10
1.75E+10
1.80E4-10
1.85E+10
1.91E+10
.1.97E+10
'2.03E+10
2.09E>10
2.15E+10
2.21E+10
2.28E+10
BUSES
6.09E+08
8.94E+08
9.20E«.Q3
9.49E*08
9.66E+08
9.91E-f08
1.02E+09
1.05E+09
1.07E+09
1.10E*09
1.13E*09
1.16E+09
1.18E+09
1.21E+09
1.24E>09
1.27E+09
1.30E+09
1.33E+09
1.36E+09
1.39E+09
1.43E+09
1.46E+09
1.50E+Q9
1.53E+09
1.57E+09
1.61E+09
1.65E+09
1.69E+09
1.73E+09
1.77E+09
1.31E+09
1.87E+09
1.92E+09
1.98E+09
2.04E+09
Total
1.16E+10
1.97E+10
1.99E+10
2.01E+10
2.03E*10
2.05E+10
2.08E+10
2.12E+10
2.16E+10
2.20E+10
2.25E+10
2.31E+10
2.35E*10
2.41E+10
2.46E+10
2.53E+10
2.59E*LO
2.66E>10
2.74E*10
2.81E+10
2.89E>10
2.97E+10
3.06E*10
3.15E+10
3.24E*10
3.34E+10
3.43E+10
3.54E+-10
3.64E^10
3.75E+10
3.86E+10
3.98E*10
4.09E>10
4.22E+10
4.34E>10

-------
                  6-A-9


               Table 6-A-6

       Urban  Travel  Fractions  [5]


Vehicle Class       Gasoline       Diesel
 Class 25           0.52           0.64
 Class 3            0.44           0.45
 Class 4            0.44           0.45
 Class 5            0.44           0.45
 Class 6            0.31           0.53
 Class 7            0.46           0.51
 Class SA           0.25           0.42
 Class 8B           0.08           0.26
 Transit Buses      1.00           1.00
 Commercial Buses   0.26           0.26
 School Buses       0.34           0.34

-------
            Table 6-B
Nationwide Particulate Emissions



Particulate Calendar Year Base
DSU





ISU





SOF





Carbon





Total Direct





Total PM





1992
1995
2000
2005
2010
2015
1992
1995
2000
2005
2010
2015
1992
1995
2000
2005
2010
2015
1992
1995
2000
2005
2010
2015
1992
1995
2000
2005
2010
2015
1992
1995
2000
2005
2010
2015
48970
51526
57896
66059
75706
87039
196310
208979
237127
271080
310931
357526
36993
70434
58565
59243
63853
70977
214683
180229
151726
154138
165984
183273
350646
302189
268187
279440
305543
341289
546956
511168
505314
550520
616474
698815
(tons)
Sulfur
100 ^
27768
29878
34620
39642
45284
51772
111382
119627
135836-
154700
176459
201531
87348
73172
60833
61638
66642
74104
216795
189513
168510
175355
191197
212610
331911
292563
263963
276635
303123
338486
443293
412190
399799
431335
479582
540017
Control
0 %
9990
10687
12791
14817
17050
19649
39276
41811
47443
54236
62208
71528
87348
73172
60833
61638
66642
74104
216795
189513
168510
175355
191197
212610
314133
273372
242134
251810
274889
306363
353409
315183
289577
306046
337097
377891
                                         Aromatics Control
                                         100 %        o %
                                           27771
                                           29901
                                           34699
                                           39745
                                           45406
                                           51915

                                          111382
                                          119627
                                          135836
                                          154700
                                          176459
                                          201531

                                           78923
                                           67522
                                           58281
                                           59882
                                           65014
                                           72294

                                          209930
                                          185255
                                          167782
                                          175758
                                          192057
                                          213707

                                          316624
                                          282678
                                          260762
                                          275385
                                          302477
                                          337916

                                          428006
                                          402305
                                          396598
                                          430085
                                          478936
                                          539447
  9993
 10710
 12870
 14920
 17172
 19792

 39276
 41811
 47443
 54236
 62208
 71528

 78.923
 67522
' 58281
 59882
 65014
 72294

209930
185255
167782
175758
192057
213707

298846
263487
238933
250560
274243
305793

338122
305298
286376
304796
336451
377321

-------
                                      6-B-2
                                     Table  6-B

                        Nationwide  SO;,  HC  and  CO Emissions
                                     "(tons)
Pollutant   Calendar Year
    SO2
1992
1995
2000
2005
2010
2015
 Base

 681633
 725621
 823359
 941250
1079620
1241409
                          Sulfur Control
                         100 \        01
 386745
 415370
 471651
 537154
 612704
 699760
136373
145178
164733
188318
215998
248362
                         Aromatics Control
                         100 %         0 %
386745
415370
471651
537154
612704
699760
136373
145178
164733
188318
215998
248362
    HC
    CO
1992
1995
2000
2005
2010
2015

1992
1995
2000
2005
2010
2015
 508337
 461757
 471735
 533077
 615087
 717888

1712905
1850598
2170306
2519754
2918358
3388653
 517759
 491613
 446353
 478250
 545413
 629909

1636825
1658341
1674549
1865366
2137011
2472360
              471338
              453705
              428271
              465077
              531951
              614953

              1565748
              1588853
              1605210
              1790367
              2052157
              2374583

-------
                           Chapter 7


              Cost Effectiveness of Fuel Controls

     This  chapter develops  the  methodology  for  and  provides
estimates  of  the cost  effectiveness of  diesel fuel  controls.
Cost effectiveness is  defined here  to  be the  net cost  per  ton
of  pollutant  removed and  is used  to  relatively  rank  control
programs.

     The chapter  is  divided into three parts.  The  first  part
gives  an  overview of  the pollutant  reductions, and cost  and
credit  elements  used in estimating net costs.   The  second  part
describes  the  specific  methodologies   used  to  estimate   net
costs.    Included  in  this  section  are  explanations  of   the
derivation of the total refinery cost,  the technology  credit,
the wear  credit  and fuel  consumption credits.  The  third  part
presents and discusses the cost effectiveness results.

     A.    Overview

     The purpose  of  this  section is  to present an  overview of
the  major  components  of   the   cost  effectiveness  equation.
Again,   the cost  effectiveness  is  expressed as  the net  cost  per
ton  of  pollutant  removed.    This   explanation will  start  by
discussing the  pollutants being  affected,  and then  the  costs
associated with fuel  controls.

     The emission analysis in  Chapter 6 showed that diesel  fuel
controls have an impact on  a number of  different  pollutants:
PM,  SC-2,   HC,  and   CO.   While  the   control  of  each of  these
pollutants is important, the purpose of the  cost  effectiveness
analysis  in  this  RIA  is  to  compare  diesel   fuel  controls  as
particulate   reduction   strategies   with  other   particulate
reduction  strategies.   Therefore,   in   this   analysis,   cost
effectiveness will be expressed  in  terms of dollars  per  ton of
particulate emission reductions  only.   The control of emissions
of  S02, HC,  and  CO,   while important,  is  only  of  secondary
importance in this  analysis, and will  be  presented  merely  as a
beneficial "by-product"  of fuel control.

     Turning first to the net cost of fuel control,  it is clear
that  the  only  actual  cost  is  the  refinery  cost  of  fuel
controls,  as was discussed  in  Chapter  2.   All  other  cost
components are  credits.   First,  there  is a  technology credit
for both  sulfur  and  aromatics control,  which  is the result of
engines  not  needing   as   much  or  as   costly  aftertreatment
technology.   Next,   there  is  a  fuel  consumption credit  which
also,  results   from fewer   trap  systems   in-use   with  fuel
controls.    (The  increase   in   volumetric   fuel    consumption
associated with  aromatics  control,  due to the  reduction in fuel
energy  density,  has  been  accounted   for  in  Bonner  and Moore's

-------
                               7-2


refinery  modeling,  and  is  included  in  the  refinery  costs
presented  in Chapter  2.)   Finally,  there  is  an engine  wear
credit  for  sulfur  control  (but  not  for  aromatics  control),
which  was discussed  in  Chapter  5.   All  of  these  individual
costs and  credits  are  combined to obtain  the  net cost  of  each
fuel control scenario.

     Another   important   aspect  of   the  cost   effectiveness
analysis  is  the  method  and  time  period  over  which  costs,
credits,  and  emission  reductions  are  estimated.   Two  basic
approaches have  been  used to  evaluate  motor  vehicle  controls,
one  focusing on  a new vehicle  over  its life  and the  other
focusing on  one  or more calendar years.   In  the  vehicle  based
approach, both costs and  emission reductions  are estimated over
the  life  of  the  vehicle.   In  the  calendar year  or  annual
approach,  net  costs and  emission reductions  are  estimated for
each  individual  calendar  year.  The  first type of  approach is
most  useful  when  the  controls primarily  affect  the technology
of.  specific  model year  vehicles.   However,  commercial  fuel
controls  affect  both  existing  and  new  vehicles  to  varying
degrees, so the calendar year approach is more appropriate.

     The annual cost effectiveness of a  control  strategy (i.e.,
fuel controls) can change dramatically  with calendar year.  For
example,  aromatics control  may have  a  significant impact .on
emissions  of the  in-use  fleet now,  but  as  more vehicles are
designed  to  meet  lower  emissions  standards  in  the  future,  it
has  correspondingly less  effect  and  the cost  effectiveness (in
absolute terms)  of aromatics control increases.   Therefore,  it
is  usually useful  to  estimate cost  effectiveness  for  several
different  years   so   that   a  single   more   appropriate  cost
effectiveness  value  may  be  estimated  from  these  individual
annual values.

     The  approach  used in a number of previous  EPA rulemakings
has  been  to estimate  the  average cost effectiveness  over 33
years.   This  approach includes  the  initial,  start-up years,
during   which   cost    effectiveness   values   can   be   either
particularly  low  or high,   as  well  as  including  a number  of
years where the controls have essentially reached steady state.

      In   the  33-year   cost  effectiveness  approach,   the  net
present  values  of  both  costs  and emission reductions  over the
33-year  period are determined  and  then the ratio taken.   This
technique,   which  weights   start-up   (near-term)   costs  and
emission  reductions slightly higher  than long-term  costs and
emission   reductions,   will  be  used   along  with   the  cost
effectiveness  values  of  several  selected calendar years  to
compare the various fuel  scenarios.

-------
                              7-3


     In this  analysis,  it was assumed that  fuel  controls  could
be implemented  as  early  as  the 1992  calendar year.   However,
implementation of  controls  in a year  other than  1992  does  not
affect the results of each  calendar year's  cost  effectiveness,
since each year's  value  is  independent  of the other years.   For
the 33-year discounted analysis  a  start  year and end year  must
be selected.  Since  1994 is  the latest  that  fuel  controls  can
be implemented  in  time  for  the  0.1 g/BHP-hr  PM standard,  the
33-year  discounted  analysis  will  focus  on  the  years   1994
through 2026.

     B.    Methodology

     This  section  discusses  the  specific   techniques  used  to
estimate  annual  costs and  credits.  Methods  used  to  estimate
total  refinery   costs  are  discussed  first,  followed  by  the
engine technology  credit, the fuel consumption credit (which is
associated with  the change engine  technology),  and  finally,  the
engine wear credit.

     1.    Refinery Costs

     The  cost   per   gallon  of   both   sulfur  and  subsequent
aromatics control  for  both the  minimum  and maximum distillate
fuel  segregation scenarios were presented  in Table 2-11,  and
are repeated  in the first  row  of  Table 7-1.   Annual  refinery
costs  are  estimated  by multiplying these costs by  annual  fuel
consumption.  With maximum segregation,  only  on-highway diesel
fuel   is   treated,   while   with   minimum  segregation,   both
on-highway and off-highway  (mobile and stationary)  diesel  fuel,
as  well  as  some  No.2  fuel  oil  is  treated,  resulting  in
significantly higher total refining costs.

     As shown in Table 7-1,  refining  costs for  sulfur control
in the maximum  segregation  case  range  from $360 million in 1992
to $830  million in  2025.   Subsequent  aromatics control  costs
range  from $470  million  in 1992 to  $1,080  million  in  2025.
Costs  in  the minimum  segregation  scenario  are  roughly double
those  in the  maximum  segregation  scenario.   As  indicated in
Chapter 2,  the  costs with  maximum  segregation do  not include
any additional distribution cost which may occur.

     2.    Engine Technology Credits

     The  engine  technology credit  is the  reduction in the cost
of  emission  control  technology  needed   to meet  the  emission
standards (1991  or 1994)  resulting from fuel control.  Vehicles
experiencing  these credits are  LDDT2s,  LHDDVs,  MHDDVs,  HHDDVs
and buses.   Vehicle model  year  groups  experiencing the credit
are the  1991-93 and  1994+  vehicles, the years coinciding with
upcoming  particulate  emission  standards.    The  total  credit in
any particular  calendar  year is  the sum of  the credit  for each

-------
                            7-4

                         Table 7-1


                   Annual Refinery Costs
             Maximum Segregation    Minimum Segregation
Calendar
  Year
 all
1992
1995
2000
2005
2010
2015
2020
2025
1992
1995
2000
2005
2010
2015
2020
2025
 Sulfur
Subsequent
Aromatics
Sulfur
Subsequent
Aromatics
                    Refinery Cost  ($/gal)
 0.0181    0.0235       0.0225     0.0209

Volume of Fuel Treated (1010 gal/yr)
2.01
2.08
2.31
2.59
2.97
3.43
3.98
2.01
2.08
2.31
2.59
2.97
3.43
3.98
3.35
4.06
4.56
5,15
5.88
6.74
7.75
3.85
4.06
4.56
5.15
5.88
6.74
7.75
 4.60
 4.60
8.89
 8.89
             Annual Refinery Cost ($million/yr)
 360
 380
 420
 470
 540
 620
 720
 830
  470
  490
  540
  610
  700
  810
  940
 1080
 870
 910
1030
1160
1320
1520
1740
2000
  800
  850
  950
 1080
 1230
 1410
 1620
 1860

-------
                              7-5


vehicle   class,   which   in  turn   is  the   product  of   the
class-specific credit times the  number of vehicles  sold  in  that
class that year.

     The  sales  data,  average   credit per  vehicle  and  total
technology  credit  by  vehicle   class  (and  for  all  vehicles
combined)  for several   selected  calendar  years  are shown  in
Table 7-2.  The sales data used  for this analysis  come  from the
MFCM, as  described  in Chapter 6.   As described there during the
calendar  years  1990-2000,  the  annual compounded  growth  rates
for  each class  are  about 2.75  percent.   For  calendar  years
beyond 2000,  this  has been rounded  to three percent per  annum
and applied in a compound fashion.

     The  average  credits  per vehicle were taken  directly  from
Tables 4-3  and 4-5.   In  1992,  the  total technology  credit for
sulfur  control  is  estimated to be  nearly  $71 million.   The
largest  credit comes from  the  LHDDVs, which  have  the  highest
projected sales.   The credit  for aromatics control, however, is
only about  $3.2 million,  which  is derived entirely  from LDDTs.
The  absence of credits  for  the other vehicles is  due  to the
fact that they already  are projected to need  no aftertreatment
with sulfur control.

     Technology credits  for  sulfur  control  from  1995  to  2025
range from  $142  -  343  million.   The  reason  for the growth • in
the  credits  is  the  projected   growth in  vehicle  sales.   For
aromatics control  over  the same  period  the credit  ranges  from
$19  -  46 million,  which  is  smaller than  the sulfur  control
credit by about a factor of 10.

     3.    Fuel Consumption Credit

     As  discussed  in Chapter  3,  vehicles with trap  systems are
estimated  to  use  two percent  more  fuel  than those with  flow
through  catalysts  or those  without  any  aftertreatment.   The
fuel  consumption  credit  represents  the savings in  fuel   costs
derived  from  those  vehicles for  which the necessity of  a trap
system is eliminated  as a  result of fuel control.

     To  estimate  the credit in  a particular  calendar year, two
items are needed.   First,  an estimate of  the  fuel  that   would
have been consumed by vehicles  that would have had traps with
current  fuel,  but  that  do  not  with  the  control  fuel,  is
needed.   The  -second  item  needed is  an estimate of the cost of
diesel  fuel minus taxes.   The  fuel  consumption credit  is the
product  of  these two  numbers.

     •The first  item  can  be estimated for  each  model year in  a
calendar year by  multiplying  the change in percent of vehicles
equipped with traps  from the base  to the control  fuel by the
total fuel  projected to be consumed by that  model year, and by
the  trap  fuel  economy  penalty  (2  percent).   The  total   fuel

-------
Calendar Year   Veh Type
1992
1995
2000
2010
2025
LDDT
LHDV
MHDV
HHDV
Bus
Total
                                       7-6

                                    Table 7-2


                            Annual Technology Credits

Vehicle
Sales
(1000's)
99
171
101
104
38
513
109
183
114
113
42

125
210
134
129
46
644
168
282
180
173
62
865
262
440
281
270
96
1,216
Sulfur
Savings
Per
Vehicle(S)
61
151
143
168
156

57
268
318
324
306

57
268
313
324
306

57
268
318
324
306

57
268
318
324
306

Control
Annual
Credit
(Smillion)
6.04
25.77
14.97
17.52
6.40
70.70
6.21
48.95
36.28
36.56
13.58
141.58
7.13
56.17
42.65
41.74
14.88
162.57
9.57
75.49
57.32
56.09
20.00
218.47
14.92
120.0
39.29
87.39
31.76
342.76
Aromatics
Savings
Per
Vehicle(S)
32
0
0
0
0

32
27
40
39
49

32
27
40
39
49

32
27
40
39
49

32
27
40
39
49

Control
Annual
Credit
(Smillion)
3.16
0
0
0
JO 	
3.16
3.48
4.88
4.57
4.38
1.63
18.94
4.00
5.60-
5.37
4.99
1.78
21.74
5.37
7.53
7.22
6.71
2.39
29.22
8.37
11.73
11.24
10.46
3.73
45.53

-------
                              7-7
consumed by vehicles  of  a specific model  year is  the  quotient
of total miles driven  by the model year and the model year fuel
economy.  The total miles driven by a model year  is  the product
of that  year's  travel fraction  and  total  VMT  for  that vehicle
class.

     The  total  credit  can  then  be  obtained  by  summing  the
credit gallons for  all model years in a calendar year,  and for
all vehicle  classes,  and multiplying this  sum by  the cost  of
diesel  fuel  minus  taxes.    This cost  was  determined  to  be
73.6^/gal for $20 per barrel crude oil, based  on  DOE distillate
fuel price information.[1]

     The  changes  in  the percentages  of  trucks equipped  with
traps,  derived from data in  Tables  4-2 and  4-4,  are  shown  in
Table 7-3.   Travel  fractions for  each model  year are estimated
as  the  product  of  the  registration,  VMT,  and  diesel  sales
fractions distributions,  which  were discussed  in Chapter  6 and
presented  in Tables  6-3 through 6-5.   VMT  and fuel  economy
estimates for each vehicle class were taken from Appendix 6-A.

     Total  fuel  consumption  by  model  year  group  and vehicle
class and  the fuel  consumption  credit for  each vehicle  class
for  selected calendar  years  is  shown  in  Table  7-4.   The fuel
consumption  credit  for sulfur control  ranges from  $12 million
in 1992 to  $400  million in 2025.  MHDDVs and HHDDVs account for
more than  50 percent  of the credit.   The  fuel credit  starts
lower and  ends higher than  the technology  credit,  because it
only affects  post  - 1991 vehicles and  grows only  as  the  fleet
turns over.   When  a truck  is  sold,   its  technology  credit  is
counted only  in  the year it is  sold.  However,  the fuel  credit
for  a vehicle accrues over every  calendar year it  is operated.
In 1992 or  1995,  there  are  few vehicles that  are  accumulating
fuel credits.  As  time  goes  on, however,  more of  the fleet is
accumulating  these  credits  until  in  the  year  2015  all of the
fleet is  accumulating fuel  consumption credits.  The growth in
the credit  between  2015 and 2025  is due  to the  growth in fuel
consumption   over   that   time  period.    The  fuel  credit  for
subsequent aromatics control ranges from almost zero in 1992 to
$41 million in 2025,  and again is about  one-tenth the size of
the fuel credit due to sulfur control.

     4.     Wear Credit

     The wear credit  is  the result  of  lower  engine  wear with
low sulfur  fuel.   The potential effects of  lower engine wear on
oil change  interval,  engine and vehicle life  were  discussed in
Chapter  5.   Generally,  the  conclusions  of Chapter  5  were that
reduced  wear could  result   in  lower  engine  oil cost  and less
frequent  oil change  intervals,  or   longer   engine  and vehicle
life,   or   longer   engine  life   with   fewer  total   rebuilds.
Benefits were estimated for each  of these possible  scenarios in
terms of tf/mile  and  were  presented  in  Tables  5-4,  5-14  and

-------
                                7-8

                             Table 7-3
                  Changes  in  Percentage of  Trucks
                  	Equipped  With Traps	
Veh
LDDT

LHDV

MHDV

HHDV

Buses

91-93
94+
91-93
94+
91-93
94+
91-93
94+
91-93
94+
25
25
27
100
20
98
8
88
100
100
.0
.0
.5

.6
.6
.9
.0
.0
.0
21.
21.
0
29.
0
36.
0
36.
67.
34.
7
6

5

0

5
8
0
	Percent Equipped	
Base    Sulfur    Aromatics
                                       19.9
                                       19.8
                                        0
                                       24.7
                                        0
                                       27.0
                                        0
                                       32.0
                                       61.4
                                       28.6
	Difference	
Sulfur   Aromatics
3.3
3.4
27.5
70.5
20.6
62.6
8.9
51.5
32.2
66.0
1.8
1.8
4.8
4.8
0
9.0
0
4.5
6.4
5.4

-------
                              7-9

                           Table 7-4


                    Fuel Consumption Credits
Calendar
  Year

  1992
  1995
  2000
  2010
  2025
Fuel Consumption
   106qal/yr)
Fuel Consumption
Credit ($/million/yr
Veh Type
LDDT
LHDV
MHDV
HHDV
Bus
Total
LDDT
LHDV
MHDV
HHDV
Bus
Total
LDDT
LHDV
MHDV
HHDV
Bus
Total
LDDT
LHDV
MHDV
HHDV
Bus
Total
LDDT
LHDV
MHDV
HHDV
BUS
Total
91-93*
57
392
1584
2779
321
5133
100
575
2154
3840
430
7099
79
234
868
1518
351
3050
40
83
327
588
127
1165
0
0
0
0
0
0
94+*
0
0
0
0
0
0
27
425
1602
2814
308
5176
256
1358
5139
8744
1032
16529
609
2201
8808
15291
2331
29240
1005
3530
14190
24669
3823
47217
Sulfur
0.03
1.59
4.80
3.64
1.52
11.58
0.62
6.74
21.29
26.36
5.03
60.04
0.17
15.04
50.00
68.28
11.69
145.18
0.32
23.18
82.16
120.00
23.25
248.91
0.50
36.63
130.0
190.0
37.47
394.27
Aromatics
0.02
0
0
0
0.30
0.32
0.03
0.30
2.12
1.86
0.65
4.96
0.89
0.96
6.81
5.79
1.15
14.80
1.72
1.55
11.67
10.13
1.97
24.49
0.27
2.49
18.80
16.34
3.04
40.94
     Vehicle model years

-------
                              7-10
5-16.  All that  is  needed  to estimate an annual wear credit  is
to multiply  these benefit  estimates  by  total VMT per  calendar
year for  each vehicle class.   As  discussed  in  Chapter 5,  the
engine wear  benefit  in  which there  is  an extension in  engine
and vehicle life, and the engine wear benefit where  there  is  an
extension  in  engine life and  a reduction in rebuilds  are both
applicable to all vehicles,  regardless of age,  once  low  sulfur
fuel is implemented.  Therefore, total VMT by vehicle class and
calendar year is used in estimating the wear  benefits  for  these
two  engine wear  benefits   analyses.   The  oil  change  interval
benefit, however,  is applicable only to  1991 and later  HDDVs.
Therefore, the  VMT  of 1991  and later model  years is needed  to
estimate this benefit.

     These two  categories of VMTs  are shown  in Table 7-5.   The
first set  of  VMTs are the  total VMTs for  each vehicle  class  in
each calendar year.   These come directly  from Table  6-A-2.  The
second  set of  VMTs  were  obtained by  multiplying  these  total
VMTs by  the  1991 and  later  travel  fractions   (as  defined  in
Chapter 6) for each vehicle class in each calendar year.

     Total wear  benefits for  each  of these  three scenarios  by
calendar year  and vehicle class  are  shown  in  Table 7-6.   The
column  headed  "Engine  and  Vehicle  Life"  shows the  estimated
credits  for  the wear   scenario  in  which  there  is  both  an
extension  in engine and vehicle  life  (benefits  are  shown  in
Chapter 5, Table  5-14).   The  total  credit ranges from $2.1- to
$5.7 billion.   This credit  is three  and  six times  the refining
cost for sulfur  controls with minimum and  maximum  segregation,
respectively.   The  column  headed  "Rebuild  Interval  and  Number
of  Rebuilds"   shows the   credit   due   to   extending   rebuild
intervals  and   reducing  the  number   of   rebuilds,   but  not
extending  vehicle  life.   The  total  credit ranges  from  $411
million to $1.1  billion,  slightly  larger than the refinery cost
of sulfur  control for maximum segregation,   and  about  one-half
the  refinery cost  of sulfur  control for  minimum  segregation.
The last column headed "Oil  Cost and  Change Interval"  shows the
total  credit from  an increase  in  oil  change  intervals  and a
slight  decrease  in  oil  cost  per  guart  (lower  total  additive
content due  to  low sulfur  fuel).   The  credit ranges  from $55
million to $611 million.  The  credit  is  only applicable to 1991
and  later vehicles,  which  explains the rapid  growth  in the
credit.   In  2025, this  credit is  between 31 percent  (minimum
segregation)  and  74  percent  (maximum  segregation)  of  the
refinery cost..  In all cases,  the credits for MHDDVs and  HHDDVs
together are at least 65 percent of the total credit.

     5.    Cost Effectiveness  Results

     'A  summary  of all the  inputs to  cost effectiveness and the
resulting  cost  effectiveness  values  for  sulfur and  aromatics
control  is  shown  for  several  calendar  years  in  Table  7-7.

-------
 Type

 All
1991 &
  later
Calendar
  Year

  1992
  1995
  2000
  2010
  2025

  1992
  1995
  2000
  2010
  2025
7-11
Table 7-5
Future VMT Projections
U-010 miles per year)
LDDV
1.64
1.47
1.68
2.26
3.50
0.12
0.34
0.93
2.26
3.50
LDDT
0.98
1.09
1.28
1.72
2.67
0.15
0.43
0.91
1.72
2.67
LHDDV
1.79
2.23
2.85
3.83
5.96
0.55
1.69
2.71
3.83
5.96
MHDDV
4.28
4.64
5.63
7.57
11.70
1.11
3.22
5.20
7.57
11.70
HHDDV
6.85
7.29
8.27
11.10
17.20
1.68
4.92
7.59
11.10
17.20
Buses

0.86
0.98
1.21
1.62
2.51
0.15
0.52
1.02
1.62
2.51

-------
                              7-12

                           Table  7-6
                      Engine Wear Credits
                          ($million/yr)
Calendar
  Year

  1992
  1995
  2000
  2010
  2025
Type
LDDV
LDDT
LHDDV
MHDDV
HHDDV
Bus
Total

LDDV
LDDT
LHDDV
MHDDV
HHDDV
Bus
Total

LDDV
LDDT
LHDDV
MHDDV
HHDDV
BUS
Total

LDDV
LDDT
LHDDV
MHDDV
HHDDV
BUS
Total

LDDV
LDDT
LHDDV
MHDDV
HHDDV
BUS
Total
Engine &
Vehicle
  Life
Rebuild
& No.of
Rebuilds
Oil Cost
& Change
Internal
                                                       264.0
                                11.3
                                 8.6
                                19,
                               144,
                               166,
                                                            1
                                                            0
                                                            0
                                                        45.2
                                                       394.2
                           5,680

-------
7-13
Table 7-7
Annual Cost Effectiveness of Sulfur
and Aromatics Control
(All Costs and Credits in Smillion/yr)
Calendar
Year Item
1992



1995



2000



2010



2025



PM Reductions (tons)
Refinery Cost
Technology Credit
Fuel Credit
Wear Credit*
Net Costs
Cost Effectiveness($/ton)
PM Reductions (tons)
Refinery Cost
Technology Credit
Fuel Credit
Wear Credit*
Net Costs
Cost Ef fectiveness($/ton)
PM Reductions (tons)
Refinery Cost
Technology Credit
Fuel Credit
Wear Credit*
Net Costs
Cost Effectiveness($/ton)
PM Reductions (tons)
Refinery Cost
Tech Credit
Fuelnology Credit
Wear Credit*
Net Costs
Cost Effectiveness ($/ton)
PM Reductions (tons)
Refinery Cost
Technology Credit
Fuel Credit
Wear Credit*
Net Costs
Cost Effectiveness($/ton)
Sulfur Control
Max Seqr
36,366
364
71
12
55
357
6,152
35,353
376
142
60
162
13
381
38,140
418
163
145
264
-154
-4,030
49,458
538
218
246
395
-321
-6,495
76,570
333
340
392
611
-511
-6,674
Min Seqr
91,696
866
71
12
55
729
7,947
94,526
914
142
60
162
551
• 5,830
105,373
1026
163
145
264
455
4,314
136,369
1,323
218
246
395
464
3,402
204,813
2,000
340
392
611
657
3,206
                                                      Aromatics Control
                                                     Max Seqr

                                                       5,601

                                                         472
                                                           3
                                                           0
                                                     	0
                                                         469
                                                      83,710

                                                       3,670
                                                       1,311
                                                         417

                                                         700
                                                          29
                                                          25
                                                           0
                                                         643
                                                      1.54xl06

                                                         561
Min Seqr

  5,601

    804
      3
      0
	g
    801
142,946

  3,670
489
19
5
0
465
126,670
849
19
5
0
825
224,846
  1,311
543
22
15
0
506
386,194
953
22
15
0
917
699,332
    417

   1,230
      29
      25
	g
   1,174
2.82xl06

    561
1,080
46
41
0
995
1.77xl06
1,860
46
41
0
1,772
3.16xl06
Extension of oil change interval only.

-------
                              7-14


Costs, emission  reductions,  and  cost  effectiveness values  for
sulfur and  aromatics  control,  for both the minimum and maximum
(NPRA) segregation scenarios, are shown therein.

     As can be seen in the Table,  the  cost-effectiveness values
shown  for  sulfur  control  in the maximum segregation  scenario
decrease with time, from  $6,152  per ton  to  -$6,674 per  ton  in
2025.   This  is  due  to   the  increase   in  the  aftertreatment
technology, fuel consumption, and wear  credits in  later years,
as  1991  and  later  model  year   engines  replace  the  current
fleet.   Beginning  in the  year  2000,  net   costs  for  sulfur
control  in  the maximum  segregation  scenario actually become
negative,  indicating  that societal  savings   will   outweigh  the
cost  to  refiners   of  producing  low  sulfur  fuel.  Under  the
minimum  segregation scenario,  cost effectiveness values  also
decrease with  time,  from $7,947  per ton in  1992  to $3,206 per
ton  in  2025.   In  the  minimum segregation  case,   fuel  economy,
technology,  and engine  wear credits  do not quite offset  the
greater costs to the refining sector associated with treating a
larger volume of fuel.

     The estimates  in Table 7-7  show  the cost effectiveness of
sulfur control assuming reduced engine wear will  result only in
an increase  in  oil change interval.   Cost effectiveness results
under the  other  two wear  benefits scenarios   are shown  in Table
7-8.   The   long  term   (2025)   cost   effectiveness  with  the
"increase in engine life and reduction  in rebuilds engine wear"
scenario  ranges  from  -$12,948  to  $860  per  ton.    For  the
"increase  in engine and vehicle  life"  scenario, long  term cost
effectiveness  ranges  from  -$72,996  to  -$21,589/ton.   These
estimates indicate that the method of  translating  the  impact of
reduced  sulfur  level  on  engine  wear  into an economic benefit
can  have a  large  impact  on -he  cost  effectiveness  of s- ifur
control.

     The calculated cost  effectiveness  of aromatics control, as
shown  in Table  7-7,  continually increases over time.   As newer
technology  engines  replace  the  current fleet,   the   emission
reductions   resulting  from   aromatics   control   continue  to
decrease,  while  refining costs  continue to grow.   Aromatics
cost effectiveness  ranges  from $83,710 per ton in  1992 to more
than $3 million per ton in 2025.

     The emission  reductions used in  the above  analysis of the
cost  effectiveness  of  aromatics control  were  based  on  the
results  of VE-1 testing on  a  Cummin's  NTCC  engine as discussed
in  Chapter  4.   However,   the   analysis  in  Chapter  4  also
discussed  emission reductions on a  Caterpillar engine  that was
tested   in  a  joint  Mobil/Caterpillar   testing  program.   The
emission  reductions resulting from  aromatics control were much
higher   on  this   engine  than   on  the  former.    While  the
Caterpillar  engine  tested  is  slightly  older technology,  the
data  obtained  from it  may be  representative of  older engines

-------
                              7-15

                            Table  7-8

            Effect  of  Higher  Engine Wear  Credits on
              Cost Effectiveness of Sulfur Control
   Wear Benefit Type

Engine Life S Rebuilds
Engine and Vehicle Life
CYR

1992
1995
2000
2010
2025

1992
1995
2000
2010
2025
  Sulfur Control Cost
  Effectiveness,$/ton
Maximum          Minimum
 -3,723
 -7,772
 -10,873
 -12,770
 -12,948

 -49,768
 -60,150
 -68,786
 -72,758
 -72,996
  3,976
  2,781
  1,837
  1,127
    860

-14,536
-16,809
-19,125
-20,630
-21,589

-------
                              7-16
still in  the  fleet in  the  mid-1990's.   Thus,  as  a  sensitivity
case, the percentage emission  reductions  resulting  from  fuel
aromatics  control  on the Caterpillar  engine have been  applied
to  all   pre-1991   heavy-duty  engines.   New  fleet  particulate
emissions  were  developed  and  new  cost  effectiveness  values
generated.  The results  are shown   in  Table  7-9.    Estimated
particulate reductions  for  1992  are increased  by a  factor  of
three, while  long  term reductions (2025)  are the  same due  to
the  sensitivity  emission  benefits  being  applicable  to  only
pre-1991  engines.   The  1992 cost  effectiveness  of  aromatics
control  with  this  sensitivity case is about  one-third the cost
effectiveness  in the previous case,  but  still over $25,000  per
ton  of  PM.   Long  term cost  effectiveness  is identical  to  the
previous case, ranging from $1.7 to $3.2  million per  ton.

     The  33-year  cost  effectiveness  analysis  for  both  sulfur
and   aromatics   control   is   shown  in   Table  7-10.    Cost
effectiveness   results   for  sulfur control   are  shown  for  all
three  engine  wear  benefit  scenarios,   as  well  as  results
assuming that  no engine wear benefits exists.   The 33-year cost
effectiveness  of  sulfur  control  ranges   from $2,800  to  $6,800
per  ton   if  no   engine   wear   benefit   is  assumed,   and  is
significantly   less   when  wear  benefits   are  claimed.    For
aromatics  control  two  cases  are shown,   the  first  based  on
emission  reductions  from  the  Cummins  engine,  and the  second
case, a sensitivity case,  based on emission reductions from the
Mobil/Caterpillar study.  The 33-year  cost  effectiveness values
based on the Cummins engine range from $310,000  to $560,000 per
ton.  The  sensitivity case  values are less than one-half of the
first case, but still are in excess of $130,000 per ton.

     The  cost  effectiveness of  aromatics  control  ($310-560,000
per  ton)  is significantly higher than that of  particulate trap
technology as  estimated in  the  NOx/Particulate  RIA.[2]  In that
document,   the   cost-effectiveness-   of   the   1994   heavy-duty
particulate standard was  estimated there to be at most $18,300
per   ton   ($18,700    in    1986   dollars),   a   difference   of
approximately $300,000 per ton.   However, as  seen  in  Chapter 6,
in  addition  to  particulate  control,  the  aromatics  control
scenario  evaluated here also results in  significant  reductions
in  HC  and CO (10,348 and 28,794  annualized tons,  respectively,
over the  33-year period).   In  order to  render  fuel  aromatics
control equivalent in  cost-effectiveness  to the 1994 heavy-duty
particulate standard,   the  33-year  discounted annual  cost  of
aromatics  control  would have to  be  only  $33.5  million, rather
than the  $560 million  shown  in Table 7-10.   If  the  remaining
$526.5  million  were  apportioned  equally  to the  control  of HC
and  CO,  the resulting  cost effectiveness  of control  for these
two  pollutants  would  be  $25,400   and  $9,100  per  urban ton,
respectively.    These  values  are  significantly  greater  than
those  of  control  strategies  which have  been  implemented or
which are currently under  consideration  by  the  Agency.   Thus,
even  considering  the  added  benefit  of  HC  and CO control,
aromatics  control  is not  cost  effective  relative to  the  1994
particulate standard.

-------
                                 7-17
                              Table  7-9
              Effect of a. Higher Emissions Effect on the
               Cost Effectiveness  Of Aromatics Control
                Emission Redactions(tons)    Cost Effectiveness($/ton)
Calendar Year       PM           HC              Max       	Min
1992
1995
2000
2010
2025
17,046
9,020
2,930
417
561
25,314
16,547
9,895
7,312
10,852
27,505
51,539
172,799
1,542,498
1,772,763
46,970
91,484
312,909
2,815,306
3,157,770

-------
               7-18
            Table 7-10

33-Year CostEffectiveness Analysis
Engine Wear
Control Credit
Sulfur Oil Interval

Eng,Veh Life

Eng Life/ Reb

No Wear Credit

AL uiuau It, a — — —

rtCUtua 1 1 1. o -,—. ~— , — ,— — —
High Emissions
Sensitivity
Item
Emiss ions ( tons )
CostsClO6)
C/E( I/ton)
Emission(tons)
Costs(106$)
C/i($/ton)
Emissions (tons)
Costs(106$)
C/E($/ton)
Emissions (tons)
Costs(106$)
C/E( I/ton)
Emissions (tons)
CostsClO^I)
C/E($/ton)
Emissions(tons)
Cost
C/E($/ton)
Level of Segregatior
Maximum Minimum
42,751
-170
-3,906
42,751
-2,900
-68,148
42,751
-460
-10,867
42,751
120
2,826
1,790
560
310,751
4,128
560
135,789
116,588
500
4,304
116,588
-2,200
-19,253
116,588
200
1,752
116,588
790
6,773
1,790
1,010
560,378
4, 128
1,010
244,868

-------
                              7-19
                      References,  Chapter 7
     1.    "Petroleum  Marketing  Monthly,"  Energy
Administration, OQE/EIA-Q38CK 88/06)» June, 1988,
Information
     2.    "Regulatory  Impact   Analysis,   Oxides  of  Nitrogen
Pollutant Specific Study  and Summary and Analysis of Comments -
Control of Air  Pollution  from New Motor Vehicles and New Motor
Vehicle  Engines;   Gaseous  Emission Regulations  for  1987  and
Later  Model  Year  Light-Duty  Vehicles,  and  for  1988  and Later
Model   Year   Light-Duty   Trucks   and   Heavy-Duty   Engines;
Particulate Emission Regulations for 1988 and  Later Model Year
Heavy-Duty Diesel  Engines," EPA,  OAR,  QMS,  March 1985.   Docket
A-80-18.
                           EPA-420-R-90-103
                              June 1990

-------

-------