Report to Congress
                   on
          Railroad  Emissions  -
     A Study  Based  On  Existing Data
              Prepared by

 U. S. Environmental Protection Agency
      Office of Air and Radiation
        Office of Mobile Sources
  Emission Control Technology Division
Standards Development and Support Branch

-------
           Report  to  Congress
                   on
          Railroad Emissions -
     A Study Based On Existing Data
              Prepared by

 U. S. Environmental Protection Agency
      Office of Air and Radiation
        Office of Mobile Sources
  Emission Control Technology Division
Standards Development  and Support Branch

-------
                              -1-


                       Table of Contents

Section                                                      Page

     1.    Introduction  	    1

     2.    Summary of Conclusions  	    4

     3.    Recommendations Pertaining to Federal Action  ,  .    7

     4.    Background	    8

           4.1     Selection of Study Areas  	    8
           4.2     Locomotive Design Characteristics
                   Pertinent to the Study	    8
           4.3     Locomotive Duty Cycles	17
           4.3.1   Switch and Transfer Locomc-ive
                   Duty Cycle	   19
           4.3.2   Line-Haul Locomotive Duty Cycle 	   21
           4.3.3   Secondary Power Source Duty Cycle ....   24

     5.    Railroad Emission Estimates 	   27

           5.1     Locomotive Exhaust Emissions  	   27
           5.2     Secondary Power Source Exhaust Emissions.   36

     6.    Environmental Impact of Railroad Emissions  ...   44

           6.1     Non-Exhaust Emissions  	   44
           6.1.1   Refueling Losses	44
           6.2     Exhaust Emissions  	   45
           6.2.1   Localized Effects  .....  	   45
           6.2.1.1 Air Quality Monitoring  	   46
           6.2.1.2 Air Quality Modeling   	   48
           6.2.2   Within AQCRs	48

     7.    Potential Emission Reduction Techniques  	   52

           7.1     Duty Cycle Modifications	52
           7.1.1   Engine Shutdown When Not
                   in Active Service	52
           7.1.2   Limiting Use of Highest Power
                   Settings When  in Urban Areas	56
           7.1.3   Composite Effect of Duty
                   Cycle Modifications  	  .....   58
           7.2     Application of Emission
                   Control Technology  	   58

-------
                        -11-
      7.2.1   Modification of Injector Design .....   59
      7.2.2   Modification of Injection Timing
              (Timing Retard)	   59
      7.2.3   Exhaust Gas Recirculation	   60
      7.2.4   Reduced Scavenging (Increased
              Internal Exhaust Gas Recirculation)  ...   61
      7.2.5   Water Injection	62

8.     Cost and Cost-Effectiveness Estimates 	   64

      8.1     Duty Cycle Modifications	64
      8.1.1   Engine Shutdown When Not In
              Active Service	64
      8.1.1.1 Engine Startability 	   65
      8.1.1.2 Use of Antifreeze and Control
              of Fuel Waxing	71
      8.1.1.3 Cold Start Emissions	75
      8.1.1.4 Lubrication Changes 	   77
      8.1.1.5 Fuel Savings from Engine Shutdown ....   78
      8.1.1.6 Composite Costs for Engine
              Shutdown   .	78
      8.1.2   Restricted use of High Power Settings
              in Urban Areas	80
      8.2     Application of Emission Control
              Technology	81
      8.2.1   Modification of Injector Design 	   83
      8.2.2   Modification of Injection Timing  ....   83
      8.2.3   Exhaust Gas Recirculation	 .   83
      8.2.4   Reduced Scavenging (Increased Internal
              Exhaust Gas Recirculation)	86
      8.2.5   Water Injection	87
      8.3     Cost-Effectiveness	90

9.     Existing State and Local Regulations  	   99

      9.1     Survey of Existing Regulations  .....   99
      9.1.1   Survey Returns	99
      9.1.2   Types of Regulations	  101
      9.1.3   Typical Regulation   	  101
      9.1.4   Compilation of State and Local
              Standards	103
      9.1.5   General Results	105
      9.1.6   Subjective Questions on Enforcement  .  . .  107
      9.1.7   Subjective Questions on the
              Need for Federal Regulations	108
      9.2     Effects of Existing Regulations 	  109
      9.2.1   Health and Welfare	  109
      9.2.2   Operational and Technical Controls   .  . .  109
      9.2.3   Interstate Commerce	  Ill

References	113

-------
                             -111-

                     List of Illustrations
Figures                                                     Page
     1.    Philadelphia AQCR	   9
     2.    Chicago AQCR	.   .  10
     3.    Central Chicago	11
     4.    St. Louis AQCR	12
     5.    Kansas City AQCR	13
     6.    Los Angeles AQCR	14
     7.    Central Los Angeles	15

-------
                         -IV-
                    List of Tables

Table                                                   Page

  1.   Switch Locomotive Duty Cycles  	  20

  2.   Line-Haul Locomotive Duty Cycles 	  22

  3.   Line-Haul Locomotive Duty Cycle Applicable to
       Operation Within Air Quality Control Regions ...  23

  4.   Line-Haul Locomotive Duty Cycles Overall Average,
       Within AQCRs and in Rural Areas  	  25

  5.   Refrigerated Rail Car Duty Cycle and Engine
       Loading	26

  6.   Average Locomotive Hydrocarbon Emissions by
       Throttle Setting 	  29

  7.   Average Locomotive Carbon Monoxide
       Emissions by Throttle Setting	30

  8.   Average Locomotive Oxides of Nitrogen
       Emissions by Throttle Setting	31

  9.   In-use Locomotive Power Rating Groups,
       Number of Locomotives by Groups and Test
       Engines Representing In-Use Groups  	 .33

 10.   In-Use Weighted Average Line-Haul
       Locomotive Emissions by Throttle Position   ....  34

 11.   In-Use Weighted Average Switch and Transfer
       Locomotive Emissions by Throttle Position   ....  34

 12.   Line-Haul Locomotive Emissions per
       Locomotive per 12-Hour Day in an AQCR	35

 13.   Switch and Transfer Locomotive Emissions
       per Locomotive,24-Hours Per day	35

 14.   Number of Locomotives in Each AQCR	37

 15.   Line Haul and Switch and Transfer Locomotive
       Emissions in Five ACQRs	38

 16.   Refrigerated Rail Cars  in Each AQCR	40

 17.   Refrigerated Rail Car Emissions by
       Power Setting	42

-------
                         -v-
18.    Refrigerated Rail  Car Emissions	42

19.    Railroad Emissions in Five AQCRs	43

20.    Summary of Air Contaminant Levels  in the Cab
      of Long-Hood, Forward Switchyard Locomotives ...  47

21.    Railroad Emissions,  Total Anthropogenic Emissions
      and Percentage Contributions by Railroads
      for Five AQCRs	  49

22.    Percentage of Total  Railroad Emissions in AQCRs
      Contributed by Idle  Mode and Notch 8 Operations.  .  51

23.    Locomotive Fuel Consumption - Average Values
      in AQCRs	57

24.    Summary of Costs:   Engine Starting Aides 	  72

25.    Summary of Costs:   Use of Antifreeze and
      Fuel Waxing Control	76

26.    Summary of Costs:   Reducing Cold  Start
      Emissions and Lubrication Changes	79

27.    Summary of Costs:   Duty Cycle Modification
      Relative to Historical Duty Cycles 	  82

28.    Summary of Costs:   Application of
      Emission Control Technology	91

29.    Percent Change in Lifetime Emissions and Fuel
      Consumption By Control Procedure  Relative
      to Historical Duty Cycles	92

30.    Lifetime Change in Mass of Emissions in AQCRs
      and Fuel Consumed for An Average  Locomotive
      Relative to Historical Duty Cycles 	  93

31.    Lifetime Costs for Emission Control
      Procedures per Locomotive Relative to
      Historical Duty Cycles  	  94

32.    Cost Effectiveness of Control Strategies
      Based on Historical Duty Cycles	95

33.    Cost Effectiveness for Controlling
      Non-Locomotive Sources  	  97

34.    Survey Returns  .	100

35.    Categorization of State and Local Regulations   .  104

-------
                       1.0   INTRODUCTION

     Section 404 of the Clean Air Act (CAA) Amendments  of  1977*
required the Environmental Protection Agency  (EPA)  to conduct a
study of emissions of  air  pollutants from railroad  locomotives
and secondary  rolling  stock power sources with respect  to:   1)
their environmental impact, 2) methods  for  control,  and 3)  the
status  and  effects of  state regulations.   The results of  the
study, together with recommendations on appropriate  legislative
action,  were to be reported to the Congress.

     This report presents  the results of the study performed by
the EPA for  five  selected Air Quality  Control Regions  and  the
recommendations based upon the findings of  the study.  Sections
6, 7, and  9 of this  report  address  the three  areas of  study
which were required by the legislation.  Background information
and computational procedures  employed  are provided  in  sections
4  and 5.   Section  8  provides the estimates  of costs  and cost
effectiveness associated with the methods which were considered
for reducing railroad emissions.

     Information on railroad  emissions,  state regulations, etc.
was gathered by literature searches,  questionnaires, interviews
and  from  the  Association  of  American  Railroads  (AAR)  and
locomotive manufacturers.  This  approach was  selected  as  being
the  most  timely  as  well  as   being   the  most  cost-effective
     "RAILROAD EMISSION STUDY

     Sec.404   (a)    The  Administrator  of   the  Environmental
     Protection Agency  shall conduct  a  study and investigation
     of emissions of  air pollutants from  railroad locomotives,
     locomotive   engines,   and   secondary  power   sources   on
     railroad rolling stock, in order to determine—

     (1)   The  extent  to   which   such   emissions  affect  air
     quality control regions throughout the United States,

     (2)   The technological  feasibility and the  current state
     of technology for controlling such emissions, and

     (3)   The status and  effect  of current  and proposed state
     and local regulations affecting such emissions.

     (b)   Within one-hundred  and eighty  days  after commencing
     such  study  and  investigation,  the  Administrator  shall
     submit  a  report of such  study and investigation, together
     with  recommendations  for  appropriate legislation,  to  the
     Senate  Committee on  Environment  and Public  Works  and the
     House Committee on Interstate and Foreign Commerce."

-------
                              -2-
approach  in  terms of  both personnel  and financial  resources.
The literature searches,  and AAR and manufacturer  supplied data
were utilized to  estimate  locomotive and  secondary power source
emissions levels,  fleet  size and  distribution,  usage  patterns
and  duty cycles,  and  fuel consumption.'  Questionnaires  were
sent  to  state  and  local   air  pollution  control  agencies  to
survey their regulations,  enforcement policies and problems and
the perceived need for Federal regulations.   Personnel of  the
Federal Railroad Administration and industry were interviewed.

     Information  gathered  from the  sources  was  compiled  to
develop   the   reported  emissions   inventories,  air   quality
impacts,  and   effects  of  state  and   local   regulations  on
interstate commerce.

     Copies of the original  draft  of  the  report were  provided,
for   review   and  comment,  to  the  Association   of   American
Railroads  and to  the  locomotive  manufacturers  (Electromotive
Division  of  General  Motors  (EMD)  and  General  Electric  (GE)).
Comments  and  recommendations provided  by  the   reviewers  have,
wherever  possible, been  incorporated into the report.   Briefly,
components of the report  which were  impacted by  the  reviewers
recommendations and components where recommendations  for change
were made but not incorporated are as follows:

     0     Locomotive   emissions    data   base   sample   size.
           Initially,  the  report  had  relied on  a very  small
           sample  of  data  collected  in  1972.   Data  collected
           from new locomotives and  supplied by  EMD in 1978 had
           not been used because  of  lack  of  information  on the
           effects  of  locomotive  aging  on emissions.   As  a
           result  of  the review,   AAR  furnished data  collected
           in 1984 from in-use locomotives.   Provision of this
           data  substantially  increased  the size  of   the  data
           base.   It  also  allowed another  increase  in the size
           of the data base by incorporation of the EMD data.

     0     Calculational  procedures  for  estimating  locomotive
           emissions.  Originally,  three methods  were employed
           to calculate  locomotive  emissions.   EPA  recognized
           that  two  of  the methods were very  weak  because of
           assumptions which had  to  be  employed to  utilize the
           available  data.  Because of  concerns  raised  by the
           reviewers  with  respect  to   the  veracity  of  the
           assumptions,  two  calculational  methodologies  were
           removed from the report.

      0     Number  of  locomotives  in AQCRs.   EMD  expressed the
           opinion that  the number  of locomotives  actually in
           operation  in  Air Quality Control Regions  are  lower
           than   the   numbers   utilized  in  the  report.   EMD

-------
                   -3-
expressed the opinion that,  for  Chicago  for example,
the  report  overestimated the  number of.  locomotives
in use by  a factor of 2.5.   This area of  the report
was  not   changed   because   of   lack   of   data   to
substantiate EMD's  opinion.   This comment  should be
borne  in  mind  when  reading  the  report,  however,
because the estimates of  locomotive  emissions in the
AQCRs are directly  proportional  to  the estimates of
the number of locomotives in use in  the AQCRs.

Emission control  technology effects  and  costs.  EMD
recommended  that  the  report  present  a  generalized
review of  technological emission control  approaches
and  the  omission  of  quantitative values  associated
with control technology  effects.  EMD  also expressed
the  opinion that  reliable  cost  estimates  could  not
be developed  lacking  better understanding  of design
changes  associated with  the application  of  control
technologies.

While EPA  shares  the concerns  raised  by  EMD,  it is
the  opinion of EPA that sufficient information is
available  to allow the  development  of first  order
estimates  of  the  effects  and costs   of  control
technologies.   These  sections  of  the  report  were,
therefore, not changed as a  result  of  the reviewers'
comments.

-------
                              -4-
                  2.0  SUMMARY OF CONCLUSIONS

     Data available  in  the literature, for  utilization in  the
performance of  the  railroad emissions estimates  and especially
for  projections  of   emissions  reductions   achievable,  were  not
extensive.   The results of  the  study must, therefore,  be viewed
as providing indications with  respect  to  the areas  of  probable
concern  and   corrections   thereof,  rather   than  an   exact
determination  of  the  impact  of  railroad  emissions  on  the
environment.

     Conclusions  resulting  from the performance of this  study
are as follows:

     1.    State and local  regulations exist  in most localities
for  some control  of locomotive  emissions.   The regulations  and
their  enforcement are  directed  almost exclusively to  visible
emissions  (smoke) with  little  if   any  attention  to  invisible
gaseous emissions (i.e., HC, CO, and NOx).   The regulations  are
directed to the steady-state  operations of the locomotives with
provisions  for  exceeding  the  standards under  those conditions
which  are  associated  with  short   term   and  generally  higher
emission levels,  e.g.,  during  maintenance, after a  cold start,
after prolonged periods of idling and during accelerations.

     2.    There  are   relatively   large   differences   in  the
stringencies  of  existing  regulations.    The  stringencies  of
existing state  and  local regulations  ranged between an opacity
limit  of 20  percent  for   the  most  stringent  standard  to  an
opacity  limit  of  60 percent  for the  least  stringent  standard.
These  differences do not  appear to pose  significant  problems
with respect to  interstate  commerce, possibly because of either
weak  enforcement  of the  regulations  or  because the  railroads
maintain separate fleets of locomotives for each region.  There
was  insufficient  data  to  either  support  or  refute  a  third
possibility, namely  that  the most  stringent standard  is easily
achievable.

     3.    The  stringency  of  enforcement  of  local regulations
varies from locality to locality.

     4.    Transferral of locomotives  with high smoke emissions
from areas  of  strict enforcement to areas of either weak  or no
enforcement appears  to be  a practice  used by railroads.  There
is  not  sufficient   data  to  determine  whether  this  practice
results  in  large numbers  of  high-emitting locomotives  being
concentrated  in  some   areas  of  the  nation  or  whether  as   a
result,  railroad  emissions  in these areas are higher than would
otherwise be predicted.

-------
                              -5-
     5.     There was  no  particulate emissions  data which  had
been  collected  from  locomotives.    A  very  small  amount  of
particulate data were  available on two  locomotive  type engines
which are  stationary  mounted at  Southwest Research  Institute.
One engine was  an EMD  product,  the other was a GE product.   The
results of what were largely single  tests per  engine  at  each
power setting,  expressed as  grams per  brake horsepower  hour,
(g/BHP-hr) ranged from a low of  0.24 g/BHP-hr to  a  high of  2.83
g/BHP-hr.     It   is   not   known   whether   these   data   are
representative of locomotive particulate  emission rates.  Since
locomotives use  diesel  engines, they should  possibly be viewed
as sources of concern for particulate emissions.*

     6.     Railroad emissions estimates  developed in this study
indicate that  railroad  contributions  to  the total anthropogenic
emissions  of  hydrocarbons,  carbon  monoxide   and  oxides  of
nitrogen  in  the  five  AQCR's   studied   are  in  the  following
ranges:   for hydrocarbons the range is from  just  over one tenth
of one  percent to  one and a third  percent;  for  carbon monoxide
the range is from under one tenth of one  percent  to one-half of
one  percent  and for  oxides  of  nitrogen  the range  is  from two
and a third percent to almost fifteen percent.   It  could not be
determined from  the data whether  areas of concentrated railroad
activity  constituted  a  significant  source  of  emissions  for
adjacent populated areas.

     7.     Technological  approaches for  the  reduction  of  NOx
emissions  from  diesel  engines   generally result  in  increased
particulate  emissions  (smoke)  and increased  fuel  consumption.
Simultaneous reductions in particulate and NOx  emissions may be
achievable by derating  the  power  of the locomotive.   However,
if derating  were to result  in  the use  of a greater  number of
locomotives  in  order  to perform  the same  function  (e.g.,  an
additional  locomotive  on  a  train),  there  could  be  a  net
increase  in  total  emissions  in  spite of there being a reduction
in emissions from each locomotive.

     8.    Modification of locomotive duty  cycles to reduce the
amount  of time that locomotive engines  are  allowed to idle has
the potential  for  reducing  railroad emissions as well  as being
a  strategy   for  fuel   conservation.    Solutions  to  several
     EPA's  Office  of  Research  and  Development  is  presently
     developing  a  health assessment document  for diesel engine
     emissions.  This assessment  is being  based,  in large part,
     on  new  epidemiological  studies   (references  19   and  20)
     involving  exposure of  railroad  workers  with  support  for
     diesel particulate exposure by railroad  employees derived
     from  data collected  from two  locomotive type  engines by
     Southwest Research Institute.

-------
                              -6-
technical  problems  must,  however,  be  developed  before  this
approach  to  emission  control  and  fuel  conservation  can  be
employed under all ambient temperatures.

     9.     The   cost-effectiveness   of    reducing   locomotive
emissions appears similar to the  cost  effectiveness of controls
for automobiles,  trucks,  and motorcycles, but  data for  making
these estimates is very limited.

     10.   Locomotives  remain  in  service   for   much   longer
periods of time  than  do trucks and  passenger cars.   If  future
emissions data warrants the  regulation of locomotive emissions,
consideration should be given to  reducing emissions from in-use
locomotives as well as new locomotives.

     11.   Enforcement  of   gaseous   emissions   standards   and
particulate  standards  cannot  be  performed   visually   as  is
apparently   the   present   practice   for   smoke   emissions.
Development  of  appropriate  test  equipment  and  test  cycles
would,   therefore,    be   necessary.     Information   currently
available does  not  indicate what level   of  difficulty  or  cost
would be associated with  the development of  test procedures and
their implementation.

-------
                              -7-
       3.0  RECOMMENDATIONS PERTAINING TO FEDERAL ACTION

     This study has  shown that there  is  relatively  little data
on gaseous  emissions  and  essentially  no data  on  particulate
emissions from  locomotives  and secondary power  sources  used by
the  railroad  industry.   The  study  has  also shown  that  data
pertaining to the application  of  emission control technology to
locomotive engines is lacking.   Estimates made in  this  study of
railroad  emissions,   their   control,   and  their   impact  on  air
quality  included  assumptions  in  those  areas where  data  were
lacking.  While the  emissions  estimates  were inconclusive, they
suggest  that  railroads could  be viewed  with some  concern  as
sources  of  oxides  of nitrogen  emissions in some Air  Quality
Control Regions.  To  some  lesser extent, hydrocarbon  emissions
from railroads  may  also be of some concern  in some  Air Quality
Control   Regions.    Preliminary   assessment   of   technology
indicates that control of  these emissions may be  cost-effective.

     It  is  recommended  that  sufficient  data   on  locomotive
emissions be  collected to permit  an  accurate determination of
railroad  emissions   and   their  effects.    Areas  where  data
collection needs  to  be emphasized are: 1) particulate  emission
rates,   2)  locomotive  duty  cycles,   3)  the  distribution  of
locomotives throughout  the country,   and  4)   the identification
of  local concentrations  of  high  emitting  locomotives.   It is
further   recommended  that  techniques   for   the   control   of
locomotive  emissions be  evaluated with  respect  to  feasibility
of  application  to  both  new  and  in-use  locomotives,   cost  of
control  and  impact  on railroad  operations.   Such studies would
reduce  the uncertainties  contained  in  present  estimates  and
allow determination  of  the need for  Federal control of railroad
emissions.

-------
                              -8-
                        4.0 BACKGROUND

     This section of the  report  gives  a general overview of the
study and  identifies some  characteristics of  locomotives  and
their  method  of  operation which  are  pertinent  to  emission
patterns  of   railroad   equipment.    Modifications  of   these
characteristics  with the  objective of  reducing .emissions  and
fuel consumption constitute part of the  section of  the  report
entitled, "Emission Reduction Techniques."

4 .1  Selection of Study Areas

     Three  criteria  were  used   in choosing  the  Air  Quality
Control  Regions   (AQCR)   which   were   analyzed.    The   first
criterion was  to meet  the Clean Air Act  (CAA)  requirement that
the effects  of railroads  on air quality  be  investigated  on  a
nationwide basis.   Regions  located in  several  geographic areas
of  the  country were, therefore,   chosen.   The  second criterion
was  that  violations   of  the   National   Ambient  Air  Quality
Standards  (NAAQS)   should  either  be  present  in  the  regions
studied  or that  the regions should have  ambient concentrations
approaching  the  standards.   The third  criteria was that  the
regions  studied  should  have   significant   concentrations  of
railroad traffic.   The  second and  third  criteria were  used to
define  areas  with   pollution  problems  that may be  partially
attributable  to  railroads.   It  was assumed  that  if railroads
were shown not to  be  significant contributors  to  pollution in
these regions, then they should not  be  major   contributors in
other areas of the  country.

     Five AQCRs  were selected for  study  on the  basis  of these
criteria.  The AQCRs are:   Philadelphia, Chicago,   St.  Louis,
Kansas  City,   and  Los  Angeles.   These AQCRs  and  the  included
railroad lines are  shown in Figures 1 through 7.   These regions
are  located   throughout  the nation.   The NAAQS  for ozone is
violated in each of these regions.  Kansas City,  St.  Louis, and
Chicago  in particular  have some of the greatest concentrations
of  railroad  traffic  in  the  nation.    The  regions  selected
should,  therefore,  represent  a  "worst  case"  for  ozone  as
related to railroad activity.

4.2  Locomotive Design Characteristics Pertinent to the Study

     Diesel  locomotives  incorporate some  design  features which
are   somewhat  different   to  those   found  in   other   ground
transportation   engines.    The   design   features   of  specific
interest  in  this   study  are:    1)  the  design  of  the  engine
cooling  systems,  and 2) the method of  controlling engine power.

     The  first  design  feature  which   is  of  interest  in  this
study  is  the  procedure  used   in  controlling  engine  coolant

-------
MllM
                                    -9-
                                                                           Princeton t!<
   I'-.
          Poetati
                   '..  T      .-v^-«C
                   \   ..-:.--•-"-'-*    'vv-:---j
             „..—••I'v.'."."—""""""*"""*'.,   ^r'**."''
           ::;.i^-              --...    -
           PENNSYLVANIA

            *
       .-•*".*
                                                                        	  Railroad Routes

                                                                        -—	  Railroad Traffic
                                                                                    Density  Tonnage
                                                                                    Class
                                  Figure  1

                            Philadelphia  AQCR
Tonnage
Class
1
2
3
A
5
6
7
Millions of
Gross ton-miles
per mile of Track
0-0.99
1-4.99
5-9.99
10-19.99
20-29.99
30-39.99
40 and over

-------
-10-

                                                                   Millions of
                                                                 Gross coo-nllci
                                                                 ier mile of Track
            >-/\:::'......
   / .-V''    ///    \l   /'
   .^---  '     •//  ;l    -..   /,
•-          '
                        -^••'••••

                            l  ,\    \  \    ""'•-..
                            1  >•.        •,        -...
         Flgur*  2

      Chicago AQCR

-------
        -11-
                                                Railroad Boutas
                                                Railroad Traffic
                                                Density Tonnagi
                                                   Millions of
                                                  Gross con-tailes
                                                 per mile of Trac
                                                     0-0.99
                                                     1-4.99
                                                     5-9.99
                                                    10-19.99
                                                    20-29.99
                                                    30-39.99
                                                       and over
                          |   •'. ^W-i\fc:
                                     ^•^^^^^
                                     •     :  *       i* •'  '"VV
                            •r
       Figure  3

Central  Chicago

-------
                                                                 Railroad  Routes
                                                                 Railroad  Traffic
                                                                 Density Tonnage
                                                                 Class
                                                                     Millions of
                                                                   Gross  ton-miles
                                                                  per mile of Track
                                                                       0-0.99
                                                                       1-4.99
                                                                       5-9.99
                                                                      10-19.99
                                                                      20-29.99
                                                                      30-39.99
                                                                      40  and over
    Figure 4

St.  Louis AQCR

-------
                                                -13-
KANSAS
                                                                                    -------  Railroad Rouces

•^-v/





••f ••"!
-'
/
.-•'7


0 S 10
Mllai
Jk
/ '
/ i
"ouch* |/
V
/ i


15


' '/ \ .V
\ \
\\
/ V
\l



1
'*"""—.. 3

/ '••"'••-. '

/ " 	 "
1

-— 	 Railroad Traffic
Density Tonnage
Class
Millions of
Tonnage Gross ton-miles
Class per mile of Track
1 0-0.99
2 1-4.99
3 5-9.99
4 10-19.99
5- 20-29.99
6 30-39.99
7 40 and over



Figure 5

-------
                                                                                       Legend
                                                                               Tonnage
                                                                                Class
Railroad Routes
Railroad Traffic
Density Tonnage
Class

    Millions of
  Gross ton-miles
 per mile of Track
      0-0.99
      1-4.99
      5-9.99
    10-19.99
    20-29.99
    30-39.99
    40 and over
(b
                                                                             J	10     10
                                     Figure  6

                                 Los Angeles AQCR

-------
                                                                                                             Legend
HI lea
                                  T
                                     -   •'•'  '"' --..;.;^-'i]Tsi;'rrfT1il*i!!ij::.U •': v*.1 '  :  ••  .       >"
   Millions ol
 Gross con~miles
per mile of Track
    0-0.99
    1-4.99
    5-9.99
    10-19.99
    20-29.99
    30-39.99
    40 and over
       PACIFIC UCliAN
                                                    Figure 7

                                         Central  Los Angeles

-------
                             -16-
temperature.    Normal  practice  for   the   control   of  coolant
temperature and, therefore, engine temperature  in  road vehicles
is to thermally separate  the  coolant  into two parts.   These two
parts or  volumes  of coolant  are  located in  the  engine and  in
the  radiator  and  are  thermally  separated  by the  thermostat.
Coolant  temperature in  the  engine  and  the  resulting  engine
temperature is controlled  by  the  thermostat which regulates the
flow of coolant to  the radiator.  Use of this design approach
results  in:   1)  rapid engine  warm-up following  a cold  start
because only a  fraction of the  coolant in the  engine  is  warmed
during   engine    warm-up,   2)   relatively   constant   coolant
temperature  and  as  a   result   relatively   constant   engine
temperature is  maintained as  changes  in engine load  occur, and
3)   relatively    constant  coolant    temperature   and   engine
temperature   is    maintained   versus   changes    in   ambient
temperature.   This  approach  to  the  design  of engine  cooling
systems requires  the use  of  an  antifreeze to  prevent freezing
of the coolant  in the  engine  and the  radiator  during  shutdown,
and in the radiator when  the  vehicle is operated in low ambient
temperatures.

     Normal practice  in locomotive design  is  to  treat all  of
the  coolant  as   a  single  volume   and   to  control  coolant
temperature by  limiting  the  capability for  heat  rejection  at
the radiator.   Two  approaches are used on  locomotives to  limit
the ability of the radiator to reject  heat.   One  approach  is  to
control the amount  of  air flowing across the radiator by use of
a  variable speed  fan  drive  and shutters   while  keeping  the
radiator  full of  coolant.  The other  approach  is  to  store that
part of  the coolant which is  not  in the  engine  in a  large
reservoir  and  to divert  coolant to  the radiator  as  necessary
for temperature control, i.e., little  or no  coolant is normally
found  in  the  radiator.   This  second approach to  temperature
control of  locomotive  engines  is  normally  referred  to as the
dry radiator method.   On  some  newer  locomotives where  the dry
radiator  method of  control  is utilized,  air  flow across  the
radiator  is  also  controlled by  use  of  a  variable-speed fan.
Many older, dry  radiator  locomotives  employ  constant-speed fan
drives  which  require  a  greater  amount   of  power   for  their
operation than that required by the variable-speed fans.

     As  a result  of  this cooling  system   design  philosophy,
locomotive  engines:   1)   require  an  extensive warm-up  period
following  shutdown  because not  only  must the great mass  of the
engine be  warmed  but all  of the coolant must also be warmed,  2)
experience relatively large changes  in coolant temperature, and
as  a  result  engine temperature,  between  idle and  full  power
operation,  and  3)  experience   relatively  large  changes  in
coolant  temperature and  engine  temperature  as  a  function  of
ambient temperature.  The  effects of the  cooling  system designs
on  locomotive  emissions,  within  railroad  yards,  have  been

-------
                             -17-
that:     1)    exhaust   emissions   (mass  basis)   from   idling
locomotives  constitute  a  significant  fraction  of  the  total
emissions  from  locomotives,  2)  start-up  emissions,  while  the
occurrences are infrequent, are higher than would  occur  if only
part  of  the coolant  needed  to be warmed  up,  and  3)  emissions
levels  following   a   prolonged  idle  period  tend  to  be  high
because of the general cooling of the engine.

     The  second  design  feature which  is  peculiar to  railroad
locomotives  and  which was of  interest in  this  study   is  the
design  and  operation  of   the  throttle.   Power  settings  for
railroad  engines   (throttle  position)  generally  involve  eight
discrete  positions or notches  on the throttle gate in addition
to  the  idle and dynamic brake  positions.   Each  notch position
is  numerically  identified, with  notch position one  being  the
lowest,  off-idle,   power  setting   and position  eight  being
maximum  power.   In the  dynamic brake position,  the  propulsion
system  provides  a   degree   of  braking.   The   dynamic  brake
position is not usually found on switch engines.

     The throttle  lever  in the  cab of the locomotive is usually
connected  to the  engine by  electrical  means as  opposed to  a
mechanical  connection.  Because  of  this  type  of  connection,
each notch on the throttle corresponds  to  a discrete  setting on
the fuel  delivery  system of the engine and there  are no engine
power  settings  which  correspond  to  throttle  settings  between
any  two  notch  positions.    The  net  effect of  this  method of
control  is  that the  engines can operate at only eight distinct
power levels, in addition to  idle and dynamic brake.

     During  accelerations,  the  usual  practice  for  throttle
operation  is for  single notch, stepwise  increases in  power as
opposed  to  a  sweeping  change to  the highest  notch  position
which will ultimately be employed.

4.3  Locomotive Duty  Cycles

     The pattern of operation followed  by  a piece of equipment,
expressed  in terms such a percent  of  time at  a  defined load,
speed  or  other  readily  identifiable  parameter,  is  usually
referred  to  as  the duty cycle  for that piece of equipment.  The
combination  of the design  of  the  throttle  control  and the usual
method  of operation  has permitted  locomotive manufacturers and
the  railroads  to  establish  historical operating patterns  or
duty   cycles   for    locomotives   based  upon   throttle  notch
position.  This information was of  assistance  in estimating the
emission  contributions of  railroads  within Air  Quality Control
Regions.   Substantial increases  in  the cost  of  fuel  over  the
past  several years has  caused  railroads to seek procedures for
conserving   fuel.   One  of   the  procedures  which   has  been
introduced  is  a reduction  in the time that locomotives spend at

-------
                             -18-
idle.  Precise identification  of  the  amount of change which has
occurred  was  not  possible,  however,  because  of  the lack  of
recent data  on the  duty cycle.  Historical  operating patterns
were, therefore,  used as the basis in this study.

     The  locomotive  duty  cycle describes  the amount of  time
spent in  each of  the throttle notch positions when a locomotive
is  available  for  service.  As  such,  the distribution  not  only
accounts  for  the  time a unit  spends engaged  in moving freight,
but  also  the time  spent  at   idle,  either  incurred while  the
locomotive  is  awaiting  assignment  and  is  not yet  an integral
part  of  a train  or  is  part  of  a train which  is  stationary.
Time  during  which  the  locomotive  is  unavailable  for  service
resulting from factors  such as the need for  a major overhaul,
scheduled maintenance,   or  inspections  would be  excluded  from
the daily duty cycle.  Data were not available  on the amount of
time  that  locomotives are  not  available for  service.   In  this
analysis it is assumed that on  an annual basis,  locomotives are
unavailable  for   service  for   approximately  5 percent of  the
total time  in each  year.   Expressed  as  the  number of days  in
which locomotives are not available for service,  this 5 percent
factor corresponds  to  20 days per  year.  On an  annual  basis
each  locomotive is  assumed,  therefore,  to be in service for 345
days.

     Within the generic definition of duty  cycle,  there are two
distinct  classes  of  locomotive duty cycle.   The  classes apply:
1)  to line-haul  locomotives,  and  2)  to  switch  and  transfer
locomotives.  Switch  and transfer  locomotives generally operate
within  localized  areas  and their  duty cycles are,  therefore,
directly  applicable  within   Air   Quality   Control  Regions.
Line-haul locomotives cross  AQCR  boundaries  and  pose problems,
therefore, with respect  to the determination of their emissions
within an AQCR.*

      In an attempt to define  the  most  representative duty cycle
for   each  type  of   locomotive  within  each  AQCR,  historical
throttle  clock  data  from the locomotive  manufacturers  were
reviewed.  The information available  was not adequate to depict
     The horsepower ratings of  locomotives  range  from a high of
     about  5,000  hp  to  under  1,000  hp.    The higher  power
     locomotives  are  usually used  for  line-haul  purposes with
     the  smaller   locomotives   used  in  switch  and  transfer
     operations.   There  is  no  firm rule  for  distinguishing
     between  the  two  types of service with respect to the power
     rating  of locomotives.   For purposes  of this  study,  all
     locomotives  of 1,500  hp and  less  were  treated as switch
     and  transfer  units with  the remainder of  the   locomotives
     classed  as line-haul.

-------
                             -19-
individual regional differences.  It  is  reasonable,  however,  to
expect  that  the  line-haul   locomotives  of  western  railroads
could spend more time  in throttle notch eight (full  power)  and
in  dynamic  brake   (descending  long  mountain grades)  than  do
their eastern  counterparts.   This expectation is  based on  the
more  mountainous terrain  found  in  the  western  areas  of  the
country which  could entail the extensive  use of  maximum  power
when  climbing  grades  and dynamic brake  when  descending grades.
Since duty cycles  for  each region could not  be  formulated,  it
was decided  to use the existing  industry duty cycles with some
modifications where appropriate.

4.3.1  Switch and Transfer Locomotive Duty Cycle

     Generally,  switch  locomotives  operate within  a switchyard
or  terminal  area,  while  transfer   locomotives  move  rail  cars
between  switchyards  and  are  involved  in  branchline  service
where rail  cars are  delivered to  or received  from customers.
Switch  and transfer  locomotives  are  typically  smaller,  lower
power units.

     The   literature   contains   two   duty  cycles  for  switch
locomotives  (Table  1) .   The  literature  does not  contain  any
duty  cycles  for  transfer  locomotives.   It  was  not possible,
therefore, to  distinguish between  duty cycles  for  switch  and
transfer  locomotives.   It  was assumed that locomotives employed
in  transfer  service  experience the  same  duty  cycle  as  those
employed  in  switch  service and  a single duty cycle was used for
these two types  of  service.   As  shown in  Table  1,  the  two
switch  locomotive   duty cycles are  nearly identical.   Dynamic
brake   is  not   included  in  these  cycles   because  switch
locomotives  are not usually  equipped with this  feature.*  The
Atchinson,  Topeka,  and  Santa  Fe   Railroad  cycle   (ATSF)  was
apparently based on only  that  railroad's  operating experience.
The  Electromotive  Division of  General  Motors  (EMD)  cycle  was
chosen  to be  more  representative of  national operations  since
it  seems  likely that  it was generated from throttle clock data
recorded  from  several  railroads.   The duty cycle for switch and
transfer  locomotives  represents operation  on  a 24  hour  per  day
basis  for  each day   of the  year   that  the  locomotive  is  in
service.
     During  dynamic   brake  operation,   the  function  of  the
     locomotive  powerplant  (engine,   generator  and  traction
     motors) is reversed and the powerplant serves as a brake.

-------
                             -20-


                           Table  1


                Switch Locomotive Duty Cycles
                                       Percent of Time
                                         in Each Notch,
                                         Per  24-Hour Day
             Throttle
          Notch  Position              ATSFl/       EMDl/
          Engine Off                     0              0

          Idle                          77             77

          Notch 1                       10              7

          Notch 2                        57

          Notch 3                        44

          Notch 4                        22

          Notch 5                        11

          Notch 6                        1              0.5

          Notch 7                        0              0.5

          Notch 8                        01
I/   ATSF - Atchinson,  Topeka and Santa Fe.
     EMD - Electromotive Division,  General  Motors.

-------
                             -21-
4.3.2  Line-Haul Locomotive Duty Cycle

     Duty  cycles  which  have  been  developed by  the  railroad
industry   and   by   locomotive   manufactures   for   line-haul
locomotives  are  shown in  Table 2.   The EMD heavy-duty  cycle,
the General  Electric  (GE)  maximum  cycle and the Association  of
American Railroads  (AAR) duty  cycles  may be most representative
of  locomotive  operation  between  cities  or  western  railroad
service  in  general.   The EMD  medium-duty  cycle  and  the  GE
average  duty  cycles  are  probably  most  typical  of  overall
line-haul freight operations on a national basis.

     Since the  throttle  notch data upon  which these cycles are
based  were   derived   from  a   combination  of   intercity  and
intracity  operations,   the  data   would  not  be  expected  to
precisely  define  operations  within  the  metropolitan/suburban
AQCRs  under  study.   It  was   concluded,  therefore,  that  the
industry duty cycles  having  high percentages of  throttle  notch
eight  operation  were  inappropriate  for use within  AQCRs.   Two
reasons predominate  in arriving at this  conclusion.   The  first
reason  is  the lower  average  speed of  trains when operating in
the  metropolitan  regions  as  compared  to   the  long  intercity
distances where  higher speeds  are  maintained.   Operation of the
locomotive  in the  highest power  settings  (notches  seven  and
eight) would, therefore,  tend  to be emphasized during intercity
operation.   Second,  although  trains  are  accelerating as  they
leave  the  switchyards  or other congested areas,  and as a result
require  the  use of  relatively  high power  settings,  it is  not
normal  operating practice  to  "sweep"  the  throttle  from  a low
notch  position  to notch  seven or  eight.    Each higher  notch
position  of  the  throttle  is  not  normally  selected until  a
significant  portion  of  the  ultimate  power in the  preceding
notch   has   been   achieved.     Because   accelerations    are
accomplished  with  an orderly and gradual progression to  higher
notch   positions,   a  relatively   small  amount  of   time  is
accumulated  in the  highest notches.   At reduced  speeds,  dynamic
brake  becomes inefficient  and its  excessive  use  could  damage
the  traction motor of  the  locomotive  because  of  poor  heat
transfer  characteristics.   The  percentage  of  time spent  in
dynamic  braking  within  an AQCR is,  therefore,  expected  to  be
less   than   that  which  exists  for  the overall  operation  of
line-haul  locomotives.

     On   this   basis,   EPA  selected   the  General   Electric
minimum-duty cycle for  line-haul  locomotives as  the basis for
representative operation within an AQCR  (Table  3).   The  values
selected  for the  percentage of time spent in  each notch were
rounded to whole numbers so as  to  avoid the appearance of great
accuracy.   In applying  the duty  cycle  selected  for line-haul
locomotive operation  in  an AQCR, it  was assumed  that 50 percent
of  the total  line-haul  locomotive  operational  time  was  spent
within AQCRs  and   the  other  50  percent  was   spent in  rural

-------
                                    -22-
                                   Table 2
                       Line-Haul Locomotive Duty Cycles
Throttle
Percent of Time in Each Notch Per 24-Hour
Notch
Position GE^/Min.
Engine Off
Idle
Notch 1
Notch 2
Notch 3
Notch 4
Notch 5
Notch 6
Notch 7
Notch 8
Dyn. Brake
0
59
6
6
6
6
2
2
2
5
1

.0
.5
.5
.5
.5
.9
.9
.5
.2
.5
GE Max.
0
40
2
2
2
2
1
1
1
38
7

.0
.5
.5
.5
.1
.7
.7
.8
.0
.0
GE Avg . 1
0
54
5
2
2
5
2
2
2
21
4

.0
.0
.5
.0
.0
.0
.0
.5
.0
.0
EMD
GE Avg. 2 Heavy!7
0
53.
5.
3.
3.
3.
2.
3.
2.
17.
5.

0
1
9
4
3
8
4
6
0
5
0
41
3
3
3
3
3
3
3
30
8

.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
Day
EMD
Med.
0
46
4
4
4
4
4
4
4
17
9

AAR!X
0
43.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
28.0
8.0
I/   AAR - Association of American Railroads.
      GE - General Electric.
     EMD - Electromotive Division, General Motors.

-------
                     -23-


                   Table 3
Line-Haul Locomotive Duty Cycle Applicable to
 Operation Within Air Quality  Control Regions
                               Percent of Time
     Throttle                     in Each Notch,
 Notch  Position                  Per  12-Hour Day

 Engine Off                            0

 Idle                                 59

 Notch  1                              7

 Notch  2                              6

 Notch  3                              6

 Notch  4                              6

 Notch  5                              3

 Notch  6                              3

 Notch  7                              3

 Notch  8                              5

 Dyn. Brake                           2

-------
                             -24-
areas.  The duty cycle for rural  line-haul  locomotive operation
was  calculated*  (Table  4)  and  compared  to  the  maximum-duty
cycles  as  a  check  of  the  assumptions  which  had  been  made.
Because  of  the similarity  between  the  calculated  rural-duty
cycle and  the AAR, EMD heavy-duty  and GE  maximum-duty cycles,
it was concluded that the assumptions were reasonable.

4.3.3  Secondary Power Source Duty Cycle

     The third  duty cycle used  in this  study  was that  of  the
relatively   small   diesel   engines   used   to   operate   the
refrigeration systems of refrigerated rail  cars.   This group of
engines constitute the secondary  power sources which are found
on trains.

     The duty cycle  for  these  engines was obtained from Pacific
Fruit Express  (PFE  1978).   This cycle  (Table 5)  is  based on
transcontinental shipments lasting  approximately  ten  days.  The
cargo in this  instance  is not pre-cooled;  hence, the  cycle is
more  rigorous  than if pre-cooled  or pre-frozen  food was to be
transported.  For the purposes of this  analysis,  the  duty cycle
is assumed to  occur  over  a 24 hour period for each refrigerated
rail car that is in service in an AQCR.
     Difference between  time  spent  in a throttle notch position
     in medium duty  cycle on a 24 hour basis  and  time spent in
     the  same  throttle notch position  in  minimum  duty cycle on
     a 12 hour basis gives the time spent  in  the  same throttle
     notch  position  in the calculated  rural duty  cycle  on a 12
     hour basis.

-------
                             -25-
                           Table  4
           Line-Haul Locomotive Duty Cycles Overall
           Average, Within AQCRs and In Rural Areas
Throttle
Position
Engine Off
Idle
Notch l
Notch 2
Notch 3
Notch 4
Notch 5
Notch 6
Notch 7
Notch 8
Dynamic
Brake
Overall
Averaqel/
0
51
5
4
3
4
3
3
3
18
Within
AQCR2/
0
59
7
6
6
6
3
3
3
5
In Rural
Areas!/
0
43
3
2
0
2
3
3
3
31
62 10
I/ Average of two G.E. average cycles and EMD medium
applicable to total time; i.e., 24 hours per day.
2/ Based on G.E. minimum cycles and applicable to 01
     total time, i.e.,12 hours per day.
3/   Based on  difference between hours  spent  in each  notch of the
     overall average and within AQCRs cycles.

-------
                               -26-
    Mode
                              Table 5
                      Refrigerated Rail Car
                   Duty Cycle and Engine Loading
              Percentage of    Engine
Maximum Cool
High Speed Cool
Low Speed Cool
Low Speed -
No Cooling
Time in Mode  Speed (rpm)
     10          1,200
     10          1,200
     50            800
     30            800
                           Rated Power
               Percent
              of Rated
 @ Speed    Power @ Speed
33 @ 1,200   95% <§ 1,200
33 @ 1,200   72% @ 1,200
20 <§ 800     67% @ 800
20 @ 800     20% @ 800

-------
                             -27-
               5.0  RAILROAD EMISSION ESTIMATES

     Three  parameters  were  used  in  the  development  of  the
estimates  of  railroad  exhaust emissions  within  the five  Air
Quality  Control  Regions  which were  studied.   The  parameters
used  were:  1)  the locomotive  and refrigerated  rail car  duty
cycles   previously  developed,   2)   exhaust   emissions   from
locomotives  and  refrigerated  rail  car  engines,  and  3)  the
number of locomotives  and refrigerated  rail cars  in use  in each
Air Quality Control Region.

5.1  Locomotive Exhaust Emissions

     Three  sets  of data  on  locomotive emissions  were utilized
in   developing   estimates    of   individual   source   exhaust
emissions.   Southwest  Research  Institute  (SwRI)  tested  three
locomotives  as  part  of  Contract Number   EHS  70-108  for  the
Environmental Protection Agency.[1]  This work  was performed in
1971-72.   The locomotives  tested were  in-use  units  obtained
from Southern Pacific  and included a  1200 hp EMD switch engine,
a 3000 hp EMD line-haul engine and 3600 hp GE line-haul engine.

     The  second  set   of  data   on  locomotive  emissions  was
furnished  by  EMD.   This data  set consisted of  data  taken from
new  locomotives  in  1978.   The  data  were presented  as  mean
results  from  nine  1500 hp engines, eighteen 2000 hp engines and
twenty-two  3000  hp engines.   The standard  deviations  of  the
data  were also presented.   The  third  and  largest set  of data
was  provided  by the  Association  of  American  Railroads  (AAR).
This  data set was  collected by AAR between 1981  and  1983.  The
data  were presented as  the  means and  standard  deviations  for
each  of the  groups  of  engines tested.   In total,  fifty  seven
in-use  locomotives were tested.   On a  manufacturer  basis,  the
locomotives  tested  were  as  follows:   EMD;   sixteen  2000  hp
engines,  two  2250  hp  engines, four  2500  hp  engines,  fifteen
3000  hp engines and  eight  3600  hp engines;  G.E; five  3000 hp
engines  and seven  3600  hp  engines.   Within  this data  set  the
engines   tested  were   divided  into   two  groups.   One  group
consisted  of engines  which  were tested   immediately  after   a
major  overhaul.    The  other  group consisted  of  engines  which
were  tested just prior to the performance  of a major overhaul.
In   utilizing this  data  (AAR  data)   EPA  treated  the  data
collected  from engines  immediately  after  a major overhaul as
being eguivalent to new  locomotive data.

     The  estimates   of  in-use   locomotive   emissions   were
developed  from  the three data sets  as  follows.   In  the  first
step,  the emission data on each  engine  type  (e.g.,  the 3000 hp
EMD  locomotive)  from  each  source were inspected  to determine
whether  the  data  were  similar or dissimilar.    Since  the test
data  on each  engine tested  in each data set  were not provided,

-------
                             -28-
an   approximate    rather    than   a   statistically    precise
determination  of  similarity  or dissimilarity  was made.   Test
results from the three data  sources  on locomotives of the  same
specification were treated as being  similar  when the  difference
between the mean results was  less  than the sum of the  standard
deviations  for the  majority of  the  data on  that  locomotive
type.   Application of this test for  similarity  or  dissimilarity
between data sets resulted in the  removal from the data  base of
one set of test results on a  single locomotive  (the 3000  hp EMD
locomotive tested by SwRI  in  1971-1972).

     The  second  step  in  the  development of  the estimates  of
emissions from in-use  locomotives was  the calculation of "new"
locomotive and "used"  locomotive (i.e.,  locomotives  approaching
a major  overhaul)  emission rates.    "New" locomotive  emissions
were derived  from  the  data  supplied  by EMD  and from  the AAR
data on  locomotives  immediately following  a  major  overhaul.
"Used"   locomotive emissions  were derived  from  the AAR  data on
locomotives approaching  a major overhaul and  from SwRI  data.
Since the means  of  the emission results  contained  in  each data
set were obtained  from differing   sample  sizes,  the  average
"new"   and  average   "used"  emission  rates were  calculated  as
sampling  size  weighted means; i.e.,  the  sum of the products of
mean sample emissions  and  sample size  was divided by the sum of
the sample sizes  to  develop  the   average  "new"  and  average
"used"     locomotive   emissions.      Implicit    within    this
calculational  methodology  is  the  assumption  that  changes  in
locomotive emissions occur linearly with time.

     The  third step  in the  development  of in-use  locomotive
emissions estimates was the calculation of the  emissions rates,
for  "average"  in-use  locomotives  of  the  types  tested.   This
calculation  utilized the  assumption  that  there  is  a  linear
change  in emission  rates  as  locomotives  age from  the  "new" to
the "used"  condition.  With  this  assumption,  emissions  from an
average  locomotive  are calculated  as  the mathematical  average
of  the  new and used  values.   The  emissions  rates for  average
locomotives  of the  eight specifications tested  are shown in
Tables  6  through 8.  Hydrocarbon  emissions  are shown  in Table
6,  carbon monoxide  emission in  Table  7  and oxides  of nitrogen
emission  are shown in Table 8.   The units  employed are grams of
pollutant per hour of operation in  each throttle notch position.

     The  fourth  step  in  the procedure  was the  calculation of
the  in-use weighted  average emissions  for  line-haul   and for
switch  and transfer  locomotives.    Since there  are  a  greater
number  of horsepower  ratings  for  in-use  locomotives  than there
are power ratings  in the  data base,  a degree of grouping of the
in-use  locomotives  was necessary.   The in-use  locomotives were
grouped  so  that  the ratings of the  test  locomotives  were equal
to  the  power  ratings  of   the   greater  number   of   in-use

-------
                                               Table 6
                     Average Locomotive Hydrocarbon Emissions by Throttle Setting
                                  Locomotive Emissions  (g/hr)
Throttle
Notch
Idle
1
2
3
4
5
6
7
8
Dyn Brake
EMD.?/
3600 hp
251
193
200
259
306
422
529
767
971
363
GE?/
3600 hp
835
822
4,511
9,728
7,922
8,781
8,786
8,840
10,044
1,941
EMD3/
3000 hp
228
201
.236
316
382
544
728
1,002
1,276
366
GE4/
3000 hp
532
608
1,177
2,220
3,359
3,850
4,029
5,330
6,234
2,271
EMD5/
2250/2500 hp
231
170
180
219
256
339
425
565
806
316
2000 hp
152
146
169
224
304
428
571
789
1,005
252
EMDZ/
1500 hp
97
93
116
145
193
271
367
517
660
—
1000 hp
387
452
638
984
1,482
1,830
2,387
2,960
3,976
—
Data Source, time frame of testing, number of engines tested
I/   AAR, 1981-1983, 9 engines.
2/   AAR, 1981-1983, 3 engines; SwRI, 1971-1972, 1 engine.
3/   AAR, 1981-1983, 12 engines; EMD, 1978, 22 engines.
4/   AAR, 1981-1983, 7 engines.
5/   AAR, 1981-1983, 6 engines.
6/   AAR, 1981-1983, 16 engines; EMD, 1978, 18 engines.
7/   EMD, 1978, 9 engines.
8/   SwRI, 1971-1972, 1 engine.
                                                                                                             VD
                                                                                                             I

-------
                                                Table 7
                   Average Locomotive Carbon Monoxide Emissions by Throttle  Setting
                                           Locomotive Emissions  (g/hr)
Throttle
Notch
Idle
1
2
3
4
5
6
7
8
Dyn Brake
EMD.?/
3600 hp
588
432
484
610
720
1,033
2,480
3,849
4,472
818
GE2-/
3600 hp
481
658
1,588
2,325
3,036
3,721
4,466
4,454
5,099
2,352
EMD!/
3000 hp
566
392
362
466
570
1,160
2,591
5,036
6,092
840
GEi/
3000 hp
1,048
962
2,236
3,858
7,964
8,504
12,038
8,339
7,892
2,780
2250/2500 hp
599
316
304
378
448
832
2,122
5,060
9,868
635
2000 hp
291
208
350
441
458
616
1,120
2,686
5,704
702
EMD!/
1500 hp
174
184
295
337
351
416
659
1,959
5,305

EMD8-/
1000 hp
160
273
341
481
560
702
768
1,052
1,844

Data Source, time frame of testing, number of engines tested
I/   AAR, 1981-1983, 9 engines.
2/   AAR, 1981-1983, 3 engines;  SwRI,  1971-1972, 1 engine.
3/   AAR, 1981-1983, 12 engines; HMD,  1978, 22 engines.
4/   AAR, 1981-1983, 7 engines.
5/   AAR, 198171983, 6 engines.
6/   AAR, 1981-1983, 16 engines, EMD,  1978, 18 engines.
7/   EMD, 1978, 9 engines.
8/   SwRI, 1971-1972, 1 engine.
                                                                                                             o
                                                                                                              I

-------
                                                Table 8
                  Average Locomotive Oxide of Nitrogen Emissions by Throttle  Setting


                                           Locomotive Emissions (g/hr)
Throttle
Notch
Idle
1
2
3
4
5
6
7
8
Dyn Brake
EMD!/
3600 hp
1,561
2,331
3,534
5,852 .
8,685
11,808
15,436
24,774
29,809
3,540
GE2/
3600 hp
978
4,083
11,880
14,944
19,343
23,427
28,061
28,666
33,050
5,043
EMD^/
3000 hp
1,448
3,105
5,367
9,091
13,282
18,626
23,337
30,344
36,409
1,771
GE*'
3000 hp
977
2,009
5,023
13,120
17,871
25,023
29,354
34,262
42,750
5,798
EMD5-/
2250/2500 hp
1,434
1,765
3,163
4,794
6,773
9,306
11,879
15,287
23,859
2,486
EMD§/
2000 hp
1,059
1,506
3,461
6,497
10,648
15,617
21,054
27,112
31,388
3,255
EM)!/
1500 hp
957
1,248
2,763
5,605
9,598
13,932
17,743
21,623
23,864
-
EMD**/
1000 hp
335
626
920
2,003
3,218
4,946
6,718
8,367
10,220
-
Data Source; time frame of testing, number of engines tested
I/   AAR, 1981-1983, 9 engines.
2/   AAR, 1981-1983, 3 engines; SwRI, 1971-1972, 1 engine.
3/   AAR, 1981-1983, 12 engines; EMD, 1978, 22 engines.
4/   AAR, 1981-1983, 7 engines.
5/   AAR, 1981-1983, 6 engines.
6/   AAR, 1981-1983, 16 engines, EMD, 1978, 18 engines.
7/   EMD, 1978, 9 engines.
8/   SwRI, 1971-1972, 1 engine.
                                                                                                              i
                                                                                                             CO

-------
                             -32-
locomotives in each in-use group.   The horsepower  groupings  of
in-use  locomotives,  the number  of in-use  locomotives  in  each
group and  the  horsepower ratings  of   the  test-engines  used  to
represent the in-use groups  are shown  in Table 9.

     The calculation  of  the in-use weighted  emissions averages
was performed by summing the products   of the  emission  rates for
the test engines  and  the number of in-use engines  in  the group
so  represented and  dividing  by  the   total  number  of  in-use
locomotives  in the group.   The results  of   these  calculations
are shown in Table 10  for line-haul locomotives and in Table 11
for switch and transfer locomotives.

     The  fifth   and   final  step   in   the  procedure  was  the
calculation of  locomotive  emissions in  each  AQCR  under study.
This  step  was  performed by  determining the   daily  emissions  of
"average" line-haul and switch and transfer  locomotives when in
an  AQCR and multiplying  by the number of locomotives  of  each
type  in the AQCR and the  number  of  days per  year  that  each
locomotive is in service (i.e., 345 days).

     The procedures  used for  developing  the duty  cycles  were
described in the Background section of  this  report.  These duty
cycles were employed in this calculation.

     The daily  (i.e.,  duty  cycle weighted) locomotive emissions
were  calculated by  determining the product of the  emissions  in
each  throttle  notch   and  the  number   of  hours  spent  in  each
throttle notch position and summing the values  so calculated.
The  results  of these calculations are  shown in Tables 12 and
13.   As can be seen  from  these  tables,  an  average  line-haul
locomotive  emits  on  a  daily  average basis  14.43 Ib.  of HC,
31.03 Ib.  of CO and 160.49  Ib. of NOx,  while an  average switch
and  transfer  locomotive emits 15.28   Ib.,  13.58 Ib.  and  77.66
Ib. of HC, CO, and NOx respectively.

     Estimation of  the number of  locomotives in each AQCR was
performed  by  extrapolating from  information  supplied  by the
Federal  Railroad  Administration   (FRA).    For  the  1978-1979
timeframe, FRA safety inspectors provided EPA  with data on the
number  of  switch  and  transfer locomotives  in  the  AQCRs  being
studied.  This  information  was used to extrapolate to the  total
number  of  locomotives in an AQCR  by  assuming  that there  is   a
direct  relationship between the number of switch  and  transfer
locomotives in a region and  the  number of  line-haul locomotives
serving  that  region.    The  assumption  is  that the fraction of
the total number of line-haul  locomotives  serving any region is
the  same  as  the  fraction  of  the total  number  of  switch and
transfer  locomotives  operating in  that region.  The  rationale
leading  to this assumption  is  that there should be a  relatively
fixed relationship  between  the number  of  line-haul locomotives

-------
                          Table  9
In-Use Locomotive Power Rating Groups, Number of Locomotives

   by Groups, and Test Engines Representing In-Use Groups
Locomotive groupings by power level
Line-Haul
In-Use
Horsepower
Group
Number of
Locomotives
in group

Test Engine
Power Rating

EMD
3300 hp
& Over


1,753

EMD
3,600 hp

GE
3200 hp
& Over


655

GE
3,600 hp

EMD
3200 thru
2800 hp


5,818

EMD
3,000 hp

GE
3000 thru
1800 hp


2,386

GE
3,000 hp

EMD
2700 thru
2250 hp


2,402

EMD
2,500 and
2,250 hp
EMD
2000 thru
1600 hp


6,192

EMD
2,000 hp

Switch and Transfer'
EMD
1500 thru
1300 hp


2,807

EMD
1,500 hp



EMD 1200 hp & under
GE 1000 hp &


2,003 EMD
72 GE
EMD
1,200 hp

under







                                                                                        u>
                                                                                        Co

-------
                       -34-


                     Table 10
         In-Use Weighted Average Line-Haul
     Locomotive Emissions By Throttle Position
Throttle         Emissions (gm/hr per locomotive)
Position           HC           CO          NOx

 Idle              264          540        1,257
  1                250          407        2,248
  2                467          637        4,489
  3                827          945        8,121
  4                961        1,535       11,976
  5              1,161        1,932       16,826
  6              1,306        3,286       21,355
  7              1,662  7     4,564       27,340
  8              2,017        6,481       33,291
 Dynamic           613        1,060        3,112
  Brake
                      Table  11
    In-Use Weighted Average Switch and Transfer
     Locomotive Emissions By Throttle Position
Throttle         Emissions (gm/hr per locomotive)
Position           HC           CO          NOx

 Idle              220          168          692
  1                246          222          983
  2                338          315        1,979
  3                502          398        4,073
  4                741          440        6,884
  5                934          538       10,110
  6              1,226          705       13,053
  7              1,556        1,573       15,984
  8              2,071        3,833       18,060

-------
                          -35-
                        Table 12
             Line-Haul Locomotive Emissions
       Per Locomotive, Per  12-Hour Day in an AQCR
Throttle        Duty
Position        Cycle
 Idle             59
    1              7
    2              6
    3              6
    4              6
    5              3
    6              3
    7              3
    8              5
 Dynamic Brake   	2

    Total        100
        Operating
          Hours

            7.08
            0.84
            0.72
            0.72
            0.72
            0.36
            0.36
            0.36
            0.60
            0.24
           12.00
             Emissions  (pounds/day)
              HC     CO      NOx
             4.12
             0.46
             0.74
             1.31
             1.53
             0.92
             1
             1
   04
   32
 2.67
 0.32
        8.44
        0.75
        1.01
        1.50
        2.44
        1.53
        2.61
        3.62
        8.57
        0.56
            14.43   31.03
        19.61
         4.16
         7.13
        12.89
        19.01
        13.35
        16.95
        21.70
        44.04
         1.65

       160.49
                         Table 13

              Switch and Transfer  Locomotive
        Emissions Per Locomotive,  24-Hours  Per  Day
Throttle
Position

 Idle
    1
    2
    3
    4
    5
    6
    7
    8
 77
  7
  7
  4
  2
  1
  0.5
  0.5
  1
Operating
  Hours

  18.48
   1.68
   1.68
   0.96
   0.48
   0.24
   0.12
   0.12
   0.24
                      Emissions (pounds/day)
                       HC      CO      NOx
 1
 1
8.96
0.91
  25
  06
0.78
0.49
0.32
0.41
1.10
6.85
0.82
1.67
0.84
0.47
0.29
0.19
0.42
2.03
    Total
100
  24.00
15.28   13.58
                77.66

-------
                             -36-
which  service  an area  and the  number of  switch  and  transfer
locomotives  which  service  the  line-haul  locomotives,  because
switch and transfer  locomotives  are  used  to process the freight
cars   which  enter   and  leave   a   region   (freightyard,   for
example).    The  numerical   values  for   these   distributions,
adjusted to  the  number of  locomotives in  service  in  1984,[2]
are given in Table 14.

     Line-haul  and switch and  transfer locomotive emissions for
each AQCR,  expressed as tons per year,  are shown in Table 15.

5.2  Secondary Power Sources Exhaust Emissions

     Emissions   from refrigerated rail  cars  (secondary  power
sources) were  calculated as described below and  the resulting
contributions in each AQCR  are shown in the  railroad emissions
summary table.

     Two basic  types of refrigerated  rail car  are currently in
service throughout  the  nation.   These types are  nonmechanical
and  mechanical   refrigerated   rail   cars.   Nonmechanical  units
utilize either  block ice,  a mixture  of crushed  ice and salt, or
solid   carbon   dioxide  (dry   ice)   as  the   cooling  medium.
Mechanical units are powered by an  internal  combustion engine.
In  this  analysis,   only  mechanical  refrigerated  units  were
considered since only these units have exhaust  emissions while
in the railroad system.

     Mechanical  refrigeration  units  use  a  relatively  small
diesel  engine  to generate  electric power  which  operates  the
refrigeration system.   In addition,  most  cars are equipped with
an  electrical heating  system.    This system is used to defrost
the  car's  interior   or, in  winter,   to protect  perishables from
freezing  temperatures.   Consequently,  most  refrigerated  cars
are  capable  of  maintaining internal temperatures  of from -20°F
to  +70°F and may be operated  throughout  the year.   These cars
are  also used without  their mechanical refrigeration devices in
operation   when   nonperishable   commodities   do   not   require
protection from temperature extremes.

     A  review of  alternative methods for quantifying the number
of  refrigerated cars  in  each  AQCR  concluded  that  the  most
accurate  assessment  could  be  made  by basing  refrigerated car
population  estimates on actual   in-use data  obtained  from the
analysis  of  rail   operations  being conducted  by  the Chicago
Terminal  Project.    This  project was jointly  funded  by  the
Federal Railroad Administration  and  the Association  of American
Railroads.   EPA  was not,  however,   able  to  secure  the desired
data from this source and had to  adopt an alternative method.

-------
                                -37-
                              Table 14
                 Number of Locomotives in Each AQCR
    AQCR
Philadelphia

Chicago

St. Louis

Kansas City

Los Angeles
     Switch and
 Transfer Locomotives
Percentage    Number
of Total1/   of Unitsi/
   1.62

  13.13

   2.06

   2.25

   1.99
 79

641

101

110

 97
            Line-Haul Locomotives
            Percentage   Number
            of Total     of Units3/
 1.62

13.13

 2.06

 2.25

 1.99
  311

2,522

  396

  432

  382
I/   Percentage of  total  switch  and transfer  locomotives  derived
     from count of  switch and  transfer  locomotives in  service in
     each city  as  furnished by Federal  Railroad Administration in
     1979,  divided  by   total  number   of  switch   and  transfer
     locomotives in service in 1979.
2/   Number  of  switch  and  transfer  locomotives  in  each  AQCR
     derived from the  percentage  of the  total number of switch and
     transfer locomotives and the total number in service in 1984.
3/   Number of  line-haul  locomotives  in  each AQCR derived from the
     percentage  of  the  total  number  of   switch  and  transfer
     locomotives and  the total number of  line-haul locomotives in
     service in 1984.

-------
                             -38-


                           Table 15
         Line-Haul and Switch and Transfer Locomotive
         	Emissions in Five AQCRs	
                                    Emissions (tons/year)
	AQCR	        HC            CO          NOx

Philadelphia:

    Line Haul                   774          1,665       8,610
    Switch                      208            185       1,058

Chicago:

    Line Haul                 6,278         13,499      69,820
    Switch                    1,690          1,502       8,587

St. Louis:

    Line Haul                   986          2,120      10,963
    Switch                      266            237       1,354

Kansas City:

    Line Haul                 1,075          2,312      11,960
    Switch                      290            258       1,473

Los Angeles:

    Line Haul                   951          2,045      10,575
    Switch                      256            227       1,300

-------
                             -39-
     Refrigerated  car  activity  was  approximated  by  weighting
the national-population of  these units by the amount of traffic
in  each   region   (Table   16).     This   method  assumes   that
refrigerated  units  are  equally  distributed   throughout  the
nation based on the amount of freight moved  in  each  area.   This
may underestimate  refrigerated  car operations   in large  market
areas  and rail  terminals  which,  because  of  high  demand  and
longer  car  residence  times,  would  tend  to  agglomerate  larger
concentrations of these units, e.g.,  Philadelphia, Chicago,  St.
Louis,  and  Kansas  City.    The   possibility  of  overestimation
exists  in  smaller  markets  and  producing  areas,  e.g., the  Los
Angeles  region.   However,   the  existence  of  a  major  port
facility in Los Angeles may moderate this effect.

     The  literature  contains inconsistent  population  estimates
of  the nation's  mechanical  refrigerator  cars.  A  statistical
summary  for  the  years  1966-76   published  by  the  AAR  (1977),
reported  9,259  mechanical   and   3,613  nonmechanical  units  in
1976.    This  compares favorably  with rail  statistics  published
by the  Interstate  Commerce  Commission (ICC 1976).   Both sources
are based  on ICC  annual  reports, which  are submitted by each
Class  I  railroad   (railroads are  grouped  into classes,  with
Class  I  railroads  being  the  largest).    In  the  Yearbook  of
Railroad Facts  (AAR 1977),   the  total  for  all refrigerated cars
owned by Class  I  railroads  is given as 74,936.  The difference
between  the   two   AAR  documents   for   refrigerated  cars  is
approximately  62,000  units.  Moreover, the figures  from AAR's
statistical summary  (AAR  1977)  are internally  inconsistent.  In
1967, the total  refrigerated car population was reported  to be
49,399,  but  one year  later  in   1968  the  total population was
reported as 15,638.

     To determine  the  exact number of refrigerated  cars  in the
nation,  the  AAR Car Service Division  was  contacted  (AAR  1978).
According to the CS-8A report from the Train II system,   30,106
diesel-electric refrigerator  cars were in service as of January
1,  1977.   This information  is   reported  directly  from the car
owners  to  the AAR.  Therefore,  this figure represents the 1976
national fleet  of  these  cars owned by Class  I  and  II railroads
(large  railroads)  as well  as other companies  and  shippers and
was  used  by  the   EPA in   the   estimation  of  emissions  from
secondary power sources.

     Table  16  shows the  average number  of refrigeration units
which  reside  in each region  at  any one  time.   As a worst case
condition, it  is  assumed that each of the diesel-electric units
is  in operation 24  hours  a  day.   The same number  of  days per
year  wherein  the units are  actually in  service was  assumed for
refrigerated cars as had been used  for locomotives.

-------
                             -40-
                           Table 16
              Refrigerated  Rail  Cars  in Each AQCR
    AQCR
Percent Traffic
(AQCR to Total)
 Total Diesel-
Electric Units!/
Regional
 Diesel-
Electric Units
Philadelphia
Chicago
St. Louis
Kansas City
Los Angeles
0.57
2.26
1.11
0.68
0.84
30,106
30,106
30,106
30,106
30,106
172
680
334
205
253
I/   AAR (1978).

-------
                             -41-
     Mechanical  refrigeration  units typically  use the  Detroit
Diesel-Allison 2-71  engine.  This  2-stroke, 2-cylinder  diesel
engine is nominally  rated  for  68 horsepower at 2,000  rpm.   The
EPA has  no  exhaust emission data on the Detroit  Diesel-Allison
2-71 engine.   Emission tests on  a  similarly configured  engine
from the  same manufacturer have, however,  been made.   Emission
rates for this engine, the  Detroit Diesel-Allison 6V-71,  were
determined  by   the   Southwest   Research  Institute  (Hare  and
Springer 1973).  Because this engine,  the  6V-71,  is essentially
three  2-71   engines  mounted  on  a   common   crankshaft,   the
mechanical refrigerated car  emission rates (Table  17)  for  each
of  the modes  of  operation in the duty cycles were developed by
dividing the  data on  the  6V-71  engine by  a  factor  of  three.
The daily  duty-cycle  weighted  emissions presented in  Table 18
for refrigerated car  engines  were derived from  this  source.
Combining the refrigerated  car  daily  emissions,   the  number of
refrigerated  cars per  AQCR and the number  of  days per  year in
which  the refrigerated  cars  are  in  service  resulted  in  the
annual emissions  contributions  of the  secondary  power sources.
These values,  expressed  as tons per year  are shown in Table 19
and constitute the  final   component  of  railroad  emissions  in
each AQCR.

-------
                             -42-


                           Table 17


      Refrigerated Rail Car Emissions by Power Settingl/


          Percent of  Rated          Emissions  (q/hour)
          Power   @ Speed           HC       CO      NOx

          95% <§  1200  rpm           14     1200      650

          72% @  1200  rpm           13       30      680

          67% @  800 rpm            10       30      580

          20% @  800 rpm             7       20      200
I/   Extrapolated from  data on  a  Detroit Diesel  6V-71  engine.
     "Exhaust Emissions  from  Uncontrolled Vehicles  and  Related
     Equipment  Using Internal   Combustion  Engines,"  Hare  and
     Springer, Southwest Research Institute,  1973.
                           Table  18


                Refrigerated Rail Car Emissions
        Emissions (pounds/day per refrigerated rail car)
                  HC          CO           NOX
                 0.52        7.62         25.56

-------
                             -43-


                           Table 19


               Railroad Emissions in Five AQCRs


                              	Emissions (tons/year)
	AQCR	        HC            CO          NOx

Philadelphia:

    Line Haul                   774          1,665       8,610
    Switch                      208            185       1,058
    Secondary Power              15            226         758
    Total                       997          2,076      10,426

Chicago:

    Line Haul                 6,278         13,499      69,820
    Switch                    1,690          1,502       8,587
    Secondary Power              61            894       2,998
    Total                     8,029         15,895      81,405

St. Louis:

    Line Haul                   986          2,120      10,963
    Switch                      266            237       1,354
    Secondary Power              30            439       1,473
    Total                     1,282          2,796      13,790

Kansas City:

    Line Haul                 1,075          2,312      11,960
    Switch                      290            258       1,473
    Secondary Power           	18            269         904
    Total                     1,383          2,839      14,337

Los Angeles:

    Line Haul                   951          2,045      10,575
    Switch                      256            227       1,300
    Secondary Power              23            333       1,116
    Total                     1,230          2,605      12,991

-------
                             -44-


        6.0  ENVIRONMENTAL IMPACT OF RAILROAD EMISSIONS

6.1  Non-Exhaust Emissions

6.1.1  Refueling Losses

     Historically,  fuel  spillage  and  leakage  have  apparently
been  widespread  and  chronic  railroad  problems.    While  the
estimates  of  fuel  losses  vary significantly,  the  estimates
always  point  to  substantial  losses.   The  Stanford  Research
Institute  (1977)  estimated  that   about   10  percent  of  the
reported  fuel  consumption  of  line-haul  locomotives  owned  by
Class  I  railroads  (the  largest  railroads)  was  spilled  or
unaccounted  for.    Individual   railroads   have  reported  fuel
spillage  and losses to  be  between  0.1 and 3.0 percent  of  the
total fuel purchased by Class  I railroads  (LMOA  1975).   Based
on  Southwest  Research  Institute  (SwRI)   figures,  up  to  367
million gallons  (9 percent of  consumption)  of diesel fuel  or
about  367   million  dollars   ($l/gal)  have  been  spilled  or
unaccounted for each year.

     To  estimate  only  the  amount   of  fuel  spilled each  year
during refueling,  a conservative 2  percent (based  on  the LMOA
reference) of the  approximately  4 billion  gallons  purchased,  is
assumed.   Based  on these  figures,  a  minimum  of  80  million
gallons  or 80 million  dollars  worth of   fuel  at present  fuel
prices,   is    spilled    during    refueling   operations.    This
represents   about   $3,300  worth  of  fuel  for  each   of  the
approximately 24,000 locomotives presently in service.

     The  causes  of  fuel losses during  refueling  operations
appear  to be  well known.  The  two  primary problems  are:   1)
improperly attended  manually  operated fuel  nozzles,  and 2) poor
maintenance  of  equipment.   Because  of  the high delivery rates
during  the  refueling  operation,  200-300  gallons   per  minute,
even difficulties  of short duration can lead to a  large  volume
spill.  Additional losses occur  when the  locomotive fuel tanks
are  filled to excess or  "topped off."   In many  cases,  excess
fuel  drains  from  the  fill pipe when  the  cap  is  not replaced
properly  or  from the overflow vents when the  train accelerates
or rounds a curve.

     Many  of  the accidental  spills can be eliminated by the use
of  automatic  shut-off  refueling  systems.   Although  some  of
these  systems are  in-use,  many  railroads  have  apparently  not
installed  them  because  of  high  initial   costs  and  increased
maintenance   costs  relative  to   those   incurred  for  manual
systems.   Apparently,  the  railroads  have,   on  a  historical
basis, not found  it  cost effective to control the loss of fuel.

-------
                             -45-
     Even  though  the  magnitude  of  refueling  losses  is  very
large,  a  preliminary   investigation  indicated  this  is  not  a
significant source of air pollution.   The high boiling point  of
diesel fuel  precludes  any large  volume  of  gaseous  hydrocarbons
entering the atmosphere  from the  spilled  fuel.   The  spilled
fuel  eventually,  however,  percolates into  the ground.   Even  if
these  HC  emissions  were  a  significant  contributor  to  air
pollution,  the  railroad companies are now,  or are  expected  in
the future,  to  reduce  refueling  losses.   As part of  the Spill
Prevention Control  and  Countermeasure Plans (SPCCP)  mandated by
the Federal Water Pollution Control Act  (40  CFR  112),  refueling
terminals  are  being  equipped  with  fuel  recovery  systems  to
prevent  these  liquid   contaminants  from   reaching  navigable
waterways.    At  the  same time, these measures  eliminate much of
the potential for fuel  evaporation.   The fuel recovery systems
consist  of concrete and metal drain  pans  under  the  rails  at
refueling  points.   After  collection,  the   fuel  is  sent to  a
waste  water  treatment  facility  or  is  reused in  a variety  of
ways,  including use as a power plant  or locomotive  fuel.   The
economical reuse  of this otherwise wasted resource helps offset
the expenses of the SPCCP system.

     The increasing cost of  diesel fuel  as  well   as  possible
shortages  are expected  to provide the  strongest impetus for the
elimination  of  refueling  and  spillage   losses.   This  growing
economic  incentive   in  conjunction  with the  fact  that  proven
technology  exists  for  reducing  this   waste,   is   making  it
increasingly  cost  effective  to  install  automatic  refueling
devices  and  provide   proper  maintenance.   The   cost-saving
potential  should be  the most effective  means  of  controlling
emissions  from this source.

6.2  Exhaust Emissions

6.2.1  Localized Effects

     Areas  which  would  typically experience  the greatest  air
pollutant   concentrations  and   longest   exposure    times   to
locomotive emissions are switch yards  and those  areas in close
proximity   to  switchyards   with   high   levels  of   railroad
activity.   The  land use adjacent  to  1,107  rail  yards,  or  27
percent  of   all   rail  yards,   is  classified  as  residential
(FRA/ORD-76/304) .

     To  assess  the potential  air quality  impact of  locomotive
operations  on a  localized  basis, the published  literature was
reviewed for information pertaining to  this  situation.   No air
quality  monitoring in  the  vicinity  of  rail  facilities  was
found,   although   analyses   of   the    train  crew's   working
environment  do  exist.   Unfortunately,  the majority of this work
has  been confined  to  determining exposure  levels during worst

-------
                             -46-
case tunnel  operations.   Only one  effort  to model  the  effects
of  locomotive  railyard  operations  on ambient  air  quality  was
found.   A discussion of the relevant  air quality  monitoring  and
modeling studies follows.

6.2.1.1  Air Quality Monitoring

     Two  studies  report  nontunnel  "background"  concentrations
(Thompson  1973  and  Hobbs  et al. ,  1977).    The  type  of  rail
operation   represented  by   Thompson's    low  exposure   level
(background concentration) is undefined in his  report  and since
it  may  have been measured  during line-haul  and  not switchyard
service, it is not discussed further.

     The    analysis    of    localized,    train-generated    air
contaminants by Hobbs  et al. ,  (1977) contains data collected in
the cabs  of switchyard  locomotives.  This  study  is summarized
only to  indicate  whether the potential  problem can be dismissed
or  if additional information is necessary.   This  is, of  course,
predicated   on   the  assumption   that    the  locomotive   cab
environment, during  switchyard operations,  represents  a  worst
case condition  because of  its  proximity  to  the  engine  exhaust
and because  air contaminants  will be diluted to  some  extent as
they  are  blown  across  the  boundary  of   the   rail  facility.
Therefore,  if   a  problem  does not  exist  in  the  cab  of  the
locomotive,  it  is probable  that a problem does not exist in the
vicinity of concentrated rail operations.

     Hobbs et al., (1977) continuously monitored  the air in the
cabs of  three  different  switchyard  locomotives for a total of
approximately 19  hours.   The 5-hour time weighted  averages  for
several  pollutants  are  shown in  Table  20.   If  these data  are
regarded  as  the  maximum  8-hour  concentration  for  CO,  the
maximum  3-hour  concentration  for HC, and  the annual  averages
for particulates and NOx  (i.e., the  worst  possible case), three
pollutants  (CO,  particulate,  and NOx) are  below  the quality of
the  ambient air  standards  and  one (HC)  is  above.    (The  HC
standard  is  not  health-based.   It  was  promulgated  because
hydrocarbons  are precursors  to  photochemical oxidants.)   The
high  concentration  of HC  in the  cab  could  be  the  result of
contributions   from   several   sources within  the  locomotive.
Therefore,   it  may   not  be  representative  of  the  external
environment  (i.e.,  there  is  probably  no  relationship   to  the
NAAQS welfare standard).

     Unfortunately,  Hobbs  et  al.,   (1977)  did  not  include   a
discussion  of  the conditions  under  which the  test  values were
recorded.   This  lack of  information makes  it difficult  to
determine  if the locomotives  being measured  were  in a worst,
typical,  or even a  best case environment.   The  most important
unknown  is  the ambient  air quality  in which the  switch units

-------
                        -47-
                      Table 20
        Summary of Air Contaminant Levels in
the Cab of Long-Hood, Forward Switch Yard Locomotives
NAAQS
Substance
Carbon Monoxide
Particulates
Nitrogen Dioxide
Hydrocarbons
5-Hour Time
Weighted Average
0.26 ppm
10 ug/m3
0 . 03 ppm
3.12 ppm
Health
Effects
9 ppm
75 ug/m3


Welfare
Effects


0.05 ppm
0 . 24 ppm

-------
                             -48-
were operating,   If  the ambient air quality  had been measured,
a  direct analysis  of  the localized  effects  could have  been
performed.

6.2.1.2  Air Quality Modeling

     The  impact  of  diesel  locomotive  exhaust  emissions  from an
Amtrak maintenance-of-way  (MOW) base was  modeled as part  of an
environmental  impact  statement which  evaluated  the facility's
effect  on localized  air  quality.   An MOW base is  a  facility
from  which   the  right-of-way  is  maintained.   It  provides  for
employee  parking  and  support  facilities,  office  space,  shops,
interior  and exterior  security  storage areas,  and  an  external
storage  area for  maintenance  of equipment.   The track  area in
the MOW  is  also  used to store idling  locomotives when  they are
not in-use along the right-of-way.

     The  ambient  air  quality  modeling in the vicinity  of  the
MOW was  performed by DeLuew,  Gather/Parsons  and Associates for
the  Department  of  Transportation  (DOT).   In  this  analysis,
Hanisch   (1978)   specifically  modeled   maximum   1-hour   N02
pollutant concentrations  only.  In the  model,   it  was  assumed
that three,  2-cycle,  roots-blown,  EMD  switch  locomotives (1,500
hp) were  allowed  to  idle  when parked  end-to-end in  the  MOW (a
normal occurrence).

     For  the worst case  meteorological  conditions (presumably
the wind was  blowing  steadily  in the direction  of the  most
sensitive receptor),  hourly concentrations of approximately 200
ug/m3  or 0.10  ppm  were  predicted  at  about  85   meters   (280
feet)  downwind.    There  is,   however, no national  short-term
N02  standard.    EPA  has   considered   the  need   for   such   a
standard  with  consideration  given  to  a maximum hourly  N02
concentration  of  approximately  0.25   ppm.  For  the  worst  case
modeled,  the  hourly  N02  concentration  was  below the  level
which was considered by EPA.

6.2.2  Within AQCRs

     Table  21 shows  the  estimated  yearly railroad emissions,
the  total  anthropogenic  emissions  inventories  for  1983 (most
recent  year  available)  and  the  percentages   of  the  totals
contributed  by   railroads  for  the  five Air  Quality  Control
Regions which were studied.

     The  estimates  of railroad  contributions  to  the  total
anthropogenic  emission  inventory  vary  by  AQCR,  i.e.,  the
estimated HC contributions  varied from a  low  of 0.12 percent in
Los   Angeles  to  a  high  of  1.33   percent  in  Chicago,  CO
contributions  varied between  0.06  percent  in  Los  Angeles and
0.53  percent  in Kansas  City,   and   NOx contributions  varied

-------
                                            Table 21
                      Railroad Emission, Total Anthropogenic Emissions and
                      Percentage Contributions by Railroads  for Five AQCR's
   AQCR

Philadelphia

Chicago

St. Louis

Kansas City

Los Angeles
Emissions (1000 tons/year)
HC
Railroad Total %
1.00
8.03
1.28
1.38
1.23
429
603
251
128
1,048
.85
.27
.05
.21
.14
0.23
1.33
0.50
1.08
0.12
CO
Railroad Total %
2.08
15.90
2.80
2.84
2.61
1,633
2,228
932
532
4,307
.91
.03
.69
.18
.33
0.13
0.07
0.30
0.53
0.06
NOx

Railroad Total %
10.
81.
13.
14.
12.
43
41
79
34
99
292.48
551.07
356.11
162.19
533.98
3.57
14.77
3.87
8.84
2.34
                                                                                                           VD

-------
                             -50-


between  2.34  percent   in  Los  Angeles  and  14.77  percent  in
Chicago.  The  estimates  suggest  that  railroads should be viewed
as  significant sources  of  NOx  emissions   in some  AQCRs,  HC
emissions may  be significant  in some  AQCRs while,  along with
CO, may be of little significance in other AQCRs.

     Inspection of Tables  12  and 13 on the basis  of  locomotive
throttle notch position,  shows that the  idle  mode  tended to be
the  single  largest  contributor  of  railroad  hydrocarbon  and
carbon monoxide emissions.  For  switch  and transfer locomotives
the  idle  mode  was   also  the  largest  contributor   of  NOx
emissions.   For  line-haul  locomotives,  the  idle  mode  was  the
third  largest  contributor  of  NOx  emissions.  After the idle
mode,  notch  8 tended  to  be  the  second major contributor  of
locomotive   emissions.    To    place   these   observations   in
perspective,  the  individual percentile  contributions of these
operational  modes  were  calculated  and  are  presented in Table
22.  The  largest  contribution to NOx emissions came  from notch
8  operations  and represented  approximately  24  percent  of  the
total   railroad   emissions   of   NOx.     Idling   locomotives
contributed  approximately  14   percent   of  total  railroad  NOx
emissions  in  AQCRs.    Idling  locomotives   were  the   largest
operational  mode   contributors  of   hydrocarbon  and    carbon
monoxide  emissions  at  approximately  34  percent and  29   percent
respectively.   Notch-8  operation  was  the  second   largest
contributor  of  carbon  monoxide  and  hydrocarbon   emissions  at
approximately  26 percent and 16 percent,  respectively.

-------
                             -51-


                           Table 22
        Percentage of Total Railroad Emissions in AQCRs
       Contributed by Idle Mode and Notch-8 Operations!/
                              Percent of Total Emissions by Mode
                                   Idle	        Notch-8
                              HC    CO    NOx    HC    CO    NOx

Switch and Transfer
 Locomotives                  12.3   5.0   3.8    1.5   1.5   1.3

Line-Haul Locomotives         22.2  24.4  10.3   14.4  24.8  23.1

Switch and Transfer plus      34.5  29.4  14.1   15.9  26.3  24.4
 Line-Haul Locomotives
I/   Calculated from emission  rates  shown in Tables 12, 13, 14,
     and 19.

-------
                             -52-
         7.0  POTENTIAL EMISSION REDUCTION TECHNIQUES

     Procedures  which   have  a   theoretical  potential*   for
reducing locomotive  gaseous  emissions  fall  into  two  general
categories.    These  categories  are  modifications to  the  duty
cycle and the application  of  emission control  technology.   The
benefits which are projected  for  these categories are discussed
in this chapter.

     Since  there was no data in the  literature  on either direct
measurements  of  particulate  emissions  from  locomotives  or  the
effects  of   particulate   control   strategies   on   locomotive
engines, quantification  of  potential particulate control  was
not  attempted.   Particulate control  strategies presently under
development by manufacturers  of heavy-duty  diesel truck and bus
engines  together  with a   reduction  in  the  sulfur  level  of
locomotive  diesel  fuel may,  to greater  or  lesser  extents,  be
applicable to locomotives.

7.1  Duty Cycle Modifications

7.1.1  Engine Shutdown When Not in Active Service

     Historically, diesel  locomotives have been allowed to idle
when not  in active service.  These  periods  of  standby idle are
often extensive; they may routinely be as high  as eight or more
hours  consecutively,  with  prolonged  periods  (48  hours or more)
not being uncommon under certain conditions.

     The  railroad  industry  has  historically   found   it  more
attractive to  allow  these  diesel  engines to  idle rather than to
shut them down.  An  idling  locomotive engine is relatively warm
and  immediately  available  for service.   When a cold locomotive
engine  is   started,   its   large  mass  (30,000  to  50,000  Ibs)
dictates a  lengthy warm-up period before maximum horsepower can
be  developed  (up to  two hours when  the engine  has  completely
cooled  to  an ambient temperature of  50°  to  60°F).  During most
of  the warm-up period,  the  locomotive  is  unavailable  for  use
and  could  present a blockage on the  section of track where the
warm-up is occurring.
     It  should be  carefully noted  that some  procedures which
     exhibit  a  theoretical  potential  for  reducing  locomotive
     emissions may  not  be practical for  use.   These procedures
     are, however,  included  in the discussion together with the
     factors  influencing  their impracticality.

-------
                             -53-
     In addition  to  the lengthy warm-up period  required,  other
time-consuming  and  potentially  expensive  problems  have  been
associated  with  starting  a  cold   locomotive   diesel  engine
(assuming that the ambient  temperature  is  high enough  to  allow
an engine start to be  achieved).   These problems are associated
with  the  size  of the  engine and,  to  a   certain  extent,  the
design  and  maintenance  of  the engine.   Coolant may  leak past
the cylinder  liner to  cylinder head  seals   into  the  cylinder(s)
and,  if it  is not manually purged, could cause a hydraulic lock
in  the  engine   during   cranking.    This   hydraulic   lock  in
combination with  the  high  inertia of  the  moving parts  of the
engine  and  the generator  rotor  will  result  in severe  engine
damage.  Leaking  of  coolant past  the cylinder liner to cylinder
block seals would contaminate  the  engine oil.   Contamination of
the  engine  oil with a  relatively large  volume of  water  would
cause bearing failures (leakage of a  relatively  small  volume of
water  during  engine  shut-down  can  be tolerated  because the
water  is  evaporated  when  the  oil  is  heated  to  operating
temperature).    During  prolonged   engine  shutdown,  lubrication
oil drains  back  into the oil  pan,  thereby  leaving  many engine
components  unlubricated  at   start-up.    This   temporary  oil
starvation  at start-up causes excessive wear rates during the
period   of    inadequate   lubrication    and,   with   repeated
occurrences,  will result  in  engine  damage.   To prevent   these
problems,  the  historical  practice   has  been  to have trained
maintenance personnel  open cylinder  test  valves  to  drain any
coolant  from  the cylinders,   pre-lubricate,  and manually turn
the engine one or two revolutions before starting is attempted.

     As  a  means  of  reducing fuel   consumption,  manufacturers
have  recently  shown that  many  newer  locomotives  can  be shut
down  with  relative  safety when the ambient  temperature is 50°F
or  above.   Protection  against damage resulting from a hydraulic
lock  is  achieved  by providing a means whereby an  extremely low
cranking speed  is used initially.   If a hydraulic lock does not
occur,  normal  cranking  speed  is then  employed  to  start the
engine.   Should a hydraulic  lock  occur,  it  is  rectified  prior
to  the use  of normal  cranking  speed  by   opening  the cylinder
test valves and draining the water.

     At  ambient  temperatures  below  50°F,  starting   of   these
large  diesel  engines is very  difficult.  At temperatures  below
35°F  starting  is usually  impossible.   This  characteristic  of
poor  startability at low ambient temperatures is shared by all
diesel  engines.   The  problem of  startability  is  addressed  in
car  and truck diesel engines  through the use of starting  aides
such  as glow plugs in  the  combustion  chambers or  intake air
preheaters.   These  technologies  have not  been used  on  diesel
locomotive    engines.     Because    of    combustion    chamber
configuration and   because   of   severe  cylinder  head   design
problems which would be associated with the  application of glow

-------
                             -54-
plugs  to  locomotive  engines   (see   Chapter   8),   intake  air
preheating  appears   to   offer  the   greatest  potential   for
incorporation into new locomotive  engines  and  appears to be the
only approach with  potential as a retrofit cold-start  aid for
in-use  locomotives.   Determination   of   the   practicality  of
applying glow plugs or intake air  preheating to  new locomotives
and  retrofitting  of  intake air  preheaters  to  older locomotives
can  only be  made  in cooperation  with the  manufacturers  and
users of the locomotives.

     Current locomotive designs  preclude any  considerations  of
shut-down at  ambient  temperatures of  32°F and below  because
there  is  no  provision for  preventing  freezing  of  the  coolant.
While the use of antifreeze in the  cooling  systems  would appear
to offer  a  solution for the control of  coolant  freezing,  three
problems are readily  identifiable  which  will  require resolution
prior  to the   successful   use  of   antifreeze  and  consequently
prior  to  low temperature shutdown  becoming practical.   Two  of
these  problems  have  already  been  identified,  i.e.,   engine
startability at temperatures  below 50°F  and  coolant  leakage.
Leakage of even a small  volume of  coolant  containing antifreeze
into the oil can result  in  component  failures.   The  need  to
prevent  contamination of  the  oil   with  coolant is,  therefore,
substantially greater  when antifreeze  is  employed.  The  third
problem  associated  with the  use  of  antifreeze in locomotives
stems  from  the  effects   on  engine   cooling.    Briefly,  the
addition  of  antifreeze to  the  engine cooling  water  decreases
the  rate  of  heat removal  in the  engine and  the  rate  of heat
rejection  at  the  radiator.   If   space  is   available  on  the
locomotive for  a larger radiator,  this segment of  the  problem
can  be resolved,  but at some cost.  It  is  not presently clear,
however,  what  methods would be  most  suitable to  increase the
heat rejection  rate in the  engine.

     If  the  problems  of  engine  startability  at relatively low
ambient  temperatures  and  coolant  leakage  can  be  overcome,
engine  shutdown  at  temperatures   approaching  32°F may become
practical.*  If solutions   to  the  problems associated with the
use  of   antifreeze  can   be  found,   locomotive   shutdown  at
temperatures lower than 32°F may become practical.
     Cyclic    starting,    warming-up   and   shutting-down   of
     locomotives  is  probably  not  practical  as  a  method  of
     reducing  idle times  and resolving  cold  starting problems
     because:   1)  repeated thermal cycling tends to  worsen the
     water  leakage problem,  2)  smoke  and probably hydrocarbon
     emissions   are  high   during  engine   warm-up,    and  3)
     additional  personnel  would  be required  to perform this
     function thereby increasing costs.

-------
                             -55-
     Table  22,  which  was   developed   previously,   shows  the
percentage contribution of  locomotive idling  emissions to  the
total railroad  emissions  for each  type  of locomotive  based on
historical operational  practices.   Estimates  of reductions  in
emissions from  locomotives through  reductions  in the time spent
at idle can  be developed by  deriving modified duty  cycles  and
the  corresponding  emissions.  The  factors which determine  the
amounts  to  which  duty  cycles  can  be  modified  are  engine
startability    versus    ambient    temperature,    operational
considerations  of  the   trains  which   limit  the   amount   of
shut-down which can be practiced and the  effects  of cold-start
emissions following  a shut-down.   Three  scenarios  are readily
identifiable  for  the relationship  between  engine  shut-down  and
restart versus  ambient  temperature.  These  scenarios  are:   1)
that  the  engine  can  be  started under essentially  all ambient
temperatures, 2) that antifreeze continues not to be  used and,
therefore, that the  inclusion  of  starting  aids is  limited to
those which are effective at 32°F and above,  and 3)  that engine
startability  is limited to 50°F and above.

     Under the  first scenario  (startability  is  possible at  all
temperatures),  practical  considerations  of train  operation  are
assumed to  limit shut-down  to  60  percent of  present idle time
and   that  emissions  during  the   cold-starts  are   equal   to
approximately  10  percent of  daily idle emissions.   Under this
scenario,  idle  emissions  would  be  reduced by approximately 50
percent.    The   resulting   reductions   in   overall   railroad
emissions in  AQCRs  would  be on the order  of  17  percent for  HC,
15 percent for  CO and 7 percent for NOx.

     Under  the  second  scenario  (engines  can be  shut-down at
temperatures  above  32°F)  the   theoretical  maximum  amount  of
engine  shut-down can be  estimated  on  a  national  annual basis
from  the  number of days  when the  ambient temperature  does  not
fall  below  32°F.   The  theoretical  maximum   reduction  in  idle
time  with  this  constraint is approximately 40 percent.[18]   If
it  is assumed  that practical  train operating  constraints  and
cold  start emissions reduce this value by  20  to  25 percent,  the
resulting  projections  for reductions  in  railroad emissions are
on  the order  of 10  to  11  percent  for  HC,  between  8  and   9
percent for CO  and between 4 and 4  1/2 percent for NOx.

      Under the  third scenario (shut-down at 50°F and above) the
reductions  in  emissions  are  estimated  to  about   one-half  of
those  estimated for  the  second  scenario,  i.e.,  reductions in
HC, CO, and NOx emissions would be  on the  order  of  5 percent,  4
percent, and  2  percent, respectively.

      Not  only do  idling locomotives contribute to the  pollutant
burden  in a  region  but they consume  fuel while  performing no
useful  work.   While  locomotive  manufacturers and operators are

-------
                             -56-
already  pursuing  engine  shut  down   to   conserve  fuel,  the
increasing  cost  of  fuel  should  favorably  dispose  both  the
manufacturers  and  the  operators  to   address  the  concept  of
maximizing locomotive shutdown when not in active  service.   It
is  presently not clear,  however,  what  level  of effort  may be
employed by  manufacturers  to address  problems associated with
engine shutdown  and  restart.  It  is  equally not clear  whether
railroads  would  utilize  locomotive  shutdown  capabilities  if
provided by manufacturers.

     Table  23  summarizes  the  fuel  consumption rates   for  an
average switch and transfer  locomotive  and  for  an average line-
haul  locomotive  as  functions of throttle notch position and by
duty  cycle   (these  in-use   locomotive weighted   values  were
calculated using the  same methodology  as  was used  previously
for   emissions).    In   the   case   of  switch  and   transfer
locomotives,  approximately 38 percent  of  the  total  daily fuel
usage  by  these  units   is   consumed  in  the  idle   mode.   The
comparable value for  line-haul locomotives  is approximately 12
percent when  operating  in  an AQCR (i.e., on  a 12 hour  per day
basis) .    Estimates  of  the  monetary  value  of  the fuel  which
could be  saved  through locomotive shut-down  are  presented in
Chapter 8.

7.1.2      Limiting Use  of Highest Power Settings  When in Urban
           Areas

     Eliminating the use of  throttle notch  positions  6,  7 and 8
in  metropolitan  areas  may  be a method of providing a measure of
emissions   reductions   from   this  source   (locomotives).  The
practicality   of   this   approach   has,   however,   not  been
established  and  it must be viewed as having a  low  potential for
success.   As a minimum, the following  conditions would have to
be  met  for  this  approach to  be  practical:   1)  that  the
locomotives   assigned  to  each train   under  present  operating
practices  could  perform the  work required  at  reduced power
settings when in urban areas, and 2)  that train schedules would
not be  significantly  affected.   If these conditions  can not be
met,  disruptions in  train  schedules   could  result  and  a net
increase  in  emissions  could  result either  from the  addition of
locomotives  to  each train so as  to maintain  schedules or from
an  increase  in  time  for each trip.  If  the conditions  could be
met,  the  estimates  of  the effects on  locomotive  emissions are
as  follows.   In the  case  of  switch  and  transfer  locomotives,
the reductions  would  be  very small  because  of   the  existing
limited  use  of  high throttle  notch  positions.  For line-haul
locomotives,   the   projected   per    locomotive   effects   of
substituting throttle notch position  5 for  the  higher power
settings   are reductions  of  11  percent,  30 percent   and 22
percent for  HC,  CO and NOx,  respectively.

-------
                             -57-


                           Table 23


     Locomotive Fuel Consumption - Average Values in AQCRs
               Switch and Transfer
                   Locomotive	      Line-Haul Locomotive
Throttle
  Notch         tt/hr      tt/dayl/       tt/hr
  Idle           24        444             38       269
    1            48         81             68        57
    2            85        143            148       107
    3           117        112            263       189
    4           216        104            369       266
    5           285         68            570       205
    6           364         44            662       238
    7           453         54            865       311
    8           545        131          1,040       624
Dynamic Brake    -          -             124        30

 Total                   1,181    '                2,296
I/   24-hour/day duty cycle.
2/   12-hour/day duty cycle in AQCRs.

-------
                             -58-
     Ignoring  any  small  benefit  attributable  to  switch  and
transfer locomotives, this  change  in the  duty cycle,  relative
to the historical cycle, offers  the  potential for about an 8 to
9  percent  reduction  in  HC  emissions,  between  a  26  and  27
percent  reduction  in  CO  emissions  and  about  an  18  percent
reduction in NOx emissions  from railroads in metropolitan areas.

7.1.3  Composite Effect  of  Duty Cycle Modifications

     Adoption  of  both of  the changes  to   the historical  duty
cycles (i.e.,  reductions in  idle time and  substitution of notch
5 for notch 6, 7, and 8  operations) would be  projected to offer
the potential  for reducing railroad HC  emissions by  between 13
and  26  percent  (the  smaller  reduction  corresponds  to  50°F
shutdown capability and the  larger  reduction corresponds  to a
shutdown   capability  at   all   temperatures).    Corresponding
potential  reductions  in  CO emissions range from 30  percent to
42 percent while  reductions  in  NOx emissions may range  from 20
to 25 percent.

7.2  Application of Emission Control Technology

     In  the  area of exhaust emissions characteristics,  large
diesel  engines  used in  locomotives  tend  to  be  similar  to
smaller  diesel  engines  used  in  trucks   and passenger  cars.
Generally,  the control of exhaust  emissions from diesel engines
pose  significant problems  with respect to  smoke  (particulate
material)  and  oxides  of  nitrogen,  while posing  lesser problems
with   respect   to   hydrocarbons.     Because   carbon   monoxide
emissions  are  inherently  quite  low,   control  of  CO  is  not
considered to be  a problem.

     The  literature did  not  contain data on  the  effects of
control  techniques  when   applied   to   full-size    locomotive
engines.    Work  which  had   been  performed   to   assess  the
effectiveness  of  control   technologies  on  locomotive  engines
employed relatively smaller  engines  and the results necessitate
an extrapolation  to the  full-size  locomotive engine (Assessment
of Control Technologies for  Reducing Emissions from  Locomotive
Engines, 1973).   It is  also  important to note that  the engines
tested  were   all  2-stroke  units  and  that  the  data  would
therefore  be most  applicable to  locomotives produced  by  EMD.
Directionally,  the  data should  be useful  with 4-stroke engines
(GE products) but quantification would be greatly in doubt.

     Emission  control techniques of general  interest  to diesel
engines  which   were   considered    in   the   study   were:    1)
modification   of  fuel  injection  timing,  2)  modification of
injector   design,  3)  exhaust  gas  recirculation,  4)  internal
exhaust  gas  recirculation  (reduced  scavenging),  and  5)  water
injection.    Each  of   these  techniques   is  discussed  below.

-------
                             -59-
Throughout these discussions,  the effects will be  expressed in
terms .of  locomotive duty cycles  as  used in this study with  the
exception of smoke which could  not  be quantified  in these terms
because  the  lack  of   an  appropriate  data base.   It  must  be
remembered that  the  duty  cycles  do  not include  any  changes
which may recently have occurred to  reduce engine  idle time.

7.2.1  Modification of Injector Design

     Results from  testing  with  three  types  of injectors  were
reported.   The  types  of   injectors  tested  were:   spherical
valve, needle valve and low-sac.  The  older  (at  the  time  that
testing  was  performed),  spherical   valve   design  was  being
replaced on in-use locomotives  by the needle valve  and  low-sac
designs.  Low-sac injectors are  presently standard  equipment on
EMD locomotives and are provided as  replacement parts  for older
EMD  locomotives.   The  reported  relative  effects  of  the three
types of injectors were:

     1.    The low-sac  injectors reduced  hydrocarbon  emissions
relative  to  the needle  valve type   on  the order  of  50  percent
for the switch and transfer duty cycle  and about  15 percent for
the  line-haul  duty  cycle  (low-sac  injector  technology  was
introduced by locomotive manufacturers  as a method  of reducing
HC emissions).

     2.    Low-sac  type  injectors   resulted   in  NOx  emissions
which were about  15  percent higher  than those from  the needle
valve  type  injector  on the  switch  and  transfer  duty cycle and
about 30 percent higher on the  line-haul duty cycle.    Relative
to   the  spherical   injectors,   the  needle  valve   type  was
essentially equal  on  the  switch and transfer duty  cycle  and
about 20 percent better on the line-haul duty cycle.

     3.    Smoke  opacity  measurements  were   essentially equal
for all  injector types.

     4.    The  low-sac  injectors  gave  benefits  of   20  to  25
percent  in carbon  monoxide  emissions  relative  to  the needle
valve type of injector on both duty cycles.

     5.    It was  not possible  to  identify any effects  on the
efficiency  of  fuel  utilization  associated  with the different
types of  injectors.

7.2.2  Modification of  Injection Timing  (Timing Retard)

     Directionally, the results  of  retarding  injection timing
(within  reasonable limits)  are  to  reduce  oxides  of nitrogen
emissions,  to   increase   smoke,  to  have   little  effect  on
hydrocarbon emissions  and to  tend  to increase carbon monoxide.

-------
                             -60-
The effect on the  efficiency  of  fuel utilization (fuel consumed
per unit  of useful  work performed)  was  not  reported  in  the
data.    Some  penalty  can,   however,  be  anticipated  and  can
reasonably be expected to be on the order of 1-2 percent.

     The  effects  of   injection  timing  retard,  based  on  the
locomotive duty cycles, are  summarized below:

     1.     NOx emissions  could be reduced  by 25 to  30  percent
on  both the  line-haul  locomotive duty  cycle  and  switch  and
transfer locomotive duty cycle.

     2.     Smoke  opacity would  be  increased  by  50  to  100
percent throughout the speed/power range of the locomotive when
needle  type  injectors  were   employed   (old  injector  design).
With recent design  (at the  time  that  the data  were  collected)
low-sac  injectors,*  the  increase  in smoke emissions resulting
from  injection  timing  retard was under  50  percent.   Derating
the maximum power  of  the engines by 10 to  15 percent tended to
alleviate the increased  smoke levels.   It  is   important  to note
that these  data  were  expressed  in terms of  opacity  and not in
terms  of  the  mass   of  particulate   emitted  (mass  emission
measurements were not performed).

     3.     Hydrocarbon  emissions  were  generally unchanged as a
result  of  injection timing  retard.   A  reduction  of  between 1
and 4 percent may, however,  be indicated.

     4.     On  the  switch and  transfer  locomotive duty  cycle,
carbon  monoxide  was almost  unaffected  (needle  valve injectors
were associated  with a  reduction  of a  few percent  and low-sac
injectors (newer  design)  were associated with an  increase of a
few percent).   On  the line-haul  locomotive  duty  cycle,   carbon
monoxide  would  be  projected   to   increase  by  approximately 20
percent as a result of retarded injection timing.

7.2.3  Exhaust Gas Recirculation

     The  referenced  study investigated  both cooled and uncooled
exhaust  gas  recirculation (EGR).  Because  the data  for  cooled
EGR were collected  under laboratory conditions,  the degree of
cooling  provided  was  not  constrained  as  it  would be  on  a
locomotive.   Those results  were not  judged, therefore,  to be
representative  of  the  effects   of  cooled   EGR  in   locomotive
operation and  are not  included  in this  report.  Application of
uncooled EGR resulted  in  the  following effects:
     Low-sac  injectors have  been standardized  for  use  on new
     EMD  locomotives  and  are provided as  replacement  parts for
     older EMD locomotives.

-------
                             -61-
     1.     Engine  operation  was  unstable  with   exhaust  gas
recirculation,  especially  at the  maximum  EGR  rate  which  was
employed,   i.e.,   30  percent  EGR   level.    Engine  power  was
generally reduced as a result of  the use of EGR.  The  amount of
power loss  was,  however,  not  clearly quantified.   The  effects
on emissions given  below are based  on a maximum EGR  rate  of 20
percent  where engine operation was  still relatively  stable.

     2.     NOx emissions   were  decreased  by  approximately  30
percent   on  the  switch  and  transfer   duty  cycle   and   by
approximately 50  percent on the line-haul duty cycle.

     3.     Smoke    opacity   was   increased  by   a    factor   of
approximately two  at low-power  conditions and  increased by  a
factor of six to  seven at high-power conditions.

     4.     Twenty  percent  EGR  rates  had   little  effect  on CO
under  the  switch  and  transfer  duty  cycle  conditions  but
increased CO by  a  factor  of between two  and  one-half and five
on  the  line-haul  duty  cycle,   dependent  upon  the  type  of
injector  being   used.     The   low-sac  injectors  which  were
introduced  to  reduce HC   emissions,  were  associated with  the
greatest increase in CO.

     5.     Reductions in  hydrocarbon emissions  associated with
the use  of  EGR  were  on the order  of 25 to  50  percent  on  the
switch  and  transfer  duty  cycle and on the  order  of 10  to 20
percent on the line-haul duty cycle.

     6.     Determination   of  the   effects  of  EGR   on  the
efficiency  of  fuel utilization could  not be made  because of
insufficient data.   The increases  in  CO  and  smoke  emissions,
are  however,   indicative   of  decreased   efficiency  in   fuel
utilization (a one percent penalty will be assumed).

7.2.4      Reduced  Scavenging  (Increased  Internal  Exhaust  Gas
           Recirculation)

     On   the  2-stroke  engine  tested,   this  approach  was
accomplished by  bleeding a portion  of the  intake air  charge at
the  air  box,  i.e., reducing the amount  of cylinder scavenging.
A  similar  effect  would be  achieved  on  a  4-stroke   engine by
increasing  valve overlap  if very little overlap were presently
used or by  air bleed as was used  on the  2-stroke  engine.  The
reported results were as follows:

     1.     With  an air  bleed  rate  of approximately 35 percent,
a  reduction  in power  output, on the  order  of 10 to  20  percent
was  observed  at  the  high  power  settings.   Engine  roughness was
experienced at the higher  air bleed  rates  investigative, i.e.,
at  air  bleed  rates  greater  than  20  percent.  The  effects on

-------
                             -62-
emissions as  summarized below  are,  therefore,  limited to  air
bleed rates  of 20 percent.

     2.    NOx emissions on the  switch  and transfer  duty  cycle
were reduced approximately 15 percent  and on the line-haul duty
cycle by between 10 and 15  percent.

     3.     Smoke opacity increased  on the  order  of  50  percent
at low power and up to 300  percent at maximum power.

     4.     Hydrocarbon  emissions were  little  changed on  both
duty cycles when the low-sac injectors  were utilized.   Needle
valve injectors  in conjunction  with the  air bleed  resulted in
approximately  a  40  percent  reduction in HC emissions  on  the
switch and  transfer  duty cycle, and approximately  a 20  percent
reduction on the line-haul  duty  cycle.

     5.     Carbon  monoxide emissions  were  reported  to decrease
by  30  percent  on the  switch  and  transfer  duty  cycle and to
increase by 30 percent  on  the line-haul  duty cycle  when  needle
valve  injectors  were  in-use.    The  results  with  the  low-sac
injectors were  a 15 percent  increase  in CO  on the  switch  and
transfer  duty  cycle and  a  factor  of  three  increase  on  the
line-haul duty cycle.

     6.     The effects  on the  efficiency  of  fuel  utilization
could not be  determined from  the data.   The increases in CO and
smoke emissions which were observed  are,  however,  indicative of
reduced  combustion efficiency and probably reduced fuel economy
(a one percent penalty will be assumed).

7.2.5  Water Injection

     Application  of  water injection  to  locomotives  will  pose
such  practical   in-use  problems   as:    1)  freezing  in  cold
weather,  2)  corrosion  in   the  water   injection  system  and
possibly  in  the engine, 3)  storage  capacity on  the  locomotive
for  the  large volume of water  required  (75  to 100  percent of
fuel tank volume),  and  4)  the availability of water of suitable
purity  (contaminants in the water will  result  in   build-up of
deposits  in  the water delivery  system and  in the engine).  The
reported effects on  emissions of water injection rates equal to
75 percent of fuel flow rates were as follows:

     1.     NOx emissions were reduced  between 15 and  20 percent
on  both duty  cycles with both the  low-sac injectors  and  the
needle valve injectors.

     2.     There was little overall effect  on smoke opacity.

-------
                             -63-
     3.     Hydrocarbon emissions  were decreased  (needle  valve
injectors)  by  about  50  percent on the switch  and  transfer duty
cycle and by 20 percent  on the line-haul duty cycle.  With  the
low-sac injectors, hydrocarbon emissions  increased  between five
and 15 percent.

     4.     Carbon monoxide decreased by  approximately  5  percent
on both duty cycles  with the low-sac injectors and  by an equal
amount  on   the line-haul  duty  cycle  when  the   needle  valve
injectors were used.  Carbon  monoxide  increased  by 25  percent
on the  switch   and  transfer duty  cycle when  the needle  valve
injectors were  used.

     5.     The   report  indicated a  slight improvement  in power
and,  as with the  other  emission control techniques, the effect
on the  efficiency of fuel  utilization  could not  be determined
(no fuel economy penalty is assumed).

     In summary,  the results of  the  study show  similar trends
for locomotive  diesel engines  as  are observed with other diesel
engines, i.e.,   control  techniques  which benefit  one pollutant,
for  example,   NOx,   often  result  in  an  increase  in  another
pollutant,   for  example,  smoke  (particulate).    It  is  worth
noting, however,  that progress  is being  made by  other diesel
engine  manufacturers,  e.g.,  those  making  heavy-duty  diesel
truck  engines,  in  controlling  the  trade-off  between  these
pollutants.  It  is  not  unreasonable  to  expect  that  similar
benefits could be realized in the case of locomotive engines.

-------
                             -64-
          8.0  COST AND COST-EFFECTIVENESS ESTIMATES

     Following  the  approach  used  in  the   previous  chapter,
estimates of  costs  (1984  dollars) are divided  into  two general
types of  control procedure.   The  costs are  those  associated
with changes  in operating  practice (duty-cycle  modifications)
and costs associated  with the  application  of emission control
technology.   The cost-effectiveness  estimates are presented  in
the section  which follows  the cost estimates.

     Estimates of  costs  shown  in this  chapter were  developed
from meetings between EPA and  locomotive manufacturers and from
costs associated with emission controls  used  on  automobiles and
trucks.    While   both  manufacturers  of  locomotives  cooperated
with EPA  in developing a  general understanding of  the effects
of adding emission control  systems  to  locomotives,  neither felt
that reliable  estimates  of  costs could be  developed at  this
time.   One basis for  their uncertainty was  the lack of detailed
design information associated  with  the  application  of emission
controls to  locomotives and consequently the unavailability of
a  basis  for  the  development  of  accurate  costs.   The  second
basis for  the uncertainty  is  the significant  difference which
exists between the high volume manufacturing processes used on
automobiles   and  trucks and  the  very  low volume  manufacturing
processes used on  locomotives.   Extrapolations  of the costs  of
locomotive  components  from  costs  associated with  automobiles
and trucks are,  therefore, subject to substantial uncertainty.

     For  the reasons  given  above,   it   is  very  important  to
stress that  the  costs shown in this chapter should be viewed as
first order  approximations which are subject to change.

8.1    Duty Cycle Modifications

8.1.1  Engine Shutdown When Not in Active Service

     Estimates of the costs of changes to locomotives which may
be  required  to  achieve  engine  shutdown  can  be grouped  into
three categories.   In decreasing order  of  magnitude  the costs
would  be associated with:  1)  engine  shutdown capability  at
temperatures below freezing, 2)  engine shutdown capabilities at
temperatures  just   above  freezing   and   3)   engine  shutdown
capabilities at 50°F and higher.

     Problems  requiring   resolution   for   engine  shutdown  at
temperatures below freezing  are:   engine startability, the need
to improve heat dissipation in the engine and  the radiator as a

-------
                             -65-
result of the use of  antifreeze,  prevention of  fuel  "waxing,"*
limiting   cold  - start   emissions,   provision   for   adequate
lubrication at  startup,  provision of  lubricants which  perform
adequately  when  both  hot   and  cold,  elimination  of  coolant
leakage into the crankcase  and  cylinders  if antifreeze  is  used
and provision  for  removal  of any coolant  which may  leak  into
the cylinders where  antifreeze  is used.   With  engine  shutdown
at  temperatures just  above  freezing,  the problems  associated
with the use of antifreeze  are  avoided and the  severity of  the
fuel  waxing  and   lubrication   problems   are  reduced.    With
shutdown limited to  50°F or  higher, the problems  are  reduced to
the control  of  cold  start emissions and coolant  leakage  and to
a lesser extent provision  for adequate lubrication  immediately
following start  up.

     Potential  methods  of   resolving  the  problems  and  the
associated  costs  are  addressed  in  the  following  paragraphs.
The  total  cost for   each  of  the categories  are developed by
summing the costs  of appropriate  remedial  actions.   Credits for
anticipated fuel savings are incorporated for each category.

8.1.1.1  Engine Startability

     Ignition of  the  fuel  in  a  diesel engine  is  achieved by
compressive  heating   of  the  air   in  the  cylinders   to  a
temperature  which  is  sufficiently high  to  cause  the  fuel to
ignite.  Two factors are primary  contributors  to the problem of
starting  a  cold  diesel  engine.   The  first  factor  is  the
relatively low temperature of the  air  after compression because
of  the  low initial temperature  of the air.   The second factor
is  the  cooling of  the  air during   compression by the   cold
surfaces  of  the  cylinder   wall   and   combustion chamber.   In
combination,  these  two  factors  play  the  critical  role  in
establishing the ambient temperature  below which a given diesel
engine will  not start  without the use  of  some form  of starting
aid.
     Fuel  "waxing"  -  diesel  fuel  contains  some  wax  which
     solidifies  at  low temperatures  and  prevents  the fuel from
     flowing.  When the engine is  in  operation,  the quantity of
     fuel delivered to  the engine is greater than that required
     to operate the engine.  The excess fuel  is  returned to the
     fuel   tank.   As   the   fuel   flows   through   the  fuel
     distribution  lines on  the  engine,  it is  warmed  and  the
     warm  excess  fuel  prevents  fuel waxing  in  the  tank  and
     supply  lines.

-------
                             -66-
     The most  readily  identifiable methods  of  improving  the
startability  of  diesel   engines  are;  the   introduction  of
electrical   heating   elements  into  the  combustion  chamber(s)
(glow plugs), heating  of  the  air  before it enters  the engine,
use of  a  readily startable  auxiliary  engine  to rapidly  crank
the main  engine for  a prolonged  period of  time coupled  with
heating of  the intake air  by the exhaust gases  of the auxiliary
engine,  the use of  an  auxiliary starting fuel and the  use  of  a
starting fluid which is introduced into the  intake air.

     Functionally,   glow  plugs   provide  a  hot-spot   in   the
combustion chamber  as  well  as  some heating of  the  metal  which
forms the surface of the combustion chamber and  some  of the air
in the chamber.  Glow  plugs  are the standardized method whereby
starting  is   achieved  in  the   relatively   small,   indirect
injection (IDI)  diesel engines  used  in passenger  cars.   Glow
plugs are seldom if  ever  used  in the  relatively  larger,  direct
injection (DI)  diesel engines  used in  heavy-duty  trucks.   The
reason for this difference  in  the application of glow  plugs is
attributable  primarily to   the  differences  in the  combustion
chambers and  to  a  lesser  extent  to the differences in  engine
size.   In an  IDI engine,  the combustion chamber consists of two
interconnected chambers with the  fuel  being injected  into  only
one of  the  chambers.  As  a  result, heating by  a glow  plug can
be  concentrated  in the  section  of  the  chamber   where  fuel
injection and  consequently  ignition occurs.  In  a  DI  engine,  a
single  chamber  is  employed  and  as a result,   the problem of
providing  sufficient  and   appropriately   located   heating  is
substantially  greater  than  is  the  case   for   an  IDI  engine.
Because of the differences  in  combustion chamber configuration,
significantly  less   electrical  energy is  required  for  the
successful  heating   and  consequently  starting   of  an  IDI  than
would  be required   in  a  DI  engine  of comparable  size.   In
addition,  air which is  warmed by  the glow plug  tends  to be
retained in  the  pre-chamber of  the IDI engine  during  cranking
while  the  air  in  the  cylinders  of the DI engine tends  to be
expelled during  cranking.   Locomotive  engines  are  of the DI
type  with  individual  cylinder volumes  which  are  between   four
and six  times larger  than   those  in heavy-duty  truck  engines.
As a  result of these  factors, it  is  reasonable to expect that
multiple glow plugs would   be  required in  each cylinder  of  a
locomotive engine so as  to  achieve the necessary  heating.  It
is  also reasonable  to  expect that  each  glow plug   would be
significantly  larger than those  used  in  passenger  car  diesel
engines.  A very substantial increase  in the size of locomotive
batteries would result from  the use of glow plugs.

     The tasks  required for the  application  of glow plugs  (the
first option)  to locomotive engines  are the  development  of  a
new cylinder  head  design that  would accommodate the glow  plugs
while  retaining  other necessary  characteristics, provision for

-------
                             -67-
an adequate electrical energy  storage  system and  an  electrical
wiring  and.  control  system  interconnecting  the  glow plugs  and
the batteries.

     Redesign of  the cylinder heads to  accommodate the  use  of
glow plugs  is expected to  be  very difficult.   The reasons  are
that space  allocated  for  the application  of glow plugs  must
come from that presently  allocated to  cylinder head cooling and
that the use  of  four valves per  cylinder  (existing  design
practice  necessary   to  facilitate   the   high   power   output
required)   dictates   glow   plug   placement   in   relatively
unfavorable  operational  positions.   Any reduction  in  either
cooling  capability  or  the  size  or  number  of  valves  will
translate into  an unacceptable  reduction in  the  power  of  the
engine.  Because  of  the severity  of  the  anticipated  problems,
the  cost  to each of the two  locomotive  builders for developing
new  cylinder  heads is  projected  to be  between  $1,000,000  and
$1,500,000.   Recovery  of  these costs  over five years with total
yearly  locomotive  production of  1,200  units  results  in  a  per
locomotive   cost   of   between   $340   and   $500.    Increased
manufacturing costs  per cylinder  head  would  be  in  the  range  of
$30  to  $40.   Locomotive  engines  employ a separate cylinder head
for  each cylinder  and  the majority  of  engines  sold   are  16
cylinder  units   (8,  12,  16  and  20   cylinder  engines  are
available).   On  a  per  locomotive  basis  the  manufacturing costs
for  the redesigned  cylinder  heads are,  therefore,  expected  to
be between  $480  and $640.   On a  per  cylinder head basis,  the
redesign  and manufacturing costs  equal between  $50  and $70.
These  costs are  between  six  percent  and   ten  percent  of  the
present price ($700-$800) of a replacement cylinder head.

     The next cost  component  of  the glow plug assisted starting
approach to  resolving  low temperature  startability  is  the cost
of the glow plugs.   Using  the price of  a replacement  glow plug
for  a passenger car diesel  ($20) as a guide, the cost  of a glow
plug for  a  locomotive  engine is projected to be between  $30 and
$40.   The  higher  cost  relative  to  the   passenger  car  unit
results  from  the  size  increase  (heating  capacity)   and  low
production  volume  of  the  locomotive  units.   If  two  glow plugs
per  cylinder  can  perform the  necessary  function,   the  per
locomotive  cost  would be  between $960 and  $1,280.    The  per
locomotive  cost  could increase to between $1,920 and  $2,560 if
four glow plugs were required per cylinder.

     The  remaining cost  components  of  the  glow  plug  assisted
starting  system  would  come  from the increase in the size of the
battery  and  the starter  and  from  the  electrical wiring  and
controls  which  would  interconnect  the battery  and   the glow
plugs.   A two fold increase in battery  size is  projected so as
to provide  an adequate energy supply for the glow  plugs  as well
as  the  increase  in engine cranking  load which occurs  at  low

-------
                             -68-
temperatures.   The price  of the batteries  which are  presently
used on  locomotives  is between $3,500 and  $5,500.   This  price,
plus an additional $100  for enlargement  of the battery  box  is
used to  represent the  cost for the  increase in battery size.
The price of a starter for a locomotive  engine  is  approximately
$1,000 (two are  used).   This price can  be  expected  to increase
by about 15 percent  because  of the increase  in  power  necessary
to   crank   the  engine   at   low  ambient   temperatures.    The
incremental  cost  increase  for   the  starter   is,   therefore,
estimated   to  be   about   $300.    The   wiring  harness   for
transmitting  the  energy  from  the  batteries to the  glow  plugs,
its  locating  parts and the  system  for automatically controlling
the  supply of  energy to  the glow plugs  could cost between  $300
and  $500.   The  first  cost of  the glow plug  assisted starting
approach is the  sum  of  the  costs of the  component parts  and  is
projected to be in the range of $5,980 to $10,100.

     Over  the lifetime  of  the locomotive  (assumed  to  be  15
years) it  is  reasonable  to expect  that replacement  batteries
will be  purchased twice  and that  25 percent of the glow plugs
will  require  replacement.    The   discounted  cost   of   these
components would be  between  $4,200 and $6,900 including a labor
cost for replacement of the defective components.

     The  second  procedure  which  has  been  identified  as  a
potential starting aid is preheating  of  the intake air.  Design
and  development  costs for  this approach should be  less  than
those  for  a   new cylinder  head.    A  per   locomotive cost  of
between $100  and  $200  is  used  assuming  a level  of  effort which
is about one-third that required for a cylinder head redesign.

     The  first set  of  components of  this  system would  be the
heating components which would consist of a fuel atomizer,  fuel
pump,  combustion   air  blower,   ignition   system,   combustion
chamber and electrical supply,  assuming  that  diesel  is the fuel
used in  the heater.   The  second set of components of the system
would consist  of  the heat delivery  unit which  could  take  more
than one  form.   If  the intake  air preheater was mounted between
the  engine  air  filter  and  the  turbocharger  with  its  hot
combustion gases  being mixed  with the air  entering the engine,
a  relatively  simple  piping  connection would  form  the  heat
delivery  unit.   A  second method  of  heat delivery  would be  to
mount the heating unit on the  intake  manifold downstream of the
turbocharger  and  the intercooler with direct mixing of the hot
combustion  products  and  the  intake   air.    This  method  of
installation  would   require  a  relatively minor  modification  of
the  intake manifold  and  a one-way valve  to  prevent  the loss  of
pressurized   intake  air  during engine   operation.   The  third
method of  supplying  heat  to the intake  air  would be by the use
of  a heat exchanger  located either before  the  turbocharger  or
after the intercooler.  While  the use  of a  heat exchanger would

-------
                             -69-
be the most complex  and costly approach of the three,  it  would
remove,  problems  caused  by   water   condensation,   intake  air
dilution  and  the build up of deposits  in  the  engine  intake
system  associated   with  the  other  approaches.    The   heat
exchanger  would have  to be relatively  large  so as  to minimize
flow restrictions in the engine intake air system.

     The  cost  of  the  heating  components  of the  intake  air
heating system can reasonably  be  expected to be between  $1,000
and  $1,500 per  locomotive.   The  cost  of  the  heat  delivery
components using a  heat  exchanger  is  projected  to be  between
$2,000 and $3,000 per locomotive  and include the heat exchanger
cost, the  incremental cost for the  redesigned intake manifold
and an exhaust pipe  for the heater.   In addition to the cost of
the intake air heating  system, provision would have to be made
for additional  battery capacity so as  to  effectively crank the
engine  at  low  ambient  temperatures.    This  additional  battery
capacity is expected  to necessitate  about a 50 percent increase
in the size of the present batteries.   This incremental cost is
projected  to  be between  $1,800 and  $2,800 including  $50  for a
larger  battery box   and  retaining hardware.   The  incremental
cost  for  the  starter would be the same as  that associated with
the use of glow  plugs and  is  approximately $300.    The  cost of
the  intake air  heating system is,  therefore, projected  to be
between $5,200 and $7,800.

     Lifetime  discounted  costs  for  battery  replacement  and
maintenance of  the  heater  system are  estimated to be between
$3,000 and $4,000.

     The  third  method  which  was  identified  as   a  potential
method  for  achieving  starting  of   a   cold  locomotive  diesel
engine was the use  of  a readily  startable auxiliary engine.*
With  this approach,  achieving the required  temperature  of the
air  in  the   cylinders   at  the  end  of  compression  would  be
obtained  by  two paths.   One   path would  be the heating  of the
air  entering  the  main  locomotive engine  by  the  hot  exhaust
gases  of  the  auxiliary engine.   The  other  path  would be the
heating of the  walls of the combustion chamber by  the  heat of
compression through  prolonged  cranking  of the main  engine prior
 *    It  is  questionable  whether   space   is  available  in  a
     locomotive   for  the   auxiliary  engine   and  associated
     hardware.    The  cost  estimate   for  this   approach  to
     achieving  low temperature startability  is,  however, based
     on  the  assumption that  space can  be made  available.   If
     space  cannot be  made available  for  the  auxiliary engine
     within  existing locomotive overall dimensions,  use  of an
     auxiliary  starting  engine will  be  precluded  because of
     size constraints which are detailed in Section 8.2.5.

-------
                             -70-
to the  introduction  of fuel.  The components  required  for this
approach  are  the  auxiliary engine,  speed  reduction  gearing
between the  auxiliary engine and the main  locomotive  engine,  a
means for coupling and decoupling between the  two  engines  and  a
heat  exchanger  in the  intake system of  the main  engine.   The
cost   of   the  auxiliary   engine,   reduction   gearing   and
engine-to-engine coupling  is projected  to  be  in the  $5,000  to
$7,000 range.  The cost  of  the  heat exchanger and  the changes
to  the  intake manifold  would  be  similar  to those  for  the
preheater approach and would be  between  $2,000 and $3,000  per
locomotive.   On  a per  locomotive basis,  the  development  costs
would   be   similar   to   those   associated  with   intake   air
preheating,   i.e.,  between  $100  and $200.   If the  use  of  an
auxiliary  engine  for  starting was   viewed  as   a   complete
changeover  from the  use of an  electric  starter,  these  costs
could be reduced  by  the cost  of the electric  starter  ($2,000)
plus  a  reduction  in  battery size by a factor  of  two  to three.
The  cost savings  in  batteries   could,  therefore,  be  between
$1,500  and  $3,500.    The  incremental   cost  increase  for  an
auxiliary engine  for  starting of the  main engine  is  developed
by combining the  sum  of the component costs and  savings  and is
projected to be between $1,600  and  $6,700.   Lifetime discounted
cost  for  maintenance  of the auxiliary engine  and  drive systems
should be between $500 to $1,000.

     The  fourth  method  which  was  identified  for  enhancing
startability  at low  temperatures  is  the  use of  an  auxiliary
starting fuel  which  will ignite more  readily  than  the primary
fuel.   The  changes to the  locomotive  which  would  be required
with  this  approach are  the  addition  of  a fuel  tank for  the
auxiliary fuel,  a system for purging the  primary  fuel from the
injection system  prior to  engine shutdown  and an  increase  in
battery  and  starter   size.   The  cost of an auxiliary  fuel tank
and  the  addition  of  a fuel selector valve  to  the  fuel delivery
line  would  not  be  large  if  space  were  available  on  the
locomotive for the fuel tank.  Assuming that  space is  available
for  a  relatively  small auxiliary fuel  tank, the cost should not
exceed  $500.   If  space is not available for the  auxiliary fuel
tank,  this  approach   for  achieving  enhanced engine startability
may  not  be  a viable  option because locomotives are  already at
the  weight  and length limits imposed by track constraints i.e.,
locomotives  cannot   be  enlarged  to  provide  space   for  the
auxiliary  fuel  tank.   The  incremental   cost for  the  larger
batteries and  starter would be equal to that  of the intake air
preheat  system;  i.e., about  $2,100  to  $3,100.  The total cost
of the  auxiliary  fuel approach is projected to be  about $2,600
to $3,600.   Maintenance  and battery replacement costs should be
equal to those for an intake air preheating system,  i.e., $3000
to $4000.   While the cost  of this  approach  is relatively low,
its  effectiveness as  a starting  aid  is also low relative to the
three  procedures  for  which  cost estimates have previously been

-------
                             -71-


developed.  It is unlikely,  therefore,  that this approach would
be   employed  where   reliable   starting   was   required   at
temperatures much below freezing.

     The  final  approach  for  improving  startability  is  that
which  is  presently  employed on some  engines  at  approximately
50°F and  for  all  engines at temperatures just below 50°F.  This
approach  is the manual  introduction of a small amount  of ether
into the  intake  system just prior  to cranking of  the engine.
While  the  cost  of  this  approach  is minimal,  it  cannot  be
considered  as  a  viable  starting aid at  temperatures below 40°F
to 45°F.

     Of the five methods for achieving low  temperature starting
of  locomotive  engines,  two  (glow plugs  and intake  air heating)
presently  appear  to  offer the greatest  potential  for  use.
Because   of  the  anticipated   problems   associated  with  the
application  of glow  plugs  and  because  of  the potential  cold
start  emission control benefits of  intake  air preheating, this
is the procedure which would probably be employed.

     The  estimates of the costs  for  the  types of starting aides
which  have  been considered  are  summarized  in Table 24 for ease
of reference.

8.1.1.2   Use of Antifreeze  and Control of Fuel Waxing

     Once  an  ability   to   start  locomotives  at   low   ambient
temperatures  is  in place the need  arises  for the  prevention of
freezing  of   the  coolant  and  fuel  waxing  during  engine
shutdown.   Freezing  of the  coolant  can be  prevented by the use
of  the appropriate quantity of  antifreeze.  Use of antifreeze
leads,  however,  to a reduction  in  the rate of heat dissipation
at  the radiators  and from the walls of  the combustion chambers
and cylinders  to the coolant.

     Increasing  the  heat   dissipation  at  the radiator  can be
achieved  by   the  use   of   a  larger  radiator.   The  required
increase  in radiator size  is expected to be  on the order of 15
to  20  percent.  The incremental cost  increase  for  the radiator
enlargement  is projected to be  between  $1,000 and $1,500 based
on  existing radiator  prices  of between  $10,000   and $12,000.
This cost increase  is  based on the assumption that there will
be no  change in the  existing fans and  fan  drives for moving air
over   the  radiator.    Obtaining the  necessary  space   in  the
locomotive  for the  larger  radiator  is  expected,   however,  to
pose  a problem.   The  cost  of  modifications  to locomotives to
secure the  necessary space  could  range  from  $1,000  to   $5,000
per  locomotive and would depend upon  the  extent of the  changes
required.

-------
                                       -72-
                                     Table 24
                     Summary of Costs:  Engine Starting Aides
         Starting Aide
Glow plugs
   Cylinder head and glow plugs
   Battery, Starter, Wiring

   Replacement batteries
   and glow plugs

        Total
                                                 Cost per Locomotive ($)
 First
1,780-3,700
4,200-6,400
Maintenance
              4,200-6,900
5,980-10,100  4,200-6,900
Applicability
                               All temperatures
Intake Air Preheat
   Heater
   Battery, Starter

   Replacement batteries
   and heater maintenance

        Total
3,100-4,700
2,100-3,100
              3,000-4,000
5,200-7,800   3,000-4,000
                               All temperatures
Auxiliary Engine
   Engine, Drive, Heat exchanger
   Starter, Battery
   Maintenance
7,100-10,200
(5,500-3,500)
              500-1,000
                               All temperatures
        Total
1,600-6,700   500-1,000
Auxiliary Fuel
   Tank, Plumbing
   Battery, Starter
        Total
500
2,100-3,100   3,000-4,000
2,600-3,600   3,000-4,000
                               Above freezing
Ether
                               Above 40°to 45°F

-------
                             -73-
     It  is  not   fully  apparent  how  heat  dissipation  to  the
coolant could be adequately increased  in  the combustion  chamber
and  cylinder  liner  regions  of the  engine without a major  and
very  costly  engine  redesign.  The   most  direct  method  for
increasing  heat  dissipation   to  the  coolant  would be  by  an
increase in the surface area of the  engine components which  are
in  contact with  the coolant  through  the  addition  of  fins  or
spines.  There  is little  if  any  potential  for  increasing heat
dissipation to  the coolant in the cylinder  head  by  this  method
because spines are  already employed  in the  design.   Increasing
the  surface area  of  the cylinder  liner which is in contact with
the  coolant  through  the  addition of  fins  or  spines does  not
appear to  be possible  without  a  major engine  redesign  because
of  the present  inability to  insert  the liners into  the  engine
if  fins or spines are employed.   Increasing the rate of  flow of
the  coolant  through the  engine may  be a method  for  resolving
the  cooling problem.  There  is,  however,  no data to support the
validity of this approach.

     A  second area  where significant  uncertainty exists  with
respect  to methods  of  resolving  identifiable problems  is that
of   achieving  control  over   coolant   leakage.    Locomotive
manufacturers  indicate  that changes have recently been  made to
the  engines which reduce  the  potential for coolant leakage into
the  crankcase and  into  the  cylinders.   These changes  do not,
however, provide  the degree  of  confidence  for  the  control  of
coolant leakage which would be required when antifreeze is used
in  the coolant.   Because  of the  high  cost of  repairing  an
engine  damaged  by   lubricant  contamination  as  a  result  of
leakage  of   antifreeze   into  the  crankcase  (an  estimate  of
$50,000   to   cover   disassembly  and  cleaning   followed   by
rebuilding  with   new   cylinders,   pistons,   crankshaft,   and
bearings was  provided by  a manufacturer)  and the time  that the
locomotive would  be  out  of  service  for   repairs,  every  effort
can  be expected to be made to prevent  this type of leak.

     If   a   major   redesign   were  necessary,   the   cost  can
reasonably be expected to  be between five  and ten  times that
which  was  estimated for  the  redesign  of  a cylinder head; i.e.,
from a low of $10,000,000 to  a high of  $30,000,000  for  the two
manufacturers.  Allowing for  an  increase in manufacturing cost
associated with the redesign, the cost  increase  per locomotive
can  reasonably be  expected  to  be  in the  range of $3,000  to
$6,000.

     The   first   cost   increase   attributed  to  the   use   of
antifreeze in locomotive engines  is  the  sum of the costs of the
engine  redesign,   radiator  enlargement  and  the  cost  of  the
antifreeze.   These  costs total  to  between  $5,300  and  $12,800
and include  200  gallons of  antifreeze   at $1.50  per  gallon.
Over  the   life   of  the  locomotive,  the  discounted cost  for

-------
                             -74-
antifreeze for  cooling  system makeup and  coolant  changes would
be approximately $2,500.

     Problems of  fuel  waxing may  be controlled  either  by  the
use  of  a fuel  which  will  not  exhibit  a waxing  problem  at
prevailing  ambient  temperatures  or  by  maintaining  the  fuel
during shut-down  at  a temperature sufficiently high  to prevent
the  formation of  wax crystals.   Changing  from  the  grade  of
diesel fuel  which is presently used  in locomotives  to a grade
which would not exhibit  the problem  of  waxing until a  very  low
ambient  temperature  was  reached  is  not  considered  to be  a
totally  viable  solution.   At a minimum,  the reasons  for  the
undesirability  of  this  approach  are   twofold.    First,  fuel
producers  would have  to  incorporate  a   portion  of  the  fuel
presently  produced  for   such uses  as  jet  aircraft   into  the
diesel fuel  for  locomotives.  This  change  in  the  locomotive
fuel blend could  result  in problems  of  adequate  availability of
other  fuel types.   Second,  the  price  of  the  fuel  would  be
higher than  that  of presently  used  locomotive  fuel.   Assuming
that the price penalty would be five percent  (i.e.,  5  cents  per
gallon)  applicable  to   the  fuel  consumed  during the  coldest
periods  of the  year,  i.e., about  one-third of the fuel  used by
the  railroads,  the annual  increase in  fuel  cost would  be about
$67  million  (present annual  consumption  of about four  billion
gallons   of   fuel  at   $1  per   gallon)  or  about  $2,800  per
locomotive per  year  (discounted cost  of  approximately $22,200
over a 15 year life).

     Maintaining   either  all  or  part   of  the  fuel   at  a
temperature  above that  at which waxing  would  occur   could  be
achieved  by:   1)  the addition  of  a  fuel  heater,  2)  insulating
the  fuel  tank to reduce  the rate of cooling  of the fuel and the
amount  of  energy supplied  by  the  fuel  heater,  and  3)  the
provision  of  a fuel drain back system which  would  remove  all
fuel  from unheated  regions  of the  fuel  delivery system (fuel
supply   pump,  filters,  distribution  and  return  lines).   An
electric  fuel  heater powered by  the locomotive battery  may be
practical  if  the  duration of each engine  shut down  was  limited
to  between 8 to  12  hours  and  if  only  a  small  fraction of the
total  fuel volume of   3,000  gallons  was  warmed.   With  this
approach,  rewarming  of   the  majority  of  the  fuel  could  be
achieved  by  piping engine  coolant  through  the  fuel tank once
the  engine had  started  and warm-up was underway.  Provision for
fuel  heating under  longer periods  of  engine shut  down would
require   the  use  of  an   external  burner  system  to  prevent
excessive discharge  of  the  batteries.    The   cost   of  the
electrical fuel heater  approach is projected to be in the range
of   $500  to  $1,000  to  cover   the  heater,   its  controls,
partitioning  of  the  fuel  tank,  fuel  drain back  system  and
provision for heating  the main  fuel tank  by  engine  coolant.
The  cost of  a burner system  would be in  the range of $1,000 to

-------
                             -75-
$2,000 for  the  burner,  its  fuel  and air delivery  systems,  its
controls and the fuel drain back system.

     The costs  associated with the  use of  antifreeze and  the
control of fuel  waxing are summarized in Table 25.

8.1.1.3  Cold Start Emissions

     Three  emission  species can  pose  problems with  respect  to
their  emission  rate  during  diesel  engine warm  up following  a
start  from  low  ambient temperatures.  The emission species  are
smoke    (particulate    material),     carbon    monoxide    and
hydrocarbons.   Primarily,  poor  combustion   in  a  cold  diesel
engine  is  caused  by  low  air  temperatures  at  the  end  of
compression which  results  in the  failure of  some of the fuel to
ignite  and  by   flame  quenching  prior  to  the  completion  of
combustion  for  the remainder of  the fuel.   Because the problem
of elevated  emission  rates following a cold  start  is primarily
the  result  of  low  temperatures  in  the"  combustion  chamber,
pre-heating of  the intake air, the  fuel and  reductions  in  the
time  required  for total  engine  warm-up would  tend  to  reduce
these  emissions.   As  was  stated  previously,  the  use  of  an
intake  air  preheater is  probably the  most viable  approach  for
enhancing   cold  engine   startability.   While   this  approach
appears to  offer the most reliable method for achieving a cold
start,  it   also   offers   a  method  of  reducing   cold  start
emissions.   Operation of  the  intake air preheater  both before
engine cranking is initiated and  for some time after  the engine
has  started  would  tend  to reduce cold start emissions.  Heating
of the fuel  by  the intake air  preheater before  engine starting
and  during  engine  warm up would  also  tend to reduce cold start
emissions.   The cost of  adding  a  fuel  pre-heating  element  to
the  intake  air  preheater  should  not exceed  a couple of hundred
dollars  for the heat exchanger  and the  necessary  piping  and
valves  to divert  the fuel  to the  preheater during  start  up.
Reducing  the   time  for   total   engine  warm-up  is  the  third
temperature   related   method    for   reducing     cold   start
emissions.   This  could  be  achieved by  the  introduction of  a
thermostat*  or  similar method  of isolating  the  coolant  in the
engine  from  the remainder of  the coolant in  the  system.   This
coolant  isolation  would  limit  the  volume of  coolant which has
to be warmed during  engine warm up and would result  in some
reduction  in warmup time.   The per  locomotive  cost  for  the
changes to  the  cooling system necessary to  achieve a reduction
*    One  locomotive  manufacturer  expressed concern with respect
     to  the  effects on  railroad  operations  of  a  thermostat
     failure.   The  concern  expressed was  for the  blockage of
     the  tracks  because the locomotive would become inoperable
     when a thermostat  failed  in the closed position.

-------
                             -76-
                           Table 25
                       Summary of Costs:
           Use of Antifreeze and Fuel Waxing Control
     System
Use of antifreeze
    Radiator
    Space Modifications
    Cylinder Redesign
    Antifreeze

    Total
 Cost Per Locomotive ($)
  First         Maintenance
1,000-1,500
1,000-5,000
3,000-6,000
    300

5,300-12,800
    0
    0
    0
2,500

2500
            Fuel
0
0
0
0
Use of different
  fuel blend
                              22,200
Fuel Heating
    Electrical
    Auxiliary heater
500-1,000
1,000-2,000
   0
   0
0
0

-------
                             -77-
in engine warmup time is expected to be on the order of  $500  to
$1,000.   Assuming  the  use of  intake  air preheating for  engine
startability,   the  incremental  cost  of  these   approaches  to
reducing cold  start  emissions is projected  to be between  $700
and $1,200 per locomotive.

8.1.1.4  Lubrication Changes

     Two  changes  in  engine   lubrication  are  expected  to  be
necessary  as   part  of  any effort  to  increase   the  amount  of
engine  shut  down  which can  be  practiced.    These  changes  are
addressed below.

     An  accelerated  rate of  engine  wear is  a problem  with all
engines during  cranking and immediately after starting because
of poor  lubrication  during this  period.   On engines where the
desired  period  of  useful   operation  is   very   long;   e.g.,
locomotive  engines,   high  wear   rates   associated  with  engine
starting  can   be  a  significant  problem  especially  when  the
number of cold  starts  is to be increased because of an increase
in the number  of engine  shutdowns.   Cold  start wear  rates may
be reduced  by the addition of an electrically driven  auxiliary
oil  pump which  would  deliver  lubricant to  all parts of  the
engine prior  to cranking.  The cost  of  adding this  system (oil
pickup,  pump,  drive  motor, plunging  and check valve)  would  be
on the order of $500 to $1,000 per locomotive.

     The   second  change   in  engine   lubrication  which  can
reasonably be expected  to be required is the  need to use an oil
which  will flow when  cold  and  which  will  also provide  good
lubrication  when  hot.   Oils of this  type  are  used  almost
exclusively  in  automobiles  and  are  known  as  multi-viscosity
oils;  i.e.,   the oil  exhibits  the  characteristics  of  a  low
viscosity  oil  when  cold  and  of   a high  viscosity oil  when hot.
Availability  of  multi-viscosity  oils  for  diesel  locomotive
engines  is  not expected  to be an insurmountable problem.   Some
increase  in  the cost  of multi-viscosity  oils  relative to the
oils which  are  presently used is, however, anticipated.  On the
basis  of a cost differential  of  20 percent  between  the two
types  of oils,*  a  cost  of  $3  per   gallon  for  oils which are
presently  used  and  a  lifetime oil usage  of  65,000 gallons  to
70,000  gallons  for  initial  oilfill, oil  changes  and  make  up
oil, the projected cost  increase  is  between $39,000  and $42,000
over  the 15  year  locomotive  lifetime.   As  a  discounted  cost,
these  values  would  represent  a  cost  of  between $21,000  and
$22,000.
     Cost differential is based upon  the  difference of about 20
     percent  which  currently exists  in  the  automotive market
     for these oils  (at discount outlets).

-------
                             -78-


     A summary  of the  costs of  reducing cold-start  emissions
and in lubrication changes is given in Table 26.

8.1.1.5  Fuel Savings from Engine Shutdown

     Under  the  three  temperature  dependent  engine  shutdown
scenarios,  the  estimated  reductions  in   idle  time   are  60
percent,   40  percent and 20  percent  for   shutdown  capability
under all temperatures,  just  above 32°F and at 50°F  and above,
respectively.   Combining  these   values  with  the  idle  fuel
consumptions  shown  in   Table  23,  weighted  by  the  ratio  of
line-haul  to switch locomotives*  and allowing  for fuel  used
during cold  starts (assumed  to  be  10  percent of fuel  saved by
shutdown)  results in  annual  fuel  savings  per  locomotive  of
approximately 8,000  gallons, 5,300 gallons  and  2,700  gallons
for  each  scenario.   With  a  fuel  cost  of  $1  per gallon,  a
locomotive  life of 15 years  and a discount  rate  of 10  percent,
the lifetime  savings in fuel are approximately $63,500, $42,000
and $21,400, respectively.

8.1.1.6  Composite Costs for Engine Shutdown

     As  was previously   indicated,  three  levels  of temperature
related  costs  can be associated  with engine  shutdown  when the
locomotive  is not in  service.  These  costs  are dependent on the
severity  of the  problems associated with shutdown and restart
and are  based on the ambient  temperature at  which shutdown is
desired.     The  three  temperature  scenarios  are:   1)  shutdown
capability   at    temperatures   below   freezing,    2)   shutdown
capabilities  at   temperatures   just  above  freezing,   and  3)
shutdown capabilities at approximately 50°F and above.

     To  achieve  shutdown  capabilities  at   temperatures  below
freezing,  costs  would   accrue  from  the  provisions  for engine
startability  at  very   low  temperatures,  use  of  antifreeze,
control  of  fuel waxing,  reductions  in cold start emissions and
changes  in  lubrication.   The cumulative cost  for  this  approach
is  expected  to be  between $38,700  and  $53,300  assuming that
intake air  preheating would  be  employed.   If  other approaches
to  achieving low temperature startability  were to be employed,
the  cumulative  cost could  range  between $32,600  and  $58,500.
The  lifetime   savings   resulting  from  reduced  fuel  usage  is
estimated to be about $63,500.

     With   engine  shutdown  constrained  to   temperatures  just
above  freezing,   it is  again  probable  that  the  intake air
 *    The  ratio  is approximately 3.93  line-haul  locomotives per
     switch   locomotive   as   derived  from   the   number  of
     locomotives  shown  in Table  14.

-------
                             -79-
                           Table 26
                       Summary of Costs:
     Reducing Cold Start Emissions and Lubrication Changes
                                      Cost Per Locomotive ($)
                                   First            Maintenance
Reducing cold start emissions
    Fuel Preheater                  200                 0
    Thermostatic Control          500-1,000             0

    Total                         700-1,200             0
Lubrication changes
    Auxiliary Pump                500-1,000             0
    Improved Oils                     0            21,000-22,000

    Total                         500-1,000        21,000-22,000

-------
                             -80-
preheater would be  preferred because of its  ability  to  provide
engine .startability  as  well as  reducing  cold start  emissions,
that changes  in  lubrication would be  used and that  the  use of
antifreeze  would  be  avoided.    The  cumulative  cost  for  this
approach  is projected  to be between $30,900  and  $38,000.   The
lifetime savings  in fuel is projected to be about  $42,000.

     Locomotive manufacturers consider  50°F  as being  the lower
limit of  the  temperature  region  in which new locomotives can be
reliably  started.    There  are,   therefore,  no  new   costs
attributable  to  engine shutdown  when  temperatures are  50°  and
above.  The lifetime savings in  fuel are projected to be about
$21,400.

8.1.2  Restricted Use of High Power Settings in Urban Areas

     The  assumption  underlying  this concept for  the reduction
of  emissions   from  line-haul  locomotives  is  that  the  power
required  to propel  the train up  grades and at  scheduled speeds
when the  train is  outside of urban  areas exceeds  similar power
requirements  when  the  train  is  in urban areas.   It  is  also
assumed  that  a reduction in  the speed of  the  train when  in
urban  areas would  not  be large  enough to  significantly impact
either  the  train  schedules or   the total time   of  locomotive
operation in urban areas.

     If  the  underlying  assumptions  are  correct,  small  cost
increases due to small  schedule  changes and  increases  in train
operator  working hours  can be expected to be  either partially
or wholly recovered through savings  in fuel usage.

     If  the  underlying  assumptions  are   in  error,  the  costs
could  be substantial and  the  net effect on  railroad emissions
could   be  negative;   i.e.,  emissions   could   actually   be
increased.  The  potential for significant  costs  stems from the
reduction  in  the  effectiveness  of  utilization  of  existing
railroad  facilities  (track, locomotives,  and  railcars)  caused
by  an  overall  slowing  of the   system  and  from the  need  to
purchase  additional  equipment  to  restore  the   speed   of  the
system  where  railroad  customers could not  accept  a slowdown.
Development of  even a  rough  estimate of the  cost attributable
to  a  reduction  in  the effectiveness of utilization  of existing
equipment would  require  information on existing  capitalization
and    amortization   schedules.    This   information   is   not
available.  If  it  is  assumed, however, that  a general   slowing
of   the  system  by   five  percent  could   result   from  this
operational change  and that just the  locomotive  fleet size was
increased  by  five  percent  to  compensate,  the  cost would be
approximately $1.5 billion for new  locomotives.  Changes in the
speed  of the  railroad system which were  either less  than or
greater   than  five  percent  would  be  expected   to   result  in
proportionally either less or greater costs.

-------
                             -81-
     The potential for the  loss  of all of the emission benefits
which were  previously estimated  for  this  approach stems  from
the  emission  and power  characteristics  of  the  locomotives.
Locomotive  emission  rates  for  HC  and NOx,   when  expressed  in
terms of  mass per  unit  of time,  are approximately halved  by
changing from throttle notch eight to  notch five.   The  power  of
the  engine  is  also  reduced  by  approximately one-half  as  a
result of a  change  between these  two  notch positions.   If  the
time  that  a  line-haul  locomotive  spends  in  an  urban  region
(leaving, entering,  or  passing through)  were  to  double* as  a
result of the  reduction  in  power  due to operation in notch five
versus  notch  eight,   all  apparent  emission  benefits   from  the
change would be lost.

     Estimates of the costs of  modifying locomotive duty cycles
are summarized in Table 27.  Lifetime  savings  are  projected for
all  shut-down  scenarios  relative  to the  historical duty cycle.
If  railroads  have already  implemented locomotive  shut-down  at
50°F  and above  to  conserve fuel,  these savings  ($21,400)  can
not be  counted as part  of  an emission control  scenario.   Under
this  modified  duty-cycle  condition,   some  cost  ($10,300  to
$17,400;  i.e.,  $21,400  savings  for 50°F shutdown  exceeds  the
$4000  to $11,100 savings  at  32°F)  would  be  associated  with
shut-down at temperatures just above 32°F and an effect ranging
from  a  small saving ($3,400; i.e.,  $24,800 minus  $21,400)  to  a
cost   ($11,200;    i.e.,   $10,200   minus  $21,400)   could   be
attributable to shut-down below 32°F.

8.2  Application  of Emission Control Technology

     Emissions  test  data  on   five control  technologies  were
presented in the  literature.  These  data  were  used in Chapter  7
to  develop  estimates  of  the  effects of  the  technologies  if
applied  to  locomotive  engines.   The   cost   estimates  for  the
application  of   each  of   the  five   control  technologies  are
developed below.
     Doubling  of  the time;  i.e.,  halving  the  speed,  that   a
     train  spends  in traversing a  section of  track as a result
     of  halving   the  power   setting  of  the  locomotives   is
     possible  because  the  power   required  to  move  a  train
     relative  to   speed   is  dictated   primarily   by  rolling
     resistance  and grade factors  which  tend  to be linear.   If
     aerodynamic  drag predominated in trains,  as  it does  in
     automobiles,  the reduction in speed would  be  less than  50
     percent  when the  power was  reduced by  one-half, because
     aerodynamic drag does not decrease linearly with speed.

-------
                                  -82-
                                Table 27
         Summary of Costs:   Duty Cycle Modification Relative to
                         Historical Duty Cycles
                                Cost per Locomotive ($)
                  First
Maintenance
  50°F and
     above
Fuel
Engine
Startability
Temperature

  Under 32°F  12,200-24,800  26,500-28,500   (63,500)I/

  Just above   6,900-12,000  24,000-26,000   (42,000)
        32°F
               (21,400)
Total
          (10,200-24,800)

          (4,000-11,100)


          (21,400)
Limiting
high power
usage.-^
I/   Discounted  at  10  percent  per  year  over  15 year  locomotive
     life.  Values in  (  ) indicate savings.
2/   Assumes  that  no  significant  slowing  of  the  rail  system would
     occur.   If  slowing by  five  percent were  to  occur,  first  cost
     would be  approximately  eguivalent  to a five percent increase in
     the  cost  of every  locomotive in  service,  i.e.,  about  $50,000
     with the potential  for no improvement in emissions.

-------
                             -83-
8.2.1  Modification of Injector Design

     Changes to the  fuel injector can  be expected to  occur  in
the number, size and  location  of the fuel  orifices at the  tip
of the  injector  and  in  changes  to the sac volume.   The  design
and  incremental manufacturing  costs   for  a redesigned  injector
is estimated  to  be no more  than $5  to  $10.   For  a  typical  16
cylinder locomotive  the  cost would,   therefore,  be between  $80
and  $160.   Development  work  leading to  the optimum  injector
design should not  be  very great and  can be expected  to cost  on
the   order   of  $100,000  to   $200,000  for  each  locomotive
manufacturer.   The development costs  for  the  two manufacturers,
when  distributed  over  locomotive  production   for  five  years
results in  a  per  locomotive  cost of  between  $35 and  $70.   The
total cost  per  locomotive for  the  use  of  redesigned injectors
is estimated  to be  between  $115  and  $230.   Relative to  the
price of  a set of  present  design injectors  ($2,500),  the cost
increase represents a change of between  five and nine percent.

8.2.2  Modification of Injection Timing

     The design and  manufacturing  costs  are not expected  to  be
large  for  the modified  hardware necessary for  the application
of a fuel  injection timing  schedule which  is  different  from
that  presently  used.   This  component  of  the  total cost  of
changing fuel injection timing can safely  be  estimated as being
no more than  $100  per locomotive.   The  cost of development work
necessary to  define  the  optimum injection timing  can,  however,
be expected to  be significant  because  of the  tradeoffs which
will have to be made between each emission  specie  (HC,  CO, NOx,
and  particulate  or  smoke)   as  well  as  fuel   economy.   These
development costs  can be expected,  therefore, to be  similar  to
but  somewhat  lower than  those for the development of a cylinder
head  for  use  with glow  plugs.   The cost  per  locomotive   is,
therefore,  estimated  to  be  between $200  and  $300  when   the
development  costs  are  spread  over   five  years  of  production.
The  total  cost per  locomotive for the application of retarded
injection  timing  is projected  to  be in  the  range of $300  to
$400.   Relative to  the   $10,000  to   $12,000  price of  the fuel
injection pumps  and  injectors used on  locomotives,  these costs
represent   a  change  of  between  two  and four percent.    The
lifetime  cost  for the   1-2  percent fuel economy  penalty  is
estimated to be between  $8,000 and $16,000.

8.2.3  Exhaust Gas Recirculation

      Successful  application  of exhaust gas  recirculation (EGR)
to   locomotive   diesel   engines  will  require   resolution  of
problems  associated  with the  design  and  development  of the  EGR
system as well as  problems of  engine durability associated with
the  use of  EGR on  these  engines.

-------
                             -84-
     The readily identifiable problems  or questions which  will
require resolution  prior to the  successful  application  of  EGR
to  locomotives  are:  1)  the choice  between  the  use of  either
cooled  or  uncooled  exhaust  gases  for  recirculation,  2)  the
degree to which  particulate matter  is  removed  from the  gases
which  are   to   be  recirculated,  3)  the  location   for   the
introduction of the recirculated  exhaust  gases into the  engine
intake  system,   and  4)   the  methods   for  either  reducing  or
eliminating   the   deleterious    effects   of   exhaust    gas
recirculation  on   engine   durability.   These   factors   are
discussed briefly  below  with  the  objective  of  identifying  a
representative system design.

     Data in Reference 3  does not show  a  substantial difference
between the  effectivenesses of  cooled  and uncooled EGR  in  the
control of NOx  emissions.  On this  basis,  it  appears  that  the
less  costly,  uncooled   EGR  method   could   be  selected   by
locomotive   manufacturers.     One   locomotive    manufacturer
expressed  the  opinion,   however,  that  cooled  EGR  would  be
required to maximize the  benefits of EGR  while avoiding thermal
problems in  the  engine which  may be  associated with  uncooled
EGR.

     Considerations of engine durability  and particulate matter
removal are  interrelated  and  will be treated  here  as  a  single
entity.  Particulate material contained in the exhaust gases of
diesel engines, if  recirculated,  can cause accelerated wear in
the  engine,   deposit  build  up   in  the  engine  intake  system
downstream of  the  point  of EGR  admission and  contamination of
the  lubricating  oil.   Reductions  in the  amount  of  particulate
materials which  is recirculated  should reduce the  deleterious
effects  of  EGR.   Two   approaches   could  be  considered  for
reducing   the   amount   of   particulate   material   which   is
recirculated.   These  approaches  are  the   application  of  a
cyclone separator where the heavier  and/or larger particles are
separated from the gas  stream or the use of filters which trap
the  particulate  material.   The  advantage  of   the  cyclone
separator  approach is  that  it  is  self  cleaning and  requires
little maintenance.  Its  disadvantage lies  in  its  inability to
remove the smaller  and/or lighter particles.   As a consequence,
the smaller particles will pass through the  separator  and would
be  admitted  into   the  engine.   The  advantage  of  the  filter
approach is  its  ability  to remove much of the  fine  particulate
material.  The disadvantage of the filter approach is the rapid
plugging  of  the  filters  and  the  subsequent  need  for  the
addition of  an automatic process for  cleaning  or  regeneration
of   the   filters.    Because  of   the  established   long  term
reliability  of   operation  of   cyclone   separators  versus  a
filtering approach,  manufacturers can  be  expected to  view  the
cyclone separator  as the preferred first  approach.  (Successful
development  of  filters  for use  on  heavy-duty  diesel  engines

-------
                             -85-
applied  to  road  vehicles  coupled   with  transfer  of   this
technology to  locomotives could result  in the displacement  of
cyclone  separators  as  the preferred  approach.)   The  use  of
filters  may  be  viewed  as a   fall  back  position  if  engine
durability  proved to  be  unacceptable  because  of  inadequate
removal of fine particulate material by  a cyclone separator and
if  the  application of  other  methods  of  increasing  durability
proved  to   be   inadequate.    Methods   for   increasing   engine
durability are  expected to  include improvements  in  lubricant
filtration and  changes  in the metallurgy  and/or  surface finish
of  such  components  as cylinder  liners,  piston rings,  camshaft
and  tappets,  bearings and  journals and  valve stems  and  valve
guides.

     Factors bearing heavily on  the selection of the  point for
admission  of the  EGR  into  the  engine   intake  system  are the
effects  of  deposits on  components of  the  intake  system  (100
percent   particulate   removal   is  not   achieved  by   either
particulate removal system) and  the requirements  for  delivering
the   recirculated  gases  to   the  engine   intake.     If   the
recirculated   exhaust   gases    were   introduced   before   the
turbocharger,  deposit  build-up  would  be  expected  to  occur
throughout the  engine  intake  system  including the  compressor
section  of  the turbocharger,  the  intercooler,  and the intake
air  preheater  (if used  as a  starting  aid).   The  accumulated
deposits  would degrade the  performance  of these  components.
The  volumetric capacity  of  the  turbocharger  would  also  have  to
be  increased to  compensate for  the  temperature  increase and
consequently volumetric  increase of the  gas  (air  plus exhaust)
being  supplied  by  the  turbocharger.    Because  of  the  small
positive pressure differential which exists  between the gas  in
the   exhaust  manifold  and  the   entry  of  the  turbocharger
compressor  section,   pumping  of  the  exhaust  gases  into  the
intake manifold could  probably  be  avoided.  If the recirculated
exhaust  gases were  introduced  downstream  of  the  compressor,
intercooler   and   air  preheater,   the   performance   of  these
components would  not  be  degraded.   Pumping  of  the recirculated
gases  into  the   intake  manifold  would,  however,  be  necessary
because   of   the  high   pressure   in  the   intake   manifold.
Manufacturers  could  reasonably be expected  to  introduce the
recirculated  exhaust   gases   downstream  of   the  compressor,
intercooler  and air preheater because  this  location would  avoid
degredation of the performance of those components.

     The configuration of an EGR system  for  a locomotive engine
can,  therefore,  be expected to  be as follows;  a compressor to
raise the pressure of the gas being recirculated,  a turbine for
driving  the  compressor,  a  cyclone separator,  a  flow control
valve,  a flow control  valve  actuator and  its  control  system,
and  plumbing between  the system components.   The compressor and
turbine  for  pumping the exhaust gases into  the intake manifold

-------
                             -86-
would  be  similar   to  a  turbocharger   and  can,   as  a  first
estimate, be expected to cost  between  15 and 20  percent of that
of the  locomotive  engine turbocharger.   With  a  price  range  of
between   $25,000   and   $30,000  for   a  turbocharger,   these
percentages correspond to a cost  of  between $3,000  and $6,000.
The  per  locomotive cost  of the  components  of the  EGR system,
including redesigned  exhaust  and intake  manifolds  is  expected
to be  in the  $8,000  to $15,000 range.   The  incremental cost  of
the  improved  lubricant  filtration system  and modified (design
and  metallurgy)  cylinder   lines,  rings,  bearings,  and  valve
train  is  estimated  to  be  between  $1,000  and  $1,500  per
locomotive.   The costs  for the development of a marketable EGR
system and the associated  engine  modifications would be higher
than  those  for  a  cylinder  head  modification but   lower  than
those  for  a   major  redesign   to   accommodate  the   use  of
antifreeze.    The  cost  is,  therefore,  estimated  to be  on the
order of  $5,000,000  to $10,000,000  for  the two  manufacturers.
When  spread  over  a  five  year production  period,  these costs
would  represent  a  per  locomotive   cost  of  between   $850  and
$1,700.   Increases  in  the cost  of  maintenance  should  not  be
large and  would  be attributable  primarily  to increases  in the
cost  of  lubricant  filtration.   A  reasonable  cost   for  this
maintenance would  be  between  $500  and  $1,000  spread  over the
lifetime  of  the  locomotive.    The  lifetime  cost   for  the   1
percent  fuel  economy penalty  is estimated  to be  $8000.   The
total projected cost  per  locomotive  associated with exhaust gas
recirculation is,  therefore, between $21,350 and $33,200.

8.2.4      Reduced Scavenging  (Increased  Internal  Exhaust Gas
           Recirculation)

     The purpose of  the emission control  approach addressed in
this sub-section is the dilution within  the  engine cylinders of
the  fresh air charge for  each power  stroke  with  some of the
exhaust  gases from the  previous power  stroke.   Dilution of the
intake  air  with  exhaust gases  is achieved by the retention of
some of the exhaust gases in the  cylinder.   In the EMD 2-stroke
engines,  removal of  the exhaust gases from  the  cylinder can be
viewed as  a two  step procedure.  The first  step  is  the  opening
of  the exhaust valves  which  allows venting  of  the relatively
high pressure gases through the valves.  The  second  step is the
opening  of  the  intake ports which allow the pressurized intake
air  to  enter  the cylinder  and blow  the  remaining  exhaust  gases
out  of  the  cylinder  through  the  open  exhaust   valves.   This
second step is referred to as scavenging.   Reducing the amount
of   scavenging  which  occurs  could  be   accomplished  by  three
methods; i.e., either by earlier  closing  of  the  exhaust  valves,
by  reducing the  pressure of the  intake  air  or by a combination
of the  two  previously identified methods.   In 4-stroke  engines
which  are used  by  GE, removal  of the exhaust  gases  can be
viewed as  a three  stage procedure.   The stages are  the  opening

-------
                             -87-
of the  exhaust  valves which  allows venting  of the  relatively
high  pressure  gases,  displacement  of  gases  in  the  cylinder
through the  exhaust  valves by  the piston  as  it  rises  in  the
cylinder  and  some  scavenging   of  the  combustion  chamber  by
pressurized  intake air  during the  short period  that  both  the
intake and exhaust valves  are partially open.  The period while
both  the  intake   and  exhaust  valves  are  partially  open  is
referred to  as  the valve overlap period.  On a 4-stroke engine
which is turbocharged (intake air is pressurized)  the amount of
exhaust  gas  which  is  retained  can  be  increased  by  either
reducing the pressure of  the  intake  air  or  by  reducing  the
valve  overlap.   In a  nonturbocharged  engine which operates at
the  low  speeds  typical  of  locomotive engines,  an increase in
valve  overlap  achieved by earlier  opening  of  the  intake valve
would be one method of increasing the quantity of  exhaust gases
which are retained.

     In the  experimental work on reduced scavenging reported in
the  literature, the desired  objective  was achieved by bleeding
intake  air  out  of the intake manifold.  While this  approach is
suitable for use  in  an  experimental program, it  is  not judged
to  be  appropriate  for  use  on  production  locomotives.   On
production  locomotives,  increasing the  amount  of exhaust  gas
which  is retained  would  probably be achieved through changes in
the timing of the valves or through changes  in the air delivery
characteristics of the turbocharger.

     Both of  these approaches would require development work to
define  the  changes  required  and  some  small  increase  in  the
manufacturing cost of the redesigned parts  to cover  changes in
production  tooling.    Development  costs per manufacturers of
$250,000 to  $500,000  should  be  sufficient   for these changes.
This cost would translate into a  per  locomotive cost  of between
$80  and  $170.    Adding  a  cost  of  between  $40   and  $50  per
locomotive to cover manufacturing costs  results in a  total  cost
estimate of  between  $120  and $220  per  locomotive.  Maintenance
costs  should not be affected by this approach to  the  control of
exhaust  emissions.  The  lifetime cost  for   the 1  percent  fuel
economy penalty is estimated to be $8000.

8.2.5  Water Injection

     Two procedures can  be considered  as offering  potential for
introducing  water  into the combustion  chambers of a  locomotive
engine.  The procedures  are:  1)  the spraying of the  water  into
the  intake  system of  the engine,  and 2)  the formation  of   a
water/fuel   emulsion  which   is   injected  into  the   combustion
chambers   by   the   fuel    injection   system.    Some   basic
considerations  pertaining  to  the  use   of  water  injection  are
discussed  below prior to the development  of  an estimate of the
cost of water injection  for a locomotive engine.

-------
                             -88-
     Fundamental  to  either  method  for  the  introduction  of
water,  is  the need  to provide  a tank  on the  locomotive  for
storage of the water.  References  3  and 4 showed that the water
injection rate should be  about 75 percent of  the  fuel injection
rate.   With  this  water  injection  rate, a  water storage  tank
volume of  about  2,250 gallons would  be required  to  correspond
to  the  3,000 gallon  fuel  tank  which  is most commonly  used.
Locomotive  fuel  tanks occupy  essentially  all  of   the  space
between the  trucks  (wheel,  motor assemblies) of  the  locomotive
and  extend the full  width of  the locomotive.   Dimensionally,
the  fuel  tanks  are about  9  feet wide,  3 feet deep and  16 feet
long.   The dimensions for  a water  tank which  would hold  the
required volume of water would  be about 9 feet by 3  feet by 12
feet.   Presently,  there  does  not  appear  to  be  even a  small
portion  of  this  space  requirement  available  on  locomotives.
Reducing fuel volume  is  not  considered to be  a  viable  option
for  securing space  for water  storage  because the present volume
of  fuel  will support  full  power  operation  for  no  more  than
about  16  to 20  hours and  railroads  are  already expressing a
desire for additional  fuel  volume.   Increasing  the size  of  the
locomotive to  accommodate  the water  tank  also  does  not  appear
to be a viable alternative.   Width and  height increases  are not
possible because  of clearance requirements between  locomotives
on parallel  tracks  and in  tunnels.   Any increase  in  the  length
of  a locomotive would be limited to the difference  between the
present  length  of  the  locomotive  and  the  maximum   length  as
defined  by the minimum  radius  of  the  turns  in  the  track  on
which  the locomotive  must  operate.    Locomotive  manufacturers
indicate  that  locomotives  are  already  as  long as  is possible
within  the constraints of  the  tracks   on  which  they will  be
operated.   At  this time,  the  addition of  a  tender to  each
locomotive appears to  be  the only option whereby  the necessary
water tank volume could be provided.

     Two  other  factors  which  would  impact  the  use of  water
injection  on locomotive engines  are the weight  increase  of the
locomotive   and   the  availability  of   water   of  appropriate
purity.   Locomotive axle  loading  is  limited to  70,000  pounds
because  of the load carrying  limitations  of the  track,   Loads
imposed  on  the  track by  each  axle  of   a  fully   fueled  and
operational  locomotive are  presently  on the  order  of  65,000
pounds  or  higher.   There  is,  therefore,  very  little  load
carrying  reserve  capacity  available.    If  the  use of  a tender
were adopted,  the  weight   constraints  would  be  eliminated
because  the  weight  of  the  tender  would  be  supported by  a
separate set of wheels.

     The  other  factor which has  been  identified  by  locomotive
manufacturers as  a  source  of concern with respect to the use of
water  injection  is  the  availability  of  pure  water.    Use  of
water  which  contains dissolved  minerals,  so  called hard water,

-------
                             -89-
in the water  injection  system would result in the  formation of
deposits  in the water delivery system, the  combustion  chambers,
the exhaust  system and  on the  turbocharger  of the  locomotive
engine.   Deposit  formation on  the  system components  must  be
avoided because of  resulting  degradation in the performance of
the  locomotive  and  of   the   water  injection  system.    The
formation  of   deposits   can   be  controlled  by   the   use  of
contaminant  free  water.   The cost  of  facilities  to  produce
water of the necessary purity  plus  the distribution and storage
of  the  water  at  each  refueling point  are  estimated  to  be
equivalent to about a one  percent increase in  the  cost of fuel,
i.e., approximately $8,000 during the lifetime  of  a locomotive.

     If  it  is  assumed  that a  water  tender would be  employed
with each  locomotive,  the cost  of  the water  injection system
would consist  of  the tender,  the water  delivery  and  metering
system and a system  to  prevent freezing of the water.   The cost
of the tender  is estimated to  be between  $20,000 and  $40,000
and  would  include  a diesel  fueled  heater system to  prevent
freezing  of  the  water.    Introduction of  the water  into  the
combustion  chambers  may  be  achievable  by  either   of  the
following methods.   The  first method  for  water delivery would
be through  the formation  of  a fuel-water emulsion which would
be  injected into  the  cylinders  by  a  suitably  modified  fuel
injection system.  The hardware  components  of  this system would
be pumps and flow  control  valves for  the  delivery of  the fuel
and  water  to   the  emulsifier,  the  emulsifier,  modified  fuel
injection  pumps and  injectors,  a   modified  fuel  return  line
which carried  unused fuel and water back to the  emulsifier and
a fuel tank  heater  to  control fuel waxing  (waxing  is  presently
controlled by  the  return of warmed fuel to the fuel tank) .  The
cost for  this  approach  can be expected to  be between  $10,000
and $15,000 per locomotive including development costs.

     The second method  for accomplishing  water  injection would
be through  the introduction of  the  water  into  the  intake air
stream.  On  the 4-stroke  engines produced by GE  this could be
accomplished   by   the   continuous   introduction   of   water
immediately   upstream   of   the   intake   valves.    Including
development  costs,  the  cost  of  the  continuous  flow  system is
estimated to  be between $3,000  and  $5,000 per  locomotive.  On
the  2-stroke  engines  produced  by  GM,  timed  water  injection
would  appear to be  necessary  so as  to  avoid puddling of the
water in the  intake manifold.  For  the timed system,  the cost
is  estimated to be  about  $2,000 higher or between  $5,000 and
$7,000 per locomotive.

     Because the cost of introducing water into the  intake air
is  expected  to  be  lower  than  the  cost of  the  water-fuel
emulsion system, it would  appear  that  it  could be  the preferred
system.   There  are,  however,  presently  unanswered  questions

-------
                             -90-
pertaining to the successful control  of  corrosion in the engine
associated with the  introduction of  the  water into  the intake
air  which may  make  this  system  impractical.   The  estimated
total cost per locomotive  of a water injection system  could  be
between a  low  of  $23,000  where the water is  injected  into the
air entering the engine  and  a  high of $55,000 with  an  emulsion
system plus an annual cost for water of  about $1,000.  Lifetime
costs are projected to be between $31,000  and  $63,000.   Because
of  problems   inherent  to  the  use  of  water   injection as  an
emission control  concept for locomotives  (addition of a tender,
freeze protection,  providing and  distributing  water  of adequate
purity and corrosion control)  it  must be viewed as  an  approach
with little practical potential.

     For ease of  reference,  the  estimates of the costs of the
five emission control technologies are summarized in Table 28.

8.3  Cost-Effectiveness

     The financial  efficiency  of an  emission  control  strategy
can be  measured  by developing  the  ratio  of  the  costs incurred
to  the  benefits  realized.   This ratio  is referred to  as the
cost-effectiveness   of  the  control  strategy  and  is  usually
expressed in terms of lifetime costs and  lifetime benefits for
the equipment involved.

     The effects on  emission  rates  and  on fuel  consumption of
the  control  strategies  which  were  analyzed  previously  are
summarized in Table  29.   These values are based  on historical
locomotive duty cycles.   Since there are no data upon  which to
base  an estimate of  the  extent  to which railroads  have taken
advantage  of   the   shutdown   capability  of  recent   design
locomotives,  i.e.,  implemented  locomotive shut-down  at 50°F and
above as  a fuel  conservation  measure,  no attempt  was  made  to
estimate benefits  under revised  duty  cycles.   Lifetime changes
in  the  mass  of emissions  contributed by  an  average locomotive
within  AQCRs  and  the  change   in fuel  consumed  by  an average
locomotive over the  15   year  locomotive  lifetime  are  shown  in
Table 30.  For  each control strategy  and pollutant, the values
shown represent the change relative to historical duty cycles.

     The  costs  which   have  previously  been   developed  are
summarized  in  Table   31.   Fuel  and  maintenance  costs  are
discounted  to   the  year  that   an   appropriately  modified
locomotive  is  placed  in  service.    Combining the costs  and
emission benefits results  in the cost-effectiveness values for
each  control  procedure  as shown in  Table 32.  In  the case of
the engine shutdown  options, total  costs  (savings)  are equally
divided   between   HC,   CO,    and   NOx   in    determining   the
cost-effectiveness values.   In  the  cases where emission control
hardware  is  employed,  costs  are  equally divided  between the

-------
                                        -91-
                                      Table 28
                                  Summary of Costs:
                     Application of Emission Control Technology
                                             Cost Per Locomotive ($)
      Technology
Modified Injectors

Modified Injection Timing

Exhaust Gas Recirculation

Reduced Scavenging

Water Injection
    First
Operating &
Maintenance
   115-230        0

   300-400        0

12,850-24,200   500-1,000

   120-220        0

23,000-55,000   8,00o!x
Fuel!7
Total
                    0           115-230

               8,000-16,000    8,300-16,400

                  8,000       21,350-33,200

                  8,000        8,12.0-8,220

                    0         31,000-63,000
I/   Lifetime cost discounted at annual rate of 10 percent.
2_/   Approximately one percent of the lifetime cost of  fuel  consumed discounted at 10
     percent per year.

-------
                                    Table 29
            Percent Change in Lifetime Emissions and Fuel Consumption
             By Control Procedure Relative to Historical Duty Cycles
Control Procedure
Engine Shut-down

HC

0 below 32°F -17
0 at 33°F and above -10
0 at 50°F and above - 5
Eliminating high power
settings
Injection Timing Retard 0
Modified Injectors
EGR
Reduced Scavenging
Water Injection

Emissions (%
CO
-13
- 8
- 4
2/
+20
change )
NOx
- 7
- 4
- 2
-25
-20 -20 +25
-20 +200 -45
0 +200 -10
+10 -5 -15
I/ N.A. - effect could not be estimated from the data
{expected to be very small).
2/ Significant uncertainties with the practicality of
Fuel
Consumption
Smoke (% Change)
N.A.V - 8
N.A. - 5
N.A. - 3
+50 +1 to +2
0 0
+200 to +700 +1
+200 +1
0 0
this system makes its
use questionable.   Values  were,  therefore, not presented.

-------
                                          -93-
                                        Table 30
                     Lifetime Change in Mass of Emissions in AQCRs
                      and Fuel Consumed For an Average Locomotive
                          Relative to Historical  Duty Cyclesl/
Control Procedure
Engine Shut-Down
"below 32°F
°at 33°F and above
°at 50 °F and above
Injection Timing
Retard
Modified Injectors
EGR
Reduced Scavenging
Water Injection

HC (tons)

-6.4
-3.8
-1.9

0
-7.5
-7.5
0
+3.8
CO (tons)

-9.2
-5.7
-2.8

+ 14.2
-14.2
+71.1
+71.1
-3.6
NOx (tons)

-26.0
-14.9
-7.4

-93.0
+ 93.0
-167.3
-37.2
-55.8
I/
                                                                          Fuel  (1,000  gal)
                                                                               -120.0
                                                                                -75.0
                                                                                -45.0
                                                                                + 30.0

                                                                                    0

                                                                                + 15.0

                                                                                + 15.0

                                                                                    0
Derived  from  emission rates and  fuel  consumption values in Chapter  5.   An average
locomotive  is  defined as  20%  of a  switch and  transfer  locomotive  and  80%  of  a
line-haul  locomotive  (derived  from the  3.93:1  ratio  for  line-haul  to  switch and
transfer locomotives.)

-------
                                          -94-
                                        Table 31
                              Lifetime Costs  for Emission
                           Control Procedures per Locomotive
                           Relative  to Historical  Duty Cycles
Control Procedure   Purchase Price    Maintenance!'
                      ($1000)           ($1000)
                                     Fuel!7
                                     ($1000)
Engine Shutdown
0 32°F and below    12.2 to 24.8     26.5 to 28.5     -63.5
                                      Total
                                     ($1000)
                                                   -24.8 to -10.2
0 at 33°F and
  above
 6.9  to 12.0
24.0 to 26.0
-42.0
-11.1 to -4.0
  at 50°F and
  above
                                  -21.4
                                  -21.4
Injection Timing
Retard

Modified
Injectors

EGR

Reduced
Scavenging

Water Injection
 0.3 to  0.4


 0.1 to  0.2

12.9 to 24.2


 0.1 to  0.2

23.0 to 55.0
                   8.0 to 16.0

0.5


0
to 1.0
0
8.0
0
8.0
8.0
0
                   8.3 to 16.4


                   0.1 to 0.2

                 21.4 to 33.2


                  8.1 to  8.2

                  31.0 to 63.0
I/   Values given  for  maintenance and fuel costs are discounted costs to  the year  that
     the modified  locomotive  is produced.  The  discount rate  used is  10 percent per
     year.  Fuel  cost of  $1  per  gallon is  assumed.   Values proceeded  by a negative
     sign are savings.

-------
                                -95-
                              Table 32

                    Cost Effectiveness of Control
            Strategies Based on Historical Duty Cycles!/
                                  Cost Effectiveness ($/ton)
Control. Procedure
      HC
        CO
      NOx
Engine Shut-Dovn
"below 32°F         (1,290) to (755)   (900)  to (530)
°at 33°F and above  (975) to (700)     (650)  to (480)
°at 50°F and above

Injection Timing
Retard

Modified Injectors

EGR

Reduced Scavenging

Water Injection
    (3,755)
    8 to is!/

1,425 to l,765l/
     (2,550)
     4 to 8!/
(320) to (185)
(250) to (180)
    (965)
                                     85 to 175l/
      	            65 to 9 5l/

      	              220I/

4,305 to 8,750£/  280 to 565&/
I/   Numbers in ( ) indicate savings rather than costs.
     Costs  are  equally  divided  between  the  pollutants  which  are
     reduced by the control procedure.
2/   Injection  timing  retard reduced  NOx while  increasing  smoke and
     CO and with minimal impact on HC.
3/   Modified  injectors  reduced  HC   and CO,  increased NOx  without
     affecting either CO or smoke.
4/   EGR reduced HC and NOx, increased CO and smoke.
5/   Reduced  scavenging  reduced NOx,   increased  smoke and  CO  without
     affecting HC.
6/   Water  injection  reduced CO and  NOx,  increased  HC and  did not
     affect smoke.

-------
                             -96-
pollutants which  would be  reduced through  the application  of
the  control   technology.   Another  method  of  calculating  the
cost-effectiveness  would  be  to  assign  the  total  cost of  the
emission  control   technology  to  the   pollutant  which   the
technology is  primarily designed to control.   For  example,  EGR
would be  applied for  the control  of  NOx  and  the HC  benefit
could be  treated  as  being  free rather  than  distributing  the
costs equally between  HC  and  NOx as was  done in Table  32.   If
this were done,  the  cost-effectiveness  value  for HC  would  be
zero and the value for NOx would be doubled  to  between $130/ton
and  $190/ton.   It  should be noted  that the application of  a
technology  which  reduces  one  pollutant   may result  in   an
increase  in  one or  more of the other pollutants,  e.g.,  NOx  and
HC emissions  were reduced by EGR,  but  smoke and  CO  emissions
were  increased   (see   Table  29).    (Note   that  none  of  the
technologies for which data were available  showed  any benefits
with respect to  the  control of  smoke  (particulate)  emissions
(Table  29)).   The   true costs  and,    therefore,  the  cost-
effectivenesses  of  technologies  which  produce  adverse effects
would  have  to  be  increased  to  cover  the   application   of
additional technologies which would neutralize some or  all  of
the penalties.  For  example,  the cost-effectiveness of control
of HC emissions through  the  use of modified injectors is  8 to
15 dollars per ton (Table 32) but with  an accompanying increase
in NOx  emissions.   If  it  were to be assumed that an increase in
NOx  emissions  could  not  be  accepted  but  that an  increase  in
smoke  emissions  could be  accepted,   then   combining  injection
timing  retard   and   modified  injectors  (Table   29   shows   a
canceling  of   the  effects  of  these  technologies   on  NOx
emissions)  could  result  in  an  acceptable  procedure.    The
cost-effectiveness value  for  the control of HC emissions would
then  be   between   $885/ton   and  $2215/ton  (cost   of   both
technologies  (Table 31)  divided by HC  benefits  from modified
injectors (Table 30)).

     Examples of  the  cost-effectiveness   of  controlling HC,  CO,
and  NOx emissions from other sources are shown in Table 33  for
purposes  of  comparison.   Comparing   the  cost-effectiveness
values  for   locomotives,  by  control procedure, with  those  for
other sources  shows  that  the cost-effectiveness of control  for
locomotives  is   not   excessive.    Combining  emission  control
technology procedures  to  limit or change  the area  of a negative
effect,    as  was  done  in  the   example   above, may  result  in
substantial    change    in   the    cost-effectiveness    values.
Combinations of  emission  control technology and engine shutdown
capability could, however, result  in the reduction in emissions
at a net  savings because  of the reductions in fuel  consumption.

     It is appropriate to note  at this point that  there were no
data on emission control  technologies  for the  control  of smoke
emissions  and  that  cost  and cost-effectiveness  estimates  for

-------
                                     -97-
                                   Table 33

          Cost Effectiveness for Controlling Non-Locomotive Sources
       Control Strategy
LDV, statutory standard
LDV, 1/M
LDT, statutory standard
Industrial Boilers - Coal Fired

Industrial Boilers - Gas & Residual
  Oil Fired

Coke-ovens (80% HC reduction)
HDGE, evap. control (evap, 5.8 to 0.5)
Motorcycle standards
 Cost Effectiveness ($/ton)
   HC        CO        NOx
508l/
943I/
207I/
49Q2/
112I/
616I/
44l/
57l/
                     130 to
                     15002./

                     500 to
                     14002/
I/   "Revised  Gaseous  Emission  Regulations   for  1985  and  Later
     Model Year  Heavy-Duty  Engines,"  U.S.  EPA,  QMS,  ECTD,  July
     1983.
2/   Federal Register, June 19, 1984, 49 FR 25144 and 49 FR 25145.

-------
                             -98-
the control of smoke emissions  are  not  included.   Incorporation
of  a  technology which  would either  negate the  smoke  emission
penalties  caused  by other  technologies or  which would  reduce
smoke   emissions   could    result   in   a   change   in   the
cost-effectiveness values which are shown.

-------
                             -99-
           9.0  EXISTING STATE AND LOCAL REGULATIONS

     This  portion  of  the  report  addresses  "the  status  and
effect  of  current  and  proposed state  and  local  regulations
affecting such emissions," i.e., Section 404(a)(3).

     To  fulfill  the  requirements,   the  study  evaluated  the
status and effects  of  the subject regulations in  the  following
manner.   First,  the  relevant  regulations were   compiled  from
political  jurisdictions   (subdivisions)   across   the   nation.
Second,  the  regulations  were  evaluated  to  determine  their
effects  on:    1)   health   and   welfare,   2)  application   of
operational   and   technical   controls  to  railroad  emissions
sources, and 3) interstate commerce.

     A  review of  the  literature  showed that  although  at  least
one  study (Sturm,  1973)  had  documented  a  number  of  typical
state  and  local regulations,  there  was no  existing  compilation
or  evaluation of  the  effects  of the standards which  would be
useful to this study.

     The  majority  of  information used  in  this  analysis  was
obtained  from a  survey  of the  50 states and 229 local (i.e.,
regional, county,  and  city)  air  pollution control  agencies.  The
address  of  each   office  was   obtained  from  the  Directory of
Governmental  Air  Pollution Control  Agencies  (APCA,  1975).   The
questionnaire  requested  three  types  of  information:   1)  all
current  and  proposed  regulations which  pertain to  or  could be
construed  to  pertain  to  locomotives,   2)   any   problems  with
enforcement which they may have  encountered,  and  3)  opinions as
to whether there was a need for Federal regulations.

     State laws which incorporated specific  language  as to the
level  of control were  obtained  directly  from  the   statutes.  The
remaining  information  came from  the  diesel-electric locomotive
manufacturers,  the Association of American Railroads,   railroad
companies, and  the  Department of Transportation.

9.1  Survey of Existing Regulations

9.1.1  Survey Returns

     Of  the  279   governmental  air  pollution  control   agencies
surveyed,  a  total  of  308 separate  replies were  received.   The
additional  responses  are accounted  for  by the fact  that  some
questionnaires  were  sent  to  regional   authorities  which  had
several  city  or   county  members within  the  region.   In  these
cases,  the region's response was tabulated for  each individual
subdivision.   Responses  were  received from  92 percent of  the
279  requests,  as tabulated in Table 34.

-------
    -100-





   Table 34



Survey Returns
Control
Authority
State
Local
Total
Total
Recruests
50
229
279
Total
Responses
47
207
254
Percent
Return
90
94
92
Jurisdictions
Represented
47
261
308

-------
                             -101-
     The  compilation  of   emission  regulations  by  this  study
should  be viewed  as  being  a  reasonable  documentation of  the
existing approaches  and not  a complete bibliography  in itself.
It represents a best effort  at cataloging responses which were,
at times,  incomplete  and fragmentary.   All  of  the  submitted
material  was reviewed and,   in some  cases,  regulations  which
could  be  construed  to   include  railroad  rolling  stock  were
identified   in   addition   to  those   pointed   out  by   the
respondents.    If   an  incomplete  text   of  the  regulations  was
included,  it is  possible that  some  standards were  mistakenly
included  or   excluded because of  a  lack  of  other  relevant
details.  Also, as  a compromise in the length of this analysis,
the  regulations  were  condensed and this  may  have unavoidably
added some ambiguity.

9.1.2  Types of Regulations

     Two  types  of state  and  local  regulations were found which
pertained  to railroad rolling  stock:  gaseous  and particulate.
The  vast majority  of the regulations  pertained  to particulate
emissions.   The  gaseous  emissions  were  defined  in  terms  of
specific   chemical   pollutants  (HC,   CO,   NOx,  and   SO^)  and
odor.  None  of  these regulations  specifically cited permissible
emissions  rates,  but  instead  made  it  illegal for the  emissions
to create  a  nuisance.  The nuisance regulations  that pertained
to excessive odor were not cataloged.

     The  particulate regulations almost exclusively  pertain to
visible  smoke  emissions,  with  only a very  small number citing
specific emission rates.

     Visible  smoke   standards  define  allowable   emissions  in
terms  of  an  acceptable   percent  opacity or  Ringelmann number.
In many cases,  the  standard  is expressed  by both  measurements.
The  percent  opacity  is   defined   as  that   fraction   of  light
transmitted  from  a  source which  is prevented from reaching the
observer  or  instrument  receiver.   The  Ringelmann  scale  was
developed  by  the U.S.  Bureau  of  Mines  as  a  measurement  for
black  and white  smoke. As originally developed,  Ringelmann ttl
was  to  equal  20   percent  opacity,  Ringelmann  82  equals  40
percent,  #3  equals  60 percent,  #4 equals  80 percent,  and tt5
equals  100 percent.

9.1.3   Typical Regulation

     Regulations  imposed  by  the State of Illinois  are  presented
as an  example  of  typical  regulations.   The  regulation provides
that:

     "Rule 707.   Diesel Engine Emission Standards.

-------
                        -102-
707(a) The visible emission standard in Rule  706  shall  not
apply to diesel engines.

707(b)  With  the  exception  of Rule  707(e) diesel  engines
manufactured before January 1, 1970, shall  not  be operated
in  such a manner as  to  emit smoke which is equal  to or
greater  than  30   percent  opacity  except  for  individual
puffs  of  smoke.    Individual puffs   of   smoke  shall  not
exceed 15 seconds in duration.

707(c)(l)  Diesel  engines  shall  be  operated  only  on  the
specific  fuels as  specified  in the engine manufacturer's
specifications  for  that  specific  engine,  or  on  fuels
exceeding engine manufacturer's specifications.

707(c)(2)  Persons  liable  for  operating  diesel  engined
fleets  wholly within  S.M.S.A.   shall   furnish   to  the
Technical  Secretary of the Illinois Air  Pollution Control
Board  once each year,  proof  that the fuel  purchased  and
used in their  operations conforms to Rule 707(c)(l).

707(d)  All diesel  engines  operated  on public  highways in
Illinois  coming  from  out  of  the  State  shall conform to
Rule 707(b).

707(e)(l)  No  person shall cause  or  allow  the  emission of
smoke  from any diesel  locomotive in the  State  of  Illinois
to exceed 30 percent (30 percent) opacity.

707(e)(2) Rule 707(e)(l) shall not apply to:

(A)    Smoke resulting  from  starting  a  cold locomotive, for
a period of time  not to exceed 30 minutes.

(B)    Smoke  emitted while accelerating under load from  a
throttle  setting  other  than idle  to  a  higher   throttle
setting; for a period  of time not to exceed 40 seconds.

(C)    Smoke  emitted   upon   locomotive  loading  following
idle;  for  a period  of  time not to exceed 2 minutes.

(D)    Smoke     emitted    during    locomotive     testing,
maintenance, adjustment,  rebuilding,  repairing or  breaking
in; for an aggregate of 10 minutes in  any  60-minute period.

(E)    Smoke  emitted by a  locomotive  which  because of  its
age   or  design  makes   replacement  or   retrofit  parts
necessary  to  achieve  smoke  reduction unavailable.  These
locomotives  shall  be  retired  at  the  earliest   possible
time."

-------
                             -103-
     In states where there  are  marked differences  in  altitude,
there   are  sometimes   different   standards   depending   upon
altitude,  e.g., below and above 5,000 feet.

9.1.4  Compilation of State and Local Standards

     The  stated  applicability  of  the  regulations  varied  from
very specific  to very broad.   In  general,  the regulations  were
expressed  in  terms  of  steady-state  emissions  with  specific
exceptions  (Table  35).   The  following  definitions  were  used in
compiling and categorizing the state and local regulations:

     1.    Stated applicability of the regulation to:

     a.    Locomotives:  Usually  specific to  railroad  industry
diesel-electric   locomotives,   but  may  include  steam-powered
locomotives and amusement park operations as well.

     b.    Generic Description:   Includes all  sources within a
general  description  class,  e.g.,  internal combustion  engine,
diesel engine, motor vehicle, mobile  source.   This  category may
include  locomotives,  locomotive  diesel  engines,  refrigerator
cars, and other railroad rolling stock.

     c.    Emissions   into   the   Atmosphere:     This   category
contains   the   most   inclusive   of  the    regulations   that
specifically cite  allowable emission levels.   A  phrase  such as
"maximum  allowable  discharge  into  the  atmosphere  from  any
source"   is   typical   of  the  regulatory  language.    When  a
regulation  was written  to   include  a large variety of  sources
and  only  incidentally mentioned  railroads,  the  regulation was
placed into this category.

     d.    Nuisance:    An   all-inclusive   category  for   any
emission  which  "causes  or  contributes  to   the  condition  of
pollution."   Because of its  vagueness,  this  type of regulation
is seldom enforced.

     2.    Exceptions

     a.    Excursions:   This  is a general category which limits
temporarily  excesses  of  the continuous standard  in duration and
intensity.

     b.    Maintenance:  These  exceptions apply when the source
is being  repaired, adjusted,  or rebuilt.

     c.    After   Idle:     Higher   allowable   emission   levels
following  a prolonged period of   idle may  be  necessary  due to
below normal operating temperatures  or carbon  loading.

-------
                                                                   Table 35
                                                 Categorization of State and Local Regulations
                     Applicability  of Regulations
Exceptions to the Rules
Subjective Evaluation
No
Regulations
Stnte
Level
Local
34%
39%
(16)
(101)
Generic
Locomotive Description
21%(10) 30%(14)
10%(27) 18%(48)
Emissions
Into
Atmosphere
34%(16)
28%(73)
Nuisance
4%(2)
13%(35)
After Cold
Excursions Maintenance Idle Start Other
57%(2) 19%(9) 17%(8) 19%(9) 11%(5)
42%(110) 4%(10) 5%(12) 8%(20) 4%(10)
Enforcement
Problems
Yes
8%(2)
15% (19)
NO
92%(23)
85%(105)
Need for
Federal
Regulation
Yes
74% (26)
84%(114)
No
26%(9)
16%(22
Level
                                                                                                                                                       I
                                                                                                                                                       1—'
                                                                                                                                                       o

                                                                                                                                                       I

-------
                             -105-
     d.    Cold-Start  Smoke:   Refers  to the  blue-white  smoke
resulting from cold combustion chambers.

     e.    Other:   This  category  contains exceptions which  are
not covered by the other classifications.

     From  the  list  of  definitions  it  should be  clear  that
emission standards  for categories other  than  "Locomotive"  are
included if:   1)  they may be  construed  to  include locomotives,
or  2)  they may limit  emissions  from  secondary rolling  stock.
For  example,  when  the  generic  classification   referred  to
diesel-powered  vehicles  or  motor vehicles,  the  standard  for
emissions  into  the  atmosphere  was   included  to  account  for
secondary rolling stock.

     Survey answers with regard to  current  enforcement  problems
and  the need  for  Federal preemptive  standards  are  cataloged
under the headings entitled  "Enforcement Problems" and  "Federal
Regulations," respectively (Table 35).

9.1.5  General Results

     The results  of  the survey are summarized  in  Table 35.  The
percentages are based  on the total number of responses  to each
of  the three  questions.   The  number  of cases upon which  the
percentage  was calculated  is  included  to  gualify the figure.
The percentiles listed under  the  "Applicability of Regulations"
and "Exceptions to the Rules"  categories  do not total  100 since
some   political   jurisdictions have   more   than  one  relevant
standard.

     The  basic  state  laws   are  clean  air  acts,  enacted  in
response  to  the  Federal  Clean  Air  Act of  1967.  Almost  all
states  have   sections   in  their  laws  which  either  directly
mention   locomotives   or   could  be   construed   to   include
locomotives and secondary  railroad  rolling  stock.   Three  states
have  a  Ringelmann  or  opacity  standard  in  the  law:    Maine,
Kentucky,  and  California.   Most  states,  however,  authorize the
state   air  pollution   control  agency  to   develop   suitable
standards.

     Of  the   state   air  pollution   control   authorities  that
responded,  66  percent or  31  had a steady-state  standard which
applied  to railroad rolling  stock.   No  states had a  nuisance
standard only.

     For local air  pollution control  authorities,  60 percent or
156 had  a  suitable steady-state  standard.  Eight  percent  or 21
had  a  nuisance  standard  only;  therefore,  52  percent  or  135
localities  had  what  might  be  termed   a  readily  enforceable
regulation.

-------
                             -106-
     There is  no  interstate or  intrastate  consensus as  to  the
most  appropriate  level  of control  among  the  air  pollution
authorities  that  do   have regulations.    Minnesota  controls
locomotives to  10  percent  opacity, while Iowa  controls  them to
40  percent.    California's  standard  is  Rtt2   or   40   percent
opacity, but  in  San  Francisco  it is Rttl or  20  percent  opacity.
The exceptions to the regulations are also inconsistent.

     There  is no  consistency  with regard  to  the  authorized
regulatory  agency.    Some  states,  Maryland   and  Texas  for
example, preempt local  control,  although localities  can  and do
adopt  and  enforce the  state  regulations.   Some states  (e.g.,
Michigan,  Kansas  and  Ohio),  have no   state  regulations,  but
individual localities do.   In  most states,  there are both state
and local regulations.

     Several of the  respondents  exhibited confusion with regard
to  the  authority for regulating this  source.    Some  county  air
pollution control agencies  in  California claimed that the state
had  preempted  local  railroad  regulations,  while  others  had
regulations which  they  were  enforcing.  (There  was also  some
confusion  as  to whether  locomotives  were mobile  or stationary
sources.)   A   county in  Maryland  indicated they  thought  the
Clean   Air  Act  had  already   preempted  other  standards  and
assigned   regulatory    authority   to   the    Department   of
Transportation or EPA.

     Preemption was  not the only  reason why some air pollution
control  agencies  did not  have regulations.   These  reasons  are
listed  below:

     1.    Railroad  operations  do  not   draw citizen complaints
or  the  number of complaints is inadequate to justify regulatory
action;

     2.    Only stationary  sources  are controlled;

     3.    The  appropriate state enabling legislation  does  not
exist;  and

     4.    There are no railroads  (Hawaii).

     Minnesota has the  only regulation  that  apparently excludes
a  portion  of  the  railroad rolling stock.   All  2-cycle  engines
are  specifically   exempted   from  meeting    the  regulatory
requirements.    The  majority of  diesel-electric  locomotives  and
perhaps all of the diesel-powered  secondary rolling stock  are,
however, 2-cycle engines.

-------
                             -107-
9.1.6  Subjective Questions on Enforcement

     Of the respondents  that  had applicable  regulations,  there
were varying degrees of  enforcement  ranging from essentially no
enforcement to  strict  enforcement.   The  reasons  given for  low
levels of enforcement were:

     1.    No citizen complaints;

     2.    Other sources are more important;

     3.    A  general  lack   of   time   and  money,  although  a
definite need for control may have existed; and

     4.    Enforcement only on  citizen complaints, but no real
active program.

     There was  no  correlation  between the specificity  of  the
regulation   (i.e.,   a   locomotive   standard   being   the  most
specific), and  the degree to  which  it was  enforced,  excluding
those  classified as nuisance.   California  and  its  localities
actively  enforce their "emission into  the atmosphere" standard
against railroad rolling stock.

     It was extremely difficult  to draw any conclusion from the
survey  with  respect to  enforcement  problems.  Many  states  and
localities  indicate they  did not consider railroads  to be  a
major  problem,   did not  try to enforce  any  regulations they
might  have,  and obviously  did  not  think  enforcement   was  a
problem.   Other localities  did enforce  their  regulations,  but
found  a  great  deal of  cooperation from the railroads,  and also
did   not  consider  enforcement  a  problem.    The   standard
enforcement  procedure  was for  the  agency to  send a  letter  to
the  railroad  stating  that  Unit XXX  was  seen  (means  visible
smoke) at  a certain time in a certain  place in violation   of the
regulations,  and   requested   that   the  railroad  rectify  the
problem.   The  railroad,  in  a return  letter,  would  report  the
steps  it  had  taken  to   end  the  emissions.   Invariably,  the
problem,  as  reported  by  the railroads,   was  a  malfunctioning
part.   Most  agencies  did  not  have  a  way  to  confirm  the
railroad's report and accepted the report at face value.

     Eight  percent  or  two  of  the   25   states  definitely  had
trouble   enforcing  their   regulations:    Maine   and  Oregon.
Fifteen  percent  or   19   out  of  124  localities  experienced
difficulties.    These   results   must  be  viewed   with  some
reservations,  however,  since  all of  the  respondents did  not
address  the  question and  railroad emissions  were not generally
viewed  as  being   a  problem.   Therefore,  the  only  definite
conclusion that can be drawn is that  at  least  15 percent  of the
localities responding experienced enforcement problems.

-------
                             -108-
     Those who  reported  enforcement problems  listed  three  main
types with two  concerning the mobility  of the  locomotives.  The
first  problem  is  that  it  is   difficult and/or  dangerous  to
follow  the  locomotives  for  long   enough periods  of  time  to
observe violations.  This is particularly true where the  train
runs at relatively high  speeds  through  a countryside where  the
highway does not parallel the tracks  long enough for  a  car  with
a smoke observer to follow it.

     The  second problem is  the tendency of  the  railroads  to
move  their  old,  less well  maintained  engines from areas  of
strict   standards   and   enforcement   to   areas   of   loose
enforcement.    Comments  to  this  effect  came  from  both  areas
which  were  receiving  the  older  engines,   and  those  which
enforced  strict  standards  and knew that  the  engines  were  being
sent somewhere else.   While many areas had solved  their  own air
pollution  problems,   they  knew  it  was  a short-term  solution
achieved at the expense of someone else.

     The  last type of problem cited  was a lack of  cooperation
from the railroads.

9.1.7  Subjective Questions on the Need for Federal Regulations

     Of the  35  state  agencies  responding, 74 percent supported
Federal regulation.  Of the 136 local agencies,  84 percent  felt
preemption was  desirable.  However, these results  are qualified
by  the fact  that about 20  percent of  the  state  and  local
agencies  answering "yes," wanted  a Federal standard only if the
EPA  found it  necessary.   The  responses did  not  distinguish
between  Federal  regulation  of  new  locomotives  and  Federal
regulation of in-use locomotives.

     Four  percent  and  17  percent  of  the  state  and  local
agencies,  respectively  preferred  to  retain  local  enforcement
power  under Federal  standards for  more  effective  and efficient
control.

     Those  that  favored a  preemptive   standard  did so mainly
because   of  the   interstate  nature   of  railroad  operations,
stressing the difficulty of  enforcing a  regulation on a vehicle
which  may only  be within one's  jurisdiction on a temporary or
occasional  basis.    Federal  regulations  could  remove   this
difficulty as well as  preventing the  transferral of  locomotives
from   jurisdiction  to  jurisdiction  which some   railroads  now
practice.

     Commenters also  suggested  that a preemptive standard would
resolve  questions concerning the   legality  of state and  local
regulations.    Four    responses   included   information  which
explains  this jurisdictional problem.

-------
                             -109-
     1.     A railroad company suggested that  local  standards do
not  apply  to  their  operations  since  they  are  engaged  in
interstate  commerce,  and  only  the   Federal   government  has
regulatory authority;

     2.     The  court upheld the  local  regulation if  the  train
originated   and   terminated   within    local    jurisdictional
boundaries, but not if it was interjurisdictional in nature;

     3.     The  court   upheld   local   authority  to   regulate
railroads  for  visible  smoke regardless of  the  train's  origin;
and

     4.     Legal  precedent  exists  for  locally  regulating  a
source engaged in interstate commerce based on  the  U.S.  Supreme
Court case of  Huron  Portland Cement Company v.  City of Detroit,
362 U.S.  440  (1960),  in which the enforcement of a  local  smoke
standard against a vessel was contested.

     Those who favored  local  or state standards  generally had
effective   enforcement,   and   saw   no   need   for   Federal
intervention.  They  also  pointed to  local  problems,  such as
large   switchyards,   which   they  thought   could   be   better
controlled by  local government.

9.2  Effects of Existing Regulations

9.2.1  Health  and Welfare

     This  evaluation focuses on  visible  emissions  (smoke)  from
railroad  sources  because existing  state  and  local  regulations
do not focus effectively on gaseous emissions.

     The  direct impact  on human  health  of  existing  state and
local regulations  is impossible to assess at this  time because
of  lack   of   data   on  the  effectiveness  of  state   and   local
regulations  and  because  of  uncertainties   pertaining  to  the
linkage  between smoke  emissions  and human health  and welfare.
The   railroad   industry  practice   of   selectively   avoiding
violations  for  excessive  visible  emissions   by  transferring
"dirty"  locomotives  into areas  where  it is  reasonably certain
no  punitive   action  will  be  taken  is  understandable,  but
undesirable.    This  practice  of  concentrating  high-emitting
rolling  stock  in  specific  areas  may  result  in  concentrated
areas of emissions,  but  insufficient  data exist to  assess the
effect on  health and welfare of those areas.

9.2.2  Operational and  Technical  Controls

     Although   many   factors   affect   visible   emissions   from
railroad   rolling  stock,  including the  quality of  fuel  and

-------
                             -110-
operational practices,  the two basic determinants  which control
smoke are engine design and maintenance.

     Diesel engines have  inherent design  features  which prevent
a smoke  free  exhaust under  all  operating  modes.   However,  by
optimizing the combustion chamber geometry  and  the fuel and air
delivery systems,  smoke  can be  reduced.   All of  the  currently
manufactured locomotive engines  have  incorporated  modifications
to control  much of the smoke  which still  plagues many of  the
older  engines.   For  this  reason,  excessive visible  emissions
from well-maintained engines are, to a large degree,  associated
with  older  locomotive  units.    The  problem  persists  either
because  low-smoke   replacement   parts  are  unavailable  or,  if
these  parts  are  manufactured,  they  have  not  been  installed
because of economic considerations.

     Inadequate maintenance is by  far the  greatest contributor
to excessive  visible exhaust emissions from this  source.   Poor
maintenance practices  are a nationwide  problem which  persists
because  the railroads  find it economically attractive  to  defer
maintenance.   It  is,  however,  not  limited   to  financially
distressed  companies.    The  problem  is  widespread  since  it
affects   engines   regardless   of   their  age    or    design
sophistication.  It is  characterized by  maladjustment  (e.g.,
fuel  injection  timing)  and malfunctioning  hardware (e.g.,  bad
fuel injectors or dirty air filters).

     There  are  exceptions,  however,   even   for  well-maintained
units.   Locomotives, in general, have visible  emissions during
or  following  a  period  of  idle  because   the combustion chamber
wall  temperatures  decrease  to  the  point   that   the  flame  is
quenched as it  nears the  walls.   When this happens, the fuel in
this  region is  not burned and these liquid hydrocarbons escape
through  the exhaust  system  appearing  as  a white  smoke.   The
other problem,  associated with line-haul  locomotives,  is termed
"turbocharger  lag."   When  more  power   is  demanded  from  the
engine,  more  fuel  is added to the combustion chamber.   This in
turn  requires  a greater  amount  of  air for  complete combustion.
However, the  turbocharger,  which relies  on exhaust gas energy
to power it,  will  not gain the  additional  speed to supply more
air until the exhaust energy  increases.   Because of this period
when  the turbocharger  is not  "up  to speed" a temporary rich
mixture  exists  in  the combustion  chamber  and results  in  black
smoke,   i.e.,   incomplete  combustion.   This problem  has  been
alleviated to some degree by locomotive manufacturers.

     Switch engines  are  potentially the  most offensive.  These
units  are  typically  of   an  older  design  and may have been
removed  from  other service because  of poor  reliability.   They
often  receive no  preventative   type  of  maintenance and may be
repaired only  after  the higher  priority  line-haul locomotives.

-------
                             -Ill-
Switch engines  undergo many  throttle excursions  during  their
active service. - They  also  idle for long periods of  time  while
awaiting assignment.   The high activity levels of these  engines
plus  the  fact  that  they   operate   in   localized  areas  may
predispose them to a high frequency of observation making them
a  greater source  of  nuisance.   The  units,  however,  do  not
operate strictly in switchyards, but  are  used in local  service
train  for  distances  often  exceeding  25  miles from  their home
base.  These  locomotives  share common problems  with  line-haul
units   although  in   nonnaturally-aspirated  switch   engines,
turbocharger   lag  is  avoided  by  using  a  mechanically  driven
roots blower.  The  use of this blower, however,  makes  a switch
engine  more   susceptible  to  excessive  visible  emissions  at
higher altitudes where the air is  less dense.

     Considerable  uncertainty  exists  in  trying  to  estimate
future   trends  which   might  affect  excessive   particulate
emissions from  railroad  rolling stock. The cost of diesel fuel
has risen rapidly in the last  decade.  If  this trend continues,
railroads may find it  cost-effective to  reduce fuel waste by
increasing attention  to  maintenance details.  (Excessive  smoke
is  generally  a sign  of  poor  combustion  and, therefore,  poor
fuel economy.)  The financial  health  of  the industry will have
a  direct affect  by  allowing the  replacement  or  retrofit  of
outdated equipment or  necessitating the  continued use  of  dirty
engines along with currently deferred maintenance practices.

9.2.3  Interstate Commerce

     At  the  present  time,  major  disruptions   of  interstate
commerce have  apparently not  resulted from the large number of
varied standards applicable to railroad  rolling  stock primarily
because  of   the widespread  lack   of   enforcement.   Some  minor
problems  have apparently been  encountered  because  of different
opacity  regulations.   Naturally,  these have occurred when rail
operations pass from  a jurisdiction with lenient or nonexistent
regulations  to a  jurisdiction of more   strict  regulation  or
enforcement.    The  best  example  occurs  between  California and
Nevada.   In   this  area,  it  has   been reported  by  government
agencies  that only "clean" locomotives proceed  into California
while the "dirty" units are uncoupled  at  the border, presumably
to   service   shipments  moving  east.   Although  this  presents
logistics problems  and time  delays for  the  affected railroad,
it has not created any extreme adverse effects.

     Although   the   current   situation   is  not  of  immediate
concern,  the potential for  disruption by  increased regulatory
enforcement  under  multiple  standards poses  a potential  threat
to  interstate  commerce.   It  is   not  possible  to   predict the
exact  degree to which rail commerce  could be curtailed.   It is
clear,  however, that  rail  operations  could  be  affected  by   a

-------
                             -112-
moderate increase  in  enforcement against  line-haul  locomotives
because of  the multitude  of  different standards  and  political
jurisdictions.

     The  potential   hazards  of   varying   state  and   local
regulations  were  recognized  by   the   railroad  industry  and
resulted  in  the  request   by   the  Association  of   American
Railroads  for   preemptive  Federal  emission  regulations.   The
request was based  solely  on the need  to prevent  disruption of
interstate commerce by removing the burden of  complying  with a
multitude of different standards.

-------
                             -113-
                          References

     1.     "Exhaust  Emissions   from  Uncontrolled  Vehicles  and
Related Equipment  Using  Internal  Combustion Engines,  Part  1:
Locomotive Diesel  Engines and Marine  Counterparts,"   Southwest
Research   Institute,   Prepared  for   the  U.S.   Environmental
Protection Agency,  October 1972.

     2.     "Railway  Motive  Power," Simmons-Boardman  Publishing
Corporation,  1984.

     3.     "Assessment  of   Control   Techniques   for   Reducing
Emissions   from   Locomotive   Engines,"   Southwest   Research
Institute,  Report  prepared   for   the   U.S.   Department   of
Transportation  and  the  U.S.  Environmental  Protection  Agency,
April 1973.

     4.     "Locomotive  Exhaust  Emissions  and  Their  Control,"
Hare, C.T.,  Springer,  K.J.  and Huls,  T.A.,   ASTM Paper  74-D
GP-3, 1974.

     5.     Workshop   on  Diesel   and   Bus  Engine  Emissions,
Southwest Research Institute,  Sponsored by Diesel and Gas Power
Division of the American Society of Mechanical Engineers, 1979.

     6.     "NOx   Studies    with  EMD   2-567   Diesel   Engine,"
Storment,  J.D.,  Springer,   K.J.  and  Hergenrother, K.M.,  ASTM
Paper 74-D GP-14,  1974.

     7.     "Studies   of   NOx  Emissions   from   a  Turbocharged
Two-Stroke Cycle Diesel Engine,"  Southwest Research Institute,
Report  Prepared for the  U.S.  Department of Transportation and
the U.S. Environmental Protectional Agency, October 1975.

     8.    Statistics  of  Railroads  of  Class   1^   in  the United
States; Years  1968 to 1978, Association  of  American Railroads,
Statistical Summary  Number  63, December 1979.

     9.    Operating and   Traffic  Statistics,  Association  of
American Railroads,  O.S.  Series  No. 220, September  1979.

     10.   "1977 National Emissions  Report," U.S. Environmental
Protection Agency, Report No. EPA-450/4-80-005, March 1980.

     11.   Yearbook    of     Railroad    Facts,    1980   Edition,
Association of  American Railroads.

     12.   Car  and Locomotive  Cyclopedia of American Practices,
1974 Edition, Association of American  Railroads.

-------
                             -114-


                    References - continued

     13.    "Railroads   and   Air   Pollution:    A   Perspective,"
Report No.  FRA-RT-73-33,  U.S. Department of Transportation.

     14.    "A  Study  of   Fuel  Economy  and  Emission  Reduction
Methods  for  Marine and  Locomotive Diesel  Engines,"  Southwest
Research Institute, Report  Prepared for the U.S.  Department of
Transportation, September 1975.

     15.    "Train Generated Air Contaminants in the  Train Crews
Working Environment,"  U.S. Department of  Transportation,  Report
No. FRA/ORD-77/08, February 1977.

     16.    "U.S.  Coast  Guard  Pollution Abatement Program:   A
Preliminary  Report  on  the  Emission   Testing  of  Boat  Diesel
Engines,"  U.S.   Department  of   Transportation,  Report   No.
CG-D-21-74, November 1973.

     17.    Compilation of  Air Pollutant Emission  Factors,  U.S.
Environmental  Protection  Agency,   Publication  No. AP-42,  March
1975.

     18.    "Weather   Atlas   of    the   United   States,"   U.S.
Environmental Data Services, June 1968.

     19.    Garshick E.,  M.B. Schenker,  A. Munoz,  M.  Segal,  T.J.
Smith,  S.R.  Woskie,  S.K.  Hammond and F.E.  Speizer;  1987.   A
case-control study  of lung  cancer  and  diesel exhaust  exposure
inrailroad workers.  Am.  Rev. Resp. Dis. 135, 1242-1248

     20.    Garshick E., M.B.  Schenker,  A.  Munoz,  M.  Segal, T.J.
Smith, S.R.  Woskie,  S.K.  Hammond  and  F.E.  Speizer,;  1988.   A
retrospective  cohort  study  of  lung  cancer  and  diesel  engine
exhaust exposure  in railroad workers.    Am. Rev.  Resp.  Dis.  137,
820-825.

-------