DRAFT


      Summary  and Analysis  of Comments

Regarding the Potential Safety Implications

     of Onboard Vapor Recovery Systems
    U.S.  Environmental Protection Agency
          Office of Mobile Sources
                 August  1988

-------
                   DRAFT


      Summary  and Analysis  of  Comments

Regarding the Potential Safety Implications

     of Onboard Vapor Recovery Systems
    U.S.  Environmental Protection Agency
          Office of Mobile Sources
                 August  1988

-------
                       Table of Contents


                                                           Page
Chapter 1 - Introduction
     A.    Background 	  1-3
     B.    Nature of the Comments 	  1-4
     C.    Overview of the Current Analysis 	  1-13

Chapter 2 - The Safety of Current Fuel/Evaporative Systems
     A.    Introduction 	  2-1
     B.    Vehicle Fires 	  2-2
     C.    In-Use Performance of Fuel/Evaporative
             Emission Control Systems 	  2-18
     D.    Summary  	  2-31

Chapter 3 - Onboard System Design and Safety Comments
     A.    Summary of Technology & Safety Comments 	  3-1
     B.    Trends in Fuel/Evaporative System Design 	  3-12
     C.    Onboard System Design and Safety 	  3-40
     D.    Other Safety Concerns 	  3-75

Chapter 4 - Potential Safety Benefits
     A.    Effects on Service Station Safety 	  4-1
     B.    Potential Vehicle Safety Benefits Due
             to Onboard System Design 	  4-13
     C.    Summary and Conclusions  	  4-16
                                 /
Chapter 5 - Summary and Net Assessment
     A.    Introduction  	  5-1
     B.    Defining the Relationship Between
             Complexity and Risk  	  5-2
     C.    EPA's Onboard System Design  	  5-5
     D.    Safety Benefits of Onboard Controls  	  5-7
     E.    Net Safety Impact  	  5-8

Appendix I - Onboard System Designs Submitted by Manufacturers

Appendix II - Safety Implications of Onboard
               Refueling Vapor Recovery Systems

Appendix III - Service Station Fire Data

-------
                           CHAPTER 1

                          Introduction

A.   Background

     Section 202(a) (6)  of  the Clean  Air  Act (as amended  1977)
requires EPA  to determine the feasibility and desirability  of
requiring  onboard   control   of  refueling   emissions  as   an
alternative  to  implementing  Stage   II  controls.   If  onboard
controls  are  found  to  be   feasible  and  desirable,  and  after
consultation  with   the  Department  of   Transportation  (DOT)
regarding the safety implications of  such  controls,  appropriate
standards and regulations are to be prescribed.   In response to
this  requirement,   as   early  as  1981,  EPA  undertook  studies
related to refueling control.   More  recently,  EPA's decision to
propose an onboard refueling  control  requirement  arose from the
results  of   an  August  1984  analysis,   the   draft  Gasoline
Marketing  Study  (GMS).tl]    The  CMS  assessed  the  technical
feasibility,  effectiveness   and  efficiency,  costs   and  cost
effectiveness of both onboard  and Stage  II control  of refueling
emissions.  Reanalysis  done  in response  to comments received on
the GMS  led the Agency to the conclusion  that onboard controls
represented the preferred approach,  in  terms  of  the criteria
set forth  in  the Clean Air  Act, to  controlling VOC emissions
from  refueling  operations.    The Act  requires  EPA to consider
the    administrative   burden    of    enforcement,    equitable
distribution  of costs, and  effects on fuel economy in addition
to  the   cost  of   the  technology  involved.   This  process
culminated  in  the  August,   1987  proposal  to  require  onboard
control  of  refueling emissions  for  gasoline-fueled  LDVs,  LDTs
and HDGVs.

      In  accordance with  the  provisions  of the  Act,  EPA began
consultation   with  DOT'S   National   Highway   Traffic  Safety
Administration  (NHTSA)  in March of  1986.  Several months later,
automotive  interests  and  the  Insurance  Institute  for  Highway
Safety  (IIHS)  raised  a number of  safety concerns regarding
onboard  controls.   In  response  to  these  concerns,  and similar
concerns  subsequently  raised  by   others,  EPA  conducted  an
evaluation of the  safety implications of onboard vapor  recovery
systems.   This  report, released  in  June,  1987,  is  found in
Appendix  II.   The  report presented EPA's  initial evaluation of
the  onboard  safety  issues   and  was  the  basis of  the  Agency's
conclusions that  safe  designs are  available and that  onboard-
equipped  vehicles  can  achieve  the  same  level   of  in-use  fuel
system  safety  as present  vehicles.   Nevertheless,  some  safety
concerns were not  resolved to the satisfaction  of  other parties
involved prior  to  issuing the NPRM.  EPA  therefore agreed to a
supplemental  proposal  of  onboard  controls,  limited to  safety
issues  and other  significant  changed  circumstances,  to  allow
the   fullest   possible  discussion   and   consideration   of   all
relevant  safety concerns  and  to  allow  additional  opportunity
for public comment  on  EPA's  reanalysis of  the  safety  issues.

-------
                              1-4


     The purpose  of  this document  is  to provide a  summary  and
analysis of  the  safety  comments received  in  response  to  the
NPRM.   It  also  provides  EPA's expanded  and updated  analyses  of
the safety issues related to onboard controls.

B.   Nature of the Comments

     A listing of the commenters is contained in Tables  l-l  and
1-2.   EPA  received  well  over  1000  sets  of  comments  on  the
NPRM.   Of  course, not all  made mention of  safety  issues,  and
many  who  did   reiterated   positions   taken   by  others.    The
comments  that  specifically  addressed  safety  concerns  can  be
subdivided into  a number  of general areas,  based largely on the
source of the comments.

     The  primary  concern  expressed  by auto  industry-related
commenters was  that  onboard systems would be more  complex than
current  evaporative  systems,   and  this   would  lead  to   an
unguantifiable   increase   in  the  risk of   crash-  and   non-
crash-related   vehicle   fires.    More   specifically,   these
commenters  pointed to an increase  in  the size  and  number  of
components, an  increased  number of vapor line connections,  and
a general  concern that the location of some of these components
could  degrade  vehicle   safety.    The   IIHS  expressed  concerns
similar   to   those   raised  by   the  auto   industry-related
commenters.   Auto industry  commenters  also  offered  a  number of
safety  comments  on  specific  hardware components  of  onboard
control systems.

     Petroleum  industry  commenters, on  the  other  hand,  stated
that  adding onboard  controls  represented a  smaller change than
the  initial  reguirement  for evaporative controls or the recent
switch  to  fuel  injection systems.   They pointed out that larger
canisters  and  vapor  lines  were the  main  differences  between
onboard systems  and current   evaporative  systems.    They also
stated  that  those systems have had almost no history  of safety
problems,  and  pointed  to  studies showing  that  the  risk of
refueling  vapor   ignition   in   either   a   crash  or  non-crash
situation  was   very  small.   In  their view,  onboard  systems
offered the  opportunity  for  a  safety benefit  over  current
evaporative  systems  because they  could be  designed to decrease
the   risk   of  fuel   tank   overpressurization,   reduce  excess
evaporative  emissions  and running losses,  and reduce  the number
of  external  fuel tank connections.

     A  number  of comments  were  received  from  other  Federal
agencies,   and   state    government  agencies.    The  National
Transportation  Safety  Board and the Department  of  Commerce,  as
well  as a number of state  governmental commenters  supported the
concerns   raised  by   the   auto   industry   commenters.     The
California  Air  Resources  Board  did  not  make  a  detailed
assessment  of   onboard   safety,  but   nevertheless  felt   that
onboard systems were  similar  to current  evaporative control
systems. They also stated that  they were unaware of any serious
safety  problems with evaporative systems.

-------
                       TABLE 1-1

            !_IST DF CCMMENTERS ON THE ONBOARD MPRfl
ORGAN I 2 AT IONAL CGMMENTERS
ACTIVATED CARBCN, INC.
ALLIANCE OF AMERICAN  INSURERS
AMERICAN AUTOMOBILE ASSOCIATION
AMERICAN INDEPENDENT REFINERS ASSOCIATION
AMERICAN NORIT CORPORATION
AMERICAN PETROLEUM INSTITUTE
AMERICAN TRUCKING ASSOCIATIONS
HilOCO
ANDERSON DEVELOPMENT COMPANY
HREA AGENCY ON  ASING FOR NORTH FLORIDA, INC.
ARIZONA AUTO ASSOCIATION
ARIZONA AUTOMOBILE DEALERS ASSOCIATION
ARKANSAS FEDERATION OF WATER AND AIR USERS
ASSOCIATED EMPLOYERS  OF ILLINOIS
ASSOCIATED GENERAL CONTRACTORS OF ICwA
ASSOCIATED GENERAL CONTRACTORS OF MAINE
ASSOCIATED MOTOR CARRIERS OF OKLAHOMA
ASSOCIATION PETROLEUM INDUSTRIES OF PENNSYLVANIA
AUDI AG
AUTOMOBILE IMPORTERS  GF AMERICA, INC.
AUTOMOTIVE DEALERSHIPS OPPOSING ONBOARD CONTROLS (6 LETTERS)
AVIS FE:NT-A-CAR COMPANY
BAY 5TATE GASOLINE DEALERS ASSOCIATION
3Mt« GF NORTH AMERICA
3RESMAN, ABELL. AND KAY FOR AMERICAN CAR RENTAL ASSOC.
BUSINESS COUNCIL OF GEORGIA
CALSON CARBON CORP.
CALIFORNIA AIR RESOURCES BOARD
CAPPCZZOLI/BRAUN PATENTS
CARBON DEVELOPMENT CORP.
CAROLINA PETROLEUM DISTRIBUTORS
CATALER INDUSTRIAL CO.. LTD.
CENTER FOR AUTO SAFETY
CHEVRON USA, INC.
CHRYSLER MOTORS
CONSERVATION LAW FOUNDATION OF NEW ENGLAND
COIMTEL SERVICE CORPORATION
COOPER OIL COMPANY
DEPARTMENT OF ENERGY
DETROIT EDISON
DOVER CORPORATION
ENTERPRISE LEASING COMPANV
EXxCN CO.
FARMERS UNION CENTRAL EXCHANGE

-------
r I f- I R & D
-LGRIDA DEPARTMENT u-  TRANSPORTATION
FLORIDA PETROLEUM COUNCIL
FLORIDA PETROLEUM MARKETERS ASSOCIATION
FORD MOTOR  COMPANY
FRESHWAV -ODD  STORES
FRIENDS OF  LVNCHBURG STREAM VALLEYS
GABEL. RUDOLPH C.,  INC
5A3 PROCESSORS ASSOCIATION
3ASOLINE DISTRIBUTORS «< STATION OWNERS IN
  -AVQR OF  ONBOARD CONTROLS <665 LETTERS)
6ENEP.AL MOTORS COMPANY
3EGR6IA ASSOCIATION OF CONVENIENCE STORES
L-EGF.GIA OILMEN'S ASSOCIATION
SIANT INDUSTRIES
SOODWIN AND GOODWIN LAW OFFICES
•zOODYEAR TIRE  AND RUBBER COMPANY
2RACG
HEALTH EFFECTS INSTITUTE
HOGEN AND HARTSEN -OR DAIMLER-BENZ AG
HONDA
HOOSIER MOTOR  CLUB
HUSKY CORPORATION
ILLINOIS COALITION FOR SAFET  BELT USE
ILLINOIS PETROLEUM COUNCIL
INDEPENDENT GASOLINE MARKETERS OF AMERICA
INDIANA AUTO SERVICE ASSOCIATION
INDIANA FARM BUREAU CO-OP ASSOCIATION
INDIANA HANUFACT'JRED HOUSING ASSOCIATION
INDIANA PETROLEUM COUNCIL
INDIANA DETAIL COUNCIL
-N3LJRANCE INSTITUTE FOR HIGHWAY SAFETY
IOWA DEPARTMENT OF PUBLIC SAFETV/TRANSPORTATION
IOWA PETROLEUM COUNCIL
-JWA TIRE DEALERS ASSOCIATION
JAGUAR CARS LIMITED
JEFFERSON COUNTY, KENTUCKY PUBLIC SAFETY CABINET POLICY DEPT.
JOE BASIL CHEVROLET
KANSAS PETROLEUM COUNCIL
KELLER CRESCENT COMPANY
KEMP SERVICE CENTER
KENTUCKY AUTOMOBILE DEALERS ASSOCIATION
KENTUCKY CHAMBER CF COMMERCE
KENTUCKY PETROLEUM MARKETERS ASSOCIATION
h.URARAY CHEMICAL CO.
LAKES REGION ASSOCIATION
LAMBERT CHEVROLET CLDSMQBILE
LIFE,- JOES
MAINE BETTER TRANSPORTATION ASSOCIATION
•1AINE FARM  BUREAU ASSOCIATION
MAINE FOREST PRODUCTS CC-'JNCIL
"1AINE MOTOR TRANSPORT ASSOCIATION

-------
MAINE PETROLEUM ASSOCIATION
MARATHON  PETROLEUM COMPANY
MARSH VILLAGE PANTRIES
McCORMICK AND COMPANY, INC.
MICHIGAN  ASSOCIATION OF CONVENIENCE  S
                                        ORES
MICHIGAN MANUFACTURERS ASSOCIATION
MICHI6AN PETROLEUM ASSOCIATION
MIDWEST PETROLEUM MARKETERS ASSOCIATION
MIDWEST SERVICE STATION ASSOCIATION
nIL3RIM, THOMAJAN AND LEE ^OR _ANGHAM-HIi_L PETROLEUM
MINNESOTA  AUTO DEALERS ASSOCIATION
           FARM BUREAU FEDERATION
           GROCERS ASSOCIATION
           HIGHWAY USERS FEDERATION
           PETROLEUM COUNCIL
           RETAIl_ MERCHANTS ASSOCIATION
             ECONOMIC COUNCIL
             FARM BUREAU FEDERATION
                                                         CLAKENDGN
MINNESOTA
MINNESOTA
MINNESOTA
MINNESOTA
MINNESOTA
MISSISSIPPI
MISSISSIPPI
MISSOURI AIR  POLLUTION CONTROL PROGRAM
MISSOURI OIL  JOBBERS ASSOCIATION
MITSUBUSHI  MOTOR COMPANY
MOBIL GIL COMPANY
MORRISON IMPLEMENT,  INC.
MOTOR VEHICLE MANAGEMENT BUREAU
MOTOR VEHICLES MANUFACTURERS ASSOCIATION
MULTINATIONAL BUSINESS SERVICES,  IMC.
MUTURN CORPORATION
NATIONAL AIR  CONSERVATION' COMMIES I ON/ AMERICAN LUNG ASSOCIATION
NATIONAL ALLIANCE OF SENIOR CITIZENS,  INC.
NATIONAL ASSOCIATION OF CONVENIENCE  STORES
NATIONAL ASSOCIATION CF R.EET ADMINISTRATORS, INC.
fimriCNAL AUTOMOBILE DEALERS ASSOCIATION
NATIONAL nIGHWttY TRANSPORTATION SAFETY  ADMINISTRATION
NATIONAL PETROLEUM REFINERS ASSOCIATION
NATIONAL SAFETY COUNCIL
NATIONAL TRANSPORTATION SAFETY BOARD
NATIONAL TRUCK EQUIPMENT ASSOCIATION
NATIONAL VEHICLE i_ES!NG ASSOCIATION
NATURAL RESOURCES DEFENSE COUNCIL
NEBRASKA PETROLEUM COUNCIL
NEW HAMPSHIRE PETROLEUM COUNCIL
NEW JERSEY  DEPARTMENT CF ENVIRONMENTAL
NEW JERSEY  GAS STATION OWNERS IN  FAVOR
NEWHALL REFINING CO.
NISSAN MOTOR  COMPANY.  i_TD.
NORTH AMERICAN CARBON. INC.
      CAROLINA ASSQCInTIJN CF CONVENIENCE STORES
      CAROLINA AUTOMCBlL.£ DEALERS ASSOCIATION
      CAROLINA PETROLEUM MARKETERS ASSOCIATION
      CAROLINA TRUCKING ASSOCIATION
                                         PROTECTION
                                         OF ONBOARD
CONTROLS  <472  LETitRS;
NORTH
NORTH
NORTH
NORTH
NORTH
      DAKOTA PETROLEUM COUNCIL
NORTHEAST  STATES FOR COORDINATED  AIR  L'SE MANAGEMENT

-------
•vQRTHWEST FLORIDA  AREA  AGENCY UN A3 ING,  INC.
SOUTHWESTERN OIL
JCEAN AUTOMOBILES  TECHNOLOGIES,  INC.
OFFICE  JF MANAGEMENT  S<  BUDGET
OHIO ASSOCIATION OF CONVENIENCE 5TORES
OHIO SELL TELEPHONE COMPANY
OHIO DEPARTMENT OF HIGHWAY SAFETV
OKLAHOMA CITY CHAMBER OF COMMERCE
OKLAHOMA HIGHWAY SAFETY OFFICE
OKLAHOMA STATE CHAMBER  OF COMMERCE AND INDUSTRY
PALMETTO SAFETY COUNCIL
PAPER INDUSTRY INFORMATION OFFICE
PASCO MOTORS,  INC.
PENNSYLVANIA CHAMBER  OF COMMERCE AND INDUSTRY
PENNSYLVANIA MANUFACTURED HOUSING ASSOCIATION
PENNSYLVANIA MANUFACTURERS ASSOCIATION
PENNZOIL
PETROLEUM MARKETERS ASSOCIATION OF AMERICA
PEUGEOT
PMH GROUP INC.
POLK OIL COMPANY
PUGMIRE LINCOLN-MERCURY-MERKUR
QUIK STOP MARKETS
QUIK-CHEK,  INC.
R *« H MAXxON,  INC.
RECREATIONAL VEHICLE  INDUSTRY ASSOCIATION
REGIONAL AIR POLLUTION  CONTROL AGENCY
RENEWABLE FUELS ASSOCIATION
RETAIL GROCERS ASSOCIATION OF FLORIDA
ROLLS ROYCE MOTOR  CARS
RYDER SYSTEM
SAAB-SCANIA
SENATE COMMITTEE ON AGRICULTURE, NUTRITION AND FORESTRY
SERVICE STATION AND AUTOMOTIVE REPAIR ASSOCIATION
SERVICE STATION DEALERS OF AMERICA
SOCIETY BANK, DAYTON  OHIO
SOCIETY OF AUTOMOTIVE ENGINEERS
SOCIETY OF  INDEPENDENT  GASOLINE MARKETERS OF AMERICA
SGHIO OIL COMPANY
SOUTH CAROLINA DEPARTMENT OF HEALTH AND ENVIRONMENTAL CONTROL
SOUTH CAROLINA HIGHWAY  USERS CONFERENCE
SOUTH CAROLINA TRUCKING ASSOCIATION
SOUTHWESTERN BELL
STATE FARM  INSURANCE
SUBARU OF AMERICA
SUN REFINING & MARKETI'-JG
SUNOCO RETAIL MARKETING
TENNESSEE OIL MARKETERS ASSOCIATION
TEXACO
TEXAS AIR CONTROL  BOARD
TEXAS AUTOMOBILE DEALERS ASSOCIATION
TEXAS DEPARTMENT OF HIGHWAYS AND PUBLIC TRANSPORTATION

-------
TIME 3AVER STORES
TCn POLAND FORD
TOYOTA MOTOR COnPANY
TRAFFIC SAFETY ASSOCIATION OF MICHIGAN
TRI-STATE GASOLINE nND AUTOMOTIVE DEALERS ASSOCIATION,  INC.
UNION OIL COMPANY
UNITED PARCEL SERVICE
UNITED STATES DEPARTMENT OF COMMERCE
UNITED STATES DEPARTMENT QF TRANSPORTATION
UNITED STATES SMALL BUSINESS ADMINISTRATION
UNOCAL REFININS AND MARKETINS DIVISION
U.S. FLEET LEASINS, INC.
VAN WATERS riND ROGERS, INC.
VERMONT RETAIL GROCERS  ASSOCIATION
VIRGINIA AGRIBUSINESS COUNCIL
VIRGINIA CHAMBER QF COMMERCE
VIRGINIA FARM BUREAU FEDERATION
V IRS IMA GASOLINE AND AUTOMOTIVE REPAIR ASSOCIATION
VIRGINIA PETROLEUM JOBBERS ASSOCIATION
VOLKSWAGEN OF AMERICA, INC.
VOLVO CARS OF NORTH AMERICA
WEST VIRGINIA GASOLINE DEALERS AND AUTOMOTIVE REPAIR ASSOCIATION
idEST VIRGINIA PETROLEUM MARKETERS ASSOCIATION
WESTVACO, CARBON DEPT.
WISCONSIN MANUFACTURERS AND COMMERCE

-------
                       TABLE 1-2

           LIST  OF  CQMMENTERS ON THE ONBOARD NPRM
MEMBERS OF CONGRESS,  STATE
LEGISLATORS &  OTHER  OFFICIALS
PRIVATE CITIZENS
ADAMS. BROCK
*FFL£R3ACH, RAY  C.
BALLENGER, CASS
BARNARD, DOUG  JR.
BOULTER, BEAU
BRANSTETTER, OLIN  R.
3ROOMFIELD, WILLIAM S.
3USTAMANTE. ALBERT G.
3YRD. ROBERT C.
CESSAR, RICHARD  J.
CHENEY,DICK
CHILES, LAWTON
CLARK, -AJ ILL I AM A.
COURTER, JIM
COY, JEFFERY w.
DARDEN, GEORGE
DAWSON. W.T.
JEWINE, MIKE
DINGELL. JOHN  P.
DIXON, ALAN J.
DOLE. SOB
DONNELLY, BRIAN
EDWARDS,T.W. JR.
EXXON, J. JAMES
-nSCELL, DANTE B.
FAWELL. HARRIS -ft.
FEIGHAN. EDWARD  F.
FESSLER, RICHARD D.
FIELDS, JACK
FISHER, D. MICHAEL
FOLEY, THOMAS  S.
GRADISON, BILL
GRANT, BILL
GRE3G, JUDD
I3UAR INI. FRANK J.
GUNDERSON, STEVE
HARRISON, DUDLEY
HASENOHRL, DONALD
HATCH, ORIN S.
HAYES, SAMUEL  E. JR.
HECHT. CHIC
HEFLEY, JOEL
HEFNER, BILL
HERTEL, CURTIS
ABEL. CYNTHIA C.
ABEL, JODY
ABEL, MICHAEL
ANDERSON, 5EVERLEE F.
ARLOGAST, DAVID
BAINES, LAVERN AND JAMES
BATHE, JOHN b.
BERRY. ANNE K.
BERRY. ROBERT C.
BIEBER, C.F.
BLOUGH, r?ARY HELEN
BREGOLI, KENNETH
BREGOLI, MARYBETH
BROWN, CYNTHIA
BRYANT. CHARLENE
BRYANT, STEVEN
BUGG, DOROTHY S.
CADVJELL, BRUCE
CAMPBELL, KAREN R.
CARROL, DIANE
COLEMAN, PAUL
COSTEILLO, JOSEPH A.
COULSON. n. FRANCIS
CROWN. JOHN J.
DAVIS, EARL '-j.
DEARLES, HAROLD D.
DEIBEL, WILLIAM T.
DEMOS, RUSSEL
DiGIACOMO, JAN
DUMAS, S.J.
DUNHAM, CHRIS
ERNST, R.G.
GABEL. RUDOLPH C.
GIESCHEN. ALICE M.
GIRARD. MIKE
GREENHAUS, DOUGLAS  I.
GREGGS, NORMAN P.
GRENON, LEO E.
GRIGSBY, EVERETT M.
GUGLIATAN, RALPH
HANIFY, DONALD F.
HARRISON, DR. KEVIN
HOLLAT2, MR. £< MRS.  KENNETH
KELLY, niLDRED

-------
HESS, RALPH W.
HOFFMAN,  PHILLIP £.
HUfiBARD.  CARROL
iivDE. HENRY J.
JONES. WALTER E.
KAFOSLIE.  NICK
KASSER,  MATTHEW
KOSTMAYER,  PETER h.
KRUELL,  RICHARD T.
KOSTEVA,  JAMES A.
KUBIER,  JULIUS E.
t-ENTOL, JOSEPH ft.
LEWIS, H.  CRAIS
LIVINGSTON,  ROBERT L.
McCAIN,  JOHN
McCLURE.  JAMES A.
McCONNELL.  MITCH
McDADE,  JOSEPH M.
MCDONALD,  NANCY H.
McEWEN,  SOB
MCMILLAN,ALEX
McNALLY,  RANDY
MELCHER,  JOHN
MIKULSKI,  BARBARA A.
NELSON,  BETTY JO
NOLEN, FRANK W.
ORR, KAY A.
OUWINGA,  SIDNEY
DXLEY, MICHEL G.
PASSERELL,  '* ILL JAM J.
PETERSON.  WILi_IAM E.
PETRI. THOMAS E.
POLINSKY,  JANET
-•ORREA,  VINCENT
RAHALL.  NICK J.
REGOLI,  JOHN to.
RHODES,  JAMES J.
RICHARDSON,  BILL
ROCKEFELLER,  JOHN D. IV
ROMANELLl,  JAMES A.
ROSE, CHARLIE
ROSHELL,  MARVIN J.
ROTH, TOBY
ROWLAND,  JOHN G.
SARBANES.  PAUL 5.
SAVATCRE,  FRANK A.
SCRUGGS,  PAUL C.
SECRE5T,  JOE
5EVERENCE,  CHARLES rt.
SLAUGHTER,  D.  FRENCH JR.
SMITH, BILLY RAY
SMITH, i_AMAR
KING. FRANCIS  W.
KNOTT, C. ROBERT
LAPHAM. DELPHINA
LUCIANO, ANTHONY J.
LUNDBERG. JAN  C.
MACK, JULIE S.
MAGNANO. DAVID A.
MARKOWITZ, ROBIN
MATTIOLI, DIANE
McBEE, HOLLY =.
MEDLEY, J.H.
MILLER, ROBERT
MITCHELL, DAVID B.
NEUFELD, SUSAN
NEWPORT. HAROLD A.
NOYES. WALTER  0.
NUSSBAUM. MRS.  SRESSEl-
O'CONNELL, DANIEL K.
PAGE, ED
PHILLIPS, HOWARD E.,JR.
RAPP, PATRICIA M.
RAPP, PETER J.
RICHMOND. ARLENE
ROBISCONE, RALPH A.
ROMBERGER. WINIFRED
RUTES, IRIS
5EAY, CYNTHIA
SEAY, JEFFREY
SHATTUCK. JAY  DEE
SHEETS. RONALD L.
SHELTON, HENRY Z.
SHEPARD, JOAN
SIEGMAN, JOSEPH
SPERLING. SHARON LEA
STAUCH, ALBERT
3TROME, IRENE
TABLER, KAREN
TOAL, CHRISTINE A.
TOAL, KATHLEEN F.
TODD, J. RICHARD
TUCKER, BARBARA
TUCKER, LISA
TUCKER, WILLIAM Q.
VOEKS. JOHN F.
WALL, CATHERINE E.
WARNER, KIMBEF.LY
WHYTE, BRIDGET J.
WHYTE, DANIEL  T.
WHYTE, MARY LOUISE
WHYTE, MICHAEL
WHYTE, THOMAS  P.
WILLIAMS, *ILL F.

-------
SMITH, ROBERT C.
SORRENTO, L.J.
SPECTER, ARLEN
STAPLETON, PATRICK  J.
3TECZG, TERRY A.
STRANGE, JAMES R.
3TRANGELAND, ARLAND
SULLIVAN, MIKE
BUNDQUIST, DON
THURMOND, STROM
VANDER JAST, GUY
VARNED. DOUG JR.
vJILK INS, 5. VANCE
WILT. ROY W.
ZEMPRELLl, EDWARD P.
WISLOCK1, JENNIPER
WISLOCKI, THEODORE M.
WRIGHT. SAMUEL H.
VERGES. JAMES J.
YOUNG, E. LEE -«< CO.

-------
                              1-13


     Several  public  interest  groups commented  on  the  onboard
safety  issue.   The most  extensive  comments were  submitted  by
the Center for Auto Safety  (CAS).   In a detailed study of  NHTSA
safety  complaint  reports  and recall  files,  CAS  found the  risk
of fires  from onboard  systems to be minimal and easily  handled
by improved technology. This  was  because onboard systems  were,
in  their  view,   only  marginally  more  complex than  current
evaporative systems, which their  study  indicated are  an  almost
insignificant  source  of  current  vehicle fires.   CAS  expressed
the  view  that   onboard   controls  combined  with   volatility
controls  would enhance vehicle  safety.  The Natural  Resources
Defense  Council  supported  the  conclusions  of   the  CAS  report
that vehicle  safety would be improved by onboard and volatility
controls.     They   stated    that   onboard    systems    were
"evolutionary,"  rather  than  "revolutionary,"  and  urged EPA  to
proceed with  the  rulemaking.   The National Safety Council  took
no  official  position,  but   indicated  that  it  also  had  some
reservations regarding possible additional fire risk.

     In addition to the above general groups, EPA received many
comments  from members  of Congress,  state  legislators,  various
special  interest  groups such as  oil marketing or auto  service
groups,  and a large number  of private citizens expressing their
views  on  the onboard  proposal.    Many of  these  commented  on
safety  issues, generally  supporting the views  of one or another
of the  groups mentioned  above.   Since  these  comments  tend  to
fall  into  the  abovementioned  categories, they  will  not  be
specifically   identified,   but  are   implicitly  considered  and
addressed  along  with  the  other   comments  dealing  with  the
general  issues outlined above.

     EPA's  initial  study  of  the potential  safety  implications
of onboard vapor recovery  systems,  released in June  1987,  was
designed  to  identify  and  evaluate both  general  and specific
onboard  safety concerns  which were raised prior to  the NPRM.
The  study  discussed  the  design  of safe  onboard   systems   and
evaluated   in-use  safety   issues   such   as   crashworthiness,
tampering,  defects,  misrepair  and  refueling  operation safety.
In   general,   there  were   few   comments  directly   addressing
technical  aspects  of EPA's  report.   The  comments  received were
general  in nature or  amounted to  a suggestion  that  additional
analysis was  needed to  support EPA's conclusions.

C.   Overview of the Current  Analysis

     The purpose of this  report  is  to consider  and evaluate  new
issues  that have been  raised by the commenters and  to provide
supplemental   analysis  for  past  concerns  where  necessary   in
response  to   the  comments  received.    For  areas   where   the
previous   analyses  is   sufficient,   the   comments   will   be

-------
                              1-14


summarized  but  referred  to  the  initial  study  for   response.
This  Summary   and   Analysis   of  Comments  consists  of  five
chapters, plus  this Introduction.   Each of these  chapters  is
discussed below.

     Chapter  2  deals  with  the  safety  of  current   fuel/
evaporative systems.   Many auto  industry  commenters,  and some
others as well, have suggested that  onboard  controls would  be
more  complex  than  current evaporative  systems  and  thus  would
lead  to  an unquantifiable increase  in the risk of  crash-  and
non-crash- related vehicle fires.  Some  have  stated  that  there
was   an   increase  in   vehicle  fires  following  the  initial
requirement  for  evaporative   systems.    The   purpose of  this
chapter  is to establish  a baseline for vehicle fire risk and to
assess  the safety  performance  of  fuel  evaporative emission
control systems since their inception.

     The  next  section,  Chapter  3,  focuses on the analysis  of
comments  received  regarding  onboard  control   system   design
considerations  and  how these  affect  safety.   It begins with a
summary  of the  technology  and  safety  comments received   in
response to the NPRM.  This is  followed by a review of trends in
fuel/evaporative  control  system  design  which  helps  provide  a
baseline  for  comparison.  The  remaining  two  sections  of  the
chapter  provide EPA's  analysis of the  comments  in  this  area.
The  third  section also includes  a discussion  of  recent  onboard
prototype system development  work completed by  EPA.

     Chapter  4  addresses  some  potential  safety benefits   of
onboard  controls.   These  include a reduction  in service  station
fires   related   to  refueling,   and   a   discussion  of   how
incorporating various  onboard system  design   features  provides
an  opportunity for  improvements  in  the  operating  safety  of
in-use fuel systems in  both crash and non-crash situations.

     Chapter  5  presents  a summary  of  EPA's   findings  in  the
previous chapters and  a  current  net  assessment of the expected
safety impact of onboard controls.

     Section 202(a)(6)  of  the Clean Air Act also provides that
if   onboard  controls  are  required,  EPA shall provide  the
manufacturers   adequate   leadtime   for   implementation.    A
substantial number of  comments were also  received regarding the
necessary  initial  leadtime or  the desirability of   a phase-in
period   for  safe  and  effective  implementation  of  onboard
controls.  EPA  recognizes that leadtime  is a critical component
of   the  manufacturers'  overall  ability  to  implement  onboard
controls  safely  and  effectively.   However,   this  report  is
designed to deal only with the technology  and  safety  aspects of

-------
                              1-15
onboard implementation.   It  is  not  designed to  deal  with  the
leadtime issue.  EPA intends to completely  resolve the  leadtime
issue before the Final Rulemaking,  and  intends  to  consider  and
account for  all factors  raised by  the commenters relevant  to
the  leadtime   issue.   The Agency  remains  fully  committed  to
providing the  industry with adequate leadtime for  the  safe  and
effective implementation of  onboard  controls,  if  such  controls
are imposed.
0365X

-------
                              1-16


                    References for Chapter 1

     1.    "Evaluation  of  Air  Pollution Regulatory  Strategies
for Gasoline  Marketing Industry," U.S.  EPA,  Office of  Air  and
Radiation, EPA-450/3-84-012a, July 1984.

-------
                           CHAPTER 2

     The Safety of Current Fuel/Evaporative Control Systems

A.   Introduction

     One of  the  key arguments  against the  promulgation  of  an
onboard  refueling  emission standard  is  the assertion that  the
implementation of  onboard refueling  control  systems will  lead
to  an   unquantif iable   increase   in   the  risk  of  crash  and
non-crash fires.   Generally  speaking, the  increased risk  that
is claimed  to  accompany  onboard  systems  refers to  a  potential
increase in fuel and/or vapor leaks which in turn  could  lead to
an  increase  in  the  number  of  vehicle  fires and  associated
consequences.

     Those  in  support   of  the  increased  risk  argument  have
further  claimed that while the  number of increased fires cannot
be quantified,  any potential increase  in  vehicle  fires  warrants
the withdrawal  of  the onboard proposal.  However,  it  should be
noted  that  independent  of  the  potential  effect  of  onboard
controls,  this  all or  nothing  philosophy seems  inconsistent
with   the   acceptance   of   fire   risk   associated  with  the
implementation  and  continued  use of other  vehicles  systems
where   the   added  risk   is   determined   to   be   finite  yet
sufficiently small to be acceptable.

     Any consideration  of incremental risk  must  start  with the
determination of the baseline  for that risk.   When considering
vehicle  fires,  the number of incidents  and the consequences of
those  fires  must be assessed,  with special attention  given to
the baseline risk  for  those  portions  of the vehicle fuel system
most  likely to  be  impacted  by  onboard  controls  (i.e.,  the
evaporative  control system).  The purpose of this section is to
characterize the current level  of fire risk that  now  occurs in
vehicles,  and to  thus  provide  an  adequate  basis to  put  the
alleged  incremental  fire  risk  of   onboard   into  the  proper
perspective.   This  section  will  examine   estimates  of  total
annual  vehicle fire rates and evaluate the  extent  to  which the
fuel tank was  a  factor in these fires.  Consequences of current
vehicle  fires  such  as  deaths,  injuries,  and property  damage
will also  be assessed.   In  addition, trends  in  the historical
fire rates will be  analyzed to evaluate the extent to which (if
any) evaporative control  systems have  affected  fire rates.

     In  addition  to  vehicle  fires,  this  section will  focus
directly on  the   in-use  performance of   current  evaporative
control  systems  to evaluate  the impact these  systems  have had

-------
                              2-2


on overall vehicle  safety.   A good  indicator  of the safety  of
evaporative systems  can be  obtained by  examining  the  failure
frequency  and  severity of  associated consequences relative  to
other types of system failures.  Failure  characteristics can  be
assessed   through  the   following  three  mechanisms:   recall
campaigns,   manufacturer    service    bulletins,    and    owner
complaints, all three of which  are maintained  in computer  files
by NHTSA.  All  of these records  will be  examined  in the  last
part of  this  section to assess the  relative safety performance
of evaporative control systems.

B.   Vehicle Fires

     The examination of past  fire  occurrences  in motor  vehicle
accidents  serves  two useful  functions.    First,  it provides  a
good baseline  for consideration  of  the   "unquantifiable  risk"
claims  of onboard  and  allows  one  to place  these risks  into
context.  Second,  historical fire  rates give a  good perspective
on  how   significantly  other  fuel  system  changes   such  as
evaporative  control  systems  (which  EPA believes   are   very
similar  to  onboard   controls)  have  affected   fire rates.   By
analyzing  the  impact  evaporative  control systems  have had  on
fire rates, a more  definitive estimate of the  impact  of onboard
control systems can be made.

     1.    Total Annual Collision Fires

     When  one considers  the  number of  vehicles  in use,  the
number of  miles  driven, and even  the number of accidents  which
occur  each   year,   motor  vehicle   fires  are  rare   events.
According  to  the  NASS data base  discussed below,  vehicle  fires
occur  in  only  0.25  percent  of  all accidents.   It  is  more
difficult  to  determine  with  precision   the   exact  number  of
vehicle  fires  that  occur  each  year.   Not   all  fires  get
reported,  and  not  all  reports  are  gathered  and  processed
through  a  single  uniform  collection  system.   Most  of  the
available  information  arises  initially   from  police  and  fire
marshall  reports  from  different  states, but   the nature  and
extent  of  the  fire  data  vary  in different  state  reports.
Nevertheless, several computer data  bases have been established
in  recent years  which  gather  sufficient  quantities of accident
data   from   the   states   to   make  possible  a  reasonable
extrapolation of national projections.

     Three  such  data  bases were  examined recently  by EPA and
NHTSA  in  attempts   to  characterize  current  and  historical
vehicle  collision fire  rates.   EPA's findings  are  summarized  in
its  "Analysis  of  Fuel  Tank-Related  Fires,"[1]  and  NHTSA1s
analysis  is  contained  in  the draft contract  report,  "Study  of
Motor  Vehicle  Fires.  [2]   The   three   data  bases  that  were

-------
                              2-3
examined are  the  National Accident Sampling System  (NASS),  the
Fatal Accident  Reporting  System (PARS),  and  the National  Fire
Incident  Reporting  System   (NFIRS).   Both  NASS and  FARS  are
operated  and   maintained  by  NHTSA  whereas   NFIRS   is   the
responsibility  of  the   Federal   Emergency  Management   Agency
(FEMA).

     NASS,  a  statistically  based  data   sampling  and  analysis
system  gathers  information  on  all types of accidents and  was
used in NHTSA's fire report[2] to project that  about  16,700  car
fires occur  annually  in  police reported accidents  nationwide.
This   estimate   is   consistent   with   the   19,500   annual
post-collision  fires  estimated  in an   earlier  NHTSA  report,
"Evaluation  of  Motor  Vehicle  Safety  Standard  301-75,  Fuel
System  Integrity:   Passenger  Cars,"[3]   and  the 15,313  annual
post-collision  fires  estimated  in a  report  by  the  Highway
Safety Research Institute (HSRI,  now  the University of Michigan
Transportation Research Institute  or UMTRI)  entitled,  "Fires in
Motor  Vehicle  Accidents."[4]    Therefore,   it  appears  that
between 15,000  and  20,000 vehicle fires   occur  annually  in this
nation.   Based  on  data obtained from FARS, EPA's fire reporttl]
estimated  that  up  to   nearly   1,700   people   are   killed  in
accidents  involving fire each  year.   In addition,  about 3,700
serious   injuries   and   3,600   moderate   injuries    occur   in
post-collision motor vehicle fire accidents each year.tl]

     2.     Non-Collision Vehicle Fires

     In addition to vehicle  fires resulting from crashes, it is
also  worthwhile to  characterize  the  number  of vehicle fires
that  result  from  some defect  or  failure in the vehicle which
did  not  involve a  crash.   One data base,  NFIRS, does   contain
information on  non-collision vehicle  fires.   However,  this data
base  contains  information  on all  non-collision vehicle fires
regardless of  whether the cause  of the  fire  originated in the
vehicle.   For  example,  NFIRS contains  data on  fires where   a
building  caught fire and then  spread to a vehicle.   Because of
the  wide  range of  causes of  vehicle  fires,  and the  level of
detail  reported by NFIRS, it is not always possible to identify
which fires originated as a  result  of a problem  in the vehicle.

     Therefore, it  is not possible to  identify the  exact number
of  true non-collision vehicle  fires  from NFIRS.  Nevertheless,
NFIRS  can provide some  interesting summary   statistics.   For
example,  in  1986,   over 350,000  vehicles were involved in fires
not  related to  crashes.   This figure is  nearly 20  times higher
than  the  number of  vehicle  collision fires.   About  100,000 of
these   350,000   fires   involved   electrical   equipment,   and
approximately 150,000  involved gasoline.

-------
                              2-4


     3.     Fuel Tank-Related Fires

     The  total number  of  annual  vehicle fires  helps  give  an
overall  view of the fire safety performance of  current  vehicles
and the current  level  of fire risk now accepted  for  the  entire
vehicle.   More specific to onboard  refueling  control  systems
however,   is  the  number  of  collision  fires related to  the  fuel
tank since many commenters have stated that onboard will  affect
the number  of connections  to  and from the fuel  tank,  and  this
in turn will  affect the fuel  system  integrity during  a  crash.
Thus, one area to  consider  in the assessment  of the potential
change in risk brought  about by onboard control  systems  is  the
number of collision fires related to the fuel  tank.

     EPA  has  already performed a  comprehensive analysis  of  the
number  of  fuel  tank  related  collision   fires.Cl]   NHTSA  has
reviewed  and commented  on  EPA's  analysis which  estimated  that
between 10-30  percent  of all  vehicle collision  fires  affected
the  fuel  tank.   This  translates  to  approximately  1,900-6,000
annual fuel  tank-related collision fires.  EPA's analysis  also
contains  ranges  of  the consequences  that accompany fuel  tank
related  fires.   These  consequences   are  summarized  in  Table
2-1.  It  should be  noted  that the number of  fires  quoted here
as  being  related  to the  fuel  tank  are   only  estimates.   Some
uncertainty  exits  in these  estimates  because it  is  not  always
possible  to  distinguish between such things  as fuel  tank fires
and  trunk fires.   Also, there  is some uncertainty with  regard
to  how  well  reported  fires represent  all   fires.   Therefore,
these figures  are simply approximations.

     Some  commenters  have  also  stated  that  onboard  controls
have  the   potential   to  affect   non-collision  fires.   EPA's
analysis   indicates  that   about   4,750-10,700  collision  and
non-collision  fuel  tank fires  occur  each year,  which  suggests
that  non-collision  fires are  at  least as frequent as collision
fires.*   The  non-collision  fires can  also  be  associated  with
serious  consequences  as indicated  in Table  2-1.   A complete
baseline  for onboard vapor  recovery  is therefore  contained in
Table 2-1,  and the change  in fire risk associated with onboard
controls  (for  collision and  non-collision)   can therefore  be
viewed relative  to  the information in this table in  addition to
the overall  fire risk data  discussed  above.

     4.     Evaporative  Control  Systems' Effect  on Fire Rates

     In addition  to estimates of overall  and  fuel tank  related
fires,  in assessing potential  onboard risks  it  is  also useful
     Because  these  estimates  were  developed  using   different
     data  sets,  it   is  not valid  to  subtract  collision  fires
     from  the total  to  get a  direct  estimate of non-collision
     fuel tank  related fuels.

-------
                              2-5


                           Table 2-1
      Fuel Tank Related Fires  and Associated Consequences*
                         (Annual Basis)
                                            Collision  and
                           Collision        Non-Collision
Fuel Tank Fires          2,000  - 6,000       4750 -  10,700
Fatalities               125 -  840           235 - 840
Serious Injuries         335 -  1,140         625 - 1,140
Moderate Injuries        1,130               1,130
Property Damage          $14 million         $32 million
     Data from reference [1].

-------
                              2-6


to examine  how fire rates were  affected by the  implementation
of evaporative control  systems  since  these  systems  are  very
similar   in   nature  to  onboard   refueling  control   systems.
Evaporative  control  systems'  effect  on  fire   rates   forms  a
useful  perspective  on  the  likely  impact of  onboard  control
systems on safety.

     Federal  regulations  requiring  evaporative  control  systems
were  first  implemented  beginning  with  model  year  (MY)  1971
vehicles  (passenger cars and light trucks to 6000 Ibs GVW) .   It
has been  suggested that by  comparing  fire rates of model  year
vehicles  a few years prior to the implementation  of  evaporative
controls  with model  year vehicles  a few  years subsequent  to
evaporative controls, it may be possible to identify whether or
not this  supposed  increased complexity  to  the  fuel system led
to a  significant change in  fire  rates.   One  of  the  commenters
(General  Motors) submitted an analysis of this  type using  FARS
data and  accident data obtained from a variety of states.   (The
results of this analysis are discussed later in  this section).

     EPA  performed  its  own analysis  of this  type using  FARS
data  because  FARS  was  one of the  data bases  utilized  in the
analysis   submitted   by  General   Motors.    However,   before
presenting  this  analysis, it should  be  noted that  EPA  is not
confident  that FARS  is  a suitable data  base  for this  type of
analysis  for  the following  reasons.   First,  FARS only gathers
data  on   all police-reported   accidents   (for   all  types  of
vehicles)  in which  a fatality  occurs.   Since   only  6  in  1000
accidents  involves  a  fatality,  a  large  number of  accidents
would  not  be considered  in  an  analysis of  FARS  data.   In
addition,  FARS  covers  all  vehicle  fires  whether  they  are
related to  the fuel system  or  otherwise,  and does  not specify
the  origin  of  the  fire.   Since  evaporative   control  systems
would only  affect  fuel system fires,  use  of  a  data  base which
contains  data on a variety of vehicle fires could mask the true
effect  of a  change to  the  fuel system.   Further,  since  FARS
data  represent  only  accidents  with fatalities, the  accidents
contained in  this  data base are generally more  severe,  and are
also  somewhat limited in  their  coverage of the  complete  range
of  vehicle fires.    Therefore,  use  of these data  alone  could
misrepresent  or  overestimate the overall  fire  hazard for  these
vehicles.   Nevertheless,  the  results  of  EPA's analysis  are
presented here for comparison with  the  results  of  the analysis
submitted by  General Motors.

     Table  2-2 shows  FARS  data for  MY 1966-86  vehicles  taken
from  the previously  mentioned draft  NHTSA contract study.[2]
As  shown in  this  table, FARS data  indicate  a  small  generally
increasing  trend in fire rates for fatal accidents between 1966
to  1975  model years.   However,  a  similar  trend  can  also  be
observed   in   the   data   for  model  years  1981  to  1986.    An
important question to ask is  whether the change in  fire  rates
that  occurred with the  implementation  of  evaporative controls
is significantly different from normal year to year variations.

-------
                         2-7


                      Table 2-2
        Car Fire Rates Per 100 Fatal Accident
       Involved Cars  and  Their  Standard  Errors*
Model Year
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
* Data from FARS
** R1- an Hard arror
Fire Rate
1.72
1.94
1.96
2.08
2.13
2.34
2.38
2.26
2.37
2.90
2.53
2.50
2.29
2.62
2.19
2.02
2.06
2.28
2.28
2.50
2.53
(1975-1986) .
J-iaeeH on c-t-a-t- i ct- i <-•
-------
                              2-8


     Using  a  multivariate  linear  regression  technique,   the
following best-fit  curve  was obtained  from the  PARS  data  for
model years 1966 to 1975:

  Fire Rate = -201.7 + 0.1035(MY)  - 0.0335(EVAP)       r2  = .84

     where:

           MY = model year
           EVAP = 0 if MY < 1971,  or
           EVAP = 1 if MY > 1970.

     The  "EVAP"  parameter was  determined  to have  a  confidence
level  of  only  14%  (i.e.,  there  is  an  86%  probability  that
"EVAP" has  no effect on fire rates), whereas  the constant  term
(-201.7)  and  model  year  parameter  were  calculated   to  be
statistically significant at  a  confidence  of 98%.  Hence, it is
highly  unlikely that  evaporative controls  had  any  effect  on
fire  rates.   Further,  even though  the coefficient of the  EVAP
term  is  negative,  EPA  does  not  view  this   result   as  an
indication  that  evaporative controls may  have decreased fires.
Our  overall conclusion  is  that  the changes  observed   in  fire
rates  from one model year  to another are due to factors other
than evaporative control systems.

     For  example,  over the two year period of  1968-1970,  fire
rates   increased  8.7   percent  without  any  influence  from
evaporative control systems (in other  words, other factors were
responsible  for  this increase).   Similarly,  from 1981-1983 and
1984-1986,   fire  rates  increased   12.9   and   11.0   percent,
respectively.  From  1970  to 1972  (during which time evaporative
control  systems were  implemented),  fire  rates  increased  11.7
percent.   This  figure  is  not  substantially  different  from
normal  year to  year variation or  the  increase  that  occurred
when  evaporative  control  systems  were  not  even  a   factor.
Therefore, the change in fire rates  from 1970  to 1972 cannot be
attributed  to  the  implementation  of evaporative control  systems
since  changes  similar  in  magnitude  occurred  in  fire  rates due
to  other   factors  even  without   the presence   of  evaporative
controls.   The  increase in fire  rates from 1968  to  1972  could
be partially  related to the fact that newer model year vehicles
were  on  average  exposed to  higher levels  of  fuel  volatility
over  their lifetime as  compared  to older  model year vehicles.
At  any  rate,  it  is  not possible  to conclude  from FARS  data
whether  evaporative control  systems had  any  significant effect
on fire  rates.

      It  should  be  noted  that vehicle  fires are  rare events
especially when considered relative to other  vehicle risks.  No
clear  trends  are   obvious  in  rate changes  between  any  model
years.   Overall,  fire  rates have fluctuated  somewhat  but  have

-------
                              2-9
continued to  remain low.  Other  changes more significant  than
evaporative control systems have occurred to fuel  systems,  such
as fuel  injection,  but even the effects of these  changes  would
be difficult to detect in rates that are so low.

     Further,   data  from  the  states  of Michigan  and  Maryland
also taken  from the  previously  mentioned NHTSA report[2]  (Table
2-3)  suggest  that   post-evaporative   fire  rates   were   even
somewhat  lower than pre-evaporative  fire rates.  For  example,
Michigan  data  show  an average pre-evaporative  (1968-70)  fire
rate  of  3.0   fires  per  1000   accident  involved  cars,  and  an
average   post-evaporative   (1971-73)    fire   rate   of   2.87.
Similarly,  Maryland  data  show  an  average  pre-evaporative fire
rate  of  3.8,  and  an  average  post-evaporative   fire  rate  of
3.23.  The  average pre-evaporative  Ohio data  shown in this same
table  indicate essentially no  difference  between the  average
pre- and post-evaporative fire rates.

     The  Michigan data  also  included  information on fuel leak
rates.    It  is  interesting  to  note   that   fuel  leak  rates
continued to  decrease from MYs 1968-73 even  though evaporative
control  systems added  additional  components and  connections  to
the  fuel  system.   For  example,  the average pre-evaporative fuel
leak  rate was  19.7  per  1000  accident  involved cars,  and the
average  post-evaporative  rate  was  14.1.   This  indicates that
fuel   system   vulnerability   in  the  form   of   additional
disconnections with  subsequent possible  leakages  does not have
to  increase  with  the  implementation   of  new  components  and
complexity.    The  Michigan   and   Maryland  data  suggest  that
evaporative  control  systems  did  not  degrade  safety,  although
one commenter  did provide analysis  to the contrary.

     Failure  Analysis  Associates  (FaAA)  prepared  an analysis
for   General   Motors  using  FARS   data,   which   stated  that
post-evaporative  fire   rates  were significantly  higher  than
pre-evaporative fire rates.   In EPA's  view, however, there are
several  problems  with  their  analysis.  To  begin  with,  FaAA
developed a technique to  neutralize  vehicle age  as  a possible
confounding  factor  in their comparison,  but by  doing so, may
have  inappropriately biased the data.*

     FaAA used   1968  through  1973  model  year  vehicles  and
separated   the   vehicles    into   pre-    (1968-70)    versus
post-evaporative  (1971-73) groups.   Next since they  wanted data
on  vehicles of the  same age and were  working with  a data base
     The  following  paragraphs  discuss FaAA's  age  neutralizing
     technique  as  it  applied specifically  to  FARS data.   FaAA
     also  applied  this  technique  to  the  data   from  several
     states,  and  in general the  discussion that  follows  also
     applies  to  FaAA's  analysis  of the state data  with specific
     differences noted  where applicable.

-------
                             2-10
                           Table 2-3
        Car Fire Rates Per 1000 Accident  Involved Cars
        in the  States of Michigan, Maryland, and Ohio


Model
Year
1968
1969
1970
1971
1972
1973
MICHIGAN
(1978-84)


Fuel Leak
No Fire Fire Rate*
20.8
19.7
18.6
16.2
14.1
11.9
3.0
3.2
2.8
2.8
2.9
2.9
MARYLAND
(1978-84)
Fire Rate**
3.5
4.4
3.5
3.5
3.1
3.1
OHIO
(1982-1984)
Fire Rate***
2.5
3.2
2.4
3.0
2.7
2.5
*
**
* * *
Irrespective of fuel leak.
All fire involved accidents.
Due to crash.

-------
                              2-11


(FARS)  which  was  started  in  1975  (the  first  year  of  data
collected from  the  various states  ranged  from  1970  to  1982),
several analytical decisions  were made  on  how the data  should
be used.  Different model  years  were paired ('71  & '70,  '72  &
'69,  and   '73  &  '68),  and  then FARS  data for  the  different
calendar years  were used  for vehicles  of  the  same  age.   For
example, as  FaAA  stated in their report, "for  a collision  of  a
1971 model-year vehicle to be counted in  a  given accident year,
the previous  accident  year must be  available  for  1970  vehicles
of  the  same age.   Similarly,  in order to  count  1972   and  1973
model-year  vehicles,  accident  data  for  three  and five years
earlier had  to  be available to  provide  the balancing  data for
1969 and 1968 model-year vehicles, respectively."

     The first  problem with this technique is that the  desire
to  use  matched model years  as described above  led to several
situations  where  apparently  valid  data  were  excluded because
there were  not  corresponding  data for the  matching model year,
even  though there were data  for vehicles  of  the  same age and
contrasting  type  (evap  or  pre-evap.)   in non-matching  model
years.

     This is  illustrated  in  Table 2-4, which shows the approach
used by FaAA applied  to FARS  data.   Across the top and bottom
of  the  table  are  the evaporative  and pre-evapprative model
years   as  matched  by   FaAA;   listed  down  the  sides  are  the
calendar years  of FARS data  available.  The entries within each
column  are  the age/vehicle  type for  each  matching model  year
set.   FaAA  used  only  data  where  there was  corresponding age
data  within the same  model  year set.   Thus as  can be seen in
the table  there were  six calendar  years of pre-evaporative or
evaporative  data  excluded  under  this   approach,  even  though
there were  data for  the contrasting vehicles of the same age in
non  matching  model  years.   The   effects of   this   apparent
omission are uncertain.   It  appears,  however, that 3  KY/CY of
pre-evaporative  and   3  MY/CY  of   post-evaporative   data  was
excluded.  This represents about  11 percent of the data used.

     Second,  since  this approach used only FARS data,  it forced
consideration  only  of  data  for  vehicles five years  of  age or
older.   (For the state analyses, vehicle  ages considered  range
from  zero  years and older up  to  twelve  years  of  age   and  older
for  the various  states.)   This alone  could  introduce  an age
bias.   Clearly,  any safety  assessment should be  based  on the
full  life and history  of  the fleets  involved.

     Third,  as discussed  earlier,   any  analysis based on  FARS
data may be  inherently unrepresentative  of  the true performance
of  the systems  since  a  fatal  accident  had to  be   involved.
Vehicle fires  are rare,  and  deaths  involving accidents and fire
are only  a  small  percentage of  that total  (<5  percent).[2]
Furthermore,   only  about  6   in  1,000   accidents  involves   a
fatality,   so  a   large  number  of   accidents  were   not   even

-------
CY
FARS
Data
                             2-12


                           Table 2-4

       Illustration  of FaAA Matching Scheme for FARS Data

                     Evaporative MYs  - "E"
                   71
75
76
77
78
79
80
81
82
83
84
85
86
El
5E
6E
7E
8E
9E
10E
HE
12E
13E
14E
15E
5P
6P
7P
8P
9P
10P
IIP
12P
13P
14P
15P
|16P

                  72
                    70              69

                   Pre-Evaporative MYs - "P1
73
                                   68
                A total of  six data points were excluded under
                MY  matching  scheme,   even though . each  point
                could  be  matched  with  a  contrasting  vehicle
                type in another MY.
           D-
A  total  of  18  data  points  were  excluded but
usable if vehicle age is not  a  factor.

-------
                              2-13


considered in the analysis of PARS data.  A valid  analysis  must
consider  all  accident  and  fire  experience  not  just   fatal
accidents.

     Fourth, the  fact that  different calendar  year data  were
used for pre-  and  post-evaporative  vehicles  introduces  a  number
of  potential  confounders  which cannot  be  easily  eliminated.
Table  2-5  shows which calendar year  information remained  for
the  comparison of  pre-evaporative  and  post-evaporative  model
years in FaAA's analysis of PARS data.

     This table  shows that for each matched  set  of  model years
there was some difference in the calendar years from which the
data  were  collected.   The  most  pronounced  effect  is  for  the
comparison  of  the  1968   MY  pre-evaporative  vehicles  and  1973
model  year  evaporative  vehicles.    For  example,  information
collected  between  1975-81  was  used for  1968  MY vehicles  in
comparison  to  data  collected  between  1980-86  for   1973  MY
vehicles.   In  other words,  for  vehicles of  like  age,  1968  MY
data was  obtained  a full five years  earlier  than  data  for 1973
MY vehicles.  The offset, while  less  pronounced,  is  three years
for  1972  MY  vehicles and  one year  for  1971 vehicles.   Any
possible  trends  observed in  this  comparison of pre- and post-
evaporative  vehicles  could  be  more due  to  changes   in other
influential  factors  that  occurred between these  very  different
time periods.

     For example,  any number of factors could contribute to the
observed  trends  in FaAA's  results  including  driving  pattern
changes,  weather  differences, vehicle population/concentration,
and  especially the  noticeable  increase  in volatility  as shown
in  Figure  2-1.   This figure shows that  summer fuel volatilities
were  roughly 3/4 psi  RVP higher in  1981-86  than they were in
1975-80.   In  addition,  winter fuel  volatilities  also  averaged
about  1 psi  RVP higher in 1981-86  compared to 1975-80.    Factors
such  as  these make comparisons  of pre-  and post-evaporative
vehicle fire performance  over  long time  periods quite difficult.

     Further,  the  data shown  in Table  2-6  from  the  states of
Michigan  and  Maryland   do  not  show  any conclusive  evidence
regarding the  effect of  age on  fire  rates.*   The Michigan data
tend  to indicate  no  appreciable  effect, whereas the  Maryland
data  do indicate  a  possible  relationship between  vehicle  age
and  fire rates.  Therefore,  given the  uncertainty  of  the true
effect  of  vehicle  age on  fire  rates,   and  the  potential  for
FaAA's  technique  to  introduce  additional  confounding  factors
     Ohio  data were not included in Table  2-6  because the  Ohio
     data  only  date back  to  1982,  and therefore,  not  enough
     information  existed  to  separate  model  year  information
     into  three distinct  age  groups.

-------
                   2-14


                Table  2-5
Calendar Year Data Used in FaAA's Analysis
   of PARS Data  for Model Years  1968-73
 Model Year           Calendar Year Data Used

   1968                       1975-81
   1969                       1975-83
   1970                       1975-85

   1971                       1976-86
   1972                       1978-86
   1973                       1980-86

-------
             Figure 2-1: Gasoline Volatility Trends*
                                                                                                NJ
*     Fran "Motor Gasolines,  Sutmer 1987,"  Cheryl L.  Dickson and




      Paul W.  Woodward, NIPER,  Report 153,  March 1988.

-------
                                       2-16
                                     Table 2-6
      Car Fires per 1000 Accident Involved Vehicles in the States of Michigan
          and Maryland (1978-84) as a  Function of Model Year and Car Age
                      MICHIGAN
                             MARYLAND
Model
Year*
1975
1976
1977
1978

All Aqes
2.4
2.0
1.8
1.9
                       <  6  Years

                         2.4
                         1.8
                         1.7
                         1.8
<  4 Years

    2.4
    1.8
    1.6
    1.7
All Aqes

    3.4
    2.1
    1.5
    1.3
< 6 Years

   2.6
   1.5
   1.2
   1.2
*    Only these  4  model  years  are shown  because  data  for  all
     three  age  classes   were  only  available  for  vehicles  in
     these 4 model years.
<  4 Years

    2.2
    1.2
    0.8
    0.8

-------
                              2-17
such  as  significant  RVP  differences,  the  technique  used  to
correct   for   vehicle  age   effects  appears   to  have   been
unwarranted.

     FaAA's  conclusion  that  evaporative  control  systems  have
increased  fire  rates is  also  based on  a  comparison  of  fire
rates for  eight  specific vehicle models  in immediately adjacent
model years of  1970  and 1971.   However,  such a  small  sample  of
models  in  comparison  to the  entire fleet of models available at
that  time cannot  yield any  significant  results  regarding  the
typical  characteristics  of   the  entire  range  of models.   In
addition,  no mention  was made  of  an adjustment to  account  for
the  fact  that  California  evaporative  control  systems  were
implemented on 1970 model year vehicles.

     In  summary,  FaAA's  conclusion that  evaporative  control
systems  increased  fire  rates  is not supported  by the analysis
presented   in  their  comments.    Even   if  the  analysis  were
reaccomplished   to   address   the   apparent   analytical   bias
introduced by  the matching  scheme,  and  the same  results were
found,  it would  still  be  necessary to  satisfactorily address
the  concerns that  1)  in their  analysis  of PARS data  (and most
of  the states)  all vehicles were  at least more than four years
old,  and  2) the data were gathered over  a  number  of calendar
years  that  did not  overlap,  so  that  other  factors  such  as
volatility could be responsible  for  the trends  observed.

     As a matter of fact, EPA's analysis  of the  data shown in
Table  2-2 was  done without  consideration of age  factors  (i.e.,
it  included all data).   This analysis supports  the conclusion
that no  difference  is   apparent between  the fire rate changes
that occurred  as a result of  implementing evaporative controls
and  those changes which  occur  normally from one  model year to
the  next.   EPA  believes that  FaAA's analysis would have reached
the  same  conclusion  if  all  the  available  data had been used.
All  other  information  available to EPA does not  indicate that
evaporative control  systems  have  had  a  noticeable  effect on
vehicle fire rates.   Some additional information  which  supports
EPA's  view is presented below.

      In addition to trying to  identify trends  in  fire  rate data
that have  resulted  from implementing evaporative controls,  it
would  be useful  in  the  evaluation of   the   safety  of  these
systems to know the  extent to which evaporative control  systems
have been involved  in vehicle  fires.   However,  no  such hard
data exist.  Indeed, one  of  the reasons  why  this  information
does not  exist  is because evaporative control  system  fires  are
extremely  rare  events.    For  example,   the  Center   for  Auto
Safety's  (CAS)  recent  review  of over  20,000 fuel  system related
owner  complaints  maintained  in  NHTSA's  files  revealed  1,501
 fires.[5]  Only six  of these were  even tangentially  related to
the  evaporative  system.   This   is  only  0.4   percent  of  the
 reports involving fire.  A similar  review by a NHTSA contractor

-------
                              2-18
revealed  23 or  less  such incidents  out  of  a  total of  2850
vehicle  fires  involving  fuel  related  components.[2]   Even  if
taken at  its maximum,  this represents only 0.8 percent of owner
complaints  involving fire.   The difference  in  the number  of
evaporative  involved fires between  CAS  and  NHTSA  (6  versus <,
23)  is   because   the   NHTSA   report   presented   only  summary
information  on  fuel   emission   control   related  fires  (code
612X000)  where  evaporative  problems  are  recorded.   However,
this  category  of  complaints  includes many other  problem areas
as well  (e.g.,  crankcase ventilation).   An EPA review  of these
23 owner  complaints  related to fire  in category 0612X indicated
that  10   or   less  involved  the   evaporative   system,   which
represents  only 0.4 percent  of  the  owner  complaints involving
fire.   The rare  occurrence of  fire problems  with evaporative
systems  is  one reason  why  the fire  risk of  onboard control
systems  is  often characterized as unquantifiable.

     In  addition,  other  sources of  information  are available
related   to the  performance  characteristics  of  evaporative
control   systems   which   can   be  used   to  provide  a   further
assessment  of  the relative safety of  these  systems.   The next
section   discusses  the    level   of   safety   that   has  been
demonstrated in-use  by evaporative control systems  as  indicated
by available performance  information.

C.    In-Use Performance  of  Fuel/Evaporative  Emission  Control
      Systems

      The effect of  evaporative  control systems on  vehicle fire
rates  is just  one of  several  useful indicators of  the  relative
safety  of  such  systems.   Another  good  measure of the  safety of
a particular  system is  the  reliability  or  in-use  performance
exhibited by the  system.  A system which  fails  infrequently or
results   in minor  consequences  when   it  does  fail   can   be
considered a relatively safe  system.

      It   is   useful  to  study  the   in-use  performance   of
evaporative control  systems  because  these  systems  are very
similar  to the  proposed onboard refueling control  systems,  and
therefore evaporative  systems provide  a  useful  perspective  on
the  safety concerns related  to onboard  controls.   As a  matter
of fact, onboard  control systems  are more of a modification to
the  current evaporative  system rather than  the addition of  a
whole  entire   new  system.   Onboard  systems   use  components
similar  to those  found on current systems such  as  vapor lines,
clamps,  charcoal  canisters,  and valves.   By studying  how these
components have performed  in-use  in the  past, it is possible to
predict  how similar components  for  onboard control  systems  are
 likely to perform in the future.

      The  in-use  performance  of  a  particular  system is  often
 measured in terms of the past failure  characteristics  of such a
 system.    Three bodies  of information  contain  data  regarding

-------
                              2-19
historical  failures/problems  with vehicle  systems/components.
These  are  NHTSA   recall   campaigns,   manufacturers  technical
service  bulletins,  and  owner  complaints.   Information  in  all
three of  these  areas  is  maintained in computer files  by NHTSA.
Manufacturers  also  maintain  their own  records  of  recalls  and
service bulletins related to  vehicles  they  have  produced.  This
section examines information gathered  from  NHTSA,  motor vehicle
manufacturers,  and other  interested  parties on  the  recalls,
service bulletins,  and owner  complaints  related to evaporative
control systems  as  part  of an overall assessment  of  the in-use
performance of these systems.

     1.    NHTSA Recall Campaigns

     NHTSA  recall  campaigns  are  one  of   the  most   important
sources  of performance  information  since   they  are   the  most
directly  related to the  safety of a system.  EPA  first began to
examine the available safety recall information in  the fall of
1986  as  part  of  an  assessment   of  the safety  of evaporative
control  systems that  appeared  in the  EPA's June 1987  onboard
safety  technical report  (Appendix II).  The analysis  in that
report was  based on file summaries provided by  NHTSA that were
classified  under their  "Fuel  Emission Control"  category.  This
category  covers   recalls   for  all  vehicle  types   (passenger
vehicles,   light-duty trucks,  and  heavy-duty  vehicles),   and
includes  recalls dating  back to 1967.

     As  of November  1986,  only 22 recall  campaigns  (out  of a
total  of  more  than  4200)  appeared  in  NHTSA's   "Fuel  Emission
Control"  category.   A closer examination of these 22 files  for
the onboard safety technical report revealed only 12  cases that
could  even   be  remotely  linked   to   the  evaporative  emission
system.    Examples  of  the  10   unrelated   recalls   range   from
defective PCV  systems to  defective valves  on  diesel trucks  to
cracked  breather tubes on  1967 vehicles.

     NHTSA  reviewed  and   commented   on   the  onboard  safety
technical report  prior  to  its  finalization,  but  provided  no
feedback  regarding the  correctness of  EPA's finding  that  only
12 recalls were found to  be relevant  to   evaporative  emission
controls.  Since   the  time  of  the   report, however,  we  have
received additional  input  from  NHTSA,   the manufacturers,  and
the Center   for  Auto  Safety  (CAS)   regarding  the  appropriate
number  of relevant  recalls.  This section  summarizes this  new
 information,  and updates our  assessment  of  the  number of recall
campaigns pertinent to evaporative emission control systems.

      a.     Additional NHTSA Input

      In October 1987, NHTSA, as  part  of testimony  given in  a
hearing  before   the  House   Subcommittee  on   Oversight   and
 Investigations, presented the  results  of a  supplemental review
 of their recall files.  In this  review, NHTSA examined all fuel

-------
                              2-20


system  recall  campaigns, not  just those  classified under  the
"Fuel Emission  Control" category.   Thus,  by  definition,  this
review  resulted  in the  identification of  additional  recalls
which  had  not  been  previously  classified  under   the  fuel
emission control category.

     In  their  testimony,  NHTSA  characterized  the  additional
recalls  they  had  discovered  as  ones   "that   might  involve
evaporative-related  fuel  systems  or  fuel  systems,"   and  as
such, not  all additional recalls  mentioned  in their  testimony
were relevant  to evaporative control  systems.   NHTSA1s purpose
in  conducting this  subsequent  review  was  to identify recalls
which might  be related  to  future onboard control  systems,  not
necessarily  to  better  characterize  past  evaporative  system
performance.*   Nevertheless,  EPA's  review  of these additional
recalls  resulted  in the addition  of  eight  more  recalls  to the
list  related to  evaporative  control  systems.   Combining these
additional  recalls  with the  original  12   in the  EPA  safety
report   resulted  in  a total  of  20   recalls  which pertain to
evaporative control  systems.  Table  2-7 contains  the 20 recalls
that appear to be relevant to evaporative controls.

     In response  to  a  recent  letter from  EPA on  the recall
issue,[6] NHTSA  suggested  that  determining  a recall's  relevance
to  evaporative control  systems is not  always clearly defined,
and that several  types of  recalls (e.g.,  those involving fires,
exhaust   temperatures,    manifold    vacuum,    or    stalling/
driveability)   may  indirectly  also  involve  the  evaporative
control system.[7]    To  help   illustrate  their  point,  NHTSA
included as part  of  their  letter  the computer file summaries of
300 recalls  involving  fire,  30  related  to  stalling,  and 20
involving exhaust emissions/temperatures.

     EPA has reviewed this  information very  carefully and the
following  points  represent  EPA's  views  on  the  relevance of
these  other types  of  recalls to  evaporative  controls.  First,
no   new  recalls  directly  relevant   to  evaporative   emission
systems were  uncovered.   In  addition,  no   new  recalls   were
discovered  in  which evaporative controls have  had any  adverse
effects  (direct    or  indirect)   on   stalling   or    exhaust
emissions/temperatures.    In   their   letter  NHTSA   noted  their
belief   that   five  recalls  in  particular   concerned  adverse
effects   of   evaporative   emission   controls    on    exhaust
temperatures  and  indicated  that EPA should  take  these   five
recalls  into  consideration   as  evidence  of  the   types of
tangential   problems  that  can  arise  from   a   vapor   recovery
system.[7]   However,  three of  these  recalls  simply involved the
addition of  a  heat shield, with  no   indication  of the  problem
being   a result of  the evaporative  control  system.   The  other
      The issue of the  effect  of  onboard controls on  recalls  is
      addressed in Chapter 3,  Section D.

-------
                                           2-21
                                        Table 2-7
                                Evaporative System Recalls
  NHTSA
Campaign   Model     Vehicle
  Number   Year(s)    Type
                   Number of
                   Affected
                   Vehicles
Description of Problem
70V137     1970     Pass. Veh.  12,550

71V051     1971     Lt. Truck   25,500


72V014     1971     Pass. Veh. 110,614

73V019     1973     Lt. Truck    1,600


75V011     1974-75  Motor Home     197


75V164     1975     Motor Home     496


76V037     1976     Pass. Veh.   2,400



76V126     1976     Pass. Veh.   9.137


78V036     1977-78  Lt. Truck   20,000



78V106     1977-78  Pass. Veh.  10,500

78V145     1973-77  Med. Truck   2,500



78V181     1978-79  Lt. Truck   23,000
                             Emission control hose may interfere with bracket.

                             Fuel/vapor  separator  (part of  evap  system)  may
                             leak.

                             Malfunction in evap system causes stalling.

                             Vent   line  from  tank  to  canister  may  contact
                             exhaust pipe.

                             Saturated   canister   can  discharge  vapor  over
                             exhaust pipe.

                             Fuel   tank  spillage  may  result in  evap  system
                             failure.

                             Misrouted  vapor  return line which  runs from  the
                             fuel  filter/vapor  separator assembly back  to  the
                             fuel  tank.

                             Erroneously installed piping for  the  "check  and
                             cut valve".

                             Blockage   of  tank   vent  system  can   lead   to
                             pressure buildup and  force fuel or vapor  leakage
                             through cracks in  tank.

                             Defective  fuel tank vent valve.        \

                             Liquid gasoline  may  discharge  from  bottom  of
                             canister  because evap  system may  lack  adequate
                             capacity under certain fuel expansion conditions.

                             Obstructed evap  line causes pressure build up in
                             tank.
 79V019
1976-78  Pass. Veh.   17,800   Possibility of kinked evaporative system hose.

-------
                                           2-22
                                    Table  2-7  (cont.)

                                Evaporative System Recalls
  NHTSA
Campaign
  Number
79V032
87V157
Model
Year(s)
           Number of
 Vehicle   Affected
  Type     Vehicles
Description of Problem
1977-79
Pass. Veh.  83.000   Obstructed evap line  causes  pressure  build up in
                     tank.
79V045
79V048
79V212
84V116
87V111
1979
1975-76
1973-78
1985
1984-87
Pass. Veh.
Pass. Veh.
School Bus
Pass. Veh.
Van
(Ambulance)
                      3,700   Misrouted vapor line to canister.

                      51,000   Defective pressure control valve.

                      2,950   Defective liquid/vapor separators.

                      2,385   Improper functioning of vacuum line valve.

                        250   Defective vapor valve grommet on fuel
                              tank.   (This  component was  installed  as part of
                              a  preliminary  attempt  at   correcting  the  fuel
                              expulsion problem of Recall No. 87V113).

1984-88  Pass. Veh.   25,000   Overfilling  of the  fuel  tank can  increase fuel
                              system  pressure to  the  point where  fuel vapors
                              escape  from  the  charcoal   filter  and  cause  an
                              engine compartment fire.

-------
                              2-23
two recalls  concerned  problems with exhaust  emission  controls.
None of  these  recalls  involved the  evaporative control  system
in any  way.   EPA sees no direct  connection between evaporative
control systems and exhaust system temperature problems.

     An examination of the  fire  recalls  revealed a  few  recalls
involving pressurized fuel  systems  (tanks).   Depending on one's
viewpoint  these could  be  viewed as  tangentially  related  to
evaporative  controls.   However,   an  assessment as to  whether  a
recall of  this  type  is  truly related to the  evaporative  system
requires  an  assessment  as to why the fuel tank was essentially
sealed  (which  by design  causes  the  tank to  operate  in-use  at
pressures  above atmospheric) and what caused the problem which
led to  the recall.   This assessment would expose  what actually
led to  the unsafe conditions that caused the recalls  to occur.
First, it  should be noted that fuel  systems  are normally vented
through  a  small  orifice  to assist  in  control  of  evaporative
emissions  and to assist  in  meeting the  rollover requirements of
FMVSS  301.   As  to  the cause  of  the  pressurization  related
recalls,  it  appears  to  EPA that fuel volatility  increases and
high  tank temperatures were  the  most likely  the  cause.   While
evaporative  control  systems and  other  system components  are
sometimes    modified   as   the   recall    fix   to   deal   with
pressurization  problems  resulting  from  high  volatility fuel,
this  is  not due to a  problem with the original  system design
but is necessary to deal  with higher  fuel  volatilities that are
beyond  the manufacturer's control.   In conclusion,  EPA believes
that  A  careful have  evaporative controls  have  not  adversely
affected  the  number  of  fire,   stalling,  or  exhaust emission
related  recalls.

      b.    Manufacturers  Comments

      In  addition  to  information  supplied  by NHTSA,   EPA  also
received views from manufacturers  on recalls  as  part of  their
comments on the NPRM.   Much  of  the information  received  from
manufacturers   was  simply  a replication  of   the  data  received
from  NHTSA.   For  example,  some manufacturers claimed  not  to
have   had  any  evaporative  emission  system  recalls.    Others
pointed  out that  NHTSA  would  be   in  the   best   position  to
determine  the  number   of  recalls  relevant  to  evaporative
emission  controls.    Several  manufacturers   provided   recalls
which  they   believed   were  relevant   to    onboard   refueling
controls, but  did not  distinguish between which files  related
strictly to onboard controls and which  files were  also  relevant
to evaporative control  systems.

      One  manufacturer   (Chrysler)   did  supply  information  on
recalls  directly  concerning the  evaporative  control  system.
Two  of  the  three   recalls  they   provided   had  already  been
 included  in EPA's earlier  review  of NHTSA's  files.    However,
the third recall  involved  a blocked vent on  1973  Dodge  trucks
 and had not been  identified prior  to Chrysler's comments.   This

-------
                              2-24
recall  was  not  found  in  NHTSA's  computer files,  nor was  it
identified by  the Center  for Auto  Safety in  their review  of
NHTSA's  recall  files.[5]   There existed  some doubt  regarding
whether  or  not  this  third  item  was   an   official   recall.
Information  received  in  response  to  a   June  9,  1988  phone
conversation with Chrysler  indicated that  this recall  had been
sent  to NHTSA,  but  that it had been classified  as a non-safety
related  defect.[8]   Thus,   it may  have simply  ended  up  as  a
technical  service bulletin  in  NHTSA's Files.   Based  on this
information this action was not classified as a recall.

      c.    Center for Auto Safety Report[5]

      The  Center  for  Auto  Safety  (CAS)   recently completed  a
study of the  safety  of  evaporative  control  systems   and  the
effects of recent increases  in gasoline volatility.   This study
was  based on a comprehensive review of  two NHTSA computer data
bases:   recall  campaigns   and  owner   complaint   records.   CAS
concluded  that  21  recalls  over the  past eighteen  years have
involved  the evaporative control system.  Of  these  21   recalls,
no  new  additional  recalls  were presented  which had  not been
identified  previously  by  EPA  or  others.   Of  the  21 recalls
identified  by  CAS,  only two  were associated with the  potential
for  post-crash fires.  Further,  no fires, injuries,  or  deaths
were  linked to  either  of  these  two  recalls.   CAS's overall
conclusion  is  that  "evaporative  emission controls   have  not
resulted  in  any significant  incidence  of  vehicle  fires  or
recalls."

      d.    Summary of Evaporative System Recalls

      After   careful   review  of  all   available   information
regarding evaporative emission control system recalls,  EPA has
determined  that  a   total   of   20  evaporative   system related
recalls (shown  in  Table 2-7)   have  occurred since  evaporative
control systems were  first  implemented eighteen  years ago  (an
average of  about one  recall per  year).   The  distribution by
model year  is  shown in Figure 2-2.  These 20  recalls  cover  all
vehicle types  including  passenger vehicles  (11),  light truck
 (4),  motor  homes (2), buses  (1),  and  other heavy-duty gasoline
vehicles (2).    These 20  recalls which involved about  470,000
vehicles represent  about  1/2 percent of  the more  than 4,200
recalls that  have occurred since 1966.  They  have affected  less
than 0.5 percent of  the   vehicles certified  with  evaporative
systems since  1971.   In other words,  problems with  evaporative
control  systems  account  for  less  than   1/2  percent  of   all
safety-related problems.   While  one might  argue  that a few  more
or  a few less  recalls should be   included,  this would not  change
the conclusion that they  represent only a very small proportion
of  safety recalls.

-------
                              Figure 2-2: Evaporative System Recalls*
4- -
3 -
r,j
in
     ~7\
       ^
           N
        ^
^
2
                                 1
                               x
^
                                        N
                                        1/5
2
                                               N:
                            1975
                                   Model Year
                                                1980
^
                                                                                                          K)
                                                  1985
                                 Pass •
                                 vehicle
                                                          Other**
            Fractional nunbers of recalls involve vehicles from more than one model year.   For
            example, a recall covering 3 model years was represented as  1/3 recall per model
       **   Includes school buses, motor homes, vans,  and heavy trucks.

-------
                              2-26
     2.     Manufacturer's Service Bulletins

     Another  important  source  of  information  regarding  the
performance  characteristics  of  a  system  is  the  number  of
technical   service   bulletins   issued   to  correct   problems/
defects.  Technical  service bulletins are  less  directly linked
to safety  than recalls because  the  problems  addressed  through
service  bulletins  are typically minor performance  adjustments
with  either minor  or no  safety  ramifications.   Indeed,  the
technical  service  bulletin file includes  a wide  assortment of
other  categories of  information  such as  product  improvement
notices,   warranty   information,    service   newsletters,   and
emission recalls.  Nevertheless, the  study of  service bulletins
does   provide   some  useful   perspective   on  the   failure
characteristics of vehicle systems.

     As  with  recalls,  EPA  first  began  to   examine  service
bulletins  back  in  late  1986  as part  of   an  assessment  of the
reliability  of  evaporative control  systems.    After  examining
NHTSA's  "Fuel  Emission Control" category for  the onboard safety
technical report (Appendix  II),  EPA  found  only 21 cases, out of
more  than  800,  that involved  the evaporative emission control
system.   Since the  time of  that report,  we have received  some
additional  information  from  NHTSA to  help refine  our initial
estimate   of  the   number  of   evaporative   control   service
bulletins.   (Note:    the CAS  study   [5]  did  not  cover service
bulletins,  and  manufacturers  did  not  provide  comments   which
specifically   address the  number  of  bulletins  relevant to
evaporative controls.)

      a.     NHTSA's  Input

      It should  first be   stated  that  as  with  recalls   NHTSA
reviewed and  commented  on the  onboard  safety  technical  report
but  did  not  address the  accuracy   of  the  number  of service
bulletins   EPA  found to  be  relevant  to  evaporative control
systems.    Since that  time,   however,  EPA  has  received  two
additional  inputs  from  NHTSA  relevant to evaporative control
service bulletins.   These are NHTSA's April  12, 1988 letter to
EPA,[7]  and   NHTSA's draft  contract  report   "Study  of   Motor
Vehicle Fires."[2]

      In their April  12,   1988  letter,  NHTSA made the same  point
about service bulletins  as they did about recalls;  that is, the
role of evaporative control systems  in contributing to  the need
for  a  service bulletin  is not  always  clearly  defined.   NHTSA
believes that it is possible for the  evaporative control  system
to indirectly  lead  to service  bulletins  for  problems  involving
fuel systems, carburetors, exhaust  systems,  and  other  emission
systems.   To  help   illustrate  their point,  NHTSA included all
service  bulletins   related  to  fuel,  exhaust,   and   emission
systems  as part  of  an  attachment  their  letter[7]  (over  6800
bulletins in total).

-------
                              2-27
     As far as any connection  with evaporative emission control
systems, NHTSA  mentioned a  total of  17 bulletins  out of  the
more  than  6800  provided  as  being  relevant.   Ten  of  these
bulletins   concerned    purge     interaction    with    exhaust
temperatures,  and  seven involved canister overloading  effects
on  stalling/driveability.    To   get  a   better  idea   of  how
evaporative control  systems  have affected  other systems,  EPA
examined the 5,900 bulletins out  of  the total  6,800  which were
classified  under   categories   other   than   "Fuel   Emission
Control."  (We concentrated on these  bulletins first,  since we
had  previously  already examined the  bulletins  in the  "Fuel
Emission Control" category.)   EPA found approximately 35  to 45
bulletins  somewhat  relevant to  the evaporative  control system
which  had been  classified under  categories  other   than  "Fuel
Emission  Control"(  ten of the bulletins are  duplicates of the
same  problem  on  a  different  model  of  the same manufacturers
vehicles).   EPA  does  not  know  the  extent  to  which  these
bulletins  overlap  the  17  identified  by  NHTSA since NHTSA did
not  provide bulletin   numbers  for each of  the  17  items  they
discussed.  These 35  to 45 bulletins were not  included  as  part
of  the bulletins  discussed in the safety report  since the 5,900
bulletins  in categories other  than "Fuel Emission Control"  were
not available at the time of that  report.

     As   far   as  service   bulletins   related   to   evaporative
emission  control systems that  fall under NHTSA's "Fuel  Emission
Control"  category,  EPA decided  to reexamine   this  issue after
receiving  NHTSA's   draft   contract   report,    "Study  of  Motor
Vehicle  Fires."[2]   This  report  identified  882  "Fuel  Emission
Control"  bulletins in  NHTSA's  computer  files  and implied by the
title  of  this  category that  all 882  bulletins  involved the
evaporative    control    system.     Since    this    number   was
substantially   different   from   the    21    files   identified
previously, EPA decided to  review this  category once again.

      A careful  review of  this  category reveals   that   "Fuel
Emission   Control"    includes   many    bulletins   other    than
evaporative emission control including exhaust emissions,  spark
knock, driveability,   high  altitude  emission  standards,  glow
plugs,  oil   consumption,   oxygen sensors,   EGR  valves,  PCV
systems,   and many  others.   In   addition,  many entries in  this
category  contain no  summary  information and  often  are missing
model   year  information making   it  essentially  impossible  to
determine whether   the  entry   is  applicable  to   evaporative
emission  controls.   These  two types of bulletins  ("other," and
 "no information") comprise  over   90 percent  of the  882 entries
 in the "Fuel Emission  Control" category.

      As was mentioned before,  it should also  be noted  that not
 all  entries  in  this  category   are  really   technical  service
bulletins.  Some are  service  newsletters,  warranty  information,
product  improvement  notices,  parts   notices,  service  manual

-------
                              2-28
corrections,  and even  "recalls."   Further,  some of  the  entries
are for the same problem but on different model vehicles or are
updates to prior bulletins.   Needless  to say,   it was  difficult
to identify with certainty  the actual  number of bulletins truly
related  to  evaporative   control   systems  and  perhaps  more
importantly the  actual number  of  separate problems which were
really identified in the bulletins.

     From  the  service   bulletins  that  did   provide   enough
information, it appears  as  though  35 to 75 could be relevant to
the  evaporative  emission  control  system.   This  estimate  is
higher than EPA's  original  estimate.  It includes problems such
as broken  or  incorrect  canisters, poor  purge  characteristics,
vapor    line    problems,    and   tank   vent/overpressurization
problems.  This is primarily  because our most  recent review was
quite  liberal  in  the acceptance  of a  relevant  bulletin.  The
range  of  35 to  75 was  used because  there  are  some  bulletins
which  are  possibly related, but could not be identified  as such
absolutely  because of  the quality  of  the  description  of the
problem.   Even with  this being the case,  adding the 35 to 45
evaporative system bulletins  revealed  in other NHTSA categories
brings the total  to 70 to  120.  It thus  appears that evaporative
control  systems  account for only  about  0.1  percent   of the
88,000  service bulletins  issued.    Because  these problems have
been  so infrequent,  identifying  the  exact  number  of   service
bulletins   related  to   evaporative  control   systems   is  not
critical.

      3.    Owner Complaints

      The   third  data  set  which  contains  information   on the
performance characteristics of vehicle  systems  is NHTSA's files
of owner  complaints.   When  this  information  was  analyzed for
the  1987 onboard  safety report, only about 100  complaints were
found which related  to  the  evaporative control  system.   Since
the   time   of   that   report,  we  have   received  two  additional
sources  of information  regarding  owner  complaints,  neither  of
which altered  our  original  findings.

      The  first piece  of  additional   information  appeared  in
NHTSA's   draft  contract   report,  "Study   of  Motor   Vehicle
Fires."[2]  This  report showed 447 cases  of  owner  complaints
 filed under the  "Fuel Emission Control"  category,  23  of which
mentioned  fire.  However,  as was  true of NHTSA's "Fuel  Emission
Control"   category  for  service bulletins,  our  review  of  the
detailed  information  concluded  that  not  all owner  complaints
 filed under this  category are  relevant to  evaporative  control
 systems.   For  example,  many of the complaints in  this  category
 related to PCV systems,  air pumps, exhaust emissions,  EGR, etc.

-------
                              2-29
     This category  was  originally searched as part  of  the 1987
safety report.  A second, more  recent,  review revealed roughly
the same  findings  as discussed in that report.   In  total, only
about  100  complaints out  of the  447  listed  in this  category
could  be  even remotely  linked  to evaporative control  systems.
Relevant   complaints    included    canister   problems,    tank
venting/overpressurization,   vapor  line,   and  purge  problems.
These  100  complaints represent less  than  0.05  percent of  all
complaints  (210,000)  in NHTSA's  computer files.  Out  of  these
100, less than 10 made mention of fire.

     This complaint  level is similar  to that determined by the
recent work  completed by the Center  for  Auto  Safety (CAS).  In
their  study,[5]  CAS searched  through  owner  complaints  filed
since  1977,  and found  only 27  cases involving  the evaporative
control  system.    Again, as was  true  of  service  bulletins,
because  there have been so  few complaints,  it  is not  essential
to  focus  on  the  exact number of relevant  complaints.   Of  these
27  cases, only 6 were found  which involved fire.

     In   summary,   of   the   hundreds  of  thousands   of  owner
complaints filed with NHTSA, only a  minute percentage (<0.05%)
have  involved the  evaporative  control  system.   Further,   only  a
handful  of  these   involved  fire.    When  these   complaints  are
considered  relative to  the  130 million plus vehicles  sold with
evaporative   control  systems,   it   appears  that  evaporative
control  systems  have not created a significant  number  of  safety
problems  for vehicle owners.

     4.    Conclusions

     The basic conclusion that  can  be drawn from an examination
of  the  recall  files,  service  bulletins,  and  owner complaints
provided  by NHTSA  and others  is  that  evaporative control
systems   have  worked   safely.    Evaporative   control  systems
account   for only  a  tiny  fraction  of  the   total  problems
appearing in NHTSA's  computer  files. Therefore, while problems
have  occurred,   they  have  been  infrequent.   Further,  when
problems  have  occurred,  the  consequences have  been minimal.
NHTSA's  records  indicate   that  no  deaths or  serious  injuries
have  resulted   from  an evaporative  control   system   failure.
Owner  complaints taken  at   face value  would indicate  less  than
 10  fires since NHTSA began keeping  records in  1977; this is  at
 a  rate   of  less   than   one   per  year.    This   information
demonstrates the effectiveness  of  NHTSA's programs to not only
prevent  system  failures, but to correct  any  infrequent mistakes
that do  occur before they lead to any significant consequences.

      Another important  result  of  this  examination of  NHTSA's
 files is  that  most  of  the evaporative system  recalls involved
 pre-1980  vehicles.   This   can  be  seen  in Figure 2-3.   This
 indicates that  as  manufacturer's  experience   with  evaporative

-------
                     Figure 2-3: Evaporative System Recalls*
130 -
1 1O
too -
90 -
x 8O -
f-H
i 7° '
1 [ IT^
cL S3
8 B 50 -
0 g
"•§ 40 -
"8 30-
1 20 -
10 -
o -






1
1
^/
//
* s
V
x\\\\\\\\\







i i i i i |
1970 1375

EZ1







•







X
\\x\\\x

^
x\\\\\\


V ~>
\
'/ rTiP/lrTiryi r7
i i i 1 . • • i i i • .
	 1980 1965
Model Year
Pass. iv \i Light Y///A Other**
Vehicle * ^ * Truck *****
                                                                                                          K)
                                                                                                          I
                                                                                                          U>
                                                                                                          O
**
involved in the recall were distributed equally among  the number of model year involved.


Includes school buses, notor hones, vans,  and heavy trucks.

-------
                              2-31


systems  has  increased,  so  has  their  ability to  design  and
produce  problem   free   systems.    Further,   because  of   the
similarity  between  onboard  and  evaporative  control  systems,
this experience should be directly applicable  in the  design and
development  of  safe,  effective,  reliable  refueling  control
systems.

     Finally,  it  is  quite  possible  that  evaporative  controls
may  have,  resulted  in  safer  fuel  systems  then  would  have
occurred without such  requirements.   The basic  requirement for
a  closed fuel system may  have contributed to  the avoidance of
numerous  fuel  leaks.   The  requirement that  carburetors  vent
their  fuel  vapors  to  a  canister  instead  of  allowing  the
continued   direct   venting  of   such   vapors   in  the  engine
compartment   may    have    avoided    numerous    engine   fires.
Unfortunately,  it  is  impossible to   quantify  such  potential
benefits.

D.   Summary

     As  was  stated  at the   beginning of  this  chapter,  auto
manufacturers  and  others  have  asserted  that onboard control
systems  would lead  to  an  unquantifiable increase  in fire  risk.
To  address  this concern,  we first attempted to  put  the current
fuel/evaporative system risk  in perspective.   An  examination of
fire   rates   in  accidents  revealed  that  evaporative control
systems  have  had   no  discernible  effect  on  vehicle  fires.
Further,  an  examination  of  past  recall  campaigns,  service
bulletins,   and  owner  complaints   reveals   that   evaporative
control  systems have created very few  problems,  and of the rare
problems that did  occur,  no serious consequences  resulted.  The
basic  conclusion  that  can be drawn from  this  information is
that  evaporative  control  systems have  worked safely  and  well.
The past  safety  performance  of  evaporative   systems  strongly
suggests that an  onboard  system which  incorporates  evaporative
system design concepts can  be made to be safe,  effective and
reliable.

-------
                              2-32
                    References for Chapter 2
     1.    "Analysis of Motor Vehicle Fuel  Tank-Related Fires,"
K. Steilen, U.S.  EPA,  AA-SDSB-88-05,  February 1988.

     2.    "Study of Motor Vehicle Fires -  DRAFT,"  prepared for
NHTSA by Data Link, Inc.,  Washington, D.C.,  February 1988.

     3.    "Evaluation of Federal Motor  Vehicle Safety Standard
301-75, Fuel System Integrity:  Passenger Cars," DOT HS-806-335,
January 1983.

     4.    "Fires  in  Motor  Vehicle  Accidents," UM-HSRI-SA-74-3,
Peter Cooley, April 1974.

     5.    "Stopping  Vehicle  Fires  and  Reducing  Evaporative
Emissions:  The  Need  to  Control  Gasoline  and Alcohol  Blend
Volatiltiy," Center for Auto Safety, March 1988.

     6.    Letter  to  George  Parker,  NHTSA,   from  Chester  J.
France, U.S. EPA, January 22, 1988.

     7.    Letter  to  Chester  J.  France, U.S.  EPA,  from George
L. Parker, NHTSA, April 12,  1988.

     8.    Conversation  with  Jim  Furlong,   Chrysler,   June  9,
1988.

-------
                           CHAPTER 3

                Onboard System Design and Safety


A.   Summary of Technology and Safety Comments

     1.     Introduction

     As  was  discussed  in  Chapter  1,  many parties  submitted
comments  concerning  the  technology   and   safety  aspects  of
onboard  vapor  recovery systems.  Most  of the comments  in these
two areas were  provided by  either auto  or  petroleum interests,
but  as  was  mentioned  in  Chapter  1,  important comments  and
information were provided by others as well.

     With a  few exceptions,  the technology  and  safety  comments
tend  to fall  into four  general  areas:  1) general views and
concerns  regarding onboard  system  technology  and  safety,  2)
specific onboard  control designs  proposed by the commenters, 3)
specific  technology   and  safety  concerns  regarding  onboard
system  designs  and hardware,  and 4) specific concerns about the
in-use  safety implications  of  reguiring  onboard controls.   The
comments  received in each of  these  areas are summarized below.
Since   there   is   an   inherent   relationship  between   the
technology/design   and   the   safety  of  onboard  systems,  the
related comments  and  concerns will be  summarized together when
appropriate.   Other comments  which fall outside this approach
are summarized  and addressed  in Section D (below).

      2.    General Views on Onboard Technology and Safety

      A  common assertion made  by  automotive  industry commenters
is  that onboard  systems are not a  simple   extension of current
evaporative  control systems  and that  it may not be technically
feasible to  meet  EPA's proposed refueling emission standard.
They  argue  that EPA cannot base the  feasibility of onboard on  a
few  prototype  systems  because the  complexity  of these  systems
will  naturally  develop  and  increase  as the  testing  and progress
towards  a   complete   vehicle-based   system  continues.   The
petroleum   industry  and the  California   Air   Resources   Board
relayed the  opposite  opinion  in their  comments,  stating  that
onboard is a  simple  extension of current   evaporative systems.
API   has   built   several   vehicle   based  prototype   systems
 (described   later  in   Section  C)   and    provided  documented
refueling  test data  from refueling tests  run  on these  systems
which  they  feel  demonstrates  the   feasibility  of   onboard
controls.

      One  reason  the   automotive  industry  commenters   do not
 consider  onboard  to  be   a  simple   extension  of   current
 evaporative  systems   is because  they  maintain  that   onboard
designs   must   have    separate    canisters   for   controlling

-------
                              3-2
evaporative and refueling emissions.   In the commenters'  view,
these  "separate"  systems  are much less  an extension of  current
evaporative systems than the "integrated" systems which  use  one
large  canister  to  capture   both   evaporative  and  refueling
emissions.  Unlike  integrated systems,  separate systems  would
have  a  greater  number  of  components  than  most  evaporative
systems.   Both  EPA's  and  API's   prototype   designs   are   of
integrated systems.

     Those commenters  who chose separate  instead of integrated
system  designs gave  several   reasons  for their choice.   Even
though  integrated  systems  are  less  complex,  and  therefore
considered more safe  by  many of  the  manufacturers who  chose
separate  systems,  they cited  perceived  drawbacks of integrated
systems which drove them to  focus  on separate  systems.   Many
said  that the  test procedure favors  a separate  system.   Most
said  that they expect  only  a short amount  of leadtime  to be
allotted  before onboard  systems  would be required  which would
force  them to  design  "retrofitted"  separate  systems.   Some said
that  they could use a  "clean  sheet" approach for later designs
and   then  may   consider  integrated   systems.    Almost  all
commenters  said  that  packaging  and  purging  the  single large
canister  of   an  integrated  system  would be  difficult,   if  not
impossible,   especially   on   smaller  cars   and   carbureted
vehicles.  On the other hand one commenter  chose an integrated
system, saying that it would be advantageous from both  a  safety
and  a  purge design  viewpoint (IV-D-340).*

     Therefore, presumably based on this view of system design,
a  widespread opinion  of  automotive industry commenters is  that
onboard  vapor  recovery  systems would be   more  complex  than
current  evaporative   control  systems,  and  that   adding  this
complexity   to   a  vehicle   fuel   system   would   cause    an
unquantifiable increase in  the risk of  both  crash and non-crash
fires.   Several complexity  and risk factors  were  cited such  as
larger  components,  more  components,  more   connections,  the
locations  and  materials  chosen   for   these  components  and
possible  effects on other  systems.  For example, complexity  is
said to be added by  increasing  the size of  components such  as
the  vapor line diameter  and carbon  canister.   Larger  components
were  thought  to  increase  crash  risks because they  are  more
vulnerable  to  damage  and  could  damage  the  fuel  tank  in  an
accident  if  located nearby.   Along the same  lines,  commenters
said  that  since  onboard  refueling  control  systems  handle a
larger amount  of  vapor than  evaporative  systems there must  be
an increased  degree  of  both  crash  and non-crash  risks. Also,
commenters   said  that   onboard  systems   would  require  more
components  and more  connections  than an  evaporative  system,
such as  a pressure relief valve  and a  liquid/vapor  separator.
      References  such  as  this   indicate  the  entry  in  public
      docket A-87-11 wherein the  document/comment can be found.

-------
                              3-3
These were  thought  to increase  risk  by providing more  targets
in accidents and a greater chance  of  disconnections.   Non-crash
risks  were  said  to  be  greater because  a  greater  number  of
components increases the  chances of manufacturing missassembly,
leaks,  and  other   failure  modes.   Commenters  said  that  some
onboard components are likely to be put  in  risky locations such
as  a  crash  zone,   near  the  fuel   tank   or  under  the  hood.
Canisters and fillneck mounted vent valves  were  common examples
of such components.   The  location  of  onboard components is also
said  to be  a non-crash concern  since some  likely  locations are
susceptible   to  tampering   and  other  failure  modes.    The
possibility that some components would  be plastic was mentioned
as  a  safety  concern  because  they  might  have  questionable
strength.  Also, according to NTSB, a static charge  could build
up  if metal components  isolated by plastic  components  are not
properly  grounded.   Finally,  one  commenter suggested  that the
complexity   issue  also   included  concerns  regarding  indirect
effects  of  onboard  controls  on other  vehicle  systems through
impacts   in  areas  such  as  manifold  vacuum,  exhaust  system
temperature, and driveability.

      3.     Specific Onboard System Designs

      A  total   of  26  different onboard   system  designs  were
submitted by  commenters.  Diagrams  of the  designs  that were
provided  in the comments are shown  in Appendix I.   As  can be
seen  in  the  appendix,   a  mechanically sealed  fillneck  was a
common  control  system  feature  found  in  the  onboard designs
submitted  by   auto   industry commenters.    Separate   activated
carbon  canisters  for  storing  the  evaporative  and   refueling
vapors  were  also  frequently  incorporated  into manufacuturers'
designs.   At least  two  commenters  indicated that they would  use
liquid    seals   (IV-D-08,IV-D-363).     There  was   no  general
agreement on the choice  of rollover/vent  valves.   Some designs
used   a  fillneck  mounted   rollover/vent   valve  while   others
suggested either  a  mechanical or solenoid-actuated tank mounted
valve.    Most   designs   involved  a   rear-mounted canister  for
reasons  discussed   later  in   Section  D   regarding   canister
safety.   Also, most  designs  submitted incorporate a  component
that  would  act  as  a  liquid/vapor  separator  which  separates
entrained liquid fuel droplets  from  refueling vapors  enroute to
the canister.

      A few  commenters discussed  methods  other  than  a  carbon
canister   based  system  for  controlling   refueling   emissions
onboard  a  vehicle.   Installing a vapor  condenser  or  a  vapor
combustor onboard  the vehicle  are  two examples  of  alternative
control  methods.   Generally,  commenters   concluded  that  these
 alternative  systems   are more  expensive  than  canister  based
 systems and that the  level of control possible is not  superior
to a canister  system.  Nevertheless,  a few manufacturers stated
 that  they  are  continuing  to  investigate  these  alternative
 systems.   Several commenters have also investigated  collapsible
 fuel bladders  as  a control  strategy.   Some problems  cited  are

-------
                              3-4
that  no  potential  bladder  material  has  been  proven  to  be
durable  and  compatible  with  all possible  in-use  conditions.
Commenters  think  that  air  entrainment  and  fuel  permeation
through  the bladder material  may be significant  problems.   It
was generally  thought  that current  bladder  technology can  not
presently   provide   control   independently   from   a   canister
system.   However,  several companies  are  continuing  research
with the  hope  that  bladders will be a viable long-term solution
to refueling controls.

     4.     Specific  Technology  and  Safety   Concerns-Design/
            Hardware

     In  addition  to the  general views on onboard technology and
safety and  the onboard system  designs  received in the comments,
a  number of  comments  were received  regarding specific onboard
system  design options  and hardware.   These  comments will  be
summarized  by grouping  them together according to  the specific
component  of  the  onboard system  to which  they  apply.   The
components  will  be  discussed in order, starting at  the fillneck
and  working  through the  entire  onboard system  to  the purge
valve.

      a.     Fillneck Vapor  Seals

      Despite  EPA's  view that  liquid seals would be preferable,
a  number of commenter' s designs used mechanical  seals.   Only  a
few   automotive  manufacturers   used   liquid  seals  in   their
designs.    BMW and  Toyota  said  they were  still  considering
liquid  seals   (IV-D-342,   IV-D-363).   Many  commenters   simply
assumed liquid seals, whether  J-tube or  flowing type, would  not
perform  adequately to  pass  the proposed  refueling  emissions
standard.  From the comments,  this  appears  to be an  assumption
since  essentially   no   actual  liquid  seal  testing   data  was
provided.  On the  other hand,  Exxon, Mobil,  and  EPA all  have
developed  onboard   systems using   liquid  seals  and  performed
tests which showed  that liquid  seals functioned  effectively  and
could reduce  refueling  emissions to  levels below  the  proposed
 refueling emission  standard (IV-D-360 (a),  IV-D-320).

      Commenters gave several  reasons why liquid seals would  be
problematic and  may not be a viable option.   For  example,  they
 said  that  evaporation  at   the   fuel/air   interface,   vapor
 diffusion  out  of  the  open fillneck,  or a  plugged  drain  hole
 could cause a  liquid  seal onboard  system  to fail  the  proposed
 refueling  emissions  test.  They  were  also concerned  about  the
 ability  to  form  a  complete   seal  at   low  flow  rates.   One
 manufacturer  said  that  tests  on  their   flowing  seal  system
 indicated  a need for dispensing rate limits of  6-10 gallons  per
 minute  (IV-D-363).   Conversely,  the testing results reported by
 Exxon  showed  that  liquid seals  functioned  adequately   at  low
 flow  rates and  another  commenter  stated  that  flowing  liquid
 seals  can  prevent vapors from  escaping,  since  the  inflow  of
 fuel being dispensed forms a vacuum in the fillneck (IV-D-01).

-------
                              3-5
     Commenters predicted that the  backpressure in  liquid  seal
systems  would  increase  spitback  spills  and  premature  nozzle
shut off.  They said this would  be a  safety  hazard unless  an
anti-spitback  or  anti-surge feature  is incorporated in  liquid
seal systems.   Very little test  data  were provided to  support
this belief  but one manufacturer reported spitback  on  a  liquid
seal  system   while  dispensing   at  8.5  gallons   per   minute
(IV-D-08).   On the  other  hand,   Mobil  was able  to fill  their
prototype onboard equipped vehicle  to  automatic  shutoff with no
spitback  (IV-D-320),   Chrysler  and GM  recognized  that  spitback
and  anti-spitback  valves   are  not a  new  problem  or  a  new
solution,  since  they  currently  use   caged-ball  valves  in the
fillnecks  of  some  of  their  vehicles.   General  Motors  did
limited  testing on sleeve seals  and  accumulators  and concluded
that neither  device  demonstrated  acceptable  performance at this
point but both provided fuel splashback control.

     Commenters  mentioned  that   liquid seal  systems would be
harder to  package than mechanical  seal systems.   Fitting the J
shaped  end  of a  J-tube seal  into a  shallow tank  (which was.
pointed  to as the  shape of most  future  tanks) is  a packaging
concern  as  is  attaching the  J   insert  to plastic  tanks.   The
possibility  that  this J-shaped end  could puncture  the tank  in a
crash  was raised up as  a  safety  concern.   Commenters also  said
that the  refueling  canister on a  J-tube system would be  larger,
and  therefore  harder  to package,  than a canister  sized for a
mechanical   seal  system.   They   said  that  a  J-tube  system
canister  would need to be  larger due to greater  entrainment,
turbulence and splashing which generates  more vapors.    General
Motors  provided  data  which show that  liquid seals caused the
generation of up to 50 percent more vapor then mechanical  seal
systems  (IV-D-360).

     Another  packaging  drawback  raised  is   that  liquid  seals
require  a certain  fill  height in  order  to contain a standing
column  of  fuel  backed up   in   the   fillneck due to   system
backpressure.  No   actual  test   data  relating   the   necessary
fillpipe height  during dispensing to  the  system  backpressure
were provided, although some simple theoretical calculations  of
standing  liquid  height were  performed.   Some commenters  said
that the elimination of the  external  vent  line  on liquid  seal
designs  contributes  to increased  backpressure.   Providing  the
necessary   fill   height  would   necessitate  • redesigning   the
fillneck and side  body panels  on some vehicles,  according  to
commenters.

     Even though  most manufacturers  prefer  mechanical  seal
systems  to  liquid  seal  systems,  they  also raised  concerns about
the technology and  safety of mechanical  seals.  Many  said that
no  material  exists  with  the   required   durability  and  cold
temperature   performance for  this  application.  Most  commenters
 also  said that  nozzle standards  (diameter,  length)  would  be

-------
                              3-6
necessary  in  order  to   ensure   nozzle  and  mechanical  seal
compatibility.   One commenter  noted  significant  damage  to  a
prototype mechanical seal  after  50 refuelings due to burred and
nicked  nozzles  (IV-D-362).   For  this  reason,  many  commenters
feel that  nozzle inspection  and  enforcement  would be necessary
if  mechanical  seal systems  are  used.   They  said  that  any
difficulty  in  inserting  the nozzle  would  lead  to  increased
tampering rates.

     Many  commenters  expressed  concern  about  the  safety  of
mechanical seal  systems.   They  said that an overpressure relief
device  would  be necessary  in case  of nozzle  shutoff  failure.
Some  said  that  if  the  pressure  relief device fails,  tank
overpressurization   could   lead   to   canister   flooding.   One
commenter  even  said that fuel tanks may need to be strengthened
in  order  to  remain  sealed  during   this  potential  excessive
pressure buildup (IV-D-342).

     b.    Refueling Vent Valves

     All  commenters recognized  the need  for a multifunctional
refueling  vent valve.  Most said  that  this valve would need to:
1)  provide rollover protection,  2)  vent adequately during both
refueling  and normal operation,  and  3) act  as  a  fill limiter.
Commenters cited  several  optional designs  and  locations for
this valve,  including  a fillneck mounted, mechanically actuated
valve  and a  tank mounted, solenoid  valve.   Several commenters
said they  were  concerned about  the safety of a fillneck  mounted
valve,  stating that body  panel modifications would be required
since   it  is  in a  crash  zone.   Nissan  and  other  commenters
expressed  concern over fillneck mounted valves  since they could
be  located in a crash zone  (IV-D-452).   There is added  concern
if  the valve is plastic,  since it could be more  susceptible  to
damage  in  a  crash.   Plastic  parts  also  necessitate  careful
grounding  of adjacent metal  components  to  prevent static charge
buildup.   Some  commenters   said  that  having  an  electrically
actuated  solenoid valve near the  fuel/tank could  be  a  potential
safety  hazard  due  to  the  proximity  of  electrical  relays  to
ignitable  liquid or vapor  fuel.

      c.   Vapor Lines

      A  larger  vapor  line diameter  is required  on an  onboard
system  than  on an  evaporative  system in  order  to  route  the
higher  flow rate of vapors  from the tank to the canister during
 refueling.  Several commenters made estimates  and/or  calculated
the necessary size of the  refueling  vapor line leading from the
 fuel  tank to the canister.   One manufacturer stated that their
vehicles   would  require a  line with  0.55 -  0.71 inches  inner
diameter   compared  to their  current  0.47  -  0.55 inches  inner
 diameter   evaporative   lines   (IV-D-377).    Others   estimated
 0.625 - 0.75 inches  and  0.98  inches  inner  diameter  (IV-D-01,

-------
                              3-7
IV-D-367).   Many  noted that  a rear-mounted  canister helps  to
minimize the required diameter  and  length of the refueling vent
line  since  less   system  backpressure  is  generated.   Various
vapor line materials were  suggested such as terne plated steel,
nylon and Buna-N rubber.

     Several  concerns  were  expressed   about   the   safety  of
refueling  vent lines.   Many  commenters said  that  the  larger
diameter  of refueling  system  vent lines  make  them a  larger
target  in  an accident  than evaporative system vent lines.  They
are more prone to punctures since more  surface  area  is exposed.
Commenters  also  claimed  that  these  lines  could get pinched,
increasing pressure during  both normal  operation and refueling
and  leading to more  frequent spitback.  Condensation in these
lines could also  increase back pressure and lead to spitback.

     d.    Liquid/Vapor Separator

     Most  commenters  suggested  that  some  component  on  an
onboard vapor  recovery system  must act to  separate  liquid fuel
droplets   from  the  refueling  vapors  before  they   reach  the
canister  so that  the  adsorptive  capacity of the carbon is not
decreased.   Most commenters'  designs   show  the  liquid/vapor
separator  as  a  separate component which forces  liquid out of
the vapor  by gravity, filtering, or inertial impaction.

     Commenters   gave  several  reasons  as  to  why  there  is an
increased  chance of  liquid  fuel   reaching  the  canister  (and
hence  a greater  need  for  separator)  with  an onboard refueling
vapor   recovery   system  than with  current  evaporative  control
systems.   One  reason is increased  tank  splash  due to turbulent
fuel   flow  in a sealed  system.    Several  commenters   thought
J-tube   seals  could  have  the effect  of   aiming  liquid  fuel
directly  at  the vapor  line  during   the  filling  process.   A
second  reason given  is  that vapor  condensation is  more  likely
in a larger diameter vapor  line.   One commenter also said  that
a  liquid/vapor   separator   is  necessary  if   the  canister  is
mounted below  the fuel tank  level,  or  if  the vapor lines  are
sloped    (IV-D-342).     Another    commenter   stated   that    a
liquid/vapor   separator  could  be   a safety hazard since it  is
virtually a miniature  fuel  tank with all of the  same associated
safety problems,  yet  no  evidence  was  given  of problems  with
fuel   expansion  tanks   and  liquid/vapor  separators  found  on
current vehicles  (IV-D-356).

      e.    Canister

      Several commenters estimated  the  canister  volume necessary
to control refueling emissions.  The  estimates  ranged  from two
to eight  times  larger  than the  current evaporative canister,
 assuming  that  the   same   quality  of  carbon   is  used.   Many
commenters specifically stated that their  estimates  account for

-------
                              3-8
deterioration of the carbon's working  capacity over the life of
the vehicle.   EPA has estimated that integrated evaporative and
refueling canisters  will be approximately  three to  four  times
larger than current evaporative canisters.

     Many commenters  expressed concern  about  the  location  of
the  larger   refueling  canister.  A  frequently  chosen  location
for  the  canister  was  the  rear  of the  vehicle.   Some reasons
given  for the  rear-mount  preference  are  that there is  not
enough room in the engine compartment for the canister  and that
locating  it  near the fuel  tank allows for  a smaller vent line
diameter  and  reduces  the  total  backpressure  in  the  system.
Obviously, lack  of  space in the engine compartment is a greater
concern  for  small  vehicles  than for large vehicles.  BMW  is an
example  of  one  commenter  who  said  that they  would  locate the
canister  in  the  engine compartment  on  their  larger  vehicles
(IV-D-342).

     However,  several commenters  stated  that all feasible rear-
mount  locations  have  unavoidable negative  affects  on  other
important vehicle  designs parameters.   For  example,  one choice
is  to mount  the  canister next  to  the  fuel  tank.  This  is
described  as  feasible  but  undesirable since it could decrease
fuel  tank size  and, therefore,  vehicle  driving  range.   Some
commenters were also concerned that  this location could  limit,
or  even  eliminate  the option of dual  fuel  tanks or dual exhaust
systems   on   some  vehicles.   A  second   possible  rear-mount
location is  below the  floor  pan,  underneath the  rear   seat.
Commenters described this  as  inconvenient  since it  must  either
reduce   the  head  space  over   the  rear  seat,  or  reduce   the
thickness  of  the  seat  cushion.  The  third  feasible rear-mount
location is  in the trunk compartment.   This  would  require  sheet
metal  changes  and would reduce  the total  volume  of the  trunk
and the total  flat  floor  area.   One  commenter  pictorially
demonstrated that  the lost trunk volume  cuts the cargo capacity
from three   to  two  suitcases  (IV-D-363).    Some  commenters
thought  that  this location could  eliminate the possibility of
carrying a spare tire.

      Manufacturers raised  many  concerns  about  the  safety of
canisters.   As  was  discussed  previously,   some   manufacturers
said that a  refueling canister  is a greater  safety concern than
an  evaporative  canister  simply  because  of  its  larger  size.
Other  canister   safety   issues   that  were  raised  are   the
 implications  of canister  location on the  crash performance of
canisters and  the potential  danger  of  the  release of  vapors
from the  canister.  One  manufacturer  said that  there  are no
safety   concerns   unique   to   a   canister    (IV-F-101).    Many
manufacturers  said  that space constraints  might   force them to
place the canister in  "crush  zones."  These are designed empty
 spaces   which  help absorb  crash  forces  and prevent damage  and
puncture of  components.    Commenters  said  both front or  rear
 locations could  affect  crush  zones.   For  instance,  a  rear-

-------
                              3-9
mounted canister  might be  placed  in  the crush  zone near  the
fuel  tank and  could  puncture the  tank  in  a crash.   However
several commenters thought a  rear-mounted  canister  location was
preferable  (IV-D-339,  IV-D-342,   IV-D-376).    Some  commenters
thought that  the  canister  and  the  activated  carbon  may  be
explosive and/or flammable.   Some feared that the carbon itself
(not vapors released  from the  carbon)  might  be a  fire  hazard.
This concern is addressed in Section D of Chapter 3.

     Most commenters  felt  that there is a potential fire hazard
if vapors are  accidentally released from  the  canister.   Damage
to  the canister in a crash,  loss  of purge,  tampering and other
malfunctions which cause  canister breakthrough were mentioned
as possible situations  which would lead to a dangerous presence
of  hydrocarbon  vapors.   Commenters  said  that  this  problem
should be taken into  account when placing the  canister on the
vehicle  since  some  locations  are  more dangerous  than  others.
According  to   commenters,   front-mounted  canisters  might  be
affected  by  the many  potential ignition  sources  in the engine
compartment.   The hot  exhaust system  could  potentially ignite
vapors from a rear-mounted canister.   A canister  located in the
trunk  was also raised  as a  safety  hazard because vapors could
potentially penetrate the passenger compartment.

     Two  commenters  did  test  work  to  help  define  the hazard
associated  with vapors  released  from  on onboard  system.   The
commenters  reached  different conclusions about the  probability
that  these vapors would cause  any  fire  hazards.  General Motors
submitted a videotape of  a tampered onboard system which caused
an  engine compartment  fire  (IV-D-360(c)).   In their  test the
purge  hose  was disconnected  and  aimed  at  some spark plug wires
which  were in  poor  condition.  The tape  shows a  fire  starting
when vapors pushed through the disconnected hose are ignited by
sparks   from   the   wires.    General   Motors  later   provided
information  which showed  that an evaporative  system under the
same  conditions could also cause  a similar fire  (IV-D-524).

      However,  API  also studied  the hazards  of released vapors
 (IV-D-358)   and   found  no   significant   difference   in  the
flammability   or   ignitability  of  refueling  and  evaporative
systems.   In  both  systems,   flammable  fuel/air  mixtures   were
found only at the exit points of the  disconnected vapor line  in
the   engine   compartment.    API   also  found   that    surface
temperatures  in  excess  of   1350°F  are  necessary  to  ignite  a
 flammable  fuel/air   mixture.   API   did  not   find   surface
temperatures    greater   than   580°F  under   normal  operating
 conditions.   API  mentioned  that  they  did  not investigate  any
turbo-charged   vehicles   which   would   have  higher   surface
 temperatures.

-------
                              3-10
     f.     Purge System Components

     Most  commenters  felt that  the purge  system on  a  vehicle
with onboard  refueling controls will be  very complex and  that
the purge  process will  be difficult to  perform.   Although there
were other comments concerning the  technology  and function  of
purge system  components the only  safety  concern raised  related
to these component  was  the potential occurrence of purge valve
failure and the  problems  that  this would cause.   The commenters
stated  that a  purge  valve  which  fails  in  the open  position
would  lead  to  driveability problems.   Valve  failure   in  the
closed  position could  lead to  canister  breakthrough  and  the
perceived  safety problems of the released vapors.

     5.    Crash Concerns

     As  was   mentioned  earlier,   many   commenters   expressed
concern about the crash resistance of  onboard refueling control
systems.   Specific  concerns mentioned  include  the questionable
strength  of plastic  parts  and  the interference  of  components
with  designed  crush  spaces.   Such components  would interfere
because  they could puncture parts protected  by  crash zones,
such  as the fuel tank.   They also may  require  added body panel
reinforcement.  A fillneck mounted vent valve was pointed to by
several  commenters as  an example of  a  component in  a highly
vulnerable crash  zone which  would  require added  body panel
reinforcement.

     General  Motors  submitted  a  videotaped demonstration  of
crash  testing  done  on  an  onboard equipped vehicle  and also
provided   written  documentation  of  the  results  (IV-D-360(c)).
The   testing   performed  by  a   contractor  (Failure   Analysis
Associates)  was  designed  to  demonstrate  that  onboard system
components are  not  crashworthy.   The test method was to  subject
the  vehicle  to a  thirty miles  per hour  side  impact crash  and
then  to measure the  leak rate of  fuel  escaping through  damaged
components.   This  is  similar to  part  of  FMVSS 301,  except that
the   crash  impact  point  on  the  test  vehicle   was   centered
directly   at the fuel  fillneck  instead  of at the centerpoint of
the  side  of  the vehicle, and  another  vehicle  was used in  the
collision  instead  of  a barrier.   The test vehicle was  equipped
by FaAA with  a  copy of an onboard system originally  designed by
Mobil   which   was   not   intended  to   be   production   quality
 (IV-D-329).   The fuel  fillneck  was chosen  as  the crash  impact
point  because, the rollover valve was  mounted  on  the  fillneck,
 and  is supposedly  one  of the most vulnerable components.  Even
though it was  not an  official  FMVSS 301  test  or a production
quality onboard system the 5.3  ounces per 5  minute leak  rate of
the  system was  compared unfavorably against the  5.0 ounces  per
 5 minute  standard  of FMVSS 301.   No  other  commenter  submitted
 crash test  results  or  challenged EPA's  finding in the  safety
 report that onboard systems could be designed to pass FMVSS  301
 crash tests.

-------
                              3-11
     6.    FMEAs

     Finally,  to support  various viewpoints  regarding onboard
 safety,  three commenters  submitted Failure  Modes  and Effects
'Analyses  (FMEAs)  for onboard  and evaporative systems.  As part
 of  the critique  of the  Mobil  System,  GM  submitted  the FMEA
 which  they originally  prepared for the  stock  system selected by
 Mobil.   Also, for  comparison  sake,  GM submitted  a  contractor
 prepared  FMEA of the  Mobil prototype onboard system  (IV-D-360,
 IV-D-360(c)).    In   addition,   Ford   submitted   an  FMEA  type
 analysis  which compared the potential failure modes and effects
 of  a  current evaporative  system to  those  of  three   different
 onboard control  approaches (IV-D-362).    API  also submitted an
 FMEA  type  analysis which compared  the  relative  risks   of   a
 simple current  evaporative  system  against  those  for several
 different onboard  configurations  (!V-D-358(e)).    These   FMEAs
 will be discussed in Section D of this chapter.

-------
                              3-12


B.   Trends in Fuel/Evaporative System Design

     1.     Introduction

     As can  be  seen  from the  comments  summarized  above,  the
"complexity"  issue  is  one of the key  safety  concerns  raised in
association with EPA's  proposal  to implement  onboard  controls.
However,  in  EPA's  view  (which  is   shared  by  several  other
commenters),  onboard  systems  are primarily  a modification  or
extension  of  the  current evaporative  emission  control  system.
Because of  the "overlap" of  these two systems,  the evaluation
of the  onboard  system  complexity issue must be made in light of
the  complexity  and  safety  of  current  evaporative  systems.
Therefore,  in  order  to  provide a baseline  for  analyzing  the
"complexity"  issue  comments,  EPA conducted a  review of  current
fuel/evaporative   system  complexity   and  design  features  by
examining  shop manuals,  comments on  the onboard proposal  and
other pertinent literature.   While this  review was  not totally
comprehensive,  a wide  variety of vehicle  sizes,  system designs
and manufacturers were  studied.

     This  section  will present  the results  of this  review.  It
will show  that  the trend in  fuel  and  evaporative system design
has  been  one of  increasing  complexity and yet as was shown in
Chapter 2,  there  apparently  has been  no  compromise in vehicle
safety  as a  result of  this  increase  in complexity.   It will
further  show that  many  of  the  design features  criticized as
potentially  unsafe for  possible onboard  system  designs  are in
fact incorporated  on present  fuel/evaporative  systems.

     The  remainder of  this  section is  broken  into three parts.
The  next  part  (2) discusses  trends  in fuel/evaporative system
design   with  respect   to   specific   design   features   and
demonstrates  the   wide  range   of  complexities   among  current
systems   which  result  from   these   design   features   being
implemented.    The  implications  these  current  system  designs
have  for  onboard  systems,  both  in terms  of  overall  complexity
and  specific design features is discussed in part  3.  A  short
summary of this section is then  presented in  part 4.

     2.    Trends  in Fuel/Evaporative  System  Design

     a-     Introduction

     As was  discussed in Chapter 2,  fuel/evaporative systems
have  not   been  a safety   concern  historically and are  not
currently  considered  to be  a  problem.    Yet over  time  these
systems have generally  become more complex  and  currently  there
is a wide  range of system   complexities  in-use.  To generally
illustrate   this   point.    Figure   3-1   shows   the  general
fuel/evaporative system layout for a Chevrolet Caprice which  is
probably   one   of  the  more   "simple"   systems  on  current
vehicles.[1]    In  contrast   to   this   is  the  Volkswagen  Golf

-------
                                                                                    FUEL TANK
CONVENTIONAL
CARBON CANISTER
  CANISTER
  PURGE HOSE
                              FUEL VAPOR LINE
        CANISTER CONTROL VALVE
                                                    Figure 3-1
                                                Chevrolet Caprice
                                                                                                           FILLER PIPE
                                                                                                           MMMMl
EXI8TINO VAPOR RECOVERY SYSTEM
     CHEVROLET CAPRICE
     ••QCIATtt. llC
                        . MD
     •CM.ti MQU£.    \ SKETCH   !»«•« j •» J .

-------
                              3-14
pictured in  Figure  3-2.   This  system  incorporates a number  of
design  features  not  included  on  the  Caprice  including  two
fillneck-mounted valves,  multiple vapor  lines exiting  the fuel
tank,  a liquid-vapor  separator  tank mounted  on  the  fillneck,
and  the  use  of  plastic  for  the  entire  fuel  tank/fillneck
assembly.   Despite  the  differences  in  design  and  complexity
between  these  systems,   both  comply   with   Federal   safety
standards (FMVSS 301) and are currently operating safely in-use.

     As  will be discussed below, there are  a  number  of factors
governing fuel/evaporative  system  design.  While  these factors
have  tended to  cause evolutionary changes  in fuel/evaporative
systems designs over the past 10 to 15 years,  not  every vehicle
model  was  affected the same  by these factors and thus not all
of the  design  features discussed below  appear  on every vehicle.
For  the purposes  of this  discussion these design trends will be
broken   down  into  the  general  categories  of  fuel  metering
changes,  system  sealing  and  venting,   component  location,
fillpipe designs, the number  of connections between components,
and  the  use  of  alternative  materials  for  system components.
Each of these  is discussed below.

      b.     Fuel Metering

      Of all  of the  changes  to fuel  and evaporative systems over
the  past several years, the  change  in the fuel metering  system
from  carbureted  to  fuel   injection  is   probably  the  most
revolutionary.    In  1980   there were  very   few   fuel-injected
vehicles available,   but   by  1990   almost  all   vehicles   are
projected to be fuel  injected.  This  change  resulted in  the  use
of   high  pressure  in-tank  fuel  pumps and  high  pressure fuel
 lines.   An  additional line and its associated  connections were
 also added  to return  the excess fuel  and vapor  from the  engine
 compartment to  the  fuel  tank.   Because  this  excess  fuel   is
heated in the engine  compartment,  its return to  the fuel tank
meant that  higher  fuel tank  temperatures  and  pressures  also  had
 to  be  accommodated.   The  resultant designs  have   clearly  added
 complexity   to  fuel/evaporative  systems.   Yet,   there  is   no
 evidence to suggest  that  this  fundamental  change  in  the fuel
 metering system has compromised vehicle safety.

      c.    Tank Sealing  and Venting

      The area of  fuel/evaporative  system  sealing and  venting
 has  seen   a  number   of   evolutionary   changes  over   time.
 Initially,  a  limiting orifice  was placed at the fuel tank/vapor
 line  outlet  to  meter   evaporative   emissions,  to  act  as  a
 liquid/vapor  separator  and  to minimize  fuel  leakage   in  the
 event of a  vehicle rollover.   Later,  a rollover valve,  such as
 those  shown  in Figure  3-3,  was  added  at  this location  to
 improve system  integrity.  Some of  these  rollover valves  also
 included float  mechanisms  which allowed  them to  also  act as an
 overfill  limiter.    Finally,  high-flow   rollover  valves  with

-------
   Figure 3-2



Volkswagen Golf

-------

-------
                              3-17
built-in pressure  relief  mechanisms were  introduced to  handle
increased tank pressures  brought on by higher  volatility fuels
and the  use  of fuel  injection.   However,  these relief  valves
sometimes operate  at very  low  pressures  and consequently vent
vapors directly to the atmosphere under many  in-use conditions.
Fillcap designs have  also  changed  in response to the need for a
better sealed system.  Tighter  sealing fillcaps  with pressure/
vacuum  relief  valves  were  introduced and  more recently two-
stage  fillcaps have been  implemented  in  response  to increased
tank  pressures.   These two-stage  fillcaps are removed  in  two
steps  which  allow  for  safe  tank  depressurization  and  thus
should help to prevent spurting upon fillcap removal.

     d.    Component Location

     Fuel and  evaporative system component  location is another
area  in which there  have  been a number  of changes  over time.
Since  the mid-1970s  there has  been a trend toward moving the
fuel  tank forward of the  rear  axle.  This change  has resulted
in  a number of other system changes.  First,  the fillpipe was
moved  from  the rear  to  the  side  of  the vehicle.   Second,  in
many  cases  the  packaging  considerations  of   the  new fillpipe
location dictated  that  the  fillpipe  diameter  be  reduced  by
approximately  25  to 50 percent necessitating the addition of an
external  vent  line  to vent refueling vapors.   This change was
also  influenced by a desire to  reduce vehicle  weight to  improve
fuel  economy.    It  should be  noted  that  these  external Vent
lines  are not  generally protected by a rollover valve.

      Finally,  in   some cases the new tank location left no room
for the tank vapor dome, which  led  to the external packaging of
the vapor dome and other  components normally  found  on the fuel
tank  itself.   Examples   of  these  external  components  include
fuel  expansion or separator tanks  (which  are  sometimes  placed
 in the  crash  zone  due  to  space  considerations)  as  shown  in
Figure 3-4, external rollover  and check  valves like those  in
Figure  3-5  which  are placed  between the fuel  tank  and  the
evaporative canister, and external  fill limiters.   Associated
with this  increase  in the use  of  external   components  was  a
corresponding  increase  in the number of hoses  and  connections
used  in  fuel  and  evaporative  systems.   Yet,   despite   the
 increased complexity brought  about by these changes the design
trade-offs  do  not appear  to have compromised  safety.

      e.    Fillpipe Related Changes

      There  have  been a  number  of  changes to  the  fillpipe  and
 related  components  over  time.   As  was discussed  previously,
 there  has  been  a  general move   from   rear-to  side-fill  tank
 designs  and  the  trend  toward   external  vent   lines  which
 accompanied it.   Because  both of  these  components  are   located
 in the crash  zone,  the  insertion  of flexible  rubber sections in
 the   fillpipe   and  vent   line  was  instituted   to   improve

-------
      Figure 3-4
Liquid/Vapor Separator
         and
 Fuel Expansion Tanks

-------

-------
                              3-20
crashworthiness.    In   order  to   improve   fill   quality   or
"fillability"   (i.e.,  the   quality   or   acceptability  of  the
vehicle refueling process in the areas of  consumer  reaction and
safety)  a  number  of  features have  been added to  the fillpipe
including anti-spitback valves like  those  in  Figure 3-6.   These
components  reduce  the  number of premature nozzle  shut-offs and
fuel   spitback   onto    the   refueler.    Another   interesting
development  in  this  area  is  that  some  manufacturers  have
designed the fillpipe/fuel tank  interface such that the  end of
the  fillpipe  is submerged in  the  fuel  during  at  least part of
the  refueling  event.   An  example  is  shown in Figure 3-7.[2]
This  creates a  "liquid seal"  that  helps  to  modulate pressure
during refueling but also  introduces the potential for spurting
if  the tank  is overpressurized and the  fillpipe  is  submerged
when the fuel cap  is removed.

     In   addition  to   the   fillpipe  modifications  discussed
previously,  the fillpipe has seen  increasing use  as a mounting
point   for   other  components.    Mounting  components  in  the
fillpipe  area  was  raised  as  a   safety concern by  several
commenters,  but  the mounting  of   components  in  this  area  on
present  vehicles  is not uncommon.    Examples  of this include the
fillneck-mounted   valves    shown   in   Figure  3-2    and  the
fillpipe-mounted  liquid/vapor separators  shown  in Figures  3-2,
3-8, and 3-9.   While mounting components in the fillpipe  places
them in the  crash zone,  there  is  no evidence  to suggest  that
this trend  has  compromised vehicle safety.

      f.     Connections

     As  fuel/evaporative systems have become  more complex  and
the number of  external  components  has  increased there has  been
a corresponding  increase  in the number  of connections  between
components.   An   example  of  the  extra  connections  associated
with  liquid/vapor  separators  is  shown  in  Figure  3-10.   The
 introduction of  rubber  sections  added  two  new  connections  to
both the  fillneck and  external vent  line.   Also, the   use  of
external rollover  and  check valves has increased  the  number  of
connections in the  vapor  line  to   the  canister.   An  important
point  to note  here  is  that  some of the  added  connections,  such
 as those associated with fuel expansion or separator  tanks,  are
direct  connections  to  the  fuel  tank  and  often   not  protected
with a rollover valve.  It  is  also worth noting that  increased
 connections  often  involve   areas   which  often  contain  liquid
gasoline  such  as  the  fillpipe,   external  vent   lines   and
 liquid/vapor separators.

      g.     Alternative Materials

      The last  trend to be  discussed is  that  of  increasing use
 of  alternative  materials,   such  as  plastic  and  rubber,  to
 fabricate  fuel and evaporative system  components.  The  use  of
 these materials  in  onboard  system  components  was  characterized

-------
                                                                    1 m   m   I 1::>';-,..
           ••' "'••• ' '•:'••v •-':. "  .  i   • ,; '•,.'-  B  '..   i  "   ;''  j^ffln ".-;-.' '":•''....' ,.'y -..•"'••""•':' •,'-•'•':' '•" mMfflnM  w^^^M
               •  l   |M!  "'"•.•   • -• ;  .:'-:"  - •  "   'V..  :  •  ;:  '   :    .-•''•':-,::.,••••'•'"::"-  ' ••'"  • '• •   '•:"'•.':::{WfSm:'":'-:





                                  ,            ,-     I  " i  I
                                    '."  ,              • '-: j
                        I   .    ••  I  •     m	 '   '
                   »i
                                                •
                                     HH ^9i !  I";  I

       ",^BKi
       ^i8p*aifc,i»
!  "n/iAiiiU. '  ^tiV^W-X" '• "': •' "  T .•"•>" ,»

i^*^"'   ;-^c«*?*
 r*-,' • '.,i  j,..•.«.-•,  ,   '' • »^» ,
A-'-VV;^' :'f---.t*-.-
x^.' '•'• .          . •'•         '  • *
             3-6


Anti-spitback Valves

-------
13"
I
       Driver's side view
                  I  °O
                                     Figure 3-7

                                 AMC Jeep Cherokee

-------
   Honda Accord
Fillpipe Assembly

-------
       Figure 3-9



lt>yota Fillpipe Assembly
*3$>£*&£fcyi£fKJ*^. v£

-------
14-76

FUEL  TANK
REMOVAL AND I
   Pre-remov«l Operation
   •  Draining of the Fuel.
                               Figure  3-10

                            Mitsubishi Montero
   Pott-instillation Operation
   •  Supplying of the Fuel.
                                                                                             oiwssi
            1 Nm
            .7 H.lb«
          12
            19
                  25-30 Nm
                  18-22 ft.lb*.
                               15-20 Nm
                               11-14 ft.Ids.
                                                                                          oivttt
   Removal steps
 1. Drain  plug
 2. Fuel filler cap
 3. Fuel filler hose  protector
 4. Vapor hose
 5. Check valve
 6. Overfill limiter (Two-way *
 7. Clamp assemply
 8. Fuel filler hose
 9. Breather hose
10. Packing
11. Fuel filler nedc
12. Main  hose
13. Return hose
                                                   '•*  Fuel gauge unit connector connection
                                                   '5.  cuel tank assembly  mounting nuts
                                                   '5  Fuel tank
                                                   • 7  P-ce assemoly
                                                   ' i  Seoarator tanks
                                                   '9  F^ei tank protector

                                                    ••OT=
                                                       Reverse  tne removal  orocedures to reinstall. H
                                                    2i  *>^   Refer to "Service Points of  installation"

-------
                              3-26
as potentially unsafe by one  commenter,  yet the use of plastics
and rubbers  in  vehicle fuel  and  evaporative systems  is  common
on  most  of  today's vehicles.   While plastic  fuel  lines  have
been  used  intermittently  on   domestic  vehicles   since   the
mid-1960's,[3]  recent  years have  seen  the more  widespread use
of  plastic fuel  tanks  and the  increased  use  of  plastic  fuel
lines.   In  fact,   one  source  estimates  that  as  much  as  50
percent of all domestic  fuel line  business  will  be  plastics
within  the next  five  or   six  years. [4]   In  addition to  fuel
lines  and  fuel tanks,  plastics  are  presently  being  used for
fillnecks,   filler  caps,   evaporative  canisters,   liquid/vapor
separators  and   other  external   hardware  such  as  rollover
valves.  The  use  of rubber components  has  also  increased over
time with  the addition of the flexible  sections in the fillneck
and external  vent line and the use of rubber for  fuel  and vapor
lines.  It appears  that this  trend of increased  use of plastic
components  and  other alternative  materials in current fuel and
evaporative system  design has not compromised vehicle safety.

     h.    Summary/Conelus ion

     This  brief  discussion,  and the  examples  provided  above,
show  that  there is indeed  a  wide range  of fuel  and evaporative
system  complexity  on  current vehicles.   These  differences  in
complexity are reflected   by  a  number  of  factors.   Fuel and
evaporative  systems vary in  their general  designs,  the number
and  function  of  the  components  used,   and  the  number   of
connections needed  to  integrate  these components  into  a working
system.    Component dimensions  vary  as  does  the  material  of
which  they are  made and their location  on the  vehicle.   Indeed,
many  components  are  commonly located  in  potential crash  zones
on the vehicle.

      Given  the  number  of  comments  received   regarding the
purported  relationship  between  safety  and complexity and the
 importance of that concern to the  commenters  opposing  onboard
controls,   one  might  ask  whether  the  fuel/evaporative  system
experience  with   complexity   led  to  an   increase  in   safety
problems.   While  detailed  information on  the  safety performance
of  specific   systems   is   generally not   available  in   public
 records,   several  conclusions are possible.   First,  the  recall
 information  in  Figures  2-1  and  2-2   shows   that  evaporative
 system recalls have historically been  rare.   The trend,  if any,
does  not  show an  increase  in recalls with later  model years  as
 systems  have become  more  complex.   Also,  the  information  in
Table 3-1, taken  from  Tables  64  and  68  of the  draft NHTSA study
mentioned  previously,  shows that  the number of owner complaints
 and service bulletins  for  fuel emission control  (category 0612)
 does  not  seem  to have any relationship to model year,  and thus
 indirectly to increasing system  complexity.[5]   Furthermore,  it
 should  be  noted  that  EPA's   review   of  evaporative  system
 recalls,   service  bulletins  and  owner   complaints  discussed  in

-------
                    3-27
                 Table 3-1
          References to Technical
  Service Bulletins, and Owner Complaints
   for Fuel Emission Control Components
Complaints                    Service Bulletins
 MY     %                        MY       %

78     12.3                   Unknown   24.0
77     12.1                     82      10.2
75     11.2                     81       8.4
79     10.7                     83       7.8
76      8.3                     79       6.8
74      6.9                     77       6.2
84      5.8                     78       5.9
83      5.6                     80       4.8
80      5.6                     84       4.8
81      4.9                     73       3.9
82      4.3                     75       3.9
pre-60  3.6                     85       3.1
73      2.9                     74       2.4
86      2.2                     76       1.9
85      1.3                     68       1.6
71      0.9                     70       1.0
72      0.4                     86       1.0
87      0.4                     71       0.9
70      0.2                     72       0.6
                                87       0.6
                                67       0.2
                                69       0.1

-------
                              3-28
Chapter 2 provided no evidence to indicate that  any  one  design,
design  approach,  or manufacturer's  systems  were safer  or  less
safe than any other.   EPA presumes  that the  manufacturers  have
concluded that  the  complexity  added  to  fuel  and  evaporative
systems  over the  recent years  was  warranted  and  the  data
available  indicates  that safety  was  not  compromised   in  the
process.

     As  is  evidenced by  the  previous  discussion, many  changes
to  fuel  and evaporative  systems  which  have  increased system
complexity have been implemented over  time and  a wide  range in
design  complexity  currently  exists.   Yet,  the discussion in
Chapter 2 indicates  that  these "evolutionary" changes have been
implemented   safely  and there  appears   to  be   no   direct
relationship between  complexity  and   safety.    Since,   in  the
opinion   of  EPA  and  others,  onboard  systems  are merely an
extension of current evaporative systems,  their implementation
is  an  evolutionary  rather   than  a  revolutionary  change.   As
such,   the   manufacturers  could  reasonably  be  expected  to
implement onboard  controls  using the  same  approaches  used to
safely  implement  the fuel and evaporative system modifications
previously discussed.

     Furthermore,  the  analysis provided  above  indicates   that
evaporative  system  safety is  generally  independent  of design
variation and  complexity.   Since  the hardware and technology
used  to  control  refueling  emissions  would  essentially be the
same   as  that  used in  evaporative  systems,  this experience
suggests  that variations  in system designs  or complexity  with
onboard controls would  also not lead to safety  problems.   Thus
systems labeled  as  "complex" by some  would not necessarily be
any less safe  than  those labeled  as  "simple."  Manufacturers
have  accommodated  a  wide variety  of  fuel/evaporative system
designs  safely,  and  this experience  indicates that  the  same
could  be done for refueling  controls.

      3.     Implications  of Specific  Design Features

      As  was discussed in Section  A  of  this  chapter,   EPA has
received numerous  comments  voicing concerns  about the safety
 implications of  specific  anticipated  onboard  system  design
features.   EPA does not necessarily agree  that onboard systems
will   require these design   changes.    However,  the  following
paragraphs   will  briefly review  some  of   these  concerns  and
demonstrate that  a number  of these features,  which have  been
 labeled  potentially unsafe   by the commenters,  are already  in
production  on current systems and,  as  was discussed  above,  are
 operating safely.   These include the  increased use of  external
 components   and  associated connections, placement of components
 in crash zones  and the use of plastic components.

      First, the  argument has been  made that some onboard system
 designs  will increase the number  of components  external to the
 fuel  tank.   This  would increase  the  number  of  connections

-------
                              3-29
between components and the  fuel  tank and therefore increase the
number  of  possible  failure  points  in  crash  and  non-crash
situations.   The  widespread  use   of  external  components  and
extra  connections on  current  designs  shows   this  not  to  be
problem.   A  specific  comment  in  this  area  came  from  the
National   Truck   Equipment   Association   which  stated   that
liquid/vapor separators could  be a safety hazard.[6]  There are
numerous  examples of  the  use  external  fuel   and  evaporative
system  components including liquid/vapor separators  on current
vehicles.   Figures  3-11,   -12,   and  -13  show  examples  of
liquid/vapor  separator  tanks and important  valves which are
located  away  from the  fuel  tank area.  Each  external component
adds  at  least  two  additional  connections  in  the  fuel/vapor
handling   system  and,   depending   on  the  design   used,   the
fuel/vapor  separator tanks  add between four and ten connections
per  fuel tank.   As  was  previously discussed  there is  a  wide
disparity  among current systems with respect  to  the number of
connections.   Some  current  systems,   such as the  system  in
Figures  3-14,[1] and -15,  have relatively  few  connections  while
others  have many, such as the system  shown in Figure 3-10.  In
addition,   components  such  as  the fuel/vapor  separator   tanks
have  added significantly to the  amount of vapor  hose  used and
the  complexity  of the layout of  the hose.  A good example of
the  increased use of  vapor tubing  and  external valves  is the
1988  Honda Accord fillpipe  assembly  shown in Figure  3-8.  In
this  case  the  extra  vapor  tubing  is  routed to  act   as   a
liquid/vapor  separator and  this is then  connected to a two way
(pressure/vacuum) valve to  the tank.

      It can be seen  from  these selected examples, that a number
of current fuel  and evaporative systems  use external components
which  handle   fuel  and/or  vapor.   These   external  components
involve an increase in the  number of  connections and the amount
of vapor line used over  some simple systems.   However, there  is
no  evidence  to  indicate   that  these  designs  have  degraded
vehicle   safety.   Once    again,   EPA    presumes   that    the
manufacturers   have   concluded   that  the   use  of  external
components and  the  corresponding increase in connections  was
warranted and the evidence  indicates  that  the design trade-offs
 involved were managed  successfully.

      The use of  plastics in onboard systems is  another  concern
that has  been  expressed  by  MVMA and others.   The concern  is
that while metal parts may bend  or  crush,  but still maintain
 integrity  in  a  crash situation, plastic  parts may  crack  or
break on  impact.   A related  concern  is  the  buildup  of  static
charge   on   isolated   plastic   components   if  not   properly
grounded.  However, as  was mentioned  in the  previous  section,
 the  use of  plastics   for  both  fuel  lines  and  fuel  tanks  is
becoming  more  widespread  and   apparently   has  not   degraded
 safety.  An  example  of  a plastic  fuel  tank  is  the Volkswagen
 Golf shown in  Figure  3-2.   In this  system,  not  only  is  the
 entire  fuel   tank plastic  but  also  the  filler neck  assembly

-------
FUEL TANK
REMOVAL AND INST;
                                          Figure 3-11


                                       Chrysler Conquest
                                                                                   14-/9
                                                                                     N14QA--
   Pra-ramoval Operation
   • Release of Residual Pressure
     from Fuel High Pressure Hose
     (Referto P.I4-38.)
   • Draining Fuet
   Post-installation Operation
   • Replenishment of Fuel
                 1.0 Nm
                 0.7ft.!bs.
                                                                               10
                      9
                          12
                              25-30 Nm
                              ta-22ftlbs.   13
al stops
  1. High floor side panel
 2. Fuel pipe cover
 3. Fuel pump'connector connection
 4. Fuel tank cap
 5. Drain plug
 6. Fuel gauge unit connector connection
 7. Fuel high pressure hose connection
 8. Return hose
 9. Vapor hose
 10. Fuel filler hose
 11. Fuel filler neck
 12. Fuel tank
 13. Electrical fuel pump
 14.-Separator tank
 15. Fuel gauge unit
 16. Pipe assembly
 17. In-tank fuel filter
                                                          15-20 Nm
                                                          11-18 ftlbs.
                                                                      OIT*M
                                                 NOTE
                                                 (1)  Reverse the removal procedures to reinstall.
                                                 (2)  »«r Refer to " Service Points of Installation".

-------
                                        Figure 3-12

                                        Range Rover
KEY TO DIAGRAM

  1.  Charcoal canister
  2.  Air Inlet to canister
  3.  Purge Int to plenum chamber
  4.  Connector hoses with restrtctors
  S.  Restrtctor In purge line
  6.  Fuel expansion tank
  7.  Fuel vapour pipe from manifold
  8.  Breather hose with anti-surge valve
  9.  Fuel tank flier neck
 10.  Filler neck breather hose
 11.  Manifold
 12.  Fuel vapour pipes from fuel tank (3 off)
 13.  Pressure retef valve and hose
 14.  Fuel tank
 15.  Hoat/roOover valve
 16.  Grommet
Tl

-------
                         Figure  ~

                         Ford Festiva
                                             FUEL TANK
                                                 FUEL
                                                RETURN
                                                 LINE
                                                          SEPARATOR
                       FUEL
                      SUPPLY    VAPOR
                       LINE      UNE
  FUEL
 VAPOR
STORAGE
CANISTER
 Fua
FILTER
                                            ROLLOVER
                                              VENT
                                              VALVE

-------
           Minna
                    VIIW A
FRONT
                                              FILLER PIPE
              /X»- PURGE. TO ENGINE
                                                           FUEL TANK
                                                SOLENOID PURGE VALVE
CARBON CANISTER 	 *-





' X





i^ /UP
//Oi
JSs
^ss&yt&S


Figure 3
Ford T£u
i 	 ^- PROMT

11
^S**-* — WIMNQ AtSCMILV

H 1
-14
irus
I 1
*3? anil »«tCM»T»O« 1 MATIMA1 STICIV.
EXISTING VAPOR RECOVERY SYSTEM
FORD TAURUS
MUFtl-EH AltOCIATB*. INC ••Illmpr*. MO
>Mf 12 (7 «6| SCAifl NONE 1 '*" 1 Ol> J


-------
         Figure 3-15

         Ford Taurus
FUEL TANK
ASSY 9002
FUEL LIN6
ASSY 9J729
FUEL FILTER
                                 FUEL LINE
                                 ASSY 9J279
                                        ADAPTER
                                        96593
              FUEL FILLER NECK
             'ASSY 9034

-------
                              3-35
including   the   fillneck-mounted   valves   and   liquid/vapor
separator.    Indeed there  are many  more  examples  of  plastic
tanks now  in use.   Furthermore,  many  other components  in  the
fuel/vapor  handling  system on  today's  vehicles  are  plastics.
Examples include the fillneck assembly  on the SAAB 9000 (Figure
3-16),  the  liquid/vapor  separator  tanks  in Figure  3-13,  and
essentially  all  of the external  valves used in  the  fuel/vapor
handling system.  These examples provide  good evidence that  the
use  of  plastics  does  not adversely  impact safety.   Concerns
expressed about the  use of plastics by MVMA are  not  supported
by in-use experience on current vehicles.

     Another  concern  voiced  by  several  commenters  is  that
onboard  systems  will require the  placement of  some  components
in the  crash zone, such as the fillpipe area, and that this may
create  a hazard  in accident  situations.   While EPA believes
this  design  approach  is  not  necessary,   there  are  numerous
examples of  rollover  valves and other  components  in current
systems  that  are  mounted  on the  fillpipe  or  in  other crash
zones.   Again  there   is   no  evidence  to  suggest  that  the
placement  of  those  components  has degraded  vehicle safety.
Examples  include  the   fillneck-mounted  rollover  valve  of  the
Volkswagen  Golf shown  in  Figure  3-2  or the rollover valve of
the  SAAB 9000  shown in Figure 3-16.  Another example,  the  1988
Mitsubishi Cordia/Tredia  in Figure 3-17,  has a fillpipe mounted
separator  assembly  with   five   separate  vapor  hoses   in  the
vicinity.   The Mazada  626  fuel  system (Figure  3-18)  has three
vapor   hoses  in  or   on  the   fillneck   plus   the   necessary
connections  to the  fillpipe.  And  both the Range Rover  (Figure
3-12)  and  the  Ford  Festiva  (Figure  3-13)  have  the  fuel/vapor
separator  tanks mounted  in the crash  zone.   Thus,  while it is
certainly  legitimate to  be concerned  about placing  components
in  the crash  zone,  the examples  demonstrate that even critical
components   have   been  placed  in  the  crash zone  on  current
vehicles.

      4.    Summary

      Fuel  and evaporative  systems have gone through  a number of
revolutionary  and evolutionary changes over time,  and  today's
vehicles  show  a wide  range in design  complexity.  Perhaps the
most  significant   change,  the   adoption  of  fuel   injection
systems,  has  complicated  fuel system  design significantly and
raised  fuel  tank temperatures  and pressures.  Many of  the other
changes discussed  above   which  have  contributed to  the  wide
range of design complexity in  current fuel/evaporative  systems
are   more   evolutionary  in  nature.    Whether   evolutionary or
revolutionary,  it is   assumed  that the  manufacturers  included
these  design  features   for  good  cause,   and  the  available
evidence suggests  that vehicle safety was  not compromised.  This
experience  strongly suggests  that onboard systems  of  various
design complexities could also  be implemented safely.

-------
Figure 3-16

 SAAB 9000
 FilLneck

-------
                                             Figure 3-17

                                     Mitsubishi  Cordia/Tredia
COMPONENTS

   Vehicle* without • turbocturfW
                                                     1.0 Nm
                                                     .7 ft
                                Fuel gauge unit and
                                pipe assembly
                        Fuel vepor ho»

                      Fuel return hot*
                                  \
                       Fuel main hoM
                         Separator assembly
              Fuel tank
    Fuel filler neck
Fuel vapor fioie
                                                                                                 . Fuel tank cap
                                                                                            Protector (TRiDIA)
                                                                             Fuel filler hose
   Fuel vepor ho*
        Fuel return hoi
             Fuel main hone
                              Tank band
                            03K566
                                         Drain plug   IS—25 Nm
    Vehfelee with a turbocharew
                                                      1.0 Nm   Cov«f
                          Fuel gauge unit and pipe assembly  .7 ft.lbt.
                             Fuel vapor hose      \    fj
                      Fuel return ho»
                                  \
                        Fuel main ho*§
                 Fuel tank
           Fuel vapor ho:
                                                                                                . Fuel tank cap
                            Separator anembly
                                                                                               Protector (TREDIA)
                Fuel return hose
                           Protector
                                Tank band
        30-40 Nm
        22-29 ftJbs.
Electric fuel pump
                                                                                          03K565
                                               Oamolug    15—25 Nm
                                      30-40 Nm           11-18ft.lb«.
                                      22-29 ft.lb..
  14-100
              STB Reviskm

-------
 Figure  3-18


Mazda  626/MX-6
                                                MU04M9S
 Not*
 Drain the fuel from the fuel tank before removing the tank.

1 . Fuel pump connectors
2. Fuel hoses
3. Steering angle transfer shaft (4-wheel

  SETS
A K  ° Se2tlon 10)
4. Cross member (4-wheel steering)
              5. Evaporative hoses
              6. Fuel filler hose
              7 Breather hose

              8 Fuel tank
              9- F^l tank

-------
                              3-39
     Furthermore,   this   review  of  current   fuel/evaporative
system design  has revealed  that a  number of  design  features
such  as  component locations,  alternative materials,  and  extra
connections which were  characterized as potentially unsafe  for
onboard systems  are  not specific or unique to onboard systems
at all,  In  fact,  the  very components or design  features  cited
as  onboard  concerns  are  present  on  current  vehicles.   Once
again,  it   is   reasonable  to  believe  that  the  manufacturers
concluded  that  these design features were  needed and  that they
would not compromise safety  .  Clearly,  if  such design features
are  used  safely  on  current vehicles  they can be incorporated
safely into onboard systems as well.

-------
                              3-40


C.   Onboard System Design and Safety

     1.    Introduction

     The  issue  of  system complexity  and  safety was  raised  by
several parties prior  to  the NPRM and in the  comments received
subsequently.   The focus  of the  issue  was that  onboard vapor
recovery systems would be  more  complex than current evaporative
systems, . and   that   adding   this  complexity  would  cause  an
unquantifiable  increase in the  risk  of  both crash and non-crash
fires.   Complexity  factors   cited by  the  commenters  included
larger  components,   more components,   more  connections,  the
location  and   materials  chosen   for   these  components,  and
possible effect on other vehicle systems.

     In  EPA's   view,  the  integrated  onboard/evaporative  system
described in the  NPRM was not  considered complex, especially in
light  of  the  range  of   complexities   in   present  evaporative
systems.  As is shown in Figure 3-19, it involved modifying the
fillneck  to provide  a liquid seal,  adding  a fillneck mounted,
nozzle actuated vapor control valve,  and  enlarging  the vapor
line  diameter  and charcoal   canister.   EPA's  initial  study of
the  safety  implications of onboard vapor recovery  systems found
that  systems such as  that  presented  in  the  NPRM  are   simple
extension/modifications  of current  evaporative systems and the
"straightforward,    reliable,    and   relatively    inexpensive
engineering solutions exist  for  each of the potential problems
identified."  The study  went on  to  conclude  that "no increase
in  risk  need  occur  or  be  accepted as a  result  of  an  onboard
system"  and that  vehicles  equipped  with   onboard  systems  can
"provide  a  level  of  in-use fuel system  integrity equal to or
better than achieved on  present  vehicles  which  incorporate
evaporative emission control systems,"  (see Appendix  II).

      Nevertheless, as was described  above,  concerns  were raised
 in  the comments  regarding the complexity  issue in general  and
more specifically with  regard to features  of the NPRM  design.
Specific  comments addressed  areas such  as   fillneck valves,  use
of  plastics, increased connections,  increased  vapor  line length
 and   diameter,   canister    location,   and   crash/crush   zone
concerns.    The potential  onboard  system  designs submitted by
the commenters  varied greatly.   As is shown in Appendix  I,  some
 commenters  submitted  design  approaches  with features similar to
that  shown  in  the  NPRM  system  such  as  liquid   seals   and
 integrated   canisters.  Other systems employed  mechanical  seals
with separate canisters for  refueling  and  evaporative emissions
 control  and with  a  variety of  different  vapor  control  valve
 approaches.

      EPA's   follow-up  on  these  comments  took  two  different
 approaches.  The  first  involved  a  review  of  evaporative system
 designs  found  on  present  vehicles.    This review  divulged  a
 great deal  of  information  regarding  evaporative  systems design

-------
                                          Figure  3-19

                           Integrated  Evaporative/Refueling  System

                                        J -  Tube Seal
PKI SSI HI / \,\( I I M
   KKUI.I CAI'
 NO//I.I-: ACTUATED
ROLLOVER/VENT VALVE
                       5/8" DIA.

                       81 LONG
                                              .05" DIA. LIMl'MNU
                                                  ORIFICE
                                     SEAL
                       \
                      DESIGNED SLOW LEAK
                                                                             PURGE VALVE
                                                                     3 LITER
                                                                     CARBON
                                                                     CANISTER
                                                          3/8" DIA.
                                                             TO PURGli
                                                             INDUCTION
                                                             POINT
                       14 GALLON FUEL TANK

-------
                              3-42
approaches  and  complexity,  and  the   hardware   used  in  the
systems.   As  was  discussed  in  Section  B,  essentially  every
specific hardware  concern  for the onboard  system was  found  to
have a  counterpart already  being used  on current  evaporative
systems.   Furthermore,  the  review found  that  there was  a wide
range of  complexity  in current  evaporative  systems   and  that
some of  the  design  concepts  used in  evaporative  systems  are
similar to that expected for  onboard  controls.  Even  with this
wide range  in design approaches  and  concepts, no  evidence  was
found  to  indicate  that  increased  complexity  has  decreased
evaporative  system safety.   Furthermore,  the wide variety  in
design complexities  in  current evaporative systems without  an
impact   on   safety   suggested   that   a   variety  of  design
complexities  in  onboard  systems  could  also  be  implemented
safely.

     Nevertheless,  even though the  premise was  debatable,  EPA
decided to confront the complexity  issue directly by developing
a  system  even simpler than that presented  in  the NPKM  and which
doesn't add  complexity  to  the current evaporative system.  The
basic  design approach used was similar  to  that shown  in Figure
18  of  the initial safety study (Appendix II)  with even  further
simplifications  and  improvements.  The   intent of this  program
was  to  show even  further  the  simple   designs   available  for
onboard  controls, and  that the more  complex  design  approaches
suggested  by some commenters are not necessary to achieve safe
and  effective control of refueling emissions.

     As  an overview,  the remainder of this section  of  Chapter  3
is  broken  into   five  parts.   Part   2   which   follows  this
introduction  explains EPA's  goals  in the onboard system design
and development  program.   Part  3  describes  the  baseline fuel
system  evaluated and  the   modifications/additions  needed  to
incorporate  onboard controls.  Part  4 presents the results of
the emission testing conducted and part  5  draws conclusions  and
discusses  why simple  systems  such as  that developed here  by  EPA
should  and would  be  preferred by most  manufacturers.   Finally,
part 6 discusses  how the  use  of  simple systems  such  as  that
developed by EPA  addresses  many  of the  general  and  specific
safety  concerns raised  by  the commenters.

     2.     Program Goal and Constraints

     The  main  goal  of the EPA program  was  to  address  the
complexity   issue   by    developing   a   simple    integrated
refueling/evaporative  control   system  which   incorporated  the
fewest  features  possible  to perform the  necessary  functions.
 It was  also EPA's intent  to utilize  system components  that  were
based  on current  production hardware.

      In   terms   of    testing,    several    constraints   were
established.   First,   it  was decided   that   the  complete  fuel
 system  should   be   tested  using   the  RVP, temperature  and

-------
                              3-43
dispensing conditions  in the proposed  test procedure  and  that
this  testing  should  be  conducted  with  the  system  in  the
geometry  and configuration  which  exists  when  mounted on  the
vehicle.  Second, other issues not  directly  relevant  to onboard
safety  such  as  canister  purge were not to be  evaluated in this
program since  for  the most  part  they  were viewed  to be  more
evaporative  control  issues.   With these goals  and  constraints
established,  the first  step of  the  process  was to  select  a
baseline  vehicle fuel/evaporative  system  to modify.  This  plus
the necessary modifications are described next.

     3.    Development of the Onboard System

     a.    Baseline Vehicle Fuel/Evaporative System

     The  fuel/evaporative  system  selected  for  modification was
that  found  on  fuel  injected 1986 General  Motors  (GM)  A-body
cars.   This  particular  system   was   selected   for  several
reasons.   First, it  is  typical  of mid-size  passenger cars in
today's fleet  in terms of fuel  tank  size,  fill location,   fuel
metering,  and  has a  relatively  simple  evaporative system.  The
use  of  the relatively simple evaporative system over some  more
complex  designs presented   a  bigger  challenge  in  terms of
demonstrating  the  same  level  of  simplicity as  compared  to the
evaporative   system.    Second,    this    general  fuel  system
configuration  was  presented by  GM  at the  public  hearing as
having  an outstanding record of  crash  integrity  in the field,
and  thus  appeared  to  be  a good base  case  vehicle in   that
regard.[7]    Third,   this   fuel  system   was   selected  for
modification over others since it  was already being used as the
base case  in  an  ongoing  failure modes  and  effects analysis
being   conducted by  EPA (see Section  D  of this  chapter).  It
should  be noted that the  concept to  be  demonstrated here was
viewed  to be equally applicable  to most other fuel systems, but
 it is  not EPA's role to design and demonstrate control  systems
for  every vehicle model.

      The  base  fuel/evaporative system  is shown  in the photo in
Figure  3-20 and is  illustrated  in Figure 3-21.   As  can be  seen
 in the figures, the fuel  system  is  configured  for   side  fill.
Both the  fillpipe (1.25" I.D.) and the external vent line  (0.5"
 I.D.)  have rubber connector  pieces.  The  fuel tank  is about 59
 liters   in  capacity  and  has  a   vapor  dome  of  about   twelve
 liters.  The fillneck enters the  tank  on the driver's side and
 continues  into  the  tank  for   about   four  inches.   It  thus
 provides  a  submerged fill   after  about 95  percent   of  capacity
has been  dispensed.

      As  is  the  case with  most  fuel/evaporative systems,  the
 vapor  line  to  the  canister is  protected by  a  rollover  valve
 which  is fitted on  the underneath  of the fuel  sending  unit.

-------

-------
                                              Figure 3-21

                                   Stock Fuel Tank/Evaporative Control System
      Vehicle
      Rear
                                                                        Vehicle
                                                                        Front
          External Vent Line
                 (1/2" I.D.)
Fillpipe
(1 1/4"  I.D.)
Internal Extension
of Fillpipe
                                                     Sending Unit
                                                       Rollover Vent Valve/Limiting Orifice (0.055")
                                                                                  Vapor Vent Line (V I.D.)  - ll  ft
                                                                                               Purge Line
                                                                                     1 .5 Liter
                                                                                     Evaporative
                                                                                     Canister

-------
                              3-46
The  outlet  into  the  vapor  line  also  incorporates  a  0.055"
limiting orifice.   The vapor line to the  canister  is  both steel
(1/4"  I.D.)  and  fuel  resistant  rubber  (5/16" I.D.).  The  1.5
liter  open bottom canister is  located under the  vehicle hood,
even  though  no  vapor  line  is  used  from  the  engine   to  the
canister.  It  is  worth  noting that this  particular  fuel  system
did  not  incorporate an anti-spitback  valve  even  though  this is
becoming  an  increasingly  common  feature  in  many  other  fuel
systems.   The  fillneck  cap  had  a  warning to  loosen the  cap
slowly to  allow the tank  to depressurize before  fully removing
the  cap in order to prevent fuel spurting,

     b.    Modifications to Make on Onboard System

     1)    Introduction

     Keeping  in  mind the  goals  of  this program  as described
above, only minor modifications  of the stock system were needed
to  incorporate refueling  controls.   A picture of the modified
system and a  labeled  sketch showing  key components are shown in
Figures   3-22  and   3-23.    The  changes  needed  to  make  an
integrated onboard  refueling/evaporative   control  system  are
listed below  and  described  in  more  detail  in  the  discussion
which  follows.  The  changes/modifications  to  the  stock system
included:

      •  Fillpipe         -   Remove the  external  vent line
                              Slightly   reduce   the    minimum
                              diameter in the  fillpipe
                           -   Add    a    current     production
                              anti-spitback  valve to  the end of
                              the fillpipe  in  the  tank

      •   Vapor Control     -   Replace  the current  rollover valve
          Valve                and   limiting  orifice  with  a
                              multi-function   valve   with   a
                               larger  orifice

      •   Vapor Line       -   Shorten and  slightly enlarge  the
                              vapor vent  line.

      •   Canister         -  Move to the   rear  of the  vehicle
                               and enlarge  canister from 1.5  to
                               2.5 liters

      2.     Description of Modifications

      a)     Fillpipe Related

      Three modifications were  needed  for the  fillpipe.   First,
 since the design used  envisioned a liquid seal and tank mounted
 vapor  control valve,  the  external   vent   line  was  no  longer
 needed.    This  piece  was  removed, and the ports  in  the  tank and

-------

-------
                                           Figure  3-23

                             Integrated Hefueling/Evaporative Control System
Vehicle
Rear
                                             Vehicle
                                            Front  _
                      2 . 5 Liter
                Evaporative/Refueling
                      Canister
Purge Line
                 Fillpipe
               (1 1/4"  I.D.)
                                                               Vapor Vent Line (V I.D.)-  3  ft.
                                                                          Sending Unit
                                                                                 Rollover/Vent Valve ami
                                                                                 Fill Limiter with a
                                                                                 0.350" Orifice
                                                             Anti-Spitback Valve in
                                                             internal extension.of
                                                             fillpipe
                             Hose Clamp

-------
                              3-49
fillneck  were  sealed.   Consequently  an  exposed  steel/rubber
vapor line in  the  crash zone was  removed and  four  connections
were eliminated.

     Second,  very  minor modifications were  necessary  to ensure
vapor would not  escape  the fillneck during  refueling  under  any
reasonably expected  in-use  dispensing  conditions.   The fillneck
seal system developed by EPA employed a liquid seal.   That  is,
vapors  were  prevented  from escaping  to  the atmosphere  due  to
the  consistent flow of  fuel  being dispensed and  the  formation
of  a liquid  seal  within  the  fillneck itself  by  the  dispensed
fuel  slightly  backing  up  in  the  fillneck during  refueling.
This  is precisely  the  situation  which  exists in many current
systems  during  refueling,  and  is  the reason  why  an external
vent line is needed  to vent refueling vapors.

     EPA's  initial  testing  found  that   the   unmodified  stock
system  provided a  liquid  seal  over  the  initial  flow  rates
tested   -  6   to  10 gpm.   The   system  was   later   tested  at
dispensing rates as  low  as  3 gpm and  it was  found that a slight
restriction  was needed  in order  to ensure a  liquid  seal over
the  full range  of  reasonably expected in-use  flow rates (3-10
gpm).   A  restrictor which reduced the  diameter of  the stock
fillpipe by  0.2 inches  (from  1.25" to 1.05")  was installed in
the  rubber section  of  the filltube adjacent to the tank.  With
this  restriction,  the  system   filled   satisfactorily,    and
controlled  refueling vapors,  at dispensing rates of  3-10 gpm.
Flow rates below  3  gpm were not  tested  because  they are very
uncommon in-use and most nozzle  manufacturers  recommend flow
rates be above this  value  so that the nozzle automatic  shut-off
function works properly.  The  10  gpm  value  is  near  the high  end
of   the  current in-use  range  and  represents  EPA's  proposed
in-use  maximum.

     As part  of EPA's  proposed test procedure,  essentially  all
fuel spitback  during  refueling must be  eliminated in  order to
comply   with   the  proposed  emission  standard.   The  unmodified
stock   tank  was  subject  to  premature  shut-offs  and  spitback
during  refueling under  some conditions,  and this problem had to
be   solved.   This  is   a common  problem   in many  fuel  systems
today.    A  similar problem,   fuel  spurting,  occurs  upon  cap
removal when  fuel in the  fillpipe is subject to  high fuel  tank
pressures.   As  was discussed  in Section  B of  this   chapter  a
number    of    devices    are   available    to    address    this
spitback/spurting  problem.  EPA simply chose  one such device,  a
Chrysler anti-spurting valve which has been in production since
 1984,[8] made  some  minor modifications,  and attached it  to  the
 stock fillpipe.   The  ball-in-cage type  device shown  in  Figure
 3-24,  was  modified by increasing  the  specific   gravity  of  the
ball to be the same as liquid gasoline,  and it  was attached by
 removing a three  inch  piece from the end of  the stock fillpipe
 and replacing  it with a  three  inch piece which  included  the
valve.    This  valve effectively  stopped  fuel spitback  during
 refueling, both with premature shut offs and  at  the  end  of  the
 refueling event.    Thus  fill  characteristics  were improved as
 compared to the stock fuel system.

-------
    STOCK
ANTI-SPITBACK
    VALVE
       Figure 3-24

    Chrysler Anti-Spitback Valve

-------
                              3-51
     b)     Vapor Control Valve

     The  next  modification  involved the  rollover/vent  valve.
The  stock system  employed  a ball  and plunger  type  approach
built beneath a  0.055"  limiting  orifice.   This design would not
provide adequate venting for an  integrated  onboard/evaporative
system  and  would  probably  not  be  sufficient  for  enhanced
evaporative  control.   Thus,  some changes  would be  needed  in
either  case.   A  tank  mounted  vapor  control  valve  for  an
integrated   system  would   ideally   need  to   perform   four
functions:   1)  venting  of  refueling vapors,  2)  fill limiting,
3)  rollover  protection, and  4)  venting of  evaporative  vapors.
The  last  three  of  these functions are performed by most current
production valves now in use.  Each of these is discussed below.

     First,  since  the  external  vent  line  was  removed  and  a
liquid  seal  existed in  the fillpipe, all refueling  vapors had
to  be  vented through the  valve.  Thus,  it  was  important that
the  valve be sized  to  allow full venting of refueling vapors,
at   a  tank  pressure which  would  avoid  premature  shut-offs.
After  conducting testing on  the system,  it  was determined that
at  a 10  gpm dispensing rate the  valve would need to  be able to
flow 2  scfm  of  vapor at  a  backpressure  of  4"  H20.   This
requires  an  orifice  of  about  0.35".

     Second, the valve  design used  needed to  incorporate  a fill
limiting  capability.   With  the removal  of  the  external vent
line and the need to enlarge the  fuel tank  outlet  orifice to
accommodate  higher vapor  flow  rates during refueling, the fill
limiting  design  of  the stock  tank was  no  longer   effective.
Thus a fill limiter was needed  which would close off the vapor
vent  outlet when  the tank was   full   thus  increasing tank
backpressure and allowing automatic  nozzle  shut-off  to  occur as
 in the  present  vehicle.

      Third,  to  provide  in-use fuel  system integrity, the valve
would need to incorporate  rollover  protection functions such as
those provided  by the stock rollover valve.   Ideally, the valve
would provide protection for partial and total rollovers.

      Finally,  since  this  was  an integrated  system, the valve
would  also  have  to  be  to vent   evaporative  vapors   to  the
 canister under  a  variety  of  fuel  temperature and volatility
 conditions.   Since  the  preferred   design  incorporated   a  fill
 limiter it would  have  to  be able to vent even when  the tank is
 full.

      As was discussed  in  Section  B of  this chapter,  rollover
 valves of  various  sizes  and designs  are in common use today.
 They vary  in  size, design  and  functional  approach.  The valve
 selected  for  this  project  is   that used  on current  Ford  350
 pickups, vans  and  ambulance chassis  (see  Figure 3-25).   This
 particular  valve  was  selected  because the orifice  was  larger
 than on  most  other valves (0.17")  and  the  rollover  mechanism

-------
                                                        I ws . mm
                                                        HH	
                                                        A;'.-;..'.";, . ,r^^
                                                          BBH , g ••/'.:

                                                          :. :'":--:-:. . •
                                                            •• E !
                                                                «G    BSalH
                                                                •^BmH^HBD  1
                                                                BH^9I' Im i
                                                                •HMjM     Wi

                                                                HmsmBMi
                                                             ^^r*H * ',
, V

-------
                              3-53
was  float  actuated.   In  fact,  the  valve used  was  originally
produced as  a recall fix which would provide more  tank  venting
and reduce tank pressures.

     Several modifications to the valve were necessary to  allow
it to  serve  the  functions  described above.   These modifications
included.

     •     enlarging the vent orifice from 0,17"  to 0.35"
     •     plugging  one of  the  two vent  ports  on the  valve
           since only one was needed*
     •     enlarging the vapor flow passages within in the valve

     Using  the  stock  grommet  designed  for  the  valve,  the
modified valve was then mounted on the fuel tank  at a point and
height  such  that the original tank  capacity was retained.  The
valve  was positioned  on  the tank  near  the  position  of  the
previous  external  vent  line to as  closely  as  possible simulate
the refueling characteristics of the stock system.

     Finally, since  this  is  an integrated approach,  the system
would  ideally need  to  be able  to  vent  evaporative emissions
under  all  conditions.   The modified valve, as  described above,
would  serve this  function under  all conditions  except  perhaps
when  the  fuel  tank  is full.   While the  need  for full tank
venting  is a relatively rare  occurrence,  two  approaches can be
used,to  accommodate  fuel  tank  venting  under  these conditions.
The  first involves  further  modification of the  existing valve
to incorporate a  .055"  bypass within the  valve body itself.  To
do this,  a two stage valve could be used.   The  main .350" vent
shuts  off  upon  tank   filling,  while  the  smaller  .055"  vent
remains  open  under  all  conditions  as  it  does  on the  stock
tank.   This  .055" vent  is  also  rollover  protected.  A schematic
of this type of  valve  is  shown in  Figure 3-26.   Since this two
stage  valve vents fuel  tank emissions under  all scenarios, it
is  not  necessary to  have the  vent in  the fuel  sending unit
which  vents  evaporative  emissions  on  the  stock  system.   For
this   reason,  the  present  fuel  vent  line  could  be   removed.
Alternatively,   the   second  approach  involves  retaining  the
current  rollover/vent  approach  and teeing the  vapor line  from
this  outlet  into  the  vapor  line from  the refueling  port.   A
sketch  of  this   approach is  shown  in  Figure  3-27.    Either
      The  stock valve  incorporated  a  pressure  relief  function
      which  allowed  uncontrolled  vapors  to  vent  at  fairly  low
      tank  pressures.   This was  deleted  since  a  larger  vent
      diameter   in  the  main orifice would  prevent  significant
      tank  pressurization.   Deletion  of  this  pressure relief
      valve  would provide  some  improvement in emissions  control
      performance  as  well,  since all evaporative emissions  would
      be routed to  the  canister.

-------
          .055
H'W
                                                          .2X>
                                                                 OUTLET
                                 •I.IOO
                                  X£
                                            \
                                                         FLOW
              Two  Sta^e Vapor Control Valve
                     Figure 3-26

-------
                                          Figure 3-27
                      Integrated  Refueling/Evaporative Control System
                      	With  Separate  Evaporative Orifice	
Vehicle
                                           Vehicle
                                           Front
                      2.5.Li.t:rr
                Evaporative/Kefuc]in
                     Canister
Purcje Line
                 Fillpipe
               (11/4"  I.D.)
                                                             Vci[x>r Vent Line (V I.D.)  - 3 ft.
                                                                     Tied-in  Evaporative Orifice
                                                                        Sending Unit
                                                                               Rollover/Vent Valve and
                                                                               Fill Limiter with a
                                                                               0.350" Orifice
                                                            Anti-Spitback Valve in
                                                            internal extension of
                                                            f i]
                            Hose Clainp

-------
                              3-56
approach would  provide tank  venting under  all conditions  and
could have been  implemented easily  into EPA's  prototype  design
evaluation program.

     Thus,   with  a   few  simple  modifications,   an  existing
rollover/vent  valve  was  modified   to  perform  all  of   the
functions  needed for   a  refueling  vapor  control  valve.   The
valve is designed and  sized  to vent  refueling emissions,  and
the  float  will  provide fill  limiting capability.   In fact,  the
larger diameter vent orifice  will provide greater  flow capacity
for  evaporative  vapors  and  will  thus  provide  the means  to
significantly reduce  in-use  fuel  tank pressures  and thus  the
possibility of  fuel  spurting as compared to the  stock  system.
The benefits associated with  lower tank pressures  are discussed
in Chapter 4.

     And  finally, since  the basic  geometry of  the  valve  and
mechanism  used  to  provide  rollover protection   in  the  stock
valve are  unchanged,  the valve  would provide  both  partial  and
full  rollover  protection,  just as  it  did   in  its  original
configuration.   Discussions  with  the  supplier   of  the  stock
valve  indicated that  they  saw  no  problem  in  closing off  the
0.35"  orifice   in  a  rollover   situation.   Overall,  the  tank
mounted float activated vapor control valve is  fully functional
in  handling  all fuel  tank  vapors.   Given  the  similarity  in
design  and functional  operation to  current valves,  it  should
provide at  least the  same level of safety in-use as  is now seen
in  present  systems.   In  fact,  the greater  vapor  flow capacity
for   evaporative  vapors   provides   protection   against  fuel
spurting as compared to the stock valve.

     c)    Vapor  Line and Canister

     Because  of  the  increased vapor flow rate  during refueling
relative  to  evaporative  flow  rates,  there  was   a  need  to
slightly  increase  the vapor line  diameter.    However,  for   a
given  tank backpressure, during refueling,  the actual diameter
needed  depends  upon  the  length  of the  vapor  line used.    A
shorter  vapor line would  allow a smaller  diameter.   The vapor
line length,  in  turn,  depends  almost  entirely  on canister
location.    There  are  basically   two  options  for  canister
location;  the front  of the vehicle  under the  hood  or the rear
of  the vehicle  in  an  underbody location or in  a rear  quarter
panel area.

     Most  vehicles  today locate the  evaporative canister under
the  hood.   This  design  originated with the need to  collect hot
soak  emissions   from   carbureted   vehicles   and   to   purge
evaporative  emissions  from  the canister  into the  engine.   It
also required less total  vapor and purge  line than would any
other  location.  However,  since the majority  of engines today

-------
                              3-57
are fuel  injected  and  require no   direct venting, vapor  line
lengths are now  governed by fuel tank  emissions,  and  there  is
no compelling  reason why  canisters have  to be front  mounted.
By locating the  canister  in the rear of  the vehicle the  vapor
line length would be less.   As is discussed below,  one of GM's
current vehicle models uses this approach.

     Given  this  design option,  EPA  selected  a  rear-mounted
location for  its canister.  This led  to vapor  line length for
the  integrated  refueling/evaporative  system  of  about  three
feet,   only  about one third of that used  for  the  stock system.
This also allowed a vapor   line of  only 1/2"  I.D.  versus  5/16"
for the stock  system.   An even  smaller  diameter vapor  line
could  have  been used,  but  1/2" I.D. was  selected  to  optimize
backpressure  within  the  total  system (vapor  control  valve,
vapor  line,  canister).   Testing  by other  commenters  suggested
that a l/2"-5/8" I.D.  vapor  line would be  needed for  an under
hood canister location.[9]

     Finally, the canister  itself,  shown in Figure  3-28,  was a
GM  2.5  liter  closed  bottom  canister,   the  size and  design
currently  used   on  large   light-duty   trucks  and  heavy-duty
gasoline  vehicles   for  evaporative  emissions  control.   Some
modifications  were needed to  improve  vapor  flow   into  the
canister and  charcoal  bed, but basically  the  canister  was used
as produced.   No efforts  were made to optimize  canister size
relative  to  the  expected  refueling  emissions   load  or  to
accommodate  the  expected  formation  of   a   heel.   Finally,
depending on  the canister  design used and  the  location on the
vehicle some  type of dust  cap may  be needed to  clean the purge
air.   However,  this would be a minor design modification to the
basic  canister,  as is  shown  in the  discussion  later  in this
section.

     d)    Design Applicability to Other Vehicles

     As  was  stated  in the beginning  of  this  discussion,  the
goal of this  program was to demonstrate the concept that it is
possible    to    construct    a   simple    integrated   onboard
refueling/evaporative   control  system  which  does  not   increase
the  complexity  of  current  systems,  using  components  based on
current production hardware.   As can be seen  in the discussion
above,  a  system was constructed which  required  essentially only
minor   modifications   of   currently  available   hardware,   and
complexity was  not  increased.

     While  this  concept  was  demonstrated for only  one fuel
system, EPA sees no  technical  reason  why  it  cannot be  applied
to  essentially all current fuel  systems.   There  is no apparent
reason why some  form of liquid  seal  cannot be used.   Most of
today's vehicles essentially have liquid  seals  now based  on  the
widespread  use  of  external or  internal  lines  to vent  refueling
vapors.   Vent valves  serving  the same  or similar  functions as

-------
Figure 3-28

 Stock 2.51
 CM Canister

-------
                              3-59
that required  for  this  system are used on  current  vehicles  and
the  relocation of  the  canister   is  a  straightforward  design
change.   Thus  while the system would need  to  be  engineered  and
optimized  for   each vehicle  model,  the  design  concepts  and
approach demonstrated have wide if not universal applicability.

     With this  discussion of the  overall  refueling/evaporative
control  system  developed  by  EPA  and   description   of  the
components  used in  the system the  next section  describes  the
results attained in tests to control refueling emissions.

     4.     Emission Test Results

     After  constructing  the  initial  prototype  system,  tests
were conducted to see  how well the system controlled refueling
emissions.   Tests  were conducted  on the  entire  fuel  system
using  the  procedures  and  conditions  prescribed  in the test
procedure portion  of  the NPRM.  All  tests were  conducted at a
nominal  9 psi  RVP,  and  the  canister was  bench  purged between
tests.  The  results of  the testing are shown  in  Table  3-2,  and
while self explanatory, several points are noteworthy.

     The  first  important  point  is  that  all  tests  yielded
emission  results   lower  than EPA's  proposed  standard  of 0.10
g/gal.  Average  emissions  for  the  thirteen  tests  were 0.04
g/gal,  less  than  half  the  level  of  the  proposed  emission
standard.   Second,  it  should  be  noted  that  the  liquid seal
worked  effectively  at  a  wide range of  dispensing rates.   A
dispensing  rate of 10 gpm was selected as  the high end based on
current  in-use practice and EPA's proposal to limit dispensing
rates to  this  level in-use.   On the  low end dispensing  rates of
3,4,  and 5  gpm were evaluated to  assess   the  effectiveness of
liquid  seals  at  lower  flow  rates.   The  3 gpm  value  was also
evaluated  to  see  if   the  fill  limiting   function  and  nozzle
automatic  shut off mechanism would work at low flow  rates or if
the  tank would overfill.  In  all cases  the system worked as
designed.   Finally,  a  third key  point  is that  the relatively
simple  anti-spitback valve   incorporated  into  the  system  was
very effective.    There  was  no  fuel  spitback   either  during
premature  shut-offs or  at final, shut-off  for any  of the above
tests.

     Subsequent to the  completion of this  testing  EPA presented
a  series  of briefings  to  interested parties  on  the details of
the  system development  described  above  and the  results of  the
emission  testing.   while  EPA has  not as  of yet  received  any
significant    comments   on   the    program,   its   results,   or
implications,   it   was   suggested   that  additional  testing  be
conducted.   Further testing  has  been conducted with the  use of
higher RVP  fuels,  higher  fuel tank temperatures, different  fuel
nozzles,   different   tank   angles   (slopes),   and   with   the
evaporative  orifice teed  into the refueling vapor  line.   In  all

-------
                                                   Table 3-2

                                 Test Data Results for Onboard Control System
Test
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
Fill
Rate
(qpm)
10
10
10
5
5
5
5
5
10
10
5
4
3
Emissions
(q/Gal)
.01
.02
.03
.05
.Ob
.06
.04
.05
.08
.04
.04
.01
.02
Average
Tank
Backpressure Premature Shut-Off
(in Hg) Shut -of fs Mechanism
.38
.38
N/A
.30
.12
.13
.12
.06
.20
.30
.09
.05
.03
N/A
None
N/A
2(a)
None
None
None
3(a)
3(a)
None
None
None
None
N/A
Manual (e)
N/A
Automatic
Automatic
Automatic
Automatic
Manual (b)
Manual (b)
Manual (b)
Manual (b)
Manual (b)
Manual (b)
Gallons
Dispensed
(c) Comments (d)
12.8
13.2
12.8
13.5
13.6
12.8
12.8
12.8
12.7
13.0
12.8
12.8
12.9



Filltube restricted with clamp
3/8" Vapor Line
Filltube restricted with clamp
3/8" Vapor Line
Filltube restricted with clamp
3/8" Vapor Line
Filltube restricted with clamp
1.025" restriction in filltube
1.025" restriction in filltube
1.050" restriction in filltube
1.050" restriction in filltube
1.050" restriction in filltube
1.050" restriction in filltube
(a)  Shut-off occurred at the start of the refueling event.
(b)  Nozzle shut off automatically upon topping off.
(c)  This does not include the 1.5 gallon initial (10%)  fill.   Nominal tank capacity is 15 gallons.
(d)  All tests were performed using a 1/2" vapor line and 1  1/4"  minimum fillpipe diameter unless noted.
N/A  Data not taken.

-------
                              3-61


cases  the system  has  performed as  expected.   The  only  real
finding,  although  not  surprising,   is  that  higher  RVP  fuels
require a larger canister.

     5.    Conclusions

     The  goal  of EPA's program  was  to demonstrate the  concept
that it is possible to construct  a simple,  yet  fully  functional
integrated    onboard    refueling/evaporative   system    using
components based on  current production  hardware.   As  can  be
seen  from the  discussion above, a  system very  much like  the
present   evaporative   control  system  was   constructed  using
currently  available  production  hardware,  which  for the  most
part  required  only  minor   modifications.   No  complexity  was
added  to  the  system  and in  fact the system was  simplified as
compared  to  the  stock  system;  the  external  vent  line  was
removed,  connections  were eliminated and the vapor  line to the
canister  was   shortened  significantly.   In   addition,   fill
performance  was improved by  the elimination  of the  spitback
which  occurred  in  the stock  system,  and  it is  reasonable to
expect  that  improved venting  and the anti-spitback  valve would
reduce  fuel  spurting  problems  related  to  high  in-use  tank
pressures.   The emission  test  results  shown  in   Table  3-2,
clearly   indicate  the  ability  of   the  system  to  meet  EPA's
proposed  refueling emission  standard with  a comfortable margin
of  safety.   These results  were accomplished with minimal time
and  resources   and   improvements  to  the  system  are  clearly
possible.  EPA  made no  real  attempt  to  improve component design
or in  any way to optimize system  performance.

     EPA  expects that because of cost and  design considerations
manufacturers  will be  motivated to. adopt the  simple  approach
presented here  and apply its concepts to  their  different  fuel
systems,  with  modifications  and improvements  engineered  into
these   components  and  systems   just  as   they  are  for  other
systems/components  used  on  the  vehicle.   For example,  just as
manufacturers    use   different   rollover/vent   valve   design
approaches  on their current  vehicles,  some manufacturers could
choose a valve approach  which is a  variation  on,   but not
identical to,  that  used by EPA.    However,  with   or  without
onboard  controls,   in   order  to   address   the  proposed   test
procedure modifications,  EPA  expects  that  manufacturers  will
have  to   incorporate  a  valve  that  provides   adequate   tank
venting,  thereby resulting  in low  tank  pressures.   Similarly,
just as  vapor  line  diameter  and   canister  locations  vary on
present  vehicles  EPA  also  expects  some  variation in  these
design features for  an  onboard  control  system.   However,  upon
full evaluation,  for simplicity and cost  considerations,  EPA
expects  that   manufacturers  will likely   select  an integrated
approach  over  separate  control  systems  .   Thus  while systems
identical to EPA's may not  be  used  by  all  manufacturers,  EPA's
development  work  suggests  that integrated  systems  with  tank
mounted valves and liquid seals are  the preferred  approach.

-------
                              3-62
     As  is  shown in  Appendix I, it  should  be noted that  many
manufacturers   presented  non-integrated   systems   in   their
comments.  A  dominant reason for this  choice was  the  proposed
refueling test  procedure,  which manufacturers believed favored
non-integrated  over  integrated  systems.    As  proposed,  they
believed  the   test   procedure   requirements  imposed   a  less
difficult  vapor  purge  requirement   on  non-integrated  systems
compared to  integrated systems.  However,  as  was  discussed  in
the NPRM, from  an environmental perspective,  integrated systems
have  definite   advantages   in   terms  of   controlling  excess
evaporative  emissions.    In  the   notice   that   announces  the
availability  of this  report,  EPA will  propose  test  procedure
modifications that  will insure  that  either  approach results  in
systems   that  deliver   equivalent   in-use   performance.    If
non-integrated  systems  are  to  be  used,  the  proposed  test
procedure changes will  insure that they provide  the same in use
performance   as  integrated  systems.   These  test  procedure
changes  are  likely to  eliminate any perceived  purge  advantage
of  non-integrated versus  integrated systems  in terms  of  test
procedure requirements  and thus lead manufacturers to  further
consider  the  advantages  of  an  integrated approach  discussed
below.

     Given  that many of the  system designs  presented  in the
comments  did not incorporate  the  concepts  used  by  EPA,  one
might  ask what  elements  of  the approach  would  ultimately make
it  preferable.   In  general,  these can  be  broken into  four
areas:    design  improvements,   cost,  packaging,  and   systems
engineering.

     First,  as  will  be  discussed further  in Chapter 4, the EPA
concept  is no more  complex than current evaporative systems and
provides  improvements by  removal  of the  external  vent line,   a
reduction  in  fuel  spitback  and  spurting,   a  shortened  vapor
line,   lower   fuel   tank   pressures,  and  control  of  excess
evaporative  emissions  which are  often  now  vented in engine
compartments.   Depending on one's view of the complexity issue,
these  design improvements  may  also  have some  positive safety
implications.   Second,  in terms of  cost,  an integrated system
is  less  expensive than a separate  system.   It requires only one
canister,  one vapor  line and one purge line,  versus two  of each
for a separate  system.   The use  of a  liquid  seal  avoids all
cost  associated  with mechanical seal  designs.   Also,   avoiding
the use of  these components reduces weight  and  improves fuel
economy  relative  to the  separate  system.  Closely related  to
cost  are advantages  related to  packaging  and engineering.  Use
of  an approach similar to that shown  by EPA will  require that
only  one canister,  purge line, and  vapor line  be packaged  on
the vehicle  instead  of two as  would be required  on a separate
system.   Finally,   in  terms  of   engineering,   developing and
implementing this type  of  system  into  the vehicle  is  likely  to
be  a  simpler task than  with  other  approaches.  There  are  fewer
components,   fewer   connections,   and   integrating  one   vapor
control  system into  the  vehicle should be  an easier  task than
is  two.

-------
                              3-63
     In summary,  an objective evaluation  of  the pros  and  cons
suggests that  the  use of  the  concepts   shown  by EPA have  a
number  of  advantages.   They  could  be  implemented  on  most
vehicles  and  would  likely  require  no  more  time  and  other
resources than systems suggested by the commenters.

     6.    Analysis of Comments

     a.    Introduction

     Chapter  2  and  Section  B  of  this  chapter provided  key
background  information on  the  design  and   safety  of  current
fuel/evaporative   control   systems.     Taken  together   these
discussions  show that despite the  wide range of  complexity in
current  fuel/evaporative  control  systems  there  is no evidence
that there is  a  safety problem with evaporative control systems
in  general  or that  the  increased  complexity of  some systems
designs  has  affected their  in-use  safety.   Furthermore,  the
analysis  in  Section  B of  this chapter  showed that most safety
concerns   regarding  various  onboard  system  components  were
essentially   rendered moot   by   the  fact  that  very  similar
components  were  now in-use,  presumably  safely,  on  current
evaporative  systems.   Given the  important perspective provided
by  this information  on  evaporative  system  design  and safety
plus  the  description  of the design,  development and testing of
EPA's  onboard control system concept, presented  above,  we turn
now  to  an   analysis of  the  technology  and  safety comments
regarding   onboard   control   systems.     These   comments  were
summarized previously in Section A  of this chapter.

     The   analysis  of  comments   is  generally broken  into  two
sections.  The  first portion addresses general comments  received
regarding   onboard   technology  and  safety  while  the  second
addresses  specific technology and safety concerns  related to
design and hardware.

     b.    Analysis of General Comments

     The  general concerns  regarding onboard system  technology
and  safety can be  summarized  by the statement  in  Section A of
this  chapter that  "onboard vapor recovery systems would be more
complex than  current evaporative  control   systems   and  that
adding this  complexity  would cause  an unquantifiable  increase
in the risk  of both crash and non-crash fires."  The  discussion
then  goes  on to cite complexity  and  risk  factors such as larger
components,   more   components,   more   connections,    component
materials,  component  locations,  and  indirect effects on other
systems.    The  commenters   then  provided  various   potential
systems designs to demonstrate the possible  complexity involved
(see  Appendix I).

      Two  key points need  to  be made in response to this  general
concern.   First, one must  question the premise.  What  evidence
is  there  of  a relationship  between  safety  and complexity?

-------
                              3-64
Since evaporative control systems  are  presently the best analog
for onboard vapor recovery  systems,  it is valuable  to  consider
what  can  be  learned  from  the  experience  with  safety  and
complexity for  current evaporative  system  designs.  First,  as
was discussed  in Section  B  of  this  chapter,  there is  a wide
range of  complexity in  current  fuel/evaporative  systems.   The
size  of  the  individual  components,   the  overall  number  of
components,  and  the  number  of  connections  required  varies
greatly  among  the  designs.   Also,  a  variety  of  different
materials were used  to make these components and  the locations
of the  components on the vehicles varies with essentially every
model.  Yet despite  the wide range of diversity  and complexity
in these  evaporative system  designs,  Chapter 2 found  very few
evaporative   system   safety  problems   and  there   was   no
relationship evident to suggest that  any one  design  or design
approach was any more or less safe  than any other.   Thus, the
available  evidence  suggests that  the  current in-use  systems
most  analogous to onboard  vapor  recovery systems  incorporate a
wide  range  of design  complexity  with  no  evident effect  on
safety.

      Second,  if there  is  still  a  concern about  complexity,
perhaps  the  most   important  point  demonstrated  by  the  EPA
onboard  system  development  program  discussed above  is  that
onboard  vapor   recovery systems  need  not  be  any  more complex
than  even  the  more  simple current evaporative  systems  and that
onboard could  indeed be viewed  as an extension  of evaporative
systems.   A visual  comparison  between  Figures  3-21  and 3-23
clearly demonstrates this point.  Very  few changes were needed
to  incorporate onboard  refueling  controls  into the  stock fuel
evaporative  system.   The concepts of  operation are essentially
the same  and essentially every  change necessary  to incorporate
refueling  controls  brings  with  it  a   design improvement  as
compared to  the stock  system.   Depending on one's  view of the
complexity  issue,  these design  improvements could bring  safety
benefits  as  well.    The  addition  of  an   anti-spitback   valve
protects  against fuel  spurting upon cap   removal  and reduces
spitback  during refueling.   Use of  an  enlarged  vapor control
valve allows  fuel  tanks  to  vent  more freely  at  lower tank
pressures   and  this   also   reduces  possible  fuel  spurting
problems.  Eliminating the external  vent line  closes off  a 1/2"
passage  from  the  tank  now unprotected  in  the   event   of   a
disconnection.   Locating  the   canister  in the   rear  of the
vehicle   allows  a   much   shorter   vapor   line,   and   allows
uncontrolled emissions to  be vented in a safer place than under
the  vehicle hood.    Finally,  increasing   canister  size  improves
the   ability  to   capture  excess  evaporative emissions  thus
preventing their escape to the atmosphere.

-------
                              3-65
     Overall,  an  onboard  vapor recovery system need be  no  more
complex and in  many cases could  be less  complex than  current
evaporative  systems.    Furthermore,  the  results  of   the   EPA
onboard design and development program  demonstrate  that  onboard
systems need  not  be as  complex as  suggested by  many of  the
commenters  design   concepts.    Much   simpler  approaches   are
possible.    EPA  recognizes  that  our  emission  standards   are
performance  standards  and  that   we   cannot   dictate  design.
However,  simplicity  concerns  and  cost   considerations   will
clearly encourage manufacturers  to evaluate  and apply  design
concepts  and   approaches  similar  to those  put  forth   and
developed  by  EPA,   More  complex  systems  are  not  inherently
driven  by  an  onboard requirement.  If  more  complex designs are
chosen, we expect manufacturers to  safely  incorporate  them  into
their  vehicles  just as  they have  done with the wide  range of
complexity in evaporative system design.

     In  addition  to   the  more   general   concerns  expressed
regarding onboard technology and safety,  EPA  received a number
of  safety comments  related  to  specific designs and  hardware.
The  comments  in these  areas  are addressed next  in essentially
the  same order as present in Section A  of this chapter.

     c.    Analyses of Specific Technology and Safety Comments

     1)    Fillneck Seals

     The onboard  system presented by EPA in  the NPRM included  a
liquid  seal   approach   to   controlling  refueling  emissions.
Despite EPA's  view that  this system  would be  preferable   from
both a cost,  performance  and safety  perspective  a  number of
commenters cited  reasons  why a  liquid seal could  not be used,
but  essentially no data  were provided  to  support  their views.
The  concerns  cited by  the commenters  included:  1)  ability to
pass the  proposed  standard,  2)  ability to  form a  seal at low
flow rates, 3)  increased  refueling  emissions,  4)  an increase in
spitback  spills,  5)  difficulty in packaging, and 6) fill height
limitations.  Each of these  is addressed below.

     First, the results of the  EPA test program presented  above
plus information developed  in test  programs  conducted by two
different  commenters all  clearly demonstrate  that  liquid   seal
systems  are  capable of  providing  the control  needed  at the
liquid fillneck  interface.   Several different  fuel  systems  have
been fitted  with a  number  of  different  liquid  seal  systems
(J-tube,  flowing  liquid seal,  submerged fill)  and all have  been
capable  of  reducing   emissions  to   a level  well   below the
proposed  standard.[9,10,11]   Concerns  raised  by the  commenters
regarding  fuel evaporation  or vapor  diffusion  at  the  fuel/air
interface did not manifest themselves  in the test results.   The
liquid seal  approach  has  been  demonstrated  to   provide the
sealing protection needed in the  fillpipe.

-------
                              3-66
     Second, test results from EPA's  recent program and that of
one  of  the  commenters  showed  that  liquid  seals  could  be
effective over a wide  range  of  dispensing rates.   EPA's program
was  capable  of  controlling  emissions   at  dispensing  rates
ranging from 3 to  10  gpm while the other commenters program was
effective  in  a  range of  4  to  11  gpm.[12]   These  results
demonstrate  that  with  proper   design  liquid  seals  can  be
effective at any reasonable lower dispensing rate.

     Third,  the  concern that  the use  of  a liquid  seal  would
increase  refueling  emissions  and  lead  to  the  need  for  a
significantly  larger   canister   appears   largely  to  be  an
assertion  not  based  on test data.   As was discussed,  most of
today's  vehicle  fuel  systems  use either  internal  or  external
vent  lines  to  route  refueling  emissions,  so  essentially by
definition  these vehicles  now fill  with a  liquid seal in the
fillneck.   Thus,  there  should  be  no  major   impact  on  the
refueling  emissions  load  to  the canister  by  using  a liquid
seal.   The  only  extremely  small incremental  affect  might be
related to  the nozzle aspiration and entrainment  of  all air in
a  liquid seal system  versus  air and vapor  in a current system
which  uses  a vent  line.   The only  real  exception would  be in
cases where the  fillpipe diameter is relatively  large  (>2") and
refueling   vapors  are  routed  out  the  fillpipe  as   fuel  is
dispensed.   Designs   of  this  type  are  uncommon  on  today's
vehicles  and will  be  rare in  the future  as  the  trend toward
side fill continues.

     EPA  acknowledges  that  vapor generation could  be reduced
with  a mechanical  seal approach  and  that  this  could  allow  a
small  decrease  in  canister  size.    While  a mechanical   seal
approach  may  be  desirable  from  this  point  of  view,  as is
discussed  in the initial safety report  (Appendix II),  there are
cost   and   other   performance  trade-offs  involved  with   this
approach  and  it  is  expected  that  their  use  will not  be as
common as now depicted in the comments.

     Fourth,  since most  fuel  systems  now  encounter  a  liquid
seal during refueling, the concern that an  onboard  system  using
a  liquid  seal would  inherently  increase  spitback  spills is
unfounded.   In fact,  the degree  to  which spitback spills  are  a
problem  on  current  vehicles  would  be addressed  with onboard
equipped  vehicles  by  the  use  of   an anti-spitback  valve or
similar piece  of hardware.  As  discussed above,  the use of  this
hardware  essentially eliminated all  spitback  spillage  for  EPA's
onboard  prototype  system  and  would be expected  to be equally
effective   for other  fuel systems.   It  is  interesting to  note
that several manufacturers now use anti-spitback type  valves in
their  fuel  systems to address problems  related  to fuel spurting
upon  cap  removal  under high  tank  pressure.   An  anti-spitback
valve  design such  as  that presented  above would  help to address
this problem as well.

-------
                              3-67
     Fifth, manufacturers expressing concerns  about  packaging a
J-tube must realize  that  the NPRM  presented the J-tube  as  one
method   of  forming  a   liquid  seal.    Subsequent   work   has
demonstrated  that  it  is  not  needed  in  many  fuel _ systems.
Nevertheless,  other  options  such  as the  submerged  fill,  flow
restricting gate, etc.,  could also be used.

     Finally,   several   commenters   stated  that   fill   height
limitations  would  prevent   the use  of  liquid  seals on  some
vehicles.  EPA agrees that  simple  fluid dynamics  suggests  that
some  minimum  fill  height  is needed to make  liquid  seals  work
without  premature  shut-offs.  However,  when considering  the
fill  height  concern,  one  must once  again  realize  that  most
vehicles  today  encounter  liquid   seals  during  refueling,  so
current  fill  heights must  be sufficient by definition.   Also,
as  the  trend toward  side   fill  designs  continues,   rear  fill
vehicles   with   inadequate   fill   heights   will  be  even  more
uncommon.   Furthermore,  the  required fill   heights  can  be
reduced   through   measures   aimed   at  reducing   turbulence,
entrainment,  and  backpressure,  and all  of  the  liquid  seal
approaches mentioned previously can be evaluated.

     EPA has  carefully  considered  the  concerns expressed by the
commenters  regarding the use  of  liquid seals.   The data  and
information available still  support EPA's  view that  liquid seal
approaches  could be used on most  if not  all vehicles  and the
cost,  durability,  tampering  and other  benefits of this approach
suggest  that  liquid  seals  will   likely   be  the  design  of
preference.   However,  as is discussed in the attached initial
safety report (Appendix  II), mechanical seal  approaches  can be
used   effectively   and  safely  if  a manufacturer  elects  this
approach  after considering  the trade-offs  associated with this
design option.

      2)    Vapor Control Valves

      Three basic valve designs were presented and discussed by
the  various  commenters.   These  included   a  fillneck  mounted
valve,   a  tank   mounted   mechanical   valve,   and   a  solenoid
activated  tank  mounted  valve.   Each   of   these  basic  design
concepts  was  also  addressed in the previous  EPA safety report
provided in Appendix II.   Comments regarding these  valves  are
addressed  below.

      Several   commenters   expressed   concern   regarding   the
crashworthiness  of fillneck  mounted  valves,  and  one commenter
suggested  that crash shields would be needed to protect  these
valves since they  would  be  mounted  in a  crash  zone.   As  was
discussed  in  Section B of this chapter, fillneck mounted valves
are  used on  several current fuel   systems with  no known  safety
problems.    This   evidence   suggests   that  any  manufacturer
electing to use  a fillneck  mounted valve  could  do  so  safely.
Furthermore,  a  review  of  the  shop manuals  for  the two  fuel

-------
                              3-68
systems  discussed  in  Section  B  indicates  that  neither  uses
fillneck crash shields.  Nevertheless,  if  a manufacturer elects
use  of these  shields  it  certainly  does  not  create  a safety
concern.   Valve   location  is a  manufacturer's  design  choice;
there  is no  inherent reason  why  valves have to be placed on the
fillneck or in other crash zones.

     A  tank  mounted mechanical  valve  similar to those  used on
most vehicles  today (see Figure  3-3)  was  discussed  by several
commenters.  This is  the  basic   approach  used  by EPA in  the
onboard prototype system discussed  above,  and  since  it  could
essentially  replace  the  current   valve,  there is  incrementally
no real change in risk.   Discussions with  several manufacturers
of these valves  indicate no  concern about  being able  to design
a valve to serve  the  functions  needed effectively  and safely,
and  since  manufacturers  perform  quality  control checks on  100
percent  of  their  production  of  these  valves,  high  in-use
reliability  would  be  expected   just  as   is  experienced  with
current valves.[13]

     With  regard  to  both   fillneck  and  tank  mounted valves,
concern  was  expressed  that  the  use  of  plastics  in  these
components reduced their crashworthiness.   The issue of the use
of plastics  was  addressed in Section B of  this  chapter, but it
is   clear  from   even  a  cursory  review  of  current   fuel  and
evaporative  system designs  that  fuel  tanks, fillneck pieces,
and  valves  are   made  of plastic.   This  clearly  suggests that
there  is no  inherent safety  risk  in their use.

     Finally,  with  regard   to   plastic   valves  and   solenoid
activated  valves,  concerns  were  expressed  regarding static
charge and electrical  relays.  Quite  simply, these concerns are
not  new to  onboard  controls.   Isolated plastic components must
be   grounded  just  as  they  are  in  present   fuel   systems.
Furthermore,   while  EPA  understands  the  concern   regarding
electrical  relays in  and  around  fuel tanks,  one  must  consider
that incremental to the widespread  use of electric fuel pumps,
potentiometers,  etc.,  within current fuel  systems  this concern
is not new or unique to  onboard systems.

     Thus,  in summary,  EPA's review of  the comments  regarding
vapor   control   valves  leads  to  the   conclusion  that   they
introduce  no  new  or  unique safety  concerns.    Any  potential
problems  with   these  valves  exist  on   current  vehicles  and
presumably have  been addressed satisfactorily.

     3)    Vapor Lines

     A number  of  commenters indicated concern that a  refueling
vent  line  requires  a  somewhat  larger   diameter   than  an
evaporative  line  and  this would  make  it  more prone to  puncture
or   rupture  in  an accident  or  problems  during  assembly and

-------
                              3-69
repair.  Several  points  should be made  in  response.   First,  it
should be noted  that  based on  EPA's  latest work  the  refueling
vapor  line  approximates  the  size of  the current  evaporative
line   it  would   replace.   Thus,  this  is  not   a  new   or
significantly  changed component.   Second, current  evaporative
lines vary  in  diameter by about  a factor of  2  to 3  (5/32"  to
3/8")  and there  is no  indication of more  problems  with larger
diameter lines.   Third,  as  is  discussed  in  SAE  Standard  J30,
rubber  vapor   line   wall   thickness   increases  with  inside
diameter,  so in contrast to the comments received, puncture and
rupture   resistance   should    increase.     And  of   course,
manufacturers  have  the option  to  use rubber, steel,  nylon,  or
some combination  of the  three materials within their vapor line
design  configuration.   Finally,  as  was  discussed  above,  in
conjunction with the  development  of  the EPA  onboard prototype
program,  a  rear  mounted canister   leads  to  a  substantially
shorter vapor  line,  and thus  can address any  perceived vapor
line safety concerns  as compared to the  current system.

     4)    Liquid/Vapor Separator

     Many  commenters   suggested that  a  liquid/vapor  separator
would  be  needed  in an onboard  system,  especially if the vapor
lines  slope downward  or  the canister is mounted below the fuel
tank outlet point for the vapor line.   However, the only safety
concerns  raised  were  with  regard   to  the  extra  connections
required with  some approaches  and an unsubstantiated assertion
that  due  to  its function  such a component  could  potentially
have all the same safety problems  as  a miniature fuel tank.

     As was discussed in Section B  of this  chapter,  fuel and
liquid/vapor  separators of  various   designs  are  in widespread
use  on  today's  vehicles.    In  addition  to  functional  design
variations,  they also   vary  in  size,  material,   number  of
connections,   and  location  on  the  vehicle.   Despite  these
differences, EPA is  not  aware of safety  problems with  any of
these  present  components.   The use  of liquid/vapor  separators
will  remain a design option with onboard  system just  as they
are  on current  systems.   Neither the  stock  system  nor the EPA
onboard  prototype  system  discussed  above   incorporated  this
function other than indirectly  through  separation caused by the
the  path  taken  by  the  vapor  enroute  to  the canister  and the
location of the canister itself.  Either way, the common use of
these  components  on  current  vehicles   without  safety problems
indicates   that,   if   desired,   these  components  can  also  be
incorporated safely into onboard systems.

      5)     Canisters

     This discussion  addresses comments received with  regard to
safety concerns  about  canister  size,   location,   and   other
packaging impacts.  More specific concerns regarding  the  safety
of the canister  itself  and vapor ignitability are  addressed  in
Section D of this chapter.

-------
                              3-70
     First, with regard to  canister  size,  there are a number of
factors which affect the size actually needed for  any given set
of fuel  tank size and  test conditions.   These  include  factors
such as  purge rate and schedule,  canister shape,  and charcoal
working  capacity.   Nevertheless,  canister size itself  is  not
really a  safety issue.   Current  evaporative  canisters vary in
size by  about a factor of  4, but  there is no  evidence  of any
relationship  between  size  and safety.   None  is  expected  for
refueling  canisters  either,   since  the  design, materials  and
technology  would  be  essentially  identical.    Other  canister
safety issues are addressed in Section D of this chapter.

     Second, with  regard  to canister location,   as was discussed
above, there are good design reasons to  place the  canister in a
rear  location,  but this  is a design  option  since an underhood
location  is  also  possible  and has  been the location of  choice
for  many manufacturers.  While  a  number  of commenters  stated
that  a  rear location  for  the  canister  would be  preferable,
several  also  stated  that  a rear  location would  have other
drawbacks  with regard  to  fuel  tank  size,  cargo   capacity,  or
head   space  in  the   passenger   compartment.    EPA's   onboard
prototype programs envisioned a  rear mounted canister location,
notably  in the left rear quarter panel  of the  vehicle,  similar
in  location  to that  of  a  current  vehicle.   General  Motors
packages  the  evaporative  canister  in  the  left  rear   quarter
panel  on its 1988 "W" body vehicles  (e.g., Pontiac Grand Prix,
Oldsmobile  Cutlass Supreme,  Buick  Regal,  etc.).  The canister,
shown  is  essentially a  slightly modified version of one  of  GM's
present  canisters.   The only modification is the dust cap added
to the bottom of the current  open  bottom canister  of the  same
size.  Figure 3-29 shows  the stock  canister, modified canister,
and dust  cap.

     A picture  of  the canister location  on the  vehicle is shown
in  Figure  3-30.   It is  interesting  to note that  the  vehicle
photographed in  Figure  3-30  was  a  dual-exhaust  vehicle   (as
shown  in Figure 3-31).   There was  adequate space  to package the
canister,  still accommodate  the  tailpipe and  muffler assembly
nearby,  and  not  affect  the  crush zone on  the vehicle.   Given
the  relatively small  size  of   this vehicle   model,  adequate
packaging  space  without   affecting   crush   zones  would  be
available on most if not all  other  vehicle models.   Similarly,
EPA  would expect  that  dual tank vehicles (only on  larger  light
trucks)   could   rear  mount  canisters  if  desired  since  rear
quarter  panel  space  is even greater  on these   larger vehicles.
While  this  is  in  the  "crash  zone,"  the  analysis presented in
Section  D  indicates that damage  to the canister  in an accident
presents no unique safety  concerns.   In fact,  a review  of  most
manufacturers  evaporative  canister  locations   indicates  that
they must be considered expendable  in accidents,  since many are
located   in  crash zones   in  underhood  areas.    This evidence
indicates rear  mounted canisters  are  clearly  feasible  from  a
safety view point  and that  manufacturers need  not  sacrifice

-------
 Figure 3-29

Front and Rear
Mounted Stock
  Canisters

-------

-------

-------
                              3-74


fuel tank size, cargo  capacity,  passenger compartment space, or
crush  zone   to   do   so.    Whether   canisters  are   front  or
rear-mounted  is  a  manufacturers  design  option;  each has  it's
advantages  and disadvantages.   Finally,  canister  location  is
not  a new  concern  brought about  by  an onboard  requirement.
Current   canisters   are   located  in   a  variety   of   places
(underhood,  fenderwells,  rear  quarter panels)  all  in crash and
crush  zones.   The  same  design  considerations  which apply  to
evaporative canister  locations  are  expected  to also apply to
onboard canisters.

     6)    Purge. System Components

     As was noted in the  summary  of  the comments,  manufacturers
cited  potential  purge  system   related  safety  problems  for
onboard systems such as purge valve  failure.   As was  the case
for  canisters and  other  components,  purge  system hardware is
used on all vehicles with evaporative  controls,  so  there are no
unique issues  for  onboard systems.   Purge system components are
expected  to remain  the same, with perhaps new calibrations, so
failure modes and  effects  would be the same as now exist for
evaporative  systems.   Onboard  controls  introduce   no  new  or
unique  safety concerns  in  this   area.   Incrementally there are
no  significant purge related safety issues  for  onboard systems
relative  to evaporative systems.

-------
                              3-75
D.   Other Safety Concerns

     In  addition  to  the  onboard  system  design  and  safety
comments summarized and  analyzed  above, comments  were  received
in several  other areas which  fell  somewhat outside  the  direct
scope of the discussion  above,   These  include  comments related
to  onboard systems  for  heavy-duty  gasoline vehicles  (HDGVs),
canister safety, the effect of onboard  systems  on  the number of
future  safety  recalls,  crashworthiness,  and failure modes  and
effects analyses.  Each of these areas  of  comment  is summarized
and addressed below.

     1.     Heavy-Duty Gasoline Vehicles

     a.     Introduction

     Relatively  few of  the many comments received regarding the
NPRM  dealt with heavy  duty  issues.   Many  of the  heavy-duty
issues  raised  could more  properly be classified as evaporative
emissions  issues,  since  they were  only marginally  related to
safety.    However,  they  will  be  addressed   in  this  section
insofar as they pertain to onboard safety.

     Heavy-duty   comments  were  received   from  four  general
categories  of  commenters, based largely  on where  they fit  into
the  production/distribution/usage  scheme  for  HDGVs.   Comments
from  the  automotive industry  reflect  their role as the primary
producers  of  truck and  bus  chassis  and  engines.  There  are
three such manufacturers, Ford, Chrysler  and  GM,  and  they are
also represented by the  Motor Vehicle Manufacturers Association
(MVMA).    Comments   were  also  received  from  two   secondary
manufacturing  interests:   truck and coach  builders who produce
commercial  truck bodies  on chassis purchased  from the primary
manufacturers,   represented  by  the  National   Truck   Equipment
Association  (NTEA),  and the recreational vehicle  (RV)  industry,
represented  by  the Recreational  Vehicle  Industry Association
(RVIA).   RV manufacturers produce  campers  and  motor homes using
chassis  purchased  from  the  primary  manufacturers.   Comments
from  these latter  two  groups  tend  to be  generally similar in
nature, but there  are  some differences  in their  concerns, as
reflected  in their comments.   Comments were also received  from
the  American  Trucking  Associations  (ATA)  and United  Parcel
Service (UPS)  representing commercial vehicle  operators.  These
comments  tend  to overlap the  first three categories.

     b.     Summary  and Analysis of the  Comments

      1)     Ability  to Use Lio^iid Seals

      Summary   of  the  Comments:   Several  commenters  expressed
doubts  about the feasibility of  liquid  seals  for HDGVs.   Ford
and  the  American  Trucking Associations  stated that  some  HDGV
fuel  tanks,  particularly the side mounted  or so-called "saddle"

-------
                              3-76


tanks, have short fillnecks  or  fillnecks  that are integral with
the tank  and so  may not  be able  to use  liquid seals.   They
stated  that  manufacturers   would  thus  be  forced to  rely  on
mechanical seals, which  could present problems with  durability
of the seal or  with inadvertent overpressurization of the tank,
through  failure  of  the  automatic  nozzle  shutoff   during  a
refueling  event.    Overpressurization could  thus result  in  a
fuel spill during refueling.   Ford stated that  a  submerged fill
might be  feasible,  but  unvented pressure buildup from diurnal
and other heating of  the fuel tank  could lead to expulsion of
fuel when the filler cap was removed.

     Response  to  the  Comments:   This  issue  was  originally
discussed  in  EPA's  safety  report,  but  will  be  summarized
briefly here.   The  reader is  referred to the  safety report in
Appendix II for additional detail.

     As  stated  in  the  safety  report,  most  HDGV's  do  not use
side-mounted  tanks.   Some  applications,  particularly  the more
numerous class  IIB  vehicles, which comprise over  75  percent of
HDGV  sales,  use tanks  mounted inside  the  frame   rails,  and
should  have   sufficient  fill   height   for  a   liquid   seal.
Moreover,  as  Ford stated, even vehicles  with  very  little fill
height may  be able  to use the  submerged  fill,  which  is a type
of  liquid  seal.  However, provision must be made for  adequate
tank  venting  when the submerged fill approach  is  used,  just as
a pressure  relief valve  is needed in the case of the mechanical
seal  to  prevent overpressurization  of  the tank.  It should be
noted that  the  issue of  tank pressures is not unique to  onboard
systems.   Current  HDGV  fuel  tanks  using   evaporative   control
systems  and  meeting  OMCS  requirements  must  now be  properly
vented,  and  adding  onboard   controls  does  not  affect  this
requirement or  the  manufacturers ability to do so.   In  fact, as
discussed  in  Chapter  4,   onboard  controls  may  offer   safety
benefits  with  regard   to   tank pressures  and  tank   venting.
Finally,  although  the   mechanical  seal,  with  proper   pressure
relief  provisions,   may  be  a  somewhat  less  desirable  design
choice  compared  to  the liquid  seal,  EPA  is  unaware  of any
inherent safety problem.

      2)     Increased  Size of Canisters and Other Components

      Summary  of  the Comments:   GM  and  MVMA  stated that  some
HDGVs use  fuel  tanks of up  to  100  gallons  capacity, and  would
require   25-30   liter   canisters   for   control  of   refueling
emissions.  GM  felt  that the risk of an underhood fire  would be
increased  because of use  of larger  canisters.   RVIA also  felt
that  the  increased  size  and complexity of onboard systems  would
increase  the  risk of fire,  which would be  further aggravated by
the  flammability   of   typical  materials   (i.e.,   wood   and
fiberglass)  used in RV  construction.  MVMA  stated that  the size
of  onboard  components  might preclude  the  use  of   integrated
systems  and  RVIA anticipated problems finding space to  locate
the larger  canisters.

-------
                              3-77
     Response to the Comments:  First  of  all,  the canister size
estimates given by GM and MVMA appear to be  somewhat  high,  even
for  extreme  cases.  Ford  estimated that  only 12-14  liters  of
canister capacity would be  required for the largest  fuel  tanks
(i.e.,  100 gallons).   Secondly,  very  few  if any  HDGVs  use 100
gallon  fuel tanks.   Information gathered  under an EPA contract
for  1985 model  year HDGVs  indicated  that  the  largest  single
tank available is 60 gallons.[14]   Some dual tank installations
may  total  100  gallons,   although  it  appears  that the  average
dual  tank  capacity is more  like  75 gallons. [15]  Moreover,  as
stated  in the  safety report,  it  is estimated that only about 15
percent of HDGVs in the over 20,000  Ibs GVWR weight  classes use
dual  tanks  providing  such  a large   capacity.   Average  tank
capacity is  likely to be more on the order  of  35-40  gallons
(single  tank)  which will require  canisters  eight to  ten liters
in size.

     Second,  with  respect  to canister  safety,  if manufacturers
are  genuinely  concerned  about the risk of  fires, the canister
should  be  located  in a  safe place,   away  from  any potential
sources  of  ignition.    Because   of  their  size  and  general
configuration, HDGVs have a great  deal  more flexibility than do
light  duty vehicles or trucks  as  to where  the canister can be
located.   Safe  location  of  the  canister  would  be  further
facilitated  by the use of  integrated evaporative and refueling
control  systems, particularly since the trend toward increased
use  of  fuel  injection in  HDGVs  virtually  eliminates any need
for  a separate hot  soak  canister.  Furthermore,  studies  cited
by  API  in  their   comments  as  well   as  EPA  in-house  testing
(discussed in  the  next section)  have shown  that  the  likelihood
of underhood  fires resulting from vapor or  canister  ignition is
extremely  small.   It is  also noteworthy  that   RVs routinely
incorporate    LP   gas   heaters   and  cooking   units    and/or
gasoline-powered   generators,  all  of   which   pose  similar
potential  safety  risks.   Yet RV  manufacturers have  apparently
been able to  safely incorporate  these devices  into  their  RV
designs.   EPA  concurs   that  manufacturers   must  carefully
consider canister  location  as well as other factors  to minimize
safety risks,  but  does not find this to be  a major obstacle for
HDGVs.

      Finally,   it   is  worth  noting   that   HDGVs  provide  an
opportunity  for development  of  vapor  limiting devices such as
bladder tanks.  Some  manufacturers,  e.g.,  Ford, have already
been investigating this alternative.

      3)   Purge Problems

      Summary  of  the Comments:   GM,  Ford, MVMA,  ATA  and the
United  Parcel  Service   stated  that  HDGVs   would   experience
difficulties   purging  the  collected  refueling  vapors  because
HDGVs typically spend considerable time at  low-vacuum wide open
throttle (WOT) operating  conditions.   Ford stated that current

-------
                              3-78


evaporative systems  are  pushing purge systems to  the  limit  due
to this factor.  Consequently, HDGVs would spend more  time with
canisters at  or  near saturation  than  would passenger cars  and
light trucks.  ATA felt  that  supplemental purge air blowers  or
heaters would  be required for proper purge  operation, and that
this could  increase  the  risk of  vehicle  fires (for example  if
sparks  came into  contact  with gasoline  vapor  as a  result  of
component failure).

     Response to  the Comments:  Manufacturers  have considerable
flexibility  under the  current EPA test  procedure to  balance
purge  rates/schedules  against  canister  sizes to  insure  that
canisters  are properly  purged.   The  same  net effect  can  be
achieved with  a  less  stringent  purge rate/schedule used with a
larger  capacity  canister  or  with  a  more  stringent  purge
rate/schedule  and   a   smaller   canister.    Thus,   there   is
sufficient  latitude  to  accommodate varying  usage  conditions.
Moreover, given  the  relative  size and frequency of  evaporative
and  refueling loads,  it becomes  evident that the  evaporative
emission  load is the  governing   factor   in  determining  purge
requirements.  The  refueling load may  be added to  the  typical
daily evaporative emission  total,  but  refueling loads are much
more  infrequent,  and  so they only  occasionally constitute the
bulk  of  the  canister  loading.    EPA  believes  that the  small
incremental effect can be managed with due  consideration to the
tradeoffs  described above  without  a  substantial  increase  in
purge rate/schedule.

     Furthermore, with respect to  the  WOT  issue,  no  data were
presented  to  support  claims  of extended  low-vacuum  or  WOT
operation.   The   CAPE-21  data base, which  forms  the  basis for
both   the  EPA   engine   dynamometer  cycle   and   the  chassis
dynamometer   cycle   used  for  evaporative  emissions  testing,
indicates  that only slightly more than  ten percent  of  typical
HDGV   operation   occurs  at  greater   than  90  percent  power
conditions.   It   should  also  be remembered  that most  HDGVs are
designed   for  intra-city  operation,   rather  than  long-haul
operation,  and   do  not  normally  carry  maximum  loads  over
extended  distances.   If the  commenters'  concerns  are valid, it
raises  a   question  of  the  in-use effectiveness  of  current
evaporative  systems   and  whether  EPA  needs   to  increase  the
amount   of   operation   at   WOT   conditions  in   the  current
certification test  procedure to  ensure  that in-use canisters
are  in fact  being  properly  purged.   However, while  there  are
certainly   exceptions,   no   evidence   has  been  presented  to
indicate  that extended WOT operation is  either  typical  or even
fairly  common.

     With  regard to  purge  hardware,  it should  be  noted that
current HDGVs use vacuum actuated/controlled purge valves which
provide minimal   sophistication in terms  of varying purge rates
and  purge schedules.  As  fuel  injection  and  electronic  engine
controls  are  phased  into  the HDGV market over  the next  few

-------
                              3-79
years,  the  use  of  electronically  controlled  purge  valves,
similar  to  those  used  in  current  LDVs  and  LDTs, offers  the
potential for  greater  flexibility  in developing  refueling and
evaporative  canister purge  schedules.   With regard  to  purge
assist hardware,  ATA presented no data in  its  comments  to show
a  need for  purge blowers  or  heaters.   Furthermore,  as  shown
above, the  use  of such devices  will be driven  by evaporative
emissions  control,  rather  than  onboard  control  requirements.
If these devices  were to be used, prudent design considerations
would  dictate  that they be constructed in such a  way  that they
would be safely incorporated into the vehicle design.

     4)    Secondary Manufacturer Concerns

     Summary  of  the Comments:   Ford, GM,  Chrysler,  NTEA and
RVIA  commented on  the  problems  arising  from  the fact that  a
significant  number of HDGVs are  not totally manufactured  by  a
single entity, but  rather have  the coachwork  installed   by  a
secondary manufacturer  on a chassis  purchased  from the primary
manufacturer.    Problems  could  arise   when  these   secondary
manufacturers  modify  fuel  tanks,   fillpipes,  and other  fuel
system  components   to   suit   their    individual    vehicle
configurations.    Most   commenters   stated  that  this  would
increase the risk of fires, fuel leaks  or malfunctions in the
onboard  control  system.  NTEA stated that  bodies might have to
be mounted  higher on the chassis to  avoid  trapping fuel in the
fillpipes,   thus   raising  the  vehicle center  of gravity and
increasing  the risk  of  rollover or  other  handling  problems.
RVIA  stated that  most  RV  manufacturers  were  small  businesses
with  limited resources  and most  lacked the necessary expertise
to safely  incorporate onboard  systems into their  designs.   They
feared that the  risk  of fires could then  be increased due to
the  increased complexity  of  the systems  and  the inability of
most   secondary   manufacturers  to   deal  with  the  increased
complexity.    Secondary  manufacturers  also  expressed  concern
over   determination  of  the   legal   and  recall   liability for
systems  that had  been  designed by  a primary manufacturer, but
which  may have been modified by a secondary manufacturer.

     Response  to   the   Comments:   While EPA  recognizes  that
difficulties can occur  anytime that more than one manufacturer
is involved  in the manufacture of a  given vehicle, this problem
is  hardly  a new  one.   It has existed  in  the past  and   still
exists now  in  conjunction  with the  modification and addition of
fuel  tanks,  fillpipes   and  other  fuel  system   components,  as
well   as  with  the  more  recent  incorporation  of  evaporative
emissions  control  systems  in  HDGV  fuel  systems.  Furthermore,
secondary manufacturers  have been required  to comply with  FMVSS
301  (for busses  and vehicles less  than 10,000   Ibs  GVWR) and
applicable    OMCS   fuel   system    safety   requirements   for
modifications  to the primary  manufacturers'  fuel systems  for a
number of  years.   Yet these standards  have  apparently been met
with   a  minimum  of   difficulty.    Onboard  control  systems

-------
                              3-80


present  few  new problems  with fuel  system design  relative  to
current evaporative/fuel systems,  and most problems  are  merely
incremental  to  those  experienced   with  current  evaporative
systems.  In fact,  as discussed in Section 3C,  incorporation  of
an  onboard  requirement   could   result   in  simplification  of
current systems.  Straightforward solutions can  be developed  to
address any problems  involved.  Onboard  controls in fact afford
an  opportunity  to  alleviate  some  of  the fuel  system  problems
currently  being experienced,  e.g.,  overpressurization of  fuel
tanks.   Vapor  lines  could also be  shorter  and  external  vent
lines could be  eliminated,  thereby doing away with some current
problem areas.

     EPA  also  sees no technical  reason  why truck bodies would
have  to be mounted higher  on the  chassis because  of  onboard
control requirements, and  no supporting  rationale was  provided
by  the commenter.    If,  for  example,  a  manufacturer wished  to
gain  additional fill height for  a  liquid seal,  it would  be
unreasonable  to raise the body when the same  result  could  be
obtained by simply extending  and/or  raising the  fillpipe on the
existing body.

     Although  the  general  certification  responsibility rests
with the  fuel  system supplier, i.e.,  the chassis manufacturer,
incorporation  of onboard  controls  into   a  modified fuel system
may be within  the abilities  of  many secondary manufacturers,
particularly   given  the  relatively  simple   nature  of  such
systems.   Ford's  suggestion  of  providing  fillneck  kits  to
secondary  manufacturers  may  represent   one  way  of  assisting
these  secondary manufacturers.   In  two  separate meetings  on
this  issue with NTEA,  EPA  has  asked for technical  input and
suggestions on how the proposed  onboard  requirement  for HDGVs
could  be  structured  to   alleviate  their  concerns  but  little
input   has  been  received   thus   far.    EPA   remains  open  to
exploring  ways  in  which the  onboard requirement  could  be met
while  minimizing the concerns  to the secondary  manufacturers.
However,   the  Agency  does   not   see  onboard  technology  as
introducing any new fuel system concerns  to these  manufacturers.

     The   question   of    legal   liability   for   fuel   system
modifications  done  to a  primary manufacturer's  fuel system by  a
secondary  manufacturer  is also not  a new problem.   This  issue
has arisen  with  current   evaporative   regulations.   However,
since  this problem does  not make a  vehicle  more  or  less  safe
but only addresses legal responsibility,  it is  a  legal problem,
rather  than a  technical  question,  and  is  not  directly  related
to  the  current  analysis.

-------
                              3-81
     2.     Canister Safety

     a.     Introduction

     As  was  mentioned  in  Section  A  of  this  chapter,   many
commenters   expressed   concern   about   the   fire  safety   of
canisters.    These  concerns  fell   in  two   general   areas:
flammability  of  refueling vapors and  canisters.   Each  of  these
is summarized and analyzed below.

     b.     Summary and Analysis of Comments

     1)    Refueling Vapor Flammability

     Summary   of  the  Comments:    Several    commenters   were
concerned  about  the  flammability of  vapors  released  from the
canister  or  vapor  line  due  to  breakthrough  or  tampering.
General  Motors  provided written  comments and  a videotape which
suggested that underhopd onboard canisters  would  create  a fire
hazard under  some conditions.

     Response to the Comments:

     i)    Location Concerns

     From  the outset  it must be clear that  if  a manufacturer
genuinely believes that there is  a  risk of underhood  fire due
to system tampering,  malmaintenance,  defects  or other problems,
then the  canister  and  any other  component in  the onboard system
thought  to  be  of  concern should be  located  elsewhere.   This
applies  not  only  to  onboard  systems,  but   also  to  current
evaporative   systems.   As   will  be   discussed   below,   some
information  presented by General Motors  suggests  possible fire
risk  with  current   evaporative systems   under  some  unique
circumstances,  but the in-use  safety  record  of  these systems
cited  in Chapter  2  reflects  no history of  fire  problems.  In
any  case, with  the  strong move  towards  fuel  injected vehicles
there  is less reason to locate  the canister under the hood than
there  was with  carbureted vehicles,  on which the  float bowl
must  be vented  to  the  canister.  There  may  even be additional
benefits  to  a  rear-mounted  canister,  since   it  allows  for  a
shorter  vapor line  from the  fuel tank  to the canister.   This
represents  a cost saving.   Thus, even  absent  safety  concerns,
EPA  would  expect  that  many integrated refueling/evaporative
canisters would  be rear-mounted by preference.   However,  it is
worth  noting that   information  regarding  vapor flammability
submitted by API  suggests  that   flammable vapor  mixtures  exist
only  at  a  few  points,  even on a  tampered  or  malfunctioning
onboard   system.[10]    With  the  exception  of a very   narrow
transition  zone, fuel/air mixtures  are either too  rich or too
lean  to  be  flammable.   Furthermore,  API's   work  shows  that
surface   temperatures  of  the  exhaust  system  and  the   engine
compartment   do  not  get  high  enough  under   normal   operating

-------
                              3-82
conditions   to  ignite   even   flammable   fuel/air   mixtures.
Evaporative  controls may  even  represent  a safety  improvement
over pre-evaporative control  vehicles  which vented  directly  to
the atmosphere at the carburetor and fuel tank vents.

     ii)   General Motors'  Vapor Flammability Tests

     Several  points need  to  be made  about  the  videotape  and
written  comments submitted  by General  Motors  with  respect  to
underhood  fires  on  vehicles  with tampered  onboard  systems.
After  closely reviewing the  tape,  EPA had discussions  with  GM
and exchanged several pieces  of correspondence concerning both
the techniques used by GM's contractor in this testing  and  the
significance   of  the  testing.    (The  interested  reader   is
referred to  these  letters,  found in Public Docket A-87-11,  for
more detail:  IV-E-30, IV-C-93,  IV-D-524, IV-C-107, IV-C-117.)

     In  their testing, GM's  contractor cut  the refueling vapor
line of  a  vehicle  equipped with a  prototype onboard system  and
then deliberately  placed the  cut  end  of  the  vapor line near a
damaged spark plug  wire.  In the course of the videotape it was
explained  that  the  refueling vapor,  vapor  line,  and  some
adjacent wires caught fire when the vehicle  was started after a
refueling  event.

     Despite  discussions and  correspondence with GM regarding
this test, GM has  not yet  satisfactorily explained the source
of  the  vapors  which  caught  fire.   GM's  videotape explained
that,  following  the end of the fuel dispensing:  1) the fuel cap
was  replaced, 2) the nozzle  was returned  to the  dispenser,  and
3)  the driver walked  to the  vehicle,  got  in and  started  the
engine.  Considering the time  it would take to accomplish these
three  steps,  it  is  difficult to understand how refueling vapors
could  still  be  flowing  out of  the cut vapor line  and  continue
to  flow   after  the  fire  started.   Furthermore,  since  the
prototype   onboard  system   used   a  nozzle-actuated   positive
shutoff valve to the refueling vapor line, vapor  flow should be
cut  off when the  nozzle was removed.   GM's initial  explanation
that  the  vapors were  those  remaining  in the  line  and being
displaced  by inertia did not seem plausible.

     Upon  inquiry  by  EPA,  GM  suggested that  the vapor  exiting
the  line  might  be  driven  by the  pressure  differential across
the  fuel   tank  limiting   orifice,  which  remained  after  the
refueling  event  ended.   (The  fillpipe   was  not  open  to  the
atmosphere after  the refueling event since  a  J-tube seal was
used.)  While  a   possible  source of vapors,   this   pressure
differential would have quickly been  diminished  at  the  rate the
vapors were exiting the end of the  line.

     Nevertheless,  if this suggested  pressure differential were
actually  the cause  of vapors  flowing out  of  the vapor line

-------
                              3-83
after dispensing  is stopped, then  a  similar  problem might also
exist with present  evaporative  systems.  Subsequent  testing on
the stock vehicle model  without onboard controls confirmed this
to be true.   Using the same  tampering  mode, GM's contractor was
able to  start and sustain a small  evaporative  vapor fire after
several minutes  of  idling following  a refueling event.   While
this result  is not  surprising,  it is important to reconize that
the entire  tampering  sequence  is  highly  unlikely,  and  in-use
experience with  evaporative  control systems  indicates  problems
of this nature do not apparently occur,

     iii)  Conclusion

     The  videotape  presented  by  GM demonstrates  a  potential
problem  which  would  have  to   be  evaluated  on  both  onboard
systems  and  present  evaporative  emission   control  systems.
Incrementally, there   appears  to  be little  difference  in  the
potential  for  occurrence  of  this  problem  between  the  two
systems, since the  GM testing showed that both  were capable of
causing a fire  under  the  given  conditions.   Therefore,   as  is
the  case  with  current   evaporative  canisters,  manufacturers
would have  to carefully evaluate potential canister  locations
and  choose   a  location  deemed  acceptable  by  vehicle  safety
considerations.

     2)    Canisters

     Summary  of  the   Comments:    In   addition   to   comments
regarding  refueling   vapor   flammability,   several  commenters
expressed concern about  the flammability  of  activated  carbon,
and  some  were  even   concerned  about  the  "explositivity"  of
canisters.   Toyota was concerned that carbon scattered out of a
canister  broken  in  a crash would  be  a fire  hazard.    On  the
other hand,  both  Ford and Nissan felt that the canister was  not
a safety problem.

     Response to the  Comments:   To  begin with,  it should  be
noted that any concerns over the  safety of carbon canisters  are
not unique  to  onboard  controls.   Carbon  canisters have  been
installed on  vehicles  since the  early 1970's.   To the  extent
that manufacturers  have  concerns  over  the  safety of  onboard
canisters,  these same  concerns  apply  to  today's   evaporative
systems.  However,  while commenters  have  raised concerns  over
onboard  canisters,   EPA   is   unaware  of   any  specific   safety
concerns with  respect to  evaporative canisters.   In  addition,
EPA has  not  been  provided  with  any evidence  to  suggest  that
evaporative  canisters  have  posed any safety hazards  to  vehicle
owners  since  they were implemented 18 years ago.  Since  onboard
controls will use  the same  activated carbon control technology
as  evaporative canisters,  there  is no  reason   to  expect  that
onboard  canisters  will   affect  safety any  differently  than
current evaporative  canisters.

-------
                              3-84
     In   addition,   to   confirm  that   no   carbon   canister
(evaporative or onboard) poses  a serious  safety hazard, EPA ran
a series of small  scale  in-house tests to address  the  concerns
raised  about  carbon  flammability and canister  explositivity.
The test  method  used was  to  saturate an evaporative  emissions
canister with  gasoline vapor, open the canister  and spill some
of the  activated carbon  from  the canister into a pile,  and try
to  ignite  the  activated  carbon with  both  spark  and  flame
sources.  Carbon that  had  been soaked in  ligiaid  gasoline was
also tested as a worst  case  situation.   Ignition  sources used
included sparks thrown from a flint  starter and  grinding wheel
on steel, and flames from a match and burning paper.

     The testing showed  that  neither  gasoline-soaked carbon nor
vapor-saturated carbon would  ignite  with  either spark source.
Carbon  in both conditions would light only after  a  flame was
held directly  to  its  surface and enough heat  was  provided to
release vapor  adsorbed onto the activated carbon.  There was no
explosive effect  to the carbon fire; it simply burned  with a
low  steady  flame  which  slowly  spread over the  surface of the
carbon pile and extinguished  as the fuel was consumed.

     The  results   of  this  testing were  precisely  as  EPA had
anticipated.   There   is  no   reason  to   expect  that   carbon
saturated with gasoline  (hydrocarbon)  vapor  would pose a  unique
or  significant safety hazard.   Hydrocarbon vapor  will  remain
adsorbed  on a bed  of  activated carbon until  sufficient  energy
is supplied to release the hydrocarbons.   Even if  a heat  source
is  present   to   supply   the   necessary  release  energy,  the
hydrocarbons   are   desorbed   slowly   enough   to  prevent  the
possibility of  rapid  combustion.  Based upon  this  information,
EPA has concluded  that the use  of onboard canisters on vehicles
does  not   introduce  any   new   or  .added  risk,  and  that  the
flammability  of  carbon  in any vapor   control canister  is not  a
fire hazard on a vehicle.

     3.    The  Effect  of  Onboard  Refueling Control  Systems on
           Future Recalls  and Technical Service Bulletins

     a.     Introduction

     As was  discussed  in Chapter   2 for  evaporative  control
systems,  one  way   to  assess  the effect  of onboard  refueling
controls  on vehicle safety is to evaluate  the extent to which
these  systems might affect future safety  recalls and  technical
service  bulletins.    NHTSA  was  the  first   to  suggest  this
approach,  proposing  that   since onboard  controls  do  not yet
exist   in-use,  EPA should  examine  past  recalls  and  service
bulletins for problems  involving  components similar  in  nature
to  those which  might be  used on onboard  systems, to  identify
the  types of  problems  that  might accompany the implementation
of onboard  systems.

-------
                              3-85
     Also with regard to  future  recalls and  service  bulletins,
NHTSA raised  the general concern about the  effect  of increased
complexity.   NHTSA  believes that  if  onboard controls  require
the  use  of  more/larger  components  or  the  modification  of
others,   recalls  and service  bulletins may  increase  in  number
due  to  a  greater   opportunity  for  design/production  errors,
in-use  problems  (e.g.,  disconnections),  and  interference  with
other systems.   While  suggesting  that this  analysis would  be
useful,  NHTSA also stated that an analysis of onboard controls'
effect   on  future   recalls   can  only  be  qualitative  (i.e.,
identify  types  of  problems),  not  quantitative,  since  onboard
systems  do  not   exist  in  use  today  and  that  any   effort  to
predict  the  number  or  severity  of  future  onboard  related
recalls  is likely to be speculative.[16]

     Both NHTSA  and the manufacturers have provided information
on  recalls and  service bulletins  involving types of problems
they  believe  to  be relevant  to onboard control  systems.   The
purpose   of   this   section   is  to   describe  the  information
received,  and discuss  the  implications of  this information for
onboard  control system safety.

     b.    Summary  and Analysis of Comments

     i.    Recalls

     Summary  of   the Comments:   After  considerable   discussion
and  correspondence  between  NHTSA  and  EPA  on  the issue  of
recalls  relevant to onboard  controls,  NHTSA  arrived  at  a list
of  38 recalls which they consider to be relevant.[17]  These 38
recalls,  shown  in   Table  3-3,  fall  into   the  following  four
categories:     1)   Fillpipe-Related    (10),   2)    Vent/Vapor
Line-Related  (9), 3) Pressurized Fuel  System/Volatility-Related
(6),  and  4)  Evaporative  Control System-Related  (13).   (These
four  categories  represent  EPA's  characterization  of  the  38
recalls  since NHTSA did not provide such classifications).

      In addition,  NHTSA  recently provided  information  on over
350  recalls which  they  believe  might  be characteristic  of the
types of problems which future onboard systems could  contribute
to  indirectly  in  the  future.[16]   These  include  31  recalls
concerning   stalling/driveability,   20   related   to  exhaust
emissions/temperatures,   and    314    involving   fire.    NHTSA
suggested a  review  of  these cases might help identify ways in
which  onboard controls might  adversely affect  the  performance
of  other vehicle  systems.

      EPA also received  information  from  several  manufacturers
on  this subject.  EPA and NHTSA  asked manufacturers   to  include
as  part of  their   comments  on the  NPRM   information  on any
recalls  and  service bulletins  which  they  believed would  be
relevant  to  onboard  control   systems.    Most   manufacturers
responded by providing  information  on all  problems   related to

-------
                                           3-86
                                        Table 3-3
                     NHTSA's 38 Recalls Relevant to Onboard Controls
  NHTSA
Campaign
 Number
Model
Year(s)
                                     Fillpipe-Related
 Vehicle
  Type
Number of
Affected
Vehicles
Description of Problem
76V069
1975-76  Pass. Veh.  67,633
78V249
79V139
80V019
81V047
82V021
82V088
82V109
83V085
1978
1979
1978-80
1982
1982
1983
1983
1982
Lt. Truck
Motor Home
Pas. Veh.
Pass. Veh.
Pass. Veh.
Pass. Veh.
Pass. Veh.
Lt. Truck
H. Truck
5,317
46
9,429
5,200
519,329
2,800
1,849
215
                     Defective   rivets  used  in  fabrication  of  fuel
                     filler  inlet assembly.

                     Fuel  filler pipe can disconnect in 301 collision.

                     Fuel   filler  hose  can  come  in   contact  with
                     exhaust pipe.

                     Cut fuel filler tube in 301  crash.

                     Poor  design of fuel filler  & vent  pipe  leads  to
                     301 failure.

                     Fillpipe hose clamp  can fracture.

                     Cut rubber  filler  pipe  connecting  hose  in  301
                     test.

                     Fuel  filler hose breaks during 301 crash.

                     Hose  type 3 piece  pipe fails in 301
                     test.
86V101
1986
Pass. Veh.  27,000   Pierced fuel filler pipe in rear collision.

-------
                                           3-87
                                    Table 3-3 (cont.)
                                 Return/Vent Line-Related
NHTSA
Campaign
Number
77V063
78V016

Model
Year(s)
1977
1973-74

Vehicle
Type
Motor Home
Pass. Veh.
Number of
Affected
Vehicles
242
20,661




Description of Problem
External vapor line
not connected to fill spout.
Cracking of fuel tank vent hose allowing fume
78V195


78V208


81V004


83V115

85V014




85V132
1979     Pass. Veh.
   13
1977-78  Pass. Veh.   16,238


1981     Pass.Veh.   14,000


1984     Lt. Truck    1,548

1984     Motor Home     750



1985     Pass. Veh.   11,000
to enter trunk area.

Fuel  line  filter  and  vapor  return  line  may
deteriorate.

Insufficient  clearance  between  floor  pan  and
vent tube can damage vent tube.

Improper design/installation of  fuel system such
that it restricts fuel supply.

Fuel or vapor line damaged due to assembly error.

Auxiliary  fuel  tank  is  subject to  overfill and
pressure  build  up  due  to  improper  vent  tube
placement.

Bumper   causes   damage  to  air   vent  hose  in
accident.
85V154
1982-85  Bus
1,520   Fuel pipe vapor line disconnects in accident.

-------
                                           3-88
                                    Table 3-3 (cont.)
                             Volatility/Tank  Pressure-Related
  NHTSA
Campaign
 Number

82V076

85V106
87V052

87V113


87V144

87V155
                    Number of
Model     Vehicle   Affected
Yearjs)    Type     Vehicles
1982
                                 Description of Problem
Pass.Veh.   24,455   Pressure buildup in tank causes fuel spitback.
1979-85  Motor Home  28,545   Filler cap  disengages suddenly when  removed for
                              refueling.   May  cause  expulsion  of  vapor  and
                              gasoline.
1986-87  Van

1983-87  Truck
         (Ambulance)
            15,500   RVP problem - expulsion of gasoline.

            16,000   RVP problem - expulsion of fuel.
1983-87  Vans       188,000   Expulsion of fuel due to use of high RVP fuel.

1985-87  Motor Home   9/041   Expulsion of fuel due to use of high RVP fuel.

-------
                                           3-89
                                    Table 3-3 (cont.)
                                Evaporative  System-Related
  NHTSA
Campaign
  Number
Model
Year(s)
           Number of
 Vehicle   Affected
  Type     Vehicles
          Description of Problem
76V126


78V036



78V106

78V145



78V181
 87V157
1976
Pass. Veh.   9,137
79V019
79V032
79V045
79V048
79V212
84V116
87V111
1976-78
1977-79
1979
1975-76
1973-78
1985
1984-87
1977-78  Lt. Truck   20,000



1977-78  Pass. Veh.  10,500

1973-77  Med. Truck   2,500



1978-79  Lt. Truck   23,000


         Pass. Veh.  17,800

         Pass. Veh.  83,000


         Pass. Veh.   3,700

         Pass. Veh.  61,000

         School Bus   2,950

         Pass. Veh.   2,385

         Van            250
         (Ambulance)



1984-88  Pass. Veh.  25,000
Erroneously installed  piping for  the  "check and
cut valve".

Blockage  of  tank  vent   system  can   lead  to
pressure buildup and  force fuel or vapor leakage
through cracks in tank.

Defective fuel tank vent valve.

Liquid  gasoline may   discharge from  bottom  of
canister  because  evap  system  may  lack  adequate
capacity under certain fuel expansion conditions.

Obstructed evap line causes  pressure build up in
tank.

Possibility of kinked evaporative system hose.

Obstructed evap  line  causes pressure build  up in
tank.

Misrouted vapor line to canister.

Defective'pressure control valve.

Defective  liquid/vapor separators.

Improper  functioning of vacuum  line valve.

Defective  vapor valve grommet on fuel
tank.   (This  component was  installed  as part of
a  preliminary  attempt  at  correcting   the   fuel
expulsion  problem of Recall No.  87V113).

Overfilling  of the  fuel   tank  can increase  fuel
system  pressure to the  point  where  fuel vapors
escape  from  the  charcoal  filter and  cause an
engine  compartment fire.

-------
                              3-90
fuel/vapor  lines  and connections,  clamps,  valves,  FMVSS  301
failures, addition  of protective shields, etc.   Many of  these
recalls  were the  same  as  those  provided by  NHTSA and  those
reviewed by  EPA  as  part  of  the characterization  of  evaporative
system safety presented in Chapter 2.

     Response  to  the Comments:   The  first group  of  recalls
addressed  in this  section  are the  38 identified  by NHTSA  as
being directly relevant  to  onboard controls.   Before describing
the  relationship of these  recalls  to onboard  controls,   it  is
important  to  have  a  perspective  on  these  38  recalls  are
relative  to  other  types of  recalls.   Since  1966,  there  have
been  over 4,200  safety  recalls  of  over 130 million vehicles.
The  38  recalls  identified  by  NHTSA represent  a  very  small
fraction  (less than one  percent)  of the total number of recalls
issued to remedy other types  of problems.  Further,  EPA did not
receive  any  information  to suggest  that  any of  the 38 recalls
involved problems that caused any deaths or serious injuries.

      It  is  also important  to  understand  that  assessing  the
safety  implications of onboard  controls  through  an examination
of past recalls of  similar components must be done with full
consideration  of two  very  important  concepts.   First,  onboard
controls'  effect on  future recalls  is highly  dependent  on the
designs  selected  by manufacturers.   Manufacturers  can  choose
unnecessarily complex approaches such  as  some  of  those shown in
Appendix  I,  or  they  can select simple approaches  such as that
demonstrated by  EPA  and  discussed in Section  3C.   This design
dependence   leads  to  the   second  key   consideration  -  the
incremental  nature  of  the  analysis.  Any  effect   of  onboard
controls  on  future  recalls must  be viewed  incrementally to the
current   recall   situation.    The  incremental  nature  of  the
analysis  applies to  different  subsystems and  components of the
fuel/evaporative system  as  well.    If  recalls  for  a  given
subsystem/component  happened   in  the  past,  it   does  not
necessarily  imply  that   adding  onboard  controls would   impact
future   recalls.    In  fact,   even  without  onboard  controls,
recalls  of  this nature  might  continue  in  the   future.   With
these  two   considerations  in  mind,  we  are   now  prepared  to
examine  the  relationship  between  onboard  refueling controls,
and   the 38 recalls identified  by  NHTSA  as  being  directly
relevant.   These 38  recalls  will be  examined  according to the
four groups  shown in Table  3-3.

      The first  set of  ten recalls shown in  Table 3-3 involve
problems  related   to  the  fillpipe.   Onboard  control  systems
should  not have any  effect on  fillpipe failure problems.  These
problems   happen without  onboard  systems,  and there   is  no
technical reason why  onboard systems will  affect the  frequency
or   severity of  FMVSS   301   fillpipe  failures,   or  fillpipe
installation,  fabrication,   and placement  problems.  Virtually
all  of  these types  of  problems  are not  relevant  to onboard
hardware.   In fact,  the system design presented in Section 3C

-------
                              3-91


suggests that  the  fillneck and  related  components can be  even
simpler  than  at  present.   If  one  accepts   the  complexity
argument, this could represent a potential safety improvement.

     Furthermore, while  some comments  have suggested that  the
information  in Table 3-3 provides  evidence that  the  potential
use  of  fillneck  mounted   valves   in  onboard  systems  would
increase  recalls,   there  is  nothing  inherent  about  onboard
controls which drives the  use of fillneck mounted valves.   Some
proposed onboard designs (including ones depicted  in  the  NPRM)
have  shown  valves mounted  at  the  fillneck.  However,  after  a
closer  examination,   EPA  has  not  identified  any  compelling
reasons  for  relocating  the  vent valve  currently  used  on  most
fuel tanks from its present position to  the fillneck.   In  fact,
it  is  anticipated  that many  manufacturers  will  elect  to simply
modify  the  current   tank-mounted  vent  valve  to  perform  the
necessary refueling  control  functions.  Even  if a manufacturer
chooses  to   use a  fillneck   mounted  valve,  this  will  not  be
unigue  to onboard  systems,   nor does  it  have  to represent  a
safety  hazard.  As  discussed in Section 3B, there are  several
current  fuel/evaporative system designs which  mount  valves and
other  hardware in  the  fillpipe area,  and presumably  without
compromising safety.

     NHTSA   also  included nine  return/vent line problems  as
being  relevant to  onboard control  systems.  Six  of  these nine
recalls  involved the  external  vent  line which  runs  along the
fillpipe.  Onboard systems are  not  expected to  have any adverse
effect   on   these  problems   since  onboard  controls  will  not
inherently increase the  number  of vent line connections (77V063
and  85V154),   or   change  material  selection  (78V016),   or
placement (85V014,  78V208, and 85V132).   In  fact,  some onboard
designs  are  likely  to eliminate this external vent line thereby
reducing these types of problems in the  future.   Further,  the
other  three  recalls in this  category  are  not  at all  related to
onboard  control systems,  since  two  involve fuel  return  lines
(78V195  and 83V115)  and the  other involves a  restricted fuel
supply  system  (81V004).   Onboard controls will not  affect the
design  of the  fuel supply or  return system.

     The third group  of recalls involves  six  problems related
to  pressurized fuel  systems  and/or  high volatility fuels.   Five
of  these six  recalls involved fuel expulsion  problems due to
pressurized  fuel systems operating with high RVP fuels.   Some
commenters   believe   these  recalls  resulted  because  of  the
evaporative  system.   However, as was discussed  in Chapter  2,  a
multitude  of  factors  are responsible  for  these  problems,  and
evaporative  systems  cannot  be  identified  as  the sole cause.
For example, most  of the vehicles  in Recall No.  85V106  and some
in   87V113  were  not  even  equipped with   evaporative  control
systems.     (A  Federal    evaporative   emission   standard  for
heavy-duty  gasoline vehicles was not in place  prior  to 1985.)

-------
                              3-92
Therefore,  these problems  were really caused  by other  factors
such as high tank temperatures and/or high volatility fuel.

     High fuel  tank  pressures are  the  result  of a  combination
of three factors,   First,  fuel tanks  normally operate  under  a
slight pressure  to  reduce vapor generation and  fuel  leakage as
required by FMVSS  301.    Second,  tank  operating  temperatures
have  increased  recently as  a result of  the trend toward  fuel
injection and fuel recirculation.   Third,  the  recent  increasing
trend  in  fuel  volatility  has  combined with  the  first  two
factors to  raise fuel tank  operating  pressures to higher  than
normal   levels.    The   fuel/evaporative   control   system   is
sometimes modified to remedy high fuel tank pressure  problems.
This  is  because the evaporative  system  is the  only one of the
three contributing factors which can be  practically modified by
manufacturers.   Therefore,  it  is  not  appropriate to  fault the
evaporative  control  system as the  cause of high tank pressure
problems.

      It should  also  be  noted that revisions to the onboard test
procedure  (discussed at  the  June  30,  1988  workshop  held at
EPA's Motor  Vehicle  Emission Laboratory  in Ann Arbor) will help
to  alleviate pressurized  fuel system problems  in the  future.
The   revised  test  procedure  will   encourage  manufacturers  to
increase tank venting thereby insuring pressure is  not  allowed
to  build up  substantially  in  the   fuel  tank.   High  fuel  tank
pressures  have  been  identified  as a  contributor  to  safety
problems such as fuel spurting,  fuel  leaks,  and increased fuel
dispersion  in the  event  of  a ruptured tank  in  an  accident.
Consequently,   by   encouraging   the  lowering  of   fuel  tank
operating  pressures, the  revised test  procedure will  enhance
safety.  As an  aside,  it  is worth noting that in  a  separate
action  EPA  has proposed  to  reduce the  volatility  of in-use
fuel.[18]   Any  level of volatility  control would help eliminate
problems of  this nature for  present  and future vehicles.

      The final  group of recalls on  NHTSA's list are  13  recalls
concerning  the  evaporative  emission control  system.   These 13
recalls  are the  most   directly  relevant  to  onboard  vapor
recovery because they involve the type of components that would
also  be used on onboard  systems such as vapor lines  and vent
valves.  Once  again though, the incremental  nature  of  onboard
relative   to  evaporative  systems  must   be   considered  when
analyzing the effects of these types of problems.

      As  was  shown  in   Section  3C,   an  onboard   refueling/
evaporative control  system  can use essentially the same valving
and  vapor  line  routing as  is currently  used in the evaporative
system.   In  fact,  onboard  designs  incorporating  rear  mounted
canisters  are likely to  reduce current  vent line lengths,  with
only  a  small  increase in  vapor line  diameter.  Consequently,
the   incremental impact   of  onboard  controls  on  vapor   line

-------
                              3-93
problems  such   as  blockage   (78V181  and   79V032),   kinking
(79V019), or misrouting  (79V045),  should  be negligible  and  may
in  fact  be positive.   Similarly,  since  onboard controls  will
have   a   minimal   effect   on   vent  valves   (the   principal
modification is a  larger orifice), there  should be virtually no
effect on recalls of this nature (78V106  and  79V048).   Pressure
control valves  such as that  listed  in recall  79V048  could not
be  used  in  an  integrated  onboard  system.   Furthermore,  with
regard  to  vacuum  valve   recalls  (84V116),   these  types  of
components are used on present vehicles and will  continue to be
used  on  future  vehicles (with or  without  onboard controls), so
no  incremental  effect  should occur.   Finally,  with respect to
canister  overloading  problems  (78V145  and  87V157),  onboard
systems  together  with  the  test  procedure   revisions  will
increase  canister  capacity  under normal operating  conditions
and  thus  reduce  these  problems  in  the  future.   Thus,  no
increase  in recalls of  this type would  be expected,  and some
decreases are possible.

     As  was stated  in the  beginning  of  this discussion,  the
potential  effect  of  onboard  controls  on future recalls  is
design-dependent.     The    same    is    true    for    current
fuel/evaporative   systems.    It   is   expected  that  for  cost,
engineering, packaging and other reasons  in addition to safety,
manufacturers   will  ultimately  implement  integrated  onboard
systems  using  liquid seals  and tank-mounted  valves.   However,
as  is the case  for  most EPA emission standards, the refueling
emission  standard  is a  performance standard and  not a design
standard.   As  such, EPA does not, by  regulation, mandate which
onboard  system  designs must  be  used.   Therefore,  it is possible
that  some manufacturers may  elect alternative designs that may
introduce more  complexity.

      However, whether a particular design will lead to a recall
appears  to  be  independent  of the  complexity of that design.  As
was shown in Section 3B, there  is  a  wide  range of current fuel/
evaporative  designs with  respect to  complexity.  In  spite of
the  complexity  that exists  in  some  current  systems,  we are
unaware  of  any information from NHTSA   or  manufacturers  that
demonstrates or suggests that  system  complexity contributes to
more  safety problems/recalls on  current  vehicles.  Absent  this
information,  we  assume  manufacturers have   accommodated   more
complexity   into   fuel  systems  without  compromising   safety.
Conseguently,   EPA  expects  manufacturers  can  accommodate   a
"complex"  onboard  design using methods similar to  those  used to
insure other added fuel  system complexities in today's vehicles
(such as fuel  injection and complex  evaporative designs)  have
not degraded safety.

      Even if  one  chooses  not  to  accept the  conclusion  that
increased  complexity  does  not  inherently  lead  to   increased
safety  problems/recalls,   as  was  explained  in  Section  3B,
onboard   systems   do   not  have  to  add  complexity.    In  fact,

-------
                              3-94


onboard  systems  can  simplify  features  of  current  systems.
Thus,  if  the complexity  rationale is  adhered  to,  it  follows
that  the  use  of   a   simple  onboard  approach  provides  the
opportunity  to  reduce  the  number of  recalls in  three of  the
four   areas   analyzed,   and  some  decrease   in  frequency  of
occurrence is possible  for  the  types  of problems  identified by
NHTSA in their list of 38 onboard related recalls.

     In addition  to the four categories of  recalls  supplied by
NHTSA,  six  manufacturers  included information  on  82  recalls
they believe to be  relevant  to onboard controls.   Some of the
recalls  had  been  identified  previously  in  EPA's  review  of
evaporative  system  recalls, or  were  included in NHTSA's list of
38 discussed above.  These  include those in  the  categories such
as fillpipes, volatility,  and canisters shown in Table 3-4, and
EPA's  view  on  the  relevance  of  these  recalls  is  unchanged.
Table  3-4  also  shows a number  of  other recall categories which
the  manufacturers   suggested  could   be  relevant  to  onboard
controls.   Many  of these  were  related  to  the  fuel  delivery
system and were included in the comments  without explanation as
to  how  they relate to  onboard  controls.    EPA reviewed the
summary  information on  these  closely,  but  could  not  see any
connection   between these   fuel  system  recalls  and  onboard
systems.

     Another area  of recalls  supplied by manufacturers  involves
problems  related  to  crash  shields.    Some manufacturers  use
crash  shields  to  protect hardware   such  as   fuel  tanks  and
fillpipes  in accident  situations,  and  it  has  been suggested
that   onboard   control  systems may  require  additional  crash
shields  which  could  lead  to  additional   problems.    However,
there  is  nothing  inherent  about onboard controls  which drives
the  use of  crash  shields.   Crash shields  are  a  design choice
which  appear on  some  vehicles and  not others  with  the  same
design  feature.    Similarly,    some   onboard  designs   may
incorporate  crash  shields,  but  incrementally,  there  is  no
reason to  believe  that onboard controls will  increase the use
of   crash  shields.   Consequently,  onboard  controls   are  not
expected to  affect crash, shield recalls.

      In  addition  to   the  38   recalls NHTSA   believes to  be
directly  related to onboard  controls,  several  hundred recalls
involving  fire, exhaust temperatures,  and driveability  problems
were also  recently  provided  by NHTSA,[16] with the suggestion
that EPA evaluate  these  recalls for  the  possibility that  that
vapor  recovery  systems  may indirectly create problems for  other
vehicle  systems.   NHTSA's letter  also provided  some  specific
examples  as evidence for  their concerns.   For  instance,   NHTSA
identified the  five recalls  in Table  3-5  as examples of how an
emission   control   system   could  adversely   affect  exhaust
temperatures.   NHTSA   further   implied that  emission  control
systems  such as onboard refueling controls could  similarly  lead
to  stalling/driveability and  fire recalls.

-------
                        3-95
                     Table 3-4
   Manufacturer Submitted Recalls Suggested To Be
   	Relevant to Onboard Control Systems	
Category                              No.  of Recalls
Fuel Lines                                  31
Fillpipes                                    3
Fuel Line Clamps/Connections                17
Fuel Tank                                    7
Fuel Pump                                    3
Crash Shields                                6
Fuel Reservoir                               1
Volatility                                   1
Fuel Line Plugs                              5
Fuel Tank Caps                               2
Diesel Fuel/Water  Separator                  2
Fuel Rail                                    1
Fuel Switching Valve                         1
Canister                                    _2
                             Total:         82

-------
                                           3-96
                                        Table 3-5
78V203


80E003

82V033

85V077


87V165
                           Exhaust System Recalls  Identified by
                          NHTSA as Relevant to Onboard Controls
NHTSA
Campaign
Number

Model
Year(s)

Vehicle
Type
Number of
Affected
Vehicles


Description of Problem
1978     Pass.  Veh.  218,500    Defective  "pulse   air  reed  valve."   (part   of
                              exhaust emission control system).
1977-78  Pass.  Veh.

1983     Lt.  Truck

1984-85  Pass.  Veh.
   49   Missing exhaust pipe heat shield.

   24   Muffler grass shield inadvertently omitted.

8,671   Heat shields for catalytic  converter  outlet pipe
        were omitted.
1983     Pass. Veh.  126,319   "Pulsair check  valve"  could  permit exhaust  gas
                              to melt plastic shut off valve.

-------
                              3-97
     EPA examined  the  more than  350  recalls provided by  NHTSA
to evaluate  the possible indirect  effects  of onboard  controls
on other types  of recalls.   We first looked at  the  20  exhaust
related recalls with particular  attention  placed  on the  five
specifically  identified by  NHTSA  (see  Table  3-5).   Three  of
these  five  recalls  (80E003,  82V033,   and  85V077) simply  dealt
with the addition  of  a  heat shield, with no relation whatsoever
to vapor recovery  controls.   The  other two  (78V203 and  87V165)
concerned  an   adverse  exhaust  emission  control  interaction.
These  two  recalls  had  absolutely  no  connection  to  either
evaporative or  onboard  controls.   Further,  none of the other 15
exhaust  recalls  supplied  by  NHTSA  were  influenced  by  the
inclusion  of   an  evaporative   control  system   or   would  be
influenced by onboard controls.

     EPA  also  examined  31   stalling/driveability  recalls  for
some possible connection  to  vapor control systems.  However, no
recalls  in  addition  to ones  included on  NHTSA's  list of  38
(Table  3-5)  were found in which  the  evaporative  control system
adversely  affected  (directly or  indirectly)  a  recall  for  a
driveability  problem.   Stalling/driveability was raised  as  an
onboard  issue   because  of  the   increased  purge  capability
allegedly  required  by  an  onboard system.   However,  increased
purge  rates  would already  be required by  the  evaporative-only
test  procedure  described  in  the August  19,  1987  volatility
proposal, regardless of whether the onboard requirement becomes
final.[18]   Incremental to  the purge rates  dictated by the test
requirements  of the  volatility proposal, onboard controls will
have no  effect  on these problems.  Indeed,  for both evaporative
and  refueling  controls,  manufacturers have a  wide  degree  of
latitude  in  trading off  purge  rate  versus canister  size and
could  keep  purge rates  at  near  current   levels  if  desired.
Therefore,  no  problems  are expected  from  either the refueling
requirements  or  the evaporative control  requirements  of  the
volatility NPRM.

     Finally, EPA examined the 314 recalls related to fire for
some  possible  relation  to  current   evaporative or  potential
future  onboard control  systems.   A  careful  examination found
about  half  a  dozen  recalls  related to   fuel  spurting  from
overpressurized  fuel   tanks,    which  some   commenters   have
attributed  to  the evaporative  control system.   However,  there
are  several  reasons  why the  evaporative  control system cannot
be  held  responsible   for  these  problems.   As discussed   in
Chapter  2  and previously in  this  chapter,  high volatility fuels
are  the major  causes of these  problems,  along  with the need  to
pressurize  the  fuel  tank to reduce  vapor  generation and  limit
spillage  during FMVSS  301  testing.  These  problems  would have
occurred   even  without   evaporative  control   systems.     As
described above,  EPA is including a provision in the  reproposal
test    procedure  to   discourage   pressurized    fuel   tanks.
Therefore,  this problem  is  likely to be reduced in the  future
as   a   result  of  the   revised   test procedure  requirements.

-------
                              3-98
Further, even  apart from onboard  and evaporative requirements,
this problem would  likely be reduced in the future  as  a result
of EPA's proposed volatility controls.[18]

     In  addition,   EPA  also  found two  fire recalls  involving
problems which some commenters have  suggested  are  relevant  to
onboard  controls.   One  concerned  a  solenoid problem  (82V091),
and the  other  involved  an  electrically ungrounded  filler  inlet
(81V092).   These  components  are  in  common  use  in  today's
systems, and  their  use  is  expected to be  continued regardless
of onboard  controls.  Problems  such as these indicate the types
of concerns that must  be  considered  by  any manufacturer  when
designing  systems  (including  ones  on today's vehicles)  which
utilize  solenoid  vent  valves  or  plastic  fillnecks.   These
concerns  are  not  new   or   unique  to  onboard  vapor  recovery.
Therefore,  it is  difficult to  conclude  that  onboard  systems
would  noticeably affect the number  of these types  of  recalls.
Furthermore,  the fact that  only  two  recalls  related  to  these
concerns have appeared  in  NHTSA's  records spanning more  than
twenty  years,  indicates that  these types  of  problems  are  not
widespread  and are  likely to continue to be  a  minor concern in
the future.

     Summary  of Recall  Information:   As  was  mentioned  in  the
beginning of  this discussion,  the number of recalls involved in
this analysis  represent  a small fraction  of the total  number of
safety  recalls.   The  38 recalls  identified by  NHTSA  as being
directly relevant to onboard  controls represent less  than  one
percent  of  the more than 4200 safety  recalls that have occurred
during  the  past two  decades.   This small  number of  recalls
demonstrates  that manufacturers have successfully designed safe
fuel/evaporative systems.

     In  addition,  although  it  has  been  shown that  onboard
system designs  can be  quite simple,  such  as  that developed by
EPA,   or more  complex  such   as  those  provided  by  several
commenters,  it  is  important to  recognize  that  a wide range of
complexity  exists  in today's  vehicles.    In  fact,  a  general
trend  has  been  toward  increased  complexity   in recent years.
Despite  this  increase complexity,  there  is no  suggestion from
the  recall data that  indicates  fuel/evaporative system safety
has  degraded.   Given  the   similarity   between  onboard  and
evaporative  systems, EPA  has concluded  that  the  addition of
onboard  controls  will  not  increase  the  number  of  future
fuel/vapor  system recalls.

     This   is  not  to  say  that  onboard controls  will  not  be
involved in any future  recalls.   We would expect  that with or
without  an onboard  requirement,  some  level   of recalls would
continue to  occur  with vapor  recovery  systems  simply  due  to
deficient    designs,    mistakes    in   production,   defective
components, etc.   Any   future  problems  would  likely  be minor
since   the  problems  that   have   occurred  with  current  vapor

-------
                              3-99
recovery   systems    have    been   infrequent   with    minimal
consequences.   Further,  since  onboard   control   systems  are
essentially   an  extension   of   evaporative   systems,   their
incremental  effect  on recalls is  expected to be  undetectable.
Manufacturers can build on  the experience gained in 18 years of
designing  and implementing  complementary evaporative  systems,
and  there is no  reason  to  expect  any  significant number  of
problems  with onboard systems over those that happened  in. the
last  and  would  occur  in  the future   regardless  of  onboard
systems.

     ii.   Technical Service Bulletins

     Summary  of the Comments:   As  part of their April  12, 1988
letter,  NHTSA  supplied EPA  with a  large  number of  service
bulletin   summaries.[16]    The  categories  contained   in  the
bulletins  included  all  fuel  system,  carburetor,  exhaust,  and
emission   system  bulletins  (over  6800  bulletins  in  total).
NHTSA  included these 6800  service bulletins so that  EPA might
be  able to  identify  ways  in  which onboard  refueling controls
might  directly or  indirectly  lead to problems  affecting other
vehicle systems.

     In addition,  EPA and  NHTSA asked manufacturers to include
as  part  of  their  comments   on   the  NPRM  information  on  any
service bulletins  which  they believed  would  be  relevant  to
onboard control  systems.   In response, 5 manufacturers provided
a total of 62 bulletins which they stated might be relevant to
onboard controls.   A summary listing  of the  service bulletin
information  provided by  the  manufacturers is  given  in Table
3-6.   Like recalls, the service bulletins that were provided by
manufacturers  were   intended  to   demonstrate   the  generic
increased complexity argument which  postulates  that the  use of
more/bigger   components  can  lead  to more service bulletins.
Examples  of  service bulletins provided by manufacturers  include
broken or  defective  canisters,   vapor  line  problems,  tank
vent/overpressurization difficulties,  and improper  purging.

     Response to  the Comments:   As with recalls,   any  analysis
of  service  bulletins with  respect  to  onboard  vapor  recovery
will  depend  on designs selected  by  manufacturers and must be
viewed incrementally to current  fuel/evaporative systems.  The
analysis  of  service  bulletins  is not as  confined to  specific
bulletins as  was  the  recall  analysis   because NHTSA  had  not
stated explicitly which individual bulletins they  believe to be
relevant  to  onboard controls.   Rather, NHTSA  provided EPA with
thousands of bulletins to  examine with the possibility of  some
being  relevant to  onboard controls.

     As  discussed  in  Chapter   2,   a  review  of  these  6800
bulletins revealed  between   70   and  120  which  were  directly
relevant  to  vapor  control  systems.   These  70  to  120  bulletins
essentially   represent  all  potential problems  which  may be

-------
                             3-100

                           Table 3-6
         Manufacturer Submitted Service Bulletins Said
           To Be Relevant to Onboard Control Systems
                                    No of  Service
Category                              Bulletins
Driveability                             15
Purge                                    10
Evaporative System Related               10
Fuel System (Tank,  Lines,
  Pump, Filter,  Switch)                  15
Fuel Fill Difficulty                      1
Noise/Odor                                6
Fuel Filler Door/Cap                      2
Water  in Fuel Light-Diesel                1
Diagnosis Information                    _2
                          Total:         62

-------
                             3-101
relevant to  onboard vapor recovery, because  no  other bulletins
were  found  which  might  even   tangentially  involve  onboard
controls without  involving evaporative  systems.  It  should be
noted that EPA identified two bulletins  involving problems that
some   commenters   suggested  would   be  relevant   to  onboard
controls.   One  involved static  charge  build  up  on  plastic
parts,  and  the  other  involved   the  grounding  of   a  solenoid.
However, as  was  discussed above  for  recalls, there is  nothing
inherent   about  onboard  controls  which  drives  the  use  of
additional ground wires,  and  incrementally,  onboard  should have
no  effect  on  these  problems.    In addition,  EPA reviewed the
service bulletin information  provided by the  manufacturers and
found  no  new potential problem areas.   Concerns with regard_to
driveability-, purge-,  and  evaporative-related  hardware  exist
now  and  will  not  be  incrementally  affected   by  an  onboard
requirement.  The remainder  of  the service  bulletins suggested
by the manufacturers  were not related to onboard systems or had
no associated safety  implications.   Therefore,  in total,  70 to
120  service  bulletins  were  identified as  relevant to  vapor
control systems.

     Even  so,  as discussed  in  Chapter  2,  70  to  120  service
bulletins   represent   a  minute   fraction   (approximately  0.1
percent) of  the  total  number  of  service bulletins  issued  over
the  years.    Because  onboard   systems  are  modifications  of
current  evaporative control systems,  and consequently  will be
similar,  additional  problems over  and  above  those  which would
normally   occur  with  evaporative  control   systems  are  not
expected.   Further,  depending  on  the  design,  onboard  systems
have  the  potential  to decrease  problems,  such as  those  with
external vent lines  or  fuel  system overpressurization and  fuel
spurting.

     c.    Conclusions

     After  a careful  review    of  all   recalls   and  service
bulletins  provided  by  NHTSA  and the  manufacturers,  EPA has
determined that  problems most relevant  to  onboard  controls are
those   which have  occurred  with  current   evaporative  control
systems,   since  the  components   of  these  two  systems  are so
similar.   Past experience with evaporative control systems  (in
the  form  of  recalls  and service  bulletins)   indicates   very
minimal problems with the types of  components envisioned for
use  on  onboard  systems.  As  a  matter  of  fact,  since onboard
systems are  expected to require only  marginal changes to the
current evaporative  control  systems,  the  incremental  increase
in  recalls/service bulletins  with onboard  systems is  expected
to  be insignificant relative to  current  systems.   Also, if one
adheres to the complexity/risk rationale, EPA has demonstrated
that  onboard  systems can be  designed  to  reduce the usage of
such  components  as external  vent lines and  certain evaporative
hardware,  so  that  current  problems  in  these  areas  can be
reduced  in  the  future  as  a   result of   onboard controls.

-------
                             3-102
Further,  because   of   test  procedure   requirements,   onboard
systems will reduce tank  operating  pressures,  and decrease  the
number  of  recalls  related  to  tank  overpressurization/fuel
spurting.   The incremental effect of onboard controls  on future
recalls and service bulletins,  relative  to  current  evaporative
controls,   is  likely  to  be of  insignificant  magnitude to  be
detected,  and may in fact be an improvement.

     4.    Crashworthiness

     a.    Introduction

     One  principle  concern  regarding  the  in-use  safety  of
onboard systems  involves  the  crash resistance  of  the  system.
In  particular,   some  commenters  questioned  the  strength  of
plastic parts and the  consequences  of  onboard  system components
interfering with designed crash spaces.   These  concerns  were
addressed  in  Sections  B  and  C of  this  chapter.  However,  one
commenter  (General  Motors) had a  contractor, Failure Analysis
Associates  (FaAA),  crash  test  an  onboard equipped vehicle to
demonstrate  the  susceptibility  of  onboard  components  in   a
crash.  The significance of this crash test is discussed below.

     b.    Summary  and Analysis of Comments

     Summary  of  the  Comments:   Aside  from General Motors, no
other   commenter  submitted  crash  test   results  or  challenged
EPA's  finding that onboard  systems could  be  designed  to   pass
NHTSA's  fuel  system  integrity  standard  (FMVSS  301).   General
Motors,  however,  submitted a  videotaped  demonstration of an
onboard equipped vehicle  being crash tested  and also  provided
written documentation  of  the  results.[19]  in  this particular
crash  test,  the onboard  equipped  vehicle  was   subjected  to  a
thirty mile  per hour  side  impact.    Following  the  crash,   a
measurement was  made of the fuel leakage rate and was  found to
be  5.3 ounces in 5 minutes.   The  test  conditions were similar
to  those  required  by  part of FMVSS  301, except that the crash
impact point  on  the test  vehicle was centered  directly at the
fuel  fillneck instead of  the  centerpoint of  the  side of the
vehicle,  and  another vehicle was used in the  collision instead
of  a barrier.  The test  vehicle  was  equipped  with a replica of
an  onboard system  prototype originally designed by  Mobil Oil
which  was  not   intended   to  be  production  quality (IV-D-329).
Even   though  it  was  not  an   official   FMVSS   301  test  of   a
production quality onboard system, the  5.3 ounces per  5 minute
leak  rate of  the test vehicle was  compared unfavorably against
the 5.0 ounces per  5 minute standard of  FMVSS  301.

      Response to  the  Comments:  Several aspects of  the crash
test  performed  by FaAA for General Motors combine to  produce  a
test  which yields  results of  questionable value.   The use  of
the test  results to characterize those  which  might be  expected
of  a legitimate production quality onboard  system is  misleading
at  best.   The key  problems with the testing are  discussed below.

-------
                             3-103


     To begin  with,  the stock  fuel/evaporative system was  not
subjected to  a crash  test  identical  to  that  imposed  on  the
onboard   equipped  vehicle   to   establish   a  baseline   for
comparison.   The  performance of  the stock  system under  these
same conditions is unknown.   It may or may not  have  passed  the
FMVSS  301 standard  under these  conditions.   While it might  be
assumed that  the stock  vehicle would  pass FMVSS  301  testing,
the onboard equipped vehicle was not  subjected to a FMVSS  301
test.   A  lack of  comparable tests  for  the stock  and  modified
vehicle makes  it  difficult  to  draw any  conclusions about  the
effect of the onboard system relative to the stock vehicle.

     Second,  the  onboard  system configuration chosen for  the
crash test was a replica of  a prototype design intended  only to
demonstrate the  feasibility  of a particular  onboard refueling
control concept.   This particular onboard  configuration  was  not
constructed with  regard  to safety or passing a crash test.  The
system  had   not   yet   been  adapted   to   "common  automotive
production  methods  and  materials  as prescribed by Mobil as a
necessary step before  implementation."   It is  obvious  that  any
production  ready  system would  not  incorporate a  common  paint
thinner  can  for  a  liquid/vapor  separator,   or  copper  tubing
connected to  rubber  hose with  radiator clamps for vapor lines.
Without  first adapting  this  prototype or   any other  onboard
system  design  to  common   automotive  production  methods  and
standards, it would not be surprising for a fuel leak to occur.

     In  fact,  taking  into  consideration the  fact  that  the
onboard  equipped vehicle  was  subjected  to   a  crash situation
involving  one of the  most  vulnerable portions   of  the  fuel
system (the fillpipe),  and  that  crash  safety was not accounted
for during the construction  of  the particular onboard prototype
tested,  a  leakage  rate of  5.3 ounces  in 5  minutes indicates
that  insuring  crashworthiness of an onboard system will  not be
a  difficult task.   If  a  system which  did not  consider  crash
safety in  its  construction  could  perform  this well, this is  a
good   indication  that  a   system  which  incorporates  correct
automotive  materials,   and  production,   design  and  assembly
methods could pass FMVSS 301 readily.

     Finally,  it  is  not   technically  valid  to  compare  the
leakage  rate  of  a  crash  test  involving  conditions  other than
those  required by FMVSS 301 against the  FMVSS 301 standard  and
then  to assert that this  system  was less safe  than a baseline
vehicle which was not  tested under the same conditions.

     Taking  into  consideration  all  of the inadequacies of this
crash  test, EPA  has concluded  that  no  results can be extracted
from   this  testing  that  apply  to  any  onboard  system  (well
designed  and  production ready or otherwise).  Nevertheless,  EPA
understands that  the crashworthiness of an onboard system is  an
important  element in  the  design and development  of  an onboard
system.   However,  no  other commenters suggested that onboard

-------
                             3-104
systems could not  be  designed to pass FMVSS 301  crash  tests  or
provide a high  level  of in-use  fuel  system integrity,  and  EPA
expects that  any onboard system  can  be designed with  the same
or  better  level  of  crash  resistance  as  current  systems  if
crashworthiness is given proper  consideration  in the design and
development process.

     5.    Failure Modes and Effects Analyses

     a.    Introduction

     Failure  Mode  and  Effect Analysis (FMEA)  is  a  structured
analytical  technique  that  is widely  used  in the  automotive
industry to  assess the potential  risks of  a  new  system.   The
primary objective  of  an FMEA is to minimize the risk and in-use
consequences  (effects)  by  determining corrective  actions  to
prevent  identified  failure  modes.   NHTSA  and  other  safety
organizations   have   suggested  an  FMEA   would  be   a  useful
technique  for  EPA to use  to  evaluate  the  risks of  onboard
systems.

     b.    Summary and Analysis of Comments

     Summary  of the  Comments:  To  support various  viewpoints
regarding  onboard  safety,  three  commenters submitted FMEAs for
onboard and  evaporative systems.   As part  of  their  critique of
the  Mobil  onboard  system,  General Motors  submitted  two FMEAs,
one   was  the   FMEA   originally   prepared   for   the  stock
fuel/evaporative system selected for  modification by Mobil, and
the  second  was a  contractor  prepared FMEA  of the  prototype
onboard  system developed  by Mobil.[19]   In  addition,  Ford
submitted  an FMEA  type analysis  which compared the  potential
failure  modes  and  effects of  a  current  evaporative  system to
those  of  three different  onboard control  approaches.[20]  API
also   submitted  an  FMEA  type   analysis  which  compared  the
relative   risks of   several   different  onboard  configurations
against those of a  simple  current  evaporative  system.[10]

     Response  to  the  Comments:   EPA  examined the  four FMEAs
submitted  by commenters and  reached  the  following conclusions.
First,  the design  FMEA submitted  by  General Motors on the stock
fuel/evaporative system provided good insight on the  scope and
depth  that  the automotive industry  typically enters  into with
this type  of  analysis,  the failure modes  and effects identified
for  a  typical   fuel/evaporative   system,   and the  manner which
risks   are  considered.   It   also   provided  good  background
information  for  the  FMEA  now  being  conducted  for  EPA  (see
below).  The other three FMEAs were  not  particularly  useful to
EPA's  safety analysis  for  the following reasons.

     The   contractor   prepared  FMEA  of   the  Mobil  prototype
onboard system is   not   directly  relevant  to  an  objective
analysis  of  onboard  safety.   The Mobil prototype system was an

-------
                             3-105
early  design   prototype   intended  only   to   demonstrate  the
feasibility of  a  refueling control  concept.   Before  a  serious
evaluation of  the  safety  of this system could  be  performed,  it
would  be  necessary to  adapt  this  system  approach  and  its
components to common automotive  industry  standards for design,
production  methods,   and   materials.    Component   design  and
material  selection for  the  prototype were  based  on ease  of
assembly for test  purposes.   It  is obvious that this prototype
was  never  intended  to  be  production  quality.    Aside  from
identifying corrective actions  for  gross  system  inadequacies,
an  FMEA  of  a concept  demonstration  prototype  system  yields
meaningless results regarding the  safety  risks of  a properly
designed, production ready system.

     EPA  also  received FMEA  type  analyses from  Ford and API.
(The  API analysis  was performed  by ICF).   The  Ford analysis
only   indicated  whether   potential   failures   and   associated
effects  were   possible for  a particular  system.    It did  not
evaluate  the  likelyhood   of   failure   or  the  severity  of  the
effects  for  comparison  among the  systems.   The   ICF analysis
compared  the  risks  of generic  onboard systems to those  of   a
generic  evaporative system.   While the  analyses   performed  by
Ford  and ICF  identified  general problem  and  improvement areas
for the  systems,  neither  analysis was  sufficiently complete to
yield  conclusive results.

     Although  the  FMEAs  submitted to  EPA did  not produce any
significant revelations  about the  safety  of  onboard controls,
EPA recognizes  the value  of a properly performed  FMEA as  a risk
analysis tool.  Subsequent to  the NPRM, based on  suggestions by
NHTSA  and others,  EPA entered into  a  contract  in  the spring of
1988   to  perform   a   comparative risk assessment   of   onboard
controls.   In  this work  assignment,   the contractor will use
FMEAs  to  evaluate the  incremental risks of  a  few different
onboard  system configurations for comparison to the  incremental
risks  associated  with  other recent fuel system  changes  such as
carburetion to fuel injection.   Once  it  is completed, EPA will
use the  findings of this FMEA  as one input in its  deliberations
regarding  a final  rule for onboard vapor recovery  systems.

-------
                             3-106


                    References for Chapter 3
     1.    "Costs  and   Cost-Effectiveness   of  Stage   II   and
Onboard Refueling  Vapor  Controls," Report Prepared  for MVMA by
Multinational Business Services, Inc., April, 1987.

     2.    "A   Study  of   Uncontrolled  Automotive   Refueling
Emissions,"  Report  Prepared  for  CRC  by  Automotive  Testing
Laboratories, January 5,  1988.

     3.    "Design Considerations  for  Plastic  Fuel Lines,"  M.J.
Harrigan, SAE 880683

     4.    "Nylon    12    in   the    Automotive    Fuel   System
Environment," Edward K. Gray, Gerhard Hopf, SAE 880684

     5.    "Study   of  Motor   Vehicle  Fires,"   Draft  Report
Prepared for NHTSA by Data Link, Inc., February, 1988.

     6.    National  Truck  Equipment  Association's  Comments on
Onboard Proposal, February  11,  1988.

     7.    Onboard/Volatility  Hearing  Transcript,  Washington/
Dulles Holiday  Inn,  Sterling, VA,  October 27-29, 1987,

     8.    Letter  to  Charles  L.  Gray,   Jr.,  U.S.   EPA,   from
Gordon  E.  Allardyce,  Manager,   Certification  and  Regulatory
Programs, Chrysler Motors, May  11,  1988.

     9.    Attachments    to   Memorandum    "EPA/Exxon   Meeting
Regarding  Onboard  Controls,"  Glenn  W.  Passavant,  U.S.   EPA,
March  1, 1988.

     10.   Comments   of   API   in   response   to  Onboard  NPRM,
February   11,   1988,   available   in  public  docket  A-87-11 at
IV-D-358 plus sub-entries.

     11.   Memorandum  "Status   of   In-House  Refueling   Loss
Measurements,"  Martin  Reineman,  SDSB  to  Robert E.   Maxwell,
SDSB,  March  6,  1979.

     12.   "Vehicle  Onboard Refueling Control," API Publication
No.  4424, March 1986.

     13.   Memorandum  "EPA  Meetings  With  Potential  Onboard
Rollover/Vent Value Manufacturers," Jean Schwendeman,  U.S.  EPA,
June 16,  1988.

     14.   "Characterization of Fuel/Vapor  Handling  Systems of
Heavy-Duty  Gasoline Vehicles  over  10,000  pounds  GVW,"   Jack
Faucett  Associates,  September 30,  1985.

-------
                             3-107
     15.   "Evaluation of  Air  Pollution Regulatory  Strategies
for Gasoline Marketing Industry -  Response to Public Comments,"
March, 1987.

     16.   Letter to  Chester  J. France,  U.S.  EPA,  from  George
L. Parker, NHTSA, April 12, 1988.

     17.   Letter to  Glenn W.  Passavant, U.S.  EPA, from Paul H.
Yoshida, NHTSA, January 7, 1988.

     18.   "Air Pollution Control;  Fuel and Fuel  Additives and
New  Motor  Vehicles  and  Engines,  etc.:  Gasoline  and  Alcohol
Blends Volatility and Evaporative  Emissions;  Notice of Proposed
Rulemaking," 52FR31274, August  19, 1987.

     19.   Comments  of  General Motors on  Onboard NPRM,  plus
attachments. Available  in public  docket A-87-11,  IV-D-360 and
additional sub-entries a through f.

     20.   Comments   of  Ford  Motor  Company  on   Onboard  NPRM.
Available in public docket A-87-11, IV-D-362.

-------
                           CHAPTER 4

                   Potential Safety Benefits

     The  previous  chapter  summarized  and  addressed  onboard
system  design  and design related  safety comments.   In  this
chapter,  potential safety benefits  of  onboard  vapor  recovery
systems will  be  discussed and the  basis  for  EPA's view  that
onboard   systems   can   enhance   automotive   safety  will   be
described.   This  includes  EPA's  analysis  of  the  potential
effects of  onboard systems  on the  safety of  service  station
refueling  operations,  as well   as  other  safety  improvements
which could result due  to the effects which the  onboard  system
design  could  have on fuel system and  evaporative system safety
and on  overall  vehicle  crashworthiness.   Each of  these  aspects
will be discussed  in this section of the safety analysis.

A.   Effect on Service Station Safety

     1.    Introduction

     As  was  discussed   in  EPA's  June  1987  report,  from  a
technical view point it  seems  reasonable to expect that onboard
vapor  recovery  systems   would  have a positive  effect  on  the
safety  of  automotive  refueling at  service  stations.  Refueling
vapors  that  are  currently  vented  to  an  area  which  poses
somewhat  of  a  safety hazard  will instead  be routed  away  from
potential  external ignition  sources  to  a safer  location  (the
charcoal  canister).   Also,  due to  test  procedure requirements,
onboard controls  would  be likely to bring about a decrease in
the   amount   of   gasoline   spilled  during  normal   vehicle
refueling.  Both  of  these are likely to  improve  service station
safety.   Refueling-related fires  are  likely  to  be  reduced in
number  and other  non-fire safety problems related to refueling
should  also decrease.  Both EPA  and Failure Analysis Associates
(FaAA,  a  contractor  for  General  Motors)  have  attempted  to
quantify  the  impact  of  onboard  systems  on  service  station
fires.   These  analyses,  along  with  a   discussion  of non-fire
service station safety effects, are presented in  this section.

     2.    Effect  of Onboard Controls  on Service  Station Fires

     a.    Service Station Fires

     EPA's  preliminary  analysis  of  refueling related  service
station  fires  is  based  primarily  on information  contained in
the  National  Fire Incident Reporting  System  (NFIRS).   NFIRS is
a fire  data base  which  is operated and  maintained by the United
States  Fire  Administration of the  Federal Emergency Management
Agency  (FEMA).   It sr.culd be noted that EPA  is currently having
a contractor   further   examine   other   information  on  service

-------
                              4-2
station safety.  This  additional work will  further refine  the
analysis  of  the  risks  and  benefits  pertaining  to  service
station  safety,  but  is  unlikely  to  significantly  alter  the
conclusions  drawn  in  the current  analysis.   The  contractor's
analysis should  be available  in time to  be considered  before
any final rule on refueling vapor controls is promulgated.

     The NFIRS  data base compiles reports  and  information  on
all  types   of fires,   including service  station  fires.    The
system receives  fire marshall  reports  from different states and
extrapolates  national  statistics from these.   Approximately  40
percent of  the total  reported  nationwide fires were reported to
NFIRS  from 1982 to  1985.   However,   a  1985  survey  of  member
service  stations by the  Minnesota Service  Station Association
showed  that there  are about  19 annual  fires  per  200  service
stations,  and historically  only about  1.75  of  the  19  service
station fires  (or  9.2%) are reported to  a  fire departmental]
If  this low reporting  rate  is a nationwide  trend, then the
number of NFIRS projected national fires  may  be much lower than
the  actual  number  of   fires  which  occur  since  NFIRS  only
contains reported fires.

     A  second problem with  the  data used  is  that  about  10
percent  of the  reported  fires in the NFIRS  data base occurred
in the  state of  California (see Appendix III,  Tables  9-15).   A
large  percentage of  California service  stations are outfitted
with  Stage  II  vapor  recovery  systems,  and therefore  already
provide  refueling  vapor  containment  which  may help  to reduce
fires  at  these  stations.  Consequently,  different assumptions
may  need to be  made when considering  how onboard systems would
affect  those fires.   Not  only  do  Stage II  systems  complicate
the  treatment of benefits from vapor  control, but the treatment
of benefits from spilled fuel control are  also  not clear.   For
this  reason,  the California statistics were  subtracted from the
nationwide statistics, and a 49 state data  base (non-stage II)
was used in this analysis.

     Table 1 -  Service  Station Fires,   in  Appendix III,  shows
the  actual service station  fire information reported to NFIRS
for  the period  1982-1985 inclusive and the national  statistics
extrapolated from  those  reports.   California  fire  data  are
contained   in  Table 9 but  the average  annual  statistics are
based  only on 1982-1984  reports and  do  not  include 1985  data,
as   the  national  average   statistics   do.   Subtracting  the
California data  from nationwide data  gives  2466 average annual
service  station  fires and  $6,900,000  estimated dollar losses
from these  fires  in  the remaining 49  states.  The  number of
fires   are  reported   by  category  (structure,   vehicle,  other)
along  with an estimate of the  dollar  value of the property loss
involved.   These two  tables also  show the  reported injuries and
deaths caused by service  station fires.   The reported estimated
dollar losses shown   do  not  include  any costs  related  to the
reported injuries  and  deaths.

-------
                              4--3


     NFIRS  data  can  also  be  used  to  characterize  fires  by
several different factors  as  shown  in  Tables 2  through 8  in
Appendix III.  The nationwide NFIRS data is categorized by what
type of  fire occurred  (Table 2),  the mobile  property  involved
(Table  3),  the  area of  fire  origin  (Table  4),  the  equipment
involved  in  ignition  (Table  5),   the  form   of  heat  ignition
(Table  6),   the  form  of material  ignited (Table  7),   and  the
ignition  factors that  caused  the  fires (Table  8).   Tables  9
through 15  categorize the California  data  in  this  same method,
except  that  a  breakdown  of the   California  fires  by  mobile
property  type  (as   in  Table 3  for  nationwide  fires)   was  not
available.   In  any event,  this  information was not used  in  the
analysis.

     Refueling-related  events  are not  one  of  the  categories
directly reported in the NFIRS  data.  The  characterizations of
service  station fires  must   be   studied  in  an  attempt  to
determine the percentage  of  fires  directly related to refueling
events.   Close  review  of the information contained in  Tables 2
through  8  in Appendix  III  indicates  the  following information
which could  have  some connection to vehicle refueling fires:

                                                    Percent of
     Type of Situation  Found - Tables  2  and 9     49 State Fires

     Outside of  Structure Fire                       - 22%
     Vehicle Fire                                   - 45%
     Outside Spill/Leak                             - 18%

     Area of Fire Origin -  Tables  4 and  11

     Fires which Started in Service/Equipment        -   8%
     Areas but  not in  Maintenance  Shop/Area

     Fires which Originated at Fuel Tank            -   5%
     Area of a  Vehicle

     Open Areas                                     ~   2%

     Equipment  Involved in  Ignition - Tables  5 and  12

     Internal Combustion Engine                      -   6%

     Other  Special Equipment                        —   5%

     Vehicle                                        -  27%

     No Equipment Involved                           =  25%


     Form of Heat Ignition  - Tables 6 and 13

     Heat from a Liquid Fuel-Powered  Object         -  2%

-------
                              4-4
     Heat  from  a  Smoking Material and                - 4%
     Match/Lighter

     Backfire from  Engine                            ~ 14%

     Form  of Material  Ignited  - Tables  7  and  14

     Fuel                                            - 78%

     Atomized/Vaporized Liquid                      ~ 1%

     Gas/Liquid from Pipe                            - 15%


     Ignition Factors - Table  8

     Fuel  Spilled Accident                          - 10%

     Improper Fueling Technique                      = 3%

     Backfire                                        ~ 14%

     If taken  at  face value,  a  summation  of  the  applicable
individual data  in each  table  suggests that refueling-related
fires could constitute between 15 and 94 percent of all service
station fires.    The 15 percent  low  is obtained by summing the
potential   refueling-related  fires  by  Area at  Fire  Origin
(Tables 4  and  11).   The  94 percent high  is obtained  when the
potential   refueling-related fires  are  tallied  from  the  fires
categorized according  to  the  Form  of  Material Ignited (Tables 7
and 14).  Even though the categories of  fires listed  above have
been   identified   as   potentially  being  associated  with  the
refueling  process,  only  a  certain  percentage of  the  fires  in
most  of  them  are  actually  directly  caused  from  refueling
emissions  and/or  spillage.   The  data  base is  only  detailed
enough  to  allow estimating a range of  gasoline  service station
fires associated with these causes  rather than a point estimate.

     In several  of the fire categories  initially  identified as
potentially refueling-related it  is probable that many  of the
fires  could  be  totally  unrelated to  refueling.  For  example,
fires  identified in Tables  2  and 10,  Outside of Structure Fires
(22 percent),  may  not  be  all  refueling-related.  Along  the same
lines,  Vehicle Fires  (45 percent) could  be  in the engine or
numerous other places  on  the  vehicle,  far  from refueling vapors
or  fuel  spilled  during  refueling.   This  is confirmed by the
information in Tables  4 and 11.  The figure  which  suggests that
internal combustion engines were involved  in six percent of the
fires   is  not  very   informative  for  two   different  reasons.
First,  since  a   separate  listing is  provided for  vehicles,
apparently  these   internal  combustion  engines  are   not  on
vehicles  but  on  other power equipment.   Second,  even if they
are  vehicles,  the engine  is far  from  the   fuel  tank, so any

-------
                              4-5
engine-related ignitions during refueling would  have to be  due
to vaporized  gasoline,  which is said  to attribute  to  only  one
percent of the fires (Tables 7 and  14).

     Several other categories, both of high-  and low-percentage
frequency,  although  related,  are far too  general to be  useful
in  estimating   the  percentage  of  refueling-related  fires.
Examples are  service/equipment  area fires  (eight percent),  open
area  fires   (two  percent),  fires  ignited  by  vehicles   (27
percent)  or  ignited with  no equipment  involved  (25  percent).
Other examples are heat ignition from a  fuel-powered object  (22
percent),  from  a  smoking  material  and  match  lighter  (four
percent) or from an engine backfire (14 percent  in Tables  6  and
13, or  13  percent in Tables 8 and 15).   The knowledge that fuel
was the  material  ignited  in 78 percent  of  the  fires  (Table 7)
is  also too  broad  a category  to  be  useful  in  determining  the
percentage of refueling-related fires.

     The  remaining  five  categories  in  the  list of  potential
refueling-related   fires   are   those   which  may  be  the  best
predictors   of   the  actual  percentage   of   refueling-related
fires.   First,  with regard  to  vapor,  gasoline  vapor (atomized/
vaporized  liquid,  Table   7)  was  the  material  ignited in  one
percent  of the  fires.   It  is  not unreasonable  to  assume that
many  of  these   fires  were  due  to  vapors  generated  during
refueling,  since the  refueling  process  is  probably  the  most
common   source  of  vapor   generation   at   a   service  station.
Second,  with regard  to  gasoline spills,   outside spill/leak
situations  (18  percent)  can be  narrowed  somewhat   by looking
instead  at  fuel-spilled  accidents  (10  percent),  which  would
seem to  be  a  subset  of  the former  category.  Furthermore,  fires
caused   by  improper  fueling  techniques  (3  percent)  are   a
separate category from  the fuel-spilled accidents.  While fires
which  occur due  to  improper fueling technique  (3  percent)  may
be  a  subset  of  fuel-spilled accidents (10 percent) one cannot
preclude the possibility  that some portion of the difference in
these  percentages (7 percent)  is  also related  to fuel spilled
during  refueling.  For  example,  the fires which originated at
the  fuel  tank   area  of  the vehicle  (five  percent,   Table 4)
likely  occurred  during  refueling.

     Thus  we  are  left  with  four useful  pieces of  data.   As
discussed  above, it appears  that  about one  percent of service
station fires are related to atomized vapor/liquid  such as  that
which   occurs with  each   refueling   event.    With  regard  to
spillage,  the data  suggests  that  ten percent  of fires are due
to  spilled fuel,  but  clearly spill-related fires could occur in
service/maintenance areas  as  well.   To narrow this  estimate,
other  characterizations  in the data  suggest that  about three
percent of   fires  are  related to  improper  fueling  technique
(presumably causing a spill)  and that  five percent  occur in the
tank  area  of the vehicle being refueled  (where spills occur).
Using  these  values, about three  to   five percent  of service

-------
                              4-6
station  fires  are  related  to  fuel spilled  during  refueling.
Assuming  that   the  vapor   and  spillage  data   are  mutually
exclusive (no overlap in  the  data)  this analysis  suggests  that
approximately four  to  six percent of all service  station fires
are  due to  refueling  emissions  and/or  spillage.   In  actual
numbers  of   fires,  this  amounts  to  approximately  99  to  148
annual  fires  in the 49 non-Stage II states  (based  upon  NFIRS
1982-1985 average  projected  national   incidents   and  1982-1984
average projected California incidents).

     b.    Impacts  of  Onboard  on  Reducing  Refueling-Related
           Service Station Fires

     Up  to this  point  in  the analysis,  an attempt has been made
to use the  NFIRS data  base to estimate  the  number  of  service
station  fires  which may  be  due  to refueling  emissions and/or
spillage.   The  goal  of  the  analysis  is   to   estimate  the
percentage  of  these  refueling-related  fires  which might  be
prevented by onboard controls.   In order to achieve  this goal,
two  additional  estimates  must be  made,  the first concerning the
efficiency  with  which onboard  controls  could  prevent  fires
related  to vapors and  the second concerning the efficiency with
which   onboard   controls   could   prevent   fires   related   to
spillage.  For  the  first  estimate it is assumed that a properly
functioning  onboard   system  could  prevent   essentially  100
percent  of the  approximately 25 fires  due to atomized vapors or
liquids  (approximately  one  percent  of  the  total  fires),  since
refueling vapors are controlled  almost completely with onboard
technology.

     In  order  to  make  the  second necessary  estimate  of  the
percentage  of  these refueling spill related  fires which would
be   prevented   if  onboard   controls   were   implemented,   EPA
referenced   the   1972    Scott   Research   Laboratories   report
"Investigation  of  Passenger  Car  Refueling  Losses".[2]   This
report  categorizes  refueling   spillage   into  the  following
groups:   prefill drip,  spitback, overfill  and postfill  drip.
Based   on   actual  field   studies  of   consumer   refueling,  it
estimates the  probability and average   amount  for  each  spillage
type.   The  study found that  over  50  percent  of  the volume of
fuel spilled  during refueling  is  due  to spitback and that the
average  emission  factor  associated with spitback spillage is
0.15 g/gallon of  dispensed  fuel.   This  is an important  finding,
since   EPA's  refueling   proposal  necessitates  the  design  of
vehicles which  can accommodate in-use  dispensing   rates without
premature nozzle shut-offs  and spitback spillage.  According to
the  test  procedure  requirements   and the  proposed   emission
standard (0.10  g/gallcn  dispensed),  a test  vehicle would fail
the  certification test  if almost  any spitback  spillage  occurred
(0.15  g/gallpn  dispensed).   Therefore,  the occurrence of  in-use
spitback spillage  should  be  substantially   reduced with  the
onboard proposal.

-------
                              4-7
     Applying the Scott Research Lab finding that  50  percent  of
refueling spills  are due  to spitback  to the  NFIRS fire  data
implies that  50  percent of  the estimated  74-123 annual  fires
due to  refueling  spills (see analysis  in Section 1),  or  37-61
fires, may be prevented by onboard controls.   This may  even  be
a  conservative  estimate,  since spills  due  to spitback  might
actually result in  a higher  rate of fires than the  other  types
of spills  (overfill,  prefill drip and postfill drip) because a
relatively larger volume  of  fuel  is spilled  at   each  spitback
event.  The Scott Report found  that the spit-back spills result
in an average of  13.7 grams of  lost  fuel  whereas  overfill,
prefill,  and  postfill  on  average result  in  8.6, 5.9,  and 1.8
grams of fuel loss,  respectively.

     Adding these estimates  of  fires which can be prevented  by
onboard refueling controls (25  fires due to atomized vapors and
37-61  fires  due  to  spitback spills  during  refueling)   gives a
total nationwide estimate  of 62-86 annual service station fires
which  could  potentially  be  prevented with  onboard  refueling
control.  This  is  2.5-3.5 percent of the  projected nationwide
service  station  fires.  It   is  about 60  percent  of the  fires
associated with  the refueling process, which make up between 4
and 6  percent of the nationwide service station  fires,  as was
determined  earlier  in  the analysis.  Table 4-1  summarizes the
breakdown  of  service  station  fire  data   included  in  this
analysis.

      In  addition  to  estimating  the number  or   percentage  of
fires  that  can be  prevented with onboard  refueling  controls, a
monetary  benefit  was  placed  on  the  occurrences  of  property
damage, injuries and lost  lives  which would be  avoided if these
fires  were  altogether prevented.  As can  be  seen in Table 4-2,
annual  losses from  service station  fires  in the 49 non-Stage II
states  are  estimated  to  be between $50.0  million and  $76.2
million dollars.  This  amount includes the property damage  (as
presented  in the  fire   marshalls'  property  damage   reports)
caused  by the fires  and  also assumed  dollar  amounts  for each
injury  and fatality  that  occurred  ($7.5 million per  life  and
$100,000-$300,000   per   injury,   depending  on  the  severity).
Since   it  is  estimated  that  onboard  controls   could prevent
2.5-3.5 percent  of  the 49-state non-Stage II fires  then  as an
initial  estimate it  is reasonable  that  2.5 to  3.5 percent of
the   $50.0-$76.2  million   annual   dollar  losses   ($1.25-$2.67
million)  could  be  saved  with the implementation  of onboard
controls.

      It  should  be   restated that  these   conclusions   are  the
result  of  a preliminary  analysis  conducted by  EPA.   A more
detailed  analysis  is   being conducted  by  an EPA   contractor.
Additional  sources  will be  used in this analysis, which  should
result  in increased confidence  in the results.   As previously
mentioned,  the low  reporting rate of fires by service  stations
which  was  made  apparent  by   the   Minnesota  Service Station

-------
                              4-8
                           Table 4-1

          Analysis of NFIRS Service Station Fire Data*
        (Average Annual Data for 49 Non-Stage II States)


Total Fire Incidents                                     2466


Refueling Related Fires

Atomized Vapors/Liquids         25        (1% of total fires)
Fuel Spillage During Refueling  74-123    (3-5% of total  fires)

Total Refueling Related Fires   99-148    (4-6% of total  fires)

Expected Reduction in
Refueling Related Fires
with Onboard Controls

Atomized Vapors/Liguids            25 <§ 100% reduction = 25
Fuel Spillage During Refueling  74-123 @ 50% reduction = 37-61

Total Reduction in Refueling Related Fires               62-86
                                     (2.5-3.5% of total fires)
     From NFIRS Fire Data Base, Appendix III Tables 1-15.

-------
                              4-9
                           Table 4-2
               Annual Dollar Losses From Service
            Station Fires in 49 Non-Stage II States
    (Data taken from Appendix III, Tables 1 and 9 NFIRS Data)


                 Average Numbers Over 1982-1985
                       ,9-State           Estimated
                       (Projected)        Dollar Loss

Fires                    2466             $6,915,588
(incidents)                               (projected)

Fatalities                  4             $30,000,000*

Injuries                  131             ($13,100,000-
*                                         $39,300,000)**
                                   Total =$50.0-76.2 mil
*    Assumes $7.5 million per life.
**   Assumes $100,000-$300,000 per injury.

-------
                              4-10
Association Survey  could  imply  that the  NFIRS fire  data  base
underestimates the  actual  number of  fires  which occur.   Also,
the  NFIRS  fire data  base only  covers  fires at public  service
stations.  It does  not  include private  stations, such as  at  an
airport  and  it is  not  clear  whether it  includes  "convenience
store"   service   stations.    Accounting   for   these   possible
problems  in  a  more  thorough  analysis  should lead  to  more
accurate results.

     c.    Failure Analysis Associates'  (FaAA) Analysis

     As  was  mentioned  earlier,  FaAA  also  analyzed  service
station  safety  and  the  possible  effects  of   refueling  vapor
control  systems  on  service  station  fires.   They  did this  in
their  report  "Safety  Issues   in  Systems  Designed to  Recover
Gasoline Vapor  During  Motor  Vehicle Refueling,"   prepared  for
General Motors and dated February 5, 1988. [3]

     FaAA  used two different  approaches  to analyze  the fire
risk reduction achieved  by California's  Stage II vapor recovery
system.  The  first  method of analysis was simply to compare the
overall  service  station  gasoline  fires   (number   of  fires
reported  to  NFIRS  per  gallons  of  gasoline  dispensed)  in
California  versus  an   aggregate   total   of  18  non-Stage  II
states.   A  50  percent  lower  overall  fire  rate  was  found in
California.   FaAA  attributed  all  of  this  difference  to  the
added  safety of  Stage  II  systems, but  failed to make  a link
between  the  cause  and  effect  of this   lower  fire  rate  in
California.   There  is no apparent  reason why Stage II equipment
would  have  caused  a  lower rate of  non-refueling  related fires
in   California,  such  as  structure  fires  and vehicle  engine
fires.   However, FaAA presented the statistics without offering
any  technical basis to support their claims.  The  FaAA analysis
fails  to account for many  important factors other  than Stage II
controls  which  influence  the  differences  between   California
fire data and fire  data  from any other state.

     FaAA's  analysis  ignores the difference  in fuel  volatility
levels in  California  as  compared  to  the  other  states.   The
Center for  Auto Safety's  "Study and Comments  on  Environmental
Protection  Agency  Rulemakings  on  Gasoline  and  Alcohol  Blend
Volatility  and  Refueling  Emissions From  Gasoline   Vehicles,"
shows  high  probability  of  a  strong  link  between states with
high fuel  volatility and  an increased frequency of fuel  system
fires,  complaints,  overpressurization and  spitback.   California
has  a lower  volatility level (9.0  ASTM  RVP in the  summer and
less  than  ASTM  levels  the rest  of the year) than  the other
states analyzed.   Also,  based on  Department of Transportation
reports  for  the period studied,  California  uses a lower  amount
of  alcohol  blend fuel;  about  4  percent  for California versus  a
weighted average  of  about  7  percent   for  the  other states.
FaAA's failure to include the effect of  differing in-use state
volatility  levels  in their  analysis  introduces a considerable
uncertainty into  the  validity  of  their conclusions.

-------
                              4-11
     There  are  additional  differences between  California  and
other  states which  could  affect  FaAA's analysis  of  service
station fires.  First, it is  likely that  California has  better
inspection  and  enforcement  programs  than most  other  states.
This  is  a  reasonable  assumption  in view  of  the  fact  that
California commits a  high level  of resources  to air  pollution
control and automotive emission compliance programs.   Also,  the
fire codes  used in California are  different than those  used in
other states.  This lessens  (by some  unknown amount) the degree
of confidence that can be assumed in making a  direct comparison
of  California  fire  statistics to  those  of  other  states  and
throws further  uncertainty  on FaAA's claim that Stage II alone
is responsible for the lower fire rate in California.

     The  second  approach  used   by FaAA,  targeted  solely  at
refueling  related  fires,  was  to  categorize   fire  reports  by
NFIRS  fire  codes  to  determine  the percentage of  all gasoline
fires  which  could  be  classified  as  vehicle   refueling  fires.
FaAA reviewed fire reports for California  and  four  non-Stage II
states  and   found  that  vehicle   refueling   fires  represent
approximately 3 percent  of  all  gasoline  fires at gas stations
in  those  five  states.   They found  that California  had  a 55
percent less  frequent occurrence of  these types of  fires than
the  other  four states,  and attributed  the  reduction  to  the
Stage  II  system.   Further,   they  reasoned  that  onboard could
only  control   these  types  of   fires  and  could  be  no  more
effective than  Stage  II  in  doing so, therefore only  55 percent
of  three  percent  (1.65  percent) of  all  gasoline fires  at gas
stations  could  be  eliminated  by  onboard  vapor   recovery
systems.  FaAA  also argued that  the reduction in  overall fire
rates attributed to Stage II could  not be achieved by  onboard.

     Several  points  must be  made  in  response to the analysis
presented by FaAA.  First,  with  regard to  the  effect of Stage
II  and  onboard  controls   on refueling-related  fires,   it  is
reasonable  to  assume  that  either  method has  the  potential to
reduce these  types of fires.   However, FaAA made  an  analytical
error which  led to a substantial underestimate of the number of
refueling related  fires  in non-Stage  II  areas.  As is discussed
in  the paragraph  directly  above,  FaAA combined  the  refueling
related fire  reports  information for California and  four other
states  (Ohio, Texas,  Michigan,  and Illinois)  and concluded that
refueling related  fires  make-up about  3 percent  of  the total in
those  five  states.  They then argue that  since  the  California
refueling  fire  rate  is  only 55  percent  of that  in  the other
four  states,  and if onboard performed as  well as  Stage  II, the
most  reduction  one could expect  in total  service  station fires
with onboard  controls  is 1.65 percent.

     However,   FaAA  inappropriately  combined   fire  information
for  Stage  II  and  non-Stage  II  states  and then   used  this
information  to suggest  that  on  a  nationwide  basis,  refuelings

-------
                              4-12
account  for  3  percent  of   all  service  station  fires.   The
California  data  should  not  have   been  combined  with   the
non-Stage II state data,  due  to  the effect of Stage II  controls
on refueling  emissions.   Using data  in  FaAA's report  for  only
the  other  4   states  shows  that  5.7 percent,   rather than  3
percent,  of  all service  station fires  are  refueling-related.
Applying  FaAA's  55 percent  reduction efficiency yields a  3.1
percent  reduction  in overall  service station fires.    Both of
these  figures are  comparable to  the  EPA estimates  presented
above,  i.e., that  4 to  6 percent of  service  station fires  are
refueling  related  (versus 5.7 percent  for the  corrected FaAA)
and that onboard controls could cause a  60  percent reduction in
these  types  of fires (versus 55 percent for  FaAA).   Obviously,
the  percentage  reductions   in  overall  fires   is  now  also
comparable,  with  EPA  estimating   2.5-3.5  percent  and  FaAA
estimating 3.1 percent.

     Despite   the  cited   problems   with   FaAA's  analysis   it
directionally  supports  EPA's  premise  that controlling refueling
vapors  reduces the risk of service  station fires.  After making
the  above adjustments  FaAA's analysis  suggests a  3.1 percent
reduction  in  overall  service station fires which  falls within
EPA's  estimated range  of 2.5 to  3.5 percent.   This  level of
agreement  is  surprisingly   good   considering  the  number  of
judgments  which had  to be made when analyzing  the NFIRS data
base.

     d.    Potential  Benefits of Onboard Controls  on  Non-Fire
           Property   Damage   and  Health   Effects  at   Service
           Stations

     As was discussed  earlier,  the  proposed  onboard  refueling
control procedure should  essentially eliminate  the occurrence
of  gasoline  spitback  during  refueling.   In  addition  to  the
reduction in  service  station  fires which  would  be   realized
because  of  this,  there  are  also  non-fire property damage
benefits  and  health  benefits which  would  occur.  According to
the  1972  Scott  Research report mentioned above [2],   spitback
spills occur  in  about  13 percent  of all  refueling events and
result in  an  average  spill  of  about  14  grams  of  fuel.   The
range  on this value  varies  from zero to in excess of  40 grams
per  refueling.  Clearly,  some fraction of spitback spills cause
damage to  the shoes and/or  clothing of  the person dispensing
the  gasoline.   As Table  7  shows,  clothing  is  the   material
ignited in 0.08  percent of  service station  fires.  Of course
only  a  small percentage  of  the  clothing damaged  by  spilled
gasoline  actually  ignites,   so  spillage  on  clothing  is  more
frequent than  this   small  figure  might  lead one  to   believe.
Also,   some  health benefits  are gained by the elimination of
spitback spills.   Repeated   or  prolonged  dermal  contact   with
liquid  gasoline   due  to  spillage   can  cause   irritation  and
dermatitis for some  individuals.   Reducing  spillage  will  help

-------
                              4-13
to  address  this problem  and will  also help  to eliminate  the
need  to  use  refueling  mitts  or  gloves  purchased  by  some
individuals or provided gratis  at  some service stations.   It is
acknowledged that these benefits may exist, but  no  attempt will
be  made  to quantify  them at  this point.   These benefits  are
expected to be relatively small but,  as mentioned  previously,
further  contract work  is being  undertaken which  may help  to
quantify them.

B.   Potential Vehicle  Safety  Benefits  Due to  Onboard  System
     Design

     1.    Introduction

     The  implementation   of  integrated  refueling/evaporative
control  systems  such  as  those  developed  by  EPA,  API,  and
described  in  comments  provided by Ford,  Chrysler,  GM,  Nissan
and  others provides  the  opportunity for  safety  enhancements
over  current   fuel  and  evaporative  systems.   As  is discussed
below,   these  enhancements   lie   in  three  separate  areas:
1)  lower fuel  tank  pressures,  2)  control of non-FTP evaporative
emissions, and 3)  simplification and  improvements  over current
fuel/evaporative systems

     2.    Lower Fuel Tank Pressures

     Over the past  5  to 10 years, several factors have caused a
significant increase in the  fuel  tank operating pressures.  One
major  reason  for this is that the volatility of in-use gasoline
has  climbed  over  the  years.    Also,   alcohol  blended  with
gasoline  now  has  fairly  widespread  use,  and these  alcohol
blends  or  oxygenated fuels  have  a  higher volatility  than the
straight  gasoline  used as  blend stocks.   Additionally,  there
has  also been a  growing percentage  of  fuel  injected vehicles
with  high  fuel  system  pressures  and recirculation  of  heated
fuel to the fuel tank, which enhances fuel evaporation and thus
increases tank pressure.

     To  assist in passing NHTSA's  FMVSS 301  rollover test and
EPA's  evaporative  emission  standard,  fuel  tanks  have  been
designed  with  extremely  small  diameter  venting   orifices  to
contain  liguid/vapor  within   the   tank.    While  this  design
approach has  apparently been successful in assisting  compliance
efforts  to meet  these regulations,  these small  orifices have
decreased   the  venting  capability   of   the   tanks  and  thus
increased tank pressures.

     These  three   factors   taken  together,  higher  volatility
gasolines,  higher  tank temperatures  due  to fuel recirculation,
and limited tank venting have  all  acted to increase  in-use tank
pressures.

-------
                              4-14
     These increased tank operating pressures  can  contribute to
several different vehicle safety problems.   As is discussed in
detail in the  comments provided by  the Center for  Auto  Safety
these include fuel  spurting,  fuel  tank pverpressurization,  and
as  mentioned   by  API,   fuel  dispersion   in  tank   rupture
accidents.  Fuel spurting occurs when gasoline  spurts  or  sprays
from  the  fuel   tank  upon removal  of the  fuel cap.  This  fuel
release is often  due  to  the excess pressure in the fuel  tank.
Fuel  tank overpressurization  problems, caused when  pressures
occur in excess of those considered in  the tank design process,
can be  manifest  as  "fuel  leaks from the  tank and  vent  lines,
poor  vehicle performance,  excessive gasoline  odor," and  other
problems.[4]  Of  course, in  the  extreme,  excess  pressures  can
blow  vent  lines and gas tank seams.    Finally, fuel dispersion
could be  increased  when  fuel  tanks are ruptured or punctured in
accidents, since  the  fuel  in  the  tank  could initially be  at a
pressure significantly in excess of atmospheric.

     Manufacturers have  clearly been  alerted to the problems of
fuel tank overpressurization as there have been many complaints
raised to  them about this problem occurring with  their current
vehicles,  several resulting   in recalls and service bulletins.
Some  manufacturers  have  taken steps  to  remedy  the  problem.
Some  General Motor' s fuel caps  have a warning label  to  alert
the owner  of the potential for fuel  spurting upon cap removal.
Other  manufacturers  such  as  Chrysler  have  incorporated  an
anti-spitback valve  in their  vehicles fillnecks to  address this
problem.

      While  an  onboard  system  alone   would  not  necessarily
address   increased   fuel   volatility   or   higher   fuel   tank
temperatures,  it  would require an increase  in the  size  of  the
fuel  tank venting orifice  and thus provide  the  opportunity to
decrease  fuel  tank   pressures.    The   onboard  system  design
contemplated by EPA  (see  Chapter  3,  Section  C)  incorporates  a
larger fuel  tank  venting orifice.   The larger activated carbon
canister  associated  with  an  integrated  refueling/evaporative
system would have excess capacity  during a  large  portion of  the
time  between refuelings, and  so  could accommodate  much  of  the
evaporative  vapor  now  contained in  the  fuel  tank  by   the
limiting  orifice.  Thus  fuel tank pressures  could be reduced,
enhancing    safety,    while   still   controlling   evaporative
emissions.   Also,  EPA  is  preparing  revisions  to  the   test
procedure  which  will ensure  that  proper   provision   for   tank
venting  is  incorporated in  evaporative/onboard  control  system
designs.

      3.    Non-FTP  Emissions

      As was  explained  in the  previous  section, there has been  a
trend  toward   increasing  fuel   volatility   and   fuel  system
temperatures and  pressures over the past  several  years.   Since

-------
                              4-15
EPA's  evaporative  emissions test  procedure  does  not  evaluate
worst  case  conditions, there  are many  in-use conditions  that
could overload the evaporative canister.  The large  quantity of
vapors generated  under such  conditions are  called  Non-Federal
Test  Procedure  or  Non-FTP  emissions,  since  actual  in-use
conditions    (ambient   temperatures,    driving   patterns   or
volatilities)  exceed  those  prescribed  in  the  Federal  Test
Procedure.*   Recent  EPA  work  suggests that  these  uncaptured
emissions can be griite high in some  circumstances.   This  is one
reason  for  EPA's  preference  for   integrated  evaporative  and
refueling control  systems,  rather than  separate  systems.   The
larger  integrated  canister   has  excess capacity  most  of the
time,  which  should allow for  non-FTP emissions to  be captured
and  controlled.   Also, emitting  gasoline  vapors such  as these
near  potential ignition  sources  in  the engine  compartment or
near  the hot exhaust  system  has  been  identified  by several
groups as a potential safety hazard  (although API  found no fire
risk  due to  these  conditions[5]).    In any  case,  the  larger
venting  orifice and  carbon canister  on an  integrated onboard
system allows  for  these  excess  vapors   to  be  captured and
controlled.   Therefore  onboard  systems  may  offer   a  safety
advantage over current evaporative emission control systems.

      4.    Simplification  of  Current Fuel/Evaporative Control
           Systems

      Several  commenters have  said that  onboard systems will be
more   complex  than   existing   evaporative  systems   and  will
therefore   inherently  increase   safety  risks.   However,  the
integrated  onboard  system  design  presented  by  EPA to  the
manufacturers is  actually significantly less complex  than  some
of  the  current  evaporative  systems discussed  in Chapter 3,
Section  B.    For  example, there  are  fewer  external fuel/vapor
carrying  components,  such  as  the  external  (or  internal)  vent
tube   which  is   found   on  even   relatively   simple   current
evaporative  systems.  Decreasing  the  number  of components  also
means that  fewer connections such  as clamps  are necessary.  The
integrated  onboard system presented by EPA  also  has   fewer and
shorter  vapor lines  and  fewer fuel  tank  connections   than  many
current  evaporative systems.  Thus,  if one  accepts the  premise
suggested  by  some  commenters  that  more   complexity   increases
risk, then  the simplifications made possible with an  integrated
onboard  system should reduce  the  crash  and  non-crash  safety
risks  associated   with   evaporative   and   refueling   control
systems.   Along the  same lines, manufacturing and misassembly
mistakes and  the  resulting  effects should also  be  reduced.


 *     Information  presented  at  June  30,  1988  EPA Onboard  Test
      Procedure Workshop.

-------
                              4-16
C.   Summary and Conclusions

     In  conclusion,  the  analyses  presented  in  this  chapter
support EPA's view that  onboard  refueling  control  systems could
enhance service  station and automotive  safety.   First,  onboard
systems have the  potential to  reduce  the number  of  service
station  fires  and  the  non-fire  property  damage  and  health
hazards  at  service  stations.   Review  of  the  NFIRS  service
station fire data  base  shows that  4 to 6 percent of all service
station  fires  are  refueling-related.   It  is  expected  that
onboard refueling  controls  will  prevent  about  60 percent of the
refueling-related  fires,  or 2.5 to  3.5  percent of  all  service
station fires.   Such  a reduction  in  fires would  also  prevent
2.5 to  3.5  percent of the $50.0 to $76.2  million  annual losses
due  to property  damage,   injuries  and  deaths  resulting  from
service station fires  which is  a  savings  of  $1.25  to  $2.67
million.   Onboard  controls  could  also  prevent  the  damage  to
shoes  and  clothing  by  preventing  spitback  spillage  during
refueling.   By  controlling spitback  spills,  onboard  systems
could eliminate  the potential  health problems caused  by dermal
contact with gasoline.

     Furthermore,   the  implementation  of   onboard  refueling
controls  could  also enhance vehicle safety  in  both  crash and
non-crash  situations.   An  integrated  onboard  systems  design
like  the  one  described in  Chapter  3,  section  C,  could lower
fuel  tank pressure  levels  thereby  decreasing  the occurrence of
fuel  spurting,   and  fuel system leaks  and  ruptures  caused  by
overpressurization.   In  addition,   the  larger  canister  of  an
integrated  onboard  system  could   control non-FTP  evaporative
emissions.   This   addresses  the   perceived  safety  risks  of
gasoline  vapor  contacting a hot engine  or exhaust  system or  a
spark  ignition source.  Finally, the  integrated onboard design
described  by  EPA  is  less  complex than  many  of   the  current
evaporative  control systems discussed  in  Chapter 3, section B.
Since  many  commenters  have equated  increased  complexity with
increased  safety  risks  than the design simplifications allowed
by  EPA's  integrated  system  design could  also be  equated with
improved  safety.

     As  stated previously,  further  contract  work is being done
to  quantify both the fire and non-fire  safety risks at  service
stations  and the  potential  impact  of onboard systems on these
risks.  Contract work is also being undertaken  to evaluate the
comparative  risk  of an  integrated  onboard system  and  current
evaporative  control  systems.

-------
                              4-17
                    References for Chapter 4

     1.    "Survey Results  of Gasoline  Retail Outlets,"  Brain
Ettesvold, Minnesota  Service Station  Association, December  6,
1985.

     2.    "Investigation  of Passenger  Car  Refueling  Losses,
2nd   Year   Program,"   Scott   Research   Laboratories,   Inc.
SRL-2874-12-0972, September 1, 1972.

     3.    "Safety  Issues   in   Systems  Designed  to  Recover
Gasoline   Vapor   During   Motor   Vehicle   Refueling,"   Failure
Analysis Associates, February 5, 1988.

     4.    "Center  for  Auto  Safety   Study  and  Comments  on
Environmental  Protection  Agency  Rulemakings   on Gasoline  and
Alcohol Blend Volatility and Refueling  Emissions from Gasoline
Vehicles," March 24, 1988.

     5.    Comments   of   API  in   response  to  Onboard  NPRM,
February   11,  1988,  available  in  public  docket  A-87-11  at
IV-D-358 plus sub-entries.

-------
                           Chapter 5

                   Summary and Net Assessment

A.   Introduction

     As prescribed  in Section 202(a)(6)  of  the Clean  Air  Act,
safety  is  one  of  the  key  factors  to  be evaluated in  the
determination of whether  onboard vapor  recovery  systems  should
be implemented,  and  indeed  the potential safety implications of
this technology  have  been a key issue in the  rulemaking.   EPA
began  its  in-depth analysis of  the  safety  issue  in the  Spring
of 1986, more  than a year before the  NPRM,  and since that  time
has been engaged in  an ongoing evaluation of all aspects  of the
issue.

     Even  prior to  the  NPRM,  EPA  completed  a  comprehensive
study  of  the onboard  safety  issue.   This study,  the June  1987
onboard  safety  report  shown   in Appendix  II,  discussed  the
design  of  safe onboard  systems  and  identified  and evaluated
both general and specific onboard safety  concerns  raised prior
to  the NPRM.    In-use  concerns  addressed by  the  1987  report
included both  crash and  non-crash situations.   In  addition to
the  general   area  of  crashworthiness,  the  study   addressed
non-crash  concerns such  as  tampering, defects,  misrepair,  and
refueling   operation   safety.    The    study   concluded   that
straightforward,   reliable,   relatively   inexpensive   solutions
exist  for  each  of  the potential problems identified,  and  that
no  increase in  risk  need occur  or  be accepted because  of the
presence  of an onboard  system.  The  study  further  concluded
that onboard-equipped vehicles could in  fact provide a level of
in-use  fuel  system  integrity  equal  to  or  better  than  that
achieved on present  vehicles,  and that  the  changes  which would
accompany   onboard  controls   could  improve  safety  on  in-use
vehicles.   Few of the  comments  received  on the  NPRM directly
addressed  technical  aspects  of EPA's safety  report.   The most
substantive comment  amounted  to  a  suggestion  that additional
analysis was needed to support EPA's conclusions.

     However,  while  EPA received few  comments  on  the  safety
report  itself,  many concerns   about  the  safety   issue  were
expressed  in the  comments.   Although  a few commenters, notably
vehicle  manufacturers,   expressed  some  specific   concerns  in
areas   such   as  potential   onboard   system   hardware,  many
commenters stated  that  one  issue in particular required  further
consideration  by  EPA:   the effect  of  increased complexity on
safety.    In   general,   manufacturers   did  not  disagree   that
solutions  could be  developed  for identifiable,  or  predictable,
types  of  problems such as those discussed in EPA's 1987 safety
report.   Rather, the  main  contention  centered on the  inability
to  foresee and  avoid  previously  unidentified  problems   that
could potentially accompany the  implementation of  a new  or more

-------
                              5-2
complex system.  Of course this  concern is based on the premise
that  onboard  vapor   recovery  systems   would   increase   the
complexity of  current  fuel systems through the addition and/or
enlargement  of  components,  connections,  etc.   Because  these
modifications  involve  the fuel  system and allegedly  introduce
the  potential  for  problems  that   cannot   be  predetermined,
manufacturers   claimed    onboard    systems    would    add    an
"unquantif iable"   increase  in  the   risk  of   both  crash   and
non-crash  fires.    While  this  risk  could not be  quantified,
manufacturers characterized it as unacceptable.

     EPA  disagrees with the  manufacturers'   concern  over  the
effect  of  increased  complexity  on  risk  for  four  reasons.
First,  even  a casual  review  of  the  increases in  fuel  system
complexity over  the past decade or  so (such  as  the  increased
complexity resulting  from the movement toward fuel-injection)
made  EPA question the  validity of  an argument  which asserts
increased    complexity    is    unacceptable    from   a   safety
perspective.    Second,  a review of  evaporative  control  system
configurations revealed  a trend toward increased  complexity in
recent  years.   Third,  EPA has  always maintained  that onboard
systems  need  only be  a  relatively  simple   extension of  the
current  evaporative  system,   and  as  such  much  of  the  added
complexity displayed in  manufacturers'  proposed designs appears
unnecessary.     Fourth,    most   manufacturers   overlooked   the
potential for  added safety benefits of onboard  controls.

     Consequently,  in  order  to  address manufacturers' concerns
regarding  complexity  and  risk,  the  analysis  in  this document
studied  onboard  safety  from  three  perspectives.   First,  since
onboard   systems   are,   in   several   respects,   analogous  to
evaporative  systems which have  been  used  for  18 years,  the
performance  and  design  of evaporative systems was assessed to:
1)  put  the onboard complexity/risk issue  into perspective, and
2)  evaluate  the  relationship between  complexity  and safety.
Second,  EPA   investigated  the  feasibility of  an  improved and
simplified   onboard refueling  control  approach  to   determine
whether  the  added  complexity  suggested by  the  manufacturers was
even  necessary.    Third,   the  potential   safety  benefits  of
onboard  systems  were characterized  to determine whether onboard
controls  can actually improve  safety.   EPA's  results  regarding
complexity/risk,  simple  onboard designs,  and  safety benefits
are summarized below.

B.   Defining  the  Relationship Between  Complexity  and  Risk

     A   major  aspect   of   assessing  any  added  safety   risk
resulting  from the implementation of onboard  controls  involves
defining the  relationship  between  complexity  and  risk.   In
order  to adequately characterize  this risk,   EPA  established  a
baseline  to  gauge  the  potential  safety   effects   that  are
possible when vapor recovery  devices such as onboard  controls
are implemented.   Because of  the inherent  similarity between

-------
                              5-3
onboard  and   evaporative   control  systems,*   an   appropriate
baseline to  put the  complexity/risk of  onboard controls  into
the proper  perspective is  the  safety performance  exhibited  by
evaporative  control   systems.    In  order   to  develop   this
baseline, the  overall  impact of evaporative  control  systems  on
safety  was  assessed  first,  and  then  the varying degrees  of
evaporative  control  system complexity  were  evaluated  in  the
context of the relationship with safety.

     1.    Evaporative System Safety

     The investigation  into the safety  of evaporative  control
systems  involved an examination of evaporative control systems'
effect  on  historical  fire  rates  and a  thorough review of  the
in-use   safety  performance  of  evaporative   systems.   In-use
performance was gauged according to  the  number of problems (and
seriousness  of  consequences)   resulting  in  safety  recalls,
technical service bulletins, and owner complaints.

     An  analysis  of  fire   rates  for post-  and pre-evaporative
model  year  accident  involved  vehicles  involving   fatalities
revealed that the  implementation  of evaporative control systems
had  no  discernible effect  on  crash  fire  rates.    Fire  rates
fluctuate to  a  small  degree from model year  to model year,  but
overall have remained rare  events.

     In  addition,   a  thorough review of  NHTSA's computer  files
of  safety  recalls,   technical  service  bulletins   and  owner
complaints   showed  that   evaporative   control   systems   have
operated   with  very  few   problems  since   they   were  first
implemented  eighteen  years  ago.    In total,  only  20  recalls
(less  than  0.5 percent of  the  more  than  4200 safety recalls  in
the  past 20  years)  involving  415,000 vehicles  (less than 0.3
percent  of the  130.8  million vehicles  recalled  since  1966),
between  70  and  120  technical  service  bulletins  (about  0.1
percent   of   the   88,000   service   bulletins   issued),   and
approximately  100  owner complaints  (which represents less than
0.05   percent  of   the  210,000  owner  complaints   in  NHTSA's
computer files), have  involved  the  evaporative control system.

     Therefore,  while  evaporative  control  systems  have been
integrated  into  the fuel systems of over 200 million vehicles,
only  a  minute  fraction may have  experienced safety problems.
Further,  of  the  rare  problems  that  did   occur,   no  serious
consequences  such   as  deaths or serious  injuries appear to have
been  reported.  The  basic  conclusion that  can be  drawn from
this  information  is  that  evaporative control systems have not
compromised fuel system  safety.
      Both systems  operate on  the same  basic  principles  using
      similar  components  performing similar  functions  (such  as
      vapor  routing and vapor  storage).

-------
                              5-4
     It  is  also  worth  noting  that  the  same  concerns  over
complexity  that  are  being  raised  in  connection with  onboard
controls  were  also  raised  15  years  ago  with   regard  to
evaporative  control  systems.   An  August   20,   1973  Federal
Register Notice  (38  FR 22417)  regarding  the implementation  of
FMVSS 301  explains NHTSA's  concern  that "Regulations of the EPA
for fuel evaporative  emission control  will  increase  the number
of components  related to fuel systems  in  all new vehicles with
a GVWR of 6,000 pounds or less, and  a  corresponding  increase in
points  of   potential  system  failure  that could  result  in the
loss  of   fuel   in   crashes."   However,   the   concerns  about
increased  risk to  fuel  systems  as  a result  of  implementing
evaporative  systems nearly  20 years ago has yet  to  materialize
as a  problem creating any  serious  consequences.   Thus,  even if
added  complexity  is  a  concern,  evaporative  control  system
experience   demonstrates   that   added   complexity    does   not
necessarily  result in any significant actual increase of risk.

     While  examining  the  overall   impact  evaporative  systems
have  had on safety,  EPA  found that levels  of  complexity vary
considerably from one vehicle model to another.   In fact, many
features  proposed for  onboard systems that were characterized
by  manufacturers as  an  increased  safety  risk  were  found  to
already   exist  in-use  on   numerous   fuel/evaporative   system
designs.   The existence  and variety  of   complexity  on current
vapor  recovery  systems  led  EPA  to   evaluate  whether  or not
complexity  has affected  current fuel  system  safety,  as many
commenters  indicated  it  invariably would.

      2.     Evaporative System Complexity/Safety

      A   review  of   manufacturers'   shop  manuals   and   other
pertinenjt    literature    revealed   that    evaporative    system
complexity has generally  increased  over  time,  but that even in
today's  systems the  complexity varies  widely.  Some  systems  are
still  fairly  simple  (few   components/connections,  etc.)   while
others  are  relatively  more  complex  and   involve  more  external
components,  multiple vapor  lines,   extra  connections and  other
design  features.  Yet despite this  increase in complexity  there
is  no  evidence  that any  one design or design  approach is  any
more  or less safe than  any  other.   EPA presumes that the  range
in  evaporative  system  complexity  exists  for  good cause,  and
that  this  complexity was  incorporated  without  compromising  the
safety  of the systems  involved.   In fact, while complexity  has
increased   over   time,   safety recalls  and service  bulletins
regarding    evaporative    systems    have   actually    decreased.
Apparently increased  complexity has  not affected  safety.

      EPA's  review  also  found  that  some  current   evaporative
control systems  designs incorporate  features  such  as  fillneck
mounted valves,  liquid/vapor  separators,  numerous vapor  lines,
and  plastic components.   When considered for use on an onboard
system,    all   of   these   features   were   characterized  by

-------
                              5-5
manufacturers as  posing  an unacceptable  risk  because of  their
increased  complexity,  and  yet,   several   manufacturers   have
incorporated   these   design   features   into   their   current
fuel/evaporative   systems,   apparently   without   compromising
safety.

     In summary, EPA has reached the  following conclusions from
its   investigation  of   evaporative  control  system   design
complexity and  safety.  First, evaporative  control  systems have
performed   safely.    Only  a   few   infrequent  problems   have
occurred, and none of the rare problems that did occur resulted
in  any  serious consequences.    Second,   EPA did not find  any
inherent connection  between increased complexity and decreased
safety.  Despite the  increased  complexity  of fuel/evaporative
systems over time and the wide variety of complexity in current
systems,  there  is  no  evidence  to suggest  a  direct,  adverse
relationship  between  safety  and  complexity  for  evaporative
systems.  Third,  given the  similarity between onboard refueling
and  evaporative  emission  control  technology,  the  experience
with  evaporative systems suggests  that onboard controls can be
implemented  safely also.   In  fact,  the   evidence suggests that
even  a wide  range of  complexity  in onboard  systems would not
impact safety.

     Evaporative  systems were  added to vehicle  fuel  systems 18
years  ago  and  apparently have  caused  no  significant  safety
problems.   This safety record was  accomplished even  though an
entire  new  system  was  added  to  the   vehicle.   Onboard  and
evaporative  system  technology are  similar  and in  many ways
onboard  is  more  an  extension or  modification of  the current
evaporative  system rather than  an  entirely new  system.  Given
this  view  of the technology involved, and  the directly  relevant
experience   gained   in implementing  safe   evaporative  control
systems,  EPA   still   believes  that  manufacturers   can,  given
reasonable  leadtime,  readily  implement onboard control systems
with  similar levels  of safety.

      Even  with the  proven safety performance  of  evaporative
control  systems   and  the  established   lack   of   a   record  of
complexity's   effect  on   safety,   EPA   recognizes   that  some
commenters   will  maintain  that   increased  complexity  would
inherently  result from onboard controls  and that this increased
complexity   would degrade  safety.   However,   EPA  has  always
maintained  that onboard controls  would   only  require  relatively
simple modifications  to the  current  fuel/evaporative system.
Consequently,  EPA initiated a development program for  a  simple
onboard  refueling  control design  which  incorporated  features
that   could even improve  safety.   The  next  section briefly
summarizes  the  results of this initiative.

-------
                              5-6
C.   EPA's Onboard System Design

     In  response   to   EPA's  proposal   of   onboard  controls,
manufacturers indicated their  preference  for designs  much  more
complex than those  anticipated by the Agency.  Along with their
suggested  designs,  manufacturers  postulated possible  problems
that  could  accompany  the   implementation   of  their  complex
designs.   Possible  solutions  to these  problems  were  rarely
discussed.   Many of  the problems postulated  by  manufacturers
concerned  added  components and complexity which did  not  appear
to be  necessary to control  refueling  emissions.   Since EPA has
always   anticipated   that    onboard   control   would   require
relatively simple designs, a development  program was undertaken
to design  a simple  onboard  system to alleviate  concerns  over
added complexity.

     In  order  to  develop  a  simplified onboard system,  EPA
established   the  following   two  basic  design   constraints.
First,    the  simple  system   should   incorporate  the  fewest
features   possible   and  should   yet   perform   all   necessary
functions.   Second,  components used in the simple system should
be based on current production hardware.  Following  these two
constraints,  EPA  modified  a typical  current  (and   relatively
simple)  fuel/evaporative  system  (Figures  3-20   and  3-21)  to
control  refueling  as well as evaporative emissions.   The system
developed  by EPA is shown in Figures 3-22 and 3-23.

     Only  a few modifications  were   necessary  to  convert the
evaporative  system  in  Figure  3-20   and 3-21  to  the onboard
system  in  Figure 3-22 and 3-23.   First, the  orifice size of the
float/rollover  valve was moderately  increased by replacing the
stock valve  with a modified version of another valve which had
been  designed for higher vapor  flow  rates.   The valve was also
modified and relocated to a  location  where  it  could serve as a
fill  limiter.    Second,  the   current  refueling vapor  vent line
was  removed since  the  vapors  were  now  to be  routed  to the
canister.     Third,    the    carbon    canister   was   enlarged
approximately  two  fold  over  its  current  size.    Fourth,  the
canister  was moved  to the  rear  of the  vehicle,  which greatly
shortened  the length of the  vent line  from the fuel tank to the
canister.   In addition,  it  should be noted   that  no  significant
modification to  the  current system  was necessary  to  form a
fillneck  seal.    The  current fillpipe  is  inherently sized to
form  a  liquid seal  and  required only minor optimizing  to  insure
effective  performance  at low fuel  dispensing rates.   Lastly,
the  particular  fuel  tank   system used  by  EPA  (from  General
Motors   "A"  body  vehicle line)  did  not actually  include the
anti-spitback  (anti-fuel  spurting) valve  shown in Figure  3-24.
Therefore,  for  this  system,  an anti-spitback valve was also
added.   The valve  chosen was a  production  Chrysler  unit  which
has been in  use since  1984.

-------
                              5-7
     The result of these modifications  was a refueling  control
system  which  was  an overall  simplification  of  the  onboard
design provided in the NPRM.   It was indeed very  similar  to  the
stock evaporative system in  terms  of  design,  configuration,  and
function.  When tested with  nominal  9.0 RVP Federal  test fuel,
this system consistently met  the proposed refueling standard by
a  substantial  margin.   In  addition,   the  added  anti-spitback
valve completely eliminated fuel spitback from the stock system.

     In  summary,  EPA  feels  that  much  of the  added complexity
suggested by manufacturers  in their proposed designs (Appendix
I)   is  not   necessary   to   successfully  control  refueling
emissions.    Onboard   systems   can  be   simple   extensions   or
modifications   of   present   evaporative  systems.    Further,
modifications  that  are  necessary can  even  simplify  certain
aspects  of the current design.   With the proper design, no risk
need  be  added,   and  in  fact,  refueling  controls  can  offer
several  safety benefits.

D.   Safety Benefits of Onboard Controls

     Onboard refueling controls offer  the potential to  enhance
safety   through  two  general   mechanisms:    improved  design
features and  reduced service  station  risks.    The  extent  to
which  benefits available through  improved design  features  are
realized will  ultimately  depend  on design choices  selected by
manufacturers.    Improved   service  station  safety,  however,
inherently   accompanies   refueling   controls   irrespective  of
design  selection.

     1.    Design Improvements

     Safety  enhancements from design improvements  could lie in
three   areas:   1)  lower  fuel  tank  pressures,  2)  control  of
non-FTP evaporative  emissions,  and 3) simplification of  current
fuel/evaporative   design  features.    These  three  areas  are
discussed below.

     First,  onboard  systems  provide the  opportunity  to  increase
the size  of  the  fuel tank  venting orifice  and thus  cause a
decrease in  fuel  tank pressures.   Fuel tank operating pressures
have  increased  significantly  in  the  past 5-10  years  due to
increased  fuel volatility,  higher  tank temperatures (e.g., due
to fuel  injection),  and  limited  tank  venting  (to  assist in
compliance  efforts  to pass NHTSA's FMVSS 301  rollover test and
reduce vapor  generation).   Increased  tank  operating pressures
have contributed  to  safety problems such as fuel spurting,  fuel
leaks,  and increased fuel  dispersion in  the event  of a  ruptured
tank   in an  accident.   The proposed   onboard  refueling   test
procedures  would  require  improved tank  venting.   This  should
lead to lower  operating  pressures  and enhanced  safety.

-------
                              5-8
     Second, integrated onboard/evaporative  systems  likely will
result in added control of certain  in-use  (non-FTP)  evaporative
emissions.  Current  evaporative  control systems do  not  contain
adequate capacity to  contain  vapors generated  under  conditions
that  exceed those   prescribed  in  the Federal Test  Procedure
(FTP).  FTP test  conditions were  designed to  represent  average
in-use  conditions.    Consequently,   it  is  not uncommon for  a
substantial   amount   of   excess    evaporative   emissions   to
"breakthrough"  the  evaporative  control system  under  worse than
average  in-use conditions.   These  vapors  are  emitted  either
through the evaporative canister  or some  pressure relief device
such  as the fuel  tank cap.  Emitting vapors such  as  these near
potential  ignition  sources in  the engine compartment has been
identified  by  several  groups  as  a potential  safety  hazard.
Integrated  onboard/evaporative   systems   would  control  these
excess vapors and provide  a safety benefit over current systems.

      Third,  onboard  systems  (such   as    the   simple   design
developed  by  EPA) can reduce the  complexity of some aspects of
the   current  fuel  tank  and  evaporative  system  design.   For
example,  an onboard system can eliminate  the external vent line
that  is  currently used to vent refueling emissions.   Also,  an
onboard  system utilizing  a rear-mounted  canister will  shorten
the vent  line from the fuel tank  to the canister.   Also, moving
the  canister  to  the rear of  the vehicle may  have  same safety
advantages  with  regard  to vapor  release due  to breakthrough,
tampering,  or  defects  relative  to under-the-hood  locations.
Thus,  if  one  accepts  the premise  that  increased  components,
connections,   etc.   increases  risk,  an  onboard  system  that
simplifies  features  of  the current fuel/evaporative system will
improve safety.

      2.     Service  Station Safety

      In  addition to  design  improvements,  onboard systems  would
have  a positive  effect  on the safety of automotive refueling  at
service  stations.   Refueling  vapors that are currently  vented
to an  area which  poses  something  of a safety hazard  would
instead  be routed away from potential  external ignition sources
to a  safer location (the  charcoal canister).  Also,  due to test
procedure requirements,  onboard  controls  are  likely  to  bring
about a  decrease  in the amount  of  gasoline  spilled  during
normal vehicle   refueling.    Therefore,   onboard  controls  are
likely to reduce  the number of  fires  that  result from  ignited
refueling vapors  or  fuel spills  and  improve service  station
safety.   EPA estimates  that onboard controls have the potential
to prevent between  63-77  service station  fires annually.

E.   Net  Safety Impact

      Overall,  EPA  still   believes  that onboard control  systems
will  have  no  negative effect upon vehicle  safety and  actually
provide  the  opportunity   to  improve safety in several  areas.

-------
                              5-9
The  added  complexity of  evaporative  control systems  was  found
not  to  affect  vehicle safety.   EPA  feels onboard  controls  can
and  will be  implemented  with the same or  a  better  safety level
as current  systems.   Further, because of the potential  design
improvements and service  station benefits,  EPA believes onboard
control  systems  will  have  the  potential  for   an  overall
beneficial impact on safety.

     Of  course,  EPA  expects to  receive  additional  comment  on
these safety issues, both as part of  its  consultations with DOT
and  from manufacturers  and  others during the comment period on
its  reproposal.   EPA will  consider  and  address   all  of  the
relevant safety-related  issues  in its final analysis, and will
continue to  consult  with  DOT before making  a  final  decision on
whether to require onboard controls.

-------
        Appendix I





  Onboard System Designs





Submitted by Manufacturers

-------
Fig,  tt      Commenter

  1       Ford


  2       Ford


  3       Chrysler


  4       Chrysler


  5       General Motors


  6       General Motors


  7       Generals Motors


  8       General Motors


  9       Volkswagen


  10      Peugeot


  11      Saab


  12      Nissan


   13      Nissan


   14      Nissan


   15      Subaru


   16      Toyota
 Citation

A-87-11



A-87-11



IV-D-39


IV-D-366(a)



IV-D-360



IV-D-360


IV-D-360


IV-D-360


IV-D-361


IV-D-340


IV-D-368


IV-D-08


IV-D-08


 IV-D-08


 IV-D-364


 IV-D-363
    Description
Separate Systems,
Mechanical Seal

Separate Systems,
Liquid Seal

Integrated System,
Mechanical Seal

Separate Systems,
Mechanical Seal

Integrated System,
Mechanical Seal

Integrated Systems,
Liquid Seal

Integrated System,
Liquid  Seal

Integrated System,
Sleeve Seal

Integrated System,
Mechanical Seal

Integrated System,
Mechanical Seal

Integrated System,
Mechanical Seal

Integrated System,
Mechanical Seal

Integrated System,
Mechanical Seal

Separate  Systems,
Liquid Seal

Separate  Systems,
Mechanical  Seal

 Separate  Systems,
 Liquid Seal

-------
Fig, ft      Commenter

  17      Toyota


  18      Toyota


  19      Mitsubishi


  20      Mitsubishi
 Citation

IV-D-363


IV-D-363


IV-D-377



IV-D-377
    Description
  21      American Petro-
           leum Institute
           (Exxon Design)     IV-D-358e
  22      American Petro-
           leum Institute
           (Mobil Design)     IV-D-358e
   23      Mobil               IV-D-329
   24      Multinational       IV-D-01
           Business Systems

   25      Multinational       IV-D-01
           Business Systems

   26      Multinational       IV-D-01
           Business Systems
Separate Systems,
Liquid Seal

Integrated System,
Liquid Seal

Integrated System,
Mechanical Seal

Integrated System,
Mechanical Seal
               Integrated System,
               Liquid Seal
               Integrated System,
               Liquid Seal

               Integrated System,
               Liquid Seal

               Integrated System,
               Mechanical Seal

               Separate Systems,
               Liquid Seal

               Integrated System,
               Mechanical Seal
 0377X

-------
                                Appendix I
                               Figure 1 - Ford
              MECHANICAL  DESIGN
                 EMERGENCY RELIEF
                       VALVE
                  VAPOR/LIQUID
                 DISCRIMINATOR
     TO CANISTERS
                                      CAS NOZZLE
               VAPOR
   VIEW IN DIRECTION
     OF ARROW Y
                                   NOZZLE SEAL
                               TRAP DOOR
                               W/SEAL
 VENT TUBE
314* • 5/8' I.D.
                                                           VAPOR/LIQUID
                                                           DISCRIMINATOR
                   VIEW IN DIRECTION OF
                       ARROW Z
 CARRYOVER
EVAP. SYSTEM
                                                     MODULAR CARBON
                                                       CANISTERS
                                            PURGE HEATER
                                  VAPOR TUBE

                                 FIGURE III-5


-------
                                 Appendix I
                                Figure 2 - Ford
               CONCEPT LIQUID  SEAL SYSTEM
 TO ENGINE INTAKE
FUEL TANK VENT A
 ROLLOVER VALVE
                                              OBVR CONTROL VALVE
                                              • LIQUID VAPOR SEPARATOR
                                              • REFUELING ACTIVATION VALVE
                                              • ROLLOVER PROTECTION
                                              • FILL HEIGHT CONTROL-OVERFILL PREVENTION
CARRYOVER
CANNISTER
      ORVR PURGE LINE
      TO ENGINE INTAKE
                                                          REFUELING ACTIVATION
                                                                SIGNAL
                                                                               SEALINC
                                                                               FLAPPEI
                                                                                DOOR
                                          ANTI SURGE
                                          CHECK VALVE
                                                                        ORVR CANNISTER
                                                                           MODULE
                              PURGE AIR HEATER -^
                                    FIGURE III-0

-------
                                                        Appendix I
                                                    Figure 3 - Chrysler
                     Detail  of Onboard
                      Fill  Pipe Inlet
•.CIILMATIC COMPARISON OF F.VAPOKATIVE  (LIGHT)  AND ONBOARD REFUELING (DARK + LIGHT)  CONTROL SYSTEMS

-------
                                       Appendix I
                                   Figure 4 - Chrysler
                      EVAPORATIVE/REFUELING
                           VAPOR RECOVERY
                                  SYSTEM
         MANIFOLD PRESSURE
              VEHICLE SPEED
       COOLANT TEMPERATURE
       , RPM
       •TIME
       1 DISTANCE
       • THROTTLE POSITION
SYSTEM COMPONENTS
i  CiJELCA?
2  FUEL F'L.ER TUBE- -TCP VENTED
3  FUEL TANK VENT - INTERNAL PASS. CAP
*  FUEL "ANK VENT - EXTERNAL TRUCK
5  DlFFcRENTlAL PRESSURE CHECK VAL\ E
6  ?UE:. TANK
7  REFUELING VAPOR CANISTER
3  FIL'SR
3  P'JP.GEAIRINLE" -PEVO-E
REFUELING FILLER TUBE FEATURES:
- NOZZLE SEALING DEVICE
- REFUELING EVENT OEViCE
- PRESSURE REUEF OEViCEtS)
10 VAPCP LINE - 5/8-! D
M "ANK VENT.'RCLLOVER VALVE
12 PURGE USE -s/ie-
•,3 PURGE L!NE -•;*'
u PURGE SOLSNCID - DUTY CYCLE CONTROLLED
15 ELECTRONIC CONTROLLING
i s EVAPORATIVE VAPOR CANISTER
i? TC BI-LEVEL PURGE
18 MANIFOLD VACUUM LINE
19 VACUUM SOLENOID - ON/OFF
 - NOZZ'.E SHUT-OFF (LIQUID)
 - VAPOR CONNECTiQN TO CANISTER
 - LIQUID CARRYOVER PREVENTION
                                      Figure 1

-------
                                                 Appendix I
                                           Figure 5 - General Motors
                          REFUELING EMISSION CONTROL
                           SEALED FILLER NECK SYSTEM
X
I
^1
VO
                                                         LIQUID/VAPOR
                                                          SEPARATOR
  TANK VENT
VALVE ASSEMBLY •
   (TWA)
                         LARGE
                        VAPOR LINE —
                                                                                  CIRCLE 'A'
                                                       LIQUID/VAPOR
                                                        SEPARATOR
                                                                              SHUTOFF
                                                                               VALVE
                                               FIGURE xv-15
                                                                                            MECHANICAL
                                                                                               SEAL
                                                                                VIEW IN CIRCLE 'A'

-------
                                           Appendix I
                                     Figure 6 - General Motors
                          REFUELING EMISSION CONTROL
                                LIQUID SEAL SYSTEM
  END OF FILL
 SHUT-OFF VALVE
                                              CIRCLE 'B1-
  TANK VENT
VALVE ASSEMBLY
x
§
    LARGE
   CANISTER
                                                                        VIEW IN CIRCLE 'A1
                                             FIGURE xv-16

-------
                                              Appendix I
                                        Figure 7 - General Motors
o>
     ANTI-EXPULSION
        VALVE
         RESERVOIR

       VIEW IN CIRCLE 'B'
 LARGE
CANISTER
                    REFUELING EMISSION CONTROL
                          LIQUID  SEAL SYSTEM
                            (PROPOSED  CONFIGURATION)
                                                                   CIRCLE 'B1
                           LARGE
                         VAPOR LINE
                                                      LIQUID/VAPOR
                                                       SEPARATOR
                                              FIGURE xv-17

-------
                                               Appendix I
                                         Figure 8 - General Motors
                             REFUELING EMISSION CONTROL
                                    SLEEVE SEAL SYSTEM
00
10
             LARGE
             CANISTER
                                                    TANK PRESSURE -
                                                     RELIEF VALVE
    TANK VENT
  VALVE ASSEMBLY
                             LARGE
                           VAPOR LINE
                                                                            -FILLER NECK
                                                                            MODIFICATIONS
                                                                                         TO
                                                                                         CANISTER
                                                                                         FILLER NECK
                                              FIGURE XV-18
SLEEVE SEAL

-------
                    Appendix I
                Figure 9 - Volkswagen
Onboard Refueling Control System A
Pressure
•relief valve
                    = Ignition key

                 Overflow valve __ 5
                     Charcoal-
                     canister
                                         Engine
 ) VOLKSWAGEN
                                         Rasearcfl

-------
                                              Appendix I
                                           Figure 10 - Peugeot


                    SCHEMATIC  OF  IMHNCIPLE  : ON-BOARD  SYSTEM

                    (SCHEMA OE PRINCIPE DE RECUPERATION  DES VAPEURS AU REUPLISSAGE)
                                                                                 VAPOR LIQUID SEPARATOR
                                CANISTER
VENTING  VALVE

    mi*e  a I air Ithrc)
       MCCANICAL SEAL

       (Efanchfi/i*/
MAX FUEL LEVEL
                                        ROLL-OVER VALVE
              OVER FILL  CONTROL
                {Ctat>*t' limifarion Je. blnn)

-------
                           Appendix I
                       Figure 11 - Saab

          PROTOTYPE ONBOARD REFUELING VAPOR CONTROL
                            SYSTEM
                           Saab  9000
1.  FUEL TANK
2.  INLET MANIFOLD
3.  THROTTLE
4.  CANISTER
5.  VENT CUT VALVE (VCV)
6.  OVERFILL LIMITER PIPE
7.  VENT LINE VCV - CANISTER
 8.  PURGE LINE
 9.  VENT LINE ROLLOVER VALVE -  CANISTER
10.  PURGE CONTROL VALVE
11.  REFUELING SEAl
12.  FILLER NECK
13.  ROLLOVER VALVE VENT
14.  PURGE CONTROL UNIT

-------
                                    Appendix I
                                 Figure  12 - Nissan
          Intake Manifold
                                          Control  System
                                                                 Vent  Cut  Valve
                            I rz. 'Titrfn^ •••••* •»-«•»•-»••"•- >-.—-*>•
Refueling  Canister
                                                                                       Seal
                                                                              Filler  pipe
                                        Fuel Tank
  K I  G.  1-1	BASIC   O N It O A K I)
                  V A I* O K   C 0 N T KOI.   SYS T K M

-------
                                      Appendix I
                                  Figure 13 - Nissan
      Water Separator
                                         Control Unit
                          Variable; Orltlcc Control
                          Rollover Valve  (with Fuel Check Valve)
                                                                   Vent Cut Valve
                                                 Vapor/Liquid Separator
Refueling Canister
                                                                                        Mechanical
                                                                                        Nozzle Seal
                                                                    Over Fill Limlter
K  I G .  a ~ 1    S Y S T K M   A   ( S K N T K A )

-------
        Appendix I
    Figure 14 - Nissan
                                ir  Valve  (wltU  Fuel Check Valve)
Viirlable Orifice Control
                                  Vapor/Liquid Separator
                                            Over  Fill Linlter
                 Canister

Kefuellng Canister



        !•• i  c .  :\    '/.   s Y r, r i-:
                                                             Seal
                  ( x o o '/_xj.

-------
                                              Appendix I
                                          Figure 15 - Subaru
                                           2	Refueling Egiisslgii^ Control Sys t
                                          Solenoid Valve
                                          (Pu.-ge Control for
                                           Refueling Vapor)
                                                                        Water Separa
                                                              Solenoid Valve —
                                                                                      Refueling Emission C
<               i
  Intake Manifold
                                            Electronic
                                            Control Unit
-Solenoid Valve
 (Purge Control
  for Evaporative Vapor)
Vapor/I iquid
   Separator

                                I  Check Valve-
                            Fvaporative
                            Emission Canister
                                                                                                       Mechanical
                                                                                                       Noizle Seal
                                                                                            Overfill  l.iiniter
                                                                    Fuel  Tank

-------
                       Appendix I
                     Figure 16 - Toyota
 ONBOARD VAPOR  RECOVERY SYSTEM
        PURGE CONTROL VALVE      REFUELING  CANISTER 6-8L
  ENGINE
VAPOR VENT VALVE
W/VAPOR LIQUID SEPARATER

                  iV
                                     WATER SEPARATER

                                     LIQUID SEAL
                      7      \
EVAP. CANISTER 1-1.8L   VALVE    OVERFILL LIMITER
     FIG 1-5 ONBOARD VAPOR RECOVERY SYSTEM NON-INTEGRATED (CARB.)

-------
                       Appendix I
                     Figure 17 - Toyota
ONBOARD VAPOR RECOVERY  SYSTEM
       PURGE CONTROL VALVE
                   REFUELING CANISTER 6-8L
 ENGINE
        VAPOR VENT VALVE
        W/VAPOR LIQUID SEPARATER
                                     WATER SEPARATER
                                     LIQUID SEAL
EVAP. CANISTER 1-1
          VALVE    OVERFILL LIMITER
        FIG
1-3 ONBOARD VAPOR RECOVERY SYSTEM NON-.NTEGSATED
                                          (EFD

-------
                      Appendix I
                    Figure 18 - Toyota
ONBOARD VAPOR  RECOVERY SYSTEM
      PURGE CONTROL VALVE      REFUELING CANISTER 6-8L
       VAPOR VENT VALVE
       W/VAPOR LIQUID SEPARATER
WATER SEPARATER
                                   LIQUID SEAL
                           \
                  VALVE    OVERFILL LIMITER
        FIG 1-4 ONBOARD VAPOR RECOVERY SYSTEM INTEGRATED (EFI)

-------
                                    Appendix I
                               Figure 19 - Mitsubishi
           CONROL UNIT

              PURGE CONTROL VALVE

                    ROLLOVER  VALVE
                   LA/AR CANISTER
                   VAPOR LINE
                    NOZZLE  ACTUATED
                            VENT VALVE
           PURGE LINE
                  WATAR SEPARATOR
                                                                           NOZZLE SEAL
        VAPOR  LIQUID SEPARATOR


OVERFILL LIMITER
                                                                             FILLER CAP
                                                                       FILLER  PIPE
                                                 FUEL TANK     \
                              REFUELING  CANISTER               OVERFILL LIMITER VALVE
                                                ( UNESTABLISHED  TECHNOLOGY \
Fig  3  ONBOARD  VAPOR  RECOVERY SYSTEM V NONINTEGRATED  SYSTEM     /

-------
                                       Appendix I
                                   Figure 20 - Mitsubishi
           CONTROL UNIT
             PURGE CONTROL VALVE
     If
                             NOZZLE  ACTUATED
                                   VENT VALVE
                            VENT LINE
                PURGE LINE
                  WATAR SEPARATOR
)ZZLE SEAL
         VAPOR LIQUID SEPARATOR

        OVERFILL LIMITER

ROLLOVER VALVE   \ VAPOR LINE
                                                                              FILLER CAP
                                  CANISTER
                                                 FUEL TANK
                                                                        FILLER PIPE
                                                                   OVERFILL LIMITER VALVE
                                              / UNESTABLISHED  TECHNOLOGY
2   ON BOA R D  VA POR  RECOVERY  SYSTEM  ( INTEGRATED  SYSTEM	

-------
                             Appendix I
                    Figure 21 - American Petroleum Institute
                             (Exxon Design)
                   PRINCIPAL ELEMENTS OF
               ENHANCED ON-BOARD SYSTEM
                            Rollover Shutoff

                             Overfill/ShutoYf
                                Entrained
                                  Liquid
                                Separator
 Standard
 Fillpipe
 With Vent
Rerouted to
 Cannister
  HC
Vapors
                                  Purged
                                 HC Vapors
                                 To Engine
                                                                        o»
                                                                        I
Enlarged
Charcoal
Cannister

-------
                                 Appendix I
                     Figure 22 - American Petroleum Institute
                                (Mobil Design)
        ONBOARD REFUELING EMISSION CONTROL SYSTEM
Vapors Purged
  to Engine
During Driving
Solenoid
Valve and J>
yM" Orifice
Control
Purging
=fl •/.' Vapor Line -^
*rh ir-


1 	 1
Vapor/Liquid -•
Productlon Vent and — *f "" | Separator
Rollover Protection

 Nozzle-Actuated
\^ Vapor Vent
'^   Valve
      4.5 Liter
   Carbon Canister
     Traps Vapors
                     Liquid Seal In
                   Prevents Escape of Vapors

-------
                                  Appendix I
                                Figure 23 - IVbbil

                               FIGURE 1
                           SCHEMATIC OF
      ONBOARD REFUELING EMISSION CONTROL SYSTEM
                         FOR OPEL ASCONA
PRODUCTION
VENT AND
CHECK VALVE
NOZZLE-ACTUATED
VAPOR VENT VALVE
                         VAPOR/LIQUID
                         SEPARATOR
                               VACUUM VALVE AND
                               PURGE CONTROL ORFICE
VAPORS PURGED
TO ENGINE
DURING DRIVING
                           BREATHING VENT
                            REFUELING VAPOR
                            OUTLET
                               19MM VAPOR LINE
                                                 4.7 LITER
                                                 CARBON CANISTER
                                                 TRAPS VAPORS
   NOT TO SCALE

-------
                                                       Appendix I
                                                        Figure 24
                                              Multinational Business Systems
                                                        VAPOR/LIQUID SEPARATOR
URGE, TO ENGINE
 LENOIO PURGE VALVE
                             AIR FILTER
 EVAPORATIVE EMISSIONS AND

 REFUELING VAPOR CARBON CANISTER
                                                 CONDENSATE RETURN LINE
PRESSURE/VACUUM

      VALVE
                                                                                  NOZZLE ACTUATED VENT VALVE
MECHANICAL NOZZLE SEAL
  )LLOVER VALVE/ PRODUCTION VENT ASSEMBLY
                                                                  ANTI-SPITBACK GATE
                                                    CONOENSATE CHECK VALVE AND

                                                    LIQUID SHUT-OFF VALVE
                                                                                                        MMIMM. I  •rtcw
                                                                                    REFUELING VAPOR RECOVERY SYSTEM
                                                                                        MECHANICAL NOZZLE SEAL
                                                                                 IIUlLLi  AtlQCIATM. IMC MHtoOft. MO
                                                                                             NONE    | 5 KETCH !**«• ' °*J
                                                    Figure  1-1

-------
NOTE! DEPENDING ON CARBURETOR VENT OPERATION,
      EVAPORATIVE EMISSIONS CANISTER MAY NOT
      HAVE TO BE RETAINED
                                                           Appendix I
                                                            Figure 25
                                                 Multinational Business Systems
                             VAPOR/LIQUID SEPARATOR
        REFUELING VAPOR
        CARBON CANISTER
                                      CONDENSATE RETURN
                         FILTER
          SOLENOID
           PURGE VALVE
                                           LIQUID SHUT-OFF VALVE
                                          AND CONDENSATE CHECK VALVE
CARBURETOR VENT
                                                                                                 LEAD RESTRICTOR DOO>
                                                                                 NOZZLE ACTUATED VENT VALVE
         EVAPORATIVE EMISSIONS
      I!  CARBON CANISTER

      \
PURGE. TO ENGINE              ROLLOVER VALVE/PRODUCTION VENT ASSEMBLY
                                 J-BEND LIQUID SEAL
                                                                                             M«Ca»TIOH    I MftTIMM.
                                                                                     REFUELING VAPOR RECOVERY SYSTEM
                                                                                                  LIQUID SEAL
                                                                                   rJUELLEB AttQCIMH. IMC »>Hlmoc». MD
                                                                                   n>t« n/it/u| »c»m NONE   | bHr.TCM  \'*a» \~or
                                                  Figure  1-2

-------
                                                   Appendix I
                                                    Figure 26
                                          Multinational Business Systems
PURGE. TO ENGINE
                     SOLENOID PURGE VALVE     VENT VALVE.  ROLLOVER VALVE.
                                           EVAPORATIVE VENT AND
                                           VAPOR/LI QIUD SEPARATOR
                                           INTEGRATED ASSY* ,
                                VENT VALVE
                                SOLENOID


                            DESIGN ULLAGE
                       CAP-ACTIVATED
                       SWITCH
        ACTIVATED
        CARBON CANISTER
                             MECHANICAL
                             NOZZLE SEAL
                                                                            FILLER TUBE VENT LINE
ANTI-SPITBACK GATE
                                                                                                     MAIBMAI.    •»•€
                                                                                 REFUELING \fcPOR RECOVERY
                                                                                        CLEAN SHEET DESIGN
                                                                                               INC Mlllmof*. MO
                                                                                                           |>«o«lor
                                          Figure  IV-1

-------
         Appendix II





Safety Implications of Onboard





   Refueling Vapor  Systems

-------
                                        EPA-AA-SDSB-87-05
                   Technical Report
       Safety Implications of Onboard Refueling
                Vapor Recovery Systems
                      June  1987
                     FINAL REPORT
Technical Reports  do  not necessarily  represent  final  EPA
decisions  or  positions.    They  are  intended  to  present
technical  analysis   of   issues  using   data  which   are
currently available.   The purpose  in  the release  of  such
reports  is  to   facilitate  the  exchange   of  technical
information  and   to  inform   the  public   of  technical
developments which may  form  the basis  for  a  final  EPA
decision, position or regulatory action.
       Standards Development and Support Branch
         Emission Control Technology Division
               Office of Mobile Sources
             Office  of  Air  and  Radiation
        U.  S. Environmental  Protection  Agency

-------
                       Table  of Contents





Section








I.     Executive Summary                                3





II.    Introduction                                     7





III.   Onboard Control System Description               9





IV.    Design Considerations for a Safe System         27





V.     In-Use Fuel System Safety                       50





VI.    Cost and Leadtime Considerations                61





VII.   Heavy-Duty Gasoline Vehicle Requirements        74





VIII.  Conclusion                                      95





IX.    References                                      97

-------
                               -3-
I.   Executive Summary

     The  purpose  of  this  technical  report  is to  evaluate  the
safety   implications   of  requiring  onboard   refueling   vapor
recovery  systems  on  gasoline-powered  passenger  cars,   light
trucks  and  heavy-duty  vehicles.    In  that   light,   special
attention is given  to the  analysis of the design considerations
for a safe  onboard  system and  the other measures  necessary to
insure that  the design  considerations  incorporated are capable
of providing a high level of in-use fuel system integrity.

     Onboard  refueling  systems  are  in  many  ways  similar  to
present  fuel  tank evaporative emission  systems.   The emissions
emanate  from the same location  on the  vehicle  and the  basic
technology used  to  control the two types of emissions  is  quite
similar.    Many   of   the  components   are  analogous,   if  not
essentially  identical,  and  the  configuration/layout  of  the
systems on the  vehicle is  also expected  to  be about  the  same.
In fact,  these two  systems and system  functions  are so similar
that  many  manufacturers  will  likely  combine  their  onboard
refueling and  fuel  tank evaporative  emission systems  into  one
integrated system which can serve both  purposes.   The fact that
these  systems  are  similar  and  will  be  integrated  has  two
effects on the  safety of  onboard  systems.   First, many of  the
approaches and  techniques  used to  safely implement evaporative
emission control systems can also be  applied to  insure  the safe
implementation  of  an integrated  onboard refueling/evaporative
emission  system.   Second,  any   safety  problems  related  to
integrated   onboard/evaporative   systems  should   be  evaluated
incremental  to  present  evaporative  systems.   Quite  simply,
there is no need to add a whole new system to the vehicle.

     Concerns over  the potential  safety implications of onboard
systems  have,  however,  been  raised.  These  concerns  can  be
grouped into four general  areas.   These include requirements to
pass the National Highway  Traffic  Safety  Administration (NHTSA)
safety  tests,   the  effects  of  tampering and system  defects,
refueling operations,   and in-use fuel system safety.

     Concerns with  the design  requirements  necessary  to  comply
with the NHTSA  safety tests  focused on the  need to integrate an
onboard system  into  a vehicle  in  a manner  which  would provide
the crashworthiness and  rollover  protection  demanded by Federal
Motor  Vehicle  Safety Standard   (FMVSS)  301.    EPA's  analysis
indicates  that  crashworthiness   for  the  key  vapor lines  and
other system components could be accomplished  using many  of the
same  approaches  and  techniques   now  applied  successfully  to
evaporative emission  systems.  Further,  the  rollover protection
now provided for the fuel tank through  the  use  of  a  limiting
orifice  can  be  gained through the application  of one of  the
several rollover valve designs now available.

-------
                               -4-
     Concerns  have  also been expressed  that  canister tampering
 and  component  defects  could  lead  to   in-use  safety problems.
 While canister  tampering is  infrequent,  the  rate can be reduced
 and  the  potential  safety  effects  eliminated through  proper
 placement.   Manufacturers  are expected  to consider " the  safety
 implications  of  tampering  when  evaluating  canister  location
 options on  the vehicle  as they do  now with  evaporative control
 system  canisters.   While  the concern has been  expressed  that
 defects   in  onboard   system   components  could   have  safety
 implications,  no  data  or  other   bases  have  been  found  that
 suggest onboard systems would  influence  the  nature or frequency
 of  such  occurrences   as  compared  to   those  seen  on  current
 evaporative  emission systems.   In fact,  given the  experience
 gained  by the manufacturers  in  safely  implementing evaporative
 controls,  it  is likely  that an  integrated  onboard/evaporative
 system  could  be  implemented  with no more  (and !perhaps  less)
 problems than present evaporative emissions systems.

     Concerns  over  the  safety  of  refueling  operations  are
 centered  on the potential to  overpressurize the  fuel  system.
 EPA's  analysis finds  that  use of a  liquid  seal   solves  all
 overpressure problems,  and that if a mechanical  seal is  used a
 simple  pressure relief device  can be  used  to  eliminate  any
 overpressure  concerns.   As  discussed  in the  analysis,  a  few
 other   less    significant    potential    problems I   have    very
 straightforward engineering solutions.            !

     Finally,  while  it  is clear  that onboard-equipped vehicles
 can be  designed to  comply with FMVSS  301  requirements,  there
has been  concern expressed  that  fuel  system  integrity  in-use
may decrease by some non-quantifiable amount because FMVSS  301
 can't cover all potential accident  situations and  an onboard
 system  requires modifications  and additions  to,  the  present
 evaporative emission system.   While no  test  procedure can cover
 all potential  situations,  it does  not  necessarily  follow  that
 system  modifications or  additions  will  cause an  increase  in
 risk over present systems.

     Both vehicle  and fuel  system safety are  evaluated as  an
 integral  part  of the  overall  design  and development  process.
This  involves   multiple trade-offs,  balances,   and compromises
with  other  key  design  considerations.   Given  the  need  to
consider  all  key  design  criteria,  manufacturers  accept  or
manage  a  certain  amount of  risk.   Since  the  safety demands  of
Federal standards  such  as FMVSS  301  must be  incorporated  into
vehicles/systems,   these standards  represent  the  minimum.   In
many cases the level of safety achieved  in-use  goes beyond that
 required by Federal  standards,  being driven  by in-use liability
concerns.

-------
                              -5-
     If a  manufacturer  perceives that the  added  risk  mentioned
above may  exist  for one  or  more of  its  vehicle  models,  there
are ways to  respond through direct measures or through keeping
the  overall  risk  in-use  at  acceptable  levels  through  other
design flexibilities.   EPA's  analysis identifies and  describes
a  number  of  these  measures.   Manufacturers  can  make  vehicles
safer  than  they  are  now;   an  onboard  requirement  does  not
increase the  amount  of  risk  a  manufacturer need  incur  or
accept.    Manufacturers   are   expected  to  integrate   onboard
controls into their fuel systems  without  compromising safety.

     Further, as part of  overall risk management,  implementing
onboard controls  provides  the  opportunity  to improve  overall
fuel handling and fuel system safety.  Refueling  spills  will be
reduced and  flammable vapors  will be  trapped in  the  canister
instead of  being  vented  out  the  fillpipe  near   the  nozzle
operator  where   inadvertent  ignition   is   possible.    Also,
installing  rollover valves  could improve the  safety for  those
vehicles  now  using   external  fillpipe   vent   lines   without
rollover valves.   The  positive  seal  provided  by  a  rollover
valve  is  an  improvement  over  the  "controlled  leak"  rollover
protection   currently  provided  by  a   limiting  orifice.    In
addition,   implementing  onboard  systems   could further  enhance
safety  by   providing  the  opportunity  to   make   other  safety
related  fuel  systems  changes  which  have  been  delayed  for
economic or  other  reasons  (e.g.,  changing  from  rear  to  side
fill).  Finally,  if a manufacturer  chooses  to  use a collapsible
fuel  bladder  to   control   refueling  emissions,   this   would
eliminate  all  of  the potential concerns  raised relative  to the
canister based onboard  system,   and  would provide  improvements
in safety  over the  present fuel system.

     Other  key considerations  include safety  related  costs and
the leadtime  needed to implement onboard  controls safely.   This
analysis estimates  that  safety  costs  related to  implementing
onboard systems will range from  $4.50-$9.00  per vehicle.   While
the cost  estimates for  the  needed hardware,  modifications and
fuel consumption impacts are reasonably accurate, there is  some
uncertainty  in  the development  and safety  crash  testing  cost
estimates.    However,  safety  related onboard  costs  are  quite
insensitive  to  even   large  changes  in   the   estimates   for
development and safety certification.

     In a  general  sense,  EPA's  estimates  are  supported  by the
fact  that  the  modifications  needed  for  present  vehicles  to
insure fuel system safety  in-use have been  acquired relatively
inexpensively,  and  vehicles  with  evaporative  emission  systems
comply with FMVSS 301 today.   Much of the groundwork needed to
implement   an  integrated onboard refueling/evaporative  emission
control system safely has  been completed and  many of  the  same

-------
                               -6-
techniques   and  approaches  can   be   used.    The   fact  that
integrated  systems  will be used means  that some costs incurred
to  implement  evaporative  emissions systems  safely  will  not
reoccur.   EPA's  analysis  has  adequately  accounted  for  safety
costs in its estimate of the  total  onboard system cost.   Safety
costs contribute  about  25  percent  of the  $20 cost estimated for
a passenger car onboard system.

     With  regard  to leadtime,  given the magnitude  of  the task
and past  experience with  implementing  evaporative emission and
fuel  system  integrity  standards   (FMVSS   301), ;this  analysis
indicates that  24 months  leadtime  is adequate.  However,  EPA is
committed to providing the  leadtime  needed to  implement onboard
controls  safely  and  effectively,   and  . is open  to  considering
additional  leadtime or a  short phase-in of contlrols  to  assist
manufacturers in dealing with problems on unique vehicle models.

     Finally, the onboard  systems  which would b;e  installed on
HDGVs are  quite similar to those  expected for  passenger  cars
and light  trucks,  even though  the  safety test requirements are
different  for HDGVs.   With the exception  of  school  buses, the
fuel  system integrity  testing  centers more  on'  evaluation of
fuel tank  integrity than vehicle crash  testing. '  Nevertheless,
many  of  the  concerns  raised  and addressed  above  regarding
onboard safety  for  lighter-weight vehicles also  apply  to HDGVs
and support the  judgment  that  onboard systems  can  be  applied
safely to this  class of vehicles within the  lea'dtime  laid out
above and for a reasonable cost.

-------
                              -7-
II.   Introduction

     EPA has  received  several comments  from  the Motor  Vehicle
Manufacturers Association, Automobile  Importers  of  America (and
their  member  companies),  and  the   Insurance   Institute  for
Highway Safety  which have  expressed  various  levels  of  concern
about  the  potential  safety  implications   of   onboard  vapor
recovery   systems.[1,2]     Also,   some   preliminary   comments
regarding  onboard  safety  have  been  received  from  NHTSA's
technical  staff.[3]   The  American Petroleum Institute  (which
has independently developed several   onboard-equipped vehicles)
and the  Center  for Auto  Safety  have   expressed  support  for the
implementation  of  onboard  vapor  recovery  systems.[4,5]   The
purpose of  this  report  is to  discuss  and  analyze  the  safety
related  concerns   raised  regarding   onboard   vapor  recovery
systems.

     Motor vehicle  manufacturers  face many difficult technical
decisions in the  design  and development  of vehicle systems and
the integration  of  these systems into new vehicle  models.  The
difficulty of  these decisions often  arises  from the  fact that
this design,  development and  integration process  requires the
simultaneous consideration of a number of key criteria.   One of
the most important  of  these criteria, safety, is normally given
a  high  priority   in  the   design   and   integration  process.
However,    the   process   also   includes   careful   and  prudent
consideration of  the trade-offs  necessary to  deal  with other
important  criteria  such  as  performance,   reliability,  cost,
styling,  and  regulatory  requirements   such as fuel economy and
emissions.     In  each   case,  manufacturers  must   find  the
appropriate balance  of  all the  important  criteria.   Since the
design of  emission  control systems has  the potential to affect
the overall safety  of  vehicles,  EPA  views  safety  as  a primary
concern when  evaluating  the feasibility of an emission control
device.

     EPA  is  presently  evaluating  the  use  of  onboard  vapor
recovery systems  (onboard  systems)  as  a  means  of controlling
refueling emissions.  The potential safety implications of such
controls  require  special  consideration,  because  implementing
onboard systems  will  involve  some minor  modifications  of the
vehicle fuel  system.  While  safety  influences  all  aspects  of
vehicle  design,  fuel  system  safety   and  integrity   is   a  key
concern in the design and integration process.

     In evaluating the safety  implications of requiring onboard
controls,  EPA  has applied  the  philosophy that  no increase  in
overall  risk  should  be  caused  or  accepted,  beyond that  now
present with today's fuel/evaporative system.   This  applies  to
both compliance with  the  applicable Federal  safety standards

-------
                               -8-
 and   the   in-use  safety  of  vehicles  equipped  with  onboard
 systems.   The   following   analysis   will   show  that  straight
 forward  engineering  solutions  are   available  for  all of  the
 potential  safety problems  which have been  identified,  and that
 while  final  choices  regarding exact  system designs lie with thQ
 manufacturers,  safe  fuel  system designs are  achievable by all.
 This  analysis of onboard  safety  issues and the associated cost
 and  leadtime generally  applies to  any canister-based onboard
 system  design.   Further,  as  will   be discussed  below,  this
 analysis indicates that  it is quite possible that  overall fuel
 system  safety  improvements  could accompany  the  implementation
 of onboard controls.

     The importance of evaluating  the  safety of onboard systems
 is highlighted  by the Clean Air Act  (Section 202 (a) (6))  which
 directs EPA  to  consult  with  the  Department of  'Transportation
 (DOT)  before   requiring  the  use  of  onboard  vapor  recovery
 systems.   This  requirement  is  intended  to  insure  that  all
 safety  issues  have   been  properly   identified  and  addressed.
 This report will also help to  assist  in the fulfillment of this
 requirement.
                                                  i
     As outlined  below,  the remainder of this report is divided
 into   five  sections.    The  first   section  following   this
 introduction  (Section III),  provides   a general  description of
 an onboard system to aid  in the  understanding of  any related
 safety  issues.   Section   IV  summarizes   and  provides  EPA's
 analysis of  the  comments  received  regarding the  design  of  a
 safe onboard  system,  and  Section V discusses onboard effects on
 in-use  fuel  system  safety.   Section  VI  discusses  the effects
 safety  considerations have  on other important factors  such as
vehicle  costs    and  leadtime.    Heavy-duty   gasoline-fueled
vehicles   (HDGV)  pose  similar  yet  distinct  onboard  control
 system  safety   issues,    and  Section  VII   addresses   these
 similarities  and  differences.    The   final  section  provides
conclusions.

-------
                              -9-
III. Onboard Control System Description

     Before considering  any safety  issues,  it is  important  to
have a clear understanding  of  onboard refueling vapor  recovery
systems (onboard  systems)  and how they work.   Likewise,  before
considering the  characteristics of  the control  system,  it  is
important to understand the  nature  of refueling emissions.   The
purpose of  this  section is  to  provide the  reader  with both  a
clear  understanding of  what refueling  emissions  are and  how
onboard systems operate to control  these emissions.

     In  many   respects,  onboard  systems  are  similar  to   the
evaporative  emission  control  systems  now  in  use  on  most
gasoline-powered vehicles.   In fact,  it has  been  suggested  that
onboard  systems  are  more  an  extension  or  modification  of
current evaporative emission systems  than  the implementation of
a new  control technology.   An explanation  of the  differences
and similarities between the  two systems  will provide  a  better
understanding  of  the   incremental   nature   of  onboard  systems
relative to current evaporative systems,  and will be  useful  in
assessing  the  design,  cost,  and   leadtime  implications  of
implementing onboard controls safely, which  are  to  be discussed
later in the report.

     This  section  will  first  briefly   describe   evaporative
emissions  and  how  they   are  currently   controlled.   Next,
refueling emissions will  be discussed  and similarities between
onboard systems  and current evaporative  emission  systems  will
be  presented.   The section  will  end  with a discussion  of  the
differences between the two control systems.

     A.    Evaporative  Emissions

     Evaporative emissions emanate from two  basic  sources:   the
fuel tank  and  the fuel metering system (either  a  carburetor or
fuel injectors).   Evaporative  emissions arising  from  the  fuel
tank are  primarily  "diurnal"  emissions  while  those  from  the
fuel metering  system  are  termed  "hot  soak"  emissions.*   This
analysis  is  primarily  concerned with fuel  tank  evaporative  or
diurnal   emissions  since   these    emissions   are   currently
controlled using  an approach similar to that  envisioned  for an
onboard system.
*    It  should  be  noted  that  a  small  amount  of  hot  soak
     emissions  come   from   the  fuel   tank;   the  fuel   tank
     evaporative control  system would  handle  these as well  as
     the diurnal emissions.

-------
                              -10-
     Diurnal    evaporative    emissions    consist    of   gasQous
 hydrocarbons  that are displaced from the  tank  when fuel  in the
 tank  is  heated.   Fuel  heating  can  result  from  changes  in
 ambient  temperature  or  during vehicle  operation  due to  the
 vehicle  exhaust system and/or  recirculation  of fuel  heated  by
 the  engine.   In either  case,  as  fuel  in the tank  and  vapor
 above the fuel  heat up, more  of the liquid fuel evaporates,  and
 the vapor  itself expands, thus causing hydrocarbon vapor  to  be
 released  into  the  atmosphere  (unless  captured  by  a  control
 system).   Fuel  volatility,   size  of the  vapor 'Space,  initial
 tank temperature,  and the degree  to which the tank  is  heated
 can  all  impact  the  quantity of hydrocarbons  emitted.  Diurnal
 emissions  occur  on  at  least  a  daily  basis,  and  a  system
 designed to control  these emissions must  be capable of handling
 repeated evaporative  emission loads.  Since  the early 1970's,
 most vehicles  have  come equipped  with a control system to limit
 the amount  of  diurnal evaporative  emissions.    The  next section
 discusses the  type of control  system typically used on today's
 vehicles.

     B.    Evaporative Emission Control System

     Figure l  depicts a fuel tank equipped with  an evaporative
 emission control system.[6]   As can be  seen from  this figure,
 the control system  is relatively  simple  in design  and requires
very  few  components.   The  purpose  of  this  section  is  to
 describe  each  of  the  system's  components  in terms  of  both
physical appearance and function.

     In  order  to  effectively prevent the escape  of  fuel  tank
vapors to  the  atmosphere,  an  evaporative control  system  must
perform  three  basic  functions.   First,  the  system  must  limit
the  number  of  exits  through  which  fuel  tank  vapors  might
 escape.  Second,  the exit that does allow fuel  tank vapors  to
 escape  must  lead  to  a  container  where  the  vapors  can  be
captured.   Third,  the   system must   eventually  restore  the
capacity of the storage container  by purging it of  the trapped
vapors.   The  discussion  below describes how  an  evaporative
emission system performs these three functions.

     The  first  function  an  evaporative   emission   system  must
perform  is  to  limit  the  outlets through  which  vapors  can
escape.  As can  be  seen in  Figure 1,  there are  only  three
openings  through  which  vapors  can  pass:   1)   the  fillpipe
opening,  2)  the external vapor vent line  to  the  fillpipe  top
 (about  1/2"  diameter),  and  3)   the  small  limiting  orifice
 (approximately  0.050-0.055  inch)  in the  top  of the tank.   The
fuel  tank  cap  is  designed  to  form  a   tight  seal  with  the
fillneck so that once the cap is secured  in place,  vapors  from
the  fillpipe   opening and the  external  vent  line  are trapped
within the  system.   Thus,  only one outlet  exists  through  which
fuel tank  vapors can  escape.   This single available  outlet  is
the small limiting orifice in the top of  the tank.

-------
                                                           Figure 1

                                             Typical Current Evaporative System
PRESSURE/VACUUM
RELIEF CAP
EXTRENAL VENT
   LINE
                                     -LIMITING ORIFICE
                       ' -TLOAT/ROLLOVER
                           VALVE
                                            Z.
                                                 7
\
                                                                                                         PURGE VALVE
                                                    DIA.
                                                 8'  LONG
    DIA.
                                                                                           1  LITER
                                                                                           CARBON
                                                                                           CANISTER
                        14 GALLON FUEL TANK
                                                                                                 TO PURGE
                                                                                                 INDUCTION
                                                                                                 POINT

-------
                              -12-
     As the  tank  undergoes temperature changes,  and hydrocarbon
 vapors  are generated, pressure  builds up in  the  tank (as long
 as  the  fuel  tank  cap  is  secure  in  place).   This  pressure
 build-up  is  slowly  relieved as gas tank vapors eventually force
 their way through the only available  exit:  the  small limiting
 orifice  in the top  of  the fuel  tank  which leads  to  the vapor
 storage device (charcoal  canister).   By limiting the  number  of
 vapor escape passages and  routing the evaporative hydrocarbons
 to   a   single  point,  the  control   system   has   successfully
 performed  the  first  of  its  three   basic  functions.   Before
 discussing  the evaporative  emission   system's second  function,
 it  is important to  understand why the orifice in the top of the
 tank is so limited in size.

     The  orifice  in  the  top of the tank  is very; small  in size
 for  three reasons.   First,  it  allows pressure  to build  up  in
 the  tank  when  vapors are  generated.   This   pressure build-up
 inhibits   further    evaporation   and   creates;   a   pressure
 differential  which eventually  leads to hydrocarbon vapor being
 forced  through  the  limiting  orifice.    Second,'   the  limiting
 orifice acts as a liquid/vapor  separator.  If liquid gasoline
 were to splash up into  the vent line  leading to the evaporative
 emission  control  storage  device  (charcoal  canister),  damage
 could  potentially  occur   to   the  storage  media  (charcoal).
 However,  the orifice in the  top of the  tank  is; so  small that
 liquid passes  through it  at only a very slow rate.   Essentially
 only vapor is allowed to  continue to  the  canister.   This point
 leads to  the  final  reason  for  limiting  the  size of  the vent
 orifice to  such an extent.   Were the vehicle eyer to be  in a
 rollover  accident, a  very  little amount of liquid fuel would be
 able to leak from the tank through  such  a  small  orifice.   Thus,
 the  limiting  orifice  is  sized  large  enough  to  allow  for
 adequate  escape of  evaporative  emissions,  but is  small  enough
 to permit only a  slow leak from the fuel  tank  in  the case of a
vehicle rollover  and thus provides  the  protection  needed  to
 comply  with  FMVSS 301.   The  cost for  this  is low.   However,
 some manufacturers  incorporate  an additional valve  for  added
protection; an example is shown in Figure 2.[7]

     Storing  the  evaporative  hydrocarbons is the  second basic
 function an evaporative emission  system must perform.   Once the
vapors  escape  from  the  fuel  tank through  the small limiting
orifice,  they  proceed  through  a  vent  line  (usually  about
 l/4"-3/8"  inside diameter  and  made  of  some  type  of flexible
 rubber compound)  that leads  to a  canister  containing charcoal.
The canister  itself  is usually made of plastic and is generally
 a  cylindrical  or  rectangular   container.    Once   inside  the
canister,  the  hydrocarbons  are adsorbed onto activated charcoal
where they are stored temporarily.

-------
                                     -13-
                                 Figure 2
                                 STANDARD
                                 VERSIONS
The tank mounted spring balanced float valve is a low cost unit designed for venting
fuel tank vapor to the carbon canister. The device employs a float which remains open
under normal conditions.  Should the tank level reach a critical height, the float will
close the canister vent line. In the event of extreme vehicle attitude or roll-over, the
float will close the canister vent  line.

A filtered tank mounted spring balanced float valve is available that performs the
same functions as the above sketches except the tank side of the part is filtered to
prevent contaminates from entering the part which might effect float closing of the
canister vent line.

For high flow applications that require a large volume of vapor venting, such as fuel in-
jection applications, a high flow valve has been developed that has more than twice
the present flow capacity without loosing other critical performance parameters.
                          FLOAT VALVE
•Qj^l Borg-Warner Automotive, Inc.
Sr^S 707 Southside Or, Decatur, Illinois 62525
m^S Phone 21 7/428-4631

437
SKETCH
NUMBER

-------
                              -14-
     The  working capacity  of  the charcoal,  the  quantity  and
 frequency  of the evaporative emissions,  and  the capability of
 the  system  to   restore  its  working  capacity  all affect  the
 amount  of  charcoal  required.   Current passenger car .evaporative
 emission  control systems  typically  utilize  a  0.85-1.5  liter
 canister.[8]  (This  size  is  sufficient for both diurnal and hot
 soak  evaporative  emissions.)   However,  a  finite  amount  of
 charcoal  is  used in the canister,  so the storage capability of
 the  canister is limited.   Once  the evaporative  hydrocarbons
 have been  adsorbed  onto the  charcoal in the  storage  canister,
 they will remain there until  removed.   The  hydrocarbons must be
 stripped  from  the  charcoal  periodically  in  order to restore
 enough  working  capacity to  adequately capture  each successive
 evaporative emission load.                        I

     While  the   vehicle  is  operated,  the  evaporative  emission
 system  performs  its  third   basic function  of  | restoring  the
 storage  capability   of  the   charcoal   canister.    After  the
vehicle's engine is  running,  manifold vacuum  is used  to draw
hydrocarbon-free    air   through    the    charcoal   canister.
Hydrocarbons  stored  in the canister  are desorbed  into  the  air
 stream which  flows  into  the  fuel metering system via a flexible
 rubber  purge line of  about  3/8"  diameter.   Once   purged,  the
evaporative  hydrocarbons  are  burned  as  fuel  through  normal
combustion in the engine.  This  process  "empties'; the  canister,
thereby preparing it for the next evaporative emission load.

     One  aspect  of   the  purge  process  which  needs  to  be
mentioned but will not be explained in great detail is  the fact
that the canister  is  not  continuously  purged Iduring vehicle
operation.[8,9]   A valve  located  between the  canister  and  the
fuel metering system  is  opened and closed at opportune times to
control the  purge  process and limit  disturbances  which affect
engine performance and exhaust emissions.        '

     To  summarize,   the current  evaporative  emission  control
system performs  three  basic  functions:   1)  it  limits  the  exits
through  which  fuel  tank vapors  can  escape;   2)  it traps  the
vapors in a  storage  device;  and 3) it restores  the capacity of
the  storage  device  to  prepare  it  for  the  next  evaporative
emission load.

     Onboard  systems  are  very  similar  to  evaporative  emission
control systems  because they  must  also  effectively perform the
same three  basic  functions  to efficiently  control  refueling
emissions.    However,  due  to  differences  in  the   quantity  of
vapors and the  rate  of generation of evaporative  and  refueling
emissions,  equipping vehicles  with onboard  systems  will require
that  some  minor  modifications  be  made to   current  fuel  and
evaporative emission control systems.

-------
                              -15-
     The  next  section  provides  additional  detail  regarding
refueling  emissions  to  help  explain the  fuel and  evaporative
system modifications  that  would be  required to equip  vehicles
with onboard systems.

     C.    Refueling Emissions

     Three processes  contribute to the  release of hydrocarbons
during a  refueling operation.  The  first  two  are collectively
termed  displacement  losses,  the  third spillage.   First,  the
hydrocarbon  vapor  present  in the tank  is  displaced  from  the
fuel tank  by liquid  fuel entering through the fillpipe.   If the
vehicle fuel tank  is  equipped with an external vapor vent line
(as shown  in Figure  1),  much of the fuel tank vapor escapes via
the external vent  line  which is  connected  to the  top  of  the
fillpipe.   However,  if no such vapor passage  exists,  the vapor
makes  its  way  out   through  the  fillpipe  concurrent  to  the
incoming  liquid  fuel.   Hydrocarbons  are   also  generated  and
released   during  refueling   as   a  result  of   liquid  fuel
evaporating  as  it  is  dispensed into  the tank.   This second type
of  displacement  loss  is  caused  by  the  turbulence  in  the
liquid/air   interface  during  the  refueling   process   and  is
enhanced  by  the  higher  volatility  of  the  dispensed  fuel
relative   to  the   fuel   in  the  tank.   A  third  source  of
hydrocarbon  refueling  emissions  is  the  evaporation  of  any
liquid  fuel  spilled  during  the  refueling  operation.    Of  the
three refueling  emission  sources,  the two displacement  sources
are  generally  much  greater   (by  far),  unless  a  large  spill
occurs.

     Because the bulk  of refueling emission  .emanate from within
the fuel  system, refueling emissions are  in  many  ways  similar
to diurnal  evaporative emissions.  Therefore,  it  follows  that
an effective onboard  system  can be designed which utilizes the
same  basic   technology   and   approach   utilized   by   current
evaporative  emission   systems.   In fact  virtually  all  onboard
systems   considered   by  manufacturers   in   their   comments
incorporate  this approach  as  do the  prototype systems developed
to date.[10,11,12,13,14,15]   The  similarities between  onboard
and evaporative emission systems are discussed below.

     D.    Onboard Refueling Control  Systems

     1.    Similarities   with   Evaporative    Emission   Control
           Systems

     In order  to control  refueling  emissions, onboard  systems
must  perform  the  same  three  basic  functions  as  described
previously  for  diurnal  evaporative  emission systems.   These
include limiting the  number  of  exits  through which  refueling
vapors can escape,  storing refueling emissions temporarily in a

-------
                              -16-
 charcoal  canister,  and  purging the  charcoal  canister of  the
 stored  refueling vapors  to  restore  its  capacity prior  to the
 next refueling operation.  Because these  three  functions  are so
 similar  to the  three  functions a diurnal  evaporative emission
 control  system  must perform  and  the emissions  arise  from the
 same  location,  extrapolation of  known  technology leads  to the
 conclusion that  an  onboard system would  use the  same approach
 and similar hardware to  that  which is currently used to control
 evaporative   emissions.     Figures   3    and   4   depict   two
 representative  onboard systems and  a comparison with  Figure 1
 shows that onboard  controls  are very similar in  overall  design
 to  current   diurnal   evaporative  emission  control   systems.
 However, while  onboard systems do use  many  of  the same  basic
 components as evaporative systems,  (i.e.,   charcoal  canisters,
 flexible rubber  tubing,  purge  control valve, etc.),   the  basic
 differences between refueling and  evaporative emissions require
 a  few  additional  components,  and an  enlargement  of  certain
 existing hardware  is  required  for the  onboard  system.   These
 are the key differences between the two  systems.

     Before   discussing   the   component    additions    and
 enlargements,   an  important   aspect   of  the  onboard  refueling
 vapor recovery system must be introduced.          ;

     Since both  emissions emanate from  the same  location,  a
 properly designed  onboard system  could control  both  refueling
 emissions  and  diurnal  evaporative   emissions.   ; Thus,  if  an
 onboard  refueling   system were  incorporated into  a  vehicle's
 fuel system,  the current diurnal  evaporative emission control
 system  would  no longer  be  needed.    This  aspect  of  onboard
 systems  has   several  implications.    First,   it  reduces  the
 conceived degree of complexity  the system adds  to the vehicle's
 fuel  system.   An  entirely new,   larger,  more  complex  system
would  not  be  needed  in  addition   to  that  which  currently
 exists.   Rather, the  current control system would be  modified
 to be  somewhat   larger with  a  small  increase in the  number of
 components.   Second,  since   onboard  systems   are   basically
modified  evaporative  emission  systems,   many   of  the  safety
design  concerns  associated with  onboard systems have  already
been addressed  in   current  evaporative  emission  control  system
designs.  These  approaches could also be used in the  integrated
 system.   One  final  effect a "dual function"  control  system has
 is it requires  less  "packaging" space and is less  expensive to
produce than two separate systems.

-------
                                                           Figure  3

                                            Integrated Evaporative/Refueling System
                                                      Nozzle Actuated Valve
                                                      Front Mounted Canister
                                                      Mechanical Seal
c
PRESSURE/VACUUM
RELIEF CAP
•NOZZLE ACTUATED
 ROLLOVER/VENT VALVE
 MECHANICAL
   SEAL
                                            r
                        5/8"  DIA.
                        3'  LONG
I
f-
                                         -PURGE
                                         sVALVE
                                   05" DIA.  LIMITING ORIFICE
                            VALVE
                                                                           ••-3/8" DIA.
                                                                                                                   TO PURGE
                                                                                                                   INDUCTION
                                                                                                                   POINT
                                                                                            3 LITER
                                                                                            CARBON
                                                                                            CANISTER
                          14 GALLON FUEL TANK

-------
                                                          Figure 4

                                             Integrated Evaporative/Refueling System
                                                        Tank Mounted Valves
                                                        Rear Mounted Canister
                                                        J-Tube Seal
•PRESSURE/VACUUM
RELIEF CAP
                   NOZZLE ACTUATED
ROLLOVER/VENT VALVE , — 5/8" DIA. | PURGE
Fa- LONG , LVALVE
' ( ( ^^l\/1 / .
&
/&" D

ffifi^
^i05" DIA. LIMIT
to\s^c*3 \ -— -
NO^— ^7 VALVE ^

^— 	 •; •>
/Lj-TUBE SEAL
DESIGNED SLOW LEAK
11 If 4 ,,o-DIA ' / ^ 1 '
ING 5' LONG T0 PURGE
CE INDUCTION
POINT
3 LITER
CARBON
CANISTER
                                                                                                                              oo
                          14 GALLON FUEL TANK

-------
                              -19-
     2.    Additions/Modifications   to   Evaporative   Emission
           Control Systems.

     The  differences   between  onboard  systems   and   current
diurnal evaporative  emission control  systems  can  be  separated
into two  broad categories:   1) those  related to  the sealing of
the system, and 2) those related to  the magnitude and frequency
of  the  refueling  emissions.   Because  of  these  differences,
onboard  systems  require  several  additional  components,  and
several components  of  the  current  evaporative  system must  be
increased in size or slightly modified.

     a.    Additions to the Present System

     Diurnal  evaporative emission  control  systems  limit  the
number of vapor  exits  by using a fuel tank cap to close off the
fillneck.   However,  during a refueling operation,  the fuel tank
cap  is not  in  place,  and  consequently,  onboard  systems must
rely on  some  other  type of sealing  mechanism  to  prevent  the
escape of  vapor  through  the fillneck opening.   Currently,  two
types  of   fillneck   seals  are  available  for  use  on  onboard
systems — liquid and mechanical.

     Liquid fillneck seals utilize modified  fillpipe designs to
route incoming gasoline in such a way  that a  column of  gasoline
is  formed which prohibits  the vapors  in the  fuel tank from
escaping to  the atmosphere  via the  fillneck.   While  this  may
sound somewhat  complicated at  first,  the  concept is fairly easy
to understand with the help  of a drawing.  Several  liquid seal
configurations  have  been developed,  but  one  design which  has
been shown to be particularly attractive from both a design and
cost perspective is  the "J-tube"  (shown  in Figure  5).[16]   As
fuel  is  dispensed  into  the fillneck,  it  is  forced  to  pass
through the  "U" shaped portion of the fillpipe.   A liquid trap
is  formed in  the  "U" shaped  portion of  the  fillpipe  which
prevents vapors from  escaping  via the fillneck.   The  "J-tube"
extension could be made of metal,  plastic  or hard rubber.

     Another type of  fillneck  seal which has  been  shown  to be
effective  is  the mechanical type seal.[14,15]   The mechanical
type seal  (see  Figure 6) is  basically  an  elastomeric  device
which  forms  a  close  connection with the inserted  fuel  nozzle
and  thereby  eliminates  any  space  in  the  fillneck  opening
through which  vapor could  escape.   While  both the  liquid and
mechanical  type  seals  perform  the  same  basic  function  of
limiting  the available  vapor  exits,  the liquid type seal  is
inherently a simpler design.

-------
        -20-




     Figure  5
j-Tube Liouid Seal

-------
                        -21-

                     Figure 6


                    Mechanical Seal
   F1LPPE MODIFICATIONS
   ROTARY SEAL
                    ROTARY SEAL-
   TRAP DOOR
           LEAD RESTRJCTOR
FILL PIPE  MODIFICATIONS
ROTARY SEAL
                 ROTARY SEAL
TRAPDOOR
        LEADRESTWCTOR
                                          SPOUT

-------
                              -22-
     If  a mechanical  type seal  were used,  excessive  pressure
could build  in the fuel tank if  the  fuel  nozzle automatic shut
off  mechanism failed, or  if  for some unusual  reason the vapor
line leading  to the  charcoal  canister became  blocked.   To avoid
the  possibility of  a fuel  spitback  which could be caused  by
this  overpressure,  a  simple pressure  relief  device would  be
needed.  More detail on this device will be provided in Section
IV.

     Therefore,  either type  of  sealing mechanism  - liquid  or
mechanical  - can  be used  to prevent the escape  of  refueling
vapors  to  the  atmosphere  via  the  fillneck.   Both  sealing
approaches   have   been  tested   and   provide  similar   control
efficiencies.[14,15]

     b.    Modifications to the Present System

     The differences in the  frequency,  magnitude,  and  rate  of
generation of  refueling  and diurnal evaporative emissions leads
to   the   need   for   several   modifications   to   the   present
evaporative system.  Each of these is discussed below.

     (1)    Charcoal Canister Size

     Generally  speaking,   on  a  per  event   basis,  refueling
emissions  are  produced  less  frequently  but  are  larger  in
magnitude  than  diurnal  evaporative  emissions.   Consequently,
more hydrocarbon storage  capacity (larger  charcoal  canister)  is
needed  to  control  refueling  emissions  than  is  needed  for
evaporative emissions.

     For   any given  vehicle,  the  size  of  the  canister  needed
depends  primarily on  the  fuel tank volume  and the  refueling
emission  rate.   The  refueling  emission  rate  is  chiefly  a
function   of  the  fuel   volatility  (RVP),   dispensed   fuel
temperature,  and  the temperature of  the fuel  in  the vehicle's
tank  prior   to   refill.    For   canister  design  purposes  the
temperatures  and  fuel  volatility  specified  in  EPA's  draft
refueling  emission test  procedure  would  be  used  to  determine
the design  emission rate which the canister   would need  to  be
able to  capture.   Canister sizing  would then  be  a function  of
tank   volume,   the   design    emission   rate,   as   well   as
considerations  for safety  and  deterioration   factors to assure
an adequate working capacity over the life of  the vehicle.

     The   size  of  the   canister  needed  for   an  integrated
refueling/evaporative   control    system   cannot   be    stated
categorically  since there  are  several  other  variables  which
must  be  considered  such   as   purge  rate,   charcoal   working
capacity,  and canister  geometry.  However,  on  average  it  is
expected     that    a     canister    for     an     integrated
refueling/evaporative system  would  be approximately 3 times  as
large as  the one  used  for  the  present  evaporative system. [17]

-------
                              -23-
While  the  larger  canister  does  not  present   any  technical
problems  it  may  cause  packaging  problems  on  a  few  smaller
vehicle models  which  could  lead to  canisters being  placed  in
locations other  than under the  vehicle  hood.   While  virtually
all evaporative emission system  canisters  are  now located under
the vehicle hood  there is  nothing inherent in the design  of  an
onboard  system  which requires  that  canisters  for  integrated
systems also be located there.   In fact,  there may be  some cost
advantage to locating  the  canister  near  the fuel  tank since the
amount of larger  vapor lines  can be minimized.  It  is expected
that  manufacturers  would  place  canisters   in  a  location  which
provides  the  optimum mix of  safety,  cost,   and  performance
characteristics.

      (2)   Refueling Vent Line Modifications

     Also, in  order  to accommodate the higher vapor flow rates
associated with refueling emissions, a larger  vent  line between
the  fuel  tank  and  charcoal  canister  is   needed  along with  a
larger opening  in the  top  of the  fuel tank to accommodate the
larger  vent  line.    The  current  vent  line  to  the  canister
associated with the  evaporative  system is  about  3/8  inch.   The
vent  line  with  the  integrated  evaporative/refueling  system
would be  approximately  1/2  - 5/8  inch in diameter.[ 16]   The
larger  vent  line (and larger opening in   the  top  of  the fuel
tank) introduce a few added complexities.

     Unlike the  limiting orifice used in   evaporative emission
systems,  the  larger  opening  required  for an  onboard  system
cannot provide  liquid/vapor  separation or  rollover  protection.
Consequently,  additional  devices  are  required  on  an  onboard
system  to  meet  these  needs.    The   liquid/vapor   separator,
examples  of  which are shown  in  Figures 7  and 8,  is  simple  in
design and purpose.[14,18]   It acts to remove  gasoline droplets
from  the  vapor stream and  returns  the liquid to  the fuel tank
to   prevent   liquid   gasoline   from   entering   the   charcoal
canister.   Many  design approaches  are available  in  addition to
those  shown  here.   The  separator  itself may  be  a  distinct
component, or  its function may be  built into  another component
such  as  shown  later  in  Figure  21.    In  terms  of  rollover
protection,   several  simple  devices   are   available which  can
prevent fuel spills  during  an  accident,   and  also  provide the
benefits  of  a  limiting orifice described  above.   These will  be
discussed in more detail in Section  IV  of this document  since
rollover  and   accident  protection  for   the   fuel  system  is
primarily a safety issue.

     Aside from the  differences  discussed  above,   onboard  and
evaporative  emission  control   systems  are   very  similar  in
design.   They both act to  direct, trap, and consume  hydrocarbon
vapor.    Onboard  systems   require   only  a  few   additional
components,  and  because  they could be integrated into vehicle
fuel   systems   to   handle  both  refueling  and   evaporative

-------
                    -24-
                  Figure 7
            VAPOR-LIQUID SEPARATOR
Mounting Holes
                                       Float Weight

-------
VAPOR OUT
                                                        Figure 8

tftrJ/
— I
I
1
V^1*^
o %
T
.66
L_i
XkOEN
o

/7X»...
<4^QAM



t
SATE RETURN
to



@RNE
NYU
/


0
*-I.O— *

.1
V
O

DIAMETER
ON MESH




rf—>
Is
s.
_£— .IZ5
~I
3"*- ADI
i ^-
i *
,™i^
\
UOUD REUEF
                                                              ADHESIVE SEAL





                                                                .75 .HOSE TYR




                                                                VAPOR/UQUD IN
                                                                                     ,	
E
                                                                                       _i_i	U	1
                                                                                           2.0
                                               MOUNTING TABS <3)
                                                                                            4   AR OPCN 6KLL FOAM
                                                                                               AR
                                                                                                  MESM
                                                                                                  UPPER
                                                                                                                     PIAST»C
                                                                                                                     PlAfcTIC
                                                                                                                     MATIMAL
                                                                                                                              •MOM1.
                                                                                                  VAPOR - LIQUID SEPARATOR
                                                                                           MU1LLJH A««QCIAT1». IMC
                                                                                                                        . MO
                                                                                           >«ft1|/IB/aJ»c*i.ii FULL
                                         |>».t | or f .

-------
                              -26-
 emissions,  overall control  system complexity  is  not increased
 significantly.    Also,   because   of  the   integration   of  the
 refueling/evaporative  emission control  functions,  it should  be
 apparent  that  many  of  the safety  concerns  associated  with
 onboard  systems have  already been  considered  in  designs of the
 present   evaporative   emission   systems.    The  experience  and
 knowledge  gained   in   implementing  safe  evaporative  emission
 systems  provides a substantial  base  of  information to  use  in
 designing  and  developing  safe  integrated evaporative/refueling
 systems.

     3.    Volatility Effects                    '

     As  was  mentioned  above, the  refueling emission rate  is  a
 key  factor  in the size of the onboard  system  canister,  and the
 refueling emission  rate itself varies  with the  fuel volatility
 and  the  dispensed  and  fuel  tank temperatures.   For  design
 purposes, the canister would be  sized  based  on, the volatility
 and  temperature  specifications  prescribed  in EPA's  refueling
 emissions test  procedure.   The  parameters  prescribed  in EPA's
 procedure   are   based   on  near   worst   case  :summer   season
 conditions,  so   the onboard canister   would   have  capacity  to
 achieve control under virtually all summer conditions.

     However, as average  temperatures  decrease;  in  the winter,
 RVP  levels  increase  and dispensed and fuel  tank  temperatures
 decrease.[19]   The  question arises as to  whether  the  onboard
 canister  would  have   adequate   capacity  to  Icapture  winter
 emissions with  higher  RVP  fuels.   If  the capacity is inadequate
 canister  breakthrough  may  occur   and  some  emissions  may  be
 uncontrolled.
                                                i
     Previous studies  and analyses conducted  by  EPA and others
 have shown that the refueling  emission  rate increases  with the
 fuel volatility (RVP)  and  fuel  tank temperature  and decreases
with the dispensed fuel  temperature.[19]  One  study  (CAPE  9)
 used   volatilities   and   temperatures   typical   of   winter
 conditions.[20]    Using  winter  season  fuel  volatilities  and
temperatures  in the relationship  derived  from this study yields
winter  refueling  emission  rates   less  than  the  design  load
 emission  rate  for  the  canister  dictated  by]  the  refueling
 emissions test  procedure.   Winter  season  values  (Dec   -  Feb)
 range  from  5.1  to 5.9 g/gal for the northern  states where RVPs
 are quite high  (14-15  psi)  while the design load value  is 7.25
g/gal.   Thus winter emissions would be controlled as well.
                                                I
     EPA  is  presently  considering a  program to  control  the
volatility level (RVP) of in-use fuels  during  the summer months
 (mid-May  to  mid-September).  As  part  of  that Iprogram,  in-use
volatility levels  nationwide would be  limited to  levels about
 21.7 percent  less  than the current  ASTM  level  for that  area
 during the  affected months.  If  that  program was  enacted,  the
volatility of the  fuel for refueling  emissions testing would be

-------
                              -27-
set  at 9.0  psi  RVP,  the  design  load  emission  rate for  the
canister could  drop to 6.0  g/gal,  and onboard canisters  could
be somewhat  smaller.   However,  as  can be seen from  comparison
with   the   emission  rate   figures   presented  above,   winter
emissions would still be controlled.                         X

     While not  the  primary  motivator,  in-use  volatility control
may  have  some   attendant   safety  benefits.    Lower  RVP  fuels
generate less vapor  and  thus could be considered somewhat safer
in a general  sense.   More  specifically,  lower volatility  fuels
generate  less fuel  tank  evaporative  emissions  and  thus  could
reduce  fuel  tank pressurization  problems which  occur on  some
vehicles with damaged or  altered  evaporative emission systems
(e.g. non-standard gas caps)  operating under  extremely atypical
conditions.   This  pressurization  could lead  to  some  fuel/vapor
being  released  from  the  fillpipe when the  gas cap is removed,
especially  if the  fuel  tank was  relatively  full at  the  time.
Lower   vapor   pressure  fuel  would  reduce   the   degree   of
pressurization  which  could  occur under these circumstances  and
thus reduce  or  eliminate the spillage which may  result.   Thus
the safety of refueling operations would be improved.

IV.  Design Considerations for a Safe System

     As  was  discussed  previously,   several  commenters   have
expressed  concern regarding  the  potential safety implications
of onboard systems.   A review of  these comments  indicates that
these concerns fall into two broad  areas:   the design of  a safe
onboard  system  and  effects  on  in-use  fuel  system  safety.
Concerns in the first area  will  be addressed in  this section.
The  section  which  follows   (Section V)  will address  the  later
area of concern.

     Comments received  regarding  the  design  of  a  safe onboard
system  fall  in  three   categories:    1)   safety  test  design
reguirements,   2)   safety   effects  of   maintenance,  defects,
tampering  and  repairs,   and  3)   refueling  operation  safety.
EPA's  summary and  analysis  of the comments in each category is
presented below.

     A.    Safety Test Design Requirements

     1.    Introduction

     Before analyzing the safety  test  design reguirements it is
interesting  to   look  at  fuel  system  safety  from  an  in-use
perspective  for passenger  cars  meeting FMVSS 301.   Presently,
about  1.6  percent of all  accidents involve  a vehicle rollover
of some type and about 0.5  percent  of the  rollover accidents
result  in  a  fire. [21]  This results  in  a fire  rate  of  0.008
percent.   Thus,  neither  rollover  accidents  or  subsequent fires
are common.   Similarly, 0.14 percent  of  all front and rear  end

-------
                               -28-
 collisions  lead to a vehicle  fire.[21]   Although vehicle crash
 fires  are  seemingly  uncommon,  approximately  1,600  fatalities
 result  each year  from  these  fires.[22]   Thus,   from  an in-use
 perspective, vehicle crash  fires  are unusual but  serious events.

     One  of the most effective ways to protect against vehicle
 crash  fires is  to restrict  fuel  leakage  during  accidents  by
 insuring  the overall  integrity  of  the vehicle's  fuel system.
 To  insure fuel  system  integrity during a  crash,  all  currently
 manufactured passenger  cars and light-duty truck's with a Gross
 Vehicle Weight Rating (GVWR) of  10,000  Ibs  or  less, must comply
 with  Federal Motor Vehicle  Safety Standard  (FMVSS)   301.[23]
 Basically,  FMVSS   301  requires  a  vehicle  to  restrict  fuel
 leakage to  less than one ounce per minute when'subjected to a
 rollover  test  following  front  and  rear collisions at  30 miles
 per hour  (mph),  and side collision(s) at 20 mph.   In a rollover
 test, a vehicle is turned  on  each  of its  sides : and  completely
 upside  down and  held  in each of  these three positions  for a
 period of five  minutes.   Onboard system designs Imust  take into
 account and protect against fuel leakage  or other  fire hazards
 which could occur  in FMVSS  301 testing.          |
                                                 i
     Along  these lines,   two  issues  exist  regarding  the design
 of  an  onboard  system  capable  of  passing  FMVSS  301.   These
 include  rollover  protection  and  the   crashworthiness  of  key
 onboard  system  components  and  connections.   As was  discussed
 previously,   onboard  systems  require  the  use  ;of  a  somewhat
 larger  vent  line  (about  l/2"-5/8"  diameter  as  compared  to
 1/4"-3/8"  on  current  vehicles)  between  the  ,fuel  tank  and
 charcoal  canister,  and  a  similar  sized orifice  would  exist  in
 the  fuel  tank.   While  the external  vent  line! used  on  many
 current fuel tanks also  requires a 1/2" orifice,; manufacturers'
 onboard system  designs may incorporate  a  rollpver  protection
 device  to  protect against  fuel leakage   during  an  FMVSS  301
 rollover  test  even  though  present  designs  do  not.   Also,
vehicle crashes present  the  possibility of direct or  indirect
damage to fuel  system  components.   In  some cases this  damage
 could  lead  to  a  fuel  leak  or  increase  the  potential  for  a
vehicle fire in  some  other portion of the  fuel ^ystem.   Thus a
properly  designed  onboard  system  must  not !compromise  the
crashworthiness of the system and key components.

     2.     Rollover Protection                  ;

     A  rollover  protection  device   is  basically  a valve  that
would close off  the  refueling vent  line whenever the risk  of
 fuel leakage existed.  Several  rollover  protection  designs have
been proposed  by  auto  manufacturers and other ! interests which
could  adequately  perform  this  safety  function.   Several  of
these are discussed below.

-------
                              -29-
     One design which  has  been proposed by several  sources  can
be  termed  the nozzle actuated valve.  The valve  is  integral  to
the  fillpipe  and  is  located near  the  top  of  the  fillpipe,
perhaps near  the  leaded  fuel restrictor.   During refueling,  the
valve  is  opened by  the  insertion of a  fuel  nozzle.  With  the
valve  open,  a clear passage through the  vent  line  is available
to  allow  for the  routing  of refueling vapors  to the  charcoal
canister.   Other  than  during  refueling,  the  valve  remains
closed  and   effectively   eliminates  the potential  for   fuel
leakage  through  the  refueling  vent   line  during  a  rollover
accident.   Figures 9  thorough  15  show  five  different  nozzle
actuated  valve assemblies capable  of  performing the  rollover
protection  function.[13,15,18,24]   Figures  9  through  13  also
demonstrate   how  nozzle   insertion  would  open  the  valve  to
provide  a  large  orifice   for  the  venting of  fuel  tank  vapors
during  refueling  and when the  nozzle  is  removed  the vent line
would be closed.

     Also,   while   a  rollover   protection   device  might   be
necessary,   it   is  interesting  to  note that   many  current
production  passenger car  and light truck models  (mostly side
fill)  employ  an external  vapor vent line of about 1/2" diameter
that connects the fuel  tank to  the  top  of  the  fillpipe (see
Figure  1).    This  external  vent line  is approaching the size
needed  for  a refueling vent line,   and yet manufacturers have
included  these   external  vent  lines   without   any  rollover
protection  device.   As will be  discussed below, depending  on
the  design  used,   a  rollover  protection  system may  actually
enhance safety over current designs.

     This analysis  has presented several basic  rollover  valve
designs  capable  of  providing  the protection reguired by  FMVSS
301  tests.    Manufacturers could  choose  to  implement  one  of
these  approaches,  or  could  develop  another.   The  approach
ultimately selected  will  be  that which provides  cost efficient
protection,   is  compatible  with the  other  components  of  the
manufacturers onboard  system,  and can  be integrated effectively
into  the  vehicle  design  from  both  safety  and  operational
perspectives.

     3.    Component/System Crashworthiness

     The second issue  regarding  safety  test  design requirements
involves  the  crashworthiness  of  the  key  components  of  an
onboard   system.     This   includes   those   components   most
susceptible  to  damage in  an accident  (nozzle  actuated rollover
valve,   charcoal  canister)  and the  structural  integrity of  the
vapor  line  (and connections)  which  may exist between the top of
the  fuel  tank  and the  rollover valve.   A  problem  in  one  of
these  three  areas  could  cause a vehicle to fail FMVSS 301 tests
and must be  addressed in  proper  system design.  Each of  these
concerns is discussed below.

-------
                       Figure 9
       SEALED FILLER NECK  SYSTEM
        TANK VENT VALVE ASSEMBLY
       (DURING NORMAL VEHICLE OPERATION)
     TO CANISTER -
                           LIQUID STOP
SHUTOFF WU.VE
                                     SEAL
                             LEADED FUEL
                              DEFLECTOR
                                                GAS CAP
                                                 OVERFILL
                                                RELIEF VALVE

-------
          Figure 10

SEALED FILLER WECK SYSTEM
TANK VENT VALVE ASSEMBLY
      (DURING REFUEUNQ EVENT)
TO CANISTER ~
                 LIQUID STOP
                    r- LEADED FUEL
                      DEFLECTOR
                                                            oo
                                                            r
                                     OVERFILL
                                    RELIEF VALVE
                            SEAL

                           FUEL NOZZLE

-------
               Figure 11
           LIQUID SEAL SYSTEM

      TANK VENT VALVE ASSEMBLY

      (DURING NORMAL VEHICLE OPERATION)
TO CANISTER -
                       SHUT-OFF VALVE
                             LEADED FUEL

                             DEFLECTOR
                                               GAS CAP
                                                                      I
                                                                      U)

-------
                  Figure 12

          LIQUID SEAL SYSTEM
     TANK VENT VALVE ASSEMBLY
          (DURING REFUEUNQ EVENT)
TO CANISTER -
  VAPOR
  FROM TANK
                   SHUT OFF
                    VALVE
                          LEADED FUEL
                           DEFLECTOR
                                  .0
                                                                      u>
                                                                      V
                                 FUEL NOZZLE

-------
                            Figure 13

NOZZLE-ACTUATED REFUELING EMISSIONS VAPOR VENT VALVE
        Vapor to
        Canister
                         Vapor from
                     Vapor/Liquid Separator
                                                                               i
                                                                               U)

-------
                                              Figure  14
NOTES:

I. REMOVE ALL BURRS < SHARP EOGCV
2. NYLOH PARTS MA* BE FABRICATED
  f^AVTlC COMPATIBLE W/ 6ASOUNE ^ METHANOL.
3. AIL DIMENSION* NOT
      ftrrn I  I  *
dUfi^
                                                              '	.O* TYU
                                                                ALL OVER
                                                                                     .19
                                                                             io | I | RIVET WASHER PIATT| s. &TCEU
                                                                                  CAM SPACER
                                                                                  CAM P%M
                                                                                  LAMIMA.TED.
                                                                                  CAM FOU.OWET*.
                                                                                  CAM
                                                                                       BALL
                                                                                       \IALVC
                                                                                  LOWER VA1VE.
                                                                                     O(«CtU»TION
                                                                                                 ALUM.
                                                                                                 T. »TEEL
                                                                                                 ALUM.
                                                                                                MYLON
                                                                                                 KYLOM
                                                                                                 MYLOM
                                                                                                         314
                                                                                                        60&I-T6
                                                                                REFUELING VAPOR VENT VALVE

                                                                                 	ASSEMBLY	
                                                                             All
                                     aSQCIATtS. IMC

                                     SCALE: FOll- |
                                                                                                     HO

-------
                    -36-
                Figure 15
            "Toyota Concept"
         Rubber Seal
          Valve
 to Canister
from Fuel
  Tank
       x"
to Fuel  Tank
Refueling
   Nozzle

-------
                              -37-
     However, before  beginning  these discussions,  it  should  be
noted that component/system crashworthiness is not  at  all  a  new
concern.  Manufacturers must address  these same  concerns in  the
design of the current evaporative  emission systems.-  Given  the
similarity  of  onboard  refueling  and evaporative controls,  and
that many systems will be integrated, there should  be  no new  or
unigue problems in this area.

     a.    Rollover Valve

     First,   the  crashworthiness  of  the  rollover  protection
device  is  a design  consideration for  nozzle actuated  valves,
since  they  would  be  located  near  the  exterior  shell   of  the
vehicle.   Integration  of  nozzle   actuated  valves  into  the
overall vehicle design  would have  to include a  consideration of
the potential to sustain damage if struck in a collision.

     However, this design  consideration  is straightforward,  and
it  is  reasonable  to  expect  that  manufacturers can  and  will
integrate rollover valves  into their  fillpipe   designs  without
decreasing  the  structural  integrity  of  the   fillpipe  while
providing crashworthiness  for  the  valve.  For  example,  it  is
worth noting that  vehicle  manufacturers  have dealt  with  similar
problems  in their designs of  fillpipes,  external vapor  vent
lines,  and  gas  caps,  and in fact,  one  would  not expect  the
nozzle  actuated  rollover  valve to  be any more  susceptible  to
damage than these  components.  As  was mentioned previously,  the
1/2" external vent line lies  in  this same  area  on  the vehicle,
and yet  manufacturers have included  such vent   lines  without a
rollover protection device.

     b.    Vapor Line

     Similarly,  manufacturers will have  to be cognizant  of  the
structural   integrity  of  the   vapor   line  and   vapor  line
connections, if  any,  between  the fuel  tank and  the  rollover
valve.   These  would  have  to  be   designed  to  withstand  the
stresses which might occur in a crash in  order  to maintain fuel
system  integrity.  However, there  is no  significant engineering
challenge to accomplishing this objective.

     The integrity of this portion  of the  vehicle's vapor line
can be  assured  through use  of  a vapor  line  material  of proper
strength, flexibility, and durability.  A  number  of vapor lines
of  different material, wall  thickness,  and construction  are
currently available.   In  addition,   routing of  this portion  of
vapor   line   is  another  design   parameter   available   to
manufacturers.    As   a  matter  of   course,  manufacturers  are
expected to  insure that the line is protected from abrasion and
normal wear and  that  it is not in a  vulnerable  location in  the
event of a  collision.   This is considered straightforward given
that  on  integrated  systems  the   refueling vapor  line  now
replaces  that   used  for  control   of   diurnal   evaporative

-------
                              -38-
 emissions.   Similar  routings  would be  expected.   Vapor  line
 integrity and connections  in  current vehicles must  meet similar
 requirements,  and   it  is  reasonable  to  expect  that  similar
 materials and connecting approaches would be used.

     Finally   with   regard  to   vapor   line   integrity   and
 connections,  it   is  worth  noting  that  many  vehicle  models now
 use  a flexible  insert between  the fillpipe  and  fuel  tank  to
 enhance  the  fuel system  safety  in-use  (see  Figure  16).[15]
 Similarly,  in  many  vehicle  models  the  external  vent  line
 actually  incorporates a flexible  vapor  line  which  connects the
 metal  portions  of the  external  vent line  from the top  of the
 fuel  fillpipe  and   the  fuel  tank  (see  Figure  16).   These
 connections are subject to  the  same performance requirements as
 would  be needed  for  onboard  system  vapor  lines  and  in some
 cases  are  even more  critical  and demanding.    Evidence  is that
 these  have   been   incorporated   safely.   The ; manufacturers'
 experience  with   current  vehicle evaporative  and  fuel  systems
 described  above   demonstrates  that  vapor  line and vapor  line
 connections can  be made to withstand the  stresses which occur
 in a vehicle accident.                           i

     c.    Charcoal Canister                     I

     Concerns  regarding  the crashworthiness  of  the  charcoal
 canister center on the possibility that a  canister ruptured  in
 an accident  could present  a  fire hazard  if  an ^ignition source
 exists nearby.

     Even   if   the   rupture   of   the    integrated   refueling/
 evaporative  canister  occurred  in  some  cases,! the  potential
hazard should  not be  overstated.   While  carbon  canisters  do
 contain  gasoline  vapor, they are  strongly adsorbed   to  active
 sites  within  the  carbon  bed  and not  easily  released to  the
 atmosphere.    Thus,   even  if  a   canister  were  crushed and  its
contents dumped,  gasoline  vapor would not be  present in  the
 atmosphere in sufficient quantity to be flammable.   There is no
 available  evidence   of  "canister    fires"  in  any   accidents
 involving  vehicles  with  evaporative  systems.   ' The   fact  that
onboard canisters would be  larger  and would hold more  vapors
 initially   than   current    evaporative    systems    makes   no
difference.   While the refueling  load to  the  canister  is larger
than the evaporative  load,  after the first few miles  of driving
the  canister  would  be purged  such that  the  amount   of  vapor
remaining  in  the canister  is  essentially  the same as  that
present in current evaporative  emission  canisters  alone.*  The
*    Due to the  nature of the charcoal used to trap hydrocarbon
     vapors,   and   strict   certification  test   requirements,
     hydrocarbons  would be  quickly  stripped from  the charcoal
     early in the purge process.  Therefore,  during  most  of the
     operation  of  the  vehicle   (90   percent),   the  charcoal
     canister  does  not  contain  enough  hydrocarbon  vapor  to
     present any safety risks.[9]

-------
                    Figure 16
BUICK CENTURY FUEL TANK AND FILLPIPE
       PRODUCTION CONFIGURATION
                                                      Nozzle
                                                      Spout
     Fuel Sending Unit -^
      and Vent Orifice  \^
12%"
I
?

-------
                              -40-
 lack  of risk from charcoal  canisters  is supported  by  a recent
 submission  from Nissan to EPA,  stating  that  no safety problems
 would be expected with refueling canisters.[25]  Thus  it could
 be  argued   that  the  hazard,   if  any,  is  not  significantly
 different than  that  now found on present systems.   Thus,  it is
 hard  to  perceive  any  added risk  from the  use  of  a  larger
 charcoal canister.

      Nevertheless, if  a  manufacturer believed  that  the canister
 posed a potential  risk,  the  risk  could be  eliminated through
 placement of the  canister  in a protected area  such  as  the  rear
 of the  engine compartment  or in some underbody area as has been
 suggested by some manufacturers.[12,13]   In  most  cases it  is
 expected  that manufacturers  would  simply place  the integrated
 refueling/evaporative  canister  where  the  present   canister  is
 now located;  in these cases no new design issues really exist.

      d.    Summary                               ;

      In  summary,  current fuel and  evaporative emission systems
 must  meet  the  same  FMVSS  301  requirements  and  much  of  the
 experience gained in  designing and  building current systems can
 be  directly  extrapolated  to  implementing  an  onboard  system.
 The   analysis  presented  above  leads  to  the  conclusion  that
 straightforward, viable  engineering solutions  exist  to address
 any potential safety design  concerns, and  that onboard systems
 can be  incorporated  into the  vehicle's  fuel/evaporative system
 without  compromising  fuel  system   integrity   or  reducing  the
 vehicle's ability to pass FMVSS 301 requirements.

     While  an onboard system  can  be designed  to provide  fuel
 system  integrity  both  in  FMVSS 301 testing  and! in-use, it  is
 prudent  to   consider  the   effects  of   maintenance,   defects,
 tampering,  and  repairs on these  systems,  and me;ans  to address
 any potential problems which  may exist.   These Issues  will  be
 addressed next.                                  ;

     B.    Maintenance, Defects,  Tampering and Repairs

     Even if  a  system is designed properly and functions safely
under  "normal"  and  "extreme"  in-use  conditions,  some  question
 remains  as  to  the  potential  effects  of maintenance,  defects,
tampering and repairs on onboard system safety.  '•.

     Maintenance  is  the prescribed actions  needed  to keep  a
 system  operating  as  designed.   Defects  involve  the  improper
 operation of  the  system  or system components caused  by design,
manufacturing,  or  assembly  errors.   Tampering  involves  the
 intentional   disablement  (partial or  total) or  removal  of  the
 system  or  a  component  within the  system,  and  repairs involve
 restoring or  replacing the system or system  components because
 of  malfunction  or  damage.    Each  of  these  events and  their
 safety effects are discussed below.

-------
                              -41-
     1.    Maintenance

     First,  an onboard  system  is  expected  to be  essentially
maintenance  free   (no scheduled  maintenance)  as  are  current
evaporative control systems.  EPA's emission  factor  testing has
found that non-tampered  fuel-injected vehicles  generally comply
with  the  evaporative  emission  standards  without  maintenance.
Furthermore,    EPA's   requirements   for   light-duty  truck   and
heavy-duty  gasoline  vehicle  emissions  certification  do  not
allow evaporative system maintenance up to 100,000 miles,  and a
similar  requirement  is  being considered for  ah onboard system.
The  technology used  here  can  be  used  for   passenger  cars  as
well.   Thus,  maintenance  will not  be  necessary  for  proper
functioning of  an onboard system over  the life  of a  vehicle.
Therefore,  lack  of   prescribed maintenance  will not  lead  to
safety problems.

     2.    Defects

     Second,  with regard to defects, the primary safety related
concern   deals  with   the  possibility  that   defects   in  the
operation  of  one  or  more  components  of  the  onboard  system
in-use  might  lead  to safety  problems  for  the vehicle.   This
includes   possible  problems   with  components   such   as   the
liquid/vapor  separator,   purge  valve,   charcoal  canister  and
rollover valve.

     Since onboard  system components  such  as  the  liquid/vapor
separator, purge  valve,  and charcoal canister  are  very similar
to those  used  in  evaporative  systems,  one method to assess the
potential  safety  effects  of  defects  is  to review the experience
seen with evaporative systems.   In  an effort  to quantify the
potential  for  defect  problems  regarding onboard systems,  three
different  computer  files  provided  by NHTSA  were reviewed for
evidence  as  to defects  pertaining  to the evaporative  emission
system which could impact  vehicle safety  in-use.[26]   The files
reviewed  covered  recalls,  service  bulletin  reports,  and owner
complaints current  as of  November, 1986  for all three vehicle
classes  (passenger  car,  light  truck, and  heavy-duty gasoline).
A review  of  the recall  files revealed only 12  cases that  could
be even  remotely  linked to the evaporative  emission system out
of an  estimated 3,000 families which  have been  certified with
evaporative  emission   systems.   Service  bulletin  reports  for
dealers added an additional 21 cases for a total  of  33  possible
problems  out  of   over  3,000  families.   None of  these  were
identified as having caused an  accident; the  vast majority were
more emission  system  performance than  safety defects.   Finally,
a  review  of  the  owner   complaints  indicated  only  about  100
problems  out of over  180 million vehicles sold with evaporative
emission  controls.  In only a  few  of  the owner  complaints did
safety  problems actually  occur, and  no significant damage was
reported.  On  a percentage basis these  potential problems are
very small.

-------
                              -42-
      Two  other valuable observations  can  be  drawn; from a review
 of  these  files.   Problems/complaints have diminished with newer
 model  year   vehicles   with   evaporative   controls,   which
 demonstrates   that   gaining   experience   leads   to   product
 improvement.   Given  the  similarity  between  onboard  refueling
 and evaporative emission  controls,   and  thQ fact  that  the two
 systems   will  be  integrated   in  most   cases,   -much  of  this
 experience will  be directly transferable to onboard systems and
 thus  improve in-use  performance.    Second,  the  ireview  of  the
 owner complaints  files  indicated  no trends  other  than those
 related to  improvements in  newer  model  year vehicles;  thus  no
 systematic problems in components or  systems were I evident.

      Further,  it is  important  to note that  the  very mechanisms
 used  to generate the files  for this survey would  actually act
 to  help  eliminate   any  potential   in-use  safety  effects  of
 onboard systems  defects.   Dealer  service bulletin  reports  are
 effective  in dealing with  problems   raised  at the  dealerships,
 and  owner complaints assist  the  manufacturers I and  NHTSA  in
 assessing  the need to  conduct voluntary or  mandatory recalls.
 Finally,  to  place  the  potential   for  defect   problems  from
 onboard systems  in context,  it should be noted that the onboard
 risk   is   essentially   incremental   to  that   now   seen   for
 evaporative  systems,   since  in most cases  the  refueling  and
 evaporative  systems  would  be  integrated.   On  an  incremental
 basis, the frequency of defects would likely be unaffected.

      Finally,  since  a  rollover valve  could  be; used on  some
 onboard system designs  specifically  to  enhance safety  and  they
 are  not  used  on current vehicles,  it  is worth ;discussing  the
 possibility  of  valve  defects.  First,  it should be  noted  that
 defects   in   these   valves   should  be   rare.  ;  Manufacturing
 engineering techniques permit  the development  and production of
 highly   reliable  valves   and  statistical  quality   control
 techniques are available to  insure  that  production  valves  meet
 design standards.  In fact,  if a rollover valve  is defective at
 the vehicle  assembly  point,  the  vehicle will  probably  not  be
 able  to  accept  the  fuel provided  at the  end of  the assembly
 line, and repairs will be needed even before the  vehicle leaves
 the   plant.   Second,   to  insure  in-use  protection,  rollover
valves must  be  designed to  fail in  the  closed position.   This
would  be  considered  "safe"   because a  closed  position  valve
 failure would never cease providing  rollover protection  and  it
would  effectively  block  the  refueling  vent  line  and  make
 refueling  the  vehicle  extremely  difficult.   This  difficulty
would  provide incentive for  the vehicle operator  to  identify
 and repair the failure.  If  the  valve failed during operation
 of the vehicle,  the  fuel  tank would vent any vapors through the
 limiting  orifice  or  gas cap  to prevent  any pressure build  up
 (See  Figures  3 and 4).  Also, rollover  valve failure might  be
 one component of  an  onboard system which could be  incorporated
 into  onboard vehicle diagnostics  and thus  allow  the  operator
notice of the problem when it  occurs  and  provide  an opportunity

-------
                              -43-
for  repair  before the fuel  level  becomes critical.  Fail  safe
designs  would be  effective  in  achieving  both protection  and
repair,  and  that the  other  measures  discussed  above  would
assist in eliminating or addressing any in-use defects.

     3.    Tampering

     A  third area  of  potential  safety  problems  involves  the
effects  of  possible  system tampering.  While several types  of
tampering  occur  with  evaporative emission  systems  (see  Table
l),  past  in-use  experience with  these systems shows that  only
one   type,   disconnection   and/or   removal  of  the  charcoal
canister, might be a  safety problem  for onboard  systems.   This
type of  tampering poses  a possible safety hazard because during
the  refueling operation  it  would lead  to a flow of  gasoline
vapor  into  the  atmosphere  at   the   point   where  the  missing
canister  had been located.   While the  gasoline vapor  mixture
reaching  the canister  location in this situation would  be well
above  the  upper  flammability  limit,   it   would   briefly  be
flammable   as  the   vapor  dissipates  and   at  the  air/vapor
transition  zones.   If  a  spark or  other  ignition  source  were
present,  the mixture  could briefly burn.   While  this situation
is  likely to be  rare,  the possible safety effects of  such  an
occurrence must be considered in the onboard system design.

     There  are several points  which  need to be made relative to
canister  tampering.    First,   this  is  not  unique  to  onboard
systems -  similar potential  problems  now exist with evaporative
emission  canisters  but  a  safety concern  regarding  tampering
with  evaporative emission  system canisters  has not surfaced.
Second,  using current  evaporative  emission canisters  as  an
indicator,  this  situation is  likely  to  be rare  for  integrated
onboard refueling/evaporative  canisters.   As  is shown  in  Table
2,  current  average  canister tampering  is only about  3  percent
of  all  vehicles,  and  similar  rate  would  be  expected  for
integrated  refueling/evaporative   emissions  canisters.   Third,
if  the  canister  were  located  in  an  area  which  would  be
difficult to access,  tampering could be further discouraged.

     Further,  the potential  problem  could  be reduced  through
proper placement of the canister in a  location  distant  from any
ignition  sources.   Possible locations  include the rear  of the
engine compartment (as is done with  some  evaporative canisters)
or  in  some  underbody  area  as   has  been   suggested  by  some
manufacturers for packaging reasons.   Placing  the  canister  in
an underbody  area would  also reduce  the potential for tampering
by making  it  less accessible  to  the  owner as  mentioned above.
While  canister  tampering  is  infreguent,  and means exist  to
discourage such actions even further,  good engineering judgment
dictates  that canisters   not  be  placed  in  a location  where
tampering  could  create a  safety  hazard.    It is expected  that
manufacturers  will  take  all  reasonable  steps  necessary  to
reduce  tampering,  and  that refueling  canisters would  not  be
placed  in locations  where their  removal  could create  a safety
risk.

-------
                              -44-
                             Table 1
                 .  Types of Tampering Problems
                 Arid Typical Rates of Occurrence
                                                Rate of
Problem                                      Occurrence (%)
Gas Cap Removed-  .   .                              1.2%
Canister Vacuum Disconnected                   .   1.7
Cap Removed & Canister Vacuum Disconnected        0.1
Canister Removed                                  0.3
Non-vacuum Canister Disconnection                 0.2
Total Disablements                                3.5%
     Tampering rates calculated  from the combined data from the
     EPA  Tampering  surveys  performed  in  1982, : 1983  and  1984
     (9,142 vehicles).

-------
                -45-

               Table 2

  Canister Tampering Survey Results
               By Year*
   Passenger Car and Light Truck**

 •Year  .                   % Tampered
 1978
 1979
 1980
 1981
 1982
 1983
 1984
 1985
Avg
                                           3
                                           2
                                           No Report
                                           2
                                           2
                                           5
                                           3
                                           4
                                 US EPA,  OAR,  QMS
*    Motor Vehicle  Tampering  Survey -  1985
     FOSD, November 1986.
**   Since  HDGVs  did not  reguire  evaporative  controls
     1985,  survey  data  is  currently  not  available  for
     vehicles.
                                              until
                                              these

-------
                              -46-
      4.    Repairs

      Finally,  repairs of  onboard systems may have  some safety
 implications.    Since   an   onboard   system   is   essentially
 maintenance  free,  any damage  to  the system (besides  that  from
 defects or tampering) would in most  cases  result  Ifrom a vehicle
 accident.  An accident which  damages  the vehicle's  fuel system
 would be  relatively  severe  and  require   critical  vehicle
 repairs.   Such  vehicle  repairs,  in  general, would  demand  a
 professional  certified mechanic  in a  licensed  facility.   These
 mechanics should be  properly  trained and have access to current
 shop  manuals  to  repair  and package the fuel system  and onboard
 components correctly to ensure effective  and  safe performance.
 They  also  should be  aware of any  potential  safety  hazards  of
 improper   installation    or   omission   of    onboard   system
 components.   Furthermore,  these  mechanics  would  normally  have
 no  economic  incentive  for  improperly  repairing  an  onboard
 system or omitting  some  components since the facility  would  be
 compensated  for  all of the parts and  time  spent repairing the
 vehicle.                                          !

      In any  repairs  of  the fuel   system with an onboard control
 system,  there  is  only  one  critical  area  with  respect  to
 safety.  This critical  area  is the connecting line  between the
 top of  the  fuel tank  and the rollover valve at 'the top of the
 fillpipe.   An improper  installation or  connection in  this  area
 could  result with  fuel   leakage  in  the  eventi  of a  vehicle
 rollover.   This  connection,  however,   is not  unique  to  fuel
 tanks with onboard  systems.   It  is very similar jto the external
vapor vent line  that appears on  many  of today's  vehicles,  and
 thus  incrementally  the situation may  be  no different  than  on
 today's vehicles.  Thus,  repairs  of onboard systems  should not
 create any potential safety hazards as compared  to  present day
 fuel  systems.                                    I
                                                 i
      5.    Summary                               ,

      In summary,  component maintenance, defects,  tampering,  or
 repairs  should  not  create  the   potential  for' in-use  safety
risks.   An  onboard  system   is   expected  not  to  require  any
scheduled maintenance.   Thus, any lack  of  maintenance by the
vehicle owner should not introduce safety hazards.

     There is no evidence  to  indicate  that possible  defects  in
other onboard system components  would  lead  to  safety problems.
There  are  very  few  defects  with present evaporative  emission
systems, and  since  it is  likely  that  refueling ;and  evaporative
emission systems would  be integrated,  the  overall  defect  rate
 is  likely  to  be  no  different  than  that  seen  in  present
vehicles.    Further,  methods  are   available   to  assure  that
reliable  rollover  valves  are  installed in  vehicles  and  to
 insure  rollover  protection in  the  unlikely  event  of  a  valve
failure.

-------
                              -47-
     While  canister  tampering  effects must  be considered,  it
should  be  noted that  it presently  is uncommon,  and this  low
rate  is  expected  to  continue  for  onboard  systems.    Also,
tampering  could actually  decrease  through  judicious  canister
placement   on   the  vehicle.    Nevertheless,   prudent   design
practices dictate  that  manufacturers not place  canisters in  a
location where  tampering could  lead  to a safety problem,  and it
is expected that this approach would be followed.

     Any repairs of  an  onboard  system, besides those  resulting
from defects or  tampering, will  probably occur as  a consequence
of accident damage to the  vehicle.   Since the damage will  most
likely  be  severe,   it  will  require  the use of  a  certified
mechanic who  is properly  trained for  such  repairs.   Further,
the only critical  area  of  the onboard system which could impose
any safety hazard if improperly repaired  are the components  and
connections between  the fuel  tank  and  fillpipe top.   Repairs
are  also critical  in  this   area  for  current  vehicles  using
external vapor  vent lines,  so  there may be  no change  in  risk
over present vehicles.   Repairs to an  onboard  system  should  not
inherently increase the potential for in-use  safety risks.

     An  onboard system  design  must also include  consideration
of potential  effects on the  safety  of  refueling  operations.
This is discussed in the next section.

     C.    Refueling Operation Safety

     1.    Fuel Tank Overpressure During Refueling

     The first  potential safety issue  involves  the possibility
of pressure build-up in the fuel tank during the refueling  of a
vehicle  equipped with  an onboard system.  Whenever a  system is
designed to be  "sealed"  from its environment,  some forethought
must be exercised  to evaluate the possibility and consequences
of an overpressure within the system.

     Although an onboard system  does not completely seal  off a
vehicle's  fuel  tank,   it  is  designed to  allow  for  only  one
opening, the  refueling  vapor vent  line.  If  for   some  unusual
reason, the vent line were to become fully or partially blocked
or the nozzle automatic shut-off mechanism failed  during  a full
refill,  excess  pressure  could build  in the  fuel tank.   This
concern  is  only associated with  an  onboard  system utilizing a
mechanical seal  as  illustrated  in  Figure 3.   With  a liquid seal
system  (see Figure  4),  excess pressure  cannot build up  in  the
tank during refueling  because  fuel would  simply  flow  out  the
fillneck  opening  (the  same  way  it  currently  does)  and  the
nozzle  operator could  then  stop  the  fuel  flow.   Liquid  seal
systems  would  function  in  the  same  manner  as  current  fuel
systems.  From the  nozzle  operator's viewpoint,  the  refueling
operation remains the same.

-------
                              -48-
      I£  a manufacturer  Qlocts thQ  mechanical  seal  design,  he
 must  incorporate a  simplQ  pressure relief  device  capable  of
 relieving  fuel  tank  pressure.   In the event of a noaale failure
 or  vgnt line  blockage,  this  device would  eliminate potential
 tank  overpressurization by opening  an  "emergency"  passage  to
 the atmosphere through which pressurized vapor  and any gasoline
 would  spill  onto the pavement  or some other location noticeable
 to the nozzle operator.  This  spillage would make1 the fuel pump
 operator  aware  of  the problem  and  fuel  flow  could  be stopped
 without  causing  damage to  the fuel  system  or  causing  fuel  to
 spitback on to the operator.                      ;

     There  have  been  several  different  designs, suggested for
 such  pressure relief  devices.   A  sample  design  is  shown  in
 Figure  17  which would be  incorporated directly into  the design
 of the  fillpipe so  that  the condition would  be, noticeable  by
 the  operator.[18]   The  operator would  then  be  prompted  to
 repair  any  problems  in   order   to  resume  normal   refueling
 actions.   (The  need  for  prompt repair  would  have  positive
 safety  and  air quality  implications.)  As  was shown  in Figure
 9, it might  also be  possible to incorporate the pressure relief
 function into  some  other  component  of  the  system such as the
 rollover valve.   Any  overpressure  concerns  can;  be  eliminated
 through a simple pressure relief device such as these.
                                                 I
     2.    Pre-Refueling Overpressure Effects    ;
                                                 i
     Another   potential    safety   issue   raised;   relating  to
 refueling  operations  has  to  do with  the "U"   bend  in  the
 "J-tube" fillneck seal.   If the  tank vent became  blocked, and
 pressure built  up substantially  in  the  tank,  upon  removal  of
 the fuel cap,  the liquid  gasoline which was  le|ft standing  in
 the "U" bend could be spit back out the fillpipe.

     This concern can  be  easily  addressed  by drilling  a  small
 hole in the bottom of  the  "U"  bend  (see Figure 4  and 5),  which
 would allow  any  fuel  left  standing  in the  fillpipe  subsequent
 to a refueling event to  drain out into  the fuel  tank.    Given
 the range  of  fuel dispensing  rates  seen in-use,  this  hole can
be sized to  quickly provide  drain  capacity  and  still  provide
 the seal  needed during  refueling.   Furthermore,, the hole size
 can be sized  so  that no  foreign object will  block it  during  a
 refueling  event.    By   evacuating   the  column i of  fuel   left
 standing in  the  fillpipe,   the  potential  for spitback  to  occur
upon  removal  of  the   fuel   tank  cap would ,be  eliminated.
Fillpipes with a "J-tube" seal  employing a drain  hole have been
 tested.  These  tests  show that  these  seals provide  refueling
 emission control efficiencies comparable to  those of  mechanical
 seals.[16]

-------
                                                        i-'iyure 17
       RIVET.
         P. 2 PtACEi
                           BOTTOM
                             VIEW
	.O* MIN      I

~         TOP VIEW
    SECT. A-A
NOTES:
        ALL BURRS AMD SHARP
  -Ol R I^IAX. OR CHAMBER
  ROTA.RY  awsFT &CAC i» A
  PART. PART NO.
                                                                                                 IWf
                                                                                                    OPPe«
                                                                                                    VALVE
ROTARY iMAFT
U3WER
                                                                                                                       VI TOM
                                                                                                  NOZZLE SCAL/RtUEF VAV-VE
                                                                                                            ASSEMBLY
                                                                                             MUELLtH
                                                                                                              . IMC •»»»•«»«.. MO
                                                                                                                              [>«•« I of I

-------
                              -50-
     3.    Summary

     The  analysis  presented  above  demonstrates  that  simple,
straightforward  engineering solutions  exist  for  the  specific
concerns raised  by  the  commenters.   In all cases,  manufacturers
have  a number  of  design  options  available  to  address  these
concerns.

V.   In-Use Fuel System Safety

     1.    Summary of Concerns

     Some  concern  has  been expressed that any  time  a  system
increases  in  size or complexity,  the potential  for  a failure
within  the  system  also   increases.   Applying  this  line  of
thinking  to  vehicle emission control  systems,  it  has  been
suggested  that   onboard   systems  would   inherently  decrease
overall  fuel  system  safety  because  several  components  are
larger  and a  few more  components  are needed than  for current
evaporative emission systems.   In-use vehicles  are  subject  to
innumerable accident situations,  and  some  concern  exists as to
whether or not  an increase in component size/number  could lead
to safety problems.

     Further,   it has been stated that  even if a  vehicle fuel
system  is  safe enough  to pass FMVSS  301,  it  does  not  insure
that it is  free  of all  safety risks in-use  as  evidenced  by the
number  of  vehicle  crash fires that  occur  each  year.   It  has
been argued that  vehicles equipped  with an onboard system could
pass all  FMVSS 301  tests  and yet  directionally increase  risk
in-use by  some unguantifiable  (presumably  small)  amount.   Thus,
it  follows that  because  some in-use  situations  differ  from
FMVSS  301  tests, onboard  systems must not  only be designed to
be  capable of  passing  Federal  safety  standards,  but   these
systems must  also  be designed  so  as not to increase in-use risk
for fuel system related hazards.

     2.    Analysis of Issues

     Fundamentally,  EPA   believes   that  overall   risk  in-use
should  not increase.   And, while   it  is   true  that  FMVSS  301
cannot  protect  against  every  conceivable  in-use  situation,
manufacturers  are  motivated  to  consider   fuel  system  safety
implications  for   reasons   other   than   insuring   that   their
vehicles  pass  Federal   safety standards.    Manufacturers  must
determine  what  they consider to  be  an  appropriate  level  of
safety and  in-use risk, and then design their  vehicles to meet
this level.   Often this  leads to different overall  levels  of
safety  in  different vehicle models.   Before discussing  how to
address  this   issue, it  is  valuable  to   discuss  how  safety
concerns  are   integrated  into the  overall  vehicle  design  and
development process.

-------
                              -51-
      First,  safety  is  an  integral  part of  the design process
 and   is  normally   not  considered   incrementally,    However,
 managing  risk  involves  a  series  of trade-offs,  balances,  and
 compromises  with  other  key  design  criteria.   Manufacturers
 choose  not  to  make  their vehicles free  of  all risk because of
 other valid  design  considerations  such as performance, styling,
 weight, cost,  and  other factors.  It is generally accepted that
 no  technological  constraints  exist which  would  prevent  the
 production   of  a  nearly  "fire-proof"   vehicle,  ; and  certainly
 vehicles  could be  made   safer  than  they  currently  are  as
 evidenced by numerous  "safety  car"  designs.[27] ' However, cost
 and other considerations  are valid and  they  prevent  "zero" risk
 (or   a   perfectly   safe   vehicle)    from   being   considered
 appropriate.   One  analyst  has  stated,  "It  is  definitely  not
 reasonable  to  expect  manufacturers  to  produce ';Sherman Tanks'
 ... as  such  vehicles would neither serve the needs  of societal
 safety, mobility, or economy."[28]                ;
                                                  i
     This  same  logic  and  risk management  process  applies  to
 fuel  system  safety.   Factoring  safety  into  fuel! system design
 is  a  complicated process that involves numerous1 tradeoffs  and
 compromises  as  above.   Fuel system  designs  are ;not  all alike,
 and fuel  system safety considerations  vary  from; one  design  to
 another.   For  example,  fuel  tank  size  and   location  on  the
vehicle have a substantial  impact on  a  vehicle's safety during
 a  collision.    Rear  fill  tanks  are  in  a more 'accident  prone
 location than  side  fill  tanks,  and  are usually located closer
to  the  exterior  shell  of  vehicle.    Side   fill   tanks  are
generally   considered   safer    than   rear   fill  tanks,   and
consequently, rear  fill fuel tanks  are gradually being phased
out of  vehicle designs.  However, it should be noted that this
change  over  has not  occurred  immediately due  to other  design
considerations  such as  cost  and  conflicting  interaction with
other aspects  of the total vehicle  design.   A similar  set  of
arguments  can  be  made with plastic versus  met;al  fuel  tanks.
These  simple   examples   demonstrate  how  risks  are  managed
relative to  other  considerations.  Even  current fuel  systems
could be safer  but some risk is accepted.

     Another  interesting  example lies  in  the  area  of  fuel
system  external plumbing such  as emission control  vapor lines
or external  vent lines  along the fillpipe.  At  one  time  added
piping  connections   similar to  the  external  vapor  vent  lines
that  appear  on some of  today's vehicles were  characterized  as
an unacceptable added  safety risk by General Motors.[29]  After
further testing and  design,  that  same  manufacturer incorporates
an external  vapor  vent line into many  of  its  current vehicle's
fuel  systems.   With  safety engineering  and  field  testing  any
potential  safety  risks  associated  with  these; external  vent
lines has been managed.

-------
                              -52-
     This   particular   design  change   illustrates   a   very
significant aspect of  fuel  system  safety.   Even  though  concern
existed  over  the  potential  safety aspects  of additional  fuel
system  plumbing,   the  mere  fact  that  these  additional  lines
appear on  today's  vehicles  confirms  that  safety concerns can be
technically addressed  if desired.   Any perceived  in-use  risk
can be managed.   Safety  does not  have to  be an obstacle to fuel
system improvements or modifications.  The  technology  to reduce
safety risks  is  currently  available,  and  the degree to which it
is utilized depends  on how  much risk a manufacturer is  willing
to accept.

     As  illustrated  in  the  discussions  above,   manufacturers
accept or  manage  varying amounts  of risk  in  order to  strike a
balance  or   compromise   with  all  of  the   important  design
criteria.  Clearly safer  vehicles  could be made,  and the amount
of  in-use  risk reduced.  As considerations change, the amount
of  risk   accepted  may  also  change.    Often  the   level   of
acceptable  risk  may  be more  constrained  by  in-use  liability
concerns  than  government   safety  tests.    For  example,  crash
testing  results  from  NHTSA's new car  assessment  program  show
that the vehicles' ability to  protect its  occupants from injury
vary  by vehicle  model.[30]   Different vehicle models  provide
different  levels  of  protection for  the  head,  chest,  and femur
during  barrier crash  testing  at  35  mph.  Some  manufacturers
chose to incorporate safer  designs on some  models  for  liability
and perhaps marketability reasons.

     Similarly,  the  safety  of an  onboard   system  on  in-use
vehicles  will  depend on  the design  decisions   made  by  the
manufacturers.  Onboard  systems would  increase  the  size  and
number  of  fuel  system  emission  control   components,  and  some
concern  has been expressed  that the safety  of these components
in  FMVSS  301  testing may  not necessarily   be  indicative  of
in-use  performance.    However, adding  these  systems  does  not
need to  affect  the level of  risk  a  manufacturer  is willing to
or  can  afford  to  accept.    As  with any  other  system  change,
manufacturers  would   integrate  onboard   systems   into  their
vehicles'  fuel  systems without increasing  overall  system risk,
and  clearly,   there   are   no  inherent   technical  constraints
prohibiting them from doing  so.

     Further,   there  is  little merit to  the  assertion  that an
onboard system must be inherently less safe  than  an evaporative
emission system because  it  is more "complicated".  Adding a few
components and  enlarging a  couple  of others  presents  no  risk
which  cannot  be managed  to  levels now deemed  acceptable.   As a
matter of  fact,  many  of the improvements  recently implemented
on   passenger  cars   and   light   trucks   have   resulted   in
vehicles/systems  which  are  increased   in   both   safety  and
complexity.    Consider    for   example   advances    made   in
vehicle/engine  control  systems.    Electronic  engine   controls
have  increased  vehicle   engine  complexity  tremendously  over

-------
                               -53-
 previous  systems, yet  there  is  no evidence  that these system
 "complications"    have    jeopardized    safety.      In    fact,
 manufacturers  are now  considering computer  controls for other
 vehicle  systems such as  the suspension  and  handling,  with the
 direct   purpose  of  improving   vehicle  safety.[31]   A  more
 complicated  system does  not  imply  a  less  safe  one  if  given
 proper consideration during  design.

     As  discussed  in  detail  earlier,  manufacturers  have  many
 options  available  in  the design of an  onboard  system which can
 manage  or  eliminate  any perceived  increase  in.  in-use  risk.
 However,  for  manufacturers  with  special  concerns  regarding
 in-use  safety  there  are  even more  design  options available.
 Fail  safe,  redundant,  or breakaway  rollover  valves could  be
 used.   The  integrity  of  the  critical portion  ;of  vapor  line
 between  the  fuel  tank  and  rollover  valve   could  be  assured
 through the  use of steel braid covered rubber hose in key areas
 or  steel  tubing.[32]   Both  rubber and steel  vapor  line  have
 been used  on past vehicle models.  If  chafing of this critical
 portion of vapor line is a concern, the affected,areas could be
 wrapped  in  a spiral spring  for protection.   Als.'o,  slack could
 be provided  in  this  critical portion of vapor  line to minimize
 the  possibility  of  separation  or  rupture  in!  an  accident.
 Improved or  additional  fittings,  adhesives,  or clamps  could be
 used  to  increase  the  strength of key vapor  line connections
 between the fuel tank and the  rollover  valve.   Concerns related
 to the charcoal  canister  can be addressed by using a reinforced
 canister shell  or  a  protective  barrier.   While  these  may  be
 somewhat extraordinary,  this  brief  listing  demonstrates  that
 further  design  options  are  available  which  iif  used  could
 improve safety over current vehicles.             |

     In summary, manufacturers  can  manage their ;in-use risk and
 can  choose  to  make  an  onboard  system  as  safe  as they  deem
 appropriate.  Onboard  systems  present  no  safety concerns which
 cannot   be   eliminated   through   proper  design,   and   each
manufacturer  will   develop   the  fuel   system!  design  which
 represents  the  best balance  for each particular  vehicle model,
with full consideration of  the  safety  risks and  all other key
 factors.

     3.     Opportunities for Improvement

     Implementing onboard  controls  could  actually result  in  a
net   improvement  in   overall   fuel   system  'safety.    since
manufacturers would  need  to  redesign   some  aspects  of  their
vehicles'   fuel  systems  to   incorporate  onboard   systems,  the
opportunity  would  be provided to  reexamine  other  aspects  o£
fuel system  safety as  well.   Some of the potential fuel system
 improvements that could  result from this  reexamination  include
 an acceleration  of the  transition from  rear  fill  to side fill,
 integration of  the current external vapor vent line  inside  the
fillpipe,  better   placement   of  the   fuel   tank,  or   even

-------
                              -54-
improvement in the  fuel  tank  integrity itself.  Also,  any number
of  other  minor modifications  or improvements  in the  fuel  or
emission  control   systems  could  be  made which  could  enhance
safety and performance and perhaps  reduce cost.  These  include
areas  such   as   tank  venting,  purge   valve  operation,   and
eliminating many  problems  identified  through owner  complaints
and other similar  survey measures.

     Also,  it  is  likely  that  an  onboard  refueling  control
requirement would  lead  to  a  decrease in  the amount  of  fuel
spilled in-use and  thus  improve  the overall  safety of refueling
events.   In  the certification refueling test,  vehicles  would
have to  be designed to  accommodate  a  refueling dispensing rate
near the  high  end  of  the present range  of  in-use values  (8-10
gallons per minute) without  any  spillage or  spitbacks.   This is
because any fuel spilled during  the test is considered  as  part
of  the   test   results.    Since  one  tablespoon  of  gasoline
evaporates to  a  substantial  amount  of vapor (about  10  grams),
almost any spillage that occurred during the certification test
would  result   in  a   failure.    Thus,   the  test   procedure
requirements   will   insure   that  manufacturers'    fuel   system
fillpipe designs are capable  of handling  dispensed  fuel  at flow
rates  up  to  10  gallons/minute without  allowing  any spitback.
The use  of these fillpipe designs  are predicted to  lead to  a
reduction  in  the  amount  of   fuel  spilled  in-use.   This  is
compared  to  some current  vehicle fillpipe  designs  which  have
difficulty accepting  fuel  at the lower end of  the  in-use range
(8-10 gpm) without spitback.   To assure this  benefit  accrues in
the  long  term,  EPA  is  considering an  in-use dispensing  rate
limit  of  10  gallons   per   minute  along  with   any   onboard
requirement.

     Also,  from  the  analysis  presented above,  it  is  evident
that implementing onboard controls would  provide  at  least three
other  direct  safety  benefits over   present  systems.    First,
depending  on   the  design used,  adding   a  rollover  protection
device may improve  the safety of  present  fuel tank  systems
which  use  a  1/2"   external  vent   line  without  rollover
protection.  Second,  adding  a rollover  valve  may enhance  the
safety for those vehicles which  now use a limiting  orifice  for
rollover  protection,   since  a rollover  valve will  provide  a
positive  seal  in  lieu  of   the   "controlled   leak"  approach
provided  by  the limiting  orifice.   Last of  all, it should be
noted  that refueling vapors  are currently  vented  to  an  area
which poses somewhat  of a safety  hazard.  This  is  because  the
potential  exists for  refueling vapor.s to ignite inadvertently
as  they  escape from the fillneck opening.   However,  as  onboard
controls  are   phased   in,  and  more  and  more  vehicles  route
refueling  vapors  away from  the fuel  pump  operator  to  a safer
point  (the charcoal  canister)  the  overall  risk  involved  in
refueling a vehicle will be reduced.

-------
                              -55-
     Finally, to  address  any special concerns regarding onboard
 system  crashworthiness and  to perhaps  improve crashworthiness
 over  current vehicles, there  is  an alternative  onboard system
 design  available  which manufacturers may elect.  As .is shown in
 Figure  18,  this system is similar  to  Figure 4,  except  all the
 needed  valves  (rollover,  vent,   liquid/vapor  separator)  are
 built into  the  top of  the fuel tank, instead of externally.

     A  solenoid activated  rollover  valve could be  used (Figure
 19)  which  is  located  on top  of  or inside  the  fuel tank. [33]
 This  valve would  normally  be closed  except during  refueling
 when it would  be  electronically opened by a switch located near
 the  opening of the  fillpipe.   The switch  could  be  activated
 either by the  opening  of the door  over the  fuel ;cap or removal
 of the fuel cap itself  (see Figure  20).
                                                 i
     Yet  another  approach  is  a  mechanical  ball  valve.   This
 device  would normally remain  open to  provide ;a  clear  vapor
 passage.  However, in  the  event of  a  rollover  accident  gravity
 causes a  metal  ball  to roll into a fitted seat and seal off the
 vent line.  One variation on this design  (see  Figure 21)  would
 be simple mechanical ball  valve built in combination with other
 needed valves.[15]                               '
                                                 i
     As is  shown  in  Figure  18,  this onboard system  design may
 need a  fill limiter  to allow for normal  refueling operations
 (i.e.,  automatic  shut-off)  and to prevent overfilling  the tank
 during  full  refills.    A  sample  design  is  shown  in  Figure
 22.[33]    The operation  of  the  fill  limiter  is! quite  simple.
When the  tank  is full  the float rises in the  fill limiter and
 closes off  the  refueling  vent line.  This  causes pressure  to
 rise in  the tank,  subsequently  fuel  runs  up the  fillpipe and
 activates  the  nozzle  automatic  shut  off  mechanism.    While
 incorporation  of  a  fill  limiter   is  quite  simple  from  an
 engineering perspective, the design would have  t;o incorporate a
 "soft  close"  to  avoid back  pressure   "spikes"   and  possible
 spills at the end of a full refill.              :

     From a  safety  perspective this  alternative is  attractive
because  all  the  external  components are  either  removed  or
mounted in  a more protected  location.   The external vent line
 (Figure 1)  can be  eliminated  and  the other system  valves and
vapor lines  are moved  away  from the  vehicle  shell  to a  more
protected  area  within  the  vehicle   body.   Also!  no  vapor line
exists between  the fuel tank  and the  rollover valve,  so  vapor
 line integrity and connections are less critical,

     Finally, depending on how  high a priority  a  manufacturer
assigns  to safety or if significant  in-use risk !is  perceived,  a
collapsible  bladder  tank  design  could  be  used  to meet  the
onboard requirement.   Bladder tanks could lead  to a substantial
 improvement  in  fuel system  safety  by providing an  additional
shell of  protection to help  reduce  fuel spillage in  case  of  an

-------
                                                       Figure  18
                                ALTERNATIVE INTEGRATED EVAPORATIVE/REFUELING  SYFTEM

                                                TANK MOUNTED VALVES
                                                REAR MOUNTED CANISTER
                                                J-TUBE
PRESSURE/VACUUM
RELIEF CAP
                         MECHANICAL OR SOI.ENOID
                         ACTUATED VENT/ROIJ./OVER
                         VALVE,  LIMITING ORIFICE
                         & LIQUID/VAPOR SEPARATOR
5/8" DIA.
   LONG
                       /   - J-TUDE SEAL

                        CONTROLLED LEAK
                                                                                   U1
                                                                                   (Ti
                                                                                                         PURGE
                                                                                                         VALVE
                                                            JC
3/8" DIA.
5'  LONG
 TO PURGE
INDUCTION
 POINT
                                                         •3  LITER
                                                          CANISTER
                           14  GALLON FUEL TANK

-------
                     Figure 19
                                                                              I
                                                                              (J1
                                                                              -J
                                            MO
SCHISMATIC or POTENTIAL ONUOAUU VAPOR RECOVEHY  SVSTLM
                                      MUELLER ASSOCIATES, INC.

                                      140t •. COOCWOOO
                                               . MAHVLAMO
                                      JAMUAN V 31. !•••

-------
                                Figure 20
                                                                CAP  OOO&
p
         CAP
         Z^D—i	r
T
   LN
—1_ jrj    «
                                                    \
                                                            7
                                                                                    CAP
                                                                          UPWAKO

                                                                                    CAP)
                                                                                                   Ln
                                                                                                   T
                I'OTIINTIAl. MOMKNTAHY SWITCH  LOCATIONS
                                                KiO
                                                          MUELLER A630CIATE3. IMC.
                                                          140* •. fOOKWOOO «TIIC«T
                                                          • AtTIMOMC. MAHttAHO 2I22T

                                                          JANUAMV  9 «.  1»«»

-------
             -59-
       Figure 21

COMBINATION VALVE
                  ROLLOVER SHUTOFF
                        SS BALL
 KXXXXXXXXXYXX
                            OVERFILL SMUTOFF
                            S.S MESH
                            VAPOR LIQUID
                            SEPARATOR

-------
Figure 22
                                                             o
UALL-IN-CAGE FLOAT VALVE
                        MO
MUELLER ASSOCIATES. INC.
• 401 •. •0«kK«*OOP tT
• ALTIMOM. MAHVLAM
JANUAHV at. 1««f
                                                     211*7

-------
                               -61-
 accident.  Also,  a  bladder  tank could eliminate essentially all
 of  the safety  concerns  raised regarding control  of refueling
 emissions.   This  is because a vapor  space would hot  be present
 in  a  bladder  tank,   and  without   a  vapor  space,,  refueling
 emissions would  not occur.   Thus neither  a  refueling emissions
 canister,  external  plumbing,  or  a  rollover valve would  be
 needed.   It  might  even  be  possible to  eliminate the  present
 evaporative   system  and  enhance   safety  even  more.    Also,
 bladders  should  be  an attractive  option for  those   who  claim
 high  costs  or  packaging problems  with  canister-based  onboard
 systems.  EPA  is  quite interested in  collapsible  bladder  tanks
 as  an  option to  canister-based onboard systems. '  This analysis
 of design and in  use safety issues  and the associated costs and
 leadtime  is  not  directly  applicable to  collapsible  bladder
 tanks.    However,  EPA  plans  to  further  explore;  the cost  and
 technological  feasibility of  bladders  as well a|s  their safety
 and emission benefits.                           ;

     In  conclusion,  the  information  and  rationale  presented
 above  refute the assertion that  adding an onboard system would
 directionally  increase   in-use  risk,  even  if  only  by  some
 unquantifiable  (presumably small)  amount.   Any perceived  risk
 is manageable, and  furthermore,  it  appears that ;the  net effect
 of  an  onboard  refueling   control   requirement   could  be  a
 potential increase  in  fuel  system safety.  As  discussed above,
 and  in Section   IV  there are  numerous design alternatives  to
 address the  safety  concerns   raised.   To  varying degrees  all
 options  have  the  potential  to  improve the safety of  fuel
 systems in-use.                                   !
                                                 i
VI.   Cost and Leadtime Considerations
                                                 i
     The  comments  received  regarding onboard  vapor  recovery
 systems also addressed  the  cost and leadtime implications  of
 implementing such controls.   More specifically, I several of  the
 comments  addressed  onboard safety  costs  in some  form (usually
 addressing hardware costs),  and several commenters  expressed
 some concern over EPA's  leadtime estimate.  An analysis of  the
costs  and leadtime necessary to  implement  onboard  controls
 safely  is  an integral  part  of the  overall  evaluation of  the
feasibility  of this  control  approach.   As was  mentioned above,
cost is  one of  the other  key  considerations 'which is  often
balanced  carefully  against  safety  concerns, ,and  the  costs
needed to  implement onboard  systems  safely must  be  reasonable
relative to other safety costs and the  overall  costs  of  onboard
systems.    Further,   the  analysis  must carefully  consider  the
manufacturer  leadtime  needed  to  implement  onboard controls  on
their  production  vehicles.   This  includes the  time  needed  to
 identify,   evaluate,  and  address all  safety  concerns  and  to
comply with the test requirements prescribed in FMVSS 301.

-------
                              -62-
     The first portion  of  this  section addresses onboard safety
costs;  the  second  discusses  leadtime and  describes the  basis
for EPA's leadtime  estimate.  Some  of the cost figures  cited in
the safety cost  analysis  are  drawn  from a broader EPA  analysis
which develops total onboard system costs in 1984 dollars.[17]

     A.    Safety Costs

     As is evident  from the discussion presented in Section IV,
the costs needed to  implement  onboard controls safely  fall  in
several areas.   R&D  type  costs will  be incurred,  some new  or
modified components will  be  needed  which  may  slightly  affect
vehicle operating  costs,  and safety  certification  testing will
be necessary.   However  before beginning  a  discussion of  these
costs,  it  is valuable  to  discuss  how the FMVSS 301  standards
and  EPA's  evaporative  emission  control  requirements  impact
onboard safety costs.

     The  control  of  refueling emissions  through  an  onboard
system would not be the first Federal regulation to  require  an
investment  to  improve  fuel  system   safety.    The  first  fuel
system  integrity  standards  (FMVSS  301)  were  implemented  by
NHTSA for 1968  vehicles,  and  since  then there have been 2 major
additions   to   these   requirements.    Each    of   these   new
requirements  has caused  a small  cost  increase,  but each  has
also  led  to  an improvement  in fuel  system  safety on  in-use
vehicles.    In  the  mid  1970's,  FMVSS  301  was  substantially
upgraded  to  extend the  coverage   of  impact  types  to  include
rollover  events and,  rear end and  side  collisions.   A  1983
NHTSA   Technical   Report   describes   the   nature   of   the
modifications made  in response  to  the upgrading of the standard
and estimates  the  costs  incurred  by vehicle  manufacturers  in
order to  meet the  revised  standard and provide a  higher  level
of in-use assurance.[21]

     Table 3  describes  modifications that  were made  to  1977
model year  vehicle  fuel  systems in  response  to the increased
requirements  of FMVSS  301.  These  modifications  ranged  from
minor changes  such as the  slight revision  of  mounting  bolts or
clips to  more major  ones  such  as  recontouring the  fuel  tank.
Based   on   information   submitted   to   NHTSA   by   vehicle
manufacturers,   the  average   (sales-weighted)   cost   increase
required to  make these modifications was  $4.60 per vehicle.*
These  modifications  were  also  estimated  to   increase  vehicle
weight slightly  (an average of  three  pounds per vehicle),  which
would tend to marginally  increase  the amount  of  fuel  consumed
over  the  life of the vehicle (about   3  gallons of  fuel).   When
these  two  costs are  added,  NHTSA  estimated  the total  safety
cost  resulting  from  the  1977  revisions to FMVSS  301  averaged
about $8.50 per vehicle (1982 dollars).
     A  Bureau  of  Labor  Statistics  analysis  estimated  that
     vehicle costs  incurred  to  meet the 1977 revision  to FMVSS
     301 were  $4.70 and costs to meet the  1976  revision  to the
     standard (added rollover test) cost $2.10.[34,35]

-------
                                 -63-

                                Table 3

                          Summary of Vehicle
                  Modifications in Response to 301-77
Vehicle Components

Fuel System
 Components

Fuel Tank
Fuel Gauge Sensor

Fuel Lines

Fuel Vapor Lines

Fuel Pump
Modification(s) to
Improve Crashworthiness

- Increase gauge of tank material
- Add protective shield
- Recontour to minimize  c
  contact/puncture by other adjacent
  vehicle components.
- Strengthen/shield filler neck
- Increase  strength  of  solder/weld
  seams
- Strengthen   mounting   by   adding
  brackets, revising  mounting bolts,
  increasing   torgue   of   mounting
  straps
- Strengthen    filler    cap    seal,
  improve impact resistance

- Strengthen mounting

- Recontour

- Recontour, revise, revise clamps

- Provide shield
Other Vehicle Components Changed to Improve Fuel System Integrity
Rear Floor Pan/Support
 Rails/Wheel Housing

Rear Suspension (Springs,
 Shock Absorbers)
Rear Axle Assembly


           /
Rear Axle Assembly

Seat Belt Brackets

Engine Mount

Power Steering Pump Bracket
- Revise, add supports


- Change  support   brackets,   revise
  mounting  bolts,   revise  mounting
  procedure, and shield

- Minor changes in  contour of lines,
  screw   heads,    mounting   clips,
  recontour vent cover

- Revise hinge assembly

- Revise anchorage

- Slight revision

- Slight revision

-------
                              -64-
     Based on  an evaluation of  in-use  accident  information for
1977  and  later  model  year vehicles,  NHTSA's  1983  Technical
Report also  estimated that the upgrading of FMVSS  301  would in
the  long term  annually prevent  400  fatalities, 630  injuries,
and  6500  post  crash  fires.  This indicates that FMVSS 301 has
been  effective  in  substantially  improving  many  aspects  of
overall  fuel  system  safety and that  these  improvements  were
purchased relatively inexpensively.

     The  second area  of  interest   is  the  effect  of  current
evaporative  emission  systems on  potential  onboard system safety
costs.   As  was  described  in  Section  III  of  the  report,  an
onboard  system  is  in  many ways  an extrapolation of  current
evaporative emission control technology and  the  two systems are
quite similar.   Many  of the control techniques and basic system
components  used would  be  similar,  and  the  same  system  and
vehicle  assembly  approaches  could  be used.    In  fact,  many
manufacturers  will  likely  integrate  their  refueling   and  fuel
tank  evaporative  control  systems.   All  current vehicle  fuel
systems  incorporate  fairly  sophisticated  evaporative  emission
control  systems.   Since  these  fuel  systems  have  all  been
designed to  meet the  most recent and most stringent version of
FMVSS  301  and  also  provide  a  high  level  of in-use  safety
performance,  it  follows  that   a  thorough  evaluation  of  the
potential  safety  implications  of  evaporative  control  systems
has  already  been   conducted.    Since   onboard   systems   are
basically  extensions  of  evaporative  emission systems,  clearly
many  of  the   safety  design   considerations   associated  with
onboard  systems  related  to  passing  FMVSS  301  or  providing
in-use  assurance  have   already  been  resolved  or   at  least
addressed    in    evaporative    emission    system    designs.
Consequently,  much of  the  "ground  work"   required  to  insure
onboard  safety has  already been performed.   Therefore, it  is
important  to keep  the  magnitude of  the  onboard safety design
process  in  perspective,  because  clearly  much  of the  safety
technology needed  for onboard  is simply an extension  of  that
which already exists.

     Remembering the  relatively  inexpensive and yet  effective
nature  of  current  fuel  system  integrity  measures   and  the
"incremental"  nature   of  onboard   safety  in   terms   of   the
magnitude  of the task  and actual cost relative to evaporative
systems,   it  is  now  possible to describe the components which
factor into  onboard  safety  costs.   Basically,  the integration
of safety  into a fuel  system  incorporating  an onboard controls
involves  four  types  of costs.    These  four  costs  are  for  1)
design and development  (R&D),  2) specific  hardware,  3)  safety
testing,   and 4) weight  penalty (or added  fuel  consumption).
The  paragraphs  that  follow  describe  how  each  of  the  cost
components are affected by onboard safety.

-------
                               -65-
      To  begin with,  some  research  and development will have to
 be performed to safely integrate onboard  controls into vehicle
 fuel  systems.   EPA  has  estimated  that the  total  design and
 development  cost  required to  incorporate  onboard  systems  in
 vehicle  fuel  systems is  about  $112,000 per  family' or  in the
 range of $0.35  to  $0.55 per vehicle  (passenger  car  and  light
 truck).   This  cost  is  for  any development  effort involved in
 combining the components of an onboard  system with the rest of
 the   vehicle  to  form   a  unit  that   interacts  safely  and
 effectively.   Because  safety is  evaluated  inherently  in the
 design  and  development process and yet  is  only  one part of the
 total  effort,  only  a  fraction of  the  total  cost  should  be
 directly allocated  to safety.   Also, because much  of the safety
 related  system development work has already been  completed it
 is  not unreasonable  to  expect that onboard safety development
 costs would  only be a small fraction  of  the total cost in this
 area.   In  addition,  because  of the  incremental; nature  of the
 onboard  system,  much  of  the research  and development  that went
 into  making  evaporative control  systems  safe  can  be applied
 directly to onboard controls.
                                                  i
     Given that manufacturers are  designing an  'onboard system
 in the context of many requirements and certain design features
 serve multiple functions,  it  is very  difficult  to isolate the
 level  of  expenditures  directly  attributable  to  safety.   For
 this  analysis  it  was  assumed that  about  20  percent of  R&D
 expenditures  relate  to safety, which  translates to about $0.10
 per vehicle.   However,  total  onboard  cost  is  quite insensitive
 to  this  assumption,   even if  the  safety  related development
 costs were  tripled,  per vehicle  costs  would increase  by only
 one percent.                                      !
                                                 i
                                                 i
     The second component  of  onboard  safety costs  relates  to
 specific hardware  that may be  required  to insute  fuel  system
 safety.  EPA  has estimated costs  for three specific items which
 have been identified  as  potential  components to ;be included as
 part  of  the  onboard  system  design  explicitly  for  safety
 reasons.   These  three  items  are   1)  a  rollover  valve,   2)  a
 pressure  relief  mechanism,  and  3) fuel  system modifications
 necessary  to  safely  incorporate   a  rollover  valve,  pressure
 relief mechanism, or  other  onboard  hardware.  EPA has estimated
 the cost of  a  solenoid  rollover  valve  (like  the one  shown  in
 Figures  19  and  20)  to  be  $4.60. [17]   This price included the
 cost of the  valve,  an actuator located at  the fillcap,  and the
 necessary wiring and connectors.    Manufacturers  estimate the
 cost  of  a valve assembly  similar  to  that described  by EPA's
 cost  estimate  would  be  in  the range of  $5.00 to  $6.00.   It
 should be noted that  these estimates  are for the most complex
 rollover valve  type, and  that the  cost  of  a  simpler  valve
 assembly such as the  fillneck  mounted type  (see  Figures 9-15)
 is  estimated  to be  more  in  the  $3.00 to  $4.00  range.   The
 available  information indicates  that   an  appropriate  rollover
valve cost  falls into a range of $3.00  to $6.00.

-------
                              -66-
     The second  safety  hardware cost  is  for  a pressure  relief
mechanism.   This  mechanism would  only  be needed  for  onboard
systems   incorporating    a  mechanical   fillneck   seal,   and
consequently not all vehicles would  require its use.   However,
for  those  systems   that  would  require  a   pressure  relief
mechanism,  EPA  has estimated  that this  device would  increase
system costs by  approximately $0.50.   This estimate is based on
pressure  relief   mechanisms    currently   used  in   automotive
applications  which perform  the  same  basic   function  and  are
similar in complexity.[36]

     The final  onboard safety  hardware cost  accounts for  any
fuel system modifications that  would  be  necessary in  order  to
safely accommodate any  onboard  control hardware.   For  example,
a vehicle's fuel tank or  fillpipe  might have  to be re-shaped or
modified in order to accept a rollover valve.   Also,  for  safety
reasons,  some  slight   re-routing  of  the  fuel system's  vapor
lines may be  required.  EPA has estimated a total  modification
cost to  be $0.50  per  vehicle.   Only part  of this  total  cost
would be required for safety purposes.  However,  because  safety
inherently enters  into  the decision to make  any modifications,
it is difficult  to access what part  of the total  modification
cost should be  allocated  to  safety;  perhaps half or more ($0.25
to $0.30 per  vehicle)  could be considered as  driven  by  safety
related concerns.

     Summing  up  the three  individual  safety hardware  costs
yields  a total  estimated  figure  in  the range  of  $3.25  to
$6.80.     However,    this   cost   estimate  does   not   include
manufacturer  overhead  and profit.    In  order  to  obtain  the
retail   price   equivalent   cost,   these   estimates   must   be
multiplied  by  a  markup  factor.   Presently,   a  markup  factor
value  of  1.26   appears  representative.[37]    Therefore,  after
inclusion of the markup factor,  a  total  retail price equivalent
safety-related hardware cost falls within  the range of $4.10 to
$8.60.

     The  third  component of   safety  costs  accounts  for  any
safety   crash   testing   that   would   be   necessary.    EPA  has
estimated the  cost of  FMVSS 301  crash testing to  assure  fuel
system  integrity for onboard  systems  to  be  about  $34,000  per
bodyline/style  or  about $0.12   per  vehicle.[38]  This  estimate
is based on  four  tests  for  FMVSS 301  only  required  per  body
line/style with two vehicles required  for  each sequence of four
tests.   Clearly  safety  crash test  costs are very minimal in the
long term  and  do  not  pose  an obstacle   to   the  adoption  of
onboard  controls.   In  some cases these costs  may  be higher  but
even if  total  costs  were  double the  estimate, the  overall  per
vehicle  cost  would rise by less than  one  percent.   Also, costs
could  be lower  if  FMVSS 301  test  were  combined  with  crash
testing required for compliance with other safety standards.

-------
                              -67-
      The  fourth component  of safety  costs is  the estimate of
 the added  fuel  consumed over  the life  of  the vehicle due to the
 increase   in   vehicle   weight  resulting   from  added  safety
 hardware.   The  amount  of  weight  added  to  a  vehicle for  a
 rollover valve  and pressure relief mechanism  is very small  (0.4
 Ibs),  and EPA  estimates  that  only about  $0.25 in  added   fuel
 costs  will   result  from  their  inclusion   into  the  onboard
 system.[17]

     A total  onboard safety  cost  is calculated 'by summing all
 four  individual   component   costs.   Total  capital  costs  per
 family average  about $56,000.   The per  vehicle safety-related
 costs range  from  $4.50 to  $9.00,  or  about  25 percent  of EPA's
 estimate of the total cost,  depending  on the  type of  rollover
 valve used.                                      '<
                                                 i
                                                 i
     One  final   point  needs  to be  made with riegard  to  these
 safety cost  estimates.   To the  degree that manufacturers   take
 the opportunity introduced  by an onboard requirement to further
 reduce  in-use  risk  beyond  that  now  accepted  with  present
 systems,  some additional  costs  might be involved which have not
 been  identified or quantified.   On  a  fleetwide  basis  these
 would be quite  small.   Also,  it should be  noted that the added
 benefits of these measures have not been included either.

     EPA   estimates  safety   related   onboard  > costs   to   be
 $4.50-9.00 per  vehicle.   While  there is some uncertainty in the
 development cost portion of the estimate, the  total range shown
 here  is  quite insensitive to any error.  These 'costs are quite
 similar to those previously incurred by manufacturers to insure
 fuel  system  safety.  Many of the potential  problems related to
 implementing   onboard   systems   safely   have ;  already   been
 considered in the  design  and  development  of present evaporative
 systems.   The manufacturers previous  experience>in implementing
 evaporative  systems  safely   and   the  incremental  nature  of
 onboard  systems  reduces  costs  and  the  level  of  potential
problems.    This  analysis  demonstrates  that  high  levels  of
 in-use fuel  system  safety  can  be  achieved  at:  low cost,  and
there is no need for a  manufacturer to "cut corners" on onboard
 safety to reduce costs.

     B.     Leadtime                             '

     If EPA were to implement an onboard requirement,  it  would
be necessary  to allow a  sufficient  period of  leadtime between
the date  the rule is promulgated and the  model  year the systems
are to  be required on production  vehicles.  This  leadtime is
provided  so   that   manufacturers  will  be  able  to  adequately
prepare for the requirement through system  design,  development,
testing,  tooling,  certification, and  safety  evaluation.   Some
of  the  tasks   involved   in  the  preparation  process could be
worked on simultaneously, while  some tasks  cannot  begin before
others are complete.   While  EPA  estimates  that  none of  the

-------
                              -68-
individual tasks  require more than  twelve months to  complete,
due to the sequential nature  of  some of  the tasks,  a  leadtime
period   of   approximately  24  months   will  be  required   by
manufacturers.

     Figure  23  shows  how  the  individual  leadtime   components
result in a total  estimate of  24 months.   First,  four to  six
months are  included for manufacturers to  develop and  optimize
working  prototype  systems  applicable to all of  their  different
vehicle models.   This is not at  all unreasonable  given  the  fact
that  working prototypes  already exist  and many  manufacturers
have  evaluated  these or  their  own  prototype  to  some  degree.
Not all  manufacturers  have developed working  prototype onboard
systems,  but the technology required to develop  such  systems is
readily  available  and  in-depth  technical descriptions  of  such
systems  have been  described  in  publicly  available  literature.
Four  to  six   months  should  be   adequate  time  for   these
manufacturers to develop and evaluate prototype systems.

     Once  the   prototype   development   is  complete,   initial
durability testing of  the  prototype could  be  conducted  under
laboratory conditions.   This  laboratory testing  is not  expected
to last more than two months.

     Following  laboratory  testing,  three  separate actions  can
begin  simultaneously.   These  three  tasks  are:  1)   in-vehicle
testing,   2)  safety optimization,  and 3)  tooling and prove  out
of the overall  control  system through efficiency and  durability
verification.    Similar   in-vehicle   testing   programs   have
required  four to six months for  completion.   Safety  evaluation
is  the   second  task   which  could  begin  subsequent  to  the
completion   of   the  prototype   laboratory  testing.    Safety
evaluation would  involve the use  of computer crash  simulation
models   and  vehicle  crash  testing   (four  tests   per   body
line/style)  to  verify  the  crashworthiness of  the  vehicle's
modified  fuel   system.   Because  this  evaluation  could  begin
immediately  after  the  completion of  laboratory  testing,  a full
14 months of leadtime  would  be  available to  manufacturers if
needed to perform  this task.  Based  on  discussions with NHTSA,
6  months  is   normally  enough   time  to   complete   a  safety
evaluation.  Therefore, 14  months  appears  more than  adequate to
perform  the  necessary  safety optimization  and  testing for  a
manufacturer's   product  line.   Tooling  could   also  begin  once
laboratory testing is complete.   Figure  23 shows EPA's estimate
that tooling could require  as little as 3 months and as much as
12 months depending on the magnitude of  the  task.   Different
factors  are  weighed before a manufacturer  commits  to various
tooling  changes.    Manufacturers  can  commit  to  some  tooling
changes  for  onboard controls  immediately after  the  in-vehicle
testing  (e.g.,  purge valves), whereas they may choose to  wait
until after  safety analysis before committing to other tooling
changes  (e.g.,  rollover valves).   However,  in  an overall sense,
12 months would provide manufacturers with enough time to delay
some  tooling changes and still  complete  the  task well  within
the 24-month leadtime.

-------
  Safety-up to 14  MO,
  Tooling!
   Hoiil,  (-0 no,
   CW, 3-4 NO,
                 IIIIIIIIIIIIIMIIIII
MIIIMIIIIIIIIIIMI

IIMIIIIIIIIHII
  Certification 18-12 HO,
                                        DevelopMent:
                                          Prototype   H NO,
                                          Lab        2 HD,
                                          JHelilcle  4-6 no,
                                                                                       CT>
 I   I   I  I   I   I  I   I   I   I  I   I   I   I   I   I   I  I   I   I   I  I   I   I   I
FfiH
      8
 '•1-2

Honths
16
24
       ==••=••: linotst cclululino flixiiilitu.
                        FIGURE 23 ONBOARD, LEADTIHE

-------
                              -70-
     The  only other  process which  requires completion  within
the 24-month  leadtime period is  emissions  certification.   EPA
has  found from  past  experience  that  a  manufacturer  normally
requires  between  10  to  12  months  to  certify  its  product
line.[39]  This  estimate is based on a 10  month engine family
certification  schedule   which   allows  time   for   durability,
emission data, fuel  economy,  and  confirmatory testing.  Because
certification  cannot  begin   prior   to   the    completion   of
in-vehicle  testing,   certification  is  critical  path,  and  EPA
estimates a total  leadtime period of  24  months  will  be needed
overall.

     Twenty-four  months   of  leadtime  is   quite   reasonable,
especially since most  of the  fundamental development work  is
already  complete.   Onboard  system  prototypes  are  presently
available, and  many  aspects of  the  system's  performance  have
already  been  tested  and  proven to be effective.   Also,  because
onboard   control   technology  is  incremental    in  nature   to
evaporative emission  controls,  there  is no need to design  and
develop entirely new systems.  As a matter  of fact,  many of  the
critical  onboard design  issues  have  already been  incorporated
into  current  fuel   system   designs   with  the   inclusion   of
evaporative emission  control  systems.   For example,  evaporative
emission  control systems  have  already added  the following  to
fuel systems:  vapor  vent lines,  vapor storage device, canister
purge    capability,    and   corresponding    safety    provisions
associated with   each of  these additions.   Since  much  of  the
development  work  is  already  complete,   implementing  onboard
systems  should be  no  more of a problem to vehicle manufacturers
than was implementing evaporative emission control systems.

     EPA's  24-month  leadtime  estimate is  supported  by  past
experience    with    three    previous   evaporative   emission
rulemakings.   These  rulemakings  included  the original  1978  6.0
g/test   LDV/LDT   evaporative   emission   standard   which   was
implemented  with just  12  months of  leadtime,  the  1981  2.0
g/test   LDV/LDT   evaporative   emission   standard   which   was
implemented  with 24  months  of  leadtime,  and  the   1985  HDGV
evaporative  standard which  was  implemented with 24  months  of
leadtime.  In each  of these three   rulemakings,  manufacturers
faced  leadtime   factors   identical   to  the  ones  that  would
accompany  an  onboard  requirement,   including   safety.    Since
manufacturers  were  able to safely  and   effectively  integrate
evaporative emission  controls into their vehicles'  fuel  systems
with  24  months  of  leadtime,   and  since  the  magnitude  of  the
onboard  implementation  task is  similar,   this  suggests  that
manufacturers  should  also  be  able   to  safely  and  effectively
integrate onboard  into vehicle fuel  systems with 24  months  of
leadtime.

     As  far  as  safety development and evaluation is concerned,
EPA's  leadtime   estimate  is  also   supported  by   the  past
experience of NHTSA  in  implementing  the various  versions  of

-------
                               -71-
 FMVSS  301.   Table  4  shows  the  chronological  history of FMVSS
 301.   The original  1968 FMVSS 301  applicable  to passenger cars
 was  implemented with less than 12  months of leadtime.  When the
 standard  was  revised  for 1976  model year  passenger  cars,  17
 months of leadtime  was provided.    For 1977 model year passenger
 cars,  manufacturers had  to  contend  with the  most  substantial
 upgrade to the standard,  and this  was accomplished with only 29
 months  of   leadtime,    and   only   12   months   between  new
 requirements.   Also, beginning  in the  1977  model  year,  FMVSS
 301  was  extended  to  include  light  trucks.   This  extension
 involved   a   29-month   leadtime    period  with  .further  crash
 requirements   in  effect  12   months  later,   thus  requiring
 recertification.   Finally,  in  1977,  FMVSS  301  was extended to
 include school  buses (with a  GVWR greater than 10,000 Ibs), and
 this  requirement was  implemented  with  17  months  of leadtime.
 This  experience  indicates  that  24 months  of  leadtime   allows
 manufacturers  sufficient  time to  factor in safety1.

     Based  on the information provided above,  24 months appears
 to be  adequate time  to  implement  onboard  controls,  with full
 consideration   of   all   safety    concerns.     Because    safety
 evaluation  can proceed in parallel to three  other  tasks, more
 than  a year   is  available  for computer  simulation  and   actual
 safety crash  testing.    This    allows  adequate   leadtime  to
 properly  integrate  safety into  onboard systems especially since
 manufacturers  can utilize and expand safety technology  used in
 current   evaporative  emission   control   systems   to  develop
 effective   onboard  systems.    Also,   much    ofi   the    safety
 development  which  would  be  required has  already  taken  place
 with the identification and resolution of such potential  safety
 issues  as rollover  protection  and  fuel  tank  pressure  relief.
 Consequently,   a   24-month   leadtime   period  ;would  provide
 manufacturers  with  sufficient  opportunity  to  develop safe and
 effective onboard systems.                       '

     While  this  analysis  indicates that the   current  leadtime
 estimate of 24  months  is  reasonable for most if not all vehicle
 models, EPA is  sensitive to manufacturers  concerns  regarding
 leadtime   requirements.    Public   comments    regarding   EPA's
 24-month  leadtime estimate  were  submitted  as  part  of comments
 on EPA's   original  Gasoline  Marketing  Study  (July 1984).[40]
 While  most  commenters  did not  object to the  24rmonth leadtime
 estimate  presented  in   the  Gasoline  Marketing  Study,  auto
manufacturers felt that a 24-month  leadtime was insufficient to
 implement  onboard controls.   The  leadtime  periods  suggested by
 these  commenters  ranged  from   three  to  six  years.    Those
 commenters  suggesting   that   four   or  more   years  would  be
necessary  also   suggested  that   onboard   controls  should  be
 phased-in  gradually   as  normal   vehicle   model;  redesign  and
turnover  occurs.   Using  this  approach,  implementing  onboard
 controls would  be less  burdensome and would allow extra time to
 deal   with  implementation  or  packaging   problems  on  unique
vehicles.    However,  it is worth  noting that  comments received

-------
                                        -72-
                                        Table 4
                          Chronology  of  FMVSS  301  Requirements
Model Year
Requirement
1968(1]
1976[2]
1977[2]
1977[2]
1978[2]
1977[2]
1978[2]
1977[3]
Vehicle
Type
PC
PC
PC
Class 1
Class 1
Class 2
Class 2
School
Promulgation
Date
2-3-67
3
3
3
3
3
3
10
-21-74
-21-74
-21-74
-21-74
-21-74
-21-74
-15-75
Effective
Date
1-1-68
9
9
9
9
9
9
4
-1-75
-1
-1
-1
-1
-1
-1
-76
-76
-77
-76
-77
-77
Leadtime Time Since
(Months) Last Requirements
11
17 7 2/3 yrs.
29 12 mos.
29
41 12 mos.
29
41 12 mos.
17
[1]   Motor  Vehicle Safety  Standard No.  301,   Fuel  Tanks,  Fuel
     Tank  Filler Pipes,  and Fuel  Tank Connections  -  Passenger
     Cars;  32  FR 2416,  February 3,  1967,  Part  571;  S  301-1.
[2]   Federal  Motor  Vehicle Safety  Standard  No.  301,  Fuel  System
     Integrity,  39  FR  10588,  March  21,  1974.
[3]   Federal  Motor  Vehicle Safety  Standard  No.  301,  Fuel  System
     Integrity,  40  FR  48352,  October  15 1975.

-------
                              -73-
 from  the  manufacturers   suggesting   the   need  for  a  longer
 leadtime were  not supported with any compelling arguments which
 would substantiate the insufficiency of a 24-month leadtime.

     While  the analysis above  indicates  that  approximately 24
 months  of  leadtime  should  be sufficient,  there are some factors
 which must  be  considered but  are  difficult to factor  into the
 analysis.   First,  as  was  mentioned  above,  some  manufacturers
 have not  developed working  onboard  prototypes due  to  resource
 or  facility constraints and the possibility  exists  that these
 manufacturers  will  take   no  definitive  action   on   systems
 development  prior  to   a   final  action  by  EPA-    Some  have
 commented  that  these   manufacturers   should  not  be  penalized
 because of  this  and may  require a greater  amount  of leadtime.
 Second, vehicles  with  atypical  duty  cycles   (ambulances,  mail
 trucks, etc.)  may  require more  leadtime  to  implement  onboard
 controls safely.  Vehicles  assembled  by secondary manufacturers
 such  as  recreational  vehicles  and   airport mini-buses  could
 also require  more time especially  if adding  an 'onboard  system
 requires other vehicle  changes.  Finally,  more  leadtime  may be
 necessary because manufacturers  may not have  the, test  facility
 and  safety  engineering resources  to  effectively  comply  with
 multiple vehicle  safety standard requirements  concurrently.   A
 similar concern  may exist  for  emissions  recertification since
 manufacturers  would  in  most  cases have to recertify virtually
 all  gasoline  powered   vehicles   for   exhaust  and  evaporative
 emissions in addition to the  new refueling requirement.  Because
 of  these   concerns,  more   leadtime  may  be  necessary  for  the
 implementation of safe onboard control systems.

     EPA is  committed  to  providing  manufacturers the  leadtime
 necessary   to   implement   onboard   controls    safely   and
 effectively.  Consequently, EPA is open to  considering  the need
 for more  leadtime  and/or  a  short  phase-in period  for  onboard
 controls.   Such  a phase-in  period  would provide  manufacturers
with additional  time  to solve any onboard  system packaging and
testing  problems  for   unique  vehicle  models. !  Also,   if  a
manufacturer had  unique  safety  concerns  on  one or two  body
 lines/styles,  this  approach  would offer  a  manufacturer  more
 leadtime  to  properly  address  them.   In  addition,  it  could
 improve   the   cost   efficiency  of   controls   by   allowing
manufacturers  to  forego  development   of  onboard  systems  for
vehicle models scheduled for  retirement or  permit  manufacturers
other flexibilities with new models being  planned and those now
 in  production.   The  implementation  of  other  unique   control
strategies,   such  as   bladder  systems,   would:  require  more
 leadtime.                                         '

     It is also important  to note that if onboard controls are
required,  the  date of  promulgation of  the  final  rule may  be
 such that more than 24-months  leadtime is actually  available.
The model  year generally  begins in  September or October.   If
the publication of  the  final rule is  much beyond that  period,

-------
                              -74-
the manufacturers  would have the  remainder  of that  model  year
in addition  to  the 24 months discussed  previously.   Therefore,
in  actuality manufacturers  could  have  substantially more  than
24 months, but  EPA's  analysis indicates that only 24 months  is
needed.

     In  conclusion,  given  the magnitude  of  the  task,  this
analysis  indicates that 24  months  of  leadtime  is adequate  to
allow manufacturers to  safely and  effectively implement onboard
controls.  This  estimate  is  supported by  EPA's  experience  with
implementing   evaporative   emission   standards   and   NHTSA's
experience   with   implementing   the   various   versions   of
FMVSS 301.  However,  EPA  is  committed to providing the leadtime
necessary  to   implement   onboard   controls  both  safely   and
effectively.    Thus EPA is  open to  considering more  leadtime
and/or  a  short phase-in  period or  other  approaches  which are
pertinent.

     Up to this  point,  this  report has addressed onboard safety
issues  from  primarily a passenger  car and light truck  point  of
view.    It should  be  noted   however  that  just  as  evaporative
emission control technology  was  extended to  heavy-duty gasoline
fueled  vehicles  (HDGVs),  onboard control  technology  could also
be applied to HDGVs.  While  many of  the  safety issues discussed
thus  far  would  be  identical  in  an   HDGV application,  some
aspects   of   HDGV  onboard   safety   would  be  distinct   from
light-duty issues.   The  next section  in  this report  has  been
included  to  address  the  similarities  and  differences  between
heavy-duty and light-duty onboard safety issues.

VII. Heavy-Duty Gasoline Vehicle Requirements

     Since an EPA  onboard  refueling control  requirement would
cover  heavy-duty  gasoline  vehicles  (HDGVs),  in  addition  to
passenger  cars  and  light trucks,   it is  important to  evaluate
any potential HDGV onboard system  safety considerations as  well
as  those  encountered   in  light-duty   applications.   (It  is
important to note that an onboard requirement  will not  apply to
heavy-duty  diesel  trucks   and buses.)   While  none  of  the
comments received  regarding  the safety  implications  of onboard
specifically   addressed  HDGVs,  overall   light-duty  concerns
discussed  earlier  are  expected  to  apply.   However,  it  is
important  to note that  HDGV fuel  system  configurations  differ
somewhat  from  those found on passenger  cars  and  light trucks,
and the fuel  system safety requirements also differ.

     This  section  of  the   report   identifies  distinct  HDGV
onboard  safety  issues and   discusses  the  implications  these
distinctions   could  have   on manufacturers  fuel  system  safety
designs.   It begins  with a  brief  description of  some  of the
more common  HDGV  configurations.   Following  these  descriptions,
a discussion of  the HDGV  fuel  system safety  standards will  be
presented,   and   differences  between   light-   and   heavy-duty
vehicle onboard  systems due  to  fuel  system  configurations and

-------
                               -75-
 safety test requirements will be discussed,  Next, HDGV onboard
 safety issues will  be  introduced and  analysed,  : Finally,  this
 section will end with  a brief  segment concerning the effect of
 HDGV onboard safety  on  costs  and  leadtime.

      BeforQ beginning  this  analysis  one  key  clarification  is
 needed.   FMVSS  301   covers  all  vehicles  with  a: gross  vehicle
 weight  rating (GVWR) of  10,000  Ibs  or less  (plus school  buses
 over  10,000 Ibs  GVWR).   For  emission   control; purposes  EPA
 classifies  all  gasoline-powered  vehicles  with  a;GVWR of  8,501
 Ibs  or more  as  HDGVs.   Out  of  EPA's  HDGV category only 90,000
 vehicles  (or approximately 25 percent)  have a GVWR greater than
 10,000  Ibs.  Thus most  (or  approximately  75 percent) of  EPA's
 HDGV class  (those vehicles with  a GVWR between  8,;501  and 10,000
 Ibs-Class   lib)   is   covered  by  the  LDT  requirements in  FMVSS
 301.  Since the  fuel  systems  on  Class  lib HDGVs are essentially
 identical to those  on lighter weight LDTs, and FMVSS covers all
 LDTs up  to  10,000   Ibs  GVWR,  the  previous  portion  of  this
 analysis  applies to  the Class lib HDGVs.    The remainder  of this
 analysis  will  focus on  gasoline-powered  vehicles  whose  GVWR
 exceeds 10,000 Ibs.

      This   analysis  addresses   compliance   costs  with   the
 assumption  that  HDGV  manufacturers will use only certified fuel
 tanks   on  their  vehicles.    Currently,   it  is  the  owner's
 responsibility to purchase and use a certified  tank if required
 by  regulation.    The current  Motor  Carrier  Safety Regulations
 exempts a vehicle or  driver used  entirely within a municipality
 or  commercial zone,  although they may voluntarily  comply with
 the  regulations.   These  regulations   may be  changed  in  the
 future  to   be   applicable  to  all  HDGVs  and  .eliminate  the
 aforementioned   commercial  zone   exemption.    Therefore,   this
 analysis  assumes that all HDGVs  will use  fuel  tanks certified
 to comply with the regulations discussed below.

     A.     HDGV  Fuel  System Configurations

     Just   as  there  are  chassis  and  drivetrain  differences
between  heavy  and  light-duty vehicles,   there  'are  also  some
differences  in   their fuel system  configurations.   Fuel  tanks
are  generally  of  a heavier  construction and  are  larger  in
volume; dual  fuel tanks are also  more  common.   Fuel tank shapes
vary  somewhat   as  does  the  location  of  the  tanks  on  the
vehicle.  Also,   it   is  often the  case  that  the fillpipe  is
 integral  with the fuel  tank, or has  a very limited length  as
compared to  lighter weight vehicles.

     As  a  part  of  a recent  contract  study,  EPA  surveyed  the
characteristics  of  the  fuel/vapor  handling  systems of  HDGVs
over  10,000  Ibs  GVWR. [41]   The key  results  of  the  survey
portion  of   that  report are  summarized in Table  5,  which will
serve as the basis for the remainder of this discussion.

-------
                                                         Table 5

        Selected Characteristics of Heavy-Duty Gasoline Vehicle Fuel/Vapor  Handling  Systems by Vehicle Model/Series
                                                                                                     Diameter   Diameter
Model or
Manufacturer Series Fuel Tank Shape
GM P4T042 Rectangular
P6T042 Rectangular
C5D042 Rectangular and
Rectangular Step
Number of Size of
Fuel Tank Location Canisters Canisters
30 gal. Mount On Right 2
Hand Frame
30 and 60 gal. Mounted 2
on Left Hand Frame
20 gal. Mounted Right 2
Hand Frame
1500 and
2500 cc
1500 and
2500 cc
1500 and
2500 cc
of Vent of Purge
Lines Lines
0.312 in. 0.375 in
OJ312 in. 0..375 in
0.312 in. 0.375 in
FORD
                C6D042
                C7D042
                C7D064
           Rectangular and
           Rectangular Step
                B6P042      Rectangular
F-Series   Rectangular
                B-Series   Rectangular
                C-Series   D-Type
50 gal. Step Mounted
Right or Left Hand Frame

20 gal. Mounted Right
Hand Frame

50 gal. Step Mounted
Right or Left Hand Frame

Dual 50 gal. Step
Mounted Left and Right
Hand Frame

30 gal. Mounted Right
Hand Frame

60 gal. Mounted Right
Hand Frame

35 gal. Right Hand
Side Frame Mounted

30 gal. Right Hand
Side Frame Mounted
                      i .
50 gal. Right Hand
Side Frame Mounted
1500 and
2500 cc
                                                                     1500 and
                                                                     2500 cc
1400 ml. ea.


1400 ml. ea.


1400 ml. ea.
0.312 in.  0.375 in.
                 0.312 in.  0.375  in.
3/8 in.    3/8  in.
                                                                                                     3/8  in.     3/8 in.
                                                                                                     3/8 in.    3/8  in.

-------
                              -77-
      First,  as can  be  seen in  Table  5,  there  are only  two
 manufacturers  which  market  HDGVs.   Between them they offer only
 about 10  different  chassis models   to  which  any  number  of
 different  bodies  or payloads  can be attached  (tanks,  dumps,
 cargo boxes, motor homes, school buses, flat beds, etc).

      The   second  area   of  interest  is   the  fuel   tanks.
 Essentially  three  different tanks shapes  are  used:   standard
 rectangular, step  rectangular,  and D-shape.  Examples of these
 tanks are  shown in  Figure  24.   The tank volumes  range  from 20
 gallons to 60  gallons, with  an average in the range  of 35 to 40
 gallons  for single  tank  HDGVs and   75  gallons  for  dual-tank
 HDGVs.  EPA  estimates  that   about  15 percent  of HDGVs use dual
 tanks,  with   most   of  those being in  heavier  weight  trucks
 (>20,000  Ibs  GVWR).[17]  Most  passenger  car  and  light  truck
 fuel  tanks  are located under the vehicle body  and this  is also
 the  case  for  some  HDGV  configurations  (e.g.,  school  buses).
 However,  on   some   HDGV  configurations,  the  fuel   tanks  are
 mounted on  the outer side of the  vehicle  frame (right  or left
 hand  side  for  single tanks, both sides for dual  tanks)  and are
 exposed to  the road rather  than  shielded  by the  vehicle body.
 As  was  alluded to  above,  most  HDGV tanks  have only  a  limited
 fillpipe length  (<8")  and  some have  essentially  none  at  all,
 with  the fuel  cap being integral to the tank.

      Finally,  with  regard   to  the HDGV  evaporative  emission
 systems two  observations  are important.   (See Figure 25  for  an
 example  of  a  HDGV  evaporative system.)   First,  for the  same
 reasons as described for passenger  cars  and light  trucks, HDGVs
 use  a  limiting orifice  in the  evaporative emission  system.
 Second, the  total  evaporative emission canister capacity  on  an
 HDGV  is more than  twice the average on passenger  cars and LDTs
 (2.8-4.0 liters).  However,   on HDGVs diurnal  emissions from the
 fuel  tank  and  hot  soak emissions from the  fuel  metering system
 are  routed  to different canisters.   Hot  soak  emissions  are
 somewhat more  of  a  concern  on HDGVs because  presently most  are
 carbureted rather than  fuel  injected.   To  the degree that  HDGV
 engines  fuel  systems   are   converted  from carbureted  to  fuel
 injected as  is now projected, concerns over hot  soak  emissions
may diminish and allow the  elimination  o£ the  second  canister
 on those vehicles.[42,43]

     With  this  brief   background  on  HDGV  fuel/evaporative
 systems we  turn now to  a discussion  of the  fuel  system safety
 standards which apply to HDGVs over 10,000 Ibs GVWR.

     B.     HDGV Fuel System Safety Standards

     Fuel system safety regulations differ  according to  vehicle
 and   fuel    system    configuration.     The    Department    of
 Transportation/Office   of   Motor   Carrier  Safety   (OMCS)   has
 requirements which apply  to  all HDGVs  over 10,000 Ibs GVWR.   In
 addition,   school  buses must meet  the requirements  prescribed
 specifically in FMVSS 301.   The OMCS  and FMVSS 301 requirements
 are summarized below.

-------
                          -78-





                     Figure  24






                 HDGV Fuel Tanks
                      D-Shape
                                Standard Rectangular
Step Rectangular

-------
                                                       Figure 25

                                            TYPICAL HDGV EVAPORATIVE SYSTEM
                        -BALL CHECK VALVE
                           TANK RESTRICIOR
         ^PRESSURE/VACUUM
              RELIEF
                                                   PURGE
                                                   VALVE
SEALED  A
GAS CAP-'
3/16" DIA
'13' LONG
                                                                                                                TO PURGE
                                                                                                                INDUCTION
                                                                                                                POINT
                                2.5 LITER
                                CANISTER
                                                                                                                             I

                                                                                                                             ?
                      30 GALLON FUEL TANK

-------
                              -80-
     1.    Office of Motor Carrier  Safety Requirements

     OMCS  safety  regulations  include  both   specific   design
requirements and actual fuel tank safety tests.[44]  The  design
requirements    contain    rules    governing    the    location,
installation, and  construction of  fuel  tanks used  on  HDGVs.
Also,  fuel  lines,  fittings,  and  fillpipes  must  conform  to
certain requirements.

     The actual  testing  requirements  depend on whether   a  fuel
tank  is  side-mounted  or  non-side  mounted.   To paraphrase  the
definition, a truck  fuel  tank  is considered side mounted  if  it
extends beyond  the outboard side  of a front tire  positioned  in
the  straight ahead  position.   This  is  shown  pictorially  in
Figure   26.    Any   fuel   tank  which   does   not  have   this
characteristic  is   considered  non-side  mounted,  and  in  this
analysis will be  referred to  as  frame  mounted.   The  testing
requirements for frame-mounted tanks will be discussed  first.

     A frame  mounted HDGV fuel tank has to be  able to pass  two
fuel  tank  safety  tests.   The  first  of these two tests,  the
safety  venting   system  test,   involves  applying  an  enveloping
flame  to an  inverted  fuel  tank  to insure that the fuel  tank's
safety venting  system activates prior  to  the tank's  internal
pressure exceeding fifty  pounds per  square  inch.  The  second
fuel tank  safety test  is a leakage test which  involves  filling
the  tank to  capacity and rotating  the tank through an angle  of
150° in any  direction  from its normal  position to insure that
neither the  tank nor  any fitting  leak more  than one ounce  of
fuel  per  minute in  any  position  the tank  assumes during  the
test.

     HDGVs  with side  mounted fuel tanks must pass  two  other
tests  which   involve  dropping the fuel  tank  to  test   impact
resistance.   The  first  test,  termed  the  drop test,  involves
dropping a fully loaded  (equivalent weight  of  water) tank from
30  feet  onto an unyielding surface,  so that  it  lands  squarely
on  one corner.   A  second similar test  (termed  the  fillpipe
test)  requires that  a  fully loaded tank be dropped from  10 feet
onto  an  unyielding surface, so that  it lands  squarely   on  its
fillpipe.   In neither  case, may the  tank nor  any fitting leak
more than one ounce per minute.

     Based  on conversations with  the  two  HDGV  manufacturers,
the  vast  majority  of   HDGV  fuel   tanks  are  frame  mounted
(non-side  mounted).    No  side  mounted  tanks  are  offered  as
standard equipment,  and  only occasionally  one  is sold as  a
special order.[45,46]  Thus, this analysis will focus  primarily
on the safety venting  and leakage  test  requirements which apply
to  frame   mounted  tanks.   However,  the  drop  tests  for  side
mounted tanks will  also be considered.

-------
                   -81-
                Figure 26
 Pictorial Definition of Side-Mounted Fuel Tank.
 If the tank extends to the left of line A or  to
 the right of line B, then the tank is side-mounted.
/Lines A and B are tangent to the outer sides  of
 the front tires.

-------
                              -82-
     2.    School Bus Requirements

     In  addition  to the  OMCS  requirements  for  frame-mounted
tanks,  outlined  above,  school  buses  are  required  to  meet
specific FMVSS  301 standards.  However, this coverage  does not
include  all  of   the   test   requirements  as   prescribed   for
passenger  cars  and  light  trucks.  FMVSS  301 for  school  buses
over  10,000  GVWR  requires  an  impact  with a contoured  moving
barrier at  any  speed up  to  and  including 30 mph,  at any point
and  angle.   Depending  on the  design  and  location  of  the  fuel
tank  and  its protective structure, the "worst   case" point and
angle of contact  is  determined for each school bus  model, and
the  contoured moving barrier impacts there.  In this test, the
fuel  system  must  be designed  so  as not  to  leak more  than one
ounce of fuel per minute.[47]

     This  briefly  summarizes   the   current   Federal  safety
standards  applicable  to  fuel  systems  on  HDGVs   over  10,000
GVWR.   It  is important to  note  that more safety  requirements
could be applied  to  HDGVs  over 10,000 GVWR in the  future.   The
Department  of   California  Highway  Patrol  recently submitted  a
petition to  NHTSA to  amend  FMVSS  301  to include fuel  system
integrity   testing   for   heavy-duty   vehicles   over   10,000
GVWR.[48]  With this background information we  are  now prepared
to   discuss   how   the  differences   in  vehicle/fuel  system
configurations  and the  Federal  safety  standards may  affect the
design  of  an onboard system for  an HDGV relative to  the design
for passenger cars and light trucks.

     C.    Distinctions in HDGV Onboard Systems

     Just  as  the  evaporative emission  control  systems used  on
HDGVs  are  very similar  to  those  used on  passenger  cars and
light trucks, it  is  also expected  that an HDGV onboard system
would be very similar  in design  and approach to that conceived
for  lighter-weight vehicles  (a  possible HDGV onboard  system  is
shown  in  Figure   27).   However,  some  minor variations  would
exist due to differences in  HDGV  fuel  system  configurations and
the  requirements  levied  by  the   applicable   Federal  safety
standards.    Before  beginning  a  discussion  of   these  minor
variations, it  is  valuable  to reiterate a few key points raised
previously  with  regard  to  the  magnitude  of the  task  of
implementing onboard controls.

     First, like passenger cars  and light trucks,  all HDGVs now
incorporate  evaporative emission control  systems  (see  Figure
25)  and  their  fuel  systems  must  meet the  present  Federal  fuel
system  integrity  standards  (OMCS and  NHTSA).   Thus,  as  before
with  the  lighter  weight  vehicles,  the  application  of onboard
systems is best  evaluated incrementally  to the  measures already
taken  to  incorporate   evaporative  emission  controls  and  meet
safety  standards.   Much of  the  ground  work  has  already  been
completed,  the  needed modifications made and  components added.

-------
                                                     Figure  27   -

                              POSSIBLE HDGV INTEGRATED EVAPORATIVE/REFUELING  SYSltM
       PKESSURE/VACUIM
          JflELIEF
                           MECHANICAL OR SOLENOID
                           ACTUATED VENT/ROLLOVER
                           VALVE, LIMITING ORIFICE
                             LIQUID/VAPOR SEPERATOR
                                                                           oo
                                                                           r
                                     I
5/8" DIA
 13' LONG
                                                       PURGE
                                                       VALVE
SEALED-'
GAS CAP
  MECHANICAL
     SEAL
t-3/8"
~ 31
DIA
TO PURGE
INDUCTION
POINT
                                7.5 LITER
                                CANISTER
                   30 GALLCN HDGV FUEL TANK

-------
                              -84-
In  many   cases   no  changes  to  present  fuel  system  safety
assurance  or  evaporative  emission  control  measures  will  be
needed.   Second,   it  is  important to  note  that  HDGV  onboard
refueling  and fuel  tank evaporative  emission  control  systems
will  likely  be  integrated  as  with  lighter weight  vehicles.
This is quite easy to accomplish on HDGVs, since they  now  have
separate  canisters and control  systems  for fuel tank  and  fuel
metering system evaporative  emissions.  Thus  a  whole  new system
will not  be  added  to  control  HDGV refueling  emissions;  instead
the  refueling  and  fuel   tank   evaporative   emission  control
systems  will  be   integrated  into one  (compare  Figure  25  with
Figure  27).   Thus  many  of the  primary design  considerations
which   applied  for  the   evaluation   of   onboard   systems  to
passenger cars and  light trucks also apply to HDGVs.

     Remembering  the  expected similarities  between  light  and
heavy-duty   vehicle  onboard  systems  and   that   the  factors
affecting  the implementation  are also  the same,  the  expected
minor variations  in HDGV onboard  systems  can now  be  discussed.
For sake of presentation, discussion will  begin  at  the fillpipe
and follow  along  the system to the canister.   The analysis will
assume  an  integrated onboard  refueling/fuel tank  evaporative
control system as discussed  above.

     To  begin with,  because the  fillpipes on HDGV  fuel  tanks
are either  relatively short  or  integral with the  tank,  liquid
fillneck  seals which  require an appreciable fill height may not
be  a  practical approach  in some  configurations.   Due  to  this
lack of  fill height, HDGV  manufacturers might  elect  to utilize
a mechanical  seal  approach  and  thus  would need to  incorporate
some  type  of  pressure  relief  device  such  as  was  described
previously.   HDGV  fuel  tanks,   which  are  made  of  steel  or
aluminum,  now use a  pressure-vacuum  relief valve,  and  it  is
conceivable  manufacturers  will  simply  modify  that  valve  for
this  application.    However,   under   the   proper   backpressure
conditions,  it might  be possible to use a  liquid  fillneck  seal
by  extending the fillpipe horizontally in  the  tank  as  has been
demonstrated  in a prototype  light-duty program.[15]
     A second potential  difference lies in the diameter  of the
refueling vapor  line  and related fuel tank vent.   From a design
perspective,  the  tank  vent  and  refueling  vapor  line  size
(diameter)  could be affected by the fuel dispensing  rate.   As
part  of  the   refueling  emissions  test  procedure,   EPA  is
proposing that  HDGV  fuel systems be designed for refueling at a
maximum  rate  of  10  gallons  per  minute,  the  same  rate  as
prescribed  for  other vehicles.*  This  10  gallon per  minute


*    Discussions  with gasoline  marketing  interests and  nozzle
     manufacturers   indicate   that   gasoline   available   to
     passenger  cars,  light  trucks,  and HDGVs  (either  at  retail
     or  private  pumps)   is  normally  not  dispensed  at  rates
     greater than 10 gpm.

-------
                              -85-
dispensing rate  results  in an  increase  in the  current  orifice
and  evaporative  vapor  line  diameter  from  about  3/8  inch  to
about 5/8 inch for an HDGV onboard system.

     However, to  minimize  spillage during  refueling,  the  OMCS
has  requirements  that  any liquid  fuel  tank over  25 gallons  in
capacity must be able to  accept  fuel  at  a  rate of 20  gallons
per  minute.[49]   For an onboard  system this  requirement  could
lead to  a  increase  in  the  diameter of the tank  vent outlet and
refueling vapor  line.   It should be noted, however, that  while
this requirement  applies  to  all  heavy-duty  liquid fuel  tanks
(both diesel  and gasoline),   fundamentally  it  is aimed more  at
diesel fuel tanks.   It  is not  uncommon  to encounter an  in-use
diesel fuel  dispensing  rate  of 20  gpm or  more to reduce the
time  needed  to   fill  a  diesel  tank  since  these  tanks  are
typically much larger than gasoline  tanks  and dual diesel  tanks
are  also more common.[50]   In-use  gasoline dispensing rates  on
the  other  hand  normally do  not  exceed 10  gpm.   Since  in-use
gasoline dispensing  rates usually  do not  exceed  10  gpm,  and
EPA's  refueling  certification test  would  involve  a  10  gpm
maximum  dispensing  rate,  OMCS's  requirement  in this  area may
not  be  needed.    EPA has  discussed  this matter  with  DOT/OMCS,
and they have expressed a willingness to consider changing this
requirement to apply only  to  diesel  fuel tanks.[51,52]   If this
standard is not  changed, and  a 10 gpm dispensing  rate limit  is
enacted,   the only   effect  would  be  that  the   refueling  vent
orifice/line for these vehicles would be over sized.

     Nevertheless,  because HDGV  fuel tanks   do not  use  long
fillnecks,  fuel dispensing operations would  not  be as  sensitive
to higher backpressure as they would be  in  light-duty.   Even if
the  refueling vent  orifice/line were sized for  a 10 gallon per
minute dispensing rate,  fuel  could  be  dispensed at a  greater
rate without  premature  shutoffs.   Thus  it may not be  necessary
to size the refueling orifice/vent line  to match the dispensing
rate requirements.   However,  in optimizing  system designs with
regard to fuel tank pressure,  manufacturers may  choose to  use a
slightly larger  refueling  vent orifice  than  seen on light-duty
applications.

     One  final   manner   in which  HDGV  onboard  systems  might
differ from  those on lighter weight  vehicles is  in the design
of  the  rollover protection  device.   The  solenoid  activated
rollover valve (Figure 19) or the combination valve (Figure 21)
could  be   applied   to  HDGV  fuel   tanks  in   their   present
configurations.    One  manufacturer's   fuel   tank   design  now
incorporates  a ball  type  check valve similar in  principle  to
the  combination  valve.[41]   Also, the  nozzle  actuated  valves
shown in Figures 9-15 could  also be  used on HDGV fuel  tanks
which have  a fillpipe  length  of  6 inches  or  more.   However,
nozzle  actuated  valve  designs  would  have  to  be   modified
slightly to perform  on fuel  tanks whose fillneck is essentially
integral with the tank.   Nonetheless,   the  basic  approach and
operation would be the same.

-------
                              -86-
     Any  of  the  three  rollover valve  designs mentioned  above
could be  used  on HDGV fuel tanks.  However the best  choice for
any  tank  would  depend   on  the  fillpipe   length   or   other
trade-offs relative to  cost,  packaging  etc.   With proper design
and  integration  any  of  these  valve  designs  could  provide
rollover protection in-use.

     With  this background  on HDGV fuel  system configurations,
safety  requirements,  and  HDGV onboard  system characteristics,
it  is  now  possible  to  address  some  of  the  unique  safety
concerns  related to  HDGV onboard.   The  next  segment of  this
report  discusses  and  addresses   potential   impacts  of  HDGV
onboard on fuel system safety considerations.

     D.     HDGV Onboard Safety Issues

     1.     Introduction

     While none  of the comments  received regarding  the  safety
implications of  onboard controls  specifically addressed  HDGVs,
it  is   reasonable  to  expect  that overall  concerns  would  be
similar because  of the  expected close resemblance between light
and  heavy-duty  vehicle  onboard  systems.   To avoid  repeating
much of  what has previously  been discussed,   this  segment  will
primarily  focus  on unique  HDGV onboard  safety  considerations.
The  analysis  presented  in  Section  IV  regarding  maintenance,
repair,   tampering and  defects and  refueling  operation  safety
apply equally  to  HDGVs  and  will  not  be  repeated  here.   The
potential problems are  similar and the same basic  approach and
straightforward  engineering solutions  can  be used.   Also,  the
extensive  analysis  in  Section V  regarding  in-use fuel  system
safety  also  applies  to  HDGVs.   As  before,   manufacturers  are
expected to  manage risk appropriately;  there   is no reason  that
adding  an onboard  system  would  directionally increase  in-use
risk over  that now accepted with present HDGV fuel/evaporative
emission systems.

     However,   as  was   discussed   above  the   fuel   system
configurations and the  safety  test  requirements  for   HDGV  fuel
tanks   are  somewhat   different   from  light-duty,    so   some
discussion of distinct safety  test  design requirement issues is
appropriate.

     2.     Safety Test Design Requirements

     As  mentioned above,  there are two separate areas of safety
test design  considerations for HDGV  fuel systems.   The  Office
of  Motor   Carrier   Safety   (OMCS)   has  fuel   system  safety
regulations which apply to all HDGVs, and NHTSA  has  additional
requirements  for  school   buses.   This  segment  begins with  a
summary and analysis of safety design considerations  related to
OMCS  requirements.   Following  this   is a  discussion  of  the
effects  of NHTSA's crash test requirements.

-------
                              -87-
     a.    OMCS Requirements/Considerations

     OMCS has established fuel  system  requirements  for HDGVs  to
insure their  structural  and in-use  integrity.   As  part of the
current  requirements,  HDGV  fuel  tanks  must  be   capable  of
passing  the  safety  venting  system  and  the   leakage   tests
described previously.  Currently HDGV fuel tanks employ a ball
check valve and pressure vacuum relief valve  to pass  these two
tests.   Since  the   refueling  vent  orifice would  be  somewhat
larger with an onboard system (5/8") the ball  check  valve  would
have  to  be  upgraded  to  provide  the  necessary   protection.
Little or no  change to the  pressure vacuum relief  valve  would
be needed.

     For an HDGV onboard system,  the protection now supplied  by
the  ball check valve could be  supplied by the  rollover  valve
designs described previously.   The  same three general types  of
rollover  protection  devices  that  were  discussed  for  use  in
light-duty   applications    (nozzle  actuated,   solenoid,    and
mechanically activated valves)  would all be feasible in various
heavy-duty  applications  as  well.   However,   for   tanks   with
little or no  fillpipe (<6")  the  nozzle  actuated valve design
would probably  have to be modified  slightly and  mounted in the
tank  instead  of  on  the  fillneck.   A  solenoid  or  mechanical
rollover (ball) valve design could essentially be used as  shown
earlier.

     HDGV and light-duty onboard systems would be  functionally
identical and would  be very similar in design and configuration
except for canister  size and vapor line length.   Of  course,  to
meet  safety  requirements   and  to  provide in use  protection,
manufacturers will  have to  consider  the  structural  integrity
and the materials used in  key  system components just as they do
now  with  other  components  of  the  fuel/evaporative  system.
Thus, some  components of  the HDGV onboard system  (notably the
rollover valve) may  need to be  constructed of metal  to provide
impact resistance  and the  flammability protection  demanded  in
the safety venting test.

     Also, with regard to  impact  resistance,  any one optional
side-mounted  tank  model,  would  be subject  to  two  additional
safety  tests   (drop  tests)  designed  to  evaluate   the  tank's
impact  resistance.    A  side-mounted  fuel  tank would  likely
utilize a rollover valve mounted integral to or within the tank
to  insure its  integrity  during  the  drop tests.   While  this
would  not be  difficult  to design  (many  current  fuel  tanks
contain interior components), it  would  represent an additional
design   consideration for  side-mounted  fuel  tanks.   From  an
in-use  safety perspective,  the  impact  resistance   and overall
integrity of  rollover valves on frame  mounted  tanks  would  be
enhanced  if  they   were  mounted   integral  or  internal to  the
tank.  Thus,  this  approach would  be  attractive for  all  HDGV
fuel tanks.

-------
                              -88-
     In conclusion,  the only HDGV onboard safety  design  feature
introduced by  the need to meet OMCS safety requirements  is  the
upgrade of the current  rollover  protection  device.   All  of  the
rollover   protection   approaches   discussed   for   light-duty
applications   (nozzle   actuated,   solenoid,   or   mechanically
activated valves) could be used to meet this  requirement.   The
design, placement,  and  construction of  the  rollover valve on  a
particular  HDGV  fuel  tank would  depend  in   part  on  fillpipe
configuration,   impact  resistance   concerns,   and  flammability
potential.

     b.    NHTSA Requirements/Considerations

     In  addition  to OMCS  requirements, all  school buses  over
10,000  Ibs.  GVWR  must  also  meet  specific  requirements   of
NHTSA's  FMVSS  301.    As  described earlier,  this  involves  a
single moving contoured barrier test at  a maximum of 30  mph  and
does  not  include a  rollover  test.   In this  test, the  barrier
impacts the school bus at the most vulnerable location  of  the
fuel  tank,  and the   fuel  system  must  be designed so  as  not  to
leak  more  than one   ounce  of  fuel  per  minute.   As was  true  of
OMCS  requirements,  an  acceptable  school  bus  onboard system  is
one which does not  impair  the fuel tank's ability  to meet  this
requirement.

     As  in  the light-duty test, the crashworthiness of  all  the
onboard  system components  (rollover  valve, charcoal  canister,
critical vapor  line  and vapor line connections  between  the  top
of the fuel tank and the rollover valve) would  all  be  evaluated
in  the test.   Design  measures  similar  to  those  described  for
passenger  cars and  light  trucks  would  have  to  be  taken  to
assure the integrity of these  three key  components.

     The  crashworthiness  discussion  in  Section  IV-A  and  the
further  options  discussed  in  Section  V   addressed  specific
safety design  approaches for  these components  which  could  also
be  applied to  school  buses,  so  this  will  not  be  addressed
further.  As  before with  light-duty applications,  evaporative
emission  systems  provide  directly  relevant  techniques   and
experience to  assist  in proper  design,  and  specific engineering
measures have been suggested to deal with potential concerns.

     Furthermore,   the   in-use   safety   of   onboard   refueling
controls for HDGVs must be considered.   The location of onboard
system   components,   as   with   the  current   fuel  tank   and
evaporative  emission  controls,  must  minimize  any  potential
safety risks.   Much  of  the HDGVs fuel system damage seen today
is  caused   by  foreign  objects   from   the   road   surfaces.
Therefore, critical   onboard control system  component should  be
located  on the  HDGV  in  a  position which will  minimize  any
foreign object damage.

-------
                              -89-
     In  conclusion,  HDGV onboard  systems  do not  introduce  any
new  or  significant  problems  to  manufacturers'   attempts   to
design  safe fuel  systems  capable of  meeting NHTSA  and OMCS
safety   requirements.    Straightforward,   viable   engineering
solutions are  available to address all problems that  have been
identified.    Therefore,  onboard   systems  are  expected  to   be
integrated  into HDGV fuel  systems  without  reducing the system's
ability to meet all applicable Federal safety requirements.

     3.     Summary

     As was mentioned in the  light-duty section of this  report,
EPA's  philosophy  in  evaluating  the  safety  implications   of
requiring onboard  controls  (including those  for  HDGVs),   is
that no  increase  in overall risk  should be  caused or  accepted,
beyond that now present with today's  fuel/evaporative  system.
This  applies to  both  compliance  with  the  applicable  Federal
Safety  standards  and  the   in-use  safety  of vehicles  equipped
with  onboard   systems.   This  portion of  the  analysis   has
addressed  the  safety  test  design   requirements  related   to
implementing  HDGV onboard  systems,  and   as  was  the  case  for
light-duty   it   concludes   that   straightforward   engineering
solutions   are  available   for   all   of  the  potential   safety
problems  which have   been   identified,  and  safe  fuel   system
designs are achievable by all.

     E.     Cost and Leadtime Considerations

     EPA  has   received no  comments  which  directly  address
specific  HDGV  onboard  safety cost  and leadtime  implications.
However,  an analysis  of the costs  and  leadtime necessary to
implement HDGV  onboard controls safely  is an integral  part of
the  overall  evaluation  of  the   feasibility of  this  control
approach.   The  first  portion of  this  section  addresses  HDGV
onboard  safety   costs;  the  second  discusses   HDGV  leadtime
requirements  and  describes   the   basis   for  EPA's   leadtime
estimates.   Some  of  the cost  figures cited  in the  safety  cost
analysis  are  drawn from a  broader EPA analysis which develops
total HDGV onboard system costs in 1984 dollars.[17]

     1.     Safety Costs

     As was true  of light-duty onboard safety costs,  the costs
needed  to  implement   HDGV   onboard   controls  fall  in  several
areas.  R&D type  costs will  be incurred,   some new or modified
components  will be needed  which  may  slightly  affect  vehicle
operating  costs,   and  safety  certification  testing  will  be
necessary.   However,   before beginning  a   discussion  of  these
costs,  it  is  valuable  to  discuss  how EPA's HDGV  evaporative
emission control requirements impact onboard safety costs.

-------
                              -90-
     As was  described in the light-duty section  of  the report,
an onboard  system (even those  for HDGVs)  is in  many ways  an
extrapolation   of    current    evaporative    emission   control
technology  and  the  two  systems  are  quite  similar.   Since
onboard   systems   are   basically   extensions  of   evaporative
emission   systems,    clearly    many   of   the   safety   design
considerations  associated  with   onboard   systems   related  to
meeting OMCS/NHTSA  requirements or  providing in-use  assurance
have  already  been  addressed  in  evaporative  emission  system
designs.  Consequently,  much of   the  ground work   required  to
insure  onboard  safety  has  already  been  performed.   It  is
important  to keep  the  magnitude  of  the  HDGV  onboard  safety
design  process  in  perspective,   because   much  of  the  safety
technology needed is  simply  an  extension of that which already
exists.   Noting  the  "incremental" nature  of onboard  safety  in
terms of  the  magnitude  of  the task and actual cost  relative  to
evaporative   systems,   it   is  now  possible  to  describe  the
components which factor into onboard safety costs.

     Basically, the  integration of  safety  into  a  fuel  system
incorporating  an  onboard system  involves  four  types  of  costs.
These four costs  are for:  1) design  and development  (R&D),  2)
specific  hardware, 3)  safety  testing,  and 4) weight penalty (or
added  fuel  consumption).   The  paragraphs  that  follow describe
how each  of the cost components  are affected by onboard safety.

     To begin with,  some research and development will have  to
be performed to  safely integrate onboard  controls  into  HDGV
fuel  systems.  EPA  has estimated that  the total  design  and
development  cost  required  to  incorporate   onboard  systems  in
HDGV  fuel systems  is  about  $34,200  per  family  or  $1.50  per
vehicle   (over  10,000   Ibs   GVWR).    This   cost  is   for   any
development  effort  involved  in combining  the components  of  an
onboard system with the rest of  the vehicle  to form  a unit that
interacts safely  and effectively.   Because  safety  is evaluated
inherently  in the design  and  development  process  and yet  is
only one  part of  the total  effort, only a fraction of the total
cost  should  be directly allocated to  safety.    The light-duty
cost  section  explained  why this  fraction is  likely  to  be
small.  The  same reasoning  is  also  applicable  for heavy-duty
applications,  and  therefore   it   was  assumed  that  about  20
percent of R&D expenditures  relate to safety, which translates
to about  $0.30 per vehicle.

     The  second component  of HDGV onboard  safety costs relates
to specific hardware that may be required to insure  fuel system
safety.   EPA  has  estimated  costs  for  three specific items which
have been identified  as potential  components to be  included  as
part  of  the onboard   system  design  explicitly  for  safety
reasons.  These   three  items are  1)  a  rollover valve,  2)   a
pressure  relief   mechanism,   and  3)   fuel   system  modifications
necessary  to  safely  incorporate  a  rollover  valve,  pressure
relief  mechanism, or  other  onboard  hardware.   HDGV  rollover

-------
                              -91-
valves should  not differ  in  cost  from light-duty valves  since
they would essentially  be  the same.   Therefore, the  light-duty
estimate of $3.00 to $6.00 will also  be used here.

     The second  safety  hardware cost  is  for a pressure  relief
mechanism.    Since this  mechanism  would be  needed for  onboard
systems  incorporating  a  mechanical  fillneck  seal,  many  HDGVs
would require  its use.   EPA's  analysis prices  this  device  at
$2.50.[13]    At this  point,  this  estimate  is considered  to  be
very conservative, since the  possibility exits that the  present
pressure relief device can be modified to perform this function.

     The final onboard  safety  hardware cost  accounts  for  any
fuel  system  modifications  that would  be  necessary in order  to
safely accommodate  any  onboard control hardware.  For  example,
a  HDGV  fuel tank or fillpipe  might have  to be  re-shaped  or
modified  in  order   to   accept  a  rollover  valve.   Also,  some
slight  re-routing of  the  fuel system's  vapor  lines  may  be
required.  EPA has  estimated a total  modification  cost  to  be
$0.50 per  fuel tank.   Only  part  of  this total  cost would  be
required   for    safety   purposes.    However,   because   safety
inherently enters into  the decision  to make  any  modifications,
it  is  difficult  to  access  what part of  the total modification
cost  should  be  allocated  to safety;  perhaps half  ($0.25  per
fuel  tank)   could be considered   as  driven  by  safety  related
concerns.

     Summing up  the  three  individual  safety hardware costs  per
fuel tank yields  a  total estimated figure in the range of $5.75
to  $8.75.    However,  this   cost  estimate   does  not   include
manufacturer  overhead  and  profit.  Consequently,  in order  to
obtain the  retail price equivalent  cost,  these estimates  must
be multiplied  by a  markup  factor.  Presently,  a markup  factor
value  of  1.27  appears representative.[37]   Therefore,   after
integration  of   the  markup   factor,  a  total   retail   price
equivalent  HDGV  safety-related hardware  cost  per   fuel  tank
falls within the  range  of  $7.30 to $11.10.  Since 15  percent of
HDGVs have  dual  tanks,   the  total  HDGV safety-related  hardware
cost range is  $8.40 to $12.80

     The third  component  of  safety costs  is  for  any  safety
testing that would be necessary.   Unlike  light-duty test  costs,
EPA  has  not   thoroughly  investigated  HDGV  safety test  costs.
However,  safety   test  costs  were  estimated  in  an  attempt  to
determine  the  approximate  magnitude  of  the  per vehicle  HDGV
safety test  cost.   Table 6 shows that even when  fairly  liberal
safety test  costs  are  assumed, the  resulting cost/vehicle  of
$0.70 is very  minimal in the  long term.

     The fourth  component  of  safety  costs  is  the estimate  of
the added fuel consumed over the  life of the vehicle  due to the
increase  in   vehicle  weight   resulting   from  added   safety
hardware.   The  amount  of  weight  added  to  vehicle  from  a

-------
                              -92-


                            Table 6

           HDGV Fuel Tank Safety Test Costs Estimate

1.   OMCS Requirements:

     2 tests per HDGV fuel system configuration
       (Safety Vent Test and Leakage Test)

     Conservative Cost/Test Estimate:  $2,000

     8 HDGV Fuel Tank Configurations

     Total OMCS Safety Test Cost:  $32,000


2.   NHTSA Requirements:

     1 test per HDGV fuel system configuration
       (30 mph moving barrier)

     Conservative Cost/Test Estimate:  $30,000

     7 School Bus Configurations (7  manufacturers,
       1 config./manufacturer)

     Total NHTSA Safety Test Cost:   $210,000


3.   Total HDGV Fuel Tank Safety Test Cost:   $242,000
4.   Cost/Vehicle (Amortized at 10 percent over 5 years of
       vehicle sales*):   $0.70
*    Assumed that  all bus  manufacturers  will crash  test  their
     vehicles.
**   Vehicle sales were estimated at 90,000/year.

-------
                              -93-
rollover valve or  pressure  relief  mechanism is very small  (0.4
Ibs),  and  because  HDGVs  are  less  sensitive to weight  changes
than  lighter  weight vehicles,  on average  less than  $0.30  in
added fuel costs will  result from their inclusion  into the HDGV
onboard system.[24]

     A total  onboard safety cost  is  calculated by summing  all
four  individual  component  costs.   Total  safety-related  onboard
costs per  family average about  $270,000,  and  the per  vehicle
costs range from $9.70 to  $14.10  or about  20-25 percent  of  the
total cost depending on the type of rollover valve  used.

     2.    Leadtime

     If EPA were to implement  an  HDGV onboard  requirement,  it
would be  necessary  to allow  manufacturers  enough leadtime  to
adequately prepare  for the  requirement.   The HDGV  preparation
process  would  involve the same  individual  tasks  that  would
enter into the light-duty process:  system  design,  development,
testing,    tooling,  certification,    and  safety   evaluation.
Although two  of  these  leadtime tasks   (certification and  safety
evaluation)  would  involve   somewhat   different  procedures  for
HDGVs, they will essentially  require  the  same  amount of  time
and would  factor into the  total  process in the same  manner  as
in light-duty.  Therefore,  it  is estimated  that  24  months would
be the total  amount of leadtime required by HDGV manufacturers,
and Figure  25 which shows  the parallel/sequential  progression
of the  individual   leadtime  components  would be  essentially the
same for HDGVs.

     Of the various leadtime components shown  in Figure  25,  all
but two would be  essentially  the  same for HDGVs  as they would
for light-duty  applications.  These two  are  certification  and
safety evaluation.   In both cases,  the HDGV processes  appear as
though  they  would  take  less  time   to  complete  than  their
light-duty counterparts because these  tasks would  be  likely to
be  less  difficult  to perform.   For  example,  in some  cases,
durability  assessments  for  certification  of  HDGVs  does  not
require any  actual  vehicle testing;   bench evaluations  can  be
substituted based   on  the  manufacturers  engineering  judgment.
This could save considerable time.

     As far  as  safety evaluation goes,  HDGV  fuel  tank  tests
performed  to  meet  OMCS  requirements   would be much simpler  to
perform than  NHTSA's safety crash  tests for passenger  cars  and
light trucks.   Also,  when  NHTSA  requirements  do  apply  (as  in
the case of school buses)  they only involve a  single crash test
with  no  rollover.   (This  is minor in  comparison to tests which
involve multiple crashes with  rollover.)  Therefore, the amount
of time allowed  for light-duty certification (10-12 months) and
safety evaluation   (>12 months)  should  also  be sufficient  for
HDGVs  since   the   heavy-duty  processes   are  less   involved.
Overall,   24  months of  leadtime  for  HDGV  onboard  is  quite

-------
                              -94-
reasonable.   This  is  especially  true  when one  considers  the
development work already completed and  the  "incremental"  nature
of onboard in relation to current  evaporative emission systems.

     EPA's  24-month  leadtime  estimate  is  supported  by  past
experience    with    previous    HDGV    evaporative    emission
rulemakings.   These  rulemakings   include  the  California  Air
Resources  Board  original  1978   6.0  g/test  HDGV  evaporative
emission standard which was  implemented  with just 21  months  of
leadtime.[53]   The  stringency  of this  standard was  increased
for  1980  model  year  HDGVs  allowing  only  2 g/test. [54]   While
this  stricter  standard  was  promulgated  with  37  months  of
leadtime, manufacturers  had to meet the  1978 standard  first,
which effectively limited the  leadtime  for  the 1980 standard to
about 24 months.  One  final  evaporative  emission  standard which
was  implemented  with  24  months of leadtime was EPA's  1985 HDGV
standard.   In  each of these  three  rulemakings,  manufacturers
faced  leadtime  factors  identical  to   the  ones  that   would
accompany  an  onboard  requirement,   including  safety.   Since
manufacturers  were   able  to  safely  and effectively  integrate
evaporative emission controls  into their  vehicle's  fuel systems
with  24  months  of   leadtime,  and  since the  magnitude  of  the
onboard  implementation task  is   similar,  manufacturers  should
also be  able to safely  and  effectively integrate  onboard into
vehicle fuel systems with 24 months leadtime.

     As far  as  safety development  and  evaluation  is concerned,
EPA's  HDGV   leadtime  estimate is  also  supported  by   the  past
experience of OMCS  and NHTSA in  implementing various  HDGV fuel
system requirements.   In 1973, OMCS extended  its  safety test
requirements to  include  previously  unaffected  non-side-mounted
(frame-mounted)   HDGV  fuel   tanks.    This  requirement   was
implemented with just 18 months of leadtime. [55]   Also in 1977,
FMVSS  301  was  extended to  include   school  buses,   and  this
requirement  was  implemented with  17 months of  leadtime.[56]
This  experience indicates  that   24  months  of leadtime  allows
manufacturers sufficient time to  factor in safety.

     Based  on  the  information provided  above,  it  appears that
24 months  is adequate time  to implement  HDGV  onboard controls,
with full consideration of all safety  concerns.   Because  safety
evaluation  can  proceed in parallel  to  three other  tasks, more
than  a  year  is  available for  actual   fuel  tank  safety  tests,
school bus  crash testing,  or  any desired  computer simulation.
This allows adequate leadtime to properly  integrate safety into
HDGV onboard systems  especially since  manufacturers can utilize
and  expand  safety  technology  used   in  current  evaporative
emission control  systems  to develop effective  onboard systems.
Also, much  of  the  safety  development  which would  be required
has  already  taken  place with  the  identification  and resolution
of such potential safety issues as rollover  protection and fuel
tank pressure relief.  Consequently, a  24-month leadtime period
would  provide  manufacturers  with  sufficient  opportunity  to
develop safe and effective onboard systems.

-------
                              -95-
     While  the  current  leadtime  estimate  of  24  months  is
reasonable  for  all  vehicle  models  including  HDGVs,   EPA  is
sensitive   to   manufacturers   concerns   regarding   leadtime
requirements.  EPA  is  committed to providing  manufacturers  the
leadtime  necessary  to  implement  onboard  controls "safely  and
effectively.   Designing  safe  onboard controls  for  some  unique
HDGVs may require more leadtime.   Such  HDGVs include those with
atypical duty  cycles,  unique  fuel  tank  or  body configurations,
and  those  HDGVs  from  secondary manufacturers.   Consequently,
EPA would include HDGVs as part of  any  overall consideration of
additional  leadtime  or   a  short  phase-in  period  for  onboard
controls.

     F.    Summary/Conelus ion

     The purpose  of this section  was  to  identify  and  address
the  potential effect^ onboard  controls  could  have on  a  HDGV
manufacturer's fuel system safety designs.   After  analyzing the
potential safety  concerns related  to implementing HDGV onboard
systems,  EPA  has  found  that  like passenger   cars  and  light
trucks,  heavy-duty  onboard  systems are  extensions of  current
evaporative systems and  corresponding safety considerations are
similar   in   nature   to  those   discussed    for   light-duty
applications.  While  a few unique  considerations  do exist  (in
part  because  of  differences  in  testing  requirements,  tank
designs/locations,   structural    integrity,    size   etc.),   no
increase  in overall risk should  be caused  or  accepted,  beyond
that  now present  with  today's HDGV   fuel/evaporative  system.
This  applies  to  both  compliance  with the  applicable  Federal
safety  standards  and the in-use  safety of  HDGVs  equipped with
onboard   systems.     As   was    the   case   for   light-duty,
straightforward engineering solutions are  available for  all of
the  potential safety  problems  which have  been identified,  and
that  while  final  choices regarding  exact  system  designs  lie
with the manufacturers,  safe  fuel system designs are achievable
by all.  EPA  estimates that  HDGV safety costs  contribute about
20-25  percent of the total ^DGV onboard system cost and should
fall  within  the  range  of  $9.70  to $14.10.    With regard  to
leadtime,  this  analysis  indicates  that  24 months appears to
provide  HDGV manufacturers with adequate  time to  prepare  for
the  safe  and effective  implementation  of  onboard controls, but
as before with passenger cars and  light trucks the possibility
of the need for more leadtime for some vehicle models may exist.

VIII. Conclusion

     EPA  has   investigated and  analyzed each of  the  potential
onboard  system safety issues raised  by the  commenters.   After
carefully  considering all of  the potential  impacts an onboard
system  could have  on the overall  safety  Of  a  vehicle's fuel
system,   it   is  concluded   that   straightforward,  reliable,
relatively  inexpensive engineering  solutions exist  for  each of

-------
                              -96-
the potential problems identified.  Furthermore, no  increase  in
risk need  occur or  be  accepted because of  the presence of  an
onboard system.  Onboard  equipped vehicles  can  be designed  to
pass  FMVSS  301 and provide  a  level  of   in-use fuel  system
integrity  equal to  or  better  than  that  achieved  on  present
vehicles   which   incorporate   evaporative   emission   control
systems.   Of  course  final  choices   regarding  exact  onboard
system   designs   lie  with   the   manufacturers,   and   each
manufacturer will choose the approach/system which provides the
best balance of  cost, safety,  and other key  factors.   EPA would
not adopt  an onboard requirement  unless it  was clear  that  safe
fuel system designs  were  available.   This  report demonstrates
this to be  the case.  Safe fuel  system designs are  achievable
by all manufacturers.

     Furthermore,  it it  is quite  possible  that  overall  fuel
system  improvements  could  accompany   the   implementation  of
onboard controls and lead  to  a net improvement in the level  of
fuel   system   safety   on  in-use   vehicles.    For   example,
collapsible  bladder  tanks  are  one  design   option  that  could
control refueling  emissions,  reduce  evaporative  emissions  and
at the same time improve fuel  system safety.

     Manufacturers  can and  are  expected to design and implement
onboard systems  in a manner which  provides  at  least the  same
level  of  fuel   system safety  as  achieved  on  present  vehicles.
In addition, a  number of  design options and other measures are
available with  onboard  systems, which  suggest  that  fuel system
safety  in-use  can  be improved  along  with  the  incorporation  of
onboard control systems.

-------
                              -97-
IX.  References

      1.   Letter, Thomas  Hanna,  MVMA and George Nield,  AIA to
Lee Thomas, US EPA, December 22, 1986.

      2.   Letter, Brian  O'Neill,  IIHS to Lee  Thomas,  US  EPA,
September 23, 1986.

      3.   Letter, Ralph  Hitchcock,   NHTSA  to Charles  L.  Gray,
Jr., US EPA, November 13,  1986.

      4.   American  Petroleum  Institute  Comments  on  US  EPA
Gasoline Marketing Study,  November 8, 1984,  Docket A-84-07.

      5.   Letter, Clarence Ditlow,   Center  for Auto Safety to
Lee Thomas, US EPA, March 20, 1987.

      6.   "Survey of Evaporative  Emission Systems  Condition of
In-Use,  High   Mileage   Automobiles",  API   Publication   4393,
February, 1985.
                                          •
      7.   Borg Warner Control Systems Catalog,  February 1986.

      8.   "Study  of   Gasoline   Volatility  and   Hydrocarbon
Emissions  from  Motor Vehicles", US  EPA,  AA-SDSB-85-5,  November
1985.

      9.   "Summary and Analysis of  Comments on  the Recommended
Practice  for the  Measurement  of  Refueling Emission",  US  EPA
AA-SDSB-87, March 1987.

     10.   Letter,  T.M.   Fisher,   General  Motors  to  James  B.
Weigold, US EPA, November 8, 1984, Docket A-84-07.

     11.   Ford  Motor  Company Comments  on  Evaluation  of  Air
Pollution  Regulatory  Strategies  for  the  Gasoline  Marketing
Industry, November 8, 1984, Docket A-84-07.

     12.   Chrysler  Corporation Comments on Evaluation  of  Air
Pollution  Regulatory  Strategies  for  the  Gasoline  Marketing
Industry, November 5, 1984, Docket A-84-07.

     13.   "Toyota  Information  on  Refueling Vapor  Recovery",
Presentation to US EPA,  March 19, 1986.

     14.   "Onboard  Control   of   Vehicle  Refueling  Emissions
Demonstration  of  Feasibility",  API  Publication 4306,  October
1978.

     15.   "Vehicle Onboard  Refueling Control",  API  Publication
4424, March 1986.

-------
                               -98-
      16.    "Evaluation  of  the Feasibility  of  Liquid Fillneck
 Seals," US  EPA AA-SDSB-86-003, December 1986.

      17.    Evaluation  of  Air  Pollution  Regulatory   Strategies
 for  Gasoline Marketing  Industry  - Response to Public Comments,
 March 1987.

      18.    "Onboard   Refueling  Vapor  Recovery   Cost  Study,"
 Mueller Associates Inc., December  1986.

      19.    "Refueling  Emissions   from  Uncontrolled   Vehicles,"
 EPA-AA-SDSB-85-6, Dale Rothman and Robert Johnson, 1985.

      20.    "Expansion   of   Investigation   of   Passenger   Car
 Refueling  Losses,"  EPA-460/3-76-006,  U.S.  EPA,   OAWM,  OMSAPC,
 ECTD,  September 1975.

      21.    "Evaluation of  Federal  Motor  Vehicle Safety Standard
 301-75, Fuel  System  Integrity:  Passenger  Cars," DOT HS-806-335,
 January 1983.

      22.    Fatal   Accident   Reporting   System,   NHTSA,   DOT,
 1980-1984.

      23.    Motor  Vehicle   Safety   Standard  No.   301-75,   Fuel
 System Integrity:  39 FR  10588,  March 21,  1974,  PART 571;  S
 301-75-5.1,  5.2,  5.3,  and  40  FR 48352,   October  15,   1975,  PART
 571:  S 301-75-5.1, 5.2,  5.3, 5.4.

      24.    Letter, David E. Martin,  GM to Barry Felrice,  NHTSA,
 March  24, 1986.

      25.    Letter,   Hiroyuki   Shinbura,   Nissan   Research   and
 Development to Charles Gray, U.S. EPA, April 14, 1987

      26.    Note from Bob Williams,  NHTSA,  to  Glenn   Passavant,
 US EPA including 3 computer file printouts, November 13, 1986.

      27.    "Design of  a  Fire  Proof  Vehicle,"  Chan,  C.Y.K./Chi,
 L.L.,  California  University,  Berkeley,   Fire  Research  Group.,
 Report No.  UCB-FRG-75-18, July, 1975.

      28.    "A  Perspective  on   Automobile   Crash  Fires",   SAE
 850092, C.  Warner, M. James, R. Wooley.

      29.    "Supplement  to  General  Motors  Commentary  to  the
Environmental Protection  Agency Relative to  Onboard  Control of
Vehicle Refueling Emissions," June 1978.

      30.    NHTSA  Press  Release,  June 11,  1987 and  NHTSA  Fact
 Sheet  on  The New  Car Assessment  Program and a Summary  of the
New Car Assessment Program Test Results,  August 5, 1986.

-------
                              -99-
     31.    Ann Arbor  News,  "'Smart'  Suspension System  Includes
Sensors;   Computers,"  Ann  Arbor News,  Newhouse  News  Service,
March 1,  1987.

     32.    "Spilled Fuel Ignition  Sources  and Countermeasures,"
Johnson,    N. ,   DOT   Contract  No.   HS-4-00872,   Report   No.
2310-75-118, September,  1975.

     33.    "Costs  of  Onboard  Vapor  Recovery  Hardware",  Jack
Faucett and Mueller Associates,  February 1985.

     34.    "Report on Quality Changes  for 1977 Model  Passenger
Cars" USDL-76-1376, BLS  November 1976.

     35.    "Report on Quality Changes  for  1976  Model  Passenger
Cars," USDL-75-626, BLS  November 1975.

     36.    "Cost   Estimations   for  Emission  Control   Related
Components/Systems    and   Cost    Methodology    Description,"
EPA-460/3-78-002, March  1978.

     37.    "Update  of  EPA's   Motor  Vehicle  Emission  Control
Equipment  Retail  Price  Equivalent  Calculation  Formula,"  Jack
Faucett Associates for U.S. EPA, September 4, 1985.

     38.    "Cost   of  Crash   Testing  to  Assure   Fuel   System
Integrity  for  Onboard Systems," EPA Memorandum,  Robert  Johnson
to  the  Record,  U.S.  EPA,  OAR,  QMS, ECTD,  SDSB,  September  2,
1986.

     39.    "Trap  Oxidizer  Feasibility Study",  U.S. EPA,  OANR,
OMSAPC,  ECTD, SDSB, March 1982.

     40.    "Characterization of Fuel/Vapor Handling Systems  of
Heavy-Duty  Gasoline  Vehicles  over  10,000  Pound  GVW",  Jack
Faucett Associates, September  1985.

     41.    "Evaluation of  Air  Pollution  Regulatory  Strategies
for   the   Gasoline  Marketing   Industry",   US   EPA,   OAR,
EPA-450/3-84-012a, July  1984.

     42.    Memorandum to File,  Review of  General  Motors  Heavy
Duty   Engine/Valued    Certification    Procedures,    Team   IV
Certification Branch,  August 4,  1986.

     43.    Memorandum to File,  Review of Ford HDE Certification
Procedures, Team IV Certification Branch, August 5, 1986.

     44.    49 CFR Part 393.65  to 393.67.

     45.    Conversation  with Jim Feiten,  GM, March 16,  1987.

     46.    Conversation  with Bob Bisaro,  Ford, March 16,  1987.

-------
                              -100-
      47.   Motor   Vehicle  Safety  Standard  NO.   301-75,   Fuel
System  Integrity:   40  FR 48352, October  15,  1975, PART  571:  S
301-75-5.4  and  41  FR  36026,   August  26,  1976,  PART  571:   S
301-75-5.4

      48.   Letter,  L.M.  Short,  Department of California  Highway
Patrol   to   Diane   Steed,   National   Highway   Traffic   Safety
Administration, May 30,  1986.

      49.   49 CFR  393.67  (c)(7)(ii).

      50.   Based  on  a  AP  of   10  psi for  Emco  Wheaton  Model
A6000 and OPW Model  7H  diesel fuel  nozzles.

      51.   Letter,  Charles L. Gray,  Jr.,  US EPA,  to Office  of
Motor Carrier Safety, May 27, 1987.

      52.   Conversation  with  Jim  Brittell,  DOT/OMCS, February
18, 1987.

      53.   Public  Hearing to Consider  Amendments to  California
Fuel  Evaporative   Emission   Test   Procedures   for  1978   and
Subsequent  Model   Gasoline-Powered  Vehicles,   Resolution  No.
76-15, March 31, 1976.

      54.   Public  Hearing  on Proposed Changes  to Regulations
Regarding Vehicle  Evaporative Emission Standards  for 1980  and
Subsequent Model Motor  Vehicles, Resolution No.  76-45, November
23, 1976.

      55.   OMCS  Regulations,  Part   393-Parts   and  Accessories
Necessary  for  Safe  Operations   (Fuel Systems):  36  FR   15444,
August 14, 1971, and 37 FR 4340, March  2,  1972.

      56.   FMVSS 301, 40  FR 48352,  October 15,  1975.
                                          * U .S. GOVERNMENT PRINTING OFFICE: 1987 - 744-622

-------
      Appendix III





Service Station Fire Data

-------
                                                         APPENDI.X Til
                                   1 90?
H;L I I'KIII'S
Sl-'UVlCli IHJUKIES
OTHER  INJURIES
SI'.I'.VICK DEATHS
OI'IIKK  DEATHS
K I KUrniKE  I-'IHKS
MUIUL!-'. I:'.I UK!!
OTIIKK  KIRKS
I-'.S'I .  I if.) 1.1 All  ''OSS
 ]
 KKHVICF. INJURIES
 OTill'.K  INJUKIKS
 SEKV1CK DKATIIS
 OTIIKK  OEATHS
 MOHJI.I:
        KIRKS
   1962

   2739
     28
    141
      0
      0
    357
   .1154
   1228
7648296
                                                         TAH1.F.  1
                                                     SERVICE STATION  FIRES

                                                         NF1KS  DATA ONLY
                                                1983          1^84
                                                      NAT IONW t DI-: PHOJKCTIONS
                                                 1983
                                                               1 984
                                                                            19H5
                                                                                     AVERAGE


21
rie
8
40
0
• 0
101
327
34H
fifil:' f
189
5
37
0
2
105
350
3U4
24594/4
8/2
9
43
0
2
88
392
392
2116459
1068
c.
J
35
0
1
117
504
44f
3140544
876
7
39
0
1
103
393
380
2470783
                                                                                      AVERAGE
2384
15
112
0
6
317
1058
1009
32530
2392
25
118
0
5
241
1075
1075
5805447
2696
13
88
0
3
295
1272
1128
7926733
2553
20
115
0
4
303
1140
1110
7203252

-------
                                                             AIM'liNMIX I I T

                                                             T/MH.K 7

                                                    TYPE or SITUATION TOUND
                                                         NFIRS DATA ONLY
STKIJ:."! UK-  I I .'I.
QUIT. in-.  • '•  . II'IJ'.TUKK  KIKE
19H2

 101
 i ;9
 377
    3
    7.
    1
 i ;'H
    0
   79

  i K,
1983

 105
 181
 350
    0
    1
    2
 131
    1
   15

 789
1984

  88
 200
 392
   2
   3
   3
 155
   2
  27

 872
1985

 117
 206
 504
    1
    4
    4
 198
    1
   33

1068
AVERAGE

      103
      192
      393
        2
        3
        4
      153
        1
       26

      876
% OF TOTAL

    11.73
    21.85
    44.88
     0.17
     0.37
     0.46
    17.46
     0.11
     2.97

   100.00
                                                     NATIONWIDE  PROJECTIONS
 I'.'l I'M"1: '•  •-  I- I id'!
 Oll'l :'. I i r  ;i  STKUCTUKK !•' IKE
 vi ill ;'i i:  -I Hi-:
 WdJ:1!! I- I 
-------
                                                            APPENDIX I LI
                                                            TAIil.t: 3
                                                  MOBILE  PROPERTY TYPE
                                                       NEIKS DATA ONLY
MOBILE PROP  UNKNOWN
NOT APPLICABLE

AUTOMOBILE
TERRAIN VEHICLES
MOTOR HOME
PA •-.!;. ROAD TRAN.S,  OTHER

: ;.,:• •.-. • ,.v!-.R 1  TON
: ':< .. r:-U?lliEK  1 TON
:.i Ml- i'RAU.ER  TRUCK
TANK TRUCK-FLAM  I.QI1
tHEIGHT RD TRANU,  OTHER
WATER TRANSPORT
HEAVY EQUIPMENT
OTHER,INVALID CODE
BLANK

TOTAL.
191)2

   i,
 333

 252
  11
   7
   5

  1 I
  29
   1
   3
   1
   1
   2
   3
 1 I 1

 776
1983

  I I
 328

 293
  12
   3
   5
   3
   0
   1
   1
   0
   3
  89

 789
1981

  12
 361

 328
  1 1
   6
   1

   9
  45
   0
   3
   0
   2
   3
   2
  I! 9

 0/2
1 98 b

  30
 119

 136
  11
  10
   3

  1 1
  3R
   3
   8
   1
   0
   1
   6
  88

.1068
                                                                                        AVh'.KAC.E
32'/
 12
  'I
  1

  9
 37
  2
  4
  1
  1
  2
  4
 91

8/6
% Of TOTA

    1.71
    41 .1

   37.33
    1 .37
     0.8
    0. 46

    0. 69
    4.22
    0.23
    0.46
    0.11
    0.11
    0.23
    0.46
   10. 73

     100
                                                   NATIONWIDE PROJECTION!,
MOnil.E PROP  UNKNOWN
NOT API'LICADLK.

AUTOMOBILE
TERRAIN VEHICLES
MOTOR HOME
PASS. ROAD TRANS,  OTHER

TRUCK-OVER 1  TON
TRUCK-UNDER  1  TON
SEMI-TRAILER  TRUCK
TANK TRUCK-FLAM LQD
FREIGHT KD TRANli,  OTHER

WATER TRANSPORT
HEAVY EQUIPMENT
OTHER, INVALID  CODE
UI.ANK
                                          1982
1175

 890
  39
  25
  18

  39
 102
   4
  1-1
   4

   4
   7
  11
 392
                                                       1983

                                                         .i!)
 U8b
  36
   9
  15

  15
 106
   9
   0
   3

   3
   0
   9
 2(,9
1984

  33
 990

 900
  30
  16
   3

  25
 123
   0
   8
   0

   5
   8
   5
 244
I 9U5

   /6
I Ob'8

1 100
   35
   25
   8

   28
   96
   R
   20
   3

   0
   3
   15
 222
                                                                                        AVI-.KAC.E
944
 35
 1 9
 1 1

 27
107
  5
 10
  2

  3
  4
 1 0
.fill'
% OV  TOTA

        2
       41

       37
        1
        1
        0

        1
        4
        0
        0
        0

        0
        0
        0
       11
TOTAL
                                          2739
                                                       2384
                                                                    2392
                                                                                2696
                                                                                                                  100

-------
                                                       T
-------
                  AI'PKNIHX 111
                    TABLE 5
        EQUIPMENT  INVOLVED IN IGNITION
UNKNOWN
HEATING SYSTEMS
AIR COND/REFK1G  EQUIPMENT
ELECTRIC DISTRIBUTION  EQUIP
  FIXED WIRING
  OTHER
AIM'1.1 AN'.T.K/KQUJ PMENT
:;i i c:;-\ • :..«iPMENT

  I !i. ;  :-i.AI  :-(.MHUST. EI1C.1 HE
  • •". ::l  i<
I-KUM M;!NG EQUIP
SEKVICE/MA INT EQUIP
OTHER
  VEHICLE
  MO EQUIPMENT  INVOLVED
  OTHER .
lil.ftHK

TOTAL
I 982

  bb
  24
   2

  II
  29
  IS

  83
  13
  sa
   n
  32

 KM
 1B8
  11
   8

 I'Ik
                NFIRS  DATA ONLY

             1983         19U4
 1]
  9
  0

 bO
 31
 10

 73
 18
 37
  I
 38

236
186
 12
 11
                                                                     n
                                                                     i

                                                                     56
                                                                     1.7
                                                                     17

                                                                     81
                                                                     39
                                                                     16
                                                                     0
                                                                     32

                                                                   250
                                                                   211
                                                                     16
                                                                     15

                                                                   872
                                     1 9Ub

                                       II
                                       12
                                        2

                                       66
                                       21
                                       11
                                       32
                                        3
                                       16

                                      2 Ufa
                                      270
                                       12
                                       13

                                     I DtU
                    AVEiiAC.E

                           fc.i
                           11
                            I

                           b.i
                           2b
                           I 1

                           H,,
                           bO
                           13
                            I
                           37

                          ^39
                          216
                           :•}
                           1 3

                          U /t>
                       % OF  TOTA

                            7.13
                            1 .60
                            0.14


                            6.OB
                            2.88
                            1 . 60

                            9.8b
                            5.71
                            4 .94
                            0.11
                            4.22

                           27.28
                           24.66
                            2.40
                            1.43

                         100.00
        MATIUtlUlDK  I'KOJ EOT IONS
UNKNOWN
HEATING SYSTEMS
AIR COND/RF.FRIG  EQUIPMENT
ELECTRIC DISTRIBUTION EQUIP
  FIXED WIRING
  OTHER
APPLIANCES/EQUIPMENT
SPECIAL EQUIPMENT
  SEPRT PUMP/COMPRESSOR
  INTERNAL COMBUST.  ENGINE
  OTHER
PROCESSING EQUIP
SERVICE/MA1NT EQUIP
OTHER
  VEHICLE
  NO EQUIPMENT  INVOLVED
  OTHER
BLANK

TOTAL
1 '.Hi 2

 1 91
  85
   7

 lib
 102
  b3

 293
 Ib2
 205
   0
 113

 6bO
 664
  49
  28

2/39
             10113

              124
               27
                0

              Ibl
              103
               30

              221
              145
              112
                3
              115

              713
              562
               36
               42
1984

 211
  30
   3

 154
  47
  47

 230
 107
 126
   0
  88

 686
 579
  41
  41

2392
 i'J4
  30
   5

 167
  53
  35

 26b
 177
  81
   8
 116

 722
 704
 I 06
  33

2696
                                AVI'.RACE

                                     181
                                       43
                                       4

                                     1 54
                                       76
                                       •11

                                     2b2
                                     lib
                                     131
                                       3
                                     1U8
                                      627

                                       36

                                     2bb3
                                                                                                      % OF TOTA

                                                                                                           "7.09
                                                                                                           1 . 69
                                                                                                           0.15

                                                                                                           6.03
                                                                                                           2.98
                                                                                                           1 .62

                                                                                                           9.88
                                                                                                           5.68
                                                                                                           5.1 3
                                                                                                           0.10
                                                                                                           4.23

                                                                                                          27.13
                                                                                                          24.57
                                                                                                           2.31
                                                                                                           1 .42

                                                                                                         100.00

-------
                                                                    APPKND1X III
IIKAIYMJ!•!.- i'WUI) OHJI-'.CT
   r.lVM'.K/'.-/..-; l-'UKI.KIl f.OP

   l;i•.."•'I'/' ',••: 1-Hr.l I-M> I-'.OI'
   :;i /.!•;:.. i  '•.>.  I uKi.r.n  V.Q\>
   li! ,YI / I ! '.I. I- l'l:l.l:'.n KQI>
   '.: ilir.l'
nrAT/i:i i '•.  i-.'.M'.   AKCING
   :'.M.i'!••!• rir.ruIT
   AH':-: .•'..'!  I Y  CONTACT
   Al'.-.:-: i Al-.K HU)M KOP/S
   i.-in; ,:
III -VI  ':.;• :' i i::   MA'IT.K.IAI,
li! ;-. i /r.i-! ;•  ii ,..:-i.,  r.PAKK
19H2

   53

   31
   26
   49
   49
   24

 I 61
   2B
   58
   10
   12

    5
   i 6
 it'i
                                                      3;
                                                      17
                                                      11
                                                       5
                                                       7

                                                     776
TAHI.K
I--OKM OK lll./\r
NFIRS DATA
1983
56
31
24
11
46
33
202
14
40
9
10
,11
25
9;
23
i

10
tl
^
20
9
6
7
6
U.NITION
ONLY
1984
97
43
18
35
50
25
190
25
48
6
14
11
28
124
23

53
12
37
4
6
11
7
5
                                                                    789
                                                                                   812
1985

   95

   48
   39
   41
   63
   35

 232
   24
   48
   10
   13
   14
   35
 177
   19

   42
   17
   64
    1
   18
   12
   11
   10

1068
AVEUAGE

     75

     39
     27
     42
     52
     29

    197
     23
     49
       9
     12

     10
     26
    125
     20

     41
     11
     43
       1
     13
     10
       7
       7

    876
%  OF  TOTA

      8.59

      4.45
      3.05
      4.74
      5.94
      3.34

    22.49
      2.60
      5.54
         00
         40
                                                                                                                                   1.17
                                                                                                                                   2.97
                                                                                                                                  14.24
                                                                                                                                   2.31
                                                                                                                                   4.97
                                                                                  .60
                                                                                  .94
                                                                                                                                   0.40
                                                                                1 .51
                                                                                1.14
                                                                                0.83
                                                                                0.83
                                                                                                                                100.00
                                                             NATIONWIDE PROJECTIONS
III- ;,!/!-:•; I.- I'liKI) (JUJKCT
   si-,-,1'.:-:, '•;,••. i-ui-'.i.iii) I-:QP
   li!'AT/<:/>!',  HJKI.lil) f.OP
   !'.! /, I. i :/'l.i ;>.  I-W.LKI)  l-'.QP
   III-.;. I1/ MO-  HIKI.ED KOI'
   U'l III-.K
ill AT/KI.!''.:.  KQP. AKCING
   SIK. I'. I1  Clhi.'UlT
   Ai'.'>i AUI.'l'Y  CONTACT
   AKu-si'Ai-K I-KIJM  KOP/SWT
   OTIII-'K
III.M'/i.!p!-:u  I--|./\HE,  SPARK
   I'.M.ri'l KI-;  I-'HOM ENGINK
   (.inn :u
Hi'AI'/llcr Ol!.)l:'.CT
   l-'l', K'l I'!IJ
   Kl.l.c.'l l< ICAI. I-'.UUIPMI'.NT
   ('I III K
li:  /'-. I /I XI I or. I VI-1.,  t'IKKWKS
11:7,1 ,';:-', . lilUil. iJOUKCK
902
187
120
92
173
173
05
579
99
205
35
42
18
56
357
56
131
60
145
25
32
1983
169
94
73
124
139
100
610
42
121
27
30
33
76
293
70
127
30
94
6
60
1984
266
118
49
96
137
69
521
69
132
16
38
30
77
340
63
145
33
101
11
16
1985
240
121
98
103
159
88
566
61
121
25
33
35
88
447
48
106
43
162
3
45
AVERAGE
216
113
78
124
152
85
574
68
145
26
36
29
74
359
59
127
42
125
11
39
% OF TOTA
8.44
4.43
3.06
4.86
5.96
3.34
22.48
2.65
5.66
1 .02
1 .41
1 .14
2.91
14.07
2.32
4.98
1.63
4. 91
0.43
1.51

-------
                                                        on           I (•
                                            50           27           30           30           29          1.13
HKAT/I I UK  UI-KKAIJlMXI-OSlWi:)                 '»           *             J9           28           21          0.81
OTIIKH                                       J"           J?           ]4           25           21          0.83
BLANK
                                          273,         2381         2392         2696         2553        100.00
TOTAL

-------
                                                         AITEMPI X
                                                           TAIU.E  I
                                                  !-'OKM  OF MATERIAL  I';:: HEP
                                                        NFIKS DATA o;;i.v
UNKNOWN
STRUCTURE  COMPONENT/FINISH
FURNITURE
CLOTHES
SUPPLIES/STOCK
POWER TRANS  EQUIP/FUEL
  ELECTRICAL WIRE
  FUEL
  OTHER
RUBBISH/TKASH
SPECIAL  E'ORM
  ATOMIZED/VAPORIZED LIQUID
  GAS/LIQUID FROM PIPF.
  OTHER
OTIIF.R

TOTAL
1911?

   6
   7
   0
   U
   r>

  i i
 59'..
   I
  11
 110
   1
  I /
1983

   4
   5
   1
   0
   1

  1 3
 61 3
   6
   1
 1 70
   1
  1 6

 789
1 9111

   3
   1
   2
   2
  I')
 6f,1
   3
   (I

  13
 1 11
   1
  I"
1985

   4
   3
   1
   1
   1

  20
 854
   3
   1

  11
 151.
   0
  i 8

1068
                                 AVERAGE

                                         1
                                         5
                                         1
                                         1
                                         3

                                        16
                                       682
                                         3
                                         1

                                        12
                                       i :n
                                         i
                                        n

                                       8/6
% OF TOTA

    0.19
    0.54
    0.11
    0.09
    0.31

    1.86
   77.80
    0.37
    0.1.1

    1.31
   1 1.98
    0.09
    1 .97

  100.00
UNKNOWN
STRUCTURE  COMPONKNT/Fl U I P.ll
KUKNITURE
CLOTHES
SUPPLrt'.S/STOCK.
POWER TRANS  EQUIP/FUEL
  ELECTRICAL WIRE
  FUEL
  OTHER
RUBBISH/TRASH
SPECIAL FORM
  ATOMIZED/VAPORiZr.D LIQUID
  GAS/LI QUID F'ROM PIPF.
  OTHER
OTI'P.R
                                          1 911?

                                            21
                                                  NATlONWmK PKO.li-'i-Tl'. :N!'.
1983

  12
  15
   3
   0
   3

  39
1852
  1 8
  . 3

  24
 363
   3
  48
                            H
                           I 1
                            b
                            5
                            It

                           11
                         1871
                            0
                            0
                          305
                            3
                           19
                         1 9H5

                           1 0
                            8
                            3
                            3
                            3

                           50
                         ?155
                            8
                            3

                           28
                          381
                            0
                           45
                     AVERAGE

                            13
                            1 5
                            3
                            2
                            9

                            48
                         19112
                            9
                            3

                            34
                          382
                            2
                            51
                         OF'  TOTA

                            0.51
                            0.57
                            0.11
                            0.08
                            0.34

                            1.87
                           77.65
                            0.37
                            0.12

                            1.34
                           14.95
                            0.09
                            1.99
TOTAL
                                                       2384
                                                                                26'JG
                                                                                             2553
                                                                                                          100.00

-------
                                                      APPEND! X  II I
UNKNOWN
INCENDIARY
SUSI'ICTOUS
MISUSE OF HEAT IGNITION
  ABANDONED  MATERIAL
  CUTTING/WELDING
  OTHER
MISUSE OF MATERIAL IGNITED
  FUEL SPILLED ACCIDENT
  IMPROPER FUELING TECIIIM'I'll;
  WASH/CLEAN/FAINT FART
  OTHER
MF.CH. FAILURE/MALFUNCTION
  FART FAILURE/LEAK/BKKAK
  SHORT CIRCUIT/GROUND FAULT
  O'i'HKR ELECTRICAL, FA I I UKH
  ! i.'~Y '"'!•' MA I M'l'l-'fu-'-!! ">•.
  I'./M'M- I I(E
  OTHER MECHANICAL FA I I.I!!'!'.
  OTHER
UEU1.GN/CONSTKCT/1NSTAL HE'-' I c I KI.'CY
OPERATIONAL  IJEFIC1 EI1CY
  COLLISION/OVERTURN/KNOCK IX •'.ill
  OTHER
NATURAL CONIJITIONK/WIIII):;
OTHER

TOTAL
              TAH1.F.  8
        IGNITION FACTORS

         NF1RT, DATA  ONLY

l"H2        19113         1IH!1

  ;•«          32           ;>5
  18          15           21
   2          1.0            7
                7
                9
               11
              76           R U
              21           ;")
                -1            5
              16           1 f,

              11V          111
              23           25
                9           ID
              1C
              '.'3          I I'j
              I G           1 G
                1            3
              11            /
  31           18           36
  10            8            1
  16           22           20

 //'G          V09          l!/2
                        1985

                          50
                          21
                           9

                           9
                           9
                          12

                         1U.3
                          11
                           9
                          20

                         209
                          3 )
                          10
                          \y.
                         I 1G
                          20
                           3
                          I 1

                         2V1
                          23
                           3
                          31

                        1060
                     AVERAGE

                           11
                           20
                            7
                            0
                            8
                            9
                           12
                            0
                           8!i
                           29
                            7
                           17
                            0
                          I 5 9
                           28
                            q
                            ')
                          1 1 5
                           1 9
                            2
                           12
                            0
                          231
                           28
                            6
                           22

                          876
                         %  OF TOTA

                             1.71
                             2.23
                             0.80
                             0.00
                             0.91
                             0.97
                             1.31
                             0.00
                             9.73
                             3.31
                             0.80
                             1.91
                             0.00
                            18.12
                             3.20
                             0.97
                             1 .00
                            13.10
                             2.20
                             0.20
                             1.37
                             0.00
                            26.71
                             3.17
                             0.71
                             2.51

                           100.00
UNKNOWN
INCENDIARY
SUSPICIOUS
MISUSE OF HEAT  IGNITION
  ABANDONED  MATERIAL
  CUTTING/WELDING
  OTHER
MISUSE OF MATERIAL 1GNITEI1
  FUEL SPILLED  ACCIDENT
  IMPROPER FUELING TECHNIQUE
  WASH/CLEAN/PAINT PART
  OTHER
MECH. FAILURE/MALFUNCTION
  PART FAILURE/LEAK/BREAK
  SHORT CIRCUIT/GROUND FAULT
  OTHER ELECTRICAL FAILURE
  LACK OF MAINTENANCE
  BACKFIRE
  OTHER MECHANICAL FAILURE
  OTHER
DEStGN/CONSTRCT/INSTAL DEFICIENCY
OPERATIONAL  DEFICIENCY
  COLLISION/OVERTURN/KNOCKDOWN
  OTHER
NATURAL CONDITIONS/WINDS
OTIIF.R
19N2

  99
  G1
   7

  28
  1.H
  21

 2U9
  78
  35
  56
  95
  18
  21
 361
  (1 8
   0
  56-

 713
 120
  35
                                                  MATTONW I DE  PROOF.'"!' I O!J.r

                                                     . 1983         1981
 97
 15
 30

 21
 27
 33

230
 73
 12
 18

1 ••• 1
 70
 27
 30
28?
 18
  3
 33

680
 51
 21
 6G
I 51
 58
 1 9

 25
 30
 17

219
 80
 11
 11

395
 C9
 27
 I 9
M5
 11
  8
 19

617
 99
 11
1.985

 126
  61
  23

  23
  23
  30

 260
 103
  23
  50

 528
  93
  25
  30
 369
  50
   8
  35

 692
  58
   8
  78
AVERAGE

     1 18
       57
       20

       21
       21
       33

     250
       83
       21
       19

     161
       82
       21
       25
     331
       58
       5
       36

     683
       83
       20
       61
% OF TOTA

    1.63
    2.22
    0.78

    0.95
    0.96
    1.29

    9.78
    3.26
    0.82
    1 .92

   18.05
    3.20

    0.95
    0.99
   13.07
    2.26
    0.18
    1.11

   26.75
    3.21
    0.76
    2.51
TOTAL
                                                      23H1
                                                                     9 2
                                                                               2696
                                                                                                          100.00

-------
                                                  AH'ENIMX  I I I
                                                  TAI'LE 9
                                        r-ALIM'KNIA SERVICE STATUUI FIRES
                                                  Ml- I RS DATA ONLY

                                                       19113
                                                                                          AVKKAGi:
    I I'r'li IT.
    VI'T ll.'.JUKlF.r;
   !• i'  MUURII'-'.i;
OlIii.H  Dl'.ATMS
f.TKUCTUKli FiRr.S
MO1ULH t'iKES
OTIIKR  FIRES
KiJT. DOLLAR LOSS
 I)
 0
11
                      71
                       u
                       1
                       o
                       u
                      ?n
                      r;
                      3;
     n
     o
     6
     o
     i
     IB
     13
     10
313501
        0
        1
        0
        0
       1  /
       19
       11
    :.915G
LNCJ DliNTS
SliKVICE INJURIES
OTHF.R  INJURIES
SERVICE DEATHS
OTHER  DEATHS
STRUCTURE FIRES
MOBILE FIRES
OTHER  FJRES
EST. DOLLAR LOSS
19H?

  '.'9
   0
   ?.
   0
   0
  If.
  :n
  b?
   TALI I-'OI-'MIA PROJKCTIONS

                  1983

                    82
                     0
                    ' 1
                     0
                     0
                    22
                    19
                    11
               266391
  1981

    79
     0
     7
     0
     1
    20
    14
    -H
1117986
AVERAGE

       i) /
        0
        1
        0
        0
       19
       21
       16
  287661

-------
                                                APPENDIX  I I I
                                                 TABLE  10
                                      TYPF.  OP1 SITUATION FOUND
                                      C:AI.IFOUHIA NFIRS DATA  ONLY

                                           1911?         1983         19H4 '
STRUCTURE  FIRE
OUTSIDE  Of STIUICTUKK
VEHICLE  FIRE
I'RUf.H FIRE
'• ; : ••••!' FIRE
     .  :<;N  (U/O FIK! )
  ...ii-.i-1  SI'ILL/LEAK
: •..'.>.! i- ici ENT i NFO
<;| ili-.K
 1
 2
 0
 0
n
n
 0
 2
 I
 0
 0
11

M
  18
  ?0
  13
   0
   0
   1
   0
   0
  19
AVMUAGK

       17
       ?3
       19
       0
       1
       1
       0
       0
       16

       'i a
OF TOTAL.

 22.22
 29.19
 21. V 9
  0.43
  1 .28
  1 .'/I
  0.00
  0.00
 20.09

100.00
                                      CAM I-'ORN IA  HRO.JKCTroNS

                                           !'>f)2        1983
STRUCTURE  FIRE
OUTSIDE  OF- STRUCTURE: FIKF.
VEHICLE  FIRE
BRUSH FIRE
REFUSE FIRE
EXPLOSION  (W/O F1KK)
OUTSIDE  SPILL/LEAK
INSUFFICIENT INFO
OTHER
1 6
29
31
 1
 1
 2
 0
 0
19
26
19
 0
 2
 1
 0
 0
12
1.904

  20
  22
  14
   0
   0
   1
   0
   0
  2.1
AVERAGE

       19
       26
       21
       0
       1
       1
       0
       0
OF TOTAL

 22.22
 29.49
 24.79
  0.43
  1 .20
     71
  0.00
  0.00
 20.09
TOTAL
                                             99
                                                          8?
                                                                      79
                                                                                   87
                                                                                              100.00

-------
                                                       i i t
          TAIll.li 1.1
   AKI A  OF  FIRE ORIGIN

CAI.lFfiUNIA  IIFTRS DATA ONLY

     1982        1983
MRANS OF  EGRESS
ASSEMBLY /SALES AREAS
FUNCTION  AREAS
STORAGE  AREAS
  PRODUCT STORAGE ROOM
  GARAGE/CARPORT/STORACF.
  OTHER
SERVICE  FACILITIES
SERVICE/EQUIPMENT AREAS
  MAINTENANCE SHOP/ARF.A
  OTHER
  NOT CLASSIFIED
STRUCTURE AREAS
TRANSPORT/ VEHICLE AKF.A
  TKI'NK AREA
  •  '.  • •:•• AKF.A
   ::i ..iiV.Y,Y/TUUI,:iC  WAY
   Li\i-ilJ/l- IF.LD/Off.N ARI-'.A
   IJ'.;T  APPI.1CAMLE
   OTIIF.R

TOTAL
 1
1 6
 T

 U
 (I
;M
 f,
 0
 1

 8
 4
 ;
 3
1 3
 2

 1
 0
11
 1
 0
                                                       74
                       1984

                           1
                           2
                           5

                           5
                           2
                           6
                           4
              1
             1 6
              3

              0
              0
             12
              1
              0
              7

              1
              4
             ' 4
              0

             VI
                      AVERAGE

                           1
                           1
                           4

                           4
                           2
                           4
                           5
                                               2
                                              15
                                               3

                                               0
                                               0
                                              16
                                               3
                                               0
                                                                                             OF TOTA
 1 .VI
 'i . '. 6

 4. '0

 4 . '0
 r.. '< 8

 4. ? /

 2.14
19.23
 (.42

 I). 4 3
 II. I)U

 .!. 12
 U. 00
 ::.  I 4

 4.  10
 'j .  13
 (•. I! 4
 I .  M
MEANS  OF EGRESS
ASSEMBLY/SALES AREAS
FUNCTION AREAS
STORAGE AREAS
  PRODUCT STORAGE  ROOM
  GAKACF./CAI!POKT/:>TO!:.V::
  OTHER
SERVICE FACILITIES
SERVICE/EQUIPMENT  ARF.AS
  MAINTENANCE SHOP/ARF.A
  OTHER
  NOT  CLASSIFIED
STRUCTURE AKHAS
TRANSPORT/VEHICLE  AI'.F.A
  PASSENGER AREA
  TRUNK AREA
  ENGINE AREA
  FUEL TANK
  EXTERIOR SURFACF.
  NOT  CLASSIFIED
OTHER
  HIGHWAY/PUBLIC MAY
  LAWN/FIE1.D/OPEN  AREA
  NOT  APPLICABLE
  OTHER

TOTAL
                                    CAI.I FflUMIA  i-'KOJKCTl'ONS
                                                     1903
                     1
                     e

                     3
                     2
                     3
                     7

                     6
                     3
                    1.1
 I
 0
1 6
 1
 0

 7.

 2

 6
 1

82
           1984

              1
              2
              e

              6
              2
              7
              4

              2
              1
             18
              .3

              0
              0
             13
              1
              0
              1
              4
              4
              0

             79
                                                                          AVERAGE
                                              IV
                                               3

                                               0
                                               0
                                              IV
                                               3
                                               0
                                               4
                                               4
                                               6
                                               1

                                              n i
                                                      I .20
                                                      I .VI)
                                                      fi . 'i 3

                                                      4. 68
                                                      2.SS
                                                      4. 6H
                                                      .•;. 95
                                                      2 . I 3
                                                     19.14
                                                      i.40

                                                      0.43
                                                      0.00
                                                     19. 99
                                                      3. 40
                                                      0.00

                                                      2.13
                                                      5.10
                                                      b.80
                                                      1 . /()

-------
                                                 TAHI.I1. 1?
                                     K.yllll'M!--|n  INVOLVED IN IGNITION
                                                       HAT/, oiii.v
UNKNOWN
HEATING SYSTEMS
AIR COND/REFR1G  EQUIPMENT
ELECTRIC DISTRIBUTION EQUIP
  FIXED WIRING
  OTHER
APPLIANCES/EQUIPMENT
SPEC I A!. KOUIPMENT
  5;i.l KT I'UMl'/COMPRESSOR
  ir:!i-.K!:/>i. COMBUST.  F.NCINK
  O! lll-.R
PROCESSING EQUIP
SERVICE/MAIN1!' EQUIP
OTHER
  VEHICLE
  NO EQUIPMENT INVOLVED
  OTHER
BLANK

TOTAL
10
 2
 9
 '1
 0
 1

21
19R3

   0
   1
   0

  20
   5
   1

   1
  10
            15
            .15
             0
             n

            '/I
1 981

   4
   0
   0

  13
   2
   2

   1
   1
   6
   0
               18
               19
                0
                0

               71
                                                                          AVER AGP.
       4
       1
       0

      14
       4
       2

       1
       8
       5
       0
              18
              18
               1
               0

              18
% OF TOTA

    1. TO
    i. n
    o.oo

   18. 38
    4. 70
    2. 56

    l.'Jl
    9.83
    5. 98
    0. 00
    ?. 99

   23.08
   23. 50
    0. 85
    0.00

  100.00
                                     CA1.I Fc>KI! I A  PROJECT IONS

                                         19(1?         1983
UNKNOWN
HEATING SYSTEMS
AIR COND/REFRIG  EQUIPMENT ,
ELECTRIC DISTRIBUTION EQUIP
  FIXED WIRING
  OTHER
APPLIANCES/EQUIPMENT
SPECIAL EQUIPMENT
  SEPRT PUMP/COMPRESSOR
  INTERNAL COMBUST.  ENGINE
  OTHER
PROCESSING EQUIP
SERVICE/MAINT EQUIP
OTHER
  VEHICLE
  NO EQUIPMENT  INVOLVED
  OTHER
BLANK
 3
 0

1 1
 4
 3

 2
10
 4
 0
 3

23
23
 2
 (I
   0
   1
   0

  22
   6
   1
  11
   4
   0
   2

  n
  T7
   0
   0
1904

   4
   0
   0

  14
   2
   2

   1
   4
   7
   0
   2

  20
  21
   0
   0
AVERAGE

       4
       1
       0

      16
       4
       2

       1
       9
       5
       0
       3

      20
      20
       1
       0
% OF TOTA

    4.G8
    1.70
    0.00

   18.29
    4. 68
    2.55

    1.70
    9.78
    5.95
    0.00
    2.98

   22.97
   ?3.39
    0.85
    o.no
TOTAL
                                                                     79
                                                                                            ino.oo

-------
                                              APPENDIX  III
                                               TAFtt.e  i)
                                     FORM OK HEAT  IGNITION
                                     CALIFORNIA NPIRS  DATA  ONLV
UIIKNONN
HEAT/FUEL-PNRD OBJECT
  SPARK/CAS  fUr.LtU UOT
  HEAT/GAS FUELED EOS'
  SPARK/LI0. FUELED EOF
  HEAT/LIO.  FURLED COP
  OTHER
HEAT/ELEC. EOF. ARCING
  SHORT CIRCUIT
  ARC-FAULTT COHTACT
  ARC-SPARK  FROM  EOP/SHT
  OTHER
HEAT/SHOKIMG MATERIAL
MEAT/OPEN FLAME,  SPARK
  TORCHES
  HATCH/Llr.HTRR
  BACKFIRE FROM ENGINE
  OTHER
HEAT/HOT OBJECT
  FRICTION
  ELECTH1CAL F.OUIPMENT
  OTHER
HEAT/EXPLOSIVE, FlRf.HKS
MEAT/NATURAL SOURCE
HEAT/FIRE SPREAD (EXPOSUIIF.)
OTHER
BLANK

TOTAL
UNKNOWN
HF.AT/FUEL-PHRD OBJECT
  SPARK/CAS  FUELED EQP
  HEAT/CAS FUELED COP
  SPARK/LIQ.  FUELED EQP
  HEAT/LIO.  FUELED EOF
  OTHER
HEAT/ELEC. EOP. ARCING
  SHORT CIRCUIT
  ARC-FAULTY  CONTACT
  ARC-SPARK  FROM EOP/SMT
  OTHER
HEAT/SKOKINC  MATERIAL
HEAT/OPEN FLAME,  SPARK
  TORCHES
  HATCH/LIGHTER

  BACKFIRE FROM ENCFNF.
  OTHER
HEAT/HOT OBJECT
  FRICTION
  ELECTRICAL  KOIIIPHENT
  OTIItR
iii:AT/i:*ei.osivt,  FIREHKS
IIPAT/NATURAL  SOURCE
HEAT/FIRE SPKF.AD (EXPOSURE)
OTHER
niJtNK
1982
€
2
4
5
6
1
18
1
6
2
3
0
4
6
1
8
0
B
4
1
I
0
0
1»83
O
0
0
4
3
2
30
4
7
2
0
0
S
3
3
8
0
1
1
0
0
0
1
1984
7
0
0
3
2
0
20
3
4
0
I
1
1
3
4
9
0
4
3
1
1
0
0
                                            '•MA  PROJF.rrtnNS

                                            I?         1983

                                             I            0
70
 1
 7
 2
 3

 0
 4

 7
 3

 9
 0
 9
 4
 1
 1
 0
 0
 0
 0
 4
 3
 2

33
 4
 a
 2
 0

 0
 6

 3
 3

 9
 0
 1
 1
 0
 0
 0
 1
1984
7
0
0
3
2
0
20
3
4
0
1
1
1
3
4
9
0
4
3
1
1
0
0
71
1984
8
0
0
3
2
0
22
3
4
0
1
1
6
3
4
10
0
4
3
1
1
0
O
AVERAGE
4
1
1
4
4
1
23
)
(
|
1
0
S
4
3
8
0
4
3
1
|
0
0
78
AVF.RACE
5
1
1
4
4
1
?5
3
C
1
1
0
5
4
4
9
0
S
3
1
1
0
0
V OF TOTA
5.56
0.85
l.M
5.13
4.70
1.28
29.06
3.42
7. 26
1. II
1.71
0.43
5.98
5.13
4.21
10.68
0.00
5.56
3.42
0.85
0.85
0.00
0.43
100.00
1 OF TOTA
5.53
o.es
1.70
S.10
4.68
1.28
28.92
3.40
7.23
1.70
1.70
0.43
5.95
5.10
4.2S
10.63
0.00
S.53
3.40
0.85
0.85
0.00
0.43
TOTAL
                                                                                 87
                                                                                           100.00

-------
                                               AITKNI) IX 111
UNKNOWN
STRUCTURE COMPONENT/FT N TSII
FURNITURE
CLOTHES
SUPPLIES/STOCK
POWF.R  TRANS EQUIP/FUEL
  ELECTRICAL HIKE
  FUEL
  OTHER
RUBBISH/TRASH
sriiciw,  i-ci:;;
  ATOMIZED/VAPORIZED LIQUID
  GAS/LIQUID FROM PIPE
  OTHER
OTHER
                                               T.M1LF.  ] 1
                                      FORM <1l-' MATKRIM. inNITFD

                                      CALIFORNIA  NF1RS DATA ONLY

                                          1911?         19R3
   0
   1
   0
   0
   0

   0
  63
   0
   1

   0
   9
   0
   0

  M
                       19H4

                          0
                          1
                          0
                          0
                          7.

                          0
                         Ml
                          0
                          0

                          1
                          8
                          0
                          1

                         71
                                                                            AVERAGE
             o
             o

             i
             9
             0
             0

             78
% Of TOTA

    0.00
    0.85
    0.00
    0.00
    0.85

    0.00
   B5.01
    0.00
    0.43

    1 .28
   11.11
    0.00
    0.43

  100.00
UNKNOWN
STRUCTURE COMPONKNT/F'l N I Sll
FURNITURE
CLOTHES
SUPPLIES/STOCK
POWF.R  TRANS EQUIP/FIJI:1.!.
  ELECTRICAL WIRE
  FUEL
  OTHER
RUBBISH/TRASH
SPECIAL  FORM
  ATOMJ./ED/VAPUIUi'.KU  l.l'.i'iii
  CAS/ LIQUID PROM P'IPK
  OTHER
OTHER

TOTAL
                                      CA1.1 fORN I A   ..EJECTIONS
1 9113

   0
   1
   0
   0
   0

   0
   10
   0
 .  I

   0
   10
   0
                                                                            AVERAGE
1 0
 II
 o
 o

 i
 9
 0
 1

70
 0
 0

 1
10
 0
 0

8'7
                                                OF TOTA

                                                  0.00
                                                  0.85
                                                  0.00
                                                  0.00
                                                  0.85

                                                  0.00
                                                 84.63
                                                  0.00
                                                  0.43
   11.06
    0.00
    0.43

  1 00. 00

-------
                                              Ai'l I-.NDIX  1 I I
                                               TAI.H.K Ib
                                            I CHIT I ON FACTORS

                                      CALIFORNIA NT IRS DATA ONLY
UNKNOWN
INCENDIARY
SUSPICIOUS
MISUSE OF HEAT IGNITION
  ABANDONED  MATERIAL
  CUTTING/WELDING
  OTHER
MISUSE OF MATERIAL IGNITED
  FUEL SPILLED ACCIDENT
  IMPROPER FUELING TECHNIQUE
  WASH/CLEAN/PAINT PART
  OTHER
MF.CH. FAILURE/MALFUNCTION
  PART FA1LURE/LEAK/BREAK
  SHORT CJKCUIT/CROUND FAULT
  OTHER ELECTRICAL FAILURE
  LACK OF MAINTENANCE
  i'.A"- :  ' i-;
    I :  .  •••: .'ilANICAL FA1 LURK

i.: ..:  :./<:I:MSTKCT/INSTAL m-:r ici KI-H-Y
i-.j.'EKAl IOHAL  DEFICIENCY
  COI.|.ir,TON/OVF.RTlJKN/KN!>::Klvr.-!ri
  OTHER
NATURAL CONDITIONS/WINDS
OTHER

TOTAL
 3
1 U
 1

 2
 0
 1

 6
 -1
 0
  I

1 (,
 0
 0
 2
 0
 G
 0
                                            (I'I
          1983

              1
              1
              7

              0
              0
              6

              5
              3
              0
              0

             12
              3
              0
              1
              0
              1
              0
              1

             33
              1
              0
              1

             71
1981

   1
   7
   2

   1
   0
   0

  10
   1
   0
   1

  10
   1
   ]
   1
   0
   1
   1
   1

  71
   4
   0
   1
                                                                           AVERAGE
 2
 7
 2

 1
 0
 2

 7
 3
 0
 1

13
 1
 0
 1
 0
 3
 0
 2

?8
 3
 0
 2

/(I
I OF IOTA

    2.14
    8.97
    2.14

    1.28
    0.00
    2.99

    0.97
    3.12
    0.00
    0.85

   1 6.24
    1.71
    0.43
    1.71
    O.OU
    3.4?
    0.43
    2.14

   36.32
    4.27
    0.00
    2. 56

  101). 00
                                          FDIUM A PROJECTIONS

                                          l»ll?         19H3
UNKNOWN
INCENDIARY
SUSPICIOUS
MISUSE OF HEAT IGNITJON
  ADANDONED MATERIAL
  CUTTING/WELDING
  OTHER
MISUSE OF MATERIAL IGNITED
  FUEL SPILLED ACCIDENT
  IMPROPER  FUELING TECHNIQUE
  WASH/CLEAN/PAINT PART
  OTHER
MF.CH. FAILURE/MALFUNCTION
  PART FAILURE/LEAK/BREAK
  SHORT CIRCUIT/GROUND FAULT
  OTHER ELECTRICAL FA FLUKE
  LACK OF MAINTENANCE
  BACKFIRE
  OTHER MECHANICAL FAII.H'M-
  OTI IK><
DESIGN/CONSTRCT/INSTAL DF.K I r I Rg,-y
OPERATIONAL DEFICIENCY
  COLLtSION/OVERTURN/KNOCKPOViU
  OTHER
NATURAL CONDITIONS/WIND.';
 :)
11
 i
 2
 0
 I

 I
 f\
 0
 1

111
 0
 0
 0
 7
 0
 3

28
 6
 0
 4
              1
              4
              7

              0
              0
              7

              6
              3
              0
              0

             13
              3
              0
              1

              0
              1
              0
              1

             37 .
              1
              0
              1
I 481

   1
   8
   2

   I
   0
   0

  11
   1
 .  0
   1

  11
   1
   0
   1
   1
   1

  30
   4
   0
   1
                                                                           AVI-.KACE
 0
 1

1 4
 1
 0
 1

 0
 3
 0
 2

31
 4
 0
 2
  OF  TOTA

    2. 13
    8. 93
    2.13

    1.28
    0.00
    2.98

    8.93
    3.40

    0.00
    0.05

   16. 16
    1.70
    0.13
    1 . 70

    0.00
    3. 40
    0.43
    2.13

   36. 15
    1.25
    0.00
    ?.'JS

-------