@eAA OM The Engineering Society
`wii. ForAdvancing Mobility
`Land Sea Air and Space,
`
`400 commoNWEALTH DRIVE, WARRENDALE, PA 15096-0001 U.S.A.
`
`901110
`
`The Developmentof a Real-Time
`Evaporative Emission Test
`
`Harold M. Haskew
`William R. Cadman
`ThomasF. Liberty
`
`Powertrain Control Center
`Current Product Engineering
`General Motors Corporation
`
`Government/Industry
`Meeting and Exposition
`Washington, DC
`
`pay 4190
`S. A. bees 9
`LIBRARY
`LO-~ PLFA
` MAHLE-1027
`* MAHLE-1027
`U.S. Patent No. RE38,844
`U.S. Patent No. RE38,844
`
`

`

`
`
`
`
`
`
`The appearance of the ISSN code at the bottom of this page indicates SAE’s consent that
`copiesof the paper may be made far personalorinternal use of specific clients. This consent
`is given on the condition, however,that the copier pay a $5,00 per article copy fee through the
`Copyright Clearance Center,Inc., Operations Center, 27 Congress St., Saiem, MA 01970 far
`copying beyondthat permitted by Sections 107or 108 oftha U.S. Copyright Law. This consent
`does not extand to other kinds of copying such as copying fer general distribution, for
`advertising or promotional purposes, for creating new collective works, or for resale.
`SAEroutinely stocks printed papers for a period of three years following date of publication.
`Direct your orders to SAE Customer Service Department.
`
`
`
`
`
`
`
`To obtain quantity reprint rates, permissionto reprint a technical paper or permission to use
`copyrighted SAE publications in other works, contact the SAE Publications Group.
`
`
`
`SAE GLOBAL MOBILITY DATABASE ©
`
`All SAEpapers are
`abstracted and indexed in
`the SAE Global Mobdity Database,
`
`
`
`
`Nopartofthis publication may be reproduced iin anyform,inant
`
`
`/ electronic retrieval system or otherwise, without the prior writ-
`atenpermission of the publisher,
`ISSN:0148-7191
`‘Copyright 1990 Society of Automotive Engineers,Inc.
`
`Positions and opinions advancedin this paperare thosaof the
`author(s} and not necessarily those of SAE. The authoris solely
`responsible forthe content of the paper. A processis available
`by which discussions will be printed with the paperif it
`is
`published in SAE Transactians, For permission ta publish this
`paperin full or in part, cantact the SAE Publications Division.
`
`
`
`Persons wishing to submit papers to be considered for presen-
`tation or publication through SAE should send the manuscript
`ora300 word abstract of a proposed manuscriptto: Secretary,
`Engineering Activity Board, SAE.
`
`Printed in USA
`
`

`

`901110
`
`The Developmentof a Real-Time
`Evaporative Emissions Test
`
`Harold M. Haskew, William R. Cadman, ThomasF. Liberty
`
`Powertrain Control Center
`Current Product Engineering
`General Motors Corporation
`
`ABSTRACT
`
`In recent years various parties have proposed
`New evaporative emission test procedures focused
`on controlling "excess" evaporative emissions,
`on hot "ozone prone" days. Studies by General
`Hotors established the need for real-time meas-
`urements of daily emissions from parked vehicles
`and of "running losses" from vehicles that are
`driven to quantify and control
`the mobile source
`contribution to VOC inventory.
`"Resting losses"
`are shown to be a previously unidentified major
`source of hydrocarbon emissions. This paper
`describes the theories, data and development of
`GM’s Real-Time Test Procedure.
`
`WRITERS NOW Characterize the 1990’s as "The
`Decade of the Environment."
`Public concern for
`the environment
`is at a new high. Many urban
`areas, for example, exceed the National Ambient
`Air Quality Standard (NAAQS) for czone.
`The
`problem is particularly severe in California.
`Figure 1
`shows the 1988 hourly ozone measure-
`ments for four California locations.
`The hori-
`zontal
`line indicates the 0.12 part per million
`{ppm} standard that should not be exceeded more
`than once a year on average.
`As Figure 1
`indi-
`cates; Azusa exceeded the standard 125 times in
`
`1988.
`
`Figure 2, on the other hand, shows the locations
`of the nine most severe ozone areas identified
`by EPA. Figure 3 shows the 1988 ozone monitor
`records for four of these cities, Chicago, New
`York, Houston, and Kenosha.
`These exceedances
`are not as chronic as the California results.
`
`Against that backdrop, additional control meas-
`ures for all organic compounds,
`including the
`
`hydrocarbons from motor vehicles, are being
`considered. The current exhaust and evaporative
`emission test procedures and standards have
`achieved significant reductions in in-use emis-
`sion leve|s, but they are now more than a decade
`old.
`[2]
`Combinations of higher ambient
`tem-
`peratures, high fuel volatility, extended driv-
`ing, and multiple day park episodes have re-
`vealed higher levels of evaporative emissions
`than were previously assumed to exist.
`
`In recent years, and particularly since 1988,
`there have been significant advances in the
`measurement and characterization of
`in-use
`evaporative emissions.
`[2]
`These advances have
`demonstrated that the current procedures and
`standard, as well as the assumptions about
`in-
`use vehicle performance upon which they were
`based, are obsolete. This important point
`is
`largely conceded by most participants in the
`current rulemaking process.
`A new focus on high
`temperature, "“ozone-prone" conditions for con-
`trol of evaporative emissions has resulted.
`
`in
`The Environmental Protection Agency (EPA)
`January, 1990 proposed new evaporative emission
`test procedures designed to further reduce the
`mobile source contribution to air pollution.
`Another proposal,
`the subject of this paper, has
`been developed by General Motors.
`[3] Another
`new test procedure from the California Air
`. Resources Board (CARB)
`[4]
`included many fea-
`tures of the GM Real-Time proposal.
`
`EPA’s new evaporative emission control proposal
`[5] differs substantially from the CARB and GM
`proposals.
`The EPA proposal’s most significant
`change from the current test protocol
`is the use
`of multiple diurnal heat builds (two rather than
`one) conducted at higher temperatures (72 to
`96°F instead of 60 to 84°F)
`thought
`to be
`
`1. Numbers in brackets designate references listed at the end of the paper.
`
`

`

`
`
`
`
`Galvestan™s
`
`a
`3
`
`PASADENA, CALIFORNIA
`
`Pus
`daN
`
`Ae aes
`+
`FES
`
`A
`eS
`HAR
`
`!
`
`a
`
`.
`
`APATL
`
`May
`
`SUNE
`
`i
`JULY
`
`AUG
`
`SEPT
`
`oer
`
`Hf
`Ma
`eea ff
`NOV
`tec
`
`SAN BERNARGING, CALIFORNIA
`
`2 2
`a
`8
`WwW
`0
`z=
`x
`aex
`a
`a
`.2
`a
`
`.
`
`.
`
`:
`
`0
`
`MAR
`
`5 intuit
`
`it} Bl
`}
`Haig
`JUNE
`
`en
`
`May
`
`APRIL
`
`
`
`Jan
`
`FEE
`
`
`MAR
`APRIL,
`MAY
`SUNE .
`JULY
`AUG
`SEPT
`CcT
`NOY
`BEE
`
`a4
`
`3 e
`
`e
`«il
`
`6
`
`Figure 2,
`
`Houston, e
`
`901110
`
`AZUSA, CALIFORNIA
`
`mh
`
`i
`
`a
`
`I
`
`Figure Il.
`HOURLY AMBIENT ATA MEASUREMENT - OZGNE
`CALIFORNIA LOCATIONS ~ GALENDAR YEAR 1986
`
`
`
`
`
`1
`Tay
`1 ht af
`I
`|
`bl
`H
`
`
`RVE)
`sain
`PrOpeMeEN Fa
`finde
`Tata
`iy
`
`
`| hie Adehilo|
`
`SEPT
`AUB
`°
`oan
`FES
`WAR
`APRIL
`MAY
`JUNE
`uULY
`oct
`hay
`DEC
`
`“
`
`a “
`
`.
`
`

`

`901110
`
`Figure 3.
`
`HOURLY AMBIENT AIA MEASUREMENT — OZONE
`NON-CALIFOANEA LOCATIONS ~ CALENDAR YEAR 1986
`
`CHICAGG,
`
`cj
`
`Laboyia
`FES
`
`doting
`MAR
`
`A
`
`JAN
`
`i
`
`APAIL
`
`MAY
`
`"UB
`
`SEPT
`
`a
`OCT
`
`coat
`
`bol
`noY¥
`
`ils Ai
`BEC
`
`JAN
`
`ee
`FEB
`
`Be nhA A ee ee hey
`wan
`APRIL
`HAY
`
`“AUG
`
`SEPT
`
`og =
`ocT
`
`NOY
`
`Busty
`Ec
`
`..4.o-nyWw»
`
`
`
`
`
`HOUALYOZONE(ppm)
`
`ILLINOIS ale Wl
`NEW YORK, NEW YORK elyHUN Ds
`HOUSTON, TEXAS
` Les
`
`rr)
`
`“APAIL i
`
`JAN
`
`FEB
`
`HAR
`
`APAIL
`
`HAY
`
`Af
`
`AUG
`
`SEPT
`
`i Lis
`
`CCT
`
`NOY
`
`DEC
`
`EPA’s new
`representative of high ozone days.
`procedure relies on various design-review re-
`quirements and test procedure changes to influ-
`ence vehicle purge rates in order to regulate
`"yunning losses".
`It does not directly measure
`vehicle running Eosses.
`It also leaves the hot
`soak measurement procedure at
`the laboratory
`ambient (i.e., 76 instead of 96°F} temperatures.
`
`inctudes direct measure-
`California’s proposal
`ments for high-temperature hot soaks and running
`losses.
`It atso measures emissions from vehi-
`cles in extended-park situations on a “real-
`time" (24 hour} basis, rather than with the
`traditional
`time-compressed diurnal
`test proce-
`dure.
`It shares many of the same features
`with the GM proposal.
`The major differences, are
`the daily high temperature (105°F vs. 96°F),
`three diurnals vs.
`two,
`the tank fill level
`(40%
`vs. 60%), and the standard.
`
`The purpose of this paper is to describe the
`theories, data, and
`tests
`that were part of
`the development of the GM "Real-Time" Environ-
`
`mentally Based Evaporative Test Procedure.
`
`There are four parts to this paper:
`
`1.
`2.
`3.
`4.
`
`Test Procedure Goals
`Test Methodologies
`Test Parameters
`the Standard and
`The Procedure,
`Estimated Benefits
`
`TEST PROCEDURE GGALS
`
`the Real-Time test procedure
`As explained below,
`would require manufacturers to optimize vehicle
`evaporative emissions control during a very high
`percentage of hot urban driving episedes and
`during extended-park "diurnal" events.
`It would
`require stringent control of "resting losses",
`which are a large and growing segment of the
`uncontrolled in-use vehicle evaporative emis-
`sions inventory.
`As a result,
`the Real-Time
`test procedure will almost certainly involve
`increased compliance risks in the early years of
`the new program.
`
`

`

`
`
`The compliance risks of a new test procedure
`like the Real-Time procedure must be considered
`in light of the alternatives. Other test proce-
`dure proposals that presented the new under-
`- standing of "excess" evaporative emissions have
`“differed from the Real-Time procedure in that
`they did not
`include direct measurement of
`either running losses or resting losses. Direct
`measurement, however, will be critical
`in ensur-
`: Ing that “excess” evaporative control
`systems
`_ perform effectively in actual use.
`
`If direct measurement procedures are inevitable,
`it is more effective to spend the resources to
`meet new test procedure requirements once,
`rather than to have multiple waves of new re-
`quirements for evaporative control over the next
`decade.
`Lower exhaust emission requirements,
`cold tests for CO, higher fuel economy, and the
`consumer’s demand for quality and reliability
`“among other demands all require the designer’s
`“full-attention. There is no shortage of chal-
`~:. Tanges for the automotive manufacturer
`in the
`1990%s.
`-
`
`
`
`The most effective techniques for the control of
`~ "axcess" evaporative emissions invalves reducing
`the. ‘quantity of vapors generated, as well as
`_
`; containing them.
`.
`Thermal’ management (less heating) of the fuel
`
`system is the most effective way to reduce vapor
`generation. Unfortunately,
`the fuel
`tank and
`- associated pipes compete for space and function
`with many other underbody components.
`The size
`of a fuel
`tank is a compromise of many factors,
`‘including ‘trunk space,
`room for suspension
`- components, and the consumer’s desire for vehi-
`“cle range {i.e.,
`tank capacity) between refuel-
`ings.
`Changes in any one component affect many
`others.
`Even minor changes can cause tooling
`and validation costs to multiply quickly.
`It is
`best
`to fix this once.
`The Real-Time test
`procedure provides a framework for achieving the
`> Maximum available reductions in excess evapora-
`tive emissions, while maintaining the manufac-
`turer’s ability to use unique,
`innovative and
`cost-effective product designs.
`‘Tn’ summary,
`the purpose and goals of the Real-
`Tima test procedure are driven by the following
`
`
`
`
`
`
`logics» Until
`
`the air quality goals of the
`70's are met, government will continue
`“to create regulations to force mobile
`source emissions to approach zero.
`Regulatory agencies may not have the
`best understanding of how to fix the
`problem, but they will try.
`
`the level of control de-
`“Given that
`* sired is highly stringent,
`it is in
`the best interest of the vehicle
`manufacturer to participate in the
`
`
`
`
`
`
`
`
`901119
`
`rule-making and help create comprehen-
`sive and realistic test procedures
`that provide the maximum reductions in
`real-world emissions.
`
`therefore is to create a test procedure
`The goal
`that achieves the maximum available reductions
`without unnecessarily limiting the use of inno-
`vative and cost-effective product designs.
`
`TEST METHODOLOGIES
`
`The Real-Time evaporative emissions test proce-
`dure integrates a series of vehicle precondi-
`tioning and measurement events with the current
`exhaust emission test.
`The purpose of the Real-
`Time evaporative test is to quantify the fuel
`vapor emissions released by the vehicle as it
`experiences a variety of conditions occurring
`during hot weather in an urban area,
`
`EVAPORATIVE EMISSIONS DEFINITIONS - There are
`four broad categories of evaporative emissions
`that occur from well-maintained (non-tampered)
`vehicies.
`These are:
`
`Running losses
`Hot Soak losses
`Diurnal
`losses and,
`Resting Losses
`
`"Running losses" are the fuel vapors that escape
`the system during operation of the vehicle,
`i.e., when the engine is running.
`These losses
`occur when the rate of fuel vapor formation
`exceeds the capacity of the vapor storage and
`purge systems.
`
`-"Hot soak" emissions are those fuel vapors that
`escape immediately after the vehicle has been
`operated, produced by residual pressure and heat
`input
`into the fuel system. Hot soak emissions
`occur within one hour or
`less from vehicle
`shutdown.
`
`losses" are the fuel vapors that escape
`"Diurnal
`from a parked vehicle as a result of the daily
`ambient
`temperature heating.
`This heating
`produces vaporization of the fuel
`liquid and
`expansion and expulsion of the vapor stored in
`the fuel
`tank.
`
`"Rasting losses" are the emissions that escape
`from the fuel system or the evaporative control
`system as a result of permeation through various
`non-metallic fuel system sources, or as a result
`of vapor migration allowed by other system
`design features such as open-bottom canisters.
`Resting losses can be affected by ambient
`tem-
`perature changes
`, but
`they are relatively
`constant,
`in comparison with other types of
`vapor loss.
`
`The Real-Time evaporative test procedure has
`three separate evaporative emission measure-
`ments. These are:
`
`

`

`901110
`
`5
`
`An example of the temperature and pressure
`measurements for a late model GM Passenger car
`is shown in Figure 4.
`
`Running Loss Measurement Techniques - Two meth-
`ods can be used to collect the running toss
`emissions,
`the enclosed dynamometer and the
`point source collectors.
`
`The enclosed dynamometer method involves sur-
`rounding the test cell with an air-tight enclo-
`sure, and measuring the increase in hydrocarbon
`(HC) concentration in the enclosure as the
`vehicle is driven.
`The basic measurement
`tech-
`niques were first outlined by EPA’s test engi-
`neering group over 15 years ago.
`[6]
`The calcu-
`lations to convert internal HC concentrations to
`grams are similar to the current Part 86 method-
`ology for evaporative emissions as measured in
`an enclosure.
`
`Figure 5.
`
`ENcLosSeD DyNaAMOMETER CELL
`
`Test Cau, frat Exouwazes
`
`—
`
`Figure 5 shows the general arrangement for the
`enclosed dynamometer methad.
`Intake air for the
`engine is supplied through a duct from outside
`the cell, and the vehicle’s exhaust
`is also
`routed to the outside.
`Care must be taken to
`insure that there are no leaks or mixing of the
`exhaust with the trapped air in the enclosure.
`Heat exchangers with active controllers are
`required to maintain the ambient
`temperature at
`the desired set-point.
`A, Jarge internal volume,
`on the order of 6000 ft’, can help to provide
`the internal
`thermal contro] necessary.
`The
`approximate internal dimensions of such an
`enclosed cell might be 18 feet wide by 34 feet
`* long by 10 feet high.
`
`A second method for measuring running losses is
`the point source collection method.
`The point
`source collection method involves placing vapor
`collectors at each of the vehicle system's
`designed vapor vents,and at any other suspected
`sources of measurable vapor losses during vehi-
`cle operation. Vapor collected by the equipment
`js transported to a critical
`flow venturi
`(CFV)
`or positive displacement pump {PDP) based dilu-
`tion and measurement system, as is currently
`ised to measure exhaust emissions.
`
`1}.
`
`2).
`
`3}.
`
`A direct measurement procedure for
`running Josses
`
`A hot soak measurement period, and
`
`An extended vehicle park episode to
`measure diurnal] and resting losses.
`
`THE RUNNING LGSS MEASUREMENT - Vehicle running
`loss measurements are made using a chassis
`dynamometer in a test cel] which is maintained
`at an ambient temperature of 96°F.
`
`Airflow in the test cell, and particularly under
`the vehicle in the area of the fuel
`tank,
`is
`managed in order to match tank temperature and
`headspace pressure measurements
`taken from
`similar vehicles that had been operated on a
`road.
`The road test vehicle is driven for three
`LA-4s at an ambient temperature in the 87 to
`100°F.
`range and corrected to an equivatent
`96°F.
`temperature profile.
`
`Road Temperature Measurements ~ As part of the
`Real-Time evaporative emission test procedure,
`the tank temperature profile is established by
`the vehicle manufacturer as part of the vehicle
`certification process. Representative combina-
`tions of powertrains and vehicle body styles are
`instrumented and tank temperatures measured
`during on-road operation. Additional details
`concerning road temperature measurements is
`contained in Appendix A.
`
`Figure 4.
`
`FUEL TEMPERATURE and PRESSURE MEASUREMENTS
`DURING ON-ROAD OPERATION
`LATE MODEL GH PASSENGER Can
`3 ba-47m WITH AIA CONDITIONING CW
`.
`TURE
`
`140,
`
`- b
`
`a =a
`
`e
`wy 229c
`
`C&R F AMBIERT TEMPERA nyuao
`
`2410
`
`w
`FA
`© s90
`
`20
`
`80
`
`o
`
`TANK PRESSURE ---->
`
`40
`
`20
`
`30
`
`4D
`
`50
`
`60
`
`3
`10 2
`<
`Su
`mnnm[2a]
`08
`
`70
`
`(mph)
`SPEED
`
`TIME IN MINUTES
`
`

`

`flow of ambient air
`The sample pump(s} draw a
`The inlet to the
`through the collector(s).
`collector(s) are placed in proximity to the
`vehicle's fuel
`system vapor vent and configured
`to capture any fuel vapor emissions without
`jmposing any significant flow or pressure im
`on the normal action of the vent.
`For vapor vents that terminate in a tube or hose
`
`Figure 6.
`
`DILUTION AIR
`
`DILUTION AIR
`
`
`arrangement for the
`Figure: 6“shows the general
`
`point source collection technique.
`The cotlec-
`tion system consists of a collector at each
`
`vehicle vapor vent,
`lengths of heated sample
`
`line (175-200°F.) connecting each collector to
`the inlet of a heated sample pump (40 scfh), and
`lengths of heated sample lines connecting the
`
`outlet of the sample pump(s)
`to the bulk stream
`of the CFV or PDP vapor measuring system.
`
`
`
`
`
`
`
`
`
`
`
`fn -
`HEATED
`Lo
`PUMP
`
`FROM OTHER
`VENTS CF ANY)
`
`“LEAKTIGHT7 \ FUEL CAP ENCLOSURE
`
`
`
`"SEAL Lf
`
`
`
`FUEL CAP
`
`FUEL TANK FILL PIPE
`
`
`
`Td
`AMBIENT BAG FILL
`
`HEATED LINE
`
`PUMP
`
`CANISTER
`VaPor
`COLLECTOR
`EVAPORATIVE
`CANISTER
`
`SILICONE
`BOOT
`
`AMBIENT AIR
`INLET
`.
`
`y,
`
`»
`&
`¥,
`
`
`
`CLAMP
`
`MIXING DEVICE
`
`TO FUEL TANK
`
`
`
`
`POINT SOURCE COLLECTION
`
` SCHEMATIC:
`
`

`

`901110
`
`such as the evaporative contro! canister,
`barb,
`a short length of tubing is used to extend the
`vent
`into the mouth of the collector (basically
`a "funnet"}.
`The dimensions of the collector
`are chosen to ensure that any vapors that might
`appear are directed into the collection system,
`but that no significant depression is imposed
`upon the vent.
`
`The fuel cap vapor vent collector consists of a
`boot that surrounds the cap and allows ambient
`air to be drawn through the boot volume.
`Fuel
`cap safety pressure relief is typically set at
`the 55 in. Ho (2.0 psi) level. Vehicle systems
`calibrated to comply with the new regulations
`should never vent
`the cap during these tests.
`Other vapor vents,
`if any, are fitted with
`similar and appropriate collectors.
`
`HOT SOAK - Within five minutes after the comple-
`tion of the running loss test (more than an hour
`of vehicle gperation),
`the vehicle is placed
`into a SHEDS
`for a one hour hot soak.
`To be
`consistent with the high temperature ozone
`focused conditions,
`the SHED ambient temperature
`is maintained between 90-100°F. after the vehi-
`cles is initially put
`in the SHED. This meas-
`urement is more stringent than the current FTP,
`insofar as it follows extended driving (22.5
`instead of 11.1 miles) and occurs at higher
`ambient temperatures (90-100°F),
`
`Figure 7 shows the test results from a vehicle
`placed in a conventional SHED for one hour
`following the running Toss test.
`The SHED,
`which is located in a laboratory with an ambient
`temperature of 76°F., was operated with the
`cooling system turned off. Putting the hot
`vehicle into the SHED rapidiy increases the
`enclosure temperature within five minutes,
`to
`levels that are appropriate (90 to 100°F.)} for
`the hot soak measurement.
`
`Figure 7 also indicates that a one hour period
`iS an appropriate measurement
`interval for a hot
`soak.
`The rapid increase in the SHED concentra-
`tion occurs within 20 minutes from the start of
`the test. After this initial
`increase,
`the SHED
`concentration continues to increase at a steady
`but much slower rate. This constant rate of
`evaporative emissions is characteristic of
`“resting losses", which will be discussed in
`more detail
`in the following section.
`
`MULTIPLE DAY PARK - The final step in the Real-
`Time procedure is a multiple day park.
`The
`vehicle is placed inte a Variable Temperature
`SHED {VT-SHED) for two days.
`The ambient
`tem-
`perature inside of the ¥T-SHED is cycled from a
`
`Figure 7,
`
`ONE HOUR HOT SOAK AFTER EXTENDED ORIVING
`22.5 MILES AT 95 ceo F
`
`te=r—~ SHEO TEMPERATURE
`
`aSHEDPPK o
`
`TEMPERATURE-F
`
`SHEG CONCENTRATION =-~->
`
`20
`
`19
`
`30
`20
`TIME IN MINUTES
`
`40
`
`50
`
`i)
`
`low of 72 °F. to a high of 96 °F, over a 9 hour
`period, and from 96 °F. back to 72 °F. over a 15
`hour period.
`The temperature profile is based
`on real hourly temperature data, as will be
`discussed in-a later section.
`
`temperature cycling inside of the
`The ambient
`VT-SHED results in large voiume changes that the
`conventional SHED is not capable of handling.
`A
`perfect gas law calculation shows that a 72 to
`96°F.
`temperature change results in a 4.5%
`volume change at constant pressure.
`For a
`normal 1600 cubic foot SHED that represents a
`change of 72 cubic feet. Daily changes in the
`barometric pressure would also compound the
`volume changes due to the temperature cycling.
`There are various solutions that can be imple-
`mented ta handle the volume changes,
`three of
`which will be briefly discussed.
`
`shown in Figure 8, has a
`One type of a VI-SHED,
`movable roof allowing for large changes in SHED
`volume. The VI-SHED roof is adjusted up or down
`such that a delta pressure of zero is maintained
`between the inside and outside of the SHED.
`The
`SHED is also well
`insulated to reduce heat
`~ Josses during high temperature operation.
`This
`type of facility was used in the work previously
`published by GM.
`[2]
`
`A second technique js a retrofit package that
`could be applied to existing SHEDs.
`The package
`
`2. The current Sealed Housing for Evaporative Determination (SHEQ) test
`methodology was derived from a SAE procedure, Jl/la, "Measurement of
`Fuel Evaporative Emissions from Gasoline Powered Cars and Light Trucks
`Using the Enclosure Technique", and replaced the previous carbon trap
`measurement method.
`
`

`

`
`
`
`
`
`Figure 8.
`
`GMR/PROVING GROUND VARIABLE TEMPERATURE SHED
`
`
`THEE
`:
`“a
`
`SOE110
`
`
`
`
`
`
`
`agcremeaeHe
`
`
`
` Jas!
`a
`28
`:
`=
`a
`
`
`The Variable Temperature SHED is unique in that it allows "real-time" measurements of
`evaporative emissions over a broad range of temperatures and time and is fully
`automated. Constructed under the guidance of C. Gus Mitsopoutos, Resident Manager and
`Norman W. Laursen, Staff Research Engineer.
`
`
`
`

`

`901110
`
`would include two large. low-permeation plastic
`bags mounted to the existing enclosure, one on
`the inside and the other on the outside.
`The
`inside bag would account for decreasing volumes
`and the outside bag for increasing volumes. At
`the start of the test the two bags would be
`positioned to a known reference volume.
`The two
`day test ends at the same ambient temperature at
`which it started, so that the only change in
`SHED volume would be due to barometric pressure.
`The ideal gas law could be used to calculate the
`velume change due to the change in the baromet-
`ric pressure.
`The retrofit package would also
`include a heating and cooling package plus
`insulation for the outside of the SHED walls.
`
`A third technique is described as a Hybrid SHED.
`The Hybrid SHED is a fixed volume enclosure that
`is maintained at a slightly negative pressure.
`
`The SHED is maintained at a negative pressure
`using a CVS pump that fills a sample bag.
`Fresh
`air enters the SHED through a controlled pres-
`sure regulator.
`The final mass
`includes the
`enclosure grams and also the amount removed into
`the sample bags.
`
`The. Two-Day Test - The vehicle is placed into
`the VT-SHED during the cooling portion of the
`ambient temperature cycle. This allows the fuel
`temperature to assume its natural relationship
`with the ambient temperature prior to the start
`of the heating cycle.
`An example of a two day
`park in a VT-SHED is shown in Figure 9.
`The
`ambient SHED temperature at
`ty corresponds to
`midnight on the hourly temperature profile.
`This allows approximately 5-6 hours of cooling
`prior to the start of the ambient
`temperature
`heating period.
`
`The emission measurement period is the second 24
`hours of the test (tag - to4)-
`Included is this
`measurement are 24 hours et resting losses and
`the 2nd days diurnal.
`An indication of the
`magnitude of resting losses is the increase in
`the SHED concentration during the portion of the
`test when the fuel
`temperature is decreasing.
`During this portion of the test there shoutd be
`no out-gassing of vapor from the fuel
`tank.
`Losses measured during this period should con-
`sist only of permeation losses and vapor migra-
`tion, plus non-fuel background emissions.
`
`Over long periods of time, such as 48 hours,
`SHED retention is a concern.
`Loss of gases from
`the VT-SHED must be minimized, and if necessary,
`accounted for in the emission measurement calcu-
`‘tation.
`The previous figure’ (Figure 9),
`showed
`‘the SHED grams in terms of "corrected" SHED
`grams.
`The correction factor applied to the
`data was 0.4%/hour, based on actual
`retention
`tests.
`An appropriate correction factor for
`individual VI-SHEDs can be established using
`propane injection techniques similar to the
`current SHED retention test).
`
`Figure 9.
`A THO OAY PARKED CAH EXPERIENCE
`
`i100
`
`oo
`
`TEMPERATURE-F oa5
`
`40
`
`f2
`TIME IN HOURS
`
`SHEDGRAMS 2a
`
`=4a2uwo
`
`hy
`
`36
`
`4B
`
`COARECTED
`
`TEST PARAMETERS
`
`The Real-Time evaporative emission test proce-
`dure outlined above requires direct measurement
`of running losses, hot soaks and extended park
`episodes.
`In addition to requiring new test
`methodologies,
`the Real-Time test procedure
`involves new test parameters.
`Six parameters
`-are discussed in this part of the paper:
`
`1. The length of the extended park episode
`2. Daily high temperature
`3. Daily temperature profile
`4. Representative tank fill level
`5.
`Fuel RV¥P, and
`6. Trip length
`
`- The
`The 24-Hour Extended Park Measurement
`intent of the Real-Time test is to simulate the
`‘in-use conditions of public concern and thereby
`require control of the emissions that contribute
`to the real-world inventory.
`The Real-Time
`measurement procedure accordingly requires
`measurement of emissions during the last 24
`hours of a 48 hour park. During both time peri-
`ods,
`the ambient
`temperature is cycled over the
`high-temperature profile described in later
`sections. First, however it is important
`to
`discuss the need for a 24 hour measurement
`interval. Shorter compressed-time measurement
`periods have a number of shortcomings compared
`to a Real-Time test.
`A few of those shortcom-
`ings are discussed below.
`
`

`

`
`
`in-
`~Emisstons that are measured during a test
`tended to simulate the multiple day park must
`include both diurnal emissions -
`those that
`
`occur during the portion of the day when the
`.
`
`fuel temperature is increasing, and resting
`
`losses (which occur 24 hours a day).
`The magni-
`
`tude of the resting losses can be quantified
`
`during the fourteen hour fuel cooling periods of
`
`“the Real-Time test.
`
`Resting losses can be a layge contributor to
`
`inventory, but are difficult to detect
`in short
`increments of time.
`The point is illustrated in
`“Figure 10.
`Two vehicles, one with an open
`
`bottom canister,
`the other with a closed bottom
`canister, were subjected to a traditional one
`
`hour: forced blanket diurnal
`fuel heating from 70
`
`to 91 °F and then sat in the SHED for an addi-
`
`tional 9 hours. These tests were done in a
`
`conventional SHED at a constant ambient
`tempera-
`
`ture of 76 °F.
`The two vehicles had similar
`
`
`“‘a@missions at the end of 1] hour, as shown in
`Figure 10. As Figure 10 also shows, however, at
`the end of the 10 hour period,
`the vehicle with
`the open bottom canister had 4.6 times higher
`
`emissions than the vehicle with a closed bottom
`’ canister.
`
`
`
`Figure 10.
`
`TEMPERATURE-F
`
`6
`
`Keme-
`
`Figure 11.
`A THREE DAY PARKED CAR EXPERIENCE
`OPEN BOTTOM CANESTER - 9 AYP FUEL
`4988 OLOSHOAILE AEGENCY 9B
`
`AHBIENT TERP
`
`CAWISTER RETBHT INCREASE
`
`901110
`
`
`
`CORRECTEDSHEOGAAMS
`
`4B
`
`fo
`
`72
`
`DAY Do <ono?
`
`q-—-+=
`
`fa
`
`2a
`
`36
`TIME IN HOURS
`DAY Lo weePeemee
`DAY 2 oseeeberee
`
`“DIURNAL AND RESTING LOSS MEASUREMENTS
`CONVENTIONAL SHED AT CONSTANT’ AHSTENT TEMPERATURE
`4 HOUR DIURNAL 70-94 F - FOLLOWEO BY A S HOVA SOAK
`
`'SHEDGRAMS
`
`.3B g/hr seee>
`
`OPEN BOTTOM CANISTER
`
`106 gfhr ---=>
`
`CLOSED BOTTOM CANISTER
`
`“>
`.
`
`$¢
`
`2
`
`8
`
`8&8
`5
`4
`TIME IN HOURS
`
`7
`
`8B
`
`48
`
`40
`
`The same open bottom canister vehicle was sub-
`jected to a real-time three day test and the
`results are shown in Figure 11.
`(In this par-
`ticular test the VT-SHED ambient temperature was
`
`cycled from 66 to 95 °F.)
`The open-bottom
`. Canister’s inherent ability to purge itself is
`
`.: evidenced. by the traces on Figure 11
`showing
`canister weight and SHED grams.
`The SHED grams
`
`
`
`continue to increase at a substantial rate even
`during the non-fuel heating periods.
`Figure 11
`indicates that
`the canister begins losing
`weight, clearly shown at
`the 36th hour of the
`test, when it is not only failing to contain the
`incoming vapors, but
`is also loosing some of the
`vapors previously adsorbed.
`
`Open bottom canisters are not the only source of
`resting losses. Rubber hoses, non-metallic fuel
`lines, and non-metallic fuel
`tanks, all have the
`potential of releasing significant resting
`losses.
`The real-time diurnal would identify
`these sources and prevent
`them from being a
`major source to inventory.
`In addition, materi-
`als or components whose resting losses are a
`function of temperature would also be accounted
`for by using a real-time temperature cycle.
`
`Another very important phenomenon that occurs
`naturally in real-time in the real world is
`backpurge.
`Backpurge,
`its occurrence and the
`effect on canister capacity, has been discussed
`jn detail.
`[2]
`The occurrence of backpurge and
`its effect on canister capacity are clearly
`demonstrated in Figure 12,
`A vehicle was parked
`in a VT SHED for eight days with the ambient
`temperature cycled from 72 to 96 °F.
`As
`the
`canister became more loaded,
`the hackpurge
`effect was much more pronounced.
`The backpurge
`effort was evidenced by the decreasing canister
`weight and only a slight increase in the SHED
`grams, during the period of time when the fuel
`is cooling.
`
`The eight day park data is also summarized in
`
`

`

`901110
`
`Figure 12.
`
`
`
`
`
`
`
`CORRECTEDSHEDendDELTACANISTEAWEIGHT-grams
`
`AN EIGHT OAV PARKED CAR EXPERIENCE
`9 RVP FUEL — 40% TANK FILL
`49GB OLOSHOBILE REGENCY 95
`
`CANTSTER SEISHT INCREASE
`
`Qo
`
`ao
`
`iho
`
`nN°
`
`TEMPERATURE~F 5
`
`24
`
`4a
`
`72
`
`420
`36
`TIME IN HOURS
`
`444
`
`468
`
`i192
`
`The lower
`the bar chart shown in Figure 13.
`portion of Figure 13 shows the resting-loss data
`for the vehicle, which ranged from .06 to .09
`grams per hour.
`The upper portion of the graph
`shows incrementat SHED grams measured during the
`fuel heating periods -- the diurnal emissions --
`for each of the eight days’. After approximate-
`ly six days,
`an equilibrium condition between
`the diurnal
`loading and the backpurge is
`reached.
`
`Because of backpurge, extended multiple day
`parks do not result in uncontrolled diurnals, as
`is currently calculated in EPA’s MOBILE4 emis-
`sions inventory model.
`[7]
`The incremental
`increase in the VT-SHED grams during the ten
`hour diurnal heating periods (days 6-8} is
`roughly 11.5 grams,
`less than half of the calcu-
`Jated uncontrolled diurnal for the vehicle who’s
`results are shown in Figures 12 and 13. MOBILE4
`sheuld be updated to add this increased under-
`standing to the inventory model.
`
`eee ee ee eee eee eee
`
`1i
`
`A final point to be made favoring real-time
`measurements concerns the time-rate of hydrocar-
`bons delivered to the canister during diurnal
`heating.
`If dupticating the quantity of vapor
`generated from a real-time diurnal was the only
`concern,
`then a time compressed diurnal could
`accomplish that. However,
`the real-world mecha-
`nisms are much more complex.
`
`Compressed diurnal heating periods increase the
`amounts of vapors per unit time that the canis-
`ter must contain. Directionally,
`this leads to
`reduced canister capacities due to the increased
`canister bed temp