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 IS25 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 1525 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^Di 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
------- |