Api publ 4650 1997 scan (american petroleum institute)

A comparison of emissions results from vehicles tested in this program to predictions from EPA’s TECH5 model which was used to generate base emission rate equations for MOBILESa indicate

Trang 1

S T D A P I / P E T R O P U B L 4 b 5 0 - E N G L 1 9 9 7 0 7 3 2 2 9 0 0 5 b 4 2 0 4 il39 I

American Petroleum Institute

Copyright American Petroleum Institute

Provided by IHS under license with API

Not for Resale

No reproduction or networking permitted without license from IHS

Trang 2

`,,-`-`,,`,,`,`,,` -S T D A P I / P E T R O PUBL 4b50-ENGL 1 9 9 7 073229U 0 5 b 4 2 0 5 T 7 5

A na lys i s of H i g h = M i I eag e-Ve h i c I e Emissions Data from Late-Model, Fuel-Injected Vehicles

Health and Environmental Sciences Department

FEBRUARY 1997

American Petroleum Institute

Trang 3

`,,-`-`,,`,,`,`,,` -S T D - A P I / P E T R O PUBL 4b50-ENGL 1777 0 7 3 2 2 9 0 05b1i20b 901

FOREWORD

API PUBLICATIONS NECESSARILY ADDRESS PROBLEMS OF A GENERAL NATURE WITH RESPECT TO PARTICULAR CIRCUMSTANCES, LOCAL, STATE,

AND FEDERAL LAWS AND REGULATIONS SHOULD BE REVIEWED

API IS NOT UNDERTAKING TO MEET THE DUTIES OF EMPLOYERS, MANUFAC- TURERS, OR SUPPLIERS TO WARN AND PROPERLY TRAIN AND EQUIP THEIR EMPLOYEES, AND OTHERS EXPOSED, CONCERNING HEALTH AND SAFETY

RISKS AND PRECAUTIONS, NOR UNDERTAKING THEIR OBLIGATIONS UNDER LOCAL, STATE, OR FEDERAL LAWS

NOTHING CONTAINED IN ANY API PUBLICATION IS TO BE CONSTRUED AS

GRANTING ANY RIGHT, BY IMPLICATION OR OTHERWISE, FOR THE MANU- FACTURE, SALE, OR USE OF ANY METHOD, APPARATUS, OR PRODUCT COV- ERED BY LETTERS PATENT NEITHER SHOULD ANYTHING CONTAINED IN

THE PUBLICATION BE CONSTRUED AS INSURING ANYONE AGAINST LIABIL-

ITY FOR INFRINGEMENT OF LE'ITERS PATENT

All rights reserved No part of rhis work may be reproduced, stored in a retrieval system or transmitted by any means, electronic, mechanical, photocopying, recording, or otherwise, withour prior written permission from the publishel: Contact the publishe6 API Publishing Services, 1220 L Street N W , Washington, D.C 20005

Copyright O 1997 American Petroleum Institute

111

Not for Resale

Trang 4

David Lax, Health and Environmental Sciences Department

MEMBERS OF THE THE VEHICLE EMISSIONS TASK FORCE

J Steve Welstand, Chairperson, Chevron Research and Technology Co

Dennis Feist, Shell Development Co

Mani Natarajan, Marathon Oil Co

Rick Riley, Phillips Petroleum

King Eng, Texaco, Inc

George S Musser, Exxon Research & Engineering Co

Charles Schleyer, Mobil Research & Development

iv

Trang 5

`,,-`-`,,`,,`,`,,` -ABSTRACT

A recent analysis of the Environmental Protection Agency’s (EPA) on-road vehicle emission factors model, MOBILESa, suggested that exhaust emissions of newer model, fuel-injected

vehicles were over-predicted at high mileage That over-prediction, which results in over-

predictions of fleet-average emission rates for future calendar years, was related to the paucity of

data from modem technology, high-mileage vehicles available for the development of

MOBILESa To bolster the database available for the next version of the MOBILE model, the American Petroleum Institute (API) sponsored a test program to investigate the exhaust emission control system deterioration characteristics of late-model, fuel-injected vehicles Seventy-five light-duty vehicles were procured and tested over the Federal Test Procedure in this program Vehicles included in the program were from the 1985 to 1992 model years, had accumulated at least 100,000 miles, and had never been subject to an inspection and maintenance (UM) program Vehicles were recruited at a site in Chicago Heights (CH), Illinois, and two separate sites in

Phoenix, Arizona The test program was conducted from April 1995 to August 1996

Analysis of the data collected in this High-Mileage Vehicle (HMV) project revealed that model year was a more important determinant of emissions than fuel-injection technology (i.e.’ port fuel injection versus throttle body injection), with vehicles in the 1988 and later model year group demonstrating substantially lower emissions than vehicles in the 1985 to 1987 model year group

A comparison of emissions results from vehicles tested in this program to predictions from EPA’s TECH5 model (which was used to generate base emission rate equations for MOBILESa)

indicated that TECH5 may be overpredicting emissions of late-model vehicles at high mileage, which has a significant impact on fleet-average emission rates calculated by MOBILESa for future analysis years Finally, a comparison of high-mileage vehicle emissions data from this project to data collected by EPA in Hammond, Indiana (which were collected approximately four to five years prior to the HMV project and were the basis of MOBILESa) indicated that newer model year vehicles are more durable (from an emissions perspective) at equivalent age and mileage

Not for Resale

Trang 6

`,,-`-`,,`,,`,`,,` -S T D A P I / P E T R O P U B L 4b50-ENGL 1997 0 7 3 2 2 9 0 0 5 b 4 2 0 9 bLO

TABLE OF CONTENTS (continued)

4 IMPACT OF NEW DATA ON MOBILESa BASE EMISSION RATES (continued)

COMPARISON TO EMFAC7F EMISSION FACTORS 4-9

5 INFLUENCE OF MILEAGE A C C m A T I O N RATES AND VEHICLE AGE 5-1

REFERENCES R-1

Appendix A

VEHICLE-SPECIFIC TEST RESULTS A-1

Trang 7

`,,-`-`,,`,,`,`,,` -S T D A P I / P E T R O PUBL q b S O - E N G L 1997 H 0 7 3 2 2 9 0 O 5 b q 2 1 0 3 3 2 =

TABLE OF CONTENTS

EXECUTIVESUMMARY e5-1

STATISTICAL ANALYSIS OF THE DATABASE ES- 1

DEVELOPMENT OF BASE EMISSION RATE EQUATIONS e5-2

IMPACT OF HMV EMISSION FACTORS ON MOBILE5a PREDICTIONS e5-3

INFLUENCE OF MLEAGE ACCUMULATION RATES AND VEHICLE AGE ON EMISSIONS e5-5

1 INTRODUCTION 1-1

BACKGROUND 1-1

ORGANIZATION OF THE REPORT 1-2

2 PROGRAM DESCRIPTION AND SUMMARY STATISTICS 2-1

PROGRAM DESCRIPTION 2-1

SUMMARY STATISTICS 2-2 COMPARISON TO HAMMOND DATA 2-8

TOTECH5 3-1

BACKGROUND 3-1

HC AND CO EMISSIONS 3-1

NOxEMISSIONS 3-3

4 IMPACT OF NEW DATA ON MOBILESa BASE EMISSION RATES 4-1

DISTRIBUTION OF EMITTER CATEGORIES BY MILEAGE FOR HIGH-MILEAGE VEHICLES 4-1

EMISSION RATES BY EMITTER CATEGORY 4-2

COMPOSITE EMISSION RATES FOR VEHICLES

WITH MORE THAN 100 O00 MILES 4-3

COMPOSITE EMISSION RATES FOR VEHICLES

WITH LESS THAN 100 O00 MILES 4-3

FLEET-AVERAGE EMISSION RATES 4-6

Not for Resale

Trang 8

`,,-`-`,,`,,`,`,,` -S T D A P I / P E T R O P U B L 4b50-ENGL 1 9 9 7 W 0 7 3 2 2 7 0 05b1i2LL 2 7 9 M

LIST OF FIGURES

ES-1 Comparison of HC Emission Factors from TECH5 and the

High-Mileage-Vehicle Study ES-3 ES-2 Comparison of Fleet-Average HC Emission Factors - MOBILE5a

Versus the High-Mileage-Vehicle Study ES-4 ES-3 Comparison of Age-Based HC Emission Rates for Fuel-Injected Vehicles with

More Than 100,000 Miles - Hammond Versus the High-Mileage-Vehicle Database ES-6 ES-4 Comparison of Age-Based CO Emission Rates for Fuel-Injected Vehicles with

More Than 100,000 Miles - Hammond Versus the High-Mileage-Vehicle Database ES-7

ES-5 Comparison of Age-Based NOx Emission Rates for Fuel-Injected Vehicles with

More Than 100,000 Miles - Hammond Versus the High-Mileage-Vehicle Database ES-7 2-1

2-2

2-3

4- 1

Mean HC Emissions by Model Year for the High-Mileage-Vehicle Database

Mean CO Emissions by Model Year for the High-Mileage-Vehicle Database

Mean NOx Emissions by Model Year for the High-Mileage-Vehicle Database

Comparison of Fuel-Injected-Vehicle HC Emission Rates -

TECH5 Versus the High-Mileage-Vehicle Study 4-5 Comparison of Fuel-Injected-Vehicle CO Emission Rates -

TECH5 Versus the High-Mileage-Vehicle Study

Comparison of Fuel-Injected-Vehicle NOx Emission Rates -

TECH5 Versus the High-Mileage-Vehicle Study

Comparison of Fleet-Average HC Emission Rates -

Comparison of Fleet-Average CO Emission Rates -

Comparison of Fleet-Average NOx Emission Rates -

Comparison of TECH5, High-Mileage-Vehicle Study, and EMFAC7F HC Emission Factors 4- 1 O

2-6 2-6 2-7

MOBILE5a Versus the High-Mileage-Vehicle Study

MOBILE5 a Versus the High-Mileage-Vehicle Study

Trang 9

`,,-`-`,,`,,`,`,,` -S T D A P I / P E T R O PUBL LibSO-ENGL 1797 = 0 7 3 2 2 7 0 05b4212 L O 5

4-8 Comparison of TECH5, High-Mileage-Vehicle Study, and

EMFAC7F CO Emission Factors 4- 1 O

4-9 Comparison of TECH5, High-Mileage-Vehicle Study, and EMFACT

NOx Emission Factors

Comparison of Mean HC as a Function of Age for the Hammond

Comparison of Mean CO as a Function of Age for the Hammond and High-Mileage-Vehicle Databases 5 -4

Comparison of Mean NOx as a Function of Age for the Hammond and High-Mileage-Vehicle Databases 5-5

Not for Resale

Trang 10

Distribution of Vehicles Included in the High-Mileage-Vehicle Database

by Model Year, Fuel-Delivery Technology, and Site 2-3

Mean FTP Emissions fiom the High-Mileage-Vehicle Database as a Function

of Fuel-Delivery Technology Compared to Predictions fiom TECH5 2-3

Mean FTP Emissions fiom the High-Mileage-Vehicle Database by Site 2-5

Mean FTP Emissions fiom the High-Mileage-Vehicle Database by Fuel Delivery Technology and Procurement Site 2-5

Mean FTP Emissions fiom the High-Mileage-Vehicle Database by Model-YearGroup 2-7

Mean FTP Emissions fiom the High-Mileage-Vehicle Database by Fuel Delivery Technology and Model-Year Group 2-8

Comparison of Mean Emission Rates, Mileages, Ages, and Sample Sizes for Fuel-Injected Vehicles in the High-Mileage-Vehicle Database and the Hammond, Indiana, Database by Mileage Increment 2-9

Distribution of Normal, High, Very High, and Super Emitters in the High-Mileage-Vehicle Database Versus TECH5 Predictions 3-2

Distribution of Emitter Categories by Model-Year Group in the High-Mileage-Vehicle Database 3-3

Mean Emission Rates of HCKO Emitter Categories in the High-Mileage-Vehicle Database Versus TECH5 Estimates 3-4

Distribution of NOx Normal and High Emitters in the High-Mileage-Vehicle

Database Versus TECH5 3-4

Mean Emission Rates of NOx Emitter Categories in the High-Mileage-Vehicle Database Versus TECH5 Estimates 3-5

Distribution of Emitter Categories by Vehicle Mileage Used in the Modeling

Mean Emission Level by Emitter Category Used in the Modeling Runs

Modeled High-Mileage-Vehicle Emission Rates Based on the

Runs for This Analysis (Based on the High-Mileage-Vehicle Database)

for This Analysis (Based on the High-Mileage-Vehicle Database) 4-3

High-Mileage-Vehicle Database 4-3

4-2

Trang 11

`,,-`-`,,`,,`,`,,` -4-4 Fleet-Average Exhaust Emission Rates Calculated with MOBILE5a with

Current and High-Mileage-Vehicle Study Emission Factors 4-7

Mean Age and Mileage of Vehicles Included in the High-Mileage-Vehicle Database 5-2 Comparison of Fuel-Injected-Vehicle Emissions as a Function of Vehicle

Age - Hammond Versus High-Mileage-Vehicle Database 5-3

Comparison of PFI Vehicle Emissions as a Function of Vehicle Age -

Hammond Versus High-Mileage-Vehicle Database 5-6

Not for Resale

Trang 12

`,,-`-`,,`,,`,`,,` -S T D - A P I I P E T R O P U B L 4 b 5 0 - E N G L 1777 0 7 3 2 2 9 0 O 5 b 4 2 L 5 711i D

EXECUTIVE SUMMARY

To investigate the emission control system deterioration characteristics of late-model, fuel-

injected vehicles, the American Petroleum Institute (API) sponsored a test program in which 75

high-mileage light-duty vehicles were procured and tested over the Federal Test Procedure (FTP)

Vehicles included in the test program had accumulated at least 100,000 miles and had never been

subject to an inspection and maintenance (UM) program Vehicles were recruited at a site in

Chicago Heights (CH), Illinois, and two separate sites in Phoenix, Arizona The test program

began in April 1995 and continued through August 1996

Based on a previous analysis of the U.S Environmental Protection Agency's (EPA's) on-road

motor vehicle emission factors model, MOBILESa, it appeared that the assumptions built into the

development of the non-UM base emission rate equations were resulting in an over-prediction of

emissions at high mileage for late-model vehicles (Sierra Research, 1994) EPA's rationale for

using some of those assumptions was related to the sparsity of data for vehicles at high mileage

(i.e., over 100,000 miles) Therefore, API sponsored this test program with the intent of

bolstering the database used by EPA to develop base emission rate equations for the next version

of the MOBILE model

When data collection in the High-Mileage-Vehicle 0 project was completed, API retained

Sierra Research, Inc., to perform an analysis of those data, which is the subject of this report The

results of that evaluation are summarized below

STATISTICAL ANALYSIS OF THE DATABASE

A general statistical analysis of the data revealed the following:

A total of 75 light-duty vehicles were included in the sample: 50 of those were equipped with port fuel-injection (PFI) systems, 25 were equipped with throttle-body injection (TEH) systems

ES- 1

Trang 13

`,,-`-`,,`,,`,`,,` -S T D * A P I / P E T R O P U B L 4b50-ENGL 1797 0 7 3 2 2 7 0 0 5 b 4 2 L b 850

Model years included in the sample ranged from 1985 to 1992 Thirty-seven vehicles

were in the 1985 to 1987 model year group; 38 were in the 1988 and later model year

group

Model year appeared to be a more important determinant of emissions than fuel- delivery technology, with the 1988 and later model year group demonstrating substantially lower emissions than the 1985 to 1987 group

A single “super” emitter was tested in the HMV project This vehicle (1 985 model year, TBI fuel-delivery system) had hydrocarbon (HC) emissions of 10.32 g/mi and carbon monoxide (CO) emissions of 149.45 gimi Because it had a significant impact on the mean emission rate calculations presented in the report, most statistics are presented with and without this vehicle included

The site at which the vehicles were recruited appeared to have little impact on the mean emission levels of the sample

The data collected in the HMV project were used to develop alternative emission factors for use

in MOBILE5a The modeling methodology used for this analysis relied on the emitter category definitions developed by EPA for MOBILESa Figure ES- 1 compares the HMV study HC

emission rates* to those predicted by TECH5 (a pre-processor model to MOBILESa that

generates base emission rate equations and I/M credits for the model) That figure, which also shows the mean emission rates of the HMV database by model-year group and mileage increment (i.e., 100,000 to 125,000; 125,000 to 150,000; and over 150,000), indicates that MOBILE5a (ie., TECHS) is significantly over-predicting HC emissions of late-model vehicles at high mileage Similar trends were also observed for CO and oxides of nitrogen @Ox) emissions

* Note that the “HMV Study” line in Figure ES-1 reflects modeled predictions described later in the report Because only vehicles with more than 100,000 miles were tested in this study, the modeled emission rates fiom O to 50,000 miles were based on TECH5, while emission rates fiom 50,000 to the frst mileage increment of the HMV data

(approximately 1 10,000 miles) were based on simple interpolation

ES-2

Not for Resale

Trang 14

Figure ES-1 Comparison of HC Emission Factors from TECH5 and the High-Mileage-Vehicle

Study (Data points reflect mean emission rates calculated from the HMV data; the “HMV Study” line represents modeled predictions based on this work.)

It is interesting to note that the mean emission rates by emitter category (i.e., normal, high, very high, and super emitters) are not dramatically different when the HMV results are compared to TECHS Rather, the difference in the distribution of vehicles among emitter categories is

responsible for the large difference in predicted emission rates between TECH5 and the HMV

study, i.e., more hgh-emitting vehicles are being predicted by TECH5 than are observed in the

HMV database (For a discussion of the basis for the TECH5 emitter distributions, please refer to

Sierra Research, 1994.)

IMPACT OF HMV EMISSION FACTORS ON MOBILE5a PREDICTIONS

The emission factors illustrated in Figure ES-1 were used in conjunction with the MOBILE5a model to generate light-duty-vehicle, fleet-average HC emission rates for calendar years 1995 to

ES-3

Trang 15

`,,-`-`,,`,,`,`,,` -2010 (Note that the HMV study emission factors were calculated only for 1985 and later model year vehicles.) The results of the MOBILE5a runs are illustrated in Figure ES-2, which shows a significant decrease in fleet-average emissions when the emission factors are based on the HMV

study Also of note in Figure ES-2 is that the impact of the revised emission factors is most pronounced on the non-I/M estimates, and the difference between the MOBILE5a rates and the

HMV study rates increases in future years as the fleet turns over

Not for Resale

Trang 16

`,,-`-`,,`,,`,`,,` -INFLUENCE OF MILEAGE ACCUhKJLATION RATES AND VEHICLE AGE ON

EMISSIONS

A final evaluation performed in this study was a comparison of FTP-based emission rates from the

EPA's Hammond, Indiana, Mî40 and emission factors test program and the emission rates from

the HMV study.' One complicating factor related to this comparison is that the HMV study was conducted about 4.5 years &er the Hammond data were collected; thus, it represents a newer fleet of vehicles (on a model-year basis) relative to the Hammond project In addition, although vehicles in the HMV project were limited to a maximum mileage accumulation rate of 33,000 miles per year, some of those vehicles accrued mileage at a much faster rate than modeled by MOBILESa In particular, vehicles in the 3- to 6-year age range had accumulated 40,000 to

To determine if emission control system in-use performance has improved with more recent model years, as the data from the HMV project suggest, emission rates from vehicles in EPA's Hammond, Indiana, emission factors program were compared to those from the HMV project at equivalent ages and mileages @e., only vehicles with more than 100,000 miles fiom the

Hammond database were included in this comparison) The results of this comparison for HCs

are illustrated in Figure ES-3, which shows that vehicles in the HMV database have lower HC emissions, on average, than vehicles in the Hammond database, at least up through about 8.5 years of age This age corresponds to a mean model year of 1982.7 in the Hammond database

and 1987.2 in the HMV database

' The Hammond database used for this comparison was that used by EPA to develop base emission rate equations for

MOBILESa and reflects the in-use fleet in the 1990 to 1992 timeframe Not all vehicles in that database received FTP

tests; rather, correlation equations were used to predict FTP values ftom IM240 results for vehicles that were not FTP tested This approach was taken because vehicles selected for FTP testing were not chosen at random (i.e., the FTP

sample was skewed with high-emitting vehicles), and this methodology ensured that an unbiased sample was obtained for subsequent analysis A detailed description of this approach can be found in Sierra Research, 1994

ES-5

Trang 17

`,,-`-`,,`,,`,`,,` -2

1.8 -

l c t -

0.8 0.6

Figure ES-3 Comparison of Age-Based HC Emission Rates for Fuel-Injected Vehicles with

More Than 100,000 Miles - Hammond Versus the High-Mileage-Vehicle Database (Note that the H M V database reflects vehicles that are 4.5 to 5 years

newer on a mgdel-vez basis.)

Results for CO and NOx are shown in Figures E S 4 and ES-5, respectively Note that the

CO results are somewhat erratic in that the HMV database has much higher CO emissions than the Hammond database for vehicles that are 9.5 years old; however, the sample size for this group

is very s d@e., 8 vehicles in the HMV database and 21 in the Hammond database) On the other hand, NOx emissions fkom vehicles in the Hammond database are much higher for a l l age groups

ES-6

Not for Resale

Trang 18

Figure ES-4 Comparison of Age-Based CO Emission Rates for Fuel-Injected Vehicles with

More Than 100,000 Miles - Hammond Versus the High-Mileage-Vehicle Database (Note that the HMV database reflects vehicles that are 4.5 to 5 years newer on a mdel-veiu: basis.)

Figure ES-5 Comparison of Age-Based NOx Emission Rates for Fuel-Injected Vehicles with

More Than 100,000 Miles - Hammond Versus the High-Mileage-Vehicle Database (Note that the HMV database reflects vehicles that are 4.5 to 5 years

ES-7

Trang 19

`,,-`-`,,`,,`,`,,` -S T D m A P I I P E T R O P U B L LibSO-ENGL 1997 .I0 7 3 2 2 9 0 0 5 b 4 2 2 2 0 5 4 W

Section 1 INTRODUCTION

BACKGROUND

To investigate the emission control system deterioration characteristics of late-model, fuel-

injected vehicles, the American Petroleum Institute sponsored a test program in which 75 high- mileage light-duty vehicles were procured and tested over the Federal Test Procedure Vehicles that were included in the test program met the following criteria:

1985 or newer model year;

passenger car;

fuel-injected (25 throttle-body and 50 port fuel-injected);

over 100,000 miles;

less than 33,000 miles per year mileage accumulation; and

never subject to an I/M program

The vehicles tested in this program were procured at three locations - an I/M lane in

Chicago Heights, Illinois, and two different IA4 lanes in Phoenix, Arizona Once procured, the vehicles were transported to Automotive Testing Laboratory’s (ATL’s) facilities in South Bend, Indiana (for the Chicago Heights vehicles) or Mesa, Arizona (for the Phoenix vehicles) for FTP

testing The test program began in April 1995 and continued through August 1996

Based on a previous analysis of the U.S Environmental Protection Agency’s on-road motor vehicle emission factors model, MOBILESa, it appeared that the assumptions built into the development of the base emission rate equations were resulting in an over-prediction of emissions

at high mileage for iate-model vehicles (Sierra Research, 1994) EPAs rationale for using some

of those assumptions was related to the sparsity of data for vehicles at high-mileage (i.e., over

1-1

Not for Resale

Trang 20

emissions of the sample by model-year group, fuel-injection technology, and site;

a comparison of the distribution of emitter categories and emission levels from the

HMV database to estimates obtained with EPA's TECH5 model;

an assessment of the impact the HMV data would have on emission rates generated with MOBILESa; and

a comparison of mileage accumulation rates between the HMV database and MOBILESa

This report documents the analyses described above

ORGANIZATION OF THE REPORT

Following this introduction, Section 2 contains a general statistical summary of the data collected

in the HMV test program Section 3 presents a comparison of the HMV database and estimates

of emitter category distributions and emissions from the TECH5 model An assessment of the impact that the new data would have on MOBILE5a emission rates is presented in Section 4, and

Section 5 reviews the mileage accumulation rates observed in the HMV sample and MOBILE5a

1-2

Trang 21

Indiana, while vehicles recruited from the Phoenix lanes were tested at ATL’s facilities in Mesa,

Arizona Testing at ATL consisted of an IM240 and a three-bag Federal Test Procedure

Vehicles recruited for this program had accumulated at least 100,000 miles, were new to the area (i.e., they were receiving their first inspection prior to registration), and had never been subject to

an VM program in the past Thus, the HMV database reflects a non-VM set of high-mileage vehicles, which is important from the perspective that EPA uses non-I/M emissions data to

generate base emission rate equations for its MOBLLE5a emission factor model

There were several other criteria placed on the vehicles before they were accepted into the

program First, they had to be fuel-injected Overall, 75 vehicles were tested in the program: 50

of those were equipped with port fuel-injection systems, and 25 had throttle-body fuel-injection systems The vehicles also had to be from the 1985 and later model years Although each vehicle had accumulated more than 100,000 miles, it could not have accumulated more than 33,000 miles per year on average (This constraint eliminated vehicles with extremely high mileage

accumulation rates.)

An issue that surfaced during the HMY program was whether the vehicles selected for testing were representative of the in-use fleet One of the concerns expressed by EPA staff is that

demographics could impact how well-maintained a vehicle is, and therefore its emission rate

In an attempt to address this issue, vehicles were selected from two different sites in the Arizona

2- 1

Not for Resale

Trang 22

`,,-`-`,,`,,`,`,,` -portion of the program One of those sites (designated AZ-6 in Table 2.1) is in a lower-income,

commerciallindustriai zone The other Arizona site (designated AZ-7) is located in a mixed

residentidcommercialíindustrial area in the Phoenix suburb of Mesa Presumably, the AZ-6 site

would provide a greater number of poorly maintained, older vehicles than the AZ-7 site

The distribution of vehicles by model year and fuel delivery technology is summarized in Table 2-

1 As described above, only vehicles equipped with PFI or TBI were tested in this program -

carbureted vehicles were specifically excluded as that fuel delivery technology is no longer used

on light-duty vehicles As observed in Table 2- 1, roughly half of the vehicles in the HMV

program were from the 1985 to 1987 model years, while the other half were from the 1988 to

1992 model years Two-thirds (i.e., 50 of 75) of the vehicles were equipped with PFI systems,

with the remainder being equipped with TE31 systems Finally, the distribution of vehicles among

sites is fairly even, with 26 vehicles tested at the AZ-6 site, 27 vehicles tested at the AZ-7 site,

and 22 vehicles tested at the Chicago Heights site

A summary of mean hydrocarbon, carbon monoxide, and oxides of nitrogen FTP emission rates

for the HMV database is given in Table 2-2, which presents those statistics for PFI and TBI

vehicles separately, and for the database as a whole (See Appendix A for vehicle-specific test

results.) As observed in the table, the mean FTP emissions from PFI vehicles are lower than for

TBI vehicles, even though the average mileage of vehicles falling into the two technology types is

the same However, emissions from one super emitter in the TEH sample (which is the only super

emitter in the 75-car database) has a significant influence on the TI31 average.’ When the mean

emission level is calculated without that vehicle, the HC results for TBI vehicles are only slightly

higher than the mean HC emissions from PFI vehicles, and the mean CO emissions from the two

technologies are nearly identical Mean NOx emissions, however, are still significantly higher for

* That vehcle was a 1985 Dodge Aries with FTP emissions of 10.32 g/mi HC, 149.45 g/m; CO, and 0.12 gími NOx

2-2

Trang 23

`,,-`-`,,`,,`,`,,` -S T D - A P I I P E T R O P U B L LibSO-ENGL 1797 O 7 3 2 2 7 0 0 5 b 4 2 2 b 7 T T =

Technology

PFI

Table 2- 1 Distribution of Vehicles Included in the High-Mileage-Vehicle Database by Model

Year, Fuel-Delivery Technology, and Site

Table 2-2 Mean FTP Emissions from the High-Mileage-Vehicle Database as a Function of Fuel-

Delivery Technology Compared to Predictions from TECH5

(74)

TECH5 - PFI TECH5 - TBI

a When rounded to the nearest 100, the mean odometer reading of PFI and TBI vehicles is identical

Numbers in parentheses reflect averages without the one super emitter in the sample

2-3

Not for Resale

Trang 24

`,,-`-`,,`,,`,`,,` -TBI vehicles relative to PFI vehicles To put those results into perspective, non-VM emission

estimates from TECH53 for PFI and TBI vehicles are also given in Table 2-2 Comparing the

TECH5 estimates to the HMV data, one observes that TECH5 is significantly overestimating

exhaust emission rates, particularly for vehicles equipped with PFI fuel-delivery systems

Table 2-3 summarizes the mean HC, CO, and NOx emission rates of the HMV database

segregated by procurement site As observed in the table, there is not a significant difference

among sites, particularly when the one super emitter is removed from the Chicago Heights

calculations In fact, mean CO and NOx emissions from the AZ-6 site (which was thought to

reflect a lower-income area) are lower than the AZ-7 and CH sites, even though the average age

of the vehicles tested there is higher Table 2-4 shows mean HC, CO, and NOx emission levels by

fuel delivery technology and procurement site Although the sample sizes of some of the

combinations shown in Table 2-4 prevent a rigorous comparison, the mean emission levels across

technology and site are not dramatically different when the comparisons are made without the

super emitter in the sample

A final analysis performed on the HMV database was an assessment of mean emissions by model year Those results are illustrated in Figures 2-1,2-2, and 2-3 for HC, CO, and NOx,

respectively Those figures indicate a fairly pronounced difference in emission rates @articularly for HC) between the 1985-1987 model years and the 1988 and later model years Based on the

results illustrated in Figure 2- 1 to 2-3, the HMV database was segregated into a 1985- 1987 model

year group and a 1988 and later model year group Summary statistics for those two groups are given in Table 2-5, which shows significantly lower mean emissions from the 1988 and later

model years compared to the 1985-1987 model years, even when the means are calculated

without the one super emitter in the sample Finally, Table 2-6 presents emissions by model-year

group and by fuel-delivery technology Although the sample sizes are relatively small when the

TECH5 is EPA?s preprocessor emission factors model that generates base emission rate equations and I/M credit

matrices for MOBLE5a

2-4

Trang 25

Table 2-3, Mean FTP Emissions from the High-Mileage-Vehicle Database by Site

Table 2-4 Mean FTP Emissions from the High-Mileage-Vehicle Database by Fuel-Delivery

Technology and Procurement Site

Site ( 4 (g/mi) (g/mi) (g/mi) Odometer Age

” The Az-ó site reflects a lower income area

Numbers in parentheses reflect averages without the super emitter

data are segregated in this fashion, the results indicate that model-year group appears to be a

more important determinant of emissions than the difference between PFI and TBI technology

or between procurement sites Again, this holds true even when the super emitter is removed

from the sample

2-5

Not for Resale

Trang 26

Trang 27

Size ( g m ( g m @/mi) Odometer Age

Figure 2-3 Mean NOx Emissions by Model Year for the High-Mileage-Vehicle Database

Table 2-5 Mean FTP Emissions from the High-Mileage-Vehicle Database by Model-Year

Not for Resale

Trang 28

`,,-`-`,,`,,`,`,,` -S T D A P I / P E T R O P U B L Yb50-ENGL 1797 m 0 7 3 2 2 7 0 0 5 b 4 2 3 1 Ob7 m

Tech Group

PFI

TBI

Table 2-6 Mean FTP Emissions from the High-Mileage-Vehicle Database by Fuel-Delivery

Table 2-7 compares the mean emission rates from the HMV project to similar data (i.e., oniy

fuel-injected vehicles with more than 100,000 d e s ) from EPA’s Hammond, Indiana, Ih4240

and emission factors test ~ r o g r a m ~ The data in that table have been segregated by model-year

group and by mileage increment @e., 100,000 to 125,000 miles; 125,000 to 150,000 miles; and

greater than 150,000 miles) As observed in the table, the pre-1988 model year mean HC and

CO emission levels from the HMV database are slightly higher than the Hammond database for

the first two mileage intervals, while results for the >150K group are lower The results for

NOx are completely opposite, indicating vehicles with high HC and CO emissions have

malfunctions that are suppressing NOx formation (e.g., rich operation) Also shown in Table 2-

7 are the results for the 1988 and later model year vehicles from the HMV database With the

exception of NOx for the >150K group, mean emission levels from the 1988+ group in the

The Hammond database used for this comparison was that used by EPA to develop base emission rate equations for

MOBILESa and reflects the in-use fleet in the 1990 to 1992 timefiame Not ail vehicles in that database received FTP

tests; rather, correlation equations were used to predict FTP values fiom lMî40 results for vehicles that were not FTP

tested This approach was taken because vehicles selected for FTP testing were not chosen at random (ie., the FTF’

sample was skewed with high-emitting vehicles), and this methodology ensured an unbiased sample was obtained for

subsequent analysis A detailed explanation of this approach can be found in Sierra Research, 1994

2-8

Trang 29

`,,-`-`,,`,,`,`,,` -S T D * A P I / P E T R O P U B L 'ib5O-ENGL L997 9 0 7 1 2 2 9 0 0 5 b 4 2 3 2 T T 3 9

Sample Size

Table 2-7 Comparison of Mean Emission Rates, Mileages, Ages, and Sample Sizes for Fuel-

Injected Vehicles in the High-Mileage-Vehicle Database and the Hammond, Indiana, Database by Mileage Increment

Includes vehicles fiom 1981 to 1987 model years (tested fiom February 1990 to Febnuuy 1992)

Includes vehicles from 1985 to 1987 model years (tested fiom A p d 1995 to August 1996)

2-9

Not for Resale

Tiêu đề	Analysis of High-Mileage Vehicle Emissions Data from Late-Model, Fuel-Injected Vehicles
Tác giả	David Lax, J. Steve Welstand, John C. Eckstrom, King Eng, Dennis Feist, Ana Rodriguez Forker, Frank S. Gerry, George S. Musser, Mani Natarajan, Rick Riley, Charles Schleyer
Trường học	American Petroleum Institute
Chuyên ngành	Health and Environmental Sciences
Thể loại	Publication
Năm xuất bản	1997
Thành phố	Sacramento

Định dạng
Số trang	58
Dung lượng	1,6 MB