WE PA
United States
Environmental Protection
Agency
Offlca of Air and Radiation
(ANR-443)
Washington, DC 20460
EPA 420-R-93-QQ7
May 1993
Federal Test Procedure
Review Project:
Preliminary Technical Report
3 Of
-------
-------
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
ANN AR8OR, MICHIGAN *81QS
OFP1CSOP
AIR AND RADIATION
DATE: May 13, 1993
MEMORANDUM
SUBJECT; Federal Test Procedure Review Project Preliminary
Technical Report -- Notification of Release --
FROM: Robert E. Maxwell, Directory
Certification Division
TO: Richard D. Wilson, Director f
Office of Mobile Sources
This memorandum serves to notify you of my approval for
release to the public of the Certification Division's Federal
Test Procedure Review Project Preliminary Technical Report. A
notice of availability and solicitation of comments will soon be
published in the Federal Register, a copy of the report has been
submitted to the Air Docket, and a limited distribution of the
report from Ann Arbor to the FTP review project mailing list will
occur as soon as the production copies are available.
-------
-------
Federal Test Procedure Review Project:
Preliminary Technical Report
May 1993
EPA 420-R-93-007
Certification Division
Office of Mobile Sources
Office of Air & Radiation
U.S. Environmental Protection Agency
DISCLAIMER: Although the information described in this preliminary technicai report has been funded in
part by the United States Environmental Protection Agency under contracts to Radian Corporation and to
Sierra Research, Inc., it has not been subjected to the Agency's peer and administrative-review processes.
It is being released for information purposes only and could be used in potential regulation development.
-------
-------
Table of Contents
Executive Summary j_
Chapter 1. Introduction ... a
Chapter 2. Background 10
2.1. The Air Quality Problem 10
2.2. History of the Federal Test Procedure 11
2.3. Statutory Provision 14
2.4. Areas of Potential Concern 14
2.4.1. Fuel 14
2.4.2. Temperature . 16
2.4.3. Altitude ..... 17
2.4.4. Driving Behavior (Including
Acceleration) 18
2.4.5. Teat Procedure Modifications 20
2.5 Heavy-Duty Vehicles and Engines 20
Chapter 3. Project Overview . 23
3.1. Research on Driving Behavior 24
3.2. Emission Assessment of In-Use Driving 27
3.2.1. Cycle Development 27
3.2.2. Emission Simulation Model 29
3.2.3. Vehicle Testing Programs 31
3.3. Notice of Proposed Rulemaking Development 31
Chapter 4. Driving Survey Methods 33
4.1. Survey Options 33
4.1.1. Choosing Survey Approach . . 33
4.1.2. Selecting Survey Sites 35
4.2. Instrumented Vehicle Approach 36
4.2.1. Key Features ..... 36
4.2.2. Instrumented Vehicle Field Results ... 39
4.2.3. Atlanta Instrumented Vehicle Study ... 42
4.3. Chase Car Approach 43
4.3.1. Key Features ......... 44
4.3.2. Chase Car Field Results . 46
4.3.3. Los Angeles Chase Car Study ....... 47
4.4. Analyses of Potential Bias . 47
4.4.1. Analysis of Instrumented Vehicle Method . 48
4.4.2. Analysis of Chase Car Method 52
4.5. Selection of Principal Data Set 55
4.5.1. Comparison of the Two Surveys Results . . 55
4.5.2. Rationale for Choosing 3-parameter
Instrumented Vehicle data 58
Chapter 5. Analytical Methods and Considerations 60
5.1. Driving Behavior Variables ............ 60
5.1.1. Speed-Based Measures of Driving
FTP Prtlimmtry Report; Miy 14, 1993
-------
Behavior 60
5.1.2. Trip-based Measures .... 62
5.1.3. Vehicle-Based Measures ......... 65
5.2. Descriptiva Methods 66
5.3. Statistical Accuracy . 68
5.4. Computer Resource Issues 69
5.5. Driving Conditions 70
5.5.1. Engine and Catalyst Cooldown 70
5.5.2. Road Grade 75
Chapter 6. Discussion and Analysis of Driving Survey
Results 77
6.1. Summary of In-Use Driving Results for Four
Cities 77
6.1.1. Driving Behavior - Second-by-Second
Analysis , 77
6.1.2. Driving Behavior - Trip Analysis .... 82
6.1.3. Choosing Representative Survey Data ... 83
6.2. Analysis of Baltimore Driving Conditions 37
6.2.1. Speed-based Measures of Driving
Behavior 87
6.2.2. Trip Measures 98
6.2.3. Vehicle.-based Measures 108
6.2.4. Vehicle Soak ..... 109
6.2.5. Trip Start Driving Activity 115
6.3. Comparisons to the Federal Test Procedure ..... 127
6,3.1, Speed-based Measures 127
6.3.2. Trip Comparison 140
6.3.3. Vehicle Soak 143
6.3.4. Trip Start Driving Activity 145
6.3.5. Road Grade 154
Chapter 7. Test Cycle Development Methods and Approach. . . . 155
. 7.1. Test Cycle Methods 155
7.2. Method Types 156
7.2.1. Segment-splicing Methods 156
7.2.2. Monte Carlo Simulation 157
7.2.3. Engineering Cycles ... 158
7.3. Cycle Criteria 159
7". 4. Cycle Validation 159
7.5. Current EPA Test Cycle Development ISO
List of Appendixes 161
Appendix A. Chase Car Method Bias Analysis: Supplementary
Tables
Appendix B. " Baltimore 3-Parameter Vehicle Characteristics
Appendix C. Summary Statistics, Distributions, and Graphs
Appendix D. Baltimore Soak Periods
Appendix E. Trip Start Driving Activity
FTP Prelcnmijy Report: Miy 14, 1993
-------
List of Figures
HO*. Title .-
5-1 Plymouth Acclaim Engine Cooldown Curve 72
5-2 t Plymouth Acclaim Catalyst Cooldown Curve 72
6-1 Distribution of Speed: Baltimore, Spokane, Los
Angeles, Atlanta 80
6-2 Distribution of Power: Baltimore, Spokane, Los
Angeles, Atlanta , 33.
6-3 Distribution of Speed: Baltimore-Exeter and
Spokane 35
6-4 Distribution of Speed: Baltimore-Rossville and
Atlanta 85
6-5 Distribution of Speed and Acceleration - Baltimore 90
6-6 Distribution of Speed and Acceleration - Baltimore,
Idle Excluded . . 91
6-7 Conditional Distributions of Acceleration:
Baltimore 3-Parameter Vehicles 92
6-8 Distribution of Trip Time: Baltimore 3-Parameter
Vehicles. 99
6-9 Distribution of Trip Distance: Baltimore 3-
Parameter Vehicles. 99
6-10 Cumulative Distribution: Maximum Speed per Trip,
Baltimore 3-Parameter Vehicles 105
6-11 Cumulative Distribution: Maximum Power per Trip,
Baltimore 3-Parameter Vehicles. 106
6-12 Cumulative Distribution: Maximum Acceleration per
Trip, Baltimore 3-Parameter Vehicles 107
6-11 _ Hill I - Average Tailpipe Emissions 110
6-14 Typical Catalyst and Engine Cooldown Curves 112
6-15 Baltimore Soak Period Distributions .....'. 113
6-16 Speed Distribution by Phase 121
6-17 Cumulative Power Distribution by Phase 122
6-18 Distribution of Miles Driven from Start of Trip.. 126
6-19 Distribution of Speed: Baltimore and FTP, Baltimore
3-Parameter Vehicles 132
6-20 Cumulative Distribution of Acceleration: Baltimore
and FTP, Baltimore 3-Parameter Vehicles 133
-------
5-21 Cumulative Distribution of Power: Baltimore and
FTP, Baltimore 3-Parameter Vehicles 134
6-22 FTP Speed-Acceleration Envelope 136
6-23 Baltimore 3-Parameter Speed-Acceleration Range with
FTP Envelope '. 137
6-24 FTP Envelope: Difference between Baltimore and FTP 138
6-25 Distribution of Speed and Acceleration: Baltimore
Driving Outside FTP Envelope. , .' 139
6-26 Federal Test Procedure Driving Trace. 142
6-27 FTP v. In-use: Start Speed Distribution. 148
6-28 FTP v. In-use: Start Cumulative Power
Distributions 149
6-29 FTP v. In-use: Distribution of Miles Driven, 152
-------
List of Tablaa
Title " p3ge
4-1 Alternative In-Use Driving Survey Methods 34
4-2 Summary of Instrumented Vehicle Field Results 41
4-3 Summary of Chase Car Field Results 47
4-4 Reasons for Chase Car Losing Target 54
4-5 Comparison of Chase Car and Instrumented Vehicle
Driving Behavior , 57
5-1 Road Gradient Distribution by Percentage of
Vehicle-Miles Traveled , 75
6-1 Comparison of Driving Behavior for Four Cities 78
6-2 Comparison of Road Type Distribution for Baltimore
and Spokane 86
6-3 . Baltimore Stratified Trip Measures 103
6-4 Temperature Distributions at Trip Start 114
6-5 Trip Start Averages by 80-Second Phase , 119
6-6 Initial Idle Time, Baltimore and Spokane 123
6-7 Initial Idle Time for Baltimore and Spokane
Combined, and Atlanta 124
6-8 Summary Statistics: Baltimore and FTP, Baltimore 3-
Parameter Vehicles 128
6-9 Distribution of Speed: Baltimore and FTP, Baltimore
3-Parameter Vehicles.. .,..... 129
6-10 Distribution of Acceleration: Baltimore and FTP,
Baltimore 3-Parameter Vehicles ... 130
6-11 Distribution of Power: Baltimore and FTP, Baltimore
3-Parameter Vehicles 131
6-12 Comparison of Trip Measures for FTP and Baltimore 141
6-13 FTP vs. Baltimore Temperature Distribution at Trip
Start 144
6-14 Comparison of FTP and Actual Start Driving
Behavior 146
6-15 Revised FTP and Actual Start Comparison 147
6-16 Initial Idle Times (Seconds) 151
6-17 FTP vs. Baltimore Comparison of Miles Driven from
Start of Trip 153
-------
-------
Executive Summary
Pursuant to Section 206(h) of the Clean Air Act, as amended
in 1990 (Act, or CAA), 'EPA has undertaken a review of the Federal
Test Procedure (FTP) used to test light duty vehicle and light
duty truck emissions to determine whether it adequately
represents actual current driving conditions. The driving cycle
used for the-FTP was adopted over twenty years ago and
accumulated research suggests that it may no longer adequately
represent overall vehicle emission control performance under
current driving conditions.
Project Overview
The principal subject of this report is driving behavior,
including acceleration and trip start patterns. After reviewing
existing research on driving behavior, the Agency determined that
new surveys were needed to assess current driving in U.S. urban
nonattainment regions. A parallel study of current technology
vehicle emissions under the full range of driving conditions is
also in progress. Results from these efforts will be combined in
determining the need for test procedure revisions.
This preliminary technical report discusses how the driving
surveys were conducted, presents analyses of the results, and
compares the data to the existing FTP. Quantitative assessments
of the emission impacts are still in progress and are not
discussed in this report.
With support from the American Automobile Manufacturers
Association (AAMA) and the Association of International
Automobile Manufacturers (AIAM) , EPA conducted surveys of driving
behavior in Baltimore, MD, and Spokane, WA. Two methods of data
FT? fnliuiiuary Report; Miy 14, 1993
-------
collection were employed. In an instrumented vehicle study, 113
Baltimore and 102 Spokane vehicles were equipped with "3-
parameter" datalogger packages that recorded second-by-second
speed and two other variables during periods of operation. As
part of the same surveys, the manufacturers recruited 79 vehicles
for study using "6-parameter" instruments designed to measure
additional variables.
The instrumented vehicles were observed for seven to ten day
periods. A separate chase car study collected similar speed data
in the two cities using a laser device mounted on a patrol car
that tracked in-use target vehicles. This produced relatively
short sequences of data on a much larger sample of vehicles.
The Baltimore and Spokane surveys are supplemented by data
collected in two other cities. EPA's Office of Research and
Development sponsored an instrumented vehicle study in Atlanta,
GA. Finally, the California Air Resources Board (GARB) sponsored
a chase car study in Los Angeles similar to the chase car studies
in Spokane and Baltimore.
For reasons relating to representativeness, availability,
and precision of the survey data, most of the discussion in this
report is confined to driving observed in the Baltimore 3-
parameter instrumented vehicle study. Participants in the
instrumented vehicle study in Baltimore were recruited from two
different centralized Inspection/Maintenance stations, one in
central Baltimore and one in a suburb just outside Baltimore.
The urban site yielded driving characteristics similar to those
in Spokane and the suburban site was similar to Atlanta and Los
Angeles. As it was not possible to obtain detailed analyses of
the Atlanta and Los Angeles data in time for this report, EPA
decided to use just the Baltimore data for purposes of this
report. This should yield a more representative picture than
FTP Ptelaamuy Report: May 14, 1993
-------
-also including the Spokane data. Future research will include
further analyses of data obtained from the other sources.
EPA also entered into a cooperative agreement with the New
York State Bureau of Mr Research to obtain test data and
analyses of the engine and catalyst cooldown processes and
factors affecting the cooldown rates. These data are used to
assess the condition of the engine and catalyst during engine
start-up.
Preliminary Indications of Actual Driving Behavior
Speed and Acceleration
Speeds were much higher in Baltimore than are represented on
the FTP. The average speed in Baltimore was 24.5 mph (median
speed was 23.7). The speeds observed ranged to almost 95 mph;
6.4% were above 60 mph and 2.6% above 65 mph. By comparison, the
FTP has an average speed of 19.6 mph with a maximum of 56.7 mph.
About 8.5% of all speeds in Baltimore exceeded the FTP maximum.
Acceleration rates in Baltimore were also significantly
higher than those on the FTP. The acceleration rates observed
ranged up to 15 mph/sec, with a standard deviation of I.5.1 The
FTP has a maximum acceleration rate of 3.3 mph/sec and a standard
deviation of 1.4. About 2.5% of all driving in Baltimore
exceeded 3.3 mph/sec.
Power-related measures also indicate that the observed
driving behavior was more aggressive than the FTP. Specific
'Mean acceleration rates generally average to zero and are not a useful
measure of comparison.
FT? Preliminary Report M*y 14, 1993
-------
power2 for the Baltimore sample ranged up to 558 mph'/sec and
averaged 46.0, with a median of 34.7, The FTP has"a maximum
power of 192, average of 38.6, and median of 21.6. An analysis
was also done of the scatter of speed-acceleration points
occurring in the Baltimore sample outside the FTP envelope of
speed and accelerations. These points represent about 18% of
total Baltimore driving time.
Driving BehaviorDeterminants
Vehicle Type - Speed distributions were fairly similar for
each of the three categories analyzed; trucks, sedans, and high
performance vehicles. However, high performance vehicles
demonstrated more aggressive driving behavior thaa the other
classes, with over twice as much operation at power levels above
200 mphVsec.
Vehicle Age - Newer vehicles (1983 and later) had higher
average speeds than older vehicles (25.1 mph v. 21.2 mph), were
driven somewhat longer and farther per day, and averaged fewer
trips and slightly fewer stops per mile. The data indicate that
newer vehicles spend more time at high speeds and are used for
longer trips than older vehicles. However, analyses of the
aggressiveness of the driving behavior, as measured by
acceleration and power distributions, indicate very little
difference between older and newer vehicles.
lThe povar needed from an engine to accelerate a vehicle is proportional
to both tha vehicle speed and the acceleration rate. Thus, neither variable, by
itself, 10 a good measure of tha load placed on the engine during acceleration.
The joint distribution of speed and acceleration is the best measure, but it oust
be examined in three dimensions, which is difficult to visualize and comprehend.
While not as good as the joint distribution of speed arid acceleration, tha best
two-dimensional measure is "specific power," which is roughly equivalent to (2
* speed * acceleration) . This measure is used extensively in this report and has
tha units mph3/sec.
FTP Prelmnuy Report: May 14, 1993
-------
Time of Day - Average speeds were lowest during the evening
rush period of 4-7 pm (23"'.0 mph) and highest during the morning
rush period of 6-9 am. and night driving period of 9 pm to 1 am
(25.4 mph). Extremes .of acceleration and specific power were not
highly associated with time of day. About twice as many trips
began during the evening rush period as during the morning rush
period. The morning peak period had the longest average trip
lengths (6.8 miles), while mid-day trips averaged only 4.4 miles.
Time of Week - Average weekend speeds were substantially
higher than on weekdays; 26.3 mph and 23.9 mph, respectively.
This is due, in part, to increased high speed driving; 13.6% of
all driving during the weekend exceeded 55 mph, compared to 9.8%-
during weekdays. Weekend driving also produced substantially
higher average time, distance, and number of trips than weekdays.
However, acceleration and specific power measures do not indicate
that weekend driving is more aggressive than on weekdays.
Trip Patterns
Average in-use trip lengths are much shorter than the FTP,
which represents a 7.5 mile trip. The average observed trip3
covered 4.9 miles. The median value of trip distance indicates
that "typical" trips are even shorter, only 2.5 miles. One of
the in-use impacts of shorter trips is that a much higher
proportion of overall driving is done within 0.67 mile of vehicle
starts (12.0% v. 8.9% on the FTP), prior to engines and catalysts
reaching normal operating temperatures. The frequency of stops
on the FTP is also uncharacteristic of in-use trips; the average
distance between stops on the FTP is only 0.41 miles compared to
3For che purposes of this report, a trip has been defined as beginning when
the engine is turned on and ending when the engine is shut off (although engine
off times of lesa than 18 seconds are ignored).
FTP Prelanmtry Report; May 14, 1993
-------
0.87 in Baltimore. Despite these differences, the FTP and
Baltimore trips disagree only slightly in the proportion of time
spent in the four operating modes: idle, cruise, acceleration,
and deceleration.
vehicleSoaks
The in-use data contains a large proportion of intermediate
soak periods (that is, the time between the end of a previous
trip and the beginning of the next one) that are not reflected on
the FTP. The FTP contains soak periods of 10 minutes and 12-36
hours; almost 40% of all soak periods in Baltimore were between
10 minutes and 2 hours. As catalysts cool off much faster than
engines and most are almost completely cold in about 45-60
minutes, this is a potential emission concern. Analyses indicate
that only about 30% of all in-use starts occur with catalysts hot
enough to be immediately effective; the FTP implicitly assumes
that 57% of all starts occur with hot catalysts. On the other
hand, the FTP implicitly assumes that 43% of all starts occur
with cold engines, while less than 25% of in-use starts occur
with cold engines.
Trj.p . gtart privinq Activity
While the FTP has lower speeds and is leas aggressive than
in-use driving behavior, overall, the reverse occurs for the
first few minutes after a vehicle start. The average observed
speed during the first 80 seconds of all trips (the initial idle
period waa not included in this period) was only 14.4 mph,
compared to 23.1 mph for the first micro-trip on the FTP. The
average in-use speed 81-240 seconds into the trip was 22.8 mph,
compared to 29.8 for a comparable period on the FTP. The
aggressiveness of the FTP was also off substantially, with the
first micro-trip on the FTP substantially lesa aggressive than
FTP PrtliaUBuy Report: M»x W. 1953
-------
in-use driving and the second micro-trip greatly overaggressive.
Under similar ambient air temperature conditions, the
initial idle time on the FTP after a cold start is similar to
observed data. However, after a hot start the initial idle time
on the FTP is much longer than observed data. The FTP uses an
initial idle time of 20 seconds after both cold and hot starts;
observed initial idle times after cold starts averaged 28 seconds
with a median of 9 seconds, while initial idle times after hot
starts averaged only 12 seconds with a median of 5 seconds.
Emission Impact Assessment Plans
The analysis of data obtained from the driving surveys
described here indicates significant differences exist between
actual driving behavior and the FTP. The driving survey data
will serve as the primary input into programs to assess the
difference between emissions predicted by the FTP and emissions
that occur in actual driving. This assessment requires the
development of driving cycles that are more representative of the
driving behavior information obtained from the surveys. EPA, in
cooperation with the California Air Resources Board, has
investigated cycle generation alternatives and developed improved
methods. Cycles generated from the driving survey data using
these methods will be used in test programs to quantify in-use
emissions.
FTP Prcinnmiry Report May 14, 1993
-------
Caapter 1. Introduction
The cornerstone of the Clean Air Act is the effort to attain
and maintain National Ambient Air Quality Standards (NAAQS).
Regulation of emissions from on-highway, area, and stationary
sources prior to enactment of the Clean Air Act Amendments (CAAA)
of 1990 has resulted in significant emission reductions from
these sources. However, many air quality regions have failed to
attain the NAAQS, particularly for ozone and CO. This is due to
many factors, including the number of vehicles on the road and a
corresponding increase in the number of miles driven by the in-
use fleet which, even though single vehicles have experienced
significant emission reductions, has increased total emissions
from the motor vehicle fleet.
The Clean Air Act, as amended (CAA, or Act) , contains a
large number of provisions to further improve ambient air
quality. Section 206(h) of the Act requires that EPA review its
regulations for the testing of motor vehicles and revise them if
necessary to ensure that motor vehicles are tested under
circumstances reflecting actual current driving conditions. This
preliminary technical report discusses the driving behavior
research conducted by the EPA in its review of the Federal Test
Procedure (FTP) under Section 206(h) of the Act. The report
discusses the need for the research, methods and approaches, and
analyses of the driving behavior observed. Results from previous
research are not included in this report; nor are quantitative
assessments of the emissions impact of such driving behavior
{although qualitative implications are discussed).
Chapter 2 discusses background information, including the
history of the Federal Test Procedure and the Act's provisions
for EPA review of the FTP. Chapter 3 provides an overview of the
goals and general approach to the entire FTP project, including
FTP Pratiraiouy Report: Miy 14, 1993
-------
'assessing the emissions impact of actual driving behavior and
development of proposed rules. Chapter 4 describes the survey
methods and programs used to determine actual driving behavior.
Chapter 5 discusses the analytical methods used to evaluate the
driving survey data. Chapter 6 presents the results of the
Agency's analyzes of driving behavior and provides comparisons to
the existing test procedures. Chapter 7 describes the
methodologies developed to generate test cycles from the survey
data.
FTP fntimaaij Report: Ml? 14, 1993
-------
Chapter 2. Background
2.1. Tha Air Quality Problem
Motor vehicles are a well known major source of volatile
organic compounds (VQC5 and oxides of nitrogen, (NQx) , both of
which are precursors of ground level ozone, or smog. Motor
vehicles are also a major source of carbon monoxide (CO)
emissions. While significant progress has been made over the
past two decades in controlling motor vehicle emissions, as of
August 1990, 96 air quality control regions still failed to meet
the national ambient air quality standard {NAAQS} for ozone, and
forty-one regions failed to attain the NAAQS for. CO.4
As an example of the impact of motor vehicles, estimates of
ozone precursors can be used. Volatile organic compounds and NOx
interact in sunlight to form ozone. Motor vehicles are estimated
to contribute approximately 25% of VOC emissions nationally
during the summer months.3 Small "area sources" such as
bakeries, dry cleaners, and consumer solvents contribute 45% and
large point sources such as petroleum refineries contribute 10%
of VOC emissions.6 Additionally, motor vehicles are estimated to
contribute approximately 42% of nationwide CO emissions during
the winter months.7 Clearly, motor vehicles are a major source
of ozone precursor and CO emissions in nonattainment areas.
*EPA, Office of Public Affairs, Environmental Newa. August 16, 1990.
5Nonroad Engine and Vehicle Emission Study- -Report, U.S. SPA, SPA-21A-2001,
November 1991.
slbid.
7Ibid.
FTP Prelaninuy Report: M*y 14, 1993
10
-------
2.2. History of the Federal Test Procedure
The FTP is the test procedure used to determine compliance
of light-duty vehicles (LDV) and light-duty trucks (LOT) with
federal emission standards.* As designed, the FTP was intended
to represent typical driving patterns in primarily urban areas.
«*«*;
Preproduction vehicles are tested using the FTP as part of
the motor vehicle certification process. The certification
process is used to establish that each vehicle is designed to
comply with emission standards for its full useful life. The FTP
is also used to test production line and in-use vehicles for
compliance with appropriate emission standards.
»
The FTP is more than just a driving cycle; it provides a way
to consistently and repetitively measure concentrations of EC,
Nox, CO, and carbon dioxide (C02)end.asions which occur when a
vehicle is driven over a simulated urban driving trip.* The
principal elements of the test are designed to test the
evaporative and exhaust emissions under several simulated
situations. Evaporative emissions are tested after heating the
fuel tank to simulate heating by the sun (the diurnal test) and
again after the car has been.driven and parked with a hot engine
(the-'hot, so^-:;test);l>_/,/^Bx^usc,:;^iLssions' are measured,..Jayr .driving- -.
"the-;vehicle'/"(pX₯<|e_dt;"dn- .^-idyziaiiMaDeicer)-- on a. 'simulated' ttrfcan-J": ','-..-
driving trip under , twct-.eoiiditions: with a cold start designed to
represent a morning start-up after a long soak (a period of non-
use) and with a hot start that takes place while the engine is
*Tha regulations that encompass the many aspects of the FTP are generally
contained in 40 CFR Part 86, Subparts A and B.
*For further discussion on the development of this cycle, see: Kruse,
Ronald B., and Thomas A. Hula, SA1 Paper #730553 "Development of the Federal
Urban Driving Schedule," 1973. A speed-time trace of this cycle is contained in
40 CFR Part 36, Appendix Z.
FT? Prctimiauy Repot M*y 14, 1993
11
-------
scill hoc. The FTP also encompasses all factors relevant to
vehicle testing, such as fuel, vehicle preconditioning, ambient
temperature and humidity, aerodynamic loss, and vehicle inertia
simulations. In addition to evaporative and exhaust emissions,
the FTP is also used in evaluating fuel economy.
The driving cycle used for the FTP was derived to simulate a
vehicle operating over a road route in Los Angeles believed to be
representative of typical home to work commuting. The original
road route was selected in the mid-1960s10 by trial-and-error to
match the engine operating mode distribution (based on manifold
vacuum and rpm ranges) obtained in central Los Angeles using a
variety of drivers and routes with the same test vehicle.
Using an instrumented 1964 Chevrolet, recordings were made
of actual home-to-work commute trips by employees of the state of
California's Vehicle Pollution Laboratory. By trial and error, a
specific street route in the vicinity of the Lab was found that
matched the average speed/load distribution on the commute trips.
That 12 mile route was called the "LA4."
In an effort to develop an improved Federal Test Procedure
(baaed on speed-time distributions rather than manifold vacuum
and rpm ranges), six different drivers from EPA's West Coast
Laboratory drove a 1969 Chevrolet over the LA4 route. The six
traces were analyzed for idle time, average speed, maximum speed,
and number of stops per trip. The total time required for the
six trips ranged from 35 to 40 minutes, with an average of 37.6
minutes. One of the six traces demonstrated much harder
acceleration rates than the other five and was discarded. The
other five traces were surprisingly similar. Of those five, the
l°Q.C. Haas, ee. al., "Laboratory Simulation of Driving Conditions in the
Los Angalea Area," SAE Paper No, 660546, August, 1966.
FTP Prcixuuiy Report; M»y 14, 1993
12
-------
trace with the actual time closest to the average was selected as
the most representative speed-time trace.. That trace contained
28 "hills" of non-zero speed activity separated by idle periods
and had an average speed of 19.2 miles per hour.
Based on a 1969 report on driving patterns in Los Angeles,11
the average trip length was estimated to be 7.5 miles. Several
of the hills and portions of others were eliminated in order to
shorten the cycle to 7.5 miles while maintaining the same average
speed. The shortened route, designated the LA4-S3, was 7.486
miles in length with an average speed of 19.8 raph. Slight
modifications to some of the speed-time profiles were also made
in cases where the acceleration or deceleration-rate exceeded the
3.3 mph/s limit of the belt-driven chassis dynamometers in use at
the time. Mass emission tests comparing the shortened cycle to
the full cycle showed very high correlation. The final version
of the cycle was designated the LA4-S4 cycle and is 7.46 miles in
length with an average speed of 19.6 mph.
This cycle is now commonly referred to as the "tA4" or the
Urban Dynamometer Driving Schedule (UDDS) . It has been the
standard driving cycle for the certification of LDVs and LDTs
since the 1972 model year» Beginning with the 1975 model year,
the cycle was modified to repeat the initial 505 seconds of the
cycle following a 10 minute soak at the end of the cycle. This
allows emissions to be collected on a "hot" start (the engine is
still warm) as well as after a cold start and during operation.
The test then provides a more accurate reflection of typical
UD.H. Dearm and R.Z.. Lanotaraux, "Survey of Average Driving Patterns in the
Los Angeles Urban Area," TM-
-------
customer service than running just one 7.46 mile cycle -from a
cold start.
2.3. Statutory Provision
Section 206 (h) of the Act directs that EPA "review and "
revise as necessary" the regulations pertaining to the testing of
motor vehicles to "insure that vehicles are tested under
circumstances which reflect the actual current driving conditions
under which motor vehicles are used, including conditions
relating to fuel, temperature, acceleration, and altitude."
This preliminary technical report documents technical
aspects of EPA's review process to date. This study does not
attempt to determine or justify the need for revisions to the
FTP. " Such a justification would be part of any regulatory
decision making that EPA may conduct on FTP modifications.
2.4. Areas of Potential Concern
Section 206 (h) of the Act specifically requires that EPA
consider actual driving conditions under which motor vehicle are
used, including "conditions relating to thejse four areas: fuel,
temperature, altitude, and acceleration. Following is a
discussion of these conditions, plus a discussion of other
driving conditions of potential concern when reviewing the
adequacy of the FTP.
2.4.1. Fuel
- The composition of the gasoline used for the FTP
(commonly referred to as indolene) was established by regulation
FTP Prclimmiry Report; May 14, 1993
14
-------
over 20 years ago.12 While it was representative of in-use fuel
at the time, commercial or in-use fuel properties have changed
significantly since then, in some cases having a major impact on
vehicle emissions, both tailpipe and evaporative.13 Studies
conducted during the 1980s indicated that vehicles tend to have
higher emissions during operation on commercial gasoline than on
indolene, particularly through evaporative losses.
To address this concern, EPA has established volatility
limits for gasoline and alcohol blended fuels. These regulations
capped the allowable Reid vapor pressure for commercial gasoline
during the summer months. The second phase of these controls
became effective in the summer of 1992.'* As a result of these
actions, the emissions of a vehicle fueled with indolene are more
representative of the emissions from vehicles fueled, with
commercial gasoline.
Diesel fuel - The Agency has taken steps to reduce the
sulfur content of in-use diesel fuel. Regulations published on
' *
May 7, 1992, to reduce the sulfur content in diesel fuel, are
scheduled to take effect on October l, 1993.l*
Alcohol and Other Fuels - The Agency promulgated
regulations in 1989 which established emission standards and test
procedures for vehicles fueled with methanol and proposed similar
regulations in 1992 for vehicles fueled with natural gas and
liquef-ied petroleum gas. At this early stage of alternative fuel
I240 CFR 186.113-94.
Evaporative emissions include diurnal, hot soak, refueling, and running
losses.
1455 FR 23658 (June 11, 1990) .
U57 FR 19535 (May 7, 1992).
FT? Preliminary Report: May 14, 1993
15
-------
development, it is impossible to know what the real-world fuel
compositions will be for any of these fuels when used in
automotive applications. In each of these rulemakings, EPA has
avoided adoption of narrow fuel specifications, specifying
instead that test fuels be representative of typical in-use
fuels.
2.4.2. Temperature
The FTP is conducted between 68°P and 86°P and includes a
cold start in its driving cycle.16 As ambient air temperatures
decrease, cold start emissions increase because a richer air/fuel
mixture must be employed to ensure the presence of sufficient
fuel vapor for combustion. In addition, colder temperatures lead
to longer warm-up times.
This is not a major concern for ozone, which is primarily a
summertime phenomenon, but it is for CO. Most CO exceedances
occur "from December to March and over half occur at temperatures
below 45°F.
To reduce the emissions generated from motor vehicles during
cold temperature operation, EPA recently issued 20°F CO emission
standards and test procedures. These regulations were issued on
July 17, 1992, and are scheduled for phase-in beginning with the
1994 model year.17 The regulations also establish interim
temperature defeat device criteria to maintain proportional CO
emission control between the 20°F standard and the warm
temperature standards. These regulations ensure that the
16An engine start is considered to be a "cold" stare if it is preceded by
a long uninterrupted soak, such as those starts chat occur after an overnight
soaJc.
I757 FR 31888 (July 17, 1992) .
FTP Prelimajjuy Report: M»y 14, 1993
16
-------
Agency's test procedures properly reflect the impact of
temperature on CO emissions. As the cold CO regulations will
prevent emission step-functions just below 68°P that could also
impact HC emissions, they will also ensure that the FTP is
representative of HC emissions at colder temperatures.
At warmer temperatures the primary emission concern is
increased fuel evaporation. The Agency has recently published
regulations revising its evaporative test procedures to address a
number of concerns, including temperature.11 The final
regulations are expected to specify ambient test temperatures of
95°F. These new test requirements should ensure that vehicles
can control evaporative emissions for most in-use events.
2.4.3. Altitude
It has long been recognized that without compensation for
the lower air density at high altitude locations engines tend to
operate at rich air/fuel mixtures more frequently and, therefore,
have excessive HC and CO emissions. Virtually all LDVs have been
required to meet emission standards at both low and high
altitudes without adjustment or modification since the 1934 model
year. Light-duty trucks and light-duty vehicles have had
separate high altitude standards since the 13S2 model year.
Regulations published on June 5, 1991, will require LDTs to meet
emission standards at both low and high altitudes without
adjustment or modification beginning with the 1997 model year.19
The cold temperature CO regulations require that both LDVs and
LDTs meet the cold temperature CO standard at both low and high
altitudes without modification. The FTP does not specify an
altitude range in which the test must be conducted. In effect.
US8 FR 16002, March 24, 1993.
1956 FR 2S724, Jun« S, 1991.
FTP Prcliniiauy Report May 14, 1993
17
-------
the regulations allow the FTP to be conducted at any altitude and
this, in fact, occurs.
2.4.4. Driving Behavior {Including Acceleration)
Current technology vehicles have achieved impressive
reductions in emissions during normal operation, primarily due to
catalyst technology development. Catalyst conversion
efficiencies (that is, the rate at which HC and CO are oxidized
into carbon dioxide (C02) and water vapor, or the rate at which
oxides of nitrogen (Nox) are reduced to nitrogen and oxygen) in a
modern, properly operating, warmed-up vehicle can simultaneously
exceed 98% for HC, 99% for CO, and 90% for Nox. This includes
typical transient operation in urban traffic situations, such as
that represented by the FTP. However, these simultaneous
catalyst conversion efficiencies are only achievable by a three-
way catalyst in a very narrow range of air/fuel ratios around the
minimum theoretical air requirement for complete combustion
(called stoichiometry). Thus, modern, properly operating
vehicles are designed to operate at stoichiometry as much as
possible during the FTP.
Two types of operation make it difficult to operate an
engine at stoichiometry. The first type of operation is a cold
start. Fuel must be vaporized with air to combust properly. When
the engine is cold, not enough heat is available to properly
vaporize the fuel, requiring the addition of more fuel for proper
operation. Cold start emissions are also increased due to the
lack of conversion activity in the catalyst until it heats up
(little catalyst activity occurs below roughly 600°F) . Thus,
emission rates during cold starts can be 20 to 100 times the
emission rates during stoichiometric operation. In fact, the
majority of emissions from modern, properly operating vehicles
operating over the FTP occur during the first 10% of the test,
FT? PrtUminjuy Rejxwt: M*)> 14, 1993
18
-------
before the engine and catalyst have warmed up. This raises a
concern within the Agency as to whether or not the cold start
portion of the FTP properly reflects the proportion of time
vehicles actually spend in the warm-up mode.
The second type of operation that makes it difficult to
operate an engine at stoichiometry is high engine loads. High
loads on an engine running at stoichiometry can dramatically
increase engine and catalyst temperatures. These elevated
temperatures increase engine-out NOx emissions and can cause
engine knock and/or damage to the.catalyst. The performance and
driveability of an engine under high load can be improved by
running with a richer air/fuel mixture. Thus, to prevent over-
temperature damage to the catalyst, and to ensure the best
possible driveability and performance, vehicles are often
designed to operate rich under high engine loads. While such a
design has little effect on NOx emissions,30 it increases HC and
CO by almost the same 20 to 100 times factor as cold start
operation. This also raises a concern within the Agency as to
whether or not a significant amount of high engine load operation
occurs in-use that is not properly reflected on the FTP. Due to
the nonlinear nature of the emission rates, this amount of
driving could actually be fairly small and still have a
significant emission impact.
A wide variety of in-use factors impact the amount of time
vehicles spend in either a warm-up or high-load mode. Factors
related to warm-up include distributions of trip length, time
between trips (referred to as "soak time"), ambient air
temperature, initial idle time, and driving behavior. Factors
20Engine-out NOx emissions decrease under rich operation, but; NOx reduction
efficiencies in the catalyst also drop. Overall, there may be a alight increase
in tailpipe NOx emissions under rich operation, but the effect is relatively
minor and varies from vehicle to vehicle.
FTP Prthminiiy Report: Mqr 14, 1993
19
-------
Chat can cause high loads on an engine include high acceleration
rates, high speeds, positive road grades, air conditioning
operation, or some combination of factors (such as moderate
acceleration up a moderate grade). Complicating the assessment
is that different vehicles have very different calibration
strategies. Thus, the impact on emissions of the exact same
driving behavior may vary widely from vehicle to vehicle.
2.4.5. Test Procedure Modifications
For repeatability, it is necessary to conduct certification
and enforcement testing in a laboratory. This gives rise to a
number of procedures designed to simulate the actual forces and
conditions a vehicle would experience on the road. The most
significant of these is the dynamometer itself, which must
simulate the inertial mass of the vehicle (which doesn't actually
move in the laboratory), aerodynamic losses, and duplicate tire-
to-road traction and rolling losses. Other potential areas of
concern, are the type and amount of external air provided for
vehicle cooling, the manner in which air conditioning losses are
simulated, and manual transmission shift points.
The impact of these test procedure factors on emissions has
been investigated at various times in the past. As a result of
these investigations, the Agency has already begun the process of
converting from small turn-roll waterbrake dynamometers to large
single-roll electric dynamometers in its own laboratory in Ann
Arbor, Michigan.
2.5 Heavy-Duty Vehicles and Engines
This preliminary report discusses 'the test procedures for
light-duty vehicles and light-duty trucks, but does not address
the very different procedures employed for testing heavy-duty
FTP Pnlimnuy Report: M»x I*. 1993
20
-------
engines.21 The Agency believes this is consistent with the
intent of Congress, as expressed in the text of the Act and its
legislative history. Section 206 (h) requires that EPA review
test procedure regulations issued under Section 206 for a
specific purpose - to insure that vehicles are tested under real
world conditions. While the test procedure for light-duty
vehicles and trucks involve, vehicle testing, heavy-duty engines
are tested differently, using an engine only test. Since the
purpose of this provision is to review and, if necessary, revise
EPA's regulations for testing motor vehicles, it is reasonable
for EPA's review to address light -duty vehicle test procedures
and not heavy-duty engine test procedures. The legislative
history confirms that the review required under Section 206 (h) is
to address EPA's test procedures for light-duty vehicles and
trucks.22 This interpretation is confirmed by the fact Congress
separately addressed the adequacy of EPA's test procedures for
urban bus engines, a subset of heavy-duty engines. Section
219 (e) requires that test procedures used for urban bus emission
standards "reflect actual operating conditions." If Section
206 (h) addressed both light and heavy-duty test procedures this
provision would amount to surplusage.23
. ,_ .EPA of 'course -has. the discretion to review .the test,
' engines ,' and in fact;' did conduct" such a
conducted, a _ large-scale,. In-depth
study' over" se;y«ralkJy^iiis-/ of -.heavy- duty vehicle driving conditions
in the _lafee i976ar and early 1980s. Based on that study, EPA _
21Light-duty is defined as lea* than or equal Co S500 pounds gross vehicle
weight.
"see Senate Report Ho. 101-228, 101st Congress, 1st Session 106-107
(19S9).
23In a recent rulemaking EPA addressed whether the current heavy-duty
engine test procedure reflected actual operating conditions for heavy-duty
engines used in urban buses, and determined there was no need to change the test
procedures. [58 FR 15781, March 24, 1993]
FTP Preliowufy Report: Mtjr 14, 1993
21
-------
revised the heavy-duty engine test procedures and driving cycle
in 1983.24 It was not considered either necessary or feasible to
re-examine the heavy-duty test procedures in this review process,
given the deadlines imposed by the CAAA and the fact that it
previously took over 8 years to assess actual heavy-duty driving
behavior and revise the heavy-duty engine test procedures.
M48 FR 52170, November 16, 1983.
FTP Prelixaiatiy Rqxxt: M«y 14, 1993
22
-------
Chapter 3. Project Overview
. It is a basic premise that motor vehicle emission levels
determined through the FTP should adequately reflect in-use
vehicle emissions. If in-use driving modes exist that generate
significant amounts of emissions that are not reflected on the
FTP, then the anticipated benefits from motor vehicle standards
are not being fully achieved.
It is also generally agreed that no test procedure can
reasonably duplicate all in-use conditions. The overall goal of
the Agency's review of the FTP is to aid in determining whether
or not the FTP should be modified to reflect in-use conditions
not currently found in the test and, if so, what modifications
should be made. To meet this goal it is not enough to simply
examine factors such as ambient temperature ranges, in-use fuel
characteristics, or driving patterns. For example, qualitative
evidence has existed for years that certain types of actual .
driving behavior are not represented on the FTP, such as high
acceleration rates. However, it would be counterproductive to
modify the FTP unless two conditions are met. First, the driving
behavior or other condition not represented properly by the FTP
should contribute a significant amount to motor vehicle..._..
emissions. -,_ I;EjJ.|y|p^ ,,.,'
ben6f^fr£^;.:Seit:|3|^gL^gfv'tB^Ij|ie^|tSt0tt- to t&», FOT.;sfe^^'"b(i?a»5ect:ed:^
to promote design improvements to vehicles and-thereby create" '
real improvements in controlling in-use emissions. If the
current FTP is already effective in reducing emissions during
non-FTP type driving or other conditions, then modifying the FTP
could again incur substantial costs with little or no air quality
benefit. Even if off-cycle emissions exist that are not properly
controlled by the FTP, it is critical to ensure that FTP
modifications will actually promote the proper design
FTP Preiiminiry Report M*y 14, 1993
23
-------
improvements. The Agency believes this approach is a reasonable
way to implement the requirements of Section 206(h) of the Act.
This chapter of the report outlines the various research
programs undertaken by the Agency to assess the FTP. The purpose
is to provide a context for the driving survey data and analyses
presented in Chapters 4-6. Section 3.1. discusses the reasons
behind the Agency's decision to conduct large-scale surveys of
in-use driving behavior. Section 3.2. outlines the methods and
test programs being developed by the Agency to assess the
emission impact of the driving survey information presented in
this preliminary report. Section 3.3 outlines some of the issues
and options that need to be addressed as part of the rulemaking
development process.
3.1. Research on Driving Behavior
Based upon the logic outlined in Section 2.4., EPA compared
conditions on the FTP and in-use for fuels, temperature,
altitude, and driving behavior and determined that driving
behavior was the most important area on which to focus the FTP
review. EPA has also already initiated work on dynamometer
improvements necessary to correct the inherent limitations of the
small twin-roll waterbrake dynamometer currently in use. This
work to upgrade the Agency's dynamometers is currently in
progress.
Shortly after enactment of the 1990 amendments to the CAA,
SPA evaluated existing data on actual driving behavior. Because
one of the primary concerns with the FTP is fuel enrichment, as
discussed in Section 2.4.4., quantitative assessments of
acceleration rates were a high priority. It quickly became
apparent that very little information existed that could be used
to assess the frequency and magnitude of high acceleration
FTP PreliBkuy Report May 14, 1993
24
-------
rates.23 The only existing sources of second-by-second speed
measurements were from the Operational Characterization Study
(OCS) conducted by EPA in Columbus, Ohio, in 1983, the European
Drive Project in six European cities, and work conducted by H. C.
Watson and his associates in Melbourne, Australia. The European
and Australian work, while interesting and informative, would not
be considered directly representative of how vehicles are driven
in the U.S.A.
The OCS study was more relevant, but contained some inherent
limitations. The study installed instrumentation on several
vehicles which were then loaned to study participants. The
driving patterns of 47 private citizens in Columbus, Ohio, were
monitored covering a total of 251 days of vehicle operation.
However, only three different loaner vehicles were used (a
Chevrolet Impala, a Ford Fairmont, and a Chevrolet Chevette),
making it difficult to match the participant's personal vehicle.
Also, some doubts are raised as to whether loaner vehicles would
be driven the same as the driver's personal vehicle. These
potential problems with the loaner vehicles used in the study
cast some doubt on how well the results represent overall driving
behavior. More importantly, significant problems in data
collection were encountered during the study, leading to concerns
regarding data validity. " ''.' - ~ - ':-i;l:~;' ;'
These problems with the OCS data, combined with concerns
about using 8-year old data and the ability of Columbus, Ohio, to
represent major nonattainment areas throughout the country, led
EPA to conclude that it would not be appropriate to evaluate
potential revisions to the FTP based upon the OCS results.
However, this left the Agency with choosing between three
uAa acceleration is the rate of change of vehicle speed, highly accurate
and frequent measurements of vehicle speed are necessary to properly compute
acceleration. '
FTP Prelinuouy Report: Mtj 14, 1993
25
-------
undesirable options. The first option was to proceed quickly
with a new cycle and standard, baaed upon best engineering
judgment, designed to control high acceleration enrichment.
However, without knowledge of actual acceleration behavior, this
approach had a high risk of both forcing excessive controls for
driving that rarely occurs (increasing vehicle costs without
attendant emission benefits) or of missing acceleration modes
that cause significant in-use emissions.26 In addition, this
approach would not allow investigation of other factors that
could potentially affect in-use emissions, such as driving
behavior during the vehicle warm-up period and catalyst cooldown
after the vehicle has been turned off. Because of the risk of
overlooking important test procedure modifications and of causing
excessive costs, the Agency concluded that this would not be an
appropriate route to take.
Another option was to take the position that there was no
available information upon which the Agency could justify
revisions to the FTP and, therefore, that no revisions were
necessary. The Agency quickly rejected this option because it
appeared to violate the spirit of Section 206(h}'s requirements
and of EPA's general goal to provide a healthy environment.
The third option was for EPA to conduct basic research into
actual driving behavior and conditions prior to deciding whether
to initiate nilemaking. The Agency concluded that this was the
only feasible alternative to simply ignoring any potential
modifications to the FTP and was important since the relative
impact of off-cycle emissions are likely to increase as emissions
represented by the FTP are reduced.
26Testing by the California Air Research Board on 10 vehicles over
different high acceleration modes had demonstrated that the mode generating the
highest emissions varied greatly from vehicle to vehicle.
FTP Prclrnjoiry Report; Miy 14, 1993
26
-------
Other CAA mandates, such as the Tier I emission standards
and longer useful life, should reduce the baseline emissions
derived through FTP testing. If adopted, Tier II emission
standards would have a like effect. Thus, any off-cycle
emissions will become relatively more important in the future.
The research program undertaken by the Agency is designed both to
quantify any emission impacts from off-cycle driving behavior,
and to provide information needed to determine whether or not EPA
should make regulatory changes to the FTP.
The program developed by the Agency to evaluate driving
behavior contains three basic components. First, to determine
how vehicles are actually driven, the Agency monitored an
extensive amount of actual vehicle operation. This is described
in the Chapter 4 of this report, "Driving Survey Methods and
Approach." Second, the data from the vehicle monitoring has been
analyzed tx> determine cycle and trip information and the impacts
of different factors on driving behavior. These analyses are
discussed in Chapter 6, "Discussion and Analysis of Driving
Survey Results." The Agency has also worked toward developing
new driving cycles that represent the complete range of actual
driving behavior. This part of the program is described in
Chapter 7, "Test Cycle Development Methods and Approach."
3.2. Emission Assessment of In-Use Driving
3.2.1. Cycle Development
To assess the impact of non-FTP driving, it is necessary to
estimate the difference between emissions predicted by the FTP
cycle and emissions that occur in actual driving. Using either
computer models or dynamometer testing., this assessment requires
the development of one or more driving cycles that are
representative of the real world. The driving survey data
FTP Preliminary Report: May 14, 1993
27
-------
discussed in Chapters 4 through 6 will serve as the primary input
to this component of the project.
Several approaches to cycle development have been used in
the past. These vary considerably in level of subjectivity. One
approach, used in developing the current FTP, is to splice
together segments of real speed patterns that are selected from
the survey data, A final cycle is obtained by matching summary
features of the resulting speed-time trace with those of the full
sample. A virtue of this and related approaches is their basis
in real driving experience that can be reproduced in dynamometer
testing. The choice of segments and matching-criteria are
potential difficulties.
The speed-time trace used- for the heavy-duty engine test
cycle was generated using Monte Carlo simulation. With this
method, each second-by-second value is chosen according to
statistical criteria derived from the survey data. Cycles are
subjected to matching criteria in order to screen out
unsatisfactory candidates. This is likely to be a more efficient
method of producing different cycles, but these cycles are wholly
"unreal" in comparison to the splicing approach described above
and have some potential of yielding driving segments that could
arguably be unrealistic.
Because of the concerns over the available cycle generation
methods, the Agency, in cooperation with the California Air
Resources Board, initiated a program to investigate cycle
generation alternatives and develop improved methods. The
results of this work are discussed in Chapter 6.
FTP Prdknioiiy Report: M*y 14, 1993
28
-------
3.2.2. Emission Simulation Model
In analyzing data from the in-use driving surveys it is
essential to consider the emissions impact of the real world
driving patterns that are not represented, by the current FTP
driving cycle. As discussed in Section 2.4.4., this requires
assessment of a wide range of driving behavior, factors
influencing emissions, and manufacturer calibration strategies.
In order to perform these large scale assessments, EPA is
developing a computer model which simulates vehicle emissions
over any desired driving cycle.
EPA is using the modeling approach because it affords
flexibility in analyzing the emission impact of the driving
survey data. Even if EPA had adequate time to do all the
emission assessments with vehicle testing, many analyses can be
done much more efficiently and quickly with a computer simulation
model. A simulation model will allow the emission assessment of
a number of unedited and/or composite driving cycles over a large
number of vehicles with relative ease. Thus, while the Agency
intends to use vehicle test results as the core for emission
impact assessments, the model will be used to fill in holes in
the test data, assess factors that were not: included in the test
program, and conduct quick assessments of potential changes to
the cycles and factors after the testing has been completed.
Vehicle testing will also be used during the course of the
emission assessment effort in order to validate the results of
the computer simulation model.
The simulation model computes instantaneous fuel and
emission rates based on instantaneous vehicle speed. This model
is currently being developed as two components, known as VEHSIM
(short for "VEHicle SIMulation") and VEMISS (short for "Vehicle
EMISSions").
FTP Prtlnniury Report: Miy 14, 1993
29
-------
The VEHSIM component was originally developed by GM and
later revised by the Department of Transportation. The VEHSIM
model takes instantaneous (generally second-by-second) vehicle
speed inputs and calculates instantaneous engine speed and load.
These calculations of engine speed and load are performed
utilizing vehicle information regarding vehicle aerodynamics,
drivetrain, transmission, and engine accessories stored in a
database known as a "part library".
The second component of the model, VEMISS, was developed by
EPA to provide fuel and emission' rate calculations based on the
engine speed and load inputs produced by VEHSIM. VEMISS uses a
series of lookup tables, known as engine maps, to simulate the
fuel and emission rates for a particular vehicle. An engine map
contains fuel and emission rates over a matrix, of engine speed '
and 3,oad. VEMISS implements an interpolation method with the
engine map in order to calculate fuel and emission rates for an
instantaneous engine speed and load.
Cold start emission simulations also need to be developed to
estimate the impact of cold start driving behavior and soak time.
As the cold start simulation will calculate emissions using the
warm engine-out emission maps as the starting point, development
of the cold start module is being sequenced behind the warm
component of the model. Once the warm component is"
satisfactorily validated, the cold operation component will be
developed based on existing test data, integrated into the model,
and validated.
Upon completion of VEHSIM and VEMISS, these components will
be linked to produce a fully functioning model capable of
simulating instantaneous fuel and emission rates for an entire
vehicle trip based on an inputted speed/time trace.
FIT PreUououy Report: Mxjr 14, 1993
30
-------
3.2.3. Vehicle Testing Programs
EPA is in the process of developing a joint testing program
with the California Air Resources Soard and the vehicle
manufacturers. This program will provide a base of test data
addressing the emission impacts of high speeds and acceleration
rates. The primary goals of the program are to quantify the
increase in in-use emissions caused by high speed/acceleration on
modern technology vehicles, investigate potential certification
cycles for control of high speed/acceleration emissions, and
investigate the feasibility of reducing the amount of enrichment
allowed during high speed/acceleration conditions. A secondary
goal of the test program is to investigate the emission benefits
and feasibility of insulating catalysts to decrease the rate at
which the catalyst cools off after the vehicle has been shut off.
This testing is expected to begin around May 15, 1993 and be
completed by the-end of July, 1993.
3.3. Notice of Proposed Rulemaking Development
The Agency will use the analyses described in the preceding
section when determining whether or not the driving cycle or
other aspects of the FTP should be revised to properly represent
vehicle emissions during actual driving conditions. However, in
any proposal to revise the FTP, EPA would also need to consider
various other issues,' 'including:
- Technology assessment
- Type of revision needed
- Lead time
- Cost and cost effectiveness analyses
The technology assessment includes determining the changes
needed for manufacturers to reduce emissions during the
identified off-cycle condition, the level of reduction achievable
FTP Prtlimintry Report: Mqr 14, 1993
31
-------
with different technologies and/or calibration strategies, and
the feasibility of making the technology changes. Closely
related issues are cost and lead time, as greater levels of
technology change or added component requirements will increase
the cost of the regulation and, possibly, increase the lead time
needed for manufacturers to implement the changes.
The type of revision to the FTP would also have to be
considered. For example, revisions to the existing cycle would
impact the usefulness of the vast array of historical data and
would require assessments of the impact on Corporate Average Fuel
Economy requirements and on the stringency of the emission
standards. It will likely be much more cost effective, instead,
to establish a new cycle and standard (much as was done for cold
temperature CO emissions, where a new 20 degree cycle was
established with a separate standard). If the emission benefits
prove to be relatively minor, or if it appears that emissions
could be effectively reduced without standards, it might be
desirable to simply promulgate stronger defeat device
requirements. The basic strategy to improve the FTP would have
to be evaluated, as well as the impacts on costs and emission
benefits.
Evaluation of this wide array of control strategies,
technology requirements, standard stringency, costs, and benefits
is a complex task. Due to the time constraints the Agency is
under, several potential new test cycles to control high
speed/acceleration emissions are being investigated as part of
the vehicle testing program described in Section 3.2. The level
of complexity will be significantly impacted by the results of
the study in regards to the level of off-cycle emissions and the
type of driving generating the emissions.
FTP Pttlimiauy Kcporc May 14, 1993
32
-------
Chapter 4. Driving Survey Methods
The process for selecting driving survey methods and sites
is reviewed in Section 4.1. A discussion of each method and how
it was implemented follows in Sections 4.2. and 4.3. Also
provided are results obtained at each survey site. Section 4.4.
reviews bias issues for the two survey methods. The chapter
concludes with a discussion of the principal data set which
serves as the basis for the analyses in the remainder of this
report .
4.1. Survey Options
4.1.1. Choosing Survey Approach
Prior to collecting data on in-use driving behavior, EPA had
two contractors. Radian Corporation and Sierra Research, review
previous efforts to obtain such information. Contractor reports
evaluated four alternative approaches to gathering driving
patterns data and included recommendations for this study. Table
4-1 presents the principal features of these approaches.27
In exploring survey method options, EPA sought input from
- interested partfie» aJid researchers working on the subject. To
:»-^» -«_r:_-_-S-"- "-" »i.f*-i TtV -v^-n-T': - 5-f=".T.i__V-« - " .'-*»- - - . .:..~..,- '
and Development {ORD>
GA, in June of 1991 to
- -" " - "-t->v>''?'-K.-<-"';5-^-"!:~ ;,'. " '. '. - " .' ' . . ~- *?
discuisa altusrrStives for collecting the in-use driving behavior
data, expert* from industry f government, and academia critically
evaluated the approaches presented by EPA's two contractors. As
a result of the meeting, EPA's Office of Mobile Sources (QMS)
concluded that two complementary approaches --chase car and
^Radian Corporation, "Evaluation of Driving Pattern Measurement Techniques:
Technical Note,* Draft Report to U.S. KPA, June 17, 1991, p. 4-14.
B*pxt: May 14.19»
33
-------
Table 4-1.
Alternative In-Dsa Driving Survey Methods
Instrumented Data' Logger on Driving Chase Stationary
Private Car Private Car piary Car Observer
3URCBS OP BIAS
Driver does not know he is being monitored
Representative of vehicle population
Representative of drivers
tfonintrusive installation of data logging equipment
Representative of all types of trip segments + +
TECHNICAL
Follows each car all the time - moving or not + +
Contains high frequency information + P
Measures average speeds + P
Measures instantaneous speeds + P
Measures accelerations + P
Measures number of cold starts + P
Measures during cold starts + P
Measures number of warm starts + P
Measures during warm starts + P
Measures distance travelled + P
Contains load information +
Permits various data analysis techniques +
Correct obeervables not needed at beginning of study +
Generates small quantity of data - +
Can measure drive train operation +
20S7
Low procurement cost
Low installation cost per car
High capacity storage medium not required - +
Data recovery does not require personnel at vehicle
Does not require data keypunching + +
P Measurable if pre-processor is programmed appropriately
-------
instrumented vehicles--were needed to address issues critical co
the FTP study.
The Atlanta meeting also helped produce an extraordinary
level of cooperation among the participants. The California Air
Resources Board (GARB) and EPA's Office of Mobile Sources (QMS)
joined together to develop the chase car approach. The auto
industry, represented by the American Automobile Manufacturers
Association (AAMA, formerly MVMA) and the Association of
International Automobile Manufacturers (AIAM), committed to
providing technical and financial support for the implementation
of the instrumented vehicle method. And finally, EPA's ORE
agreed to sponsor an instrumented vehicle study, in Atlanta, GA.
4.1.2. Selecting Survey Sites
After settling on survey methods, the next step was to
select the cities where the surveys would be conducted. EPA's
budget for the driving surveys necessitated limiting the study to
two sites. EPA gave consideration to several factors in
selecting the survey sites: geographical representation,
population of nonattainment area, and the type of nonattainment
area (CO or ozone) . The site selection process was complicated
by special requirements ot the two survey methods. The
x, , '. ^L, ~ ^- -'i.JT^r^T- *-,»r; - x ;'~v~ 'A:^j<4^^7,-.rr^-=u'^ire**'-':-'.-i"" , ';' - - _"""""- : -,\ -' . v>**-::^'^£;i:^T"\7~*~*"*lr'~'~' -~' '
have -
.-.kJ.^=^.-^ :~* ..- !"' -==*" r:-^l..;^fi-*»>2_,- 5-Srt: ;-' - ": -
urban^ traMptatibn^netwcik. model was a ' requisite' of the' chase
car method*. --'.. ;
SPA's criteria and the survey method requirements quickly
limited the list of candidate cities. Baltimore, MD, was chosen
to represent a Northeast, medium-sized, ozone nonattainment area.
For the second site, a suitable midwest city was desired, but
could not be found. The best alternative was Spokane, WA. This
FTT Prttamwy Report: Mff 14, 19*3
35
-------
Northwest city is fairly small, but owing to its geographical
features, it is typical of a CO nonattainment area. In addition
to these two cities, two other cities had already been selected
for related studies. CARS was sponsoring a chase car study in
Los Angeles, CA, and Atlanta, GA, was the site of an instrumented
vehicle study sponsored by EPA's 01D. Overall, EPA felt these
four cities provided a good representation of urban nonattaiment
areas.
4.2. Instrumented Vehicle Approach
This section discusses the principal features of the
instrumented vehicle approach and presents the field results.
The reader should note that a draft report prepared by the
contractor, Radian Corporation, provides a much more
comprehensive review of the instrumented vehicle study.21
4.2.1. Key Features
EPA chose the method of instrumenting privately-owned
vehicles for several principal reasons. First, this approach
collects second-by-second driving behavior data. Real-time
information is crucial for studying the full range of vehicle
operation. The instrumented vehicle method also obtains
important information on the start and finish time of trips,
driving behavior following the start of a trip, and the amount of
time vehicles are shut off between trips (soak time).
EPA had concerns with certain features inherent to an
instrumented vehicle approach. The cost and time required to
recruit and instrument privately-owned vehicles necessarily
2lRadian Corporation, "Light-duty Vehicle Driving Behavior: Private
Vehicle Instrumentation, Volume 1: Technical Report," Draft Report to O.S, SPA,
August 24, 1992.
FTP PnUmioaiy Report: Miy 14, 1993
36
-------
limited the number of vehicles to be instrumented. Equally
important, the driver's knowledge of the instrumentation and the
possible impact it could have on driving behavior was a concern.
The contractor, Radian Corporation, took these concerns into
account in the design of the actual project plan for collecting
instrumented vehicle data.
The basic features of EPA's instrumented vehicle study can
be summarized as follows. Participants were recruited from
centralized I/M stations. A monetary incentive was offered to
drivers for participating in the study. While the car owner
waited, mechanics installed a small datalogger under the hood of
the vehicle to collect vehicle speed, engine RPM, and manifold
vacuum. The participants were truthfully informed on the nature
of the instrumentation; however, the contractor did not
explicitly discuss the monitoring of vehicle speed. The
dataloggers remained on the vehicle for about one week. At the
end of the week, the driver returned to the I/M station for
removal of the datalogger.
The contractor installed two types of instrumentation
packages: a three*parameter datalogger and a 6-parameter
datalogger. The latter was sponsored by a group of domestic and
foreign auto manufacturers. In fact, the auto industry's
participation was of great value in the overall instrumented
vehicle study. Through; the Ad Hoc Panel on FTP, comprised of
AAMA and AIAM members, the industry worked together with EPA in
the development and implementation of the study. Financial
support from participating auto manufacturers significantly
increased the number of vehicles instrumented in Spokane and
Baltimore. .
FTP PnSauBMrf Report M*y 14, 1993
37
-------
3-parameter Sasg Program
Under EPA contract, Radian Corporation custom built 55
compact 3-parameter dataloggers, each capable of recording1 54
hours of second-by-second data. Initially, the contractor
manufactured 10 prototype dataloggers for testing. After testing
and making some refinements, the remaining dataloggers were
produced. The dimensions of these 45 datalogger boxes were 5" x
7" x 1.5"; the 10 prototypes were slightly larger. All of the
logger components were tested to 85 degrees Centigrade. The
compact size and durability of the dataloggers allowed mechanics
to install them under the hood of the vehicles, completely out of
the sight of the driver. The dataloggers recorded vehicle speed,
engine RPM, and manifold vacuum each second of vehicle operation.
In addition, the logger recorded the date and time on a real-time
basis.
Driver Recruitment
Centralized I/M stations were selected an recruitment sites
in order to obtain the moat representative sampling of drivers
and vehicles possible. In Spokane, all gasoline-powered light-
duty cars and trucks are required to be tested at the sole I/M
station. Baltimore requires testing of all vehicles less than
20 years old. There are a number on I/M stations in and around
Baltimore. EPA selected two stations for recruitment: a
suburban site (Rossville station) and an urban site (Exeter
station).
EPA's contractor solicited the driver after the vehicle
passed emission testing. A monetary incentive of up to $100 was
offered to drivers to compensate for their time and
inconvenience. If the driver passed an initial screening and
agreed to participate, the contractor mechanics installed the
FTP Pretiamu? Report: M»y 14, 1993
38
-------
logger at the I/M station-. Typically, the installation of the
datalogger took between 30 minutes and an hour.
. In order to try and minimize sampling bias due to refusals,
the recruitment procedure included a driver\vehicle replacement
strategy. The solicitor recorded information on vehicle age,
vehicle type, vehicle's country of origin (foreign or domestic),
and driver age for all solicited drivers. If a driver refused to
participate, a replacement driver\vehicle was found which matched
the four recorded characteristics.
Supplemental 6-parameter Program
The program of installing 6-parameter dataloggers on
manufacturer-sponsored vehicles solicited from I/M stations was
conducted in parallel with the 3-parameter study. The purpose of
this supplemental program was to collect additional vehicle
operating information. Like the 3-parameter datalogger, speed,
RPM, manifold vacuum were recorded, but, in addition, the 6-
parameter dataloggers obtained coolant temperature, throttle
position, and air/fuel ratio collected downstream of the
catalyst. Each participating auto manufacturer developed
customized interfaces for the 6-parameter dataloggers.
Typically, the loggers were installed in the vehicle's trunk.
Two major differences existed between the 3-parameter and 6-
parameter program. Only late-model vehicles were eligible for
the 6~-parameter program and candidates were further limited to
high-volume models among the participating manufacturers.
4.2.2. Instrumented Vehicle Field Results
EPA's instrumented vehicle study was the largest study of
real-time, in-use driving patterns ever conducted in the U.S.
Nearly 300 vehicles were instrumented in Baltimore and Spokane.
FTP PKlimsauy Report; Miy 14, 1993
39
-------
The dataloggers recorded over 6 million seconds of driving
behavior in the two cities,
IPA's contractor carried out a one week pilot study in
Spokane in January, 1992. The pilot study tested alternative
recruitment strategies arid logger installations in the field.
The success of the pilot study permitted the start of the full
Spokane Instrumented Vehicle Study on February 3. Datalogger
retrieval in Spokane was completed by the firsit week in March.
At that time, the contractor began work in Baltimore. The
contractor finished all field work by the first week of April.
In the two cities, a total of 730 drivers were solicited for
participation in 'the program. 331 of the solicited drivers
agreed to participate, and from that group, 294 vehicles were
successfully instrumented. Table 4-2 gives a summary of the
field study results.
FIT Prt licoiury Report-- Mty 14, 1993
40
-------
Table 4-2.
Summary of Instrumented Vehicle Field Results
City
Both
cieies
Spokane
Baltimore
Datalogger
Type
3 and 6
3
6
3 and 6
3
6
3 and £
3
6
Vehicles
Solicited
727
479
243
246
161
as
481
318
163 '
Screen
Passes
571
374
197
222
144
78
349
. 230
119
Drivers
Participating
331
226
105
168
111
57
163
115
48
Participa-
tion
Rate {%}
58
60
S3
76
77
73
47
50
40
Complete
Instal -
lations
294
215
79
144
102
42
ISO
113
37
The driver participation rate was substantially higher in Spokane
than in Baltimore. About three-quarters of the solicited Spokane
who passed the screening agreed to participate in the driving
patterns study. In contrast, the participation rate in Baltimore
was only 47 percent. .
After completing the field work, the contractor began the,,
processing of this immense data set. All data were passed
through a rigorous quality control software program to check for
errors in the data. This procedure identified 216 error-free
vehicle data sets and 78 "suspect* vehicle data sets. A
subsequent review of the "suspect" vehicles resulted in the
recovery of a number of vehicles. The final count of vehicles
for which there was good data was 217, including 168 3-parameter
and 59 6-parameter vehicles.
FTP Prttmmury Report: M»y 14, 1993
41
-------
4.2.3. Atlanta Instrumented Vehicle Study
Atlanta, GA, is the site of an intensive emission inventory
study by EPA's ORE. The objective of the instrumented vehicle
study in Atlanta was to provide important vehicle operation data
for ORD's mobile source emission inventory analysis, while also
serving to supplement OMS's instrumented vehicle database.
Under a cooperative agreement with ORD, the Georgia
Institute of Technology conducted the instrumented vehicle study
in the summer of 1992. The 3-parameter dataloggers were
installed on a fleet of privately-"owned vehicles. The Atlanta
study followed the procedures used in Spokane and Baltimore,
whenever possible. The most significant difference in Atlanta
was the method for recruiting drivers. Atlanta's use of a
decentralized I/M program made it impossible to use I/M stations
as a driver recruitment location. However, Atlanta does have a
cenj;ga.lizecl driver's license renewal system which requires
drivers to renew their licenses in person at one of a limited
number of Driver's License Stations. The recruitment method
utilized three such stations to recruit a representative cross-
section of Atlanta drivers.
Georgia Institute of Technology conducted a successful pilot
study in the first week in July of 1992. The full study was
carried out from July through the beginning of October; mechanics
instrumented a. total of 101 vehicles. The successful completion
of the study served to add a third city to the instrumented
vehicle database.
EPA contracted with Radian Corporation to process the
Atlanta data set using the exact procedures employed in the
processing of Baltimore and Spokane data. Due to contract
delays, only initial processing of the data is complete.
FTP fn&siotif Rqsoo: Mty 14, 1993
42
-------
Preliminary data are available for 68 vehicles from Atlanta.
This limits the usefulness of the Atlanta data for this
preliminary report; however, it is expected that further-analyses
will use data from a finalized Atlanta data set.
4.3. Chase Car Approach
The use of chase cars to collect driving patterns data has a
fairly long history. The EPA Highway fuel economy driving cycle
was based on driving patterns data collected by driving a chase
car over 1000 miles of non-urban roadways.29 For the FTP study,
EPA felt the chase car approach could complement the instrumented
vehicle approach in several areas. First, a chase car allows for
the collection of driving behavior data in a non-intrusive
manner. The chase car method also provides a cost-effective way
to collect data for a large sample of vehicles. Finally,
information on the driving environment--road type, congestion
level--can be obtained using chase cars.
The traditional qhase car approach has a major drawback. It
is assumed that the driver of,the chase car can simulate typical
driving patterns, either by following specific "target" vehicles
or by "flowing" with traffic. However, EPA's objective was to
not only capture "typical" driving patterns, but also to study
.the entire spectrum of drivers and driving patterns, including
low frequency behavior found in the "tails" of the driving
behavior distribution. Clearly a single chase car driver could
not be expected to adequately achieve such a goal.
^Austin, Thomas C., Karl H. Hellman, and C. Don Paulsell, "Passenger Car
Fuel Economy During Non-Urban Driving," SAB Paper 740592, August, 1974.
FT? Preliminary Report; Mtjr 14, 1993
43
-------
4.3.1. Key Features
Under contract with both CARB and EPA, Sierra Research
developed an enhanced chase car equipped with a grill-mounted,
laser rangefinder which can infer the speed of the "target"
vehicle.30 The laser rangefinder utilizes the same technology
found in hand held laser guns currently used by many police
departments to measure vehicle speed. The laser measures the
distance between the "target" vehicle and the chase car (patrol
car). The patrol car itself is instrumented to collect the
vehicle speed; knowing the patrol car's speed and the change in
the distance between the patrol car and target: car, the target's
car's speed can be calculated. The laser measures the distance
20 times a second and calculates a one second 'average. The
laser-enhanced chase car method makes it possible to get an
accurate representation of the driving behavior of the target
vehicle.
The use of the a laser rangefinder is the most prominent
feature of Sierra Research's chase car approach. Other important
characteristics of this approach are* in the following discussion.
T.najt rumentation
The patrol car was a 1991 Chevrolet Caprice with sufficient
room behind the grill to install the custom-built laser
rangefinder. The car was equipped with a videotape recorder
mounted between the front bucket seats. In ciddition, the patrol
car speed and manifold pressure were collected once per second.
A road grade measurement system was developed using two
30Siarra Research Inc., "Design and Operation o£ an Instrumented "Chase
Car" for Characterizing the Driving Patterns of Light-Duty Vehicles in Customer
Service," Draft Report to U.S. EPA, February 28, 1992.
FT? Prttimnary Report: M«y 14, 1993
44
-------
unidirectional accelerometers mounted on the vehicle floor pan to
collect longitudinal and lateral acceleration. All of the data
were directed to an on-board' laptop computer operated by the
"co-pilot." The co-pilot was responsible for working a handheld
switch box for recording road type, traffic level, and target
vehicle information.
Route Selection
It was imperative that the routes over which the patrol car
would travel and collect 'information reflect road conditions,
trip conditions, and vehicle/driver variation which are
characteristic of the overall driving patterns for the particular
city being studied. The chase car study employed transportation
planning models for the generation of representative routes.
These computer-based transportation planning models are used by
urban areas to track travel activity over the road network.
Information from periodic travel surveys are a principal input
into these models, providing trip information such as trip
origin, destination, purpose, and length. Additional information
on land-use patterns are also incorporated into the model.
Representative routes are generated using a four-stage process of
trip generation, trip distribution, model choice, and route
choice. The contractor worked with local transportation
authorities to generate 300 routes for each city.31
' Target Selection . ' '
A detailed road map was generated for each route. At a pre-
determined time and start location, the patrol car began the
route and sought its first target vehicle. Target vehicles were
selected at random from candidate vehicles (cars and light
FT? Prelimaujy Report: NUy 14, 1993
45
-------
trucks) traveling on the same route as the chase car. If the
target vehicle left the pre-defined route, the chase acquired a
new target vehicle according to prescribed protocol.
Trip ends
Inherent to the chase car approach is an inability to
measure vehicle activity at very beginning and very end of a
trip. Recognizing this, Sierra Research conducted trip ends
surveys in Sacramento and Los Angeles.32 Information on idle
time and driving time and behavior prior to accessing the road
network were recorded by observers in the field. The
information can be used in conjunction with the chase car data to
give a more complete profile of driving behavior.
4.3.2. Chase Car Field Results
Sierra Research conducted the first chase car study in
Baltimore during November and December of 1991. After completion
of the data collection effort in Baltimore, the contractor
processed and quality checked the data. Chase car survey work
continued in the spring of 1992 with the Los Angeles study
sponsored by CARS. The Spokane chase car study was carried out
in July of 1992.
Table 4-3 presents a summary of the field results for
Spokane and Baltimore. The chase car study collected data on
over two hundred routes in each city. A large number of targets
were acquired in each city, although the fraction of total route
time with a target was disappointingly low. The biaa analysis,
in Section 4.4, discusses the significance of these results.
32Sierra Research Inc., "Characterization of Driving Patterns and Emissions
from Light-Duty Vehicles in California," Draft Final Report to California Air
Resources Board, March 5, 1993.
FIT Prtlmaury Report; Mix I*. 1993
46
-------
Table 4-3.
Summary of Chase Car Field Results
Characteristic
Number of Routes
Total driving
time (seconds)
Number of targets
Total target time
(seconds)
Target time as
percent of total
Both Cities
467
366,256
1,641
143,668
39
Baltimore
218
191,119
770
69,528
36
Spokane
249
175,137
871
74,140
42
4.3.3.
Los Angeles Chase Car Study
CARB's Los Angeles chase car study ran concurrent with EPA's
studies in Baltimore and Spokane. The first chase car work in
Los Angeles was in conducted in the fall of 1991. Due to data
acquisition problems and revisions to the route selection
methodology, these data were not acceptable to GARB. The
contractor returned to Los Angeles to conduct additional chase
car runs in April and May of 1992. The chase car collected data
f gr_t02 -rqigtes , encompassing 28 hours of vehicle operation. The
L°s Angeles
P8e : en^lbyeS ijrIBaltimore and
a 1 imited analysia of the Los
.Angeles
4.4. Analy««» of Potential Bias
The objective of both driving survey approaches was to
obtain the most representative characterization of in-use driving
behavior. While the design features of the two survey methods
attempted to explicitly account for all areas of foreseeable
FIT PrabaiMfy Report: Kby 14, 1991
47
-------
biases, it is the surveys' results which ultimately determine the
success of the two approaches. This section provides a largely
qualitative assessment of the "representativeness" of the chase
car and instrumented vehicle data for Baltimore and Spokane.
4.4.1. Analysis of Instrumented Vehicle Method
EPA's contractor, Radian Corporation, conducted a post-
survey study of several issues pertaining to possible survey
bias.33 Sampling bias sources included improper representation
of vehicle age, make and manufacturer, and performance type.
Nonsampling sources included accuracy of speed measurement and
the effect on driving behavior ^of the presence of the datalogger.
Vehicle Selection
The sampling frame in the 3-parameter instrumented vehicle
surveys was the set of vehicles scheduled for inspection during
the month in which the survey waa conducted. Radian obtained
lists from Spokane and Baltimore of all vehicles tested during
the survey periods. For both cities these vehicles were
classified by model year (age) and make and compared to the
corresponding distributions found in the sample. Goodness-of-fit
statistics showed no significant difference except in the case of
the Spokane vehicle make variable, in which Mazdas were somewhat
overrepresented.
Vehicle performance type is a rather subjective measure, and
EPA was unable to obtain sufficient information on performance
criteria for all vehicles tested in the two cities during the
survey periods. Therefore, the representativeness of the sample
33Radian Corporation, "Private Vehicle Instrumentation Survey: Data Bias
Analysis Technical Note," Draft Report to U.S. SPA, September 30, 1992.
FIT Frsfinuntjy Report; Mix 14, 1993
48
-------
was judged by comparing the final sample vehicles co the set of
all vehicles chat were randomly solicited, which are assumed to
be representative of the target population.
Using power -to -weight values and some subjective judgment
about the "image" of the vehicle, vehicles were classified as
"High -Performance" or otherwise. In Spokane, 3 of 99 (3.0%)
randomly solicited vehicles were classified as High- Performance?
in the final sample, 1 of 75 (1.3%) were High- Performance. For'
Baltimore, there were 6 of 86-14434 (4.2-7.0%) High- Performance
random solicitations, and 4 of 93 (4.3%) High -Performance
vehicles in the final sample. These numbers suggest the
possibility that High -Performance vehicles were underrepresented
in the final sample; EPA has not yet developed a reasonable
adjustment method for this factor. This may be an area for
further study prior to completion of a NPRM.
Speed Measurement
Accurate vehicle speed data were critical to the success of
the instrumented vehicle study. The 3 -parameter dataloggers used
three different methods for obtaining vehicle speed: magnets on
a drive shaft, OEM speed sensor, or an aftermarket cruise control
attached - to_the^ spee^onsater cabler; ; The acceptable^ accuracy for
' ^:\^i^f^'^^&^:^i~'-^^'^ndy' was.
contractor to teat the7 accuracy of the three speed measurement
methods. The OEM sensor and drive shaft magnet methods showed
""The selection method was not recorded for 58 vehicle solicitations in
Baltimore.
3iRadian Corporation, "Light-Duty Vehicle Driving Beh*vior: Private
Vehicle Instrumentation, Volume li Technical Report,* Draft Report to U.S. EPA,
August 24, 1992.
FT? Pretaauy Report Miy 14, 1993
49
-------
very good correlation with the fifth wheel at all tested speeds
(1 to 50 mph). The speedometer "cable approach performed well at
speeds greater than 10 mph; however, at low speeds the
speedometer cable approach proved leas accurate due to jerky
rotation of the cable.36 In the field, every attempt was made
to keep the speedometer cable straight to minimize the potential
for such erratic low-speed data. Upon quality checking of -the
data, EPA does not consider this to be a significant problem.
The vehicle speed resolution for 6-parameter vehicles varied
among manufacturers. The speed data were obtained directly from
the engine's computer (ECU), and were therefore limited by the
resolution required by the vehicle's production specifications.
Typically, a fine speed resolution is not necessary for a
vehicle's dashboard speedometer. Thus, many of the 6-parameter
vehicles only recorded speeds to the nearest 1 mph or less. As a
consequence, the lower precision of the 6-parcimeter vehicles'
speed is a potential problem, particularly at low speeds. More
importantly, this problem results in low resolution (that is, l
mph/sec) and an upward bias in the calculated acceleration
rates.37
Datalogger Presence
A driver may alter his or her vehicle operation in response
to the datalogger presence. Reasons for this change in behavior
might include concern that the data could be used as evidence of
illegal or unsafe driving. It is most likely that bias
"ibid.
37Systems Applications International, "Stratified Comparison and Rounding
Effect Analysis of Driving Operation Patterns and Event: Characteristics Between
Three-Parameter and Six-Parameter Instrumented Vehicle Data," Draft Report to
American Automobile Manufacturers Association, January *?, 1993.
FTP Prtliruairy Report: Miy 14, 1993
50
-------
introduced in this way would tend in the direction of more
conservative driving, that is lower speeds or lower rates of
acceleration.
EPA and Radian conjectured that, if it exists, this type of
bias might be most evident during the initial phase of
instrumentation and would decline as the driver adjusted to the
presence of the datalogger. This motivated an examination of
possible bias based on the "Observation Phase* in which speed and
related values were observed.
For each vehicle in the Baltimore 3-parameter sample, the
speed observations were classified as belonging to the first day
of instrumentation (Observation Phase 1} or later (Observation
Phase 2). Various driving behavior measures were computed for
the two phases. (Section 5.1 contains a discussion of speed-
related driving variables analyzed in this study.) It was found
that during Observation Phase 1, drivers operated at lower
average speeds than in Phase 2, which appears to support the
claim that the datalogger influenced driving behavior in the
initial phase of instrumentation. However, further examination
showed that the time of week also affects average speed:
substantially higher speeds are observed in weekend driving than
on weekdays. (See* Section 6.2 for details.) Moreover, due to most
vehicles being instrumented, on weekdays, the proportion of
Observation Phase 1 driving on weekends (17.9%) was lower than
for Phase 2 (25.6%). If the average speed during Phase 1 is
adjusted for this difference, it actually is higher than the same
value in Phase 2. Thus, it was decided that driving behavior
probably did not change substantially over the instrumentation
period and that data for both observation phases should be
included in subsequent analyses. Of course, this analysis leaves
unanswered the question of whether instrumented vehicle driving
was more conservative than normal over the full survey period.
Report Mqr 14, 1993
51
-------
4.4.2. Analysis of Chase Car Method
In terras of potential bias, the three areas of concern for
the chase car study are the representativeness of the routes
selected, vehicles, speed measurement, and driving behavior.
Each of these issues were reviewed following the completion of
data collection.
RouteSelection
The method for selecting routes was designed to match the
overall distribution of trips by travel period or trip type, zone
type, and trip length. This method assumes that the
transportation model produces an accurate portrayal of in-use
trip behavior. Accepting this, the subset of the selected routes
which were actually run, closely matches the desired distribution
(For a qualitative comparison see Appendix A, tables A-i and A-
2) . However, the assumption that the transportation model
produces an accurate portrayal of in-use trip behavior may not be
appropriate, as discussed in Section 4.5.1.
Vehicle Select ion
As discussed in Section 4.3.1, the protocol for selecting
target vehicles was designed to obtain a random sampling of
vehicles. In the field thia method proved practical; however,
obtaining and maintaining a lock 'on the target vehicles was only
a limited success. The overall proportion of total route time
with a target locked on was 36 percent in Baltimore and 42
percent in Spokane. The general availability of candidate
vehicles was a significant factor in the low percentage. In
fact, 10 percent of the routes in Baltimore had no target vehicle
data. The corresponding percentage for Spokane waa 13.
FTP Prclinmiiy Report: Miy 14, 1993
52
-------
Maintaining a constant laser lock-on for a given target
vehicle also contributed to the lack of target data. A break in
the laser-lock of ten"" occurred due to change in grade, on turns,
or lane changes. In order to investigate the lock-on problem,
EPA contracted Sierra Research to analyze the video tapes for
Baltimore and Spokane to identify reasons for losing targets."
Table 4-4 presents the results of this analysis.
A second issue related to vehicle selection is the
representativeness of the target vehicle sample relative to the
vehicle population. The video tape review also helps address
this topic. The contractor tabulated the distribution of target
vehicles by vehicle type, model year, and manufacturer, for
Baltimore and Spokane (For a summary of the results see Appendix
A, Table A-3).
"sierra Research Inc., "Chase Car Video Tape Review,"-Draft technical
memoa to O.S. SPA, October 9, 1932 and October 29, 1992.
FT? Prelimiaiiy Report: Miy 14, 1993
53
-------
Table 4-4.
Reasons for Chase Car Losing Target
Reason for losing target
Target turned off route
Patrol changed roadways to continue
route
End of route
Lost target over hill or around
corner
Target and patrol car out of
alignment while in same lane
Changed lanes {patrol car)
Lost target while merging with heavy
traffic
Patrol stuck at traffic light
Another vehicle came between
target and patrol car
Lost target because of aggressive
driving
Suspect target driver realized
they were being followed
Target changed lanes
Turned off switch to read map
Navigator bumped switch
Laser error, sunlight interference
Other reason
Baltimore
Number
250
249
14
684
216
27
15
19
36
14
7
186
21
23
95
81
Percent
12. S
12.8
0.7
35.2
11.1
1.4
.,0-8
1.0
1.9
0.7
0.4
9.6
1.1
1.4
4.9
4.2
Spokane
Number
331
211
24
653
131
98
5
23
23
0
4
235
22
19'
0
19
Percent
18,4
11.7
' 1.3
36.3
7.3
5.5
0.3
1.3
1.3
0.0
0.2
13.1
1.2
1.1
0.0
1.1
Aggressive-
From the start, EPA was aware of, "and concerned about, the
ability of the chase car approach to capture aggressive driving
behavior. There was some small fraction of vehicles which were
FTP frtlmmtiy Report: May 14, 1993
54
-------
operated in such a fashion that it was either impossible or
unsafe for the patrol car to "chase" them. It is not possible to
quantify the incidence of these occurrences. A partial answer
can be found, however, from the review of the target vehicle
videotape. As shown in Table 4-4, less than 1 percent of the
cases in which the target was lost was attributable to aggressive
driving behavior of the target {the classification was based on
the videotape reviewer's best judgment). In addition, some
fraction of the reason "lost due to lane change" {10 percent in
Baltimore and 13 percent in Spokane), can be attributed to
aggressive driving on the part of the target vehicle.
Speed Measurement
The use of the laser was designed to produce an accurate
measurement of the target vehicle's speed on a second-by-second
basis. The preliminary testing of the laser suggested that the
laser provided an accurate measurement of target speed. However,
in the process of comparing the chase car data and instrumented
vehicle data, EPA identified a problem with the target vehicle
speed resolution which resulted in skewed acceleration data under
certain situations. Post-processing of the target speed data
using a smoothing algorithm greatly improved the "fit" of- the
laser-based target data.39
4.5. Selection of Principal Data Set
4.5.1. Comparison of the Two Surveys Results
EPA's contractors conducted the chase car and instrumented
vehicles studies in Baltimore and Spokane within a nine month
3*Sierra Research Inc. "Characterization of Driving Patterns and Emissions
from Light-Duty Vehicles in California," Draft Final Report for California Air
Resources Board, March 5, 1993.
FTP freliaaauy Report: M«y 14, 1993
55
-------
period in 1991 and 1992. A variety of constraints prevented
simultaneous data collection in each city. Except for potential
seasonal differences, it is valid to compare the results for the
chase car and instrumented vehicle approach. Such a comparison
cannot validate or invalidate an approach, since the absolute
truth is.not known. Rather, the comparison serves-to identify
substantial differences and allows for an evaluation of these
differences.
In looking at the-results from the two fundamentally
different survey approaches, it was reassuring to see that they
paint like portraits of driving behavior, especially in terms of
differences between Baltimore and Spokane. Both methods indicate
higher average speed in Baltimore than Spokane (see Table 4-5) .
This is also true for average specific power.40 The surveys also
agree in the direction of the differences of trip time and
distances; Baltimore trips averaged about 0.5 miles longer than
Spokane trips and the duration of trips in Baltimore were about
three minutes longer than in Spokane.
While the two surveys agree in terms of differences between
the cities, the actual estimates of various driving behavior
measures are clgarly different. The chase car method estimates
much higher average speeds than the instrumented vehicle
approach. This difference is, however, consistent with
expectations, given the methodological differences. The chase
car's lack of trip end information (see Section 4.3.1) is likely
to be the primary reason for the higher average speed.
^Specific power is a single measure for the combined effect of speed and
acceleration. See Section 5.1.1 for more detailed discussion of this measure.
Ft? Frclaaiouy Reject M»)f 14, 1993
56
-------
, Table 4-5.
Comparison of Chase Car and Instrumented Vehicle Driving Behavior
Driving
Behavior
Measures
Sp*«d (mph)
Average
Maximum
Standard
deviacion
Number of
observations
Acceleration
(mph/s«c.)
Minimum
Maximum
Standard
deviation
Power
(mphV**c)
Average"
Maximum
Averagartrip
length (alias)
Average trip. -
tiBM (minutes)
Baltimore
Chase Car
30.70
79.50
21.00
191,1X9
-2Q.SO
8.13
1.39
22.2
577.30
T.4S
14.61
Instrumented
Vehicle
24.50
94.46
20.52
3,365,504
-19. 4i
15.19
1.50
19.23
557.63
4.89
12.03
Spokane
Chase Car
29.80
83.20
19.50
175,137
-11.30
7.79
1.42
20.0
403.00
5.83
11.72
Instrumented
Vehicle
23,24
77.55
17.71
2,081,199
-15.46
15.95
1.46
17.01
672.28
3.56
9.18
^ Thii
diffoti ftora *v*r*|« pow«r V*!UM cjtod in Chapur 6, which UM tottl'Mcoadi of powtiv* *ec«l«ntioa *« th« dwwmiiuior. Th«
ilimiMiv* d*finhio(iuM4 bM» WM twewmiy in orctot to h«v« coa*i«M* WMiga pow«e VB!UM for th* cfau* car tad ia*uunMnt«i
vehid«d«li.
The differences in the surveys' estimates of trip time and
distance suggest two distinct perspectives on trip-making
activity. Using engine on/off to define a trip (see Section
5.1.2), the instrumented vehicle approach identified a large
number of short trips. In contrast, the chase car routes are
developed using the transportation network model's concept of a
Report: M«y 14, 1993
57
-------
trip which, is based on crip purpose. A given chase car route may
be a result of the implicit connection of two or more engine
on/off events. As an example, a single home-to-work chase car
trip could be viewed as 2 separate trips in the instrumented
vehicle approach--home-to-day care and day care-to-work. The
instrumented vehicle's trip patterns perspective, which captures
the full-range of trip behavior, is most consistent with the
objectives of the FTP study.
4.5.2. Rationale for Choosing 3-parameter Instrumented
Vehicle data
For this preliminary technical report, EPA feels the 3-
parameter instrumented vehicle data are the best representation
of in-use driving behavior currently available. This data set
provides detailed driving behavior for a large cross-section of
vehicles and drivers. All substantive issues on bias or data
integrity have been satisfactorily addressed for the 3-parameter
data set.
The 6-parameter instrumented vehicle data are a valuable
addition to the overall data set; however, two factors limit
their usefulness for this preliminary report. As discussed in
Section 4.2, speed measurement problems impact the accuracy of
the acceleration-based measures for the 6-parameter data set. In
addition, the vehicle selection process for the 6-parameter
vehicles did not attempt to obtain a representative sample;
weighting factors will need to be developed in working with the
data in the future.
In choosing the 3-parameter instrumented vehicle data over
the chase car data, one factor was just the sheer volume of data
obtained by the instrumented vehicle approach. In Baltimore and
Spokane, 3-parameter instrumented vehicles collected over 1500
hours of driving behavior data compared to 100 hours of chase car
FTP fniauaay Report Miy 14,1993
58
-------
data. EPA also considered the surveys' ability to provide data
on a representative cross-section of vehicles and driving. While
we are largely satisfied with the results of the instrumented
vehicle bias analysis, the results of the chase car surveys are
less reassuring. The majority of data from the chase car is for
only one vehicle--the patrol vehicle. The target vehicle data
does capture a large number of vehicles, but typically, the data
are obtained for a very small amount of time. Finally, the
routes (or trips) driven in the chase car approach are
considerably longer than typical trips from the instrumented
vehicle data, casting doubt on whether the chase car routes
properly reflect the range of in-use driving conditions.
FTP Prclimmuy Report: M«JF 14, 1993
59
-------
Chapter 5. Analytical Methods and Considerations
This chapter describes driving behavior variables and
methods for their analysis based on the survey data discussed in
Chapter 4. Analytical methods consist primarily of traditional
descriptive techniques: arithmetic summary measures, variable
frequency distributions, and supporting graphs. Formal
statistical inference is omitted, in part due to large sample
features of the survey. Nevertheless, choices and emphases in
describing data can easily color the reader's impression of what
those data say about the larger population of drivers and their
vehicle operation. The final section reviews a study of engine
and catalyst temperature cooldown undertaken as part of the
project.
5.1. Driving Behavior Variables
Despite longstanding recognition of the importance of mobile
sources to air quality, surprisingly little research has been
done on driving behavior. This report gives a preliminary, broad
description of the current survey results while concentrating on
those issues that pertain most to vehicle emissions and test
drive cycle policy. The data collected for this report offer a
rich source for future study of driving behavior. EPA
anticipates that researchers in different fields, with other
perspectives, skills, and computing resources, will contribute
new understanding of the information contained in these data.
5.1.1. Speed-Based Measures of Driving Behavior
In planning the survey, EPA attempted to collect basic
measures known or believed to explain vehicle emissions. As
noted earlier, the most fundamental variable, vehicle speed, was
recorded for each second that a vehicle's engine was running.
FT? Prtlininuy Report; Mjiy 14, 1993
60
-------
These "time series" data have been used to derive two other key
measures: acceleration and specific power.
Acceleration
Acceleration, the change in speed per unit time, is computed
as the simple difference in two successive speed values. This
calculation results in a missing value for acceleration at the
start of each (engine on) trip. By its very nature, the
distribution of acceleration is centered at zero and tends to be
symmetric and unimodal.. Thus, the mean and median are not useful
for comparing two or more distributions. Average measures of
dispersion, such as the standard deviation, help describe how a
set of accelerations vary. Percentiles are useful for
summarizing extreme levels of acceleration and deceleration.
The problem with analyzing the acceleration rate is that, by
itself, it has dubious significance. For a given acceleration
rate, the load placed on the engine increases with the vehicle
speed. Thus, the most appropriate measure is the joint
distribution of speed and acceleration. However, this requires
three dimensions to display, making it harder to visualize. As a
-useful coppji'OBiise.j - thft. joint "distribution of speed and
to
-^Ir^i-iow"^^ j?le.*SI^*|Pf below.
- Sec i'f lev- 'Power '*"
Specific power is defined as the per second change in the
square of vehicle velocity during positive accelerations:
specific power - Vf2 - v,af Vr > Vs
FIT PtthouB«ry Report; Ml? 14, 1993
61
-------
where Vf and Vf are the initial and final velocity, respectively,
in a one-second interval. Interest in this variable is motivated
by the work of Watson, et. al.,41 who introduce a measure called
"positive acceleration kinetic energy change per unit distance,"
or PKE. In their work, it is presumed that fuel economy and
emissions are proportional to PKE when measured jointly over a
given trip. Thus, PKE is acumulative measure of increases in
kinetic energy over a fixed travel distance.
As used in this study, specific power is determined for each
second of driving, unadjusted for-distance traveled. It is
closely related to per second speed and acceleration by simple
multiplication:
specific power 2 x speed x acceleration
(for acceleration greater thaja zero) . Preliminary work by EPA
suggests that, for explaining second-by-second emissions,
specific power performs better than either speed or acceleration
alone (although the joint distribution of speed and acceleration
is best). In conclusion, specific power appears to give a useful
univariate composite measure of the more intuitive speed and
acceleration variables.
5.1.2. Trip-based Measures
Trip-based analyses provide an alternative, summarized
perspective of driving behavior. The concept of what constitutes
a trip appears intuitively simple; however, a number of practical
issues require some discussion. This section reviews features of
4lWatson, H.C., E.B.Milkina, M.O. Preston, C. Chittleborough and B.
Alimoradian, "Predicting Fuel Consumption and Emiasions-Transferring Chassis
Dynamometer Results to Real Driving Conditions," SAI Technical Paper Series,
830435, 1983.
FTP Pntkniouy Rcjxxt; Miy 14, 1993
62
-------
the instrumented vehicle trip definition and describes the main
concepts used in the trip-based analyses.
Trip Definition
The instrumented vehicle survey used engine on- engine off to
define the beginning and end. of a trip. In reviewing the initial
trip data it became evident that this definition resulted in
including vehicle, stalls as separate trips. To exclude stalls,
EPA directed the contractor to make a slight modification to the
original definition. Using the time between trips, or soak time,
apparent stalls were connected to the appropriate contiguous
trip. As an example, a vehicle sits in the driveway all night
and in the morning the car is started, but stalls after 5 seconds
because it is cold. It then takes 15 seconds to start the car
again and the driver leaves on a 10 minute trip. Using just
engine on -engine off as the trip definition criteria would result
in having the stall treated as a separate, 5 second trip.
However, if one uses the criteria that if the time between trips
is less than a certain number of seconds, then the "stall trip"
would be combined with the following "real* trip. Using IS
seconds as the soak time threshold, the 5 second "stall "trip"
:^
~'~''
---.--«.- >.
The' formal :.trip,. definition is:. ' .
if Soak, <, 18 then,
New trip « Trip^ + Soak, + Tripj
ttpott; M«y 14, 1993
63
-------
The selection of is seconds as a soak time threshold is based on
the distribution of in-use soak times and it is consistent with
similar analyses contracted by AAMA and AIAM.
An additional modification to the trip data was made to
insure a representative sample of trips. As discussed in Section
4.2, the vehicles in the survey were recruited and instrumented
at centralized I/M stations. After installation and testing of
the datalogger, the datalogger was placed in a "run" to begin
collecting in-use driving data. However, on many occasions the
vehicle was driven by the mechanic to the front of the I/M
station and parked. At such time the vehicle was shut off and
the keys were turned over to the customer. This trip was
recorded as the first trip for the vehicle. A similar scenario
could occur when the vehicle was returned. EPA felt such trips
were not representative of in-use operation and should be
deleted. To accomplish this task, I/M station trips at the
beginning and end of a vehicle's trip file were identified and
deleted using a distance threshold of 0.2 mile (1,056 feet).
This resulted in an elimination of 55 trips in Spokane and 67
trips in Baltimore.
Vehicle Operating Mode Definitions
A common measure in trip-based analyses is the proportion of
time spent in the four vehicle operating modes: cruise,
acceleration, deceleration, and idle. In defining the modes, EPA
started with a definition for cruise. Driving at a constant
speed is what is typically thought of as a cruise. However, a
vehicle's speed is rarely constant. For this report, the cruise
mode is defined as a driving period at least three seconds long
during which all acceleration and deceleration are of an average
magnitude of 0.5 mph per second or less over two second
intervals, with speeds measured at one second intervals. The
FTP FreliBunuy Report: Mix 1*. WS
64
-------
definitions for acceleration, deceleration, and idle modes fall
out of this cruise definition. Computationally, the four modes
are defined as follows:
l. Find the average 2 -second acceleration:
A,- (St+I - SM)/2
where, St- vehicle speed at time t
2. Cruise: If A,.,, A,, and A^, are less than or equal to
0.5 mph/sec. in absolute value, and St is
greater than 1, then values for time t-l, t,
and t+1 are classified as Cruise.
Acceleration: If A, > 0.5 and St > l, then values for time t
is an acceleration.
Deceleration: If A, < -0.5 and S, > 1, then values for time
t is an deceleration.
Idle: If the value for time t does not meet the
conditions for cruise, acceleration, or
deceleration it is classified as idle. This
definition of idle includes some non-zero
speed driving and, thus, differs slightly
from the idle definition used in the
speed/acceleration sections of this report.
Sto Mod,e
The number of stops per trip or the distance between stops
is another commonly used trip measure. A stop at a traffic: light
or stop sign typically involves the vehicle' speed going to zero,
but this is not always true, as in the case of "rolling stops."
For this report, the stop mode is defined as beginning when the
vehicle slows from above 10 mph to below 4 mph. The stop
concludes when the vehicle's speed exceeds 4 mph, conditional on
the .speed exceeding 10 mph without falling below 4 mph. This
definition allows for a "creep" mode within the stop mode.
5.1.3. Vehicle -Based Measures
FT? Preliniury Report: M*y 14, 1993
65
-------
An alternative measurement unit for describing driving
behavior is the vehicle that generates the speed and trip
patterns discussed previously. Vehicle-based measures are needed
to judge the representativeness of the vehicle sample and for
analyzing various vehicle factors that may influence driving
behavior. Because this view of the data is less important to
emission test cycle development than the speed- and trip-based
measures, its discussion in this report is comparatively limited.
The final Baltimore and Spokane 3-parameter samples
contained a total of 168 vehicles, by no means an extremely large
sample. In the analyses presented in Chapter 6. The following
variables measured for each vehicle are reviewed:
* Miles driven per day
* Number of trips per day
** Number of stops per hour of operation
* Number of minutes of operation per day
These and other vehicle-based variables are derived from the
speed-time data or from trip-based variables constructed from
those data.
5.2. Descriptive Methods
Summary statistics developed from the raw data are intended
to paint a basic picture of in-use driving and to distill the
large mass of detail into forms that are readily managed in
searching for important and/or unexpected patterns. Standard
arithmetic measures include a variable's mean, standard
deviation, minimum, maximum, and count. Distributions of
frequency, both percentage and cumulative percentage, furnish
additional detail, along with the joint distribution of speedand
acceleration. These also provide the basis for approximating
distribution percentiles, such as the median. Two and three
FT? frtlmsury Report: M»y 14, 1993
66
-------
dimensional graphs of the distributions give a visual image of
variation patterns.
For reasons discussed in Section 5.4, exact percentile
calculations of speed-related data are not part of the standard
measures generated for this report. These values can be
approximated by interpolation from the frequency distribution of
a given variable. The median, a widely used measure of central
tendency, equals the 50th percentile of a. distribution. The
median usually is preferred to the mean (average) when describing
data that are distributed highly asymmetrically, or skewed.
Unlike the mean, it is unaffected by the tail values and outliers
of the distribution and, thus, is more representative of the
central mass of data. Of the three speed-related variables,
specific power exhibits large positive skewnessr speed data
typically display more modest positive skewness; and acceleration
data show little or no skewness. Therefore, in the analyses of
Chapter 6, both mean and median often are reported for the
specific power and speed variables. The distributions of several
trip variables, such as trip time and distance, also tend to be
skewed.
One of the principal goals of this study is to identify the
presence of "aggressive" driving behavior patterns that may
account for disproportionately high levels of vehicle emissions.
While there is no commonly accepted measure of driving
aggressiveness, it is generally accepted that this condition
corresponds to high speed and/or acceleration levels. It follows
that high levels of specific power also characterize aggressive
driving. For all three variables, then, the upper percentiles,
or "tails," of the frequency distribution provide a useful
measure of this important driving feature.
FTP Preliminary Report May 14, 1993
67
-------
"Tail" percentages corresponding to selected values of the
measured variable are used to describe and compare driving
aggressiveness patterns. For example, it was found that in the
3-parameter Baltimore'sample, 2.61 percent of driving time is
spent at speeds greater than 65 miles per hour. This is
equivalent to saying that the 97.39th percentile of the speed
distribution is 65 mph.
Note that percentiles associated with the distribution of
specific power are based only on those data for which that
variable is defined: during positive acceleration. This
constitutes less than half of all driving time. In discussions
of this variable, percentiles sometimes are also re-expressed in
terms of the total of all driving time.
5.3. Statistical Accuracy
In describing and comparing results, it is important to
recognize the basic unit that generates a particular measure.
The "finest" unit of interest is the one-second period of time
during which a vehicle's speed and other characteristics were
recorded. These data form the basis for other variables
aggregated by individual trip (engine-on to engine-off) and by
vehicle.
Of course, a large volume of second-by-second data were
obtained from the Spokane and Baltimore instrumented vehicle
study: over S.4 million observations for the 3-parameter vehicles
alone. Such a large sample virtually guarantees that observed
values give very accurate estimates of the true population
measures that they represent. Questions about sampling error
arise from the vehicle and trip sample sizes. For the Baltimore
and Spokane 3-parameter vehicles, some 8,459 trip observations
were made on 168 vehicles. Other potential problems involve
FTP Prtlanjaiiy Report; M»y 14, 1993
68
-------
nonsampling sources: Che representativeness of the vehicle sample
and the manner in which the vehicles were driven during
instrumentation. These issues were addressed earlier in the
discussion of possible bias sources.
In this report, primary attention is given to per-second and
per-trip driving behavior measures because these bear directly on
emission test cycle development. Driving measures based on the
vehicle unit are reviewed cursorily.
5.4. Computer Resource Issues
Analysis of the instrumented vehicle speed data presents a
substantial challenge to the management of computer resources.
Routine statistical computations on the large data set collected
in the study call for efficient, high-speed processing. At the
same time, the unique nature of the problems and their analyses
requires considerable application of non-routine exploratory
techniques, which tend to be relatively inefficient.
EPA approached these issues by combining private contractor
and Agency resources in order to draw on the strengths of both.
At the contractor level, processing of the. raw data on mainframe
computers reduced the data to the level of "sufficient"
statistics: means, standard deviations, frequency distributions,
and other potentially useful descriptive measures. These outputs
were"then studied in more ad hoc fashion by Agency staff working
on personal computers.
Resolving these resource concerns necessarily involved
certain compromises in terms of analytical methodology. For
example, exact calculation of some common descriptive statistics,
such as the median and other quantiles (percentiles), is very
computer-intensive.- In most cases, these are excluded from the
FTP PreliaiMfy Report: M*y 14, 1993
69
-------
current analysis or are approximated from other summary results.
In general, stratification of speed, acceleration, and specific
power across two or more factors also are excluded.
5.5. Driving Conditions
The speed, trip, and vehicle-based measures discussed
previously describe driving behavior over which the vehicle
operator exercises more or less conscious control. The typical
^driver is less conscious of two other factors known to influence
emissions: engine and catalyst temperature and road grade. As
part of this study, EPA investigated the temperature issue in
some depth. Road grade requires additional effort.
5.5.1. Engine and Catalyst Cooldown
In order to predict the effect of engine and catalyst
temperature on emissions, it is necessary to estimate the
proportion of"time these vehicle systems spend at different
temperature levels. Those levels are essentially constant after
the vehicle is driven for a certain period. However, at engine
start-up, engine and catalyst temperatures vary aa a function of
several factors, including the length of the preceding soak.
Starting temperatures decrease with soak time. Knowledge of
the rate of these temperature drops, combined with the
distribution of soak times, enables estimation of the
distribution of start-up temperatures in-use. When this is
further combined with information on emission-temperature
relations, a picture of overall emissions during start-up
emerges.
FT? Prclimimy Report M*y 14, 1993
70
-------
Ngw Researc^
This section describes research into cooldown rates
described above. In August, 1991, EPA entered into a cooperative
agreement with the New York State Bureau of Air Research to
obtain test data and analysis of the engine and catalyst cooldown
processes and factors affecting the cooldown rates. Researchers
at the Bureau's Albany Emissions Laboratory (AEL5 performed a
series of forty tests on six different vehicles. In each test,
the vehicle was driven under actual road conditions until the
engine and catalyst temperature were stabilized, typically for a
period of about thirty minutes. The engine was then turned off
to begin a soak, which varied from 2 to 10 hours. Throughout
each test, component temperatures and other variables were
recorded at intervals of fifteen seconds.
Figures 5-1 and 5-2 display measured temperature for engine
coolant and catalyst, respectively, for one such test. After the
engine has been turned off the engine coolant continues to rise
for several minutes before beginning its decline toward ambient;
catalyst temperatures begin the descent immediately.
The goal of this -study was to construct functions of engine
coolant and catalyst temperature versus soak time. ABL
researchers attempted to- -account for several vehicle and
environmental factors expected to influence these functions,
including engine size, ambient air temperature, wind speed, and
pavement condition (w«t or dry) . In a quasi-experimental design,
these factors were varied in an effort to fit a response surface
over a reasonable range of operating conditions.
FTP Pictimiauy Report: May 14, 1993
71
-------
O
if)
UJ
UJ
ex
C3
UJ
Q
UJ
UJ
Q.
2
UJ
§
Figure 5-1
Plymouth Acclaim Engine Cooldown Curve
a
1 2
SOAK TIME CHOURS}
o
UJ
UJ
UJ
(X
UJ
Q.
UJ
C/)
<
o
Figure 5-2
Plymouth Acclaim Catalyst Cooldown Curve
600
500
1 2
SOAK TIME CHOURSD
FTP Pirlkniuiy Report; Miy 14, 1993
72
-------
Results . - .' . .
In a series of progress reports, AIL researchers adopted a
two-stage approach to the problem of temperature prediction. The
first step was to fit component temperature as a nonlinear
function of soak time for each of the tests. This produced a set
of equation coefficients that varied across tests. In the second
step, these coefficients were regressed against the vehicle and
environmental variables to enable prediction of the coefficients
for any prespecified values of those variables,
Proposed forms for the engine coolant and catalyst
temperature-time functions were based on known physical
properties. Both are exponential functions of~ Che form:
Temperature - K*exp{S*soaktime} + C
where K, S, and C are constants for a given set of vehicle and
environmental factors. In general, fi must be negative in order
for temperature to decrease as soaktime increases. Using linear
regression, the relations between the fitted coefficients and the
vehicle and environmental variables were estimated in equations
of the form:
Coefficient » b0 + bt* (engine size) + bj* {ambient temp)
+ bj*{wind speed) + b4* (pavement condition)
In validating the AEL coolant analysis, EPA applied this
method to the 6-parameter data from the Baltimore and Spokane
studies. The 6-parameter vehicles were instrumented to record
engine coolant temperature (but not catalyst temperature). A
data file was created with values for soak time and for coolant
and ambient temperature at start and end of trip. While these
data were"generated in uncontrolled experimental conditions, they
FTP Preliauttiiy Report: Miy 1*, 19*3
73
-------
produced temperature-time equation fits similar to those of the
AEL experiments.
In adapting the AEL results for application to the survey
driving data, EPA considered several modifications of the AEL
equations to better conform with the requirements of the current
study. For the engine coolant formulas, this review failed to
produce an appreciable improvement and the AEL equations were
left intact for subsequent analyses. For several reasons, the
catalyst equations were altered to some degree.
It was learned that one of the test vehicles, an Oldsmobile
Cutlass, was equipped with a pelletized catalyst," an obsolete
technology. It was therefore decided to delete from
consideration the 10 tests performed on this vehicle. In
addition, catalyst cooldown for the single truck, a Ford Ranger
used for 11 tests, behaved very differently than in the cars.
These tests were analyzed separately.
The f'inal equations are used in Chapter 6 to help assess the
impact on vehicle emissions of cold start driving behavior. They
have the following form (t equals soak time):
Engine
Coolant - k*exp{S*|t-r|} + C
Temperature
where
k - - 0.6221*(Average Ambient Temp)
- 0.0237*(Pavement)
- 0.0058*(Number of Cylinders) * 0.9695
S « - 0.0179*(Average Wind Speed)
+ 0.0394*(Number of Cylinders) - 0.705&
T - 0.1097
C - 0.9660*(Average Ambient Temp) * 0.0582
FTP PrdimJBuy Report; May 14, 1991
74
-------
Catalyse
Temperature - ki*exp{-4*t} -t- Jcz*exp{S*t} + C
where '.." " . .. . .
" S - 1.6269* (Average Ambient Temp)
- O.l58a*( Average Wind, Speed)
+ 0.2486* (Number of Cylinders) - 3.0453
kt - 0.5224- ."'-'
kj - 1.2482 . "
. 5.5.2. Road Grade
AST mentioned previously, the road gradient data collected
through the chase car study may be of questionable" value due to
potential inaccuracies introduced by a variety of sources .
Consequently, the Agency is currently not in a position to
present data with respect to the distribution of road gradients
in Baltimore. A 1980 EPA report to Congress summarized some U.S.
Department of Transportation (DOT) data on the nationwide
distribution of road gradient by the percent of vehicle-miles
traveled (VMT) (See Table 5-1) .47 From these data it is also
possible to calculate a nationwide VMT -weigh ted average for road
gradient of !»&.. percent. Even though over time; the mileage up
U^ .emissions^ to
--
-
...
due to ^
"f 0* by the
. '>>r.t:/ .:-^;.^:t^l-;-,-'"" ^ .-..'-' - .:-' - '
downgrade ., ..,..
42U.S. Environmental Protection Agency. "Paaaenger Car Fxiel Sconooiy: EPA
>ad." Report No. EPA 460/3-80-010, September 1980, p.119.
and Road.
Report: Stay 1*. 19W
75
-------
Table 5-1
Road Gradient Distribution
by Percentage of Vehicle-Miles Traveled
(Miles Traveled Up » Miles Traveled Down)
Road
<
0.
1
2
3
4
5
Gradiant
0.5
5-1
- 2
- 3
- 4
- 5
- 6
> 6
P«rc«nt of
VMS
35
20
15
10
8
6
4
2
FTP Ptilinuotty Rtfxxt: M«y I*, 1993
76
-------
Chapter 6. Discussion and Analysis of Driving Survey
Results
This chapter provides a brief summary of the results
obtained from driving surveys in all four cities: Baltimore,
Spokane, Atlanta, and Los Angeles. A detailed discussion follows
of the Baltimore survey results; EPA feels that Baltimore results
are the most representative at this point. The chapter concludes
with comparisons between the FTP and the Baltimore results.
6.1. Summary of In-Use Driving Results for Four Cities
The four cities for which driving patterns data were
collected serve as the basis for EPA'3 study of urban, in-use
driving behavior. In looking at the results, it is important to
keep in mind that differences among cities are attributable to
not only driver behavior differences, but also to transportation
network differences, in terms of both the road network
configuration and usage. The road network can be thought of as
the driving environment faced by each driver and, as such, it can
have a significant impact on driving behavior.
~rae*sure»|tor- Mch^^;-jrJi^B'iB^^idl ^i₯'
Section 4. i; 3^:~the ^O^iM^Sa^SiLtiiIcor £o» ^gel'eiif^re j»'£t-v'r-'»-'::^"!
directly "coffl^femble ilitfi" the data for the other cities; however,
certain qualitative evaluations are still valid across cities.
14, 1M1
77
-------
Table 6-1
Comparison of Driving Behavior for Four Cities
Driving
Behavior
Measure
Speed (mph)
Average
Maximum
Standard
deviation
Number of
seconds
Acceleration
(mph/««c. )
Minimum
Maximum
Standard
deviation
Number of
seconds
Power
(ophV**c.)
Average
Maximum
Standard
deviation
Number of
seconds
Average trip
length
(siles)
Average trip
time
(minute*)
Average
distance b/w
topi
Percent idle
operation
Baltimore
Both
sites
24.50
94.46
20.52
3,365,504
-19.49
15,19
1.50
3,360,550
46.02
557.63
42.96
1,407,908
4.89
12.03
0.87
21.12
Exeter
20.90
83.72
18.47
1,686,890
-13.31
15.19
1.54
1,684,228
45.38
557.69
41.51
686,347
3.99
11.55
0.64
23.91
Rossville
28,11
94.46
21.80
1,678,614
-19.49
14.65
1.46
1,676,322
46.62
47S.82
44.29
721,561
5.89
12.56
1.13
18.32
Spokane
23.24
77,55
17.71
2,081,199
-15.46
15.95
1.46
2,077,008
40.14
672.28
40.82
880,258
3.56
9.18
0.81
17.91
Atlanta
28.80
96.48
22.60
3,010,672
-18.62
16.69
1,53
3,00«,675
-
51.72
723.12
48.11
1,318,072
6.04
12.59
1.09
17.38
Los
Angeles
28.35
80.30
20.15
99,729
-15.00
10.41
1.74
99,625
58.97
769.10
49.11
45,251
7.78
16.45
1.26
11,78
FTP fniaufLuy Report M»y 14, 1993
78
-------
Speed ' , . '
Atlanta had the highest average speed among the four cities,
followed by Los Angeles, Baltimore, and Spokane. A look at the
frequency distribution of speed affords a more interesting
picture of driving behavior. Figure 6-1 compares the
distribution of speeds for the four cities. The cities share a
characteristic tri-modal distribution. The first mode, at idle,
is similar for three of the cities: 17 percent for Atlanta, 18
percent for Spokane, and 21 percent for Baltimore. The low
percentage of idle in Los Angeles,, at 12 percent, is likely due
to the chase car survey methodology which excludes trip ends
information (see Section 4.3.1). -
In Spokane, a second mode is very obvious for speeds between
30 and 35 mph. For Los Angeles and Baltimore this mid-speed mode
is less significant; it is actually shifted out to 35 to 45 mph
speed range for Atlanta. A third mode at 55 to 60 mph is
distinct for Baltimore and Spokane; Los Angeles and Atlanta also
have a slight mode at 60 to 65 mph. In both cases, this third
mode is indicative of freeway operation.
Ac ce 1 era t i on .._'.'...
- "--.As'presented" in- Table- 6-1, there are not large differences .
in the summary acceleration measures for the four cities. In
addition, the analytical importance of acceleration" rates is '
limited' (see discussion in Section 5.1.1.).
Specific Power
The power term serves to combine speed and acceleration into
a single variable, as discussed in Section 5.1.1. The average
specific power wa_s highest for Los Angeles (58.97 mphVsec) and
FTP Prtlanauuy Report Miy 14, 1993
79
-------
Figure 6-1
Distribution of Speed
Baltimore, Spokane, Los Angeles, Atlanta
25.0%
00
o
n
p
o
Baltimore
~ Spokane
Los Angeles
~ Atlanta
o m
o o
o in
m o
o
CM
in
CM
o
CM
O
m
04
in
m
o
o
m
o
m
in
IO
O
M
O
co
in
to
in
CO
o
CO
in
co
in
o
o
CO
in
in
co
o
co
Speed (mph)
ooo
o
tn
in
en
o
en
o
o
-------
Figure 6-2
Distribution of Power
Baltimore, Spokane, Los Angeles, Atlanta
00
100.0% T
90.0%
80.0%
CO;
l_
f» 70.0%
O'
o
CM
o
O
CO
o
00
o
(O
o
o
o
oo
o
CM
o
o
o
CM
O
o
o
CM
o
03
O
CM
CM
o
o
CM
O
*
CM
O
-------
Atlanta (51.72 mph2/sec) , and lowest in Spokane (40.14 mphVsec) .
Baltimore's average power, 46.02 mphVsec, placed it in the
middle.
Figure 6-2 presents the frequency distribution for specific
power for the four cities. Since power is calculated, only for
positive values of acceleration (see Section 5.1.1), the power
distributions are based on total number of seconds of positive
acceleration (1,407,908 seconds out of 3,36S,504 seconds of total
vehicle operating time, or 42 percent). Differences in the
frequency distribution of the power- value across cities gives an
indication of the varying "intensity" of driving in the four
cities. For example, 63 percent of driving time in Spokane
occurred with power values less than or equal to 40--such power
values correspond to low speed and near-cruise driving. In
contrast, only 44 percent of power values in Los Angeles were 40
or less. The corresponding percentages for Atlanta and Baltimore
were 52 and 57 percent, respectively.
The upper tail of the power distribution is of particular
interest, as it may be the best single measure of aggressive
driving. The percentage of specific power values greater than
200 is highest for Los Angeles (1.5 %); Atlanta, with 1.25
percent, was next, followed by Baltimore (0.77%) and Spokane
(0.57%).
6.1.2. Driving Behavior - Trip Analysis
Speed, acceleration, and power measure driving behavior on a
second-by-second basis. Alternatively, driving behavior can be
examined over a more extended period of time. Trip-based
measures summarize driving behavior from the time the vehicle is
started until it is turned off-again.
FTF Prelinanuy Report: M»j» 14, 1993
82
-------
Overall, trip measures for the four cities show, an ordering
of the cities.similar to that found for the speed and
acceleration measures (these are also in-Table 6-1). The'average
trip length in Spokane was 3.56 miles while Baltimore's average
trip distance was 4.89 miles and Atlanta's was 6.04 miles. Not
surprisingly, trip durations reiterate this pattern with trips
averaging about 9 minutes, 12 minutes and 13 minutes in Spokane,
Baltimore, and Atlanta, respectively. The Los Angeles trip
measures were considerably longer than the other cities. The
average L.A. trip covered 7.78 miles in about 16 minutes. It is
likely that these long trips can be. attributed in large measure
to methodological differences between the chase car and
instrumented vehicles approaches (see Section 4.5..l) .
6.1.3. Choosing Representative Survey Data
The analysis of in-use driving behavior requires a more
detailed examination than provided by the summary statistics
described previously. For this preliminary report, the 3-
parameter instrumented vehicle data from Baltimore and Spokane
are the only practical candidates for such analysis, as discussed
in Section 4.5.2. However, the summary statistics presented in
the previous section suggest a certain uniqueness to Spokane
driving patterns relative to the other three cities. This calls
into question Spokane's appropriateness for representing other
ozone nonattainment areas. The discussion following outlines
features of Baltimore's instrumented vehicle data which make it
the best candidate for representing ozone nonattainment areas in
this preliminary technical report on in-use driving behavior.
As discussed in Section 4.2.1, the Baltimore data can be
categorized according to the I/M stations from which the vehicles
were recruited. The Exeter station is located in the central
part of Baltimore. In contrast, the Rossville station is in a
FT? PraliaiBuy Report: May 14, 1993
83
-------
suburban location. A comparison of the driving behavior from the'
two sites gives a view of urban/suburban driving differences.
The assumption is made that vehicles recruited in a suburban site
will have a tendency to be driven primarily in an suburban
environment, likewise for the urban-station vehicles.
The average speed for Exeter (urban) was much lower than
Rossville (suburban), 20.90 mph and 28.11 mph, respectively. The
average trip length was 3.99 miles for Exeter compared to 5.89
for Rossville. These differences are consistent with
expectations of suburban and urban driving patterns; higher
speeds and longer trips are presumed for a suburban road network
with a large percentage of freeways and extensive land-use
patterns.
After taking this disaggregated view of Baltimore, driving
patterns for the Exeter station turn out to be similar to those
found in Spokane, and the Rossville station's driving behavior is
very consistent with that found in Atlanta. Figure 6-3 compares
the speed distribution of vehicles from the Baltimore-Exeter
station to Spokane and Figure 6-4 compares Baltimore-Rossville
with Atlanta. Furthermore, the average trip length for
Baltimore-Exeter is similar to that in Spokane, 3.99 miles-vs.
3.56 miles. Baltimore-Rossville and Atlanta also have"very close
meanxrip lengths of 5.89 miles and 6.04 miles, respectively.
Thus, it appears that Spokane driving most closely represents
driving patterns found in a urban, central city location, while
failing to,, capture the suburban driving patterns common in many
of the major metropolitan areas.
FT? Prtlmaufy Reporu M«y 14, 1993
84
-------
Figure 6-3
Distribution of Speed
Baltimore-Exeter and Spokane
25.0%
Baltimore-Exeter
Spokane
Speed (mph)
Figure 6-4
Distribution of Speed
Baltimore-Rossville and Atlanta
Speed (mph)
85
-------
The driving behavior differences described for Spokane and
Che other cities have many potential explanations. One
compelling explanation is the differences in the configuration
and utilization of the road network. For example, the Spokane
road network is much smaller than Baltimore's. Spokane has only
one east-west freeway and a handful of principal arterials. In
contrast, Baltimore has an extensive and diverse road-network
system. Data from the chase car study can be used to gather some
insight into the differences in the road-network systems of
Spokane and Baltimore. Table 6-2 shows the fraction of chase car
operation time by road type. The Baltimore chase car data
contain a much larger fraction of freeway operation than the
Spokane data, 18.9 percent vs. 11.2 percent. Almost 80 percent
of Spokane driving took place on arterials and collectors
compared to 64 percent of total driving time in Baltimore.
Table 6-2
Comparison of Road Type Distribution for Baltimore and Spokane
I Road Type
Private :
Local
Arterial/
Collector
Freeway
Ramp
Carpool lana
Other
Total
Baltimore
Number of
seconda
2,407
22,932
122, 1S6
36,113
7,382
3
126
191,119
Percent of
total driving
1.26
12.00
63.32
18.90
3.36
0.00
0.07
100.00
Spokane
Number of
seconda
1,571
12,678
138,249
19,592
3,047
0
0
175,137
Percent of
total driving
0,90
7.24
78.94
11.19
1.74
0
0
100.00
Based on the these factors, EPA feels that it is
inappropriate to use a combined Spokane and Baltimore data set to
represent urban in-use driving behavior. A more accurate
FTP Prtlinucu? Rtjxxl: M«y 14, 1993
86
-------
representation can be made analyzing Baltimore separately.
(Adding Atlanta to the Spokane and Baltimore data set might be
even better, but as the Atlanta data are not yet available for
detailed analyses, this is currently not a viable option.) The
similarity between Spokane and the Baltimore-Exeter station
ensure that Spokane-type driving is included, and the Baltimore-
Rossville data helps to capture the type of driving which is more
consistent with that seen in Atlanta and Los Angeles. Thus, the
remainder of the driving behavior analysis will focus on just
Baltimore driving patterns. (One exception is the trip start
activity section; all instrumented vehicle data are used.) We
expect future analyses of the instrumented vehicle data will
include Spokane and Atlanta.
6.2. Analysis of Baltimore Driving Conditions
The comparison of driving across four cities gave several
descriptive measures of Baltimore driving behavior. This section
provides additional detail on overall patterns found in the 3-
parameter instrumented vehicles sampled in that city, along with
comparisons of driving over the levels of several factors likely
to influence behavior. The 93 vehicles and some of their
characteristics are listed in Appendix B.
6.2.1. Speed-based Measures of Driving Behavior
The speed, acceleration, and specific power variables form
the basis for analyzing second-to-second driving behavior.
Methods of analyzing data for these variables were discussed in
Chapter Five. It is useful to consider these variables in
absolute terms and in comparisons across the levels of driving
determinants that influence driving behavior.
FTP Pretmuniry Report: M«)r 14, 1993
87
-------
Overall Driving Behavior
In discussing overall -speed-based descriptive measures.for
Baltimore, two data sets are useful: the complete three-parameter
second-by-second sample, and the subset of these data for non-
zero speeds corresponding to that portion of operation when the
vehicle is moving. In Appendix C, Tables C-l to C-4 and Figures
C-l to C-3 provide summary statistics, distributions, and graphs
for speed, acceleration and specific power of the Baltimore data
for the full and vehicle-moving sets. After removing idle-mode
data, average speed rises from 24.50 mph to 31.06 mph. The
median speed is less affected by extreme values, and thus is
smaller in the positively skewed speed distribution. For
Baltimore, the approximate median is 23.70 mph for all driving,
and 30.2S mph for non-idle driving.
Speeds ranged to nearly 95 miles per hour; 2.61% of total
driving time, and 3.30% of time while moving, were spent over 65
mph. Accelerations varied from about -20 to +15 mph per second?
4.25% of driving time is spent in decelerations below -3 mph/sec
(5.18% of non-idle); 3.17% of the time is positive accelerations
above +3 mph/sec (4.03% of non-idle); the remaining 92.58%
(90.79% non-idle) of driving time lies between -3 and +3 mph/sec.
The distribution graphs reflect driving time-in-mode
patterns quite typical of vehicle second-by-second data. The
tri-inodal speed distribution was commented on earlier.
Accelerations exhibit a standard uniraodal pattern around zero.
Cumulative graphs of acceleration are used in this report because
they tend to draw a clearer distinction between two or more
distributions.
The less intuitive specific power is a multiplicative
composite of speed and acceleration. Mean and median specific
*
FTP Prelmoaiy Report: Miy 14, 1993
SB
-------
power are 46.02 mphVsec and 34.69 mphVsec. Appendix Figure C-3
shows the cumulative distribution of specific power for
Baltimore. (Because it is defined only for positive
acceleration, there is no comparable distribution that includes
the idle mode.) The distribution of specific power is highly
skewed, reflecting the joint tailing off of speed and
acceleration. As with acceleration, the cumulative graph
presents a more useful picture of the specific power
distribution; this form is used throughout this section. Of all
driving time, 0.22% is spent at specific power above 200 (0.77%
of driving in the positive acceleration mode) .
The joint distribution of speed and acceleration describes
the percentage of driving time spent in different modes. It's
importance to emissions is discussed in Chapter 5. Appendix
Table C-5 gives this distribution for the entire Baltimore 3-
parameter data set. Figure 6-5 presents a three dimensional view
of this distribution. This surface graph is characteristic of
such data, with a large spike at zero speed and acceleration
(idle mode) and a ridge of varying height corresponding to the
speed pattern observed earlier. In Figure 6-6, the same data are
portrayed with the.. idle^ portion of driving removed. This better
pattern of distribution during
^S>r:;'£:~^ the interact ibis --between speed
and ' acc₯ieratc^s": to- compare "the "conditional * distributions of
acceleration -at ; "different speed levels. Figure 6-7 presents this
comparison for several speed categories . Each distribution
corresponds" to a specific row in the speed-acceleration matrix of
Appendix Table C-5. These are essentially cross -sectional views
of acceleration -frequency plane from the speed- acceleration
surface distribution. They reveal the pattern of relatively low
FTP Prclitnmrr Report: M*y 14, 199}
89
-------
Distri
ibution of Speed
figure 6-5
and Acceleration
100
-------
BALTIMORE K
i.a H
o.a
0.8
0.3
o.o
Figure 6-6
'Speed and Acceleration -Baltimore
" Idle Excluded .' ,
100
60
SPEED (MPH)
to
-------
Figure 6-7
Conditional Distributions of Acceleration
Baltimore 3-Parameter Vehicles
(O
ru
LU
(3
CL
Q
LU
HI
a
in
LU
H
LD
DC
Q
-6 -5 -4-3-2-10 1 2 3
ACCELERATION CMPH/SEQ
.0 -5 mph _*_15-20 mph
45-50 mph _a_60-65 mph
mph
mph
-------
variation in acceleration at both low and high speeds, where
speed changes are physically or legally constrained.
Driving Behavior Detejrrninants/Factors
While a number of factors are known to influence individual
driver behavior, the survey conducted for the current study
obtained only limited information on such factors. In
particular, no data were collected on driver demographics. Some
vehicle characteristic information was identified, and other
variables relating to time of observation were generated from the
datalogger output. This section examines the speed,
acceleration, and specific power results across several available
factors in terms of speed-based and vehicle-based measures.
Stratification of speed-based measures includes vehicle type
and age, time of day and time of week.
Vehicle Type. One possible factor in explaining how
vehicles are driven is the vehicle itself. This is a complex
issue, since vehicles can be characterized in many different
ways. One possible set of variables might be described as
engineering-based: engine size, power-to-weight ratio,
transmission type, or body style. A second sec relates to the
vehicle's primary functions: work commuting, long-distance;-../.
travel, family errands, freight transport. Another set of
vehicle variables might in some way reflect the owner's self-
image: conservative, outdoors-oriented, sporty.
These and other potential factors are correlated and
choosing factor levels is rather subjective. In this study,
virtually no information is available on the second and third
factor types discussed above. Limited data have been obtained
concerning the survey vehicle engineering characteristics. In ,
FTP Prelimmmiy Report: Mijr 14, 1993
93
-------
Che analysis chat follows, a. "Vehicle Type" factor is defined on
Che basis of power/weight, body type, and assumptions concerning
owner/driver "perception" of the vehicle's performance
capabilities.
Using this approach, three vehicle categories were
identified:
Luxury/Sedan/Station Wagon
» Pickup/Van/Utility
Sports Car/High Performance
The Baltimore 3-parameter sample consisted of 68 Luxury/Sedan/
Station Wagon vehicles, 21 Pickup/Van/Utility, and 4 Sports
Car/High Performance. (See Appendix B for a list of survey
vehicles.)
In Appendix C, Tables C-6 to C-9 and Figures C-4 to C-6
present summary statistics, distributions, and graphs describing
each of the vehicle categories. Average speed is lowest for the
Sedan class (24.07 mph) and highest for the Pick-up class (25.59
mph) . Mean specific power is somewhat higher for the High
Performance category.
Comparing the upper percentiles for the speed distributions
gives essentially the same picture as the mean values. If
aggressive driving is judged by high speed levels, then no clear
picture emerges. The percentage of driving over 55 mph is
greater for the Sport category (12.48*) than for the other
vehicles (Sedan, 10.65%, and Pickup, 10.60%). This order changes
for speeds over 65 mph: the Sport vehicles exceeded that level
only 1.15% of the time; the Sedan class, 2.98%; Pick-up, 1.82%.
Similarly, high acceleration levels do not clearly identify
one vehicle type as being more or less aggressively driven.
FT? Prtlnuojuy Report: May 14, 1993
94
-------
Accelerations over 3 mph/sec occurred with highest frequency
within the Sedan group (3.32%) and lowest for the Pick-up class
(2.85%)
The specific power variable presents a somewhat more
definitive result. Using high specific power as a measure of
aggressive driving, the Sport group displays by far the greatest
proportion of high values. For example, Sport vehicles exceeded
a specific power level of 200 mph2/sec during 0.76% of driving
(1.89% while in positive acceleration mode) of driving. This
compares with 0.26% and 0.39% (0.63% and 0.90% in positive
accelerations) for the Sedan and Pick-up groups, respectively.
It might also be instructive to look at how low performance
vehicles are driven. However, these data are not currently
available due to the lack of a systematic definition of "low
performance." Developing this definition and analysis of driving
behavior for these vehicles is planned for the future.
Vehicle Age. The age of the sample vehicle was .used to
define three categories corresponding approximately to the
evolution of engine and emission control technologies. These
categories include:
Age Category Model Years
Old Pre-1975
Medium 1975-1982
New 1983-1992
The Baltimore 3-parameter vehicles included only members of
the Medium and New age groups. Of the 93 vehicles, 15 fall in
the Medium class, and 78 are New.
FT? Prclmaury Report Msy 14, 1993
95
-------
It might be assumed that vehicle age is a. proxy, for various
factors influencing driving behavior. For example, older
vehicles tend to be used for purposes resulting in lower mileage
accumulation and less highway driving. Comparing driving
behavior across age categories should illuminate these factor
effects.
Appendix C Tables C-1Q to C-13 and Figures C-7 to C-9 show
summary statistics, distributions, and graphs for the speed-
related variables in these categories. On average, newer
vehicles were driven some 4 mph faster than those in the medium
age category (25.10 mph vs. 21.15 mph). Accelerations among the
medium age vehicles actually vary somewhat more than for the new
category (standard deviation 1.53 mph/sec vs. 1.49 mph/aec).
Average specific power is slightly higher in the new vehicle
category (46.15 vs. 45.22 mph/sec). Median specific powers
compare similarly: New, 34.77 mphVsec and Medium, 34.18.
High speed driving is more common in newer vehicles, with
11.33% of driving time spent over 55 mph compared to 7.32% for
the Medium category. As with the summary statistics, the
acceleration and specific power distributions differ less
dramatically.
Tiaa of Day. Because the dataloggers recorded the exact
time of day for each speed observation, it is a simple matter to
categorize speed-related measures by that factor. As a first cut
in relating driving behavior to time of day, the twenty-four hour
period was subdivided into the following classes:
1:00 am to 6:00 am
6:00 am to 9:00 am Morning Rush
9:00 am to 4:00 pm
4:00 pm to 7:00 pm Evening Rush
7:00 pm to 1:00 am
FIT frttkaiQuy Report; May U, 1993
96
-------
An observation was included in a given time interval if the.
trip containing the value began during that interva^. Thus,
summary statistics of actual driving for a given class may
include data values for driving that occurred in a later
interval. Presumably, vehicle operation depends on trip purpose
and traffic conditions, both of which in turn are related to trip
start time.
In Appendix C, Tables C-14 to C-17 and Figures C-10 to c-12
summarize the speed, acceleration and specific power levels for
the five time categories. Average speed was lowest during the
evening rush hour of 4:00 to 7:00 pm (22.98 mph) and highest
during the morning rush period (25.42 mph) and night driving
period of 9:00 pm to 1:00 am (25.43 mph).
Daytime driving, between- 6:00 am and 7:00 pm exhibits the
largest percentage of high speed driving (over 55 mph), Extremes
of acceleration and specific power do not appear to be highly
influenced by time of day.
Time of Week. A second time classification of interest is
weekday versus weekend. As with time of day, there is reason to
suspect that different trip and traffic congestion patterns for
the two periods may influence driving behavior.
Summary statistics, distributions, and graphs comparing
these categories appear in Appendix C Tables C-18 to C-21 and
Figures C-13 to C-15. The hypothesis that differences exist for
the two parts of the weeks appears to be supported in terms of
the Baltimore 3-parameter speed variable. Average weekday speed
is substantially lower than on weekends: 23.92 mph and 26.35 mph,
respectively. Differences between acceleration and specific
power statistics are less pronounced.
FIT Prelmm»ry Report: M»y 14, 1993
97
-------
The distributions of speed-related variables give further
insight into their relation to time of week. For example, 9.81%
of weekday driving exceeded the speed 55 mph; this increased to
13.56% on weekends. This reinforces the impression of higher
speed weekend driving. However, using acceleration and specific
power to indicate driving aggressiveness, no evidence indicates
that weekend driving is more extreme than on weekdays.
6.2.2. Trip Measures
The nature and patterns of vehicle trips are another
important area in the study of driving behavior. This section
examines such trip characteristics as trip duration, trip
distance, time between trips, and vehicle operating modes
summarized over the trip. In addition, extremes of speed-based
measures are analyzed from a trip perspective.
For this study, a trip basically begins when the engine is
turned on and ends when the engine is shut off. (See Section
5.1.1 for a more detailed discussion of trip definition.) This
definition allows for very short trips, such as moving a vehicle
in the driveway. Even zero distance trips are possible with this
definition--a car is started, probably with the intent of driving
it somewhere, but never moved. In fact, 64 zero mile trips were
recorded in Baltimore.
The 3-parameter instrumented vehicles in Baltimore generated
a total of 4,681 trips. The average trip covered 4.89 miles and
lasfefed an average of 12 minutes. The median values for trip
distance and duration indicate a much shorter "typical" trip of
8.8 minutes in duration and 2.53 miles in distance.
Figures 6-8 and 6-9 present the frequency distribution for
trip time and distance. A little more than a quarter of the
FTP pnUmmuy Report: May 14, 1993
98
-------
Figure 6-8
Distribution of Trip Time
Baltimore 3-Parameter Vehicles
Trip Time (minutes)
Figure 6-9
Distribution of Trip Distance
Baltimore 3-Parameter Vehicles
Trip Distance (miles)
99
-------
crips are a mile or less. Only about 12 percent: of the trips
were longer than 10 miles. In terms of trip duration, about 5
percent of the trips were very short, lasting a minute or less.
Over the course of a trip, a vehicle will spend a certain
fraction of time in each of four operating modes: idle, cruise,
acceleration, and deceleration. In Baltimore, the average trip
involved 23 percent idle operation, 34 percent cruise, 22 percent
acceleration and 20 percent deceleration. (See Section 5.1.2 for
mode definitions.)
TripFactors/Determinants
The characteristic features of a trip are dependent upon
numerous factors including trip purpose, driver attitude, trip
location, and the driving environment encountered during the
course of the trip. None of these factors can be determined from
the survey data. However, certain known variables can be used as
a proxy for one or more of these unknown factors. Knowing the
time of day when a trip took place, it may be possible to make
some crude aggregation of such factors as trip purpose, driving
environment, and possibly driver attitude. For example, a trip
occurring between Sam and 9am could be considered a morning
commute trip driven under congested conditions by a driver who
may react to a very familiar driving environment in an aggressive
manner.
This section considers several factors for which trip data
were collected. Time of day and time of week (weekday vs.
weekend) are examined along with vehicle type and vehicle age.
These preliminary analyses serve as a starting point for further
work in understanding the complex nature of trip patterns.
FTP PreEiQUBuy Report May 14, 1993
100
-------
Time of Day. As in Section 6.2.1, time of day is separated
into 5 categories, selected to roughly represent 2 periods of '
peak vehicle operation (Sam-9am, 4pm-7pm) and 3 periods of off-
peak operation (lam-Sam, 9am-4pm, 7pm-lam) . Table 6-3 presents
summary statistics for trip time, distance, and idle time
stratified by time of day. A large fraction of the trips
occurred during the off-peak times, with nearly half of all
trips taken during the mid-day period. Almost twice as may trips
were taken in the afternoon peak period than in the morning peak.
The morning peak period had the longest average trip distance
(6'.65 miles) while mid-day trips averaged only 4.45 miles.
Average trip"time was greatest during the early morning off-peak
period. From these summary measures it appears that"the
afternoon peak-period looks more like the mid-day off peak than
the morning peak.
Time of Week. Trips were also stratified by weekend and
weekday to examine the need for isolating weekday operation for
future analyses. Trip time and distance are typically longer on
weekends than weekdays. These differences are not particularly
large; the median trip length of 2.56 miles for weekends compares
to a median of 2.35 miles for weekdays. The variation in
distance and duration.weekend trips is considerably greater than
found for trips taken during the week. The standard deviations
shown in Table 6-3 illustrate these differences.
Vehicle Type. It is reasonable to expect that the type of
vehicle could have an impact on trip behavior. For example, a
commuting car may only be used to make long trips to and from the
workplace and a sports utility vehicle may be used for long trips
on weekends. The vehicle type classification for this report is
very broad, categorizing vehicles as either: luxury, sedan,
station wagon; pick-up, van, utility; and sport/high performance
(see Section 6.2.1).
FTP Prelimauuy Report; May 14, 1993
101
-------
The average trip time of sport/high performance vehicles
(12.71 minutes) was longer than for either pick-ups (12.56
minutes) or sedans (11.8 minutes). Trip length differences among
the vehicle types were less distinct. The average trip length
for pick-ups was 5.32 miles, slightly higher than mean trip
length of 5.22 for sports/high performance (see Table 6-3), The
sedans average trip length was just 4.72 miles. Interestingly,
the median trip length was longer for sport/high performance
vehicles than pick-ups. The contrasting results for the mean and
median suggest that the pick-up class had some extremely long
trips which only impacted the mean. In fact the longest trip for
pick-ups was 113 miles compared to only 37 miles for the
sport/high performance vehicles'.
V«hicla Age. The age of a vehicle can affect the way the
vehicle is driven in terms of the type and frequency of trips.
Concerns regarding a vehicle's reliability may lead people to
limit their driving of older vehicles to short trips. To take a
very cursory look at the influence of vehicle age on trip
patterns, EPA categorized vehicles into three broad age groups:
pre-1975, 1975-82, 1983-1992. For Baltimore, the I/M program
restricted recruitment to 1977 and newer model year vehicles;
thus, only the two newer categories are considered here.
The newer vehicles' (1983-92) average trip time was 12.36
minutes and covered an average of 5.14 miles. Both measures were
considerably shorter for older vehicles(1975-82), with a mean
trip time of 10.48 minutes and a trip length average 3.7 miles
(see Table 6-3).
FTT» Prelkuuuy Report May 14, 1993
102
-------
Table 6-3.
Baltimore Stratified Trip Measures
Characteristic
Tiae of Day
lam to 6am
6am to' 9am
9am to 4pm
4pm to 7pm
7pm to lam
Tima of Week
Weekday
Weekend
Vehicle Type
Luxury, sedan,
station wagon
Pick-up trucks,
van, utility
Sports, high
performance
Vehicle Age
1975 to 1982
19S3 to 1992
Trips
Number
100
572
2,284
1,047
678
3,578
1,103
3,294
1,207
180
816
3,865
Percent
of total
2.14
12.22
48.79
22.37
14.48
76.44
23.56
70.37
25.79
3.85
17.43
82.57
Trip trine (minutes)
Mean
14.37
15.78
10.89
11.97
12.44
11.94
12.32
11.80
12.56
12.71
10.48
12.36
Median
15.06
13.28
7.74
8.80
9.05
8.87
8,53
8.43
9.28
10.88
7.43
9.05
Std.
9.23
12.46
11.39
10.85
12.80
10.85
14.06
11.77
11.70
9.72
10.82
.11.83
Trip Distance (miles)
Mean
5.85
6.65
4.45
4.55
5.27
4.75
5.35
4,72
5.32
5.22
3.70
5.14
Median
3.77
3.91
2.18
2.45
2.80
2.56
2.35
2.37
2.80
3.89
1.81
2.66
Std.
5.87
8.95
8.29
6.07
8.74
7.04
10.49
8.15
7.80
5.82
6.62
8,23
Distance
between
stops
(miles)
1.33
1.08
0.80
0.78
1.01
0.84
0.97
0.81
1.04
0.96
0.64
0.92
-------
Trip-'baaed Analysis of Speed. Acceleration, and Power
The driving behavior exhibited over the course of a given
trip varies greatly and is affected by a host of factors. Thus,
even on a trip basis, one would expect to see considerable
variation in driving behavior variables such as speed,
acceleration, and power. A trip-based analysis of the extremes
of these driving behavior variables provides an alternative
perspective of the incidence of aggressive driving.
High speed operation(> 60 mph) accounts for 7 percent of
total driving time. On a trip basis, it is clear that the
majority of trips in Baltimore do not involve driving at high
speeds. Only about 14 percent of all trips had a maximum speed
greater than 60 mph. Figure 6-10 shows the cumulative
distribution of maximum speed per trip.
In marked contrast to high speeds, extreme values of power
are found on disproportionately large fraction of trips relative
to their overall frequency. About 30 percent of all trips had a
maximum power value greater than 200, while as a percentage of
total instances of positive power values, a power value of 200 or
greater accounts for only 0.77 percent (0.22 percent of total
vehicle operation). (See Figure 6-11)
High acceleration rates illustrate this point as well. Nine
out of ten trips contain accelerations greater than 3 mph/second,
yet these accelerations only account for about 3 percent of total
driving time. (See Figure 6-125
These differences illustrate how certain driving parameters
are closely related to the type or purpose of a trip (high speed
operation), while others (speed and acceleration) are a more
FTP Pitlooiauy Report; May 14, 1993
104
-------
Figure 6-10
Cumulative Distribution: Maximum Speed per Trip
Baltimore 3-Parameter Vehicles
"C
H
"5
"c
0)
o
0)
Q.
IUU.U70 '
90.0% -
80.0% -
70.0% -
60.0% -
50.0% -
40.0% -
30.0% -
20.0% -
10.0% -
n no/ J
^~
/
jr
::::::::::::::/:::::::::::::::
::::::::::::::V::::::::::::::::::::::
rr , i , , i i - » .- l - -i i i- - --I-
o
CM
O
CM
O
o
n
o
Ui
o
to
o
in
o
h-
o
CO
o
CO
o
o
o
en
Speed (mph)
-------
Figure 6-11
Cumulative Distribution: Maximum Power per Trip
Baltimore 3-Parameter Vehicles
100.0% T
90.0% --
80.0%
-------
O)
Q.
*C
<
o
**
r
0)
u
k.
4)
0.
4 AA no/ _
1 UU.UYo
QA AO/ ,
yu.u /o
MAO/ -
.U /o
7A AO/
/U.0%
ftA AO/ -
OU.UYo
eft AO/
OU.UVo
ilA AO/
4U.U/O
QA AO/
ou.UTo
OA AO/
UU.Uyo
'[ !.'[« ;':'.i", " ^^P^**^^^^
: ' , ' :>ll|-;:t:. : x^^
" ' "^ll|l|l'i:; ' /
: ; '!|li|l:;' /
.................... .'l!. .'-.'I .{a'l'.I.' . .'.... ^. ..'............. -.,..
i-n^fi / "
- -;i.^«:. /
J^.^IL / 1
' --^'^^iir^j*
li'! !'- j^/ .
................... ....' L--.- «>!;,. !.V^'* .-............-..--.
::,.v:1?ryr-n
-^pj^:'il:
';;^|fli!->--
1 ₯ . T 1 I 1 -1 - 4- _ I- J - -1 1- t 1 1 t - 1
CM
«o
n
r-
*
-------
universal crip characteristic. From this perspective, the
"tails" of the acceleration and power distribution take on
greater significance.
6.2.3. Vehicle-based Measures
Several summary statistics were compiled from the survey to
describe driving behavior variation among vehicles. These are
reviewed briefly here, for all Baltimore 3-parameter driving, and'
for several factor comparisons. Refer to the Radian Corporation
report for additional detail on these and related variables.43
The Baltimore 3-parameter sample contained 93 vehicles.
Variables described in this section were listed earlier: miles
driven per day, number of trips per day, number of minutes of
operation per day, and number of stops per hour of operation.
In summarizing the data for these variables, it was
appropriate to weight their values to account for differing
amounts of instrumentation time and operation time. Therefore,
in the outcomes reported below, means and standard deviations for
the first three variables are weighted by the number of days the
vehicle was instrumented; for the last variable, those statistics
are weighted by the number of hours the vehicles was operated.
In Appendix C, Table C-22 gives summary ressults for the 93
Baltimore vehicles. Mean daily mileage is 30.56, while one
vehicle averaged over 105 miles per day. Daily time of operation
averaged over 75 minutes; number of trips, 6.25 per day; and
number of stops, over 35 per hour of operation.
43Radian Corporation, "Light Duty Vehicle Driving Behavior: Private Vehicle
Inntrumentation, Volume 1," Draft Report to U.S. IPA, August 24, 1932.
FT? Pttliouuiy Report; M«y 14, 1993
108
-------
The vehicle-based measures used in the discussion of overall
Baltimore driving behavior were summarized by the levels of
several factors, including vehicle type and age, and time of
week. As before, means and standard deviations are weighted
appropriately.
Comparisons of these statistics for several stratification
factors are found in Appendix C Tables C-23 to C-25. The factor
levels are defined in Section 6.2. Some general observations
follow. Newer vehicles were driven somewhat longer and farther
per day, and averaged fewer trips and fewer stops per hour.
Weekday driving produced substantially higher average time,
higher distance and number of trips than weekdays. For the
vehicle type factor, average daily usage is considerably higher
for the Pickup/Van/Utility class than for the Sedan and Sport
classes; the per hour stop rate of Sedans was somewhat higher.
Presumably, these outcomes relate to differences in trip purpose
associated with these vehicle types.
6.2.4-. Vehicle Soak
The soak period44 is a critical factor in the level of
emis8ions_generated_'after^a; vehicle. start. For illustration','.'
'Eio^re/u£^3jrs^
" rangi-ngi. ;f rd8;- l^^^il^.'ifeo'^weraigEfe-" f ojr -^
: first- ix^m^j^^^l^^iSi^^^^^t''^^- .the
the cycle).4* .The data indicate that HC and CO ""emissions during
**"Soak period" 10 the time after the vehicle has been turned off until it
is restarted." " '
data is from a test program run at KPA's National Vehicle and Fuel
Emissions Laboratory in Ann Arbor, MI. All the soaks were conducted with an
ambient temperature of 75OF and second-by- second emission data were gathered on
a modal test bench. The vehicles tested were a 1992 Ford Crown Victoria with a
4.6L engine, a 1992 Dodge Dakota pickup with a 5.2L engine, a 1991 Honda Accord
with a 2.2L engine, a 1991 Dodge Caravan with a 3.0L engine, and a 1990 Pontiac
6000 with a 3. IX. engine.
FTP PntiauoMry Report: M*jr 14, 1991
109
-------
Figure 6-13.
Hill 1 - Average Tailpipe Emissions
3.5
25.00
3 --
^ 2.5
CO
E
a
3 2
X
0 15
z 1
-------
the first 2 minutes of operation can increase by a factor of 10
after long soak periods, relative to 10 minute soak periods, and
NOx by a factor of 5, Even relatively short soaks of 45 minutes
increased all emissions by a factor of 4-5.
This emission increase occurs due to engine and catalyst
cooldown. Typical engine and catalyst cooldown curves can be
generated using the equations developed from the Albany Emission
Laboratory work, as discussed in Section 5.5.1. Figure 6-14
shows the predicted cooldown curves for a 6-cylinder vehicle at
75°F with a 4 MPH wind and dry pavement. The curves demonstrate
that the catalyst cools off much faster than the engine. Thus,
most of the emission increase found when increasing soak time
from 10 minutes to about an hour is probably due to catalyst
cooldown. After an hour, engine cooldown becomes increasingly
important as the soak time continues to increase and the catalyst
cooldown becomes less important (as the catalyst light-off
temperature is around 65Q°F, the catalyst is close to being
completely cold after 45-60 minutes).
Because of the potential emission impacts, one of the
primary objectives of the driving surveys was to examine the
distribution of actual soak times. Figure 6-15 summarizes the
distribution of soak times found for the Baltimore 3-parameter
data set. A more detailed distribution is presented in tabular
form in Appendix D.
The soak distributions include a large number of relatively
short soak periods. This indicates that drivers frequently
string trips together. In fact, only about 18%'of the trips
occurred after a soak period greater than 8 hours, indicating
that single trips directly to work (or some other destination)
and then home again for the night are more the exception than the
rule.
FTP Prelimmiiy Report: M*y 14, 1993
1H
-------
Figure 6-14.
Typical Catalyst and Engine Cooldown Curves
975
875
775
lT
£ 675
3
O
a
E
0}
m
575 --
475
« 375
-------
Figure 6-15.
Baltimore Soak Period Distributions
U)
I Distribution
Cumulative Distribution
0-10 min 10-30 min 30-60 min 1-2 hrs
Soak Period
2-4 hrs
4-8 hrs
8+ hrs
-------
To help put the soak time information into context, the
cooldown formulas developed in Section 5.5.1. were applied to the
Baltimore soak distributions to develop a profile of engine and .
catalyst temperatures at vehicle start (assuming an ambient
temperature of 75°F, wind of 4 MPH, and dry pavement) . These
temperature distributions are presented in Table 6-4. It should
be noted that these temperature distributions overstate the
initial temperatures to some degree, as it was assumed that the
engine and catalyst had reached normal operating temperatures
prior to the, soak period, which is not true for every soak.
Table 6-4
Temperature Distribution at Trip Start
Catalyst
Temperature
(°P)
800+
700-800
600-700*
500-600
400-500
300-400
200-300
<200
Frequency
( Percent 5
22
9
7
6
5
5
6
40
Engine
Temperature
<°P)
180 +
160-180
140-160
120-140
100-120
<100
Frequency
(Percent)
SI
7
6
6
7
23
Catalyst light-off range
The data indicate substantial differences between catalyst
and engine conditions upon vehicle start-up. A high proportion
of vehicle starts (at least 62%) occur with catalyst temperatures
Ft? Pitlimkuuy Report M»y 14, 1993
114
-------
below the light-off range. However, che engine is quite warm
for most vehicle starts. Only about 23% of starts occur with
engine temperatures below 100°F and about 64% of all starts had
engine temperatures above 140°F.
6.2.5. Trip Start Driving Activity
Vehicle soak periods impact emissions by affecting engine
and catalyst temperatures at the start of the next trip; the
longer the soak period, the more the engine and catalyst cool
off. How the vehicle is operated after start-up may also impact
emissions, as it affects the rate at which the engine and
catalyst warm up and the level of emissions from the vehicle
during the warmup period.
Most of the exhaust emissions from modern, properly
operating vehicles occur during the first 1-3 minutes of
operation after the vehicle has been started. However, this was
not the case when the current FTP driving cycle was developed.
Vehicles in the late *60's always ran with rich air/fuel ratios
and did not have catalysts. Start emissions from these vehicles
were not very different from hot stabilized emissions and how the
vehicles were driven during the engine warmup period had little
impact on overall emission levels. Thus, there was little need
to evaluate whether vehicles are driven differently after they
are started.
The introduction of the catalyst and more sophisticated fuel
*
control systems dramatically altered the relationship between
emissions during the warmup period and hot stabilized emissions.
Hot stabilized emissions have been reduced to relatively low
levels with modern catalysts, plus improvements in fuel systems
ensure efficient catalyst operation even during moderate
transient operation. However, emissions shortly after cold (and
FTP Preliminary Report; May 14, 1993
115
-------
warm) starts have not been reduced proportionally due to the need
to warm up the catalyst and the inability to immediately enter
closed-loop operation. Thus, how vehicles are driven shortly
after they have been started is a much more important
consideration today than when the FTP was developed.
One concern that EPA investigated was whether or not the
length of the soak period influenced driving behavior after
start-up (for example, are vehicles driven differently after
overnight soaks at home than they are after short soaks at
stores). Therefore, the first step'of the analysis was to divide
the start data into cold, warm and hot starts. A cold start is
where both the engine and catalyst are cold, a warm start is
where the engine is warm and the catalyst is cold, and a hot
start is where both engine and catalyst are warm (as catalysts
cool off much faster than engines, starts with warm catalysts and
cold engines are not a concern). Based on previous test
programs, it is known that catalysts warm up after about two
minutes of driving. The work by New York's Albany Emission
Laboratory indicated that catalysts cool off to fairly cold
levels in about 45 minutes. Thus, a start was classified as
having a warm catalyst if two conditions were met: (1) the
previous trip prior to the soak was less than two minutes in
duration, and (2) the soak period was less than 45 minutes. If
either the previous trip was less than 2 minutes or the soak
period was longer than 45 minutes, the start was classified as
having a cold catalyst.
For engine classification, Radian developed engine warmup
and cooldown algorithms from the 6-parameter data vehicles (which
monitored coolant temperature). The warm-up algorithm related
engine temperature rise to idle time and vehicle speed while
moving. The cooldown algorithm related engine temperature fall
to ambient air temperature, engine displacement, and soak time.
FTP Prclnakuuy Report Mijf 14, 1993
116
-------
These algorithms were then used by Radian to predict engine
temperature at the beginning of every trip in the 3-parameter
data set (both Spokane and Baltimore). If the predicted engine
temperature at the beginning of the trip was within 10°C of the
ambient temperature, the start was classified as having a cold
engine. If the difference was greater than 10°C, the start was
classified as a warm engine.
The next step in assessing start driving behavior was to
define the length of time46 after a vehicle start that driving
should be classified as "start", as opposed to hot stabilized.
To facilitate this analyses, EPA decided to strip off the initial
idle period and analyze it separately from the data after the
vehicle began to move. This was done because of the substantial
impacts of ambient temperature on the initial idle time and to
equitably compare driving after cold, warm, and hot starts.
The criteria established for the start period was to
encompass the portion of the warm-up period that relates to
elevated emission levels. Based upon an analysis of trip start
emission activity by Sierra Research,47 a coolant temperature
threshold of 140°F was selected as the point at which elevated
emission levels largely disappear during vehicle warmup.
Using the 140°F coolant temperature threshold, Radian
analyzed the 6-parameter coolant temperature data for the amount
of time it took to reach 140°F on each trip, excluding the
initial idle. The distribution of these times are listed in
was used instead of distance because the data was gathered in
second-by-second increments. A distance criteria would also create problems when
generating new testing cycles, which are time based.
47Reference: Technical Note Under Contract No. 68-Cl-0079 In Response to
Work Assignment 1-05, "Evaluation of Trip Start Activity, Subtasfc Ib, Definition
of Trip Start Condition", by Sierra Research, dated April 12, 1993.
FTP Prelimmuy Report: Miy 14, 1993
117
-------
Appendix E, both for all trips (Table E-l) and broken down by
cold, warm, and hot starts (Tables E-2 to E-4, respectively).
The data indicate that over 50% of all cold starts reach 140°F
coolant temperature after 240 seconds of operation. Also, almost
95% of all warm starts reach 140 °F coolant after 240 seconds and
over 90% of all starts, combined, reach 140°F after 240 seconds.
Thus, 240 seconds, excluding the initial idle period, was
selected as the "start" portion of overall driving. To further
distinguish the sequence of driving behavior after starts, this
240 second period was broken up into three equal 80 second
"phases" for analysis purposes.
Radian analyzed speed, acceleration, and power distributions
for each combination of the 80 second phases and the different
trip start definitions. The averages and standard deviations for
each group, as well as for overall driving, are presented in
Table 6-5.
The averages and standard deviations indicate that driving
behavior 'is not significantly affected by the length of the soak
time. For any given time phase, the averages of speed,
acceleration, and power are very similar for each start category
(i.e. cold, warm, and hot). Graphs comparing the actual
distribution of speed and power are presented in Appendix E
(Figures E-l to E-3 show the speed distribution by start category
for phases 1-3, respectively,- Figures E-4 to E-6 show the
cumulative power distributions). The distributions for each
individual phase show remarkable similarity across all three
start categories. Thus, the only driving behavior that needs to
be considered as a function of soak time is the initial idle
period.
FTP PnlimBuy Report: M*y 14, 1993
118
-------
The data in Table 6-5 also indicate that driving behavior
immediately after a start (be it cold,'warm, or hot) is different
Table 6-5
Trip Start Driving Behavior by 80-Second Phases
Start
Phase
1
2
-..
3
Hot*
Type
o£
Start
Cold
Warm
Hot
Wgt.
avg.
Cold
Warm
Hot
Wgt.
avg.
Cold
Warm
Hot
Wgt.
avg.
Number of
Seconds
134,397
171,764
313,433
145,684
156,818
273,546
134,348
141, 53S
233,245
3,365,299
Speed
-------
than all the other phases, although phase 2 and phase 3 actually
have higher average power values than hot stabilized driving.
Figure 6-16 displays the speed distribution for Baltimore
and Spokane by phase (i.e. each of the first three 80 second
phases of driving, plus driving after the first 240 seconds, with
the different start types combined). This figure shows that the
speed distribution during the initial 80 second phase is
completely different than for subsequent driving, with the most
frequent driving speed range being 0-5 mph and the frequency
dropping steadily with increasing speed. The speed distributions
for all subsequent driving phases have more similar shapes, with
the most frequent speed range being 30-35 mph. Phase 2 and Phase
3 of start driving behavior are especially similar, although
Phase 3 has a little less driving in the 5-25 mph range and a
little more in the 45-65 mph range. The speed distribution for
hot stabilized driving (that is, the "rest of trip") differs in
having substantially more driving in the 50-75 mph range.
Cumulative power distributions for each phase of driving are
presented in Figure 6-17. These distributions demonstrate that
the aggressiveness of the driving for each phase is more similar
than the speed distributions. Phase l spends substantially more
time at very low power levels, but the distribution of power
levels above 80 is very similar to hot stabilized driving* Phase
2 and Phase 3 are virtually identical to each other, although
both are a little more aggressive than either Phase 1 or hot
stabilized driving.
FTP fttliasarf Report: Mi? 14, 1993
120
-------
Figure 6-16.
Speed Distribution by Phase
18.00
16.00
Rest of Trip
Phase 1
Phase 2
* Phase 3
0.00
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80
Speed (Top of range - MPH)
-------
Figure 6-17.
Cumulative Power Distribution by Phase
to
100 -T-
90
£. 80 ~~
o
c
0)
3
er
0)
0)
>
*-
a
a
E
3
a
70
60 --
50
40
30
20
40 60 80 100 120 140 160 180
Power (maximum of Range)
200 220 240
260
-------
Initial Idle
The distribution of the initial idle time for all trips (3-
parameter, Spokane and Baltimore) has been analyzed and is
presented in Appendix E for each start category (Tables E-5 to E-
7 for cold; warm, and hot starts, respectively}. Table 6-6
summarizes the average and median initial idle times for Spokane
and Baltimore.
Table 6-6
Initial Idle Time, Baltimore and Spokane
Start Type
Cold start
Warm start
Hot start
Overall
Mean
(seconds)
59
28
15
29
Median
(seconds)
13
8
5
diacrepancy between;: t^e .mean and .the .median ^
Leant number of extended
had, initial
TllSiiB-C-£p^ ' This
discrepancy needs 'to*--be addressed" in the course of generating
representative test cycles. : ' .
The data also indicates that cold starts have much longer
initial idle times than hot starts. This also needs to be
addressed during cycle development.
FTP Pntiaaaff Report: Miy 14, 1993
123
-------
One potential problem with the Baltimore and, especially,
Spokane data is that the studies were conducted during colder
weather when people frequently let their cars warm up before
driving off. The average daily temperature during the survey
period was roughly 40°F and 50°F for Spokane and Baltimore,,
respectively. The initial idle periods would probably be shorter
during the summer months more representative of ozone
nonattainment, especially after a cold start.
One way to mitigate this concern would be to use the Atlanta
data to calculate initial idle time, as the Atlanta survey was
conducted during August and early September. The average daily
temperature during the survey period in Atlanta was between 70
and 75QF. A comparison of the Atlanta to the combined Spokane
and Baltimore initial idle times is presented in Table 6-7 (the
distribution of initial idle times for Atlanta are presented in
Appendix E, Tables E-8 to E-10 for cold starts, warm starts, and
hot starts, respectively).
Table 6-7
Initial Idle Time for Baltimore/Spokane vs. Atlanta
Start Type
Cold start
Warm start
Hot start
Baltimore and Spokane
Mean
(seconds)
59
28
IS
Median
(seconds)
13
8
5
Atlanta
Mean
(seconds)
23
18
12
Median
(seconds)
9
7
5
Proport;ion gf Drlvincf Occurring After Starts
The final factor that will be needed to assess the emission
impact of start driving behavior is the proportion of overall
FIT Prelmmuy Report: Mty 14, 1993
124
-------
driving represented by the start driving behavior. A breakdown
of the survey data -into-miles driven from the start of each 'trip
is presented"in Figure 6-18 (the cumulative distribution can be
read as "the proportion of all driving that was driven less than
or equal to the number of miles on the X-axis after the start "of
a-trip").. The reduction in the miles spent in each speed range
as the miles since the start of the trip increases reflects the
high percentage of short trips in the data base.
FTP Prtlnnjowy Report: May 14, 1993
. .' . 125
-------
CD
*!
C/3
£
o
CO
<0' g
si
ll
I
o a
cw -C
+*
"o
»*
IB
a
O
c
a
CO
a
-------
6.3. Comparisons to the Federal Teat Procedure
The Federal Test Procedure is intended to represent in-use
driving behavior. The Baltimore 3-parameter data offer an
empirical source for judging the success of the FTP in capturing
real driving characteristics. This section compares the FTP and
Baltimore data using the various descriptive tools employed
earlier.
6.3.1. Speed-based Measures
Using the speed-related data, driving on the FTP is more
conservative than that observed in Baltimore. Table 6-8 compares
summary statistics on speed, acceleration and specific power for
the two sources. Typical FTP speeds are lower than for the
Baltimore data: the mean FTP speed is 19.59 mph (24.14 mph when
the vehicle is moving), compared to 24.50 mph (31.06 mph when
moving) for Baltimore. Accelerations in Baltimore range from -15
to -1-20 mph/sec, with standard deviation 1.50; the FTP range is
much narrower, -3.3 mph/sec to +3.3 mph/sec, with standard
deviation 1.40. Specific power (during positive acceleration)
for the Baltimore sample averaged 46.02 mph2/sec; the median is
34.7. For the. FTP, the mean and median are much lower, 38.60 and
21.6, respectively..- . " ' .-.-'. "..;,''
Tables 6-S to 6-~ll and Figures 6-If to 6-21 contrast the
distributions of speed-related variables for Baltimore with those
of the FTP. As is well documented, the FTP contains no speeds
above 56.7 mph, aad no positive or negative accelerations in
excess of 3.3 mph/sec. In Baltimore, roughly 8.5% of speeds
(between 6.35% and 10.73%) fell above 56.7 mph. For
accelerations, about 2.5% of driving exceeded 3.3 mph/sec, while-
around 3.5% of decelerations were under -3.3 mph/sec.
FTP Pretinuuiy Report: Miy 14, 1993
127
-------
The largest FTP specific power is 192 mphVsec. In the
Baltimore survey, about 0.3% of all driving time exceeded that
level (about 1% of time in positive acceleration mode).
Table 6-8
Summary Statistics
Baltimore and FTP
Baltimore 3-Parameter Vehicles
Mean
St Dev
Max
Min
Count
Speed (mph)
FTP
19.59
14.69
56.70
0.00
1,369
Baltimore
24.50
20.52
94.46
0.00
3,365,504
Acceleration (mph/»«e)
Mean
St Dev
Max
Min
Count
Mean
St Dev
Max
Min
Count
FTP
0.00
1.40
3.31
-3.31
1,361
Power (mphV«ec
FTP
38.60
31.74
191.85
0.01
544
Baltimore
- 0.00
1.50
15.19
-19.49
3,360,550
)
Baltimore
46.02
42.96
557.69
0.00
1,407,908
FIT httinuniry Report: May 14,
128
-------
Table 6-9
DISTRIBUTION OF SPEED
Baltimore and FTP
Baltimore 3-Parameter Vehicles
Spaed
(mph)
0-0
0-5
5-10
10-15
15-20
20-25
25-30
30-35
35-40
40-45
45-50
50-55
55-60
60-65
65-70
70-75
75-80
80-85
85-90
50-95
Total
Baltimore
Count
710,890
207,698
181,442
201,708
209,420
231,917
280,040
308,521
253,681
173,105
125,294
121,004
147,398
125,917
66,826
17,485
2,803
253
91
11
3,365,504
Percent
21.1
6.2
5.4
6.0
6.2
6.9
8.3
9.2
7.5
S.I
3.7
3.6
4.4
3.7
2.0
0.5
0.1
0.0
0.0
0.0
100.0
"Cum.
%
2J..1
27.3
32.7
38.7
44.9
51.8
60.1
69.3
76.8
82.0
85.7
89.3
93.7
97.4
93.4
99.9
100,0
100.0
100.0
100.0
FTP
Count
258
76
74
80
129
265
248
86
42
8
27
52
24
0
0
0
0
0
0
o
1,369
Percent
18.8
5.6
5,4
5.8
9.4
19.4
18.1
6.3
3.1
0.6
2.0
3.8
1.8
0.0
0.0
0.0
0.0
0.0
0.0
0.0
100.0
Cum.
V
18.8
24.4
29.8
35. 6
45.1
64.4
32,5
as. a
91.9
92.5
94.4
98.2
100.0
100.6
100.0
100.0
100.0
100.0
100.0
100.0
FT? PrelinuBuy Report: M»y 14, 1993
129
-------
Table S-IQ
DISTRIBUTION OF ACCSI^SRATIONS
Baltimore and FTP
Baltimore 3-Parameter Vehicles
Accel
fmrVH /Rpf*)
-20 to -19
-19 to -18
-18 to -17
-17 to -16
-16 to -IS
-IS to -14
-14 to -13
-13 to -12
-12 to -11
-11 to -10
-10 to -9
-9 to -8
-8 to -7
-7 to -6
-6 to -5
-5 to -4
-4 to -3
-3 to -2
-2 to -1
-1 to Q
0
0 to 1
1 to 2
2 to 3
3 to 4
4 to 5
S to 6
6 to 7
7 to 8
8 to 9
9 to 10
10 to 11
11 to 12
12 to 13
13 to 14
14 to 15
i«5 to ifi
m-_fc -I
Baltimore
fntl*"1^"
1
0
0
0
0
3
2
11
34
44
149
54 5-
1,801
5,756
17,199
41,575
76,005
120,151
219,036
767,792
702,538
899,795
267,355
133,587
62,415
27,375
10,639
3,712
1,649
776
325
187
62
20
8
2
1
*3 3 fin "5^0
0.0
0.0
0.0
0.0
0.0
O.Q
0.0 ~"
0.0
0.0
0.0
0.0
0.0
0.1
0,2
0.5
1.2
2.3
3.6
6.5
22.8
20.9
26.8
8.0
4.0
1.9
. 0.8
0.3
0.1
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
no
1 00 0
f*i?m %
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.1
0.2
0.8
2.0
4.3
7.8
14.4
37.2
58.1
84.9
92.8
96.8
98.7
99.5
99.8
99.9
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
inn_o
FTP I
Pritini"
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
76
56
85
258
349
314
125
43
56
0
0
0
0
0
0
0
0
Q
0
0
0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
5.6
4.1
6.2
18.9
25.5
23.0
9.1
3.6
4.1
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
no
fiim * II
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
5.6
9.6
15.9
34 . 7
60 . 2
83.2
92.3
95.9
100.0 1
100 . 0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100 . 0
inn 0 II
1368 inn o l|
FTF Pnlimioaty Report: M*y 14, 1993
130
-------
Table 6-11
DISTRIBUTION OF POWER
Baltimore and FTP
Baltimore 3-Parameter Vehicles
Power
(mphVsec)
0
20
40
60
. 80
100
120
140
160
180
200
220
240
260
280
300
320
340
360
381*
400
420
440
460
480
500
520
540
- 20
- 40
- 60
- 80
- 100
- 120
- 140
- 160
- 180
- 200
- 220
- 240
- 260
- 280
- 300
- 320
- 340
- 360
- 380
- 400
- 420
- 440
- 460
- 480 ^
- 500
- 520
- 540
- 560
Total
Baltimore
Count
477,725
308,000
215,625
149,080
100,442
64,170
38,864
22,726
13,099
7,386
4,306
2,566
1,565
1,023
582
302
187
103
72
42
20
a
7
7
0
0
0
1
1,407,908
Percent
33.9
21.9 .
15.3
10.6
7.1
4.6
2.8
1.6
0.9
0.5
0.3
0.2
0.1
0.1
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
100.0
Cum. %
33.9
55.8
71.1
81.7
88.8
.93.4
96.2
97.8
98.7
99.2
99.5
99.7
99.8
99.9
99.9
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
FTP
Count
182
154
104
49
30
11
5
5
1
3
0
0 ..
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
544
Percent
33.5
28.3
19.1
9.0
5.5
2.0
0.9
0.9
0.2
0.6
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
100.0
Cum. \
33.5
61.8
80.9
"8 9 '.'9
95.4
97.4
98.3
99.3
99.4
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
lOff.O
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
FTP Prclnninuy Report: M*y 14, 1993
131
-------
Q.
U.
c
(0
_ 53 i-
O)
-------
Figure 6*20
Cumulative Distribution of Acceleration: Baltimore and FTP
Baltimore 3-Parameter Vehicles
100.0%
90.0%
80.0% --
70.0% --
£ 60.0% .--
1 50.0% --
Baltimore
FTP
a
o*
o>
£
40.0% -
30.0% -
20.0% -
10.0% -
0.0% C
o
c
' m
/
/
Vi H- rr^ 1 1 1 1 1 1 1 1 1 I
rJ U M 1 1 | | 1 | I | 1 | 1 -J 1
> 10 it CT ««- o o T- < CM co ^f u> to r«-
O OOOOOQO
> O O Q O S *-*.*.*.«-*. ^
I*-
-------
Figure 6-21
Cumulative Distribution of Power: Baltimore and FTP
Baltimore 3-Parameter Vehicles
Ul
*>.
c
S3
CO
o
Q>
o
o
<
100.0% -I
90.0% -
80.0% -
70.0% -
60.0% -
"** M A.J
»
c
01
0)
Q.
40.0% +
30.0% --
S 20.0% --
10.0% --
0.0%
o
CM
O
CM
O
<0
o
CO
O
-------
Combinations of speed and acceleration also can be compared
in order to judge how well the FTP "covers" the range of in-use
driving. To make this comparison, it is necessary to define in
the two-dimensional sense what is meant by FTP and non-FTP
driving. Figure 6-22 displays the speed-acceleration plane with
points corresponding to all realized combinations of these two
variables. Both variables are given to an accuracy of 0.1.
These points can be said to define an "envelope" of driving, but
there currently is no generally accepted method of defining the
boundary of this region.
Typical two-dimensional displays of speed and acceleration
imply that the maximum absolute value of acceleration tends to
decrease as speed increases. This suggests one possible
definition for the FTP envelope: maximum acceleration (or
deceleration) at a given speed equals the largest acceleration
(deceleration) occurring at or above that speed. The outline in
Figure 6-22 is drawn using this definition. All points on or
inside this line represent FTP conditions; all points outside are
considered. non-FTP driving.
«*'
Figure 6-23 displays the scatter of speed-acceleration
points occurring in the Baltimore sample outside the FTP envelope
as defined above,. These points represent about 18% of total
Baltimorei driving time*; :
Two other.views of the FTP/non-FTP contrast are useful.
Figure 6-24 depicts the difference in the distribution of driving
time found in the FTP and the Baltimore in-use data. This draws
attention to those combinations of speed and acceleration that
are under- or over- represented by the FTP. The distribution of
Baltimore non-FTP driving is shown in the surface graph of Figure
6-25, which helps emphasize regions of relative importance.
FT? Prelimmuy Report: Mqr 14, 1993
135
-------
Figure 6-22
FTP Speed-Acceleration Envelope
01
G
Ul
in
x
a.
z
O
<
EC
Ul
_l
LU
O
O
60
-------
Figure 6-23
Baltimore 3-Parameter Speed-Acceleration Range
' with FTP Envelope
U)
-4
u
111
Q.
2
LU
_l
UJ
o
u
-15
-20
10
20
40 50
SPEED
80
90
100
-------
Figure
6-24
«*1-T * '
V/7'/ //////
-10 -a .1 .4 -a a ^
*««-
138
-------
Figure
6-25
139
-------
6.3.2. Trip Comparison
The FTP driving cycle was based on a driving route developed
to represent: driving patterns found on a morning commute trip in
Los Angeles, circa 1966. The original route was over 12 miles
long, considerably longer than the average trip length of 7.45
miles. In generating the FTP driving cycle, EPA shortened the
actual driving trace to 7.45 miles while maintaining the
characteristics of the longer cycle.4*
Table 6-12 compares the FTP "trip" to overall trip data for
Baltimore, as well as to Baltimore's typical morning commute
trips. The FTP is longer in duration and distance than either of
the Baltimore trip measures, although the FTP most closely
resembles the Baltimore morning trips.
Comparisons can also be in terms of the proportion of time
spent in the four operating modes: idle, cruise, acceleration,
and deceleration. The percentage of idle time is slightly higher
for Baltimore compared to the FTP. Cruise operation accounts for
a little more than a third of operating time for the FTP and on
the Baltimore trips. The differences between the in-use data and
-the FTP for the other operating modes are also minor.
4tKruse, Ronald I., and Thomas A. Hula, "Development of the Federal Urban
Driving Schedule," SA1 Paper #730553.
FTP Pretaniouy Report: Miy 14, 1993
140
-------
.:''.-.-"".:-f. £--;..,. V Table 6-12 . / .: .
Contpa'riaon of Trip Measures for FTP and Baltimore
Trip , Measure -
. " _"-. -.' "-- ' - -.
Duration (minutes)
Average ' . . ..
Median . ~ "
Distance (miles)
Average ' - -".-..-
Median ' .'-"' s-- x-.'.~ . . ' ~
Percent '.tin* in mod**
Idle
''Cruise'-"" ' ' ,; ''"'"- '" ' .-
Acceleration ' : "'..'."'
Deceleration ' '-
Average distance '.
betw««n stops ' "
FTP
. _.. ^ -
22.32
22.82
'<.
7.45
".. 7.45
-
19.58
3&.23
24.18
20.01
0.41
' Bal
All Trips
.
12.03
8.80
4.89
2.53
22.81
34.38
22.46
20.36
0.87
timore
Morning
commute trips
15.78
13.28
6.65
. "" 3.91-
24.34-
34. IS
21.89
19.59
1 .08
fTO "s" 0'; 41mile.'?'For ail-
ia the average distance between atops,
more than double^ the" FTP. distance. During the morning commute
period,'.the"corresponding 'distance grows to r.08 miles. ' '-
FTP fntmnaaj tttporc Mxy 14, 1993
141
-------
o
u
05
O5
.£
*>
-------
At 7.5 miles, the FTP is longer than the average Baltimore
trip (4.9 miles); the median Baltimore trip length indicates that
"typical" trips are even shorter, only 2.5 miles. The frequency
of stops on the FTP is also uncharacteristic of in-use trips,
with more frequent stops on the FTP. Despite these differences,
the FTP and Baltimore trips disagree only slightly in the
proportion of time in various operational modes.
6.3.3. Vehicle Soak '"""
The FTP includes two soaks; a 10 minute soak and an
overnight soak. Clearly, when compared to Figure 6-15 of Section
6.2.4, the 10 minute soak in the FTP is appropriate, as 0-10
minute soaks were the largest mode in Baltimore. However, almost
40% of the soaks in Baltimore occurred with soak periods between
10 minutes and 2 hours, the most critical period for catalyst
cooldown. To illustrate this point, Table 6-13 compares the
temperatures at trip start calculated for the Baltimore data
(from Table 6-4) with engine and catalyst temperatures calculated
from the cooldown formulas for the FTP.
The data demonstrate that the FTP substantially overstates
both the proportion of starts with hot catalysts (that is,-
catalyst temperatures abcy^lth^liglit-of f ! range) and the ,_ r >-T ;
proportion of starts ;w£^: ^id'^ehgrihes . " The "FTP
assumes that : 57% of all starts occur with hot cataiystsV^ whiie
the actual data indicate that only about 30% of all starts occur
with hot catalysts (as discussed in Section 6.2.4., the Baltimore
figures in Table 6-13 are likely to overstate the actual catalyst
and engine temperatures upon start-up). The.. FTP also assumes
that -
FTP Ptctimnsry Report: M*y 14. 1993
-- 143
-------
Table 6-13
FTP vs. Baltimore Temperature Distribution at Trip Start
Catalyst
Temperature
Rang* (*F>
800+
700-800
600-700*
500-600
400-500
300-400
200-300
<200
Frequency (%)
. Bale.
22
9
7
6
5
S
6
40
FTP
57
--
. -
43
Engine
Temperature
Range (°F)
180+
160-1SO"
140-160
120-140
100-120
<100
Frequency (%)
Bait.
51
"1
6
6
7
23
FTP
57
43
Ctulyit light-off rug*
43% of all starts occur with cold engines, while the data'suggest
that less than a quarter of all starts occur with engine
temperatures within 25°F the ambient temperature.
From an overall emission inventory point of view, the
overstatement of hot catalysts and cold engines in the FTP are
offsetting factors which may largely cancel each other out.
However, while the 10 minute and overnight soaks in the FTP
represent the most frequent modes of soak periods, they offer no
incentive for manufacturers to delay catalyst cooldown.
Catalysts remain above light-off temperatures after a 10 minute
soak and it is not practical to delay cooldown enough to obtain a
benefit after an overnight soak. Thus, there may be a "lost
opportunity" for inclusion of an intermediate soak time in the
FTP. Such an intermediate soak time, if it proved to be
FTP Pnlxnxury Report May 14, 1993
144
-------
feasible, would offer an incentive to.improve catalyst light-off
times or delay catalyst cooldown (which could be done -with
catalyst insulation). Thus, the feasibility and potential
emission benefits of an intermediate soak period should be
included in future emission assessments.
6.3.4. Trip Start Driving'Activity
As indicated earlier, the FTP, overall, has lower speeds and
is less aggressive than actual driving behavior. However, the
reverse occurs during the first few minutes of driving after
vehicle starts. Both the FTP and actual driving begin with an
initial idle period (20 seconds for the FTP and variable for
actual driving, depending on the type of start and ambient air
temperature). The FTP then goes into a 105 second micro-trip
with an average speed of 23.1 MPH. The average speed during the
first 80 seconds of actual driving after the initial idle is only
14.4 mph (as reported in Section 6.2.5). The FTP follows the
first micro-trip with a 39 second idle, then launches into the
most aggressive section of the entire cycle, reaching speeds over
» 50 mph 200 seconds after the initial idle. The actual driving
activity during the second and third 80-second phases averages
about 23 mph, including 27 seconds of idle, and only has about 6%
of total operation above 45 mph.
A comparison between the FTP and the actual driving activity
from the Spokane and Baltimore data (combining cold, warm, and
hot starts), both broken down into 80-second phases, is presented
in Table 6-14.
FT? Prtlisuoiry Report M»y 14, 1993
145
-------
Table 6-14
Comparison of FTP and Actual Start Driving Behavior
(by 80-second phases after initial idle)
Phaaa l
Accual
FTP
Phase 2
Accual
FTP
Phase 3
Accual
FTP
Speed
Avg
14.4
22.7
21.6
11.9
24.1
46.0
St Dav
12.3
S.9
15.3
13.2
16.6
12.1
Acceleration
Avg
.24
.38
.00
-.06
-.01
.35
St Dav
1.82
1.15
1.6S
1.51
l.SS
1.03
Power
Avg
40.7
34.1
46.9
39.9
45. 7
68.2
St Dev
43.1
43. S
43.0
% Tims
at
Idle*
13. S
0.0
17.3
49.0
16.1
0.0
«N«* including th* initial idte tim*.
These 80 -second phase comparisons are skewed by the fact
that all of the idle time during the first 240 seconds of the FTP
(after the initial idle) occurs in one long 39 -second segment
that falls in phase 2, while idle time is distributed across all
3 phases of actual driving. Thus, a more meaningful comparison
would be to compare the first micro- trip on the FTP (Hill 1} with
phase 1 of the in- use driving and to compare the next section of
the r^P (106 to 240 seconds after the initial idle) with phase 2
and 3 combined. As the first micro- trip on the FTP is already
longer than the first phase of actual driving (105 seconds v.
30) , the 39-second idle should be assigned to the second section
of the FTP to keep the initial driving comparisons as compatible
as possible. These comparisons are presented in Table 6-15.
FTP PrcUccunuy Report: Miy 14, 1993
146
-------
Table 6-15
Revised FTP and Actual Start Comparison
Phase 1 - Actual
Hill 1 of FTP
Phase 2+3 Actual
2nd Phase of
FTP**
Speed
Avg st Dev
14.4 12.3
22.8 15,9
29.8 22.6
Acceleration
Avg St Dev
.24 1.32
.00 1.45
.00 1.60
.40 1.06
Power
Avg St Dev
40.7 43.1
31.8
46.3 43.3
64.8
% Time
at
Idle*
13. S
1.0
16.7
28.1
Not including the initial idle period.
**SU(t* 106 seconds «fter the initial idl* ind endt 240 seconds after the initial idle. Include* the 39-second idle between the let 2 micro
trips.
The data clearly show that the speeds on the FTP are much
higher than encountered in-use after a vehicle start. Figure 6-
27 presents the speed distributions. The solid lines compare
Hill 1 of the FTP to phase 1 of in-use driving; the FTP has far
less driving in the 0-15 mph range and much more in the 15-35 mph
range. The dashed lines compare the 2nd section of the FTP to
phase 2 and 3 of actual driving combined; the FTP has less
driving in both 0-15 mph and 30-40 mph ranges, plus far more in
the 45-60 mph range.
The aggressiveness of the driving is also skewed, with the
first micro-trip on the FTP less aggressive than Phase 1 and the
second micro-trip on the FTP clearly overaggressive. Figure 6-28
compares the cumulative power distributions (that is, the amount
of driving occurring at or below a given power level). Hill 1 of
the FTP spends less time at the higher power levels (i.e. the
line rises faster) than Phase 1 of actual driving, while the 2nd
section of the FTP spends far more time at the higher power
levels. The speed and power data clearly show that the FTP
FT? Pnlimtasiy Report; May 14, 1993
147
-------
Figure 6-27.
FTP v. In-use: Start Speed Distribution
30.00
0.00
- Phase 1
Phase 2+3
FTP-Hili 1
- - - FTP- 2nd
tO 15 20 25 30 35 40 45
Speed (Maximum of range - MPH)
50
55
60
65
-------
Figure 6-28.
FTP v. In-use: Start Cumulative Power Distributions
100.00 T-
0.00
20
40
60
BO 100 120 140 160
Power (Maximum of Range)
780
200
220
-------
contains speed/acceleration events that are seldom encountered
in-use during the first 240 seconds of operation.
The impact of driving behavior immediately after a vehicle
start on exhaust emissions is subject to interpretation. Higher
speeds and loads increase the mass flow through the engine, which
would cause more engine out emissions to pass through the
catalyst unconverted during warm or cold starts. On the other
hand, the engine and catalyst heat up faster during higher mass '
flows, decreasing the length of time needed for fuel enrichment
and catalyst light-off. Compounding the assessment is the fact
that the first micro-trip of the FTP has much higher speeds than
are encountered in-use during a similar period after a start, but
is also less aggressive. The overall impact of these offsetting
factors should be investigated in future emission assessment and
testing programs. Of course, for compatibility with cycles
developed for hot stabilized operation, any start cycle developed
for use in emission assessments should be baaed only on the
Baltimore 3-parameter data.
Initial idletime
Table 6-IS presents the initial idle times for
Spokane/Baltimore (combined), Atlanta, and the FTP.
FTP PreUmxuuy Report: Mijf 14, 1993
150
-------
Table 6-16.
Initial idle Times
Start.
Typ«
Cold Stare
Warm Start
Hot Start
City
Spokane /Baltimore
Atlanta
FTP
Spokane/Baltimore
Atlanta
FTP
Spokane /Baltimore
Atlanta
FTP
Mean
(seconds)
59
28
20
28
18
»/A
IS
12
20
Median
(seconds)
13
9
20
a
7
N/A
5
5
20
The data show that the initial idle time after a cold start
on the FTP falls between the mean and median initial idle times
for Atlanta. Thus, the cold start initial idle time appears to
be reasonable. However, the initial idle time after a hot start
on the FTP is substantially longer than the mean initial idle
time in Atlanta, and four times as long as the median time in all
the cities.
Proportion of Qrj.vinqr Occurring After Starts
Figure 6-29 compares the distribution of miles driven since
the start of the trip on the FTP to the actual distribution. As
the FTP is one trip of 7.5 miles, its distribution appears as a
straight lin* up to 7.5 miles. Due to substantial numbers of
trips
FTP Prtlanauy toport: Mqr 14, 1993
151
-------
Figure 6-29.
FTP v. In-Use: Distribution of Miles Driven
m
10
100 T~
90
34567
Miles Driven from Start of Trip
10
-------
longer than 7,5 miles in-uae, the actual distribution shows a
large discrepancy in the range of 5-10 miles from the start of
tha trip. However, this is not likely to be of concern from an
emission point of view, as the engine and catalyst are fully
wanned up and stabilized aft'er 1-3 miles of operation. Whether
the driving after this point occurs 3-7.5 miles from the start of
the trip o,r extends out to higher mileages should not impact the
level of hot stabilized emissions.
The differences between the FTP and actual driving less than
one mile from the start of the trip are of much more interest, as
this relates to the amount of driving that occurs before
catalysts or engines have warmed up. While these differences are
much smaller in terms of absolute frequency than the differences
at higher miles, the relative differences are substantial. Table
6-18 compares the FTP distribution of the first mile driven from
the start of the trip to the distribution recorded in Baltimore.
Table 6-17
FTP v. Baltimore Comparison of Miles Driven from Start of Trip
Mileage Range from
Start of Trip
0-0.25
0.25-0.50
O.SO-0.75
0.75-1.00
*
0-0.67
(Equivalent to Hill 1
of FTP)
FTP
Proportion
3.33 %
3.33 %
3.33 %
3.33 %
8.93 %
Baltimore
Proportion
4.78 %
4.42 %
4.09 %
3.81 %
12.01 %
In-Use
Increase (%)
43 %
. 33 %
23 %
14 %
34 %
FIT Preliminary Report; May 14, 1993
153
-------
It is clear that the FTP substantially underestimates the
amount of time vehicles actually spend in the initial portions of
a trip.
.3.5. Road Grade
The current FTP speed-time trace waa*. developed without
attention to the effects of road gradient. The original road cycle
in Los Angeles contained at least one freeway on-ramp (Harbor
Freeway, southbound) , which was no doubt approached at some rate of
acceleration, yet the transcription of the road cycle to a
dynamometer trace assumed a flat surface. Thus, the additional
engine loads caused by operation on freeway on-ramps and other
gradients are not adequately represented.
Regardless of the current limitations of road gradient data,
a comparison of in-use driving with the FTP is relatively easy: the
FTP is flat, the world is not. The DOT data presented in Table 5-1
illustrates, in nationwide terms, that fully 30% of VMT are
traveled on gradients greater than 2 percent, and almost half occur
on gradients greater than 1 percent. It is quite clear that
emission effects due to road gradient are not represented on the
current FTP and that this should be an area of concern for the
Agency, but it is also clear that more work is needed to better
identify these effects and develop a better understanding of the
amount of driving that occurs in urban areas on various road
gradients.
FTP Preta»«r lUpcxc MIT K. 1993
154
-------
Chapter 7. Teat Cycle Development Methods and Approach
7.1. Test Cycle Methods
A driving cycle is a time series of vehicle speeds occurring
at successive (equally spaced) time points. For emission testing,
a test driving cycle attempts to synthesize real driving conditions
with respect to a number of measures, including speed,
acceleration, specific power, trip patterns, road grade, and
temperature. As a practical matter, the cycle must be of
reasonable length and it must be "driveable," that is, consist of
speeds and accelerations that can be physically realized by
vehicles running on a dynamometer.
Following review of the adequacy of the current Federal Test
Procedure, EPA will determine whether the driving cycle represented
by the FTP should be modified to ensure that future testing better
represents actual in-use driving. The comparison of the FTP with
results of the driving behavior research presented in Chapter 6 o'f
this report will be a central consideration in developing and
judging new test cycles. A full assessment will require combining
these pure driving measure comparisons with emission test data.
j^^^^n^rj^^
~" an^l pracititioners
-. These are
briefly reviewed!;'"; in the current section. In the following
sections, development of a sample cycle is discussed, together with
EPA 's plans for new cycle generation.
Drive cycle development aims to satisfy two general goals of
vehicle emission testing; emission control and emission inventory.
Emission control strategies are designed to limit emissions
occurring across the range of actual driving behavior and
FTP Prdmauury Report: M»y 14, 1993
155
-------
conditions. Emission inventory estimates are obtained by combining
vehicle test results with fleet and mileage information. This
distinction has an important implication for test cycle
development. A cycle" used to achieve emission control does not
necessarily have to be representative of overall in-use driving.
It may artificially emphasize aspects of driving behavior most
critical to emission control.
7.2. Mathod Types
The report bibliography cites previous work on drive cycle
method and development. Most of this research is of an applied,
empirical nature, with little substantial theoretical basis.
Method that have been studied fall into two general categories:
segment-splicing and Monte Carlo simulation. A third method might
be called the "engineering" approach, whereby a cycle is developed
using criteria that do not necessarily include the frequency of
occurrence of in-use driving patterns.
7.2.1. Segment-splicing Methods
Probably the most widely accepted approach to cycle generation
is to select and splice together segments of observed speed-time
trace data that satisfy some set of driving behavior criteria. The
current FTP was created by applying this method to speed-time data
collected during morning rush-hour driving in Los Angeles,
California.4*
Perhaps the greatest appeal of the segment-splicing approach
lies in its use of real driving sequences that are known to be
driveable (although it is not necessarily true that all on-road
4*iCru»e, R.I. and T.A. Hul», "Development of the Federal Urban Driving
Schedule,1* SAB Technical Paper Series, 730553, 1973.
FIT fttinuuiy ftaport: Miy 14,1993
156
-------
driving can be replicated on laboratory dynamometers) . This method
does, however, raise a number of questions and problems. The most
obvious issue is the selection of segments in order to adequately
meet cycle requirements. In practice, this has been performed by
sampling in a variety of ways, at random or otherwise, from an in-
use database.
>
Once a set of segments is chosen, they must be connected to
produce a single cycle^ If the segments are full "microtrips"
(extending from one idle to another) , this splicing operation is
straightforward and is nr,;: likely to jeopardize the driveability of
the full cycle. Howev^ i-t sometimes is desirable to use only
- ^ * - £7' *,. '
parts of microtrips (cal_"i^" kinematic sequences or modal segments)
which begin and/or end at non-zero speed. Connecting such segments
must be done more carefully in order to achieve realistic driving
behavior.
., v- .-'.-'
7.2.2. Monte Carlo Simulation
Simulation provides an alternative method for generating a
drive .cycle. In a Monte Carlo simulation, a model is established
for describing how actual driving occurs^. The model includes both
deterministic aind probability, or^ "stochastic , ": elemenjta to account
"f^^^H^i^ -
~ _driyixig ~
~ '" '
:-- -» -~.::~^- - ___ ----~. "-';r?-- .--- --
characteriatlca resembling those of the underlying model.
The principal challenge in applying this technique is to
create a suitable model. The typical model will be estimated by
fitting a set of assumptions to in-use data. For example, in
FTP PrdJmnuy Report: Miy 14, 1993
157
-------
developing EPA'a heavy duty test cycle,30 second-to-second speed
transitions were assumed to obey a "Markov" process, in which the
probability of a given speed at the end of the second depends only
on the vehicle's speed at the end of the previous second. These
conditional probabilities were estimated from in-use driving data
frequency distributions. However, evidence exists that the model
used in that work is «pot sufficiently complex to capture all
important emission-related elements of real driving, such as the
duration of acceleration events. Even the current large sample of
second-by-second data *!say not permit accurate estimation of
parameters needed to describe such behavior.
f
The Monte Carlo ?.,- .lation approach offers considerable
flexibility in modeling driving behavior and generating cycles. It
is less popular than the segment-splicing method for the reasons
given-and because resulting cycles are "artificial" - the second-
to-second sequences have not actually been driven in-use. This
raises the fear that a simulated cycle may not be driveable or, in
emission testing, may yield emissions that do not or cannot occur
in real driving.
7.2.3. Engineering Cycles
An engineering cycle is a speed-time trace that satisfies
certain criteria not generally based on the frequency distribution
of in-use driving behavior. For example, it may be the designer's
judgment that a cycle should include certain extreme levels of
speed, acceleration, or specific power. These objectives can be
achieved without-reference to in-use data, possibly using simple
"straight-line* construction methods.' This approach to cycle
50Smith, M. Heaw-Dutv Vehicle Cvele Developiqei|t. ^-s- Environmental
Protection Agency (EPA-460/3-78-008), 1978.
FT? PrtUmjoiry Report: M«y 14, 1993
158
-------
development is useful as a way of forcing certain conditions that
may occur Infrequently, or to implement feasibility testing.
7.3. Cycle Criteria
Most applications of cycle methodology have attempted to
represent the full range of second-to-second driving behavior; that
is, all combinations of speed and acceleration likely to occur in-
use. In some circumstances, it may be desirable to construct a
cycle that overrepresents various types of driving. In particular,
in considering modifications to the FTP, a cycle with a
disproportionately high amount of non-FTP driving has potential
advantages. This poses special problems, especially for the
microtrip-splicing method, because it may be difficult to find
largely non-FTP micro trips that are sufficiently short in length to
obtain a cycle of realistic duration. In splicing of modal
segments or Monte Carlo methods, this is less of a concern.
7.4. Cycle Validation
Once a cycle is generated by either the segment-splicing or
Monte Carlo method, it is necessary to test how well it satisfies
the objectives of the problem. Usually, this involves matching
various characteristics of the cycle to comparable features of in-
use driving. The features may include the summary statistics and
distributions, used in Chapter 6 to describe the Baltimore 3-
parameter instrumented vehicle sample. A cycle is considered valid
if its statistical properties are reasonably close to those of the
data supporting its generation. In practice, this involves
selecting a set of arbitrary statistical tests (or "filters"). A
cycle that passes these tests at a desired level of significance
becomes a candidate for testing. This process is normally followed
by a more qualitative evaluation to assure selection of a realistic
cycle.
Mqr 14.19W
159
-------
7.5. Current BPA Teat Cycle Development
As part of the current study of non-FTP driving, EPA has
undertaken the development of new test cycles using the survey data
described in this report. This effort includes reviewing cycle
development methods and cycle validation techniques and programming
of cycle generation algorithms. Through two of its contractors,
the Agency currently is engaged in creating cycles for use in
emission testing and as candidates for new regulation, pending
final conclusions from this study.
FIT tnbua*rr JUpcrt Mqr 14, 1993
160
-------
List of Appendixes
The following appendixes are presented in their order of first
occurrence in the text of the report.
Appendix A.
Appendix B.
Appendix C.
Appendix D.
Appendix B.
Chase Car Method Bias Analysis: Supplementary Tables
Baltimore 3-Parameter Vehicle Characteristics
Summary Statistics, Distributions, and Graphs
Baltimore Soak Periods
Trip Start Driving Activity
FTP Prelimauy Report: M«jr 14, 1993
161'
-------
-------
Appendix A
Chase Car Method Bias Analysis: Supplementary Tables
Table A-1. Comparison of chase car route characteristics for Baltimore
Trip type
Home-based
nonwork trips
Home-based
work trips
Non-home
based trips
Intrazonal
trips
Number of trips
Master data
set
2,507,325
1,789,950
1,117,031
638,688
Selected
124
89
55
32
Driven
93
62
42
20
Average Trip length (miles)
Master data
set
7.8
11.1
7.9
. NA
Selected
7.6
10.6
7.0
NA
Driven
7.2
* 9.8
7.4
2.1
Table A-2. Comparison of chase car route characteristics for Spokane
Travel
Period
Interzonal
AM Peak
PMPeak
Off Peak
Intrazonal
AM Peak
PMPeak
Off Peak
Number of Trips
Master data
set
103,996
139,273
1,228,423
5,017
9,608
112,441
Selected
20
26
231
1
2
20
Driven
18
23
186
2
2
18
Average Trip length (miles)
Master data
set
7.6
6.5
6.3
NA
NA
NA
Selected
7.9
6.8
6.5
NA
NA
NA
Driven
7.1
5.2
6.4
NA
NA
NA
Note: The local planning agencies in Baltimore and Spokane did not have estimates of intrazoflal trip
lengths.
A - 1
-------
Table A-3. Target Vehicle Characteristics for Balitmore and Spokane
Vehicle
Characteristic
Vehicle Typ»
Luxury, sadan
Pick-up, van, utility
Spores
Hot determined
Vehicle Age
Pre-1979
1980-1989
1990 +
Not datermined
Continent of Origin
Asia
Bur op a
Domestic
Not determined
Manufacturer
CMC
Ford
Chrysler
Nissan
Toyota
Honda
Mazda
Not determined
Baltimore
Number
559
167
37
IB
32
464
248
37
179
53
495
43
226
181
75
34
50
52
15
148
Percent
72
21
5
2
4
59
32
5
23
7
64
6
29
23
10
4
6
7
2
19
Spokane
Number
552
307
12
9
125
527
213
15
203
46
612
19
319
173
117
40
68
43
24
96
Percent
63
35
1
1
14
60
24
2
23
5
70
2
36
20
13
5
8
5
3
11
A-2
-------
Appendix B
Baltimore 3-Param»ter Vehicle characteristics
Sample
t**t Sit* Number Vehicle ID Numb«r
1FTOUOA1HUC48311
7VUX8B5XJ1793960
lClrP21S3IOJ.04««9
Y33At75L8N700*092
10irP21«lJL19318«
2P4IB413XJR««2190
1P3BK4«D2KC4*98«9
lQiU*983D92C6i34
10CIQ25BXC7125742
1HQC87CC1MA0495C1
102AB2705H7332521
ieiJVll*2K7180*ri
1PAPP14J2MH2.9094
1XVC031BSM51387Q*
irrcRiOAiouK8ioo4
1Q1JC1113KJ171405
162AB2707I72S03C4
Mod«l
1
2
3
4
5
£
7
8
9
10
11
12
13
14
IS
16
17
18
19
20
21
22
tp 23
L 24
25
26
27
28
29
30
31
32
33
34
35
3C
37
38
39
40
41
42
43
44
45
4C
Roaavill*
Roaavill*
Roaavill*
Roaavill*
Roaavill*
Roaavill*
Roaavill*
Roaavill*
Roaavill*
Roaavill*
Roaavill*
Roaavili*
Roaavill*
Roaavill*
Roaavill*
Roaavill*
Roaavill*
Roaavill*
Roaavill*
Roaavill*
Roaavill*
Roaaville
Roaavill*
Roaavill*
Roaavill*
Roaaville
Roaavill*
Roasviil*
Roaavill*
Roaavill*
Roaavill*
Roaavill*
Roaavill*
Roaaville
RoaavilU*
Roaavill*
Roaavill*
Roaavili*
Roaavill*
Roaavill*
Roaavill*
Roaavilia
Roaaville
Moaavill*
Roaaville
Roaaville
B002
B010
B016
B017
BO 19
B037
B039
B042
BOSO
BOS2
B058
BOS9
B0«2
B0££
BO 72
B081
B084
B086
B088
B098
BIOS
B123
B124
B130
B139
B143
B1SO
B153
B15S
B163
B164
B169
B178
B184
B190
B191
B193
8195
B206
B208
B210
B233
B236
B243
B247
B255
1GCDC14I8KZ211S65
10TBS14R2QS533452
JT4RM50R7ClOn3173
fS725009680
1B3BP48D5KH589752
1FABP44K2KF117098
164U.1937PQ4C339C
JT2RA44C6B0001018
1B3YA44K6JG4485 4(
1GMCT18X3K0152538
l
-------
ix B (continued)
Baltimore 3-f*ram«t»e V»hlcl« Characteristics
f
Saopl*
T«»t ait* Humb«c V*bicl« ID Huaixr
1G1AM3SK5CR22SC18
1B7B01417BS461S22
iriciuiTijai262So
JT2AL3107E0225492
CRM1498264964
1237(798412061
lHKBP9237GBt54018
1G3OC69B9G4332S06
lfTSX15a2GU941S8
8G87B1C9118
lFABP259xaNllG408
JT4RM50R7H032S017
Mod»l
47 Ez«t«ir
48 Ezat*C
49 Er.«t«r
SO Ez*t*c
SI Ex*t«C
52 Ex*t*c
S3 Kx«t«r
54 Rx«t«r
SS Ez*tac
56 Sx*t*c
57 Ex*t«c
58 Ex*t*c
59 Rr.«t»c
60 Ex»t«r
61 Ez*t*c
62 EX*t*r
63 Ex*t*i:
64 Ex*t*c
65 Ex*t*c
66 Ex*t«r
67 Ex*t*r
68 Ez»t«r
69 E£*t*t
70 Ez*t*t
71 Bi«t«r
72 E**t*r
73 Bx*t*c
74 Br.at«c
75 Ex*t*c
76 Er*t«t
77 Ex*t*c
78 Ez*t*r
79 Ex*t*r
80 Exate£
81 E£«t«r
2 Exat*c
3 Bxat*r
84 Ez*t*r
SS Sxetec
86 Ez.t.r
87 Ex*t*c
88 Ex«t*r
» Ex.t.c
90 Er«t«r
91 Er*t«c
92 Bz*tar
93 Ex«t*c
B268
S269
B273
B275
B286
B287
B297
B314
B31S
B317
B319
B325
B329
B337
B338
B344
B3S1
B361
B365
B367
B368
B369
B371
B37S
B376
B381
B386
B389
B39S
B406
B410
B413
B419
8420
B436
B43B
B441
B44S
B447
B451
B452
B460
B466
B467
B468
B472
B482
JT2HX62EXI001 4430
1C6AD478XC91S0406
1Y1SK5167LI072182
1S4U47A8EB477206
JT2AE92N6J3076220
J12AE83E403279911
1C2JB6903Q7338623
IQ«BUr31J3Ja420717
082AE72SXD2040393
1M6MD06SOBC332168
2GlAN3SX3alll41l(
JN1HB1155C0006220
!G3B7373Sai874913
JB2A882E1G3308647
JT2TE72N2C5109018
JKlPB1254rD631495
lG2NE27l.irC74t917
JM2UT1135J0369027
4J47AAC110524
4J47NAG128222
1AB08C3DI262608
4P3ca34TXlttl09053
JT2IL3laOHOO«9»50
lalAB08CSEA12S860
JT4YR2»a5V5007140
la4AMS9A1CB143803
1P3BF36K4HT153374
KMCU06D2LT200403
lfABP46r6D4168830
JT2AE72EXD2098772
lBaBA743XOAOC3308
JT2BL31D5H0075958
1P3BX18C6BD304919
1C3CJ51E3HG185086
Chevrolet
dodge
ford
TOYOTA
CHEVROLET
CHEVROLET
MERCURY
OLDSHOBXLE
FORD
fORD
fORD
TOYOTA
CHEVROLET
toyota
CADILLAC
GEO
buick
TOYATA
TOYOTA
rOMTUC
BTfUUDAI
TOYOTA
ni**an
Chevrolet
NISSAN
OLD3MOBILB
TOYOTA
TOYOTA
NISSAN
PONTIAC
MAZDA
BOICK
BOXCK
Chevrolet
tLXMOQTB
TOYOTA
CHEVROLET
TOYOTA
MXCX
PLYMOUTH
POMIIAC
fORD
TOYOTA
BOMDA
toyota
PLYMOUTH
CHRYSLER
Halibu_Cla»»ic
cam *"
Ranger .
TERCEL
LDV
HONTECARLO
COOOAR
98
F150
TBOHDERBIRD
ESCORT
PICKUP
MOMTBCARLO
cre»«ida
COOPDEVILLE
PRIIM
cega
COROLLAMAaON
CAROLLA
30KBIRD
EXCEL
CAROLLA
pickup
celebcitygl
SENflUk
88
COROLLA
COROLLA
3EMTRA
GRAND AM
b2000
REGAL
REGAL
Chevette
LASER
TERCEL
CBEVET
vm
REOAL
RELIANT*
TRANSPORT
THDMDERBZRD
COROLLA
ACCORD
tercel
HORIZON
LEBARON
V»blcl«
Luxucy/S«dan/Statlon Hagon
Pickup/Van/oti li ty
Pickup/Van/Otiiity
Luxury/3«dan/Station Wagon
Pickup/Van/Otility
Luxury /8«dan/3 tat ion Wagon
Luxury/S»dan/atation Magon
Luxucy/3*dan/statlon Magon
Pickup/Van/Otility
Luxury/s«dan/Station Wagon
Luxury/3»dan/3tation Magon
Pickup/Van/Otility
Sporta Car/High P«r£ormanc«
Luxury/3«d»n/3tation Hagon
Luzury/S«dan/3tation Magon
Luxury/3«dan/station Magon
Ioixury/3«d»n/Station Magon
Lurury/3«dan/Station Magon
Luxucy/S«dan/Station Magon
Luxury/S«dan/3tation Nagon
Luxury /3»d»n/3 tat ion Magon
Luxury/S*d*n/Station Magon
Pickup/Van/Otility
Lurury/S»dan/3tation Magon
Luxury/3«dan/3tation Magon
Luxury /S*dan/Station Magon
Luxury /S«dan/Stat ion Magon
Luxury/3«d»n/3tation Wagon
Luxury/S«dan/3tation Hagon
Luxury/S«dan/3tation Nagon
Pickup/Van/Otility
Luxury/3«dan/3t*tion Nagon
Luxury/Sadan/3tation Nagon
Luxury/Sadao/Station Wagon
Luxury /Ssdir:/ Station Nagon
Luxury /dan/Station Nagon
Luxury/S«dan/3tation Nagon
Pickup/Van/Otility
Luxury/a»d*n/3tation Nagon
Luxury/Stdau/Statiop. Nagon
Pickup/ Van/at i H ty
Luxury/3«d*n/3tation Nagon
Luxury /3«d«n/3t at ion Wagon
Luxury/8«dan/Station Nagon
Luxury /3»dan/Stat ion Nagon
Luxury/3«d»n/3tation Wagon
Luxury/S*dan/Station Nagon
Mod«l
Transmission
1982
1987
1988
1984
1979
1979
1986
1988
1986
1978
1987
1987
1987
1981
1982
1990
1984
1988
1986
1986
1988
1983
1984
1986
1982
1986
1986
1982
198S
1985
1988
1980
1980
1983
1991
1987
1984
1986
1982
1987
1990
1983
1983
1986
1987
1984
1987
Automatic
Automatic
Automatic
Hanual
Hanual
Automatic
Automatic
Automatic
Automatic
Automatic
Automatic
Hanual
Automatic
Automatic
Automatic
Automatic
Automatic
Automatic
Manual
Automatic
Automatic
Hanual
Automatic
Automatic
Manual
Automatic
Automatic
Automatic
Manual
Automatic
Manual
Automatic
Automatic
Automatic
Manual
Manual
Automatic
Manual
Automatic
Automatic
Automatic
Automatic
Automatic
Automatic
Manual
Manual
Automatic
-------
Appendix C
Summary Statistics, Distributions and Graphs
"' Table C-1
SUMMARY STATISTICS
ALL DRIVING & VEHICLE MOVING
Ba^imore 3-Parameter Vehicles
Speed (mph>
AIL
Mean
St Dev
Max
Min
Count
Driving
24.50
- 20.52
94.46
0.00
3,365,504
Vehicle Mffvina
31.06
18.17
94.46
0.02
2,654,614
Accel (mph/sec)
Mean
St Dev
Max
Min
Count
All Driving
- 0.00
1.50
15.19
-19.49
3,360,550
Vehicle Moving
0.03
1.67
15.19
-19.49
2,654,298
Power (mphA2/sec)
Mean
St Dev
Max
Min
Count
All Driving
46.02
42.96
557.69
0.00
1,407,908
Vehicle Moving
46.02
42.96
557.69
0.00
1,407,908
C-1
-------
Table C-2
DISTRIBUTION OF SPEED -
AH Driving and Vehicle Moving
Baltimore 3-Parameter Vehicles
Speed
(mph)
0 - 0
0 - 5
5 - 10
10 - 15
15 - 20
20 - 25
25 - 30
30 - 35
35 - 40
40 - 45
45 - 50
50 - 55
55 - 60
60 - 65
65 - 70
70 - 75
75 - 80
80 - 85
85 - 90
90 - 95
Total
V
All IT" ig
Count
710,890
207,698
181,442
201,708
209,420
231,917
280,040
308,521
253,681
173,105
125,294
121,004
147,398
125,917
66,826
17,485
2,803
253
91
11
3.365.504
"*'
Per -?
21,1%
6.2%
5.4%
6.0%
6.2%
6.9^
8.3%
9.2%
7.5%
5.1%
3.7%
3.6%
4.4%
3.7%
2.0%
0.5%
0.1%
0.0%
0.0%
0.0%
100.0%
Cumulative
Percent
21.1%
27.3%
32.7%
38.7%
44.9%
51.8%
60.1%
69.3%
76.8%
82.0%
85.7%
89.3%
93.7%
97.4%
99.4%
99.9%
100.0%
100.0%
100.0%
100.0%
Vehicle Moving
Count
0
207,698
181,442
201,708
209,420
231,917
280,040
308,521
253,681
173,105
125,294
121,004
147,398
125,917
66,826
17,485
2,803
253
91
11
2.654.614
Percent
0.0%
7.8%
6.8%
7.6%
7.9%
8.7%
10.5%
11.6%
9.6%
6.5%
4.7%
4.6%
5.6%
4.7%
2.5%
0.7%
0.1%
0.0%
0.0%
0.0%
100.0%
Cumulative
Percent
0.0%
7.8%
14.7%
22.3%
30.1%
38.9%
49.4%
61.1%
70.6%
77.1%
81.9%
86.4%
92.0%
96.7%
99.2%
99.9%
100.0%
100.0%
100.0%
100.0%
C-2
-------
Table C-3
DISTRIBUTION OF ACCELERATIONS
All Driving and Vehicle Moving
Baltimore 3-Parameter Vehicles
Accel
(mph/s]f
- -20 to -19
-19 to -18
18 to -17
-17 to -16
-16 to -15
-15 to -14
-14 to -13
; '--/ ; -13 to "-12
-12 to -11
. 7 "-11 to -10
-10 to -9
... ' ' ... -. . : -9 "to -a
-:7:"t- .:, '-_ .8 to -7
;.''. " -- - '; ' -7. to "-6
-6 to -5
,'V ' -. .':.-. - >s to -4
'' ' '""'? - \' -4 to -3
"-..-.' ' -3 to -2
; '..:" 7" -:' '' _- -2 tb-1
..:. --- -v.-v ';:-7-1~ to 0
- ''""' :-'' - ' ^ ^,t!7;">7' -7f ^9
ri=j^SSs^^^^^^§^^
.C ^r~<«~~-~'~'~ ^JjTI ^jr^i(^,yS '^Mt.-'jt^iMi^lft'
~':"'"' ;'"~^!2l;tSiJ
_ \ ^^Tl^lfe?
''" . --" ~:9'rto'iO
10 to 11
11 » -12
12Jo 13
13 to 14
14 to 15
15 to 16
Total
All
Driving
Count Percent
1
0
0.
0
o
3
' --. '..- 2.
. tl
- '"-'a*
. v " '4'4 .'--
149
".." 54S-"
1,801
; 5.7SS
17,199
41-,575"
78,005
120,151
219,036
767,792/
^,:r,70.2J.3I.C,l.
0.0%
0.0%
0.0%
0.0%"
- 0.0%
0.0%
0.0%
0.0%
0.0%
. 0.0%
0.0%
0.0%
0.1%
02%
0.5%
' -.',t2%
2J%
3.6%
' 8J%
22.8%
_JO;9M
AJr;fig^i'.2S^
i^r^-t^iiit-:"
~--""::; "~77((v -.
:;r":-:" ;-3:25' '
.-\^. 187
-. ' 62
20
8
'" ' 2,.
. i
3,360.550
zrf&L
"i. itfcl^fe;
0-0*
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
100.0%
Cumulative
Percent
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
:. 0.0%
0.0%
0.0%
0.1%
0.2%
.0.8%
2.0%
4.3%
7.8%
14.4%
372%
);^-.5jM%
fe3^*
'- _ -\99.9%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
Vehicle Movlrm
Count
1
0
0
0
0
- 3
2
11
30
42
145
533
1,753
5,611
18,704
40,176
72,353
112.253
205,116
758,565
33,092
iH
'^:"^S[
3,712
1,849
778
325
187
62
20
8
2
1
2.654,298
Percent
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
. 0.0%
. 0.0%
0.0%
0.0%
0.0%
0.1%
02%
0.8%
1.5%
2.7%
42%
7,7%
28.6%
'" :'.;r12%
8
^---(x*&
0.1%
0.1%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
100.0%
Cumulative
Percent
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.1%
0.3%
0.9%
2.4%
5.2%
9.4%
17.1%
45.7%
47.0%
80.9%
90.9%
98.0%
98.3%
99.3%
99.7%
99.9%
99.9%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
C-3
-------
Table C-4
DISTRIBUTION OF POWER
All Driving
Baltimore 3-Parameter Vahicias
Power
(mph*
0
20
40
60
80 -
100 -
120 -
140 -
130 -
1SO -
200 -
220 -
240 -
260 -
280 -
300 -
320 -
340 -
360
380 -
400 -
420 -
440 -
460 -
480 -
500 *
520 -
540
2/s)
- 20
- 40
- 60
- 80
too
120
140
160
180
200
220
240
260
280
300
320
340
360
380
400
420
440
460
480
500
520
540
560
Total
AH Driving
Count
477,725
308,000
215,625
149,080
100,442
64,170
38,864
22,726
13,099
7,388
4,308
2.566
1,565
1,023
582
302
187
>S03
72
42
20
8
' 7
7
0
0
0
1
1,407.908
Percent
33.9%
' 21.9%
15.3%
10.6%
7.1%
4.6%
2.8%
1.6%
0.9%
0.5%
0.3%
0.2%
0.1%
0.1%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
100.0%
Cumulative
Percent
33.9%
55.8%
71.1%
81.7%
88.8%
93.4%
96.2%
97.8%
98.7%
99.2%
99.5%
99.7%
99.8%
99.9%
99.9%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
C-4
-------
£
Speed
>
0
5
10
15
20
25
30
35
40
45
50
55
60
65
70
75
80
15
90
ALL
>
<-
nf*>
<,
0
5
10
15
20
25
30
35
40
45
SO
55
60
65
70
75
80
85
90
95
10
ft
0400
0400'
0.000
0.001
0400
0.000
0.000
0400
0400
0400
.
m
.
.
.
t
,
aoo3
1
10
-9
ft
0400
0.000
0.001
0401
0401
aooi
aooo
0,000
0.000
0400
t
.
4
t
t
. '
.
, !
-
0.004
-9
4
%
0.000
0401
0403
0.004
0404
.0.002
0.001
0.001
0.000
0400
0.000
[_ 0400
0400
0.016
s
i -
' 4
-7
ft
0401
0402
0412
04|4
0411
0407
0.003
0402
0.001
0.000
aooo
0.000
0.000
t
t
,
t
.
0.054
-7
' -6
%
0.004
0408
0440
0444
0435
0422
0410
0405
0402
0401
0400
aooo
0400
0400
f
.
.
(
f
.
ani
''. ';*y
'"»'' -'III'
:"i. -
;; ; : , ; .: ;:. |
' I : ':' '
' '. '.': '
:''v-S,
':»'
0419
0434
0.11?
0.126
0.102
0.062
0431
0414
0406
0403
040|
0400
0400
0400
0400
.-
'
i , '
...
Kj
, '
>;^1|
.';$:'&
!-vV:$
a r1
ft
0.1*7
0.2J3
aiio
w
13
oat*
04«
0419
040J
0404
0401
0401
0400
r Tr^r
adao
I^Tii
0400
.*i '"
'r*
:V'i '
* *
Hi
'vf.-:M
0.512 Iyp3ff
Iji
Ill'f!
$ ';
lit
M$
^!JK
ijjjl^'j
^^.
' V$$
'ii'fe
iJffeli!'
';o:**
iO3S8:
4400!
':-:IQJ^
:;o-39^
y*m.
01*1
iww
0049
'pit
ooiio
0405
33£.
0401
rOflOO;
':o^o».
Mil'
l:«fp:
life
jl i'.
'$
|:
gg!
-------
Table C-fi
SUMMARY STATISTICS
BY VEHICLE TYPE
Baltimore 3-Parameter Vehicles
Speed (mph)
Mean
St Dev
Max
Mln
Count
-».
Mean
St Dev
Max
Mln
Count
Luxury/Sedan/
Station Waagn
24.07
20.58
94.46
0.00
2,324,449
-
Luxury/Sedan/
Station Wagon
0.00
1.52
15.19
-19.49
2,320,940
Pickup/Van/
Utility
25.59
20.19
83.72
0.00
903,688
Accel (mph/sec)
Pickup/Van/
Utility
0.00
1.44
13.11
-11.88
902,442
Sports Car/
High Performance
24.61
21.33
72.20
0.00
137,367
Sports Car/
Hlgfi Performance
0.00
1.48
11.75
-11.84
137,168
Power (mphA2/sec)
Mean
St Dev
Max
Min
Count
Luxury/Sedan/
Station Waaon
48.02
42.3d
557.69
0.00
966,213
Pickup/Van/
Utility
45.63
43.51
457.01
0.00
386,314
Sports Car/
High Perforrnanca
48.58
48.80
407.04
0.00
55,381
C-6
-------
MflK
Tabla C-7
O
DISTRIBUTION OF SPEED
By Vehicle Typo
Baltimore 3-Parameter Vehicles
Speed
(mph)
0 0
0 - 5
5 - 10
10 - IS
15 - 20
20 25
25 - 30
30 - 35
35 - 40
40 45
45 50
50 - 55
55 - SO
60 - 65
65 - 70
70 - 75
75 - 80
80 85
85-90
80 - 85
Total
Luxury/Sedan/Station
Count
508,282
145,405
126,874
139.811
146,487
158,302
184.792
215,640
172,105
113.136
77,083
77,798
95,575
82,830
51.381
15,048
2,541
246
91
11
2,324,449
Wagon
Cumulative
Percent
21.9%
6.3%
5.5%
6.0%
6.3%
6,8%
8.4%
9.3%
7.4%
4.9%
3.3%
3.3%
4.1%
,3.6%
2.2%
0.6%
0.1%
0.0%
0.0%
0.0%
100.0%
Percent
21.9%
28.2%
33.6%
39.6%
45.9%
52.8%
61.1%
70.4%
77.8%
82.7%
86.0%
89.3%
93.5%
87.0%
99.2%
99.9%
100.0%
100.0%
100.0%
100.0%
Pickup/Van/Utllity
Count
168,764
53,660
47,585
54,386
55,278
65,268
75,649
83,646
72,586
51,896
41,531
37,512
43,002
36.343
13.933
2,380
262
7
0
0
903,668
Percent
18.7%
5.9%
5.3%
6.0%
6.1%
7.2%
8.4%
9.3%
8.0%
5.7%
4.6%
4.2%
4.8%
4.0%
1.5%
0.3%
0.0%
0.0%
0.0%
0.0%
100.0%
Cumulative
Percent
18.7%
24.6%
29.9%
35.9%
42.0%
49.2%
57.6%
66.9%
74.9%
80.6%
85.2%
89.4%
94.1%
98.2%
99.7%
100.0%
100.0%
100.0%
100.0%
100.0%
Sports Car/High Performance
Count
32,844
8,633
6,983
7,511
7.645
8,347
9,599
9,235
8,990
8,073
6,680
5.693
8.821
6,744
1,512
57
0
0
0
0
137.367
Percent
23.9%
6.3%
5.1%
5.5%
5.6%
6.1%
7.0%
6.7%
6.5%
5.9%
4.9%
4.1%
6.4%
4.9%
1.1%
0.0%
0.0%
0.0%
0.0%
0.0%
100.0%
Cumulative
Percent
23.9%
30.2%
35.3%
40.7%
46.3%
52.4%
59.4%
66.1%
72.6%
78.5%
83.4%
87.5%
93.9%
98.9%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
-------
Table c-a
DISTRIBUTION OF ACCELERATIONS
By Vehicle Type
Baltimore 3-Parameter Vehicles
Acc*l
(mph/s)
20 to -19
10 to -18
-18 to -17
-17 to -16
-18 to -IS
IS to -14
-14 lo -13
-13 to -12
12 to -11
-11 to -10
-10 to -9
9 to -a
9 to -7
-7 to -«
-« to -5
5 to -4
-4 to -3
3 to' -2
-2 to -1
-1 to 0
0
0 to 1
1 to 2
2 to 3
3 to 4
4 to 5
5 to 6
6 to 7
7 to8
8 to 9
9 to 10
10 to 11
11 to 12
12 to 13
13 to 14
14 CO 15
15 to 16
Total
Uuxury/Sedan/Statlon
Count
1
0
0
0
0
3
2
11
25
37
121
410
1.361
4,324
12.671
29.641
53.672
83,651
150.477
516.758
501,562
613,053
182,595
93,270
44,517
19.325
7,868
2.755
1.238
S16
230
150
51
15
7
2
1
2.320,940
Wtgon
Cumulative
Percent
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.1%
02%
05%
1.3%
2.3%
3.6%
6.5%
22.3%
21.6%
26.4%
7.9%
4.0%
1.9%
0.9%
0.3%
0.1%
0.1%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
100.0%
Percent
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.1%
0.3%
0.8%
2.1%
4.4%
8.0%
14.5%
36.8%
58.4%
84.8%
92.7%
96.7%
98.6%
90.4%
99.8%
99.9%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
Plckup/V«n/Utllltv
Count
0
0
0
0
0
0
0
0
a
s
28
110
35*
1,154
3,828
10.226
19.528
32.208
60.135
219,964
168,582
250,606
74.447
35.745
15,575
6.371
2.222
792
333
122
60
29
6
S
1
0
0
902.442
Percent
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.1%
0.4%
1.1%
22%
3.6%
6.7%
24.4%
18.7%
27.8%
82%
4.0%
1.7%
0.7%
02%
0.1%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
100.0%
Cumulative
Percent
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.1%
02%
0.6%
1.7%
3.9%
7.5%
14.1%
38.5%
57.2%
85.0%
932%
972%
98.9%
99.6%
99.9%
99.9%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
Sport* Car/High Performance
Count
0
0
0
0
0
0
0
0
1
2
2
25
84
278
700
1,708
2,807
4.292
8.424
31 .070
32.394
36,136
10.313
4,572
2,323
1,179
549
165
78
38
15
a
s
0
0
0
0
137.168
Percent
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.1%
02%
0.5%
12%
Z0%
3.1%
6.1%
22.7%
23.6%
26.3%
7.5%
3.3%
1.7%
0.9%
04%
0.1%
0.1%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
100.0%
Cumulative
Percent
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.1%
0.3%
0.8%
2.0%
4.1%
72%
13.4%
36.0%
59.6%
36.0%
93.5%
96.8%
98.5%
99.4%
99.8%
99.9%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
-------
Tabla C-9
DISTR1BUTTON OF POWER
By Vehicle Type
Baltimore 3-Pararneter Vehicles
Power
(mph*2/s)
0 - 20
20 - 40
40 60
60 - 80
80 100
100 120
120 - 140
140 - 160
ISO - 180
180 - 200
200 - 220
220 240
240 - 260
260 - 280
280 300
300 320
320 - 340
340 - 360
360 - 380
380 - 400
400 420
420 - 440
440 - 460
460 480
480 ~ SOO
SOO - 520
S20 - 540
540 - 560
Total
Uuxury/3»
-------
Table C-1Q
SUMMARY STATISTICS BY VEHICLE AGE
Baltimore 3-Parameter Vehicles
Mean
St Dev
Max
Min
Count
Spaed (mph)
MY 75-82
21.15
19,20
81.15
. 0,00
513,342
MY 83-92
25..10
20.69
94.46
0.00
2,852,162
Accel (mph/sec)
Mean
St Dev
Max
Min
Count
MY 75-82
0.00
1.53
15.19
.-19.49
512,438
MY 83-92
0.00
1.49
14.88
-14.70
2,848,112
Power (mphA2/sec)
Mean
St Dev
Max
Min
Count
MY 75-82
45.22
42.08
379.87
0.00
207,215
MY 83-92
46.15
43.11
557.69
0.00
1,200,693
C-10
-------
Tabla C-11
DISTRIBUTION OF SPEED
By Vehicle Aga
Baltimore 3-Parameter Vehicles
Speed
(mph)
0 - 0
0 - 5
5-10
10 - 15
15 - 20
20 - 25
25 30
30 35
35 - 40
40 - 45
45 - 50
50 55
55 - 60
60 - 65
65 - 70
70 - 75
75 - 80
80-85
85 90
90 * 95
Total
Model
Count
124,485
37.962
31.840
33,343
34,553
38,127
43,822
45,369
36,97§
21.391
14,900
12.931
15,795
15,861
5,384
580
18
- " 2-
0
0
513,342
Year 75
Per -
2- >
7.4%
6.2%
3.5%
6.7%
7.4%
8.5%
8.8%
7.2%
4.2%
2.9%
2.5%
3.1%
3.1%
1.0%
0.1%
0.0%
0.0%
0.0%
0.0%
100.0%
82
Cumulative
Percent
24.2%
31.6%
37.8%
44.3%
51.1%
58.5%
67.0%
75.9%
83.1%
87.2%
90.1%
92.7%
95.7%
98.8%
99.9%
100.0%
100.0%
100.0%
100.0%
100.0%
Model
Count
586,405
169,736
149,602
168,365
174,867
193,790
236,218
263.152
216,702
151.714
110,394
108,073
131,603
110,056
S 1,442
16,905
2. 785
251
91
11
2.852.162
Year 83
Percent
20.6%
6.0%
5.2%
5.9%
6.1%
6.8%
8.3%
9.2%
7.6%
5.3%
3.9%
3.8%
4.6%
3.9%
2J2%
0.6%
0.1%
0.0%
0.0%
0.0%
100.0%
- 92
Cumulative
Parcent
20.6%
26.5%
31.8%
37.7%
43.8%
50.6%
58.9%
68.1%
75.7%
81.0%
84.9%
88.7%
93.3%
97.1%
99.3%
99.9%
100.0%
100.0%
100.0%
100.0%
C-11
-------
Table C-13
DISTRIBUTION OF ACCELERATIONS
By Vehicle Age
Baltimore 3-Faraffletsr Vehicles
Accel
(mph/s)
-20 to -19
-1i to -18
-18 to -17
17 to -16
-16 to -15
-15 to -14
-14 to -13
-13 to -12
-12 to -11
-11 to -10
-10 to -9
-9 to -8
-8 to -7
-7 to -8
-6 to -5
-5 to -4
4 to -3
-3 to -2
-2 to -1
-1 to ff
0
0 to 1
1 to 2
2 to 3
3 to 4
4 to S
5 to 6
8 to 7
7 to 8
8 to 9
9 to 10
10 to 11
11 to 12
12 to 13
13 to 14
14 to 15
15 to 16
Total
Modal
Count
1
0
0
0
0
0
1
1
5
7
15
61
222
S2S
2,719
6,762
12,546
19.837
34,278
105,514
122,429
125,325
42,420
21,687
10,651
4,877
1,602
532
184
89
22
19
4
2
0
0
1
512,438
Year 75
Percent
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.2%
0.5%
1.3%
2.4%
3.9%
6.7%
20.6%
23.9%
24.5%
8.3%
4.2%
Z1%
0,9%
0.3%
0.1%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
100.0%
82
Cumulative
Percent
0.0%
0.0%
0.0%
0,0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.1%
0.2%
0.8%
2.1%
4.5%
8.4%
15.1%
35.7%
59.6%
84.0%
92.3%
98.5%
98.6%
99.5%
99.8%
99.9%
100.0%
100,0%
100.0%
1 00.0%
100.0%
100.0%
100.0%
100.0%
100.0%
Model
Count
0
0
0
0
0
3
1
10
29
37
134
484
1,579
4.931
14,480
34,813
63,459
100,314
184,758
662,278
530,109
774,470
224,935
111,900
51,764
22,898
9,037
3,180
1,465
687
303
168
58
18^
8
2
0
2,848,112
Year 83
Percent
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.1%
0.2%
0,5%
1.2%
2J2%
3.5%
6.5%
23.3%
20.4%
27.2%
7,9%
3.9%
1.8%
0.8%
0.3%
0.1%
0.1%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
100.0%
- 92
Cumulativa
Percant
0.0%
0.0%
0.0%
0.0%
0.0%
0,0%
0.0%
0.0%
0.0%
0,0%
0.0%
0.0%
0.1%
0.3%
0.8%
2.0%
4.2%
7.7%
14.2%
37.5%
57.8%
85.0%
92.9%
96.9%
98.7%
99,5%
99.8%
99.9%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
-------
Tabla C-13
DISTRIBUTION OF POWER
By Vehicle Age
Baltimore 3-Parameter Vehicles
Powar
(mpn*2/s)
0 - 20
20 - 40
40 30
SO - 80
80 - 100
100 - 120
120 - 140
140 - 160
160 * 180
180 - 200
200 - 220
220 240
240 260
260 - 280
280 300
300 - 320
320 340
340 - 360
360 - 380
380 - 400
400 - 420
420 - 440
440 - 460
460 - 480
480 - 500
500 - S20
520 - 540
540 - 560
Total
Model
Count
72,234
44,243
31,423
21,785
14,979
9,427
5394
3,255
1,846
1,059
566
319
190
119
52
14
7
1
2
'. '. ' :" ' °
0
0
0
0
0
0
0
0
207,215
Year 75
Percent
34,9%
21.4%
15.2%
10.5%
7.2%
4.5%
2.7%
1.6%
0.9%
0.5%
0.3%
0.2%
0.1%
0.1%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0,0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
100.0%
- 82
Cumulative
Percent
34.9%
56:2%
71.4%
81.9%
89.1%
93.7%
96.4%
98.0%
98.9%
99.4%
99.7%
99.8%
99.9%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
Model
Count
405,491
263,757
184,202
127,295
85,463
54,743
33,170
19,471
1 1 ,253
6,327
3,740
2,247
1,375
904
530
288
180
102
70
42
20
8
7
7
0
0
0
1
1,200,693
Year 83
Percent
33.8%
22.0%
15.3%
10.6%
7.1%
4.6%
2.8%
1.6%
0.9%
0.5%
0.3%
0.2%
0.1%
0.1%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
100.0%
- 92
Cumulative
Percent
33.8%
55.7%
71.1%
81.7%
88.8%
93.4%
96.1%
97.7%
98.7%
99.2%
99.5%
99.7%
99.8%
99.9%
99.9%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
-
C-13
-------
Table C-14
SUMMARY STATISTICS
BY TIME OF DAY
Baltimore 3-Parameter Data
Speed (mph)
Mean
St Dev
Max
Min
Count
1-6 AM
24.80
20.15
72.36
0.00
84,906
6-9 AM
25.42
21.44
83.72
0.00
538,780
9AM-4PM
24.59
20.77
94.46
0.00
1,488,941
4-7 PM
22.98
19.44
80.61
0.00
746,958
7PM- 1AM
25.43
20.26
89.43
0.00
505,919
Accel (mph/sec)
Mean
St Dev
Max
Min
Count
Mean
St Dev
Max
Min
Count
1-$ AM.
0.00
1.53
13.78
-14.70
84,796
1-6 AM
47.24
45.26
469.68
0.00
35,328
6-9 AM
0.00
1.45
12.89
-19.49
538,163
Power
6-9 AM
47.65
43.50
448.12
o.oa
221,077
9AM-4PM
0.00
1.51
15.19
-12.64
1,486,519
(mphA2/sec)
9AM-4PM
45.89
42.79
557.69
0.00
629,305
4-7 PM
0.00
1.51
13.96
-11.88
745,870
4-7 PM
46.15
42.91
426.85
0.00
309,734
7PM-1AM
0.00
1.47
13.92
-13.31
505,202
7PM-1AM
44.30
42.52
479.82
0.00
212,464
C-14
-------
I DBIRBUTONOFSftH)
'; By Tint ot My ' .
.tumor* 3-ftnimur
AU-4PM
4-7 PM
7PU-1AU
Count P*K*M
CumulMv*
ftictnt
Count
Klt'j" P»c.
'-ML <«;!?;'
P»ce«ni
CouM P«rc«nt
Cumuiatlv*
Ptfc«nl
Count
' P»fC«nl
Cufflulwiv*
Parcwii
Count Percent
Cumulative
1 Parc«nt
0-0
0- »
8- 10
10-15
19 20
20-29
29 30
30 39
39 40
40 45
49 90
SO - 95
99 - 80
80 -89
65 . 70
70 79
79 - 88
88 89
89 - 80
80 89
18.708
3,717
3.839
4.290
4.84?
9,398
8,881
1,349
7,784
4,818
3,921
3.188
3.818
3,341
1.184
70
0
0
a
o
23.2%
4.4%'
43*
9.1%
9.7*
8.3%
7.8%
11.0%
81*
9.4%
i 4.1%
18%
4.3%
3.8%
1.4%
0.1*
04%
04ft
84%
04ft
23.2%
2?4ft
314ft
42.8%
484%
98.8%
7.8%
774%
12.4%
8M%
90J%
84.8%
884%
884%
1000%
100.0%
1004%
100.0%
1000*
:..j 78.7%
i1|!i!i'83.8%
W-
'''§.'i
100.0%
308.182
8,827
88.022
83,018
83,780
101,249
118.888
134,892
110,280
73.485
91.284
91.007
7,23$
95.824
33,884
11.843
2,078
138
93
II
20.6%
4%
5-8*
8J%
8J%
8,8*
11%
34%
7.4%
44ft
3.4%
14%
44%
17%
2J*
at*
0.1%
oo*
84%
0.0%
204%
27.1%
328%
39.1*
454*
822%
80.3%
88.3%
79.7%
81.7%
85.1%
88.8%
83.1%
88.8%
88.1%
884%
1000*
100.0%
100.0%
100.0%
180.351
50,314
42.248
48.833
48,482
54,630
88.358
71.952
57,225
37,884
27,817
23,384
25.703
21.884
8.818
1,307
115
3
0
0
21.5%
8.7*
5.7%
8.3ft
*5%
7.3%
89%
8.6%
7.7%
9.1%
17%
3.1%
14%
2.8%
1.3ft
0.2%
04ft
00%
04ft
00*
21.8%
28.3%
33.8%
40.2%
48.7%
54.0%
62.0*
725*
80.2%
15.3%
884*
82.1*
85.6%
96.5*
88.8*
100.0*
100.0*
100.0*
100.0*
100.829
28,148
25,022
28.784
30.485
34,480
41.583
46,457
41,203
27.683
21.822
24.583
28,238
18,551
4.551
48
241
5
38
0
199*
'5.8*
4.9%
9.7%
8.0*
6i*
8.2%
8,2%
81%
55%
4.3%
4.9%
5.6%
3.8%
0.9*
0.2%
00%
0.0%
00*
00*
18 9V
25.7r
30.7%
38.3*
42.4%
49.2%
57.4*
666%
74.7%
802%
846%
89.4%
5.0*
98.9%
99.8*
89.8%
100.0%
100.0%
100.0%
1000*
Tou
4.804) 100.0*
531.710 100
t2*
1.488841 100.0%
746,858 1000*
505,919 100.0*
-------
Tifrli C-16
DISTRIBUTION OF ACCELERATIONS
By Tm> erf Day
V»MC|«
ACC«I
(mph/»)
-20 to -IB
19 to -II
-11 to -1?
-17 lo -16
18 to -IS
-IS 10 -14
-14 to -13
-13 (a -12
12 to -11
-11 la -10
-10 to -9
« to -1
- 10 -7
-7 10 -8
-8 lo 5
-S 10 -4
-4 10 -3
-3 10 -2
-2 IB -1
1 10 0
0
0 to 1
1 *>2
2 to3
3M4
4 to S
5 to «
0 to 7
7 to 1
toO
8 IQ 10
10 10 11
11 10 12
12 to 13
13 lo 14
14 10 IS
15 !0 18
total
1-6 AM
CumulaUvt
Count P*rc*nt P»rc»nl
o ao% 0.0%
0 O.O% 0.0%
0 00% 0.0%
0 0.0% 00%
0 0.0% 0.0%
2 0.0% O0%
1 0.0% 0.0%
a 0.0% ao%
s 0.0% ao%
s 0.0% ao%
i* 00% ao%
34 0.0% 0.1%
82 0.1% O2%
114 0.2% a4%
SIS 0.6% 14%
1.152 1.4% 2.4%
1.87S Z2% 4.8%
2,718 13% 7.0%
4.510 S.4% 113%
18.514 21.8% 35.1%
19,888 212% 5*3%
22,802 27.0% 853%
«.S42 7.7% 93.1%
3,222 i»% 96.9%
1.415 t.8% 88.8%
710 at% 09.5%
282 0.3% 89.8%
88 ai% 93.8%
41 ao% 100.0%
25 0.0% 100.0%
8 ao% 100.0%
2 0.0% 100.0%
2 0.0% 100.0%
3 0.0% 100.0%
2 ao% 100.0%
0 0.0% 100.0%
0 0.0% 100.0%
84,796 1000%
8-9 AU
CumuUUva
Count Purcsni Pirc»nl
1 0.0% 0.0%
0 00% 0.0%
0 0.0% 00%
0 00% 0.0%
o ao% 0.0%
1 0.0% 00%
0 0.0% 0-0%
0 00% 0,0%
7 0.0% 0.0%
I 0.0% 00%
1 4 0.0% 00%
73 00% 00%
244 0.0% 0.1%
902 0.2% 02%
2.74S 05% ar%
8,353 1.2% 1.9%
11.532 2.1% 4.1%
17.8 IS 3.3% 7.4%
33.380 8.2% 136%
121.277 225% 38.1%
122.724 22.8% 519%
143.178 28.8% 855%
41, 8SS 7.8% 93.3%
20.98$ 3.8% 87.1%
8.209 1.7% 88.8%
3.841 0.7% 988%
1.445 03% 98.8%
£08 ai% 88.9%
2S4 00% 100.0%
118 0.0% 100.0%
S3 0.0% 100.0%
21 0.0% 100.0%
S 0.0% 100.0%
3 0.0% 100.0%
0 00% 100,0%
0 0.0% 100.0%
0 0.0% 100.0%
531,163 1000%
8AU-4PM
Cumulailv*
Count PaiCMl P»rc»m
o ao% 0.0%
o ao% 0.0%
0 0.0% 00%
0 0.0% 00%
0 0.0% 0.0%
0 0.0% 0.0%
0 0.0% 0.0%
2 0.0% 0.0%
13 0.0% 0.0%
18 0.0% 0.0%
70 0.0% 0.0%
251 0.0% 0.0%
848 0.1% 0.1%
2.814 0.2% 0.3%
7,880 0,5% 0.8%
18,800 1.3% 2.1%
34,213 2.3% 4.4%
54.194 18% 80%
97.880 8.8% 14.8%
339.237 228% 37.4%
301,113 20.3% $7.7%
400.887 27.0% 84.8%
118,882 8.1% 92.7%
80,144 4.0% 98.8%
28,070 1.9% 98.8%
12.191 0.8% 99.5%
4,798 0.3% 88.8%
1.711 0.1% 99.9%
780 ai% 100.0%
377 0.0% 100.0%
1S1 00% 100.0%
91 0.0% 100.0%
30 00% 100.0%
8 0.0% 100.0%
2 0.0% 100.0%
2 0.0% 100.0%
1 0.0% 100.0%
1,488.513 100.0%
4-7 PU
Cumulative
Count Percent Percanl
0 0.0% 0.0%
0 00% 00%
0 00% 00%
0 00% 00%
o ao% 00%
0 0.0% 00%
0 0,0% 00%
0 0.0% 00%
7 0.0% 00%
3 0,0% 00%
24 ao% 00%
92 ao% 0.0%
358 0.0% 0.1%
1,214 02% 02%
3.6S4 0.5% Q7%
9.188 1.2% 1.0%
17.30S 2.3% 4.3%
28.057 3.8% .8.0%
S2.234 74% 15.0%
188.238 22.3% 37.3%
157,788 21.2% 58.5%
182.073 258% 84.3%
81,944 43% 82.8%
30,481 4.1% 08.7%
14,233 1.9% 98,6%
6.445 09% 09.5%
2,490 0.3% 998%
871 0.1% 99.9%
340 0,0% 100.0%
148 0.0% 100.0%
61 0.0% 100.0%
47 0.0% 100.0%
1 S 0.0% 100 0%
2 0.0% 100.0%
3 0.0% 100.0%
0 00% 100.0%
0 00% 1000%
745,870 1000%
7PM- 1AM
CumuUUva
Couni Parcenl Parcant
0 0.0% 0.0%
0 0.0% 00%
o ao% 00%
0 0.0% 0.0%
0 00% 00%
0 0.0% 0.0%
1 ao% 00%
3 00% 0,0%
2 0.0% 00%
1 o ao% 0.0%
25 0.0% 00%
95 00% 0,0%
258 0.1% 0.1%
842 0.2% 0.2%
2,425 0.5% 0.7%
6.004 1.2% 10%
11,080 2.2% 4.1%
17,296 3.4% 7.5%
30,952 8,1% 13.7%
122.S28 24.3% 37.9%
101,217 20.0% 57.9%
140.054 27.7% 85.7%
36,952 7.3% 93.0%
19.375 3.8% 96.8%
9,418 1.9% 9».7%
4.082 08% 995%
1,814 03% 998%
533 0.1% 99,9%
234 00% 100.0%
107 00% 100.0%
54 OQ% 100.0%
26 0.0% 100.0%
10 0.0% 100.0%
4 0.0% 100 0%
1 00% 1000%
0 0.0% 1OO.O%
0 00% tOO 0%
505.202 1000%
-------
i OSmeUTIQNOFPOWER
' :. 8y nn*olO*y
BalUffiora 3-PMUMUt V*1MM
Power
(mpti'T'*)
0- 20
20 -40
40 - M
80- M
80 - 100
100 120
120 140
140 180
180 1M
1*0 - 200
200 220
220 240
240 - 2*0
2M - 2*0
no - 900
900-920
920-940
940 * 9*0
9M -980
9M - 4M
400 -420
420 « 440
440-4*0
4*0 - 4M
4M MO
800 520
820 * 940
540 * 8*0
ToUl
V*. All '- ' ''! '?
Cunul****
Count Portent Percent
' 11.13* . 394* ' 33*%
7.701 21 J* ' MJ%
9.27* 14** ' 70.2%
9.711 104* 80**
1,791 7,7% *U%
1,119 5,1* 834*
M* 34* MA*
S27 14% *7.7*
244 07% 88.4*
149 0.4* 8*4*
117 03% 68.1*
'88 02*' . M.4%
19 02* MA*
98 02* M.7*
49 01* 894*
1* 01* MA*
14 OAK MJK
* OAK ; MA*
8 OO* ' 100.0*
8 OAK 100.0*
. 1 00* 100.0*
1 00% 100.0*
9 OO* 100.0*
2 OAK 100.0*
0 00* IMA*
OO* 100.0*
0 00* 100.0*
0 OO* IMA*
39.928 100.0*
'^^Ifl'fl;*!!!^:-.-'
' - '' ;'Vi!i:^j 'jp^^i* ^"M***1
'' Co««iV^ilpUfeeVrt ' ' Petceni
' :' 'fwHi, ';' 8i'J|4!'?|i:':\-J,Sii**
,; '^Mliigli' a2.9*|ijl'::;i M.I*
'\';^lf|[*j|'l.^J»!jJSi:\:)IM%
§S!a'j;:';;*ot*
l"i[l ,: ;. j**.S%
j|-r';,'."M\a%
Ijt ',(95.8*
' v^l^i^'ir-wi*
/rj.'^pipwfi-!1- "**
. i!i.S*j( '|-|'"j*i*p{'V r''M2*
!': ' :-^ iFi**'^'i; *»*
', ,tMj|'>Mi,'fti^',i'.f,'.i;;M.7*
24* i: o'i* '''' , Ml*
|jl1!'i'-'lfMt4 '.':;M.8%
:..p8.J>;;ftO%;!!;; MB*
58:!r::':>».0*;:-j. 100.0*
(jilA'i 8.0*1 1' 100.0*
'\^f"v^4M^.,'-'- 100.0*
-, , . ^iMliwIiiif J'. joo.o%
' :.4WiW m&***
' ' >\ j. *!, Ji !,.[ 0(^»[' ; i , \10OO*
'.'.'\r «K||w|i| H< MM%
. ''^.ta^IJTOAte'j^lOOA*
' ' !, J!,a]rf| 'iBJiiijiii .f,ioo.o*
r :A ,';|di ||;|*3wf'SPH|IOttA*
. ' i'.!ji'ljll j'lf jfWwIi '; AL'i00-8*
- - ' it! 0' ''' 5"oo*' 'ii-"' 1100.0*
a^olir^iyfto*."'' -" :''
'' ': 9AM-4PH
' CumuJeilve
Count Percent Percent *
214.43* 34.1% 34N*
137.141 21.** 55.8*
88.301 153* 71.2*
' 88.44* 104* 81.7*
44.M8 7,1* 88.8*
; 28.242 45* 834%
17.51* 1»* 96.1%
10.333 14* 874%
8.001 1.0* 8*7*
,. . 9.384 ';, 05% M^*
' ' 1.800 : ;03* ' : 888%
1,084 0.2* 898%
858 0.1% MA*
404 01* MA*
245 OO* 1000%
111 OO* 100,0%
91 00* 100.0*
27 OO* 100.0*
18 0.0* 100.0*
i 11 OAK 100.0*
* OO* 100,0*
4 OAK 100.0*
2 : OO* 100.0*
2 OAK 100.0*
0 OAK 100.0*
', 0 OO* 100.0*
0 OO* 100.0*
1 OAK 100.0*
29,305 100.0*
4-7 PM
Cumulative
Count Percent Percent
104,111 33.8% 33.**
8,535 ' 21.5* ' / 55 3%
47,8*8 154* TOT*
93.897 10,9% 81.8*
22.36* 7.2* 18.8*
14,543 4.7* 83.5*
8.515 2.7* 883%
4,843 1.0* 97.8*
2.839 Of* 98.7*
1.571 05* 89.2%
882 03* M.5*
583 02* 99.7%
356 01* 89.9*
217 01* 89.9%
119 00* 999%
78 0.0* 100.0*
89 00% 100.0*
24 0.0* 100.0*
14 OO* 100.0*
a 0.0* 100.0*
4 00* 100,0*
t 00* 100.0*
0 00* 100.0*
0 0.0* 100.0*
0 0.0* 100.0*
0 0.0% 100.0*
0 0.0* 100.0*
0 00* 100.0*
309,734 100.0%
7PM-1AH
Cumulative
Count Percent Percent
78,260 355* 35.9%
47.312 22.3* 58.2%
31.558 14.9* 73.0%
20,119 9.1* 82.6*
14,10* 6.6% 89,5%
9,092 4.3* 937%
5.625 2.6* 963%
3,245 1.5* 97,9%
1.142 0.9% 88.7%
1,041 0.5* 99.2%
S3 0.3* 995%
397 0.2* 93.7%
242 O1* 998%
153 01* 999%
65 0.0* 100.0%
38 00* 100.0%
21 0.0* 100.0%
14 0.0* 1000%
19 00* 100.0%
5 0.0* 100.0%
3 0.0* 100.0%
1 0.0* 100.0%
1 OO* 100.0%
9 . 0.0* 100.0%
0 OO* : 1004%
0 0.0* 100.0%
0 00* 1090%
0 0.0* 1000%
212,484 100.0*
-------
Table C-18
SUMMARY STATISTICS
BY TIME OF WEEK
Baltimore 3-Parameter Vehicles
Speed (mph)
Mean
St Dev
Max
Min
Count
Weekday,
23.92
20.24
94.46
0.00
2,559,047
Weekend
26.35
21.29
83.14
0,00
806,457
Accel (mph/sac)
Mean
St Oev
Max
Min
Count
Weekday
0.00
1.51
15.19
-19.49
2,555,267
Weekend
0.00
1,47
14.88
-12.82
805,283
Power (mphA2/soc)
Mean
St Oev
Max
Min
Count
Weekday
46.57
43,40
557.69
0.00
1,064,896
Weekend
44.30
41.54
457.01
0.00
343,012
C-18
-------
Table C-19
DISTRIBUTION OF SPEED
By Time of Week
Baltimore 3-Parametar Vehicles
Speed
(mph)
0 - 0
0 - 5
5". 10
10 - 15
15 - ZO
20 - 25
25 30
30 - 35
35 * 40
40 45
45 - 50
50 - 55
55-60
60 - 65
"65-70
70 - 75
75 -:8b
80 - 85
~85 - 90
90-95
Total
Weekday
Count
552,407
159,416
139,732
156,539
162,361
179,891
218,756
238,220
193,424
128,989
90,224
87,687
101,185
86,200
48.73S
12.457
. "-"- '1,572-
"-" 16i
91
11
2.559.047
Percent
21.6%
6.2%
5.5%
6.1%
6.3%
7.0%
8.5%
9.3%
7.6%
5.0%
3.5%
3.4%
4.0%
3.4%
v /Ul%
~' 0.5%
*~:;'~ai%-
0.0%
0.0%
0.0%
100.0%
Cumulative
Percent
21.6%
27.8%
33.3%
39.4%
45.7%
52.8%
61.3%
70.6%
78.2%
83.2%
86.7%
-* 90.2%
94.1%
97.5%
: 99.4%
99.9%
too.o%
100.0%
100.0%
100.0%
Weekend
Count
158,483
48,282
41.710
45,169
47,059
52,026
61,284
70,301
60,257
44,118
35,070
33,337
46,213
39,717
17,090
5,028
1.231
84
0
0
806.457
Percent
19.7%
6.0%
5.2%
5.6%
5.8%
6.5%
7.6%
8.7%
7.5%
5.5%
4.3%
4.1%
5.7%
4.9%
2.1%
0.6%
0.2%
0.0%
0.0%
0.0%
100.0%
Cumulative
Percent
19.7%
25.6%
30.8%
36.4%
42.2%
48.7%
56.3%
65.0%
72.5%
78.0%
82.3%
86.4%
92.2%
97.1%
- ;.-- 99.2%
. "...;; 99*8%
100.0%
100.0%
100.0%
100.0%
C-19
-------
Tabla C-2Q
DISTRIBUTION OF ACCELERATIONS
By Tim a of Week
Baltimore 3-Paramatar Vehicles
Accal
(mph/s)
-20 to -19
-19 to -18
-18 to -17
-17 to -16
-16 to -15
-15 to -14
-14 to -13
13 to -12
-12 to -11
11 to -10
-10 to -9
-9 to -8
-8 to -7
-7 to -6
-6 to -5
5 to -4
-4 to -3
-3 to -2
-2 to -1
-1 tb 0
0
0 to 1
1 to 2
2 to 3
3 to 4
4 to 5
5 to 6
6 to 7
7 to 8
8 to 9
9 to 10
10 to 11
11 to 12
12 to 13
13 to 14
14 to 15
15 !o 16
Total
Weekday
Count
1
0
0
0
0
3
2
6
27
32
110
401
1355
4352
13146
31670
58583
93303
1 70202
572896
544282
672150
207292
1 03228
47953
20929
8088
2890
1292
586
259
153
51
17
6
1
1
2,555,267
Parcant
0.0%
0.0%
0.0%
0.0%
. 0.0%
%
0.0%
0.0%
0.0%
0.0%
0.1%
0.2%
0.5%
1.2%
£3%
3.7%
6.7%
22.4%
21.3%
26.3%
8.1%
4.0%
1.9%
0.8%
0.3%
0.1%
0.1%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
100.0%
Cumulative
Percent
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.1%
0.2%
0.8%
2.0%
4.3%
7.9%
14.6%
37.0%
58.3%
84.6%
92.7%
96.8%
98.7%
99.5%
99.8%
99.9%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
Wotktnd
Count
0
0
0
0
0
0
0
5
7
12
39
144
446
1,404
4,053
9,905
17,422
26,848
48,834
194,896
158.256
227,645
60,063
30,359
14.462
6,448
2,551
822
357
190
66
34
11
3
2
1
0
805,283
Percent
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
...0.0%
0.0%
0.0%
0.0%
0.1%
0.2%
0.5%
1.2%
2£%
3.3%
6.1%
24.2%
19.7%
28.3%
7.5%
3.8%
1.8%
0.8%
0.3%
0.1%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
100.0%
Cumulative
Percent
Q.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0,1%
0.3%
0.8%
2.0%
4.2%
7.5%
13.6%
37.8%
57.4%
85.7%
i3.1%
96.9%
98.7%
99.5%
99.8%
99.9%
100.0%
100,0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
020
-------
Table C-21
DISTRIBUTION OF POWER
By Time of Week
Baltimore 3-Parameter Vehicles
Power
lmphA2/s)
0 - 20
20 - 40
40 - 60
30 - SO
80 - 100
100 - 120
120 - 140
140 - 160
160 - 180
180 - 200
200 - 220
220 - 240
240 - 260
260 - 280
280 - 300
300 - 320
320 - 340
340 360
360 380
380 * 400
400 420
420 - 440
440 - 460
430 - 480
480 - 500
500 - 520
520 - 540
540 - 560
Total
Weekday
Count
356,443
231,895
164,114
114,158
76,923
49,005
29,897
17,704
- 10,236
5,827
3,461
2,006
1,281
827
477
250
161
95
61
37
17
8
5
7
0
0
0
1
1,064,896
Percent
33.5%
21.8%
15.4%
10.7%
7.2%
4.6%
2.8%
1.7%
1.0%
0.5%
0.3%
0.2%
0.1%
0.1%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
100.0%
Cumulative
Percent
33.5%
55.2%
70.7%
81.4%
88.6%
93.2%
96.0%
97.7%
98.6%
99.2%
99.5%
99.7%
99.8%
99.9%
99.9%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
Weekend
Count
121,282
76,105
51,511
34,922
23,519
15,165
8,967
5,022
2,863
1.559
845
560
284
196
105
52
26
8
11
5
3
0
2
0
0
0
0
0
343,012
Percent
35.4%
22.2%
15.0%
10.2%
6.9%
4.4%
2.6%
1.5%
0.8%
0.5%
0.2%
0.2%
0.1%
0.1%
0.0%
0.0%
0.0%
0.0%.
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
100.0%
Cumulative
Percent
35.4%
57.5%
72.6%
82.7%
89.6%
94.0%
96.6%
98.1%
98.9%
99.4%
99.6%
99.8%
99.9%
99.9%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
021
-------
Table C-22
SUMMARY VEHICLE STATISTICS
ALL BALTIMORE
Baltimore 3-Parameter Vehicles
Minutes/Pay
Mean 75.15
St Dev 33.78
Max 228.05
.,_.Min 17.76
Count 24.50
Miles/Day
Mean
St Dev
Max
Min
Count
30.56
13.85
105.68
4.77
93.00
Trips/Day-
Mean
St Dev
Max
Min
Count
6.25
2.83
19.91
2.01
93.00
Stops/Hour
Mean - 35.23
St Dev 9.70
Max 70.99
Min 12.97
Count 93.00
Note: Means and standard deviations weighted:
for per day measures, by number of days
vehicle was instrumented; for per hour
measures by number of hours of operation.
C-22
-------
Table C-23
SUMMARY VEHICLE STATISTICS
BY VEHICLE AGE
Baltimore 3-Parameter Vehicles
Minutes/Day
Mean
St Dev
Max
Min
Count
MY 75-82
73.08
31.31
132.72
27.94
15.00
MY 83-92
75.53
34.21
228.05
17.76
78.00
Miles/Day
Mean
St Dev
Max
Min
Count
MY 75-32
25.78
15.53
56.57
8.72
15.00
MY 83-92
31.45
19.27
105.68
4.77
78.00
Mean
St Oev
Max
Min
Count
Trips/Day
MY 75-82 MY
6.98
3.12
13.49
2.91
15.00
83-92
6.11
2.75
19.91
2.01
78.00
Stops/Hour
Mean
St Dev
Max
Min
Count
MY 75-82
40.12.
8.05
52.24
28.49
15.00
MY 83-92
34.35
9.71
70.99
12.97
78.00
Note: Means and standard deviations weightet
for per day measures, by number of days
vehicle was instrumented; for per hour
measures by number of hours of operation.
-------
Table C-24
Luxury/Sedan/
Station Wagprj
Mean 71.21
St Dev 29.90
Max 173.32
Min 17.76
Count
Mean
St Dev
Max
Min
Count
SUMMARY VEHICLE STATISTICS
BY VEHICLE TYPE
Baltimore 3-Paramtter Vehicles
Minutes/Day
Pickup/Van/ Sports Car/
Utility High Performance
89.15 68.25
43.46 9.10
228,05 82.92
32.93 57.68
68.00
21.00
4.00
Miles/Day
Luxury/Sedan/
Station Wagon
Mean 28.47
St Dev 17.89
Max 105.68
Min 4.77
Count 68.00
Pickup/Van/ Sports Car/
Utility High Performar\$^
37.78 28.01
21.19 10.80
82.70 39.88
11.66 13.09
21.00 4.00
Trips/Day
Luxury/Sedan/
Station Wagon
6.04
2.60
15.54
2.01
68.00
Pickup/Van/
Sports Car/
H/gft Performance
7.10 5.37
3.51 1.17
19.91 6.79
3.51 4.23
21.00 4,00
Stops/Hour
Luxury/Sedan/
Station Wagon
Mean 36.30
St Dev 9.89
Max 70.99
Min 15.30
Count 68.00
Pickup/Van/ Sports Car/
Utility High Perfofrnancp
32.97 31.89
9.09 6.50
51.70 38.24
12.97 21.72
21.00 4.00
Note: Means and standard deviations weighted: for per day
measures, by number of days vehicle was instrumented; for pi
hour measures by number of hours of operation.
C-24
-------
Table C-25
SUMMARY VEHICLE STATISTICS
BY TIME OF WEEK
Baltimore 3-Parameter Vehicles
Minutes/Day
Mean
St Dev
Max
Min
Count
Weekday
80.14
37.37
254.45
11.72
93.00
Weekend
62.86
45.31
166.57
0.00
91.00
Miles/Day
Mean
St Oev
Max
Min
Count
Weekday
31.90
20.78
104.76
2.37
93.00
Weekend
27.29
24.96
108.15
0.00
91.00
Trips/Day
Weekday Weekend
Mean
St Oev
Max
Min
Count
6.69
3.12
23.20
1.48
93.00
5.10
3.38
13J3
0.00
91.00
Stops/Hour
Weekday Weekend
Mean
St Oav
Max
Min
Count
35.69
10.18
81.10
. 12.76
93.00
33.72
12.09
100.93
0.00
91.00
Note: Means and standard deviations
for per day measures, by number of days
vehicle was instrumented; for per hour
measures by number of hours of operatic*
025
-------
Figure C-1
Distribution of Speed: All Driving and Vehicle Moving
Baltimore 3-Parameter Vehicles
25.0%
20.0%
o
K»
OV
Ail Driving
Vehicle Moving
0.0%
Speed (mph)
-------
Figure C-2
Cumulative Distribution of Acceleration:' All Driving and Vehicle Moving
Baltimore 3-Parameter Vehicles
All Driving
Vehicle Moving
3 &
i 50.0%
(0
o
in
o
-------
Figure C-3
Cumulative Distribution of Power: Ail Driving
Baltimore 3-Parameter Vehicles
I
o
100.0% T
90,0% --
80.0% --
S 70.0%
1
2 60.0% +
» 50.0%
o
40.0% --
30.0% --
All Driving
** \JtVt\f 49
S
£. 20.0% -
10.0% -
0.0% -
c
c
c
1
-
_ - j
l I l I I i l i II
L.
1 1 l.l 1 I 1 1 1 1 r~ l
a O O O O O O O O O O O c
vi^rtocoocM^rtocoocMTi- ^
₯-r-«r-r-CMCMCM C
3000 ' ' ' '
CMfWOOOOOOOOc
ooocM^rcoooocM'q
»-»-»-i-«-CMCMcs
Power
-------
25.0%
HP
;Pi^'}!li'&:Jl:''';' .''
llll&alti
Figure C-4 'V
of Speed: By Vehicle Type
Baltimore 3-Parameter Vehicles
0.0%
o
in
o
in
o
CO
o
Luxury/Sedan/Station Wagon
D Pickup/Van/Utilily
H Sports Car/High Performance
o
in
in
o
04
o
CO
m
o
o
in
m
in
o
to
in
to
tn
o
o
00
in
oo
o
co
o
a>
CO
01
o
o>
Spaed (mph)
-------
Figure C-5
Cumulative Distribution of Acceleration: By Vehicle Type
Baltimore 3-Parameler Vehicles
£
o
100.0% -
90.0% -
80.0% -
70.0% -
g 60.0% -
g 50.0% -
tr
f
Luxury/Sedan/Station Wagon
Q Pickup/Van/Utilily
~ t~ " Sporls Car/High Performance
40.0% --
30.0% --
20.0% --
10.0% --
0.0%
to
o
in
o
-------
0
i
. :: lIirM'
Cumulate Distribution O| Power: By Vehicle Type
..-..; r;;':j'i";iil||y|ftaltimore 3-Parameter Vehicles '..'
- . -
Luxury/Sedan/Station Wagon
i
Pickup/Van/Utility
Power
-------
Figure C-7
Distribution of Speed: By Vehicle Age
Baltimore 3-Parameter Vehicles
25.0%
MY 1975-1982
D MY 1083-1992
0.0%
Speed (mph)
-------
Figure C-8
Cumulative Distribution of Acceleration: By Vehicle Age
Baltimore 3-Parameter Vehicles
100.0% --
90.0% --
80.0%
70.0% +
o * 60.0%
i »
w -
ui >
g 50.0%
or
£ 40.0%
30.0% --
20.0% --
10.0% --
MY 1975-.1982
MY 1983-1992
0.0% V Vs- -
;-:. » ;
' l^'e' i "' *7
! '"' -1
& 2
. ** . *^
' *? *S*
,
'o
$
T,
1
1 1
o »-
2
o
1
CM
s
r-
1
OT
2
CM
1 1
<* Ul
2 2
« »
f.
(D
o
U)
5
CO
Acceleration (mph/sec)
-------
Figure C-9
Cumulative Distribution of Power: By Vehicle Age
Baltimore 3-Parameter Vehicles
LO
c 30.0% +
20.0% --
-Q-
-Q-
-£>
-Q
MY 1975-1982
MY 1983-1992
10.0% -
0.0% -
c
1 1 1 1
1 1 I 1
3 O O O O
xj V (0 CO O
* i
o o o o *
CM V U> 0
CO
t
1
I
O
CM
I
1
I |
I 1
O 0
* co
T" *"
*
0 0
CM ^
1
1
O
CO
o
to
1
1
o
0
CM
'
O
f
1
I
o
CM
CM
O
O
CM
j
1
O
"*
CM
*
O
CM
CM I
O
to
CM
Power
-------
Figure C-10
Distribution of Speed: By Time of Day
Baltimore 3-Parameter Vehicles
25.0%
0.0%
1-6 AM
n 69 AM,.
dAM 4PM
4-7PM
G 7PM- 1AM
(
0 -
0 -
5
s -
10
10
15
15
20
20
25
25
30
30
35
40
35 40 45
Speed (mph)
-------
Figure C-11
Cumulative Distribution of Acceleration: By Time of Day
Baltimore 3-Parameter Vehicles
1AM - 6AM
6AM - 9AM
9AM - 4PM
4PM - 7PM
7PM - 1AM
(O
o
in
8
so
t
o
m
o
CM
o
o
o
S
CM
O
p»
U)
o
to
o
r-.
o
CM
Acceleration (mph/sec)
-------
Figure C-12
Cumulative Distribution of Power: By Time of Day
Baltimore 3-Parameter Vehicles
100.0% T
tp
oi
1AM - 6AM
6AM SAM
* 9AM
4PM - 7PM
7PM .
Power
-------
0)
<*> CO
0
-------
10
o
in
o
(O
Cumulative Distribution of Acceleration: By Time of Week
Baltimore 3-Parameter Vehicles
.j'/y^i^^t;;1'/ . -
o
in
«
o
f
o
o
S
CM
O
CO
O
CM
O
co
in
o
o
co
Acceleration (mph/sec)
-------
Figure C-15
Cumulative Distribution of Power; By Time of Week
Baltimore 3-Parameter Vehicles
o
CM
O
IP
O
CM
O
00
o
(O
o
o
a
CO
o
CM
o
o
o
-------
Appendix D. Baltimore Soak Periods
Soak Time Classes
3-Parameter Trtps
So«k
SrtlH Frequency
0-2
1-4
4-6
S-B
S-10
10-20
20-30
30-40
40-50
50-60
60-70
70-80
ao-io
90-100
100-200
200-300
300-409
400-500
500-600
600-700;
259
442
263
236
148
531
282
210
157
123
98
86
72
78
433
198
108
112
169
" '' 98'"
j£-'-}^~^^-^~^-'-' r '
'^Tjt^?* iU'Jll'^lff : ":
Tim* {minutes)
CuniUtiv* CtmiUttve
Percent frequency Percent
5.S
9.5
5.7
5.1
3.2
11. 5
6.1
4.5
3.4
2.7
2.1
1.9
l.«
1.8
9.3
4.2
2.3
2.4
3.6
2.1
2:1
2S9
701
9S4
1200
1348
1879
2161
2371
2528
26S1
2747
2833
290S
2981
3414
3610
3718
382S
3997
*095
0->J8fei
5.6
1S.1
20.8
2S.9
29.1
40. S
46.7
SI. 2
54.6
57.2
59.3
61.2
62.?
64.4
73.7
77.9
80.2
82.8
86 3
88.4
J1.8 ^
-------
-------
Appendix S. Trip Start Driving Activity
Table E-l
Time to 140° Coolant Temperature
(< = # seconds)
10
20
30
40
SO
60
78
80
90
100
110
120
uo
no
ISO
180
170
130
190
200
210-
220
210
240
259
2W
270
210
290
300
310
320
330
340
ISO
380
370
380
390
410
<20
438
4M
4ta
810
8M
119
100
810
9SO
1030
1 HO
uas
1230
1390
1470
1620
1740
1321
10
15
JO
34
36 '
41
29
<2
" 39
38
44
39
IS
34 ,.
28
34
33
4S
37
45
38
37
39
40
Z2
28
21
23
9
7
10
7
1
3
1
1
1
1
1
3
2
1
1
1
1
I
1
1
I
. 1
w.s
' 3 4
0.5
I 3
t 5
1.5
t-»
12
1.8
1.7
1.8
1.9
1.7
1.5
l.S
1.2
l.S
1.4
2.0
l.S
1.9
l.«
1.8
1.7
1.7
0.9
l.l
0.9
i. a
0.4
0.3
0.4
0.3
O.I
0.3
0,1
O.I
0.0
0.0
0.0
0.0
O.I
0.0
0.0
0.0
0.0
0.0
0.1
0.1
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
1321
1331
134S
137S
1410
1448
1487
ISIS
1558
1597
1S3S
1679
1718
1753
1787
ISIS
1349
1882
1928
198S
2010
2040
208S
2124
2184
2188
2212
2U3
2258
2281
2271
22tt
22M
229]
I29S
2303
2308
2301
2301
2301
2110
2311
2314
231S
2318
2317
2311
2321
2321
2324
2329
2328
2327
2321
2329
2330
2331
2332
56 S
57.1
57 7
59.0
so.s
S2.0
63.8
65.0
66.8
68.5
70.1
72.0
73.7
75.2
78.8
77.8
79.3
30.7
82.7
84.3
as. 2
87 8
39.4
91.1
92.8
93.7
94.9
9S.8
98.7
97.1
97.4
97.9
98.2
98.3
38.8
98.8
98.9
98.9
M.O
99 0
99.1
99.2
99.2
99.3
99.3
99.4
99.4
99. S
99.1
99.7
99.7
99.7
99.8
99.8
99.9
99.9
100.0
100. 0
E-l
-------
Table E - 2
Time to 140° Coolant Temperature
(< = § seconds)
.. ., . >«iFCAT»CQid Start
Cumulative Curtwldt i VB
rjHECt. Freautncy
10
40
SO
60 ,
70
80
90
110
130
140
ISO
ISO
170
180
. 190
200
210
220
230
240
2SO
260
270
280
290
300
310
320
330
340
3SO
360
370
390
430
490
610
690
800
810
990
1030
1230
1470
1740
13
I
I
I
2
2
1
3
2
3
5
3
6
S
16
is
20
20
19
23
31
17
21
16
17
8
7
9
6
3
4
1
2
1
1
1
1
1
1
1
1
1
1
1
1
3ercenc i
5,8
0.3
0.3
0,3
0.6
0.6
0.3
0.9
0.6
0.9
1.6
0.9
1.9
l.S
s.o
4.7
6.3
6.3
5.9
7.2
9.7
S.3
6.6
S.O
S.3
2.5
2.2
2.8
1.9
0.9
1.3
0.3
O.S
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
requtney
18
19.
20
21
23
25
28
29
31
34
39
42
48
S3
69
84
104
124
143
166
197
214
23S
251
268
271
283
292
29t
301
30S
301
30t
3QS
310
311
312
313
314
315
316
317
318
319
320
?*rctnt
5 6
S.9
6.3
6.6
7.2
7.3
a.i
9.1
9.7
10.6
12.2
13.1
15.0
IS.S
21.8
26.3
32.5
38.8
44.7
51.9
61.6
68.9
73.4
78.4
83.8
38.3
88.4
91.3
93.1
94.1
9S.3
95.6
98.3
96.6
98.9
97.2
97.5
97.8
98.1
98.4
98.8
99.1
9i.4
9i.7
100. 0
E - 2
-------
Table E - 3
Time to 140° Coolant Temperature
(< = # seconds)
rweCl
.10
20
30
40
50
60
70
ao
90
too
no
120
130
140
ISO
ISO
179
iao
190
200
210
220
230
240
2SO
2SO
270
280
290
300
320
330
340
3SO
3SO
370
380
420
430
aio
1620
... .. ,K,
f-eauency
162
4
9
23
25
23
35
21
37
34
35
43
31
29
25
23
29
28
29
21
25
17
11
16"
8
5
5
5
5
1
1
1
I
2
3
i
i
i
i
i
i
ipuu»«*n«
Percent
20,7
0 5
1.1
2.3
3.2
29
4,5
2.7
' 4,7
43
45
5.5
4.0
3.7
3.3
2.9
3.6
3.3
3.7
2.7
3.2
2.2
2.3
2.0
i. a
0.8
0.8
0.8
0.8
0.1
0.1
0.1
0.1
0.3
0.4
0.1
O.I
0.1
0.1
0.1
0.1
i jurt
CunuUttvt
"""eoutncy
182
186
1/5
198
223
248
211
302
* ''9
'1
' 3
*5l
482
511
537
SSO
5§«
81*
643
664
* 689
708
724
740
74S
753
751
783
761
769
770
771
772
774
777
771
?7t
780
781
782
783
Cumulative
Percent
20.7
21.2
22.3
25.3
28.5
31.4
35.9
38.6
43.3
47.6
52.1
57.6
61.6
65.3
68.6
71.5
75.1
78.4
82.1
34.8
88.0
90.2
92.5
94.5
95. 5
96.2
96.8
97.4
98.1
98.2
98.3
98.5
98.6
98.9
99.2
99.4
99.5
99.6
99.7
99.9
100. 0
i - 3
-------
Table E - 4
Time to 140° Coolant Temperature
(< » f seconds)
CuRUlitly*
TIHKL
10
20
JO
40
SO
60
70
80
90
100
120
130
140
ISO
160
180
190
200
220
250
290
410
430
460
710
800
1080
1180
1390
rrequtney
1141
6
8
6
a
12
4
6
4
S
I
6
3
3
Z
2
1
1
1
1
1
1
1
1
I
2
I
1
1
=erct«
92,8
as
0 5
0.5
0.7.
1.0
0.3
a.s
0.3
0.4
O.I
o.s
0.2
0,2
0,2
0.2
0.1
0.1
0.
0.
0,
0.
a.
0.1
0.1
0,2
0.1
O.I
0.1
Frequncy
1141
1147
1153
1159
11S7
1179
1183
1139
1193
1198
1199
1205
1208
1211
1213
1215
1218
121?
1218
1219
1220
1221
1222
1223
1224
1228
1227
1228
1229
Cumilativ*
P«re«ni
92.8
93.3
93.8
94.3
95.0
9S.9
96.3
96.7
97.1
97.5
97.6
98.0
98.3
98.5
98,7
98.9
98.9
99.0
99.1
99.2
99.3
99.3
99.4
99.
99.
99.
99.
99.
100.0
E - 4
-------
Figure E-1.
Phase 1 - Speed Distribution by Start Type
Ul
18.00
16.00
0.00
10 15 20 25 30 35 40 45 50 55 60 65 70 75
Speed (Top of range - MPH)
u
80
-------
Figure E-2.
Phase 2 - Speed Distribution by Start Type
18.00
16.00
0.00
0- - - 0
10 15 20 25 30 35 40 45 50 55 60 65 70 75 80
Speed (Top of range MPH)
-------
18.00
16.00
o.oo --
Figure E-3.
Phase 3 - Speed Distribution by Start Type
H h 1 1
u
5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80
Speed (Top of range - MPH)
-------
Figure E-4.
Phase 1 - Cumulative Power Distribution by Start Type
100.00
Cold-Phase 1
Warm-Phase 1
Hoi-Phase 1
30.00
20 40 60 80 100 120 140 160 180
Power (maximum of Range)
200
220
240
260
-------
i . Figure E-5.
Phase 2 Cumulative Power Distribution by Start Type
100.00
Cold-Phase 2
Warm-Phase 2
Hot-Phase 2
30.00
20< 40 60 80 100 120 140 160 180
Power (maximum of Range)
200,
220
240
260
-------
Figure E-6.
Phase 3 - Cumulative Power Distribution by Start Type
100.00
Cold-Phase 3
Warm-Phase 3
Hot-Phase 3
30.00
20 40 60 80
100 120 140 160 180 200 220 240
Power (maximum of Range)
260
-------
Table E - 5
Initial !dl« (MinuttsJ Classes during Trip
3p Atlanta Entirt Trtp Statistics
OBS
31
32
33
34
3S
36
37
38
39
40"
41
42
43
44
45
46
47
48
49
50
51
52 ^
53""
54
Jfttv
/u*,
0.040
0.050
0.063.
0.080
0.100
0.128
0.1S9
0 200
0.212
0.318
0.400
0.504
0.634
0.798
l.OOS
1.265
1.592
2.00S
2.524
3.177
4.000
5.036
6.340
'..-.HI.
io:Mt2":-;:
..._. )Kji
UPPER
0.050
0.063
0.080
0.100
0.128
0.159
0.200
0.252
0.318
0.400
0.504
0.634
0.798
1.005
1.26S
1.592
2 005
2.524
3.17?
4.009
5 038
6.340
7.981
10.04*
i&gffc"-
*uti*varm
COUNT
317 -
59
107
48
77
93
SS
84 '-
' 52
39
25
"23
19
15
16
9
S'_-
4
3
I
I
* f
igr:'~^-^
start
PERCENT
30.6576
0.0000
5.7060
10.3482
4.6422
7 4468
.3.9942
5.3191
6.1896
5.0290
3.7718
2.4178
2.2244
1.8375
1.4507
1.5474 -
0.8704
0,4836
0.0000
0.3888
0.2901
0.0967
0.0967
_~0,Q987__;_..
^**9*k-,-
CUM
30.66
30.66
36,37
46.72
51.36
58.81
67.80
73.12.
79.31
34.34
38.11
90.53
92.71
94.59
96.04
97.59
98.46
98.94
98.94
99.33
99.82
99.72
99.82
99.it
9»-9*T~:r
O'CUM
69.34
69.34
83.63
53.28
48.64
41.19 """
32.20 - "
26.88
20.69
15.66
11.89
9.47
7.21
5.41
3.96
2.41
1.54"
1 .06
1.08
0.67
0.38
0.28
0.18
0-08
-ijkb- .
0.0000 100.02. -0.02
E-ll
-------
Table S - 6
Initial Idle (Minutes) Classes during Trip
3p Atlanta Entire Trip Statistics
OBS
51
62
63
64
65
66
67
60
69
70
7i
n
n
74
75
76
77
73
79
80
8!
82
83
84
OS
86
a?
as
89
90
LOWER
0.040
0 050
0,063
0.080
O.iOO
0.126
0.159
0.200
0.252
0.318
0.400
0.504
0.634
0.798
1. 005
1.265
1.592
2.005
2.524
3.177
4.000
5.036
6.340
7.981
10.048
12.649
15.924
20.047
25.238
31.773
iKin.Ai»not start
UPPER COUNT PERCENT
0.050
0.063
0.080
0.100
0.126
0.159
0.200
0.252
0.318
0.400
O.S04
O.S34
0.798
l.OOS
1.255
1.592
2.005
2.524
3.177
4.000
5.036
6.340
7.981
10.048
12.649
15.924
20.047
25.238
31.773
40.000
891
153
245
108
155
151
!17
56
73-
52
43
27
22
18
14
10
8
2
5
2
2
2
I
1
41.0978
0.0000
7.0572
11.3007
4.9815
7.1494
6.9649
1.396?
3 . 0443
3.3672
2.3985
1.9834
1.2454
1.0148
0.8303
0.64S8
0.4613
0.3690
0.0923
0.2306
0.0923
0.0923
0 . 0000
0.0923
0.0000
0.0000
0.0461
0.0461
0 0000
a. oooo
CUM
41.10
41.10
«8.16
S9.46
64.44
71.59
78.55
83.95
36.99
90.36
92.76
94.74
93.99
97.00
97.83
98.48
98.94
99.31
99.40
99.63
99.72
99.81
99.81
99.90
99.90
99.90
99.95
100.00
100.00
too. oo
QECUK
58.90
53.90
51.84
40.54
35.56
23.41
21.45
16.05
13.01
9.64
7.24
5.26
4.01
3.00
2.17
1.52
1.06
0.69
0.60
0.37
0-28
0.19
0.19
0.10
0.10
0.10
0.05
0.00
0.00
0.00
S-12
-------
Tafcle 1 -7
Initial Idl« (Minute*) C1*ss« during Trie
3p Spokane ft Baltimore Entire Trip
OSS
I
2
3
4
5
6
7
a
9
10
11
12
13
14
IS
16
17
11
19
20
21
22
23
24
:-"zi,
">-"ii
iiV^St?
13HER
0.040
0 050
- 0.063
0.080
0.880
0.126
6.1 Si
0 20*
0.:^
0 . . >'.'
0.400
0.504
0.834
0.798
1. 005
1.265
1.S92
2. COS
2.524
3.177
4.000
5.038
6.340
7.981
...^ lOlWft,,,.
"Vlf-iBif^
-yr.".f!^.-fi
^li.-iM?^'
«.- IK1
UPPER
3.050
0.063
0.080
0.100
0.128
0.159
0.200
0.252
0.318
* 3 400
0.504
0.634
0.798
t.QOS
1,265
1.592
V2.aos
2.524
3.177
4.000
5.038
6.340
7.981
10.048
I!*-84*
-iMt*
te "-"^rjf-^y^j
^?fcS?
rLAI'LOIQ
count
419
63
130
69
128
144
128
. IOS .
104
84
92
80
68
67
.' 51
46
54
38
30
22
27
30
23
..... 17
^-:_a -
r^afs-"^ -
SHfSL
. ««pt
"ERCENT
20.4092
0 0000
4.0429
6.3322
3.3609
6.1374
7.0141
6.1374
5. 1145
S.06S8
4.0916
4.4812
3.8967
3.2148
3.263S
2.8790
2.2408
2.8303
1.7S3S
1.4613
1.0711
1.31SI
1.4613
1.1203
0.8281
'-1 1.1:0. §«>..
-'- ' *'£* ^>
CUM
20.41
Z0.41
2«.4S
38.78
34.14
40.21
47. »
53.43
56.54
63.61
67.70
72. !
76.08
79.29
82.SS
85.23
87.47
90.10
91.8S
93.31
94.31
95.70
97.18
98.26
.».»'
9S.fi
. 91K-WL .
,"«s ..i^S»~rvffc
OECUN
?9.59
79.59
75.55
69.22
65.66
59.72
52.71
46.57
41.46
36.39
32.30
27.82
23.92
20.71
17. 4S
14.77
12. S3
9.90
8. IS
6.69
5.62
4.30
2.84
1.72
0.89
"04S_^
£^ovii?r
:;»;!"
B-13
-------
Table E - 8
Initial Idlt (Minutes) Classes during Trip
3p Spokane & Baltimore Entire Trip
cas
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
4?
48
49
50
51
52
S3
54
55
56
57
SI
59
60
LOWER
0.040
0.050
0.063
0,080
0.100
0.126
0.15S
0.200
0.252
0.318
0.400
0.504
0.634
0.798
1.005
1.26S
1.592
2. 005
2.524
3.177
4.000
S.03S
6.340
7.9M
10.048
12.649
15.924
20.047
25.238
31.773
I HIP
UPPER
0.050
0.063
o.oao
0.100
0.126
0.159
0.200
0,252
0.318
0.400
0.504
0.634
0.798
1.005
1.2SS
1.592
2.005
2.524
3.177
4.000
S.038
6.340
7.981
10.048
12.64S
15.924
20.047
25.238
31.773
40.000
LAI 'Warm
COUNT
604
141
259
LOO
182
225
142
L25
100
89
92
72
56
36
32
28
18
i;
is
20
11
11
1
I
5
2
1
Start
PERCENT
25.3143
0.0000
S.909S
10.3550
4.1911
7 . 6278
9.4300
5.9514
5.2389 '
4.1911
3.7301
3.3SS8
3.0176
2.3470
1.5088
1.3411
1.1735
0.7544
0.712S
0,6708
0.8382
0.4610
0.4610
0.0419
0.0419
0,2098
0.0838
0.0000
0.0000
0.0419
CUM
ZS.31
25.31
31.22
42.07
46.26
53.89
S3. 32
S9.27
74.51
78.70
82.43
88.29
39.31
91.68
93.17
94.51
95.68
96.43
97.14
37.81
98.65
99.11
99.17
99.61
99.65
99.88
99.94
99.94
99.94
99.98
OECUH
74.89
74 , 69
68,78
57.93
53.74
48.11
36.68
30.73
25.49
21.30
17.57
13.71
10.69
3.34
6.83
5.49
4.32
3.57
2.86
Z.19
1.3S
0.39
0.43
0.39
0.3S
0.14
0.08
0.08
0.06
0.02
B-14
-------
Table E - 9
Initial Idle (Minutss) Classti during Trip
3p Spok*n« I Baltimore Enttrt trip
oas
61
62
S3
64
65
61
67
6»
89
70
71
72
, 73
74
75
71
"** 77
78
79
80
«
82
83
84
as
'--.'.' - .... 81 .
- - - "'-':,»'..
-.-''. . .:, -- 8» '.
8i
90
LOWER
0.040
0.050
0.053
0.080
0.100
0.128
0.159
0.200
0.252
0.318
0.400
0.504
0.634
0.798
l.OOS
1.265
1.592
2.005
2.124
3.177
4.000
5.031
6.340
7.911
10.041
12.649
15.924
20.047
25.238
31.771
IK
UPPER
0.050
0.063
0.080
0.100
0.126
6.159
0.200
0.252
0.318
0.400
0.504
0.634
0.798
l.OOS
1.265
1.S92
2.005
2.524
3.177
4.000
5.036
6.340
7.981
10.048
12.849
15.924
20.047
25.238
31.773
40.000
1 PtA 1 »«0t
COUKT
1580
346
SS6
200
334
364
2SO
207
155
138
94
65
n
44
39
19
23
12
12
12
2
5
5
5
2
itart
PIRCEKT
36.2225
0.0000
7.4601
11.9879
4.3122
7.2014
7.8482
5.3903
4.4631
3.3420
2.9323
2.0287
1.4015
1.5308
0.9487
0.8401
0.4097
0.4959
0.2587
0.2587
0.2587
0.0431
0.1078
0.1078
0. 1078
0.0431
0.0000
0.0000
0.000ft
0.0000
"CUH
36.22
36.22
43.68
55.57
59.98
67.18
75.03
30.42
84. U
38.22
91.15
93.18
94.58
98.11
97.06
97.90
98.31
98.81
99.07
99.33
99.59
99.83
99.74
99.8S
99.98
100.00
100.00
100.00
100.00
100.00
OCCUN
63.78
63 . 78
56.32
44.33
40.02
32.82
24.97
19.58
15.12
11.78
3.85
6.82
5.42
3.89
2.94
2.10
1.69
1.19
0.93
0.87
0.41
0.37
0.2S
0.1S
0.04
o;oo
0.00
0.00
0.00
0.00
E-15
-------
Table E - 10
initial Id1« (Minutes) Cl*ss«s during Trip
3p Atlanta Entirt Trip Statistics
DBS LOWER UPPER
1 0,040 0,050
2 0.050 0.063
3 0.063 0.030
4 0.080 . 0.100
5 0 100 0.126
6 0.126 0.1S9
7 O.ISi' *' 1.200
8 0.200 '"' 252
9 0.252 -""' US
10 0.319 3.400
11 0.400 0.504
12 0.504 0.634
13 0.634 0.798
14 0.798 1 005
IS l.OOS 1.265
16 1.2SS l".S92
17 1.592 2.005
ia 2.00S 2.524
19 2.524 3.177
20 3.177 4.000
21 4.000 S.03S
22 S.036 8.340
23 S.340 7.981
24 7.981 10.048
25 10.048 12.649
28 12.649 15.924
27 15.924 20.047
28 20.047 25.238
29 25.238 31.773
30 31.773 40.000
* 1 ~VU 1 t
:OUHT
169
51
71
34
60
SI
36
' 51
39
36
34
24
27
12
9
9
a
5
2
a
4
4
I
2
PERCENT
22.3545
0.0000
6.7460
9.3915
4.4974
7.9365
9.0688
4.7619
6.7460
5.0265
4.7619
4.4974
3. 1748
3.5714
1.5873
1 . 190S
1 . 190S
1.0582
0.6614
0.2646
1.0582
0.5291
Q.S291
0.1323
0.2648
0.0000
0.0000
0.0000
0,0000
0.0000
CUK
22.35
E2.3S
29.10
38.49
42.99
50.93
59.00
63.76
70.51
75.54
30.30
34.80
37.97
91.54
93.13
94.31
95.il
96.57
97.23
97 49
98.55
99.08
99.61
99.74
100.00
100.00
100.00
100.00
100.00
100.00
QECUM
77.65
77.65
70.90
61.51
57.QI
49.07
41.00
36.24
29.49
24.48
19.70
15.20
12.03
3.46
6.97
5.68
4.49
3.43
2.77
2.51
1.45
0.92
0.39
0.28
0.00
0.00
0.00
0.00
0.00
0.00
E-16
-------
|