Final Technical Report on Aggressive
            Driving Behavior for the Revised
            Federal Test Procedure Notice of
            Proposed Rulemaking
&EPA
United States
Environmental Protection
Agency

-------
              Final Technical Report on Aggressive
                 Driving Behavior for the Revised
                 Federal Test Procedure Notice of
                       Proposed Rulemaking
                              Certification Division
                            Office of Air and Radiation
                             Office of Mobile Sources
                         U.S. Environmental Protection Agency
&EPA
United States
Environmental Protection
Agency
EPA-420-R-95-102
January 1995

-------
  Final Technical Report on Aggressive Driving Behavior for the
Revised Federal Test Procedure Notice of Proposed Rulemaking

Section 1. Need for controlling emission from non-FTP Driving
           Behavior                     .
   1.1  History of  the  FTP Driving  Schedule
   1.2  Concerns with., the  FTP  Driving  Schedule
   1.3  CARB Testing

Section 2. In-use Driving Behavior
   2.1  Speed  and Acceleration
   2.2  Driving Behavior Determination
      2.2.1 Transmission  Type
      2.2.2 Vehicle .Performance
      2.2.3 Heavy, light-duty Trucks

Section 3. Emission Impact of Non-FTP Driving
   3.1  Cycle  Development
      3.1.1 Non-LA4 Driving  (REP05)
      3.1.2 Start  and  "Remnant"'Driving  (REM01)
      3.1.3 Other  cycles:  HL07 and ARB02
   3.2  Vehicle Testing           .
      3.2.1 EPA Test Program
      3.2.2.   Vehicle  Manufacturer Test  Program
           3.2.2.1 Load-adjusted Testing
      3.2.3 ARE Test Program

Section 4. Controlling Emissions from Non-FTP Driving Behavior
   4.1  Causes of Emission Increases
      4.1.1 Commanded  Enrichment
           4.1.1.1 Impact of Commanded Enrichment
           4.1.1.2 Comparison of Production and Stoich Bag
                   Results
      4.1.2 Transient  Fuel Control
           4.1.2.1 Throttle Movement
           4.1.2.2 Throttle Movement Modelling

-------
                                           Revised FTP-r-aggr. rpt
           4.1.2.3 Heavy Deceleration  Enrichment
      4.1.3 High Combustion Temperatures
      4.1.4 NOx Catalyst Conversion Efficiency
   4.2   Approaches  to Compliance Testing for Non-FTP Driving
        Behavior-
      4.2.1 Air-Fuel  control cycle
      4.2.2 Representative  cycle
      4.2.3 High speed/load transient control cycle(US06)
      4.2.4 Justification for selecting US06 as preferred option
      4.2.5 US06 adjustments
        4.2.5.1 Vehicle  Performance and Transmission Type
        4.2.5.2 HLDTs
        4.2.5.3 Adjustment  approach
   4.3   Potential Strategies for Controlling Emissions  from Non-
        FTP Driving Behavior
      4.3.1 Improved  Fuel Control Through Calibration
        4.3.1.1 Commanded Enrichment
        4.3.1.2 Transient Enrichment
        4.3.1.3 Heavy Deceleration Enrichment
        4.3.1.4 NOx Catalyst Conversion Efficiency
      4.3.2  Improved Fuel  Control Through Sequential-Fire Port
             Fuel Injection
      4.3.3 Drive-by-Wire Systems
      4.3.4 Improved  Catalysts
      4.3.5 Reasonable Conclusion
   4.4   Level of Control
      4.4.1 HC Control
      4.4.2 CO Control
      4.4.3.  NOx Control

Section 5. Technological Feasibility
   5.1  Impact on Performance
      5.1.1 Power                 .
      5.1.2 Driveability
   5.2.  Impact on Durability    •   .. '   .

-------
                                           Revised FTP--aggr. rpt
      5.2.1  Catalyst Durability
        5.2.1.1   Catalyst  Thermal  Degradation
        5.2.1.2   Production Catalyst Temperatures
        5.2.1.3   Stoich Catalyst Temperatures •
        5.2.1.4   Temperature Differences Between Production and
                 Stoich Calibrations
        5.2.1.5  Summary
            5.2.1.5.1 Underfloor Catalysts
            5.2.1.5.2   Close-Coupled Catalysts
      5.2.2. Engine Durability
   5.3.  Excessive Temperatures and Road Grade
   5.4.  Projected Vehicle  Modifications
      5.4.1. Modifications Necessary to Comply  with Proposed
             Standards
      5.4.2. Modifications Necessary to Offset  Performance and
             Durability Impacts

Section 6.  US06  Procedures
   6.1  Preconditioning
   6.2  Sequencing
   6.3  Manual transmission shift points
   6.4  Adjustments for  vehicle performance
      6.4.1  Performance Criteria
        6.4.1.1  W/P measure
        6.4.1.2  Measure based  on 0 to 60 acceleration  time
      6.4.2  Performance categories
      6.4.2  Adjustment  Calculation
      6.4.3  Requirement for "Super" performance  vehicles
   6.5  Adjustments for  Heavy,  Light-duty Trucks
   6.6  Driving Schedule Tolerances

-------
                                           Revised FTP--aggr. rpt
Final Technical Report on Aggressive Driving Behavior for the
Revised Federal Test Procedure Notice of Proposed Rulemaking

   This  technical report documents the need for certain  proposed
additions to the Federal Test Procedure(FTP)  to  ensure that it
reflects current driving behavior.    The first section provides
background information on the current FTP driving cycle  and
discusses the need for the proposed  modifications to the FTP.  In
section two, information is presented on the differences between
in-use driving and the FTP.  This  is followed, in section three,
by a summary of.  emission testing  results,  which quantify the
emission impact of the non-FTP driving.   Methods for controlling
emissions from non-FTP driving are discussed in  section four.
This section also discusses the  appropriate level of control, as
well as adjustments for special  cases.    Section five reviews
feasibility issues, followed by  a  cost and benefits discussion in
section six.  The final section  presents a discussion of the
required test procedures.

Section 1. Need  for Controlling  Emissions from Non-FTP Driving
           Behavior

1.1  Background on the FTP Driving Schedule
    The  FTP  driving schedule is  a principal component of the
exhaust emission test.   As designed,  the FTP was intended to
represent typical driving patterns in primarily  urban areas.  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-1960s1 by trial-and-error  to
     1G.C. Hass, et. al.,  "Laboratory Simulation of Driving Conditions in the
Los Angeles Area," SAE  Paper No. 660546, August, 1966.

-------
                                           Revised FTP--aggr. rpt
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 a  1970   effort to develop an improved Federal Test
Procedure  (based on  speed-time distributions rather than manifold
                      j
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 more speed
variation than the other five  and  was discarded.   The other five
traces were surprisingly similar.   Of those  five,  the trace with
the actual time closest to  the average was selected as the most
representative speed-time trace.   That trace contained 28
segments of non-zero speed  activity separated by idle periods
(these segments are  commonly referred as  hills,  or microtrips)
                                      »
and had an average speed of 19.2 miles per hour  (figure 1-la).
   Based on a  1969  report  on driving patterns in Los Angeles,2
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 mph.   Slight
     2D.H. Dearm and R.L.  Lamoureux, "Survey of Average Driving Patterns in
the Los Angeles Urban Area," TM-(L)-4119/000/01, February 28, 1969.

-------
                                           Revised FTP--aggr. rpt
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  officially called the   the Urban  Dynamometer
Driving Schedule (UDDS), but more commonly referred to as the
LA4.   (The remainder of this report will use the term LA4 when
referring to the current FTP driving schedule).  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
customer service than running just one 7.46 mile cycle from a
cold start.

1.2  Concerns with the FTP Driving Schedule
   The LA4 has been an  critical  component of EPA's strategy for
reducing vehicle exhaust emissions; however, the driving schedule
was developed over twenty-five years ago and EPA's initial review
identified a number of  concerns:

     Speed.   The maximum .speed on the LA4 is 57 mph.  Even in
     urban areas,  limiting the speed to 57 mph is clearly
     missing a significant portion of in-use operation.

     Acceleration.   Acceleration rates on the LA4  were
     artificially reduced to accommodate the capabilities of the

-------
                                           Revised FTP--aggr. rpt
     testing equipment.  The LA4 was targeted to  average  driving,
     thus it fails to capture aggressive driving    Current-
     technology dynamometers permit higher, more  representative
     accelerations.

     Road grade. There is no attempt to account for road  grade  in
     the LA4.   In some urban areas, the extra  load placed on the
     engine can be considerable.

     Speed variation..  The methodology used in  the  development  of
     the LA4 led to a fairly smooth driving schedule and  may not
     represent small timescale variation in vehicle speed.

The above concerns regarding the LA4's representativeness of in-
use driving behavior is ultimately a concern that the emission
control demonstrated by a vehicle when tested on  the LA4  may not
be  translated into the same level of emission  control  in use.

1.3  CARB Testing
   At the  start  of  the  FTP  Review project,  limited information
existed on the emission impact of non-LA4  driving  behavior.  In
1990,  the California Air Resources Board (CARB) conducted
emission testing in order to get a preliminary  assessment of the
emission impact of  high acceleration rates; acceleration rates
greater than those on the LA43.   Ten late-model vehicles were
tested over an engineered driving schedule consisting of  nine
acceleration modes developed to simulate various  types  of
acceleration events.  Relative to LA4-like accelerations,  CO
emissions increased very dramatically during most of the  other
accelerations. HC and NOx also showed large increases,  although
there was large variation in the emission response  across
     3State of California Memorandum,  from Mark Carlock to K.D. Drachand

-------
                                           Revised FTP--aggr. rpt
vehicles and acceleration modes.   It  was calculated if the FTP
was modified to add a  single  acceleration mode lasting only 16
seconds, CO  emissions could  double and HC emissions could
increase by nearly 20%.
   The CARB results provided preliminary  evidence  that  non-LA4.
operation can result in high  emissions.   However,  the emission
response of the 10 vehicles was  extremely varied across the 10
acceleration modes and the CARB  data  did not address the in-use
frequency of such behavior  Thus,  information was still needed to
identify the range and frequency of non-LA4  driving which occurs
in-use.  At the start  of FTP  Review Project, EPA surveyed
existing driving behavior data to access the importance of the
above concerns.  It quickly became apparent  that very little
information existed on the real  world driving behavior.   As a
result,  a major portion of the  project involved conducting and
analyzing results from a large scale  in-use  driving survey.

Section 2.   In-use Driving Behavior
         \
   In  a coordinated research effort,  EPA collaborated with the
American Automobile Manufacturers Association (AAMA),  the
Association of International  Automobile Manufacturers (AIAM),  and
the California Air Resources  Board (CARB)  over the spring and
summer of 1992 to conduct surveys of  in-use  driving and soak
behavior in four major U.S. cities.4   The Agency employed  two
survey methods to gather basic data on the  speeds and
accelerations found in actual in-use  driving.   In the
"instrumented vehicle"  approach,  digital dataloggers were
installed in private owner vehicles to record second-by-second
speed and engine parameter data  over  a period of seven to ten
days.   Separate "chase car" studies used laser rangefinder
     4For a detailed description of the driving surveys and results, see
"Federal Test Procedure Review Project: Preliminary Technical Report," May
1993, EPA 420-R-93-007

-------
                                           Revised FTP--aggr. rpt
technology in a "patrol" vehicle  to  calculate  vehicle speed of
targeted in-use vehicles operated over  predetermined routes.

   The  instrumented vehicle surveys were conducted on a sample of
150 vehicles ,in Baltimore, Maryland,  and  144 vehicles in Spokane,
Washington.  An additional 101 vehicles were instrumented in
Atlanta, Georgia, in a cooperative effort  between EPA's Office of
Research and Development and  the  Georgia  Institute of Technology.
Chase car studies funded by EPA were  conducted on 218 routes in
Baltimore and 249 routes in Spokane;  CARB-funded chase-car work
was performed on 102 routes in Los Angeles.  The critical
findings for the FTP review project  are discussed below.

2.1  Speed and Acceleration
   In May of  1993,  EPA published its initial conclusions
regarding aggressive driving  behavior in  the "Federal Test
Procedure Review Project:  Preliminary  Technical Report."5
These findings were largely based on  the  Baltimore,  instrumented
vehicle survey data; subsequent analysis  has been completed on
the larger, 3-city instrumented vehicle database,  and the 3-city
results were found to be consistent with  the Baltimore only
results(table 1-la).  The 3-city  analysis  showed that nearly 13
percent of vehicle operation  time occurs  at combinations of speed
and acceleration that fall outside the  matrix  of speeds and
accelerations found on the LA4.   The  maximum observed in-use
speed in was 95.5 mph, compared, to the  LA4 maximum speed of 56.7
mph, and sightly more than 7  percent  of in-use vehicle operation
time was spect at speeds greater  than 60  mph.   Average speed from
the 3-city in-use data was 25.9 mph  compared to 19.6 mph over the
LA4 .
     5U.S. EPA, "Federal Test Procedure Review Project:  Preliminary Technical
Report,"  May 1993, -EPA 420-R-93-007

-------
                                    Revised FTP--aggr. rpt
                    Table 1-la  .
           i

Comparison of Driving Behavior  for Four Cities
Driving
Behavior
Measure
Speed (mph)
Average
Maximum
Standard
deviation
Number of
seconds
Acceleratio
n
(mph/sec. )
Minimum
Maximum
Standard
deviation
Number of
seconds
Power
(mphVsec.)
Average
Maximum
Standard
deviation
Number of
seconds
Average
trip length
(miles)
Average
trip time
(minutes)
Average
distance
b/w stops
Baltimore
instr.
veh

24.50
94.46
20.52
3,365,504

-19.49
15.19
1.50
3,360,550

46.02
557.69
42.96
1,407,908
4.89
12.03
0.87
Spokane
instr. veh

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
Atlanta
instr. veh

28.84
96.48
22.61
3,339,489

-18.57
16.69
1.54
3,335,057

51.99
723.12
48.06
1,463,313
6.32
13.16
1.08
3-City
instr.
veh
average

25.85
96.48
20.87
8,786,192

-19.49
16.69
1.50
8,772,615

46.97
723.12
44.79
3,751,479
4.99
11.59
0.88
Los
Angeles
chase
car

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

-------
                                             Revised FTP--aggr. rpt
Percent
idle
operation
21.12


17.91


17.47


18.97


11.78


   Another speed-based measure, specific  power6,  is  useful when
analyzing  in-use driving behavior.   Measures of  power also
indicated  that in-use driving  behavior was more  aggressive than
reflected  in the LA4.  Specific  power in the 3-city sample ranged
up to  723  mph2/sec and averaged  47.0 mph2/sec; the  LA4 has
maximum power of 192 mph2/sec  and an average of  38.6 mph2/sec.

2.1.1  Microtransient Operation
   The previous  discussion of in-use  speeds  and accelerations
presents a snapshot of driving behavior.  While the  acceleration
measure, .which looks at the  chancre in speed from one second to
the next,  partially characterizes the transient  nature of
driving, there are other measures which expand the  time interval
to examine the small-scale deviations in speed  (microtransients).
One measure,  referred to as  jerk,  is equal to the change in
accelerations.   Using speed  data collected and averaged on a one
second basis,  the jerk measure expands the picture  of driving out
to three seconds.  A related measure is change in specific power.
which  is second-to-second change in power.  Conceptually, this
measure captures the change  in the power requirement imposed by
the driving behavior.
   EPA used the  3-parameter instrumented  vehicle data from
Baltimore,  Spokane, and Atlanta,  to calculate the microtransient
measures for in-use driving  behavior and compared the results to
     6The power needed from an engine to move a vehicle is proportional to
both the vehicle speed and the acceleration rate.  Thus, neither variable, by
itself, is a good measure of the load placed on the engine.  The joint
distribution  of speed and acceleration is probably the best measure, but it
must be examined in three dimensions, which is difficult to visualize and
comprehend.  The concept of specific power, calculated as the difference in
the square of velocity from one second to the next, provides a two-dimensional
measure which is roughly equal to 2*speed*acceleration and has the units of
mph2/second.

                                  8

-------
                                          ' Revised FTP--aggr. rpt
the LA4's representation on in-use driving behavior.  The
measures of jerk and change in power are shown below in the table
1-ib:
                            Table 1-lb
           Measures of microtransient driving behavior
Measure

Jerk
Change in power
Mean of the
absolute values
In-use
.47
20.48
LA4
.36
14.96
Standard deviation
In-use
.89
34.36
LA4
.63
22.96
For both jerk and change in power, the mean of the absolute
values were used in order to look at both the positive and
negative values  (the mean of the signed values of jerk is always
equal to zero).   The in-use means were higher than those for the
LA4 indicating larger in-use changes in acceleration and power,
as well as reflecting, in part, the LA4's acceleration rate
cutoff of 3.3 mph/sec and the maximum speed of 57 mph.  The
standard deviations of jerk and change in power is probably a
better measure of microtransient behavior.  Again, in-use data
show larger values for both measures.  The greater variation
around the mean demonstrated by the in-use data suggests that the
LA4 does not adequately represent the microtransient nature of
in-use driving behavior.

2.2  Road Grade
   To properly evaluate  the  need  to  incorporate  road  into  the
existing FTP, information is needed regard in-use road grade and

-------
associated driving, as well as  the  emission  impact  of road grade.
Ideally, the in-use data would  include  the fraction of in-use
operation driven on roadways by level of  grade.   It is also
important to understand the relationship  between road grade and
driving behavior.  If driving behavior is  independent of road
grade, then the presence of road grade  will  increase the severity
of the driving by  increasing the engine load.   If road grade
affects driving behavior then the impact  of  road grade will be
diminished or eliminated. It is likely  that  a  severe grade will
require an additional load on the engine  even  with  "conservative"
driving.

EPA's review of existing data found comprehensive information on
road.grade to be quite limited.  A  1980 EPA  report  summarized
Department of Transportation Data on the  nationwide distribution
of road gradient by the percent of  vehicle-miles-travel (VMT).7
From these data-r-.the average positive road grade was 1.66
percent.  Roughly  6 percent of  VMT  was  spent on grades of 4
percent or higher.  EPA did not find any  information on road
grade and driving  behavior.  As a.result, the  in-use driving
survey included instrumentation to  collect realtime road grade
data.   As part of  the chase-car study,  a  gyroscope  was installed
to collect such data.  Unfortunately, a combination of limited
equipment precision and noise introduced  by  vehicle movement
resulted in inadequate road grade data  to evaluate  the
relationship between driving behavior and road grade.   Thus,  EPA
has/very limited in-use data for evaluating  road grade.

2.3  Driving Behavior Determinants
   There  are  a  host of  factors  which can influence driving
behavior; vehicle  characteristics which may  impact  driving
behavior need to be considered  when looking  at revisions to the
     7U.S. Environmental Protection Agency,  "Passenger Car Fuel Economy:  EPA
and Road," Report No.  EPA 460/3-80-010,  September 1980, p.119.
                                10

-------
FTP.  In analyzing in-use driving behavior, EPA identified three
important vehicle characteristics:  transmission type
(manual/automatic); vehicle performance(high/low powered); and
vehicle type  (ie. cars/trucks).

2.3.1  Transmission Type
   Among  the  166  vehicles  from Spokane  and Baltimore,  60  vehicles
had manual transmissions and 106 vehicles had automatic
transmissions.  In looking at the in-use vehicles with aggressive
driving,  it appeared that the most of the vehicles which were
driven aggressive were manual transmission vehicles.  This
finding suggested the need for a more detailed look at
differences in manual and automatic transmissions.  A comparison
of aggressive driving  for automatic and manual transmission
vehicles was made using specific power.  Specifically, for each
vehicle we calculated fraction of vehicle operation above
specific power .of. 200 (FTP maximum is 192).  This was repeated
using specific power of 300, as a somewhat arbitrary measure of
extreme operation.  Table 1-2 presents summary statistics for
these two measures of aggressive driving.  Manual transmissions
had  higher mean values for both measures than automatic
transmission vehicles and manuals showed much more of a spread
across vehicles as indicated by the larger standard deviations.
Figures 1-1 and 1-2  present the frequency distributions for high
power operation  (power=300)  for manual and automatic
transmissions vehicles.   These charts show this greater variation
for/manual transmission vehicles relative to automatic
transmission vehicles, as well as showing the manual transmission
vehicles with the largest fraction of high power operation.

2.3.2  Vehicle Performance
   The -performance  of  the in-use  fleet  of vehicles  was measured
by calculating the ratio of a  vehicle's weight(W) to its peak
horsepower (P).   In general terms, performance relates to a
vehicle's  ability to deliver power from the engine when demanded
                                11

-------
by the driver, such as during an acceleration.   A common
performance measure is the time to accelerate from 0 to 60;  the
faster the time, the higher the performance.  In using W/P as a
performance measure, a low W/P value is associated with a high
performance vehicle while a low performance vehicle would have a
high W/P value.  Weight to power ratios were calculated using
information supplied by the vehicle manufacturers.   The W/P is a
good indicator of performance, although it does not account for.
differences in torque curves or aerodynamic design.  Further, an
in-use vehicle's performance is also'determined by the physical
condition of the vehicle--poorly maintained or malfunctioning
vehicles will show a performance loss relative to the
manufacturers specifications.  Nonetheless, the ratio of weight
to power is a useful indicator.  The W/P measure also does
account for performance differences of automatic and manual
transmission.  For a given W/P a manual, transmission vehicle will
show a higher performance(as measured by 0 to 60 times) than a
comparable automatic transmission vehicle.  The analysis below
treats automatic and manual vehicles separately.
   The  analysis  looked  at differences  in driving behavior  as  a
function of vehicle performance. Specifically, the analysis tried
to answer the question of whether higher performance vehicles are
driven more "aggressively"  than lower performance vehicles.
Driving aggressiveness was  measured by first calculating,  for
each vehicle, the fraction of time spent at or above specific
power values, using intervals of ten.  The next step was to rank
the^vehicles according to their percent time spent in each of
these categories.   Finally,   in each category,  the vehicles were
separated into three groups:   the bottom ten percent(least
aggressive); the middle 80  percent (normal); and the top ten
percent (most aggressive).    Once the vehicles were categorized
as either high,  middle or low, the average W/P for  each group in
each power category was obtained.
   For  manual  transmission vehicles,  figure 1-3  shows that  the
most aggressive vehicles consistently had a lower mean  W/P  than
                                12

-------
the normal vehicles, for operation in the non-FTP operation  (the
intervals above 200).  The least aggressive manual transmission
vehicles had a higher mean W/P than the normal vehicles  and
spent no time at all at power values above 200.  For automatic
transmission vehicles, there was only a small difference in mean
W/P between most aggressive and normal vehicles  (figure 1-4).
The mean W/P for the least aggressive vehicles, however, was
substantially higher than the normal vehicles
   For manual  transmission  vehicles  these  results indicate  that
W/P, as a proxy for vehicle performance, is correlated with
driving behavior, higher performance vehicles  (low W/P) tend to
be driven more aggressively than lower performance vehicles  (high
W/P).   The in-use data for automatic transmission does suggest
that lower performance vehicles (high W/P.) are driven less
aggressively than "normal" vehicles, but  the vehicles driven
most aggressive aren't necessarily  high performance vehicles.
The results suggest that the driver is an important, but
unquantifiable, factor in the driving behavior.
   EPA considers  the conclusions on  vehicle performance to  be
preliminary.  The Spokane/Baltimore database had a very limited
number of high performance vehicles.  There were only 3 vehicles
with a W/P below 20--two automatic and one manual transmission.
To increase the sample of high performance vehicles, EPA is
updating the analysis  to include vehicles from the Atlanta in-.
use driving survey.

2.3*3   Heavy,  light-duty Trucks
   The  initial  analysis presented  in the Preliminary Technical
Report (footnote) showed very little difference between cars and
trucks.  As a result of the vehicle test program (discussed in
section 3), additional analysis of the in-use data was conducted
using -a further categorization of trucks since  the light duty-
truck classification covers a broad range of vehicles.  EPA
classifies light duty trucks into light,  light-duty(LLDT) and
heavy, light-duty(HLDT).   It became apparent during the test
                               13

-------
program discussed below that the HLDT may need to be treated  in a
different manner than LLDTs and LDVs.
   The in-use data  provided a  limited data  set to .examine the  in-
use behavior of HLDTs.  For this analysis,  eight vehicles from
the in-use data set were identified as HLDTS  (gross vehicle
weight > 6,00lbs.).  It is important to note  that there  is no
information of the physical load these vehicles carried.  It  is
possible some or all of the vehicles, for some or all of the
time, were subject to extra load which may impact the driving
behavior.
   The speed/acceleration  distribution of the HLDTs  were compared
to the speed/acceleration distributions, for all vehicles.  The
fraction of time spent at speeds above 50 mph was much  smaller
for the HLDTs compared to the rest of the fleet. This difference
increases with increasing speeds above 50 mph.  In terms of
acceleration, there was one HLDT which was driven very
aggressively, bu£_ only at speeds below 50 mph  Thus, the
acceleration distribution by speed range was  fairly similar for
the two groups up to 50 mph.  However, above  50 mph, the
distribution at higher acceleration rates dropped off faster  than
the overall decrease for the speed range,  indicating that HLDTs
were also driven less aggressively during the limited time spent
•at higher speeds.  This  limited data suggests that the driving
behavior of HLDTs is likely to be'different than the fleet at
large.
   In  summary,  high performance, manual transmission vehicles
wer^ driven in a more aggressive manner than the broad, mid-
performance category.   At the other end,  the low performance
vehicles were driven somewhat less aggressively.   EPA also
found that the heavy,  light-duty trucks(HLDTs) tended to be
driven at lower speeds than other light-duty vehicles,  and when
driven^at higher speeds, their accelerations were typically less
severe.
                                14

-------
Section 3.   Emission Impact of Non-FTP Driving
   After analyzing  the driving patterns data, the next  step is
to assess the resulting  exhaust  emissions.   This section
discusses the vehicle  emission test programs carried out by EPA,
ARE, and the vehicle manufacturers.  A fleet of  vehicles were
tested on new cycles developed from the in-use driving survey
data.  The testing  served  two  objectives:  a quantitative
assessment of the emission impact  of non-FTP driving,  and an
evaluation of alternative  control  cycles.

3.1  Cycle Development
   The  first  step in assessing the  emissions impact of the  non-
LA4 driving, is  the reduction  and  synthesis of the driving data
into representative driving cycles for use  in vehicle testing.
EPA's approach to cycle  development involved the selection of
actual segments  of  in-use  driving  which best matched the joint
distribution of__Ln-use speed and acceleration.  In order to
maintain a high  level of coordination between the EPA and CARB,
the data set used in developing  the> in-use  cycles was the driving
survey data from EPA's Baltimore instrumented vehicle study and
data from the ARB's Los  Angeles  chase car  study.8 EPA developed
separate cycles  for start  driving  and aggressive driving(non-
LA4), and to complete the  representation of in-use driving
behavior for emission assessment purposes,  a third cycle,  the
Remnant cycle, was developed to  characterize in-use  driving
behavior not represented by either the start or  aggressive
driving cycle.   EPA developed  individual cycles  rather than a
single "representative"  cycle  in'order to  evaluate EPA's areas of
concern independently.   This is  most critical  in the case of
aggressive driving where both  capturing the diversity of
aggressive driving behavior and  representing it  proportionally in
      .A follow-up analysis compared the Baltimore/Los Angeles database to the
3-city, instrumented vehicle database.  The differences were not large and EPA
does not believe they would materially effect the cycles.
                                15             .

-------
a single cycle covering all  in-use  operation would lead to a very
long cycle.

  Under contract with EPA,  Sierra Research developed driving
cycles  intended  to represent  the  range  of  in-use vehicle
operation.  In generating a  cycle,  entire  micro-trips  (idle-to-
idle) were the basic building blocks  used  to match the  "target
surface" of the  joint distribution  of speed and acceleration.
The cycle generation software developed for.this task uses an
iterative technique to find  the combination of  microtrips which
best match the target surface.  The first  step  involves the
random  selection of a specified number  of  microtrips.   Their
speed-acceleration surface is computed  and compared to  the target
surface.  The software then  searches  for the microtrip  which
provides the best incremental fit to  the target surface.  This
micro-trip is then added to  the cycle and  the process  is repeated
until the desired cycle length is reached.   In  this manner, a
large number of  cycles were  generated (several  thousand).   The
candidate cycles were then ranked according to  how well they
matched  speed-acceleration  distribution of the target  surface i
order to select  the best cycle.9

3.1.1   Start Driving   (ST01)
  For  the  Start  (ST01)  cycle, three target surfaces were
developed from the database,  representing  three successive 80-
second  segments of in-use driving immediately following the
initial idle.  The combinations of  speed and acceleration found
in these distributions could  largely  be found in the LA4,  but
with different percentages and in a different sequence.   The
microtrips that produced the  best fit to these  surfaces, together
with an initial  idle period  that  best matched in-use initial
     9For a detailed discussion of the cycle development see the contractor
report,  "Development of Driving Cycles to Represent Light-Duty Vehicle
Operation in Urban Areas."
                                16

-------
idles, generated a start cycle that was  257  seconds  long (figure
1-5) .  Testing using ST01 allowed separate determination of  start
driving emissions;  ST01 was also used to quantify the  emissions
effects of varying soak duration.

3.1.2  Aggressive Driving  (REP05)
   The second cycle,  characterizing aggressive driving,  was the
Representative Non-LA4 cycle  (REP05) . This cycle  targeted speeds
and accelerations, as well as microtransient  effects, not covered
by the current LA4 .  The in-use data points  used  in  developing
the REP05 target surface were those with combinations of speed .
and acceleration that were not represented on the LA4 cycle (non-
LA4) and, in addition, were not part of  the  ST01  target surfaces.
These points tended to be either high-speed  or high-acceleration
(or both) .   By assembling the cycle from actual idle-to-idle
driving segments, however, the cycle necessarily  included some
speed/acceleraticn combinations that were represented on the  LA4,
amounting to about 30 percent of the cycle's  1400 seconds.  The
average speed of REP05 is 51.5 mph, the  maximum speed is 80.3
mph, and the maximum acceleration. rate is 8.5 mph/sec  (figure 1-
6a) .                          '

3.1.3  Remnant Cycle
   The Remnant  cycle  was; intendfe^^fef%eprfesent the balance of in-
use driving riot already rcaverMC-^^STpl-%^i(p05: ; Thus,  the
        .       :•;-:••:• v~.v'vvXv.v.r -.': ~r^'i^"V-''V;-' : '""'•'•" ; "-'"""
Remnant target surf ace"' Wad obfcained^by using? the  remaining
spe^d/acceleration ^distribution after subtracting that  found  in
the in REP05  and STO1 .  Though much of the 1237  seconds  in the
Remnant cycle is LA4-like driving, there are some non-LA4
segments (at low speeds, with high acceleration rates)  which  were
not captured by the REP05 cycle  (figure l-6b) .  In addition,  the
Remnant cycle has greater speed variation than is found on the
LA4.
                                17

-------
3.1.4  Representation of Microtransients
   The  three  in-use  cycles  had in common a representation of
microtransient operation.  Table l-2b summarizes the  -
characteristics of the cycles as well as the LA4's
characteristics.
                            Table l-2b
                Microtransient characteristics  of
                      in-use driving cycles

Mean:

speed
power
jerk
change in
power
Standard
deviation; -~—-
speed
power
jerk
change in
power
Start

19.37
21.59
0.56
19.45

13.49
37.82
0.88
28.92
Remant

1.8.87
17.66
0.49
17.25

16.93
57.17
0.98
30.26
REP 05

51.50
35.54
0.64
54.97

20.16
57.92
1.02
78.07
LA4

19.60
15.30
6.36
14.96

14.70
27.47
0.63
22.96
3.1.5  Other cycles:   HL07 and ARB02
   The ARE02  cycle  (figure-1-7h was  developed  by  GARB  based on
data from their Los Angeles chase car study.  The purpose of the
cycle is to test vehicles over in-use operation outside  the
boundary of the LA4,  including  extreme in-use driving events.
The HL07 (figure 1-8)- is an engineered cycle developed by EPA in
coordination with the auto manufacturers.  The purpose of this
cycle is to test vehicles on a series of acceleration events over
a range of speeds.   The severity of the accelerations are such
that most vehicles  will go into wide open throttle.
                                18

-------
3.2  Vehicle Testing
   The  coordinated  effort developed in the driving survey
program carried over to the vehicle testing phase.  Beginning  in
the Spring of 1993, EPA, ARE, and the vehicle manufacturers
worked together to develop a cooperative vehicle test program.
Emission testing was conducted at each of  the agencies'
facilities while all of the testing sponsored by the vehicle
manufacturers' was carried out at GM's Milford facility.

3.2.1  EPA Test Program
   The  principal  objective of  EPA's  FTP Test Program was to
assess the emissions from well-maintained, current  technology
vehicles over the in-use cycles.  Nine, 1991-1993 model year
vehicles representing a range of vehicle and engine types  were
selected for the test program.  Table 1-3  describes the eight
vehicles which completed the program  (one  vehicle was lost due to
malfunctions)	Baseline FTP tests were run first to ensure
vehicles met current standards.   All of the in-use testing was
designed to test the vehicle in a hot, stabilized condition--both
the engine and catalyst are stabilized at  their normal  (hot)
operating condition.  These conditions were selected in order  to
look at emissions, and differences in emissions,  associated with
driving behavior.  Replicate .tests; were run for each cycle.
Testing was conducted on EPA'.;sl;iri-.use cycles (REP05, REM01) ,
ARE's ARB02 cycle,; ^n^:-'theV:Mi'b^;-'C^6|e-/-;';-'-Vehicles were also
tested on  the LA4, again: :in^a^yhbtv':"sta"bilized condition, in order
to Compare emissions from LA4.driving to emissions  from in-use
driving.   EPA completed the base testing  in August of 1993.
   Figures  1-9 and 1-10  presents-a summary of the  emission
results for the 8 vehicles in EPA's test program.   The large
differences seen between the non-FTP emissions and  the LA4
emissions must be placed in the proper context by applying
appropriate weighing factors to reflect the fraction of in-use
operation represented by the specific cycle/  The weights  shown
in table 1-4  correspond to the fraction of in-use  miles
                                19

-------
represented by the cycles; these weights were developed from the
in-use driving survey data as part of the cycle development
effort..   The weighted emission results are shown in table 1-5.
The weighted "in-use" emissions are significantly higher than the
weighted, warm,  stabilized  FTP emissions.  NMHC increased by
0.04 grams/mile,  CO rose by 2.8 grams/mile, and NOx rose by 0.08
grams/mile.  This increase in in-use emissions relative to the
FTP cannot be attributed to a single driving mode or condition.
In evaluating the relative significance of the in-use components,
their contribution to the increase was calculated as:
    (In-use component  -  warm, stabilized  FTP)   x   in-use
component weighing factor

The total difference between in-use and hot, stabilized FTP
emissions is the sum of the weighted differences.
   Table  1-6  shows each  in-use  component's contribution as  a
percent of the-total increase.   The results suggest that while
"REPQ5" driving accounts- for a large fraction of the increase  (31
to 58%),  the other in-use driving modes also make significant
contributions,  with more for HC and NOx than CO.   Some of the
observed emission increases were unexpected.  For example,
substantial emission increases for all three pollutants were
observed on the Remnant cycle;  such increases were not predicted
given the Remnant cycle's sil&ilaJrity to the FTP in speed and
acceleration.

3.2/2.  Vehicle Manufacturer Test Program
   Vehicle  testing sponsored  by the American Automobile
Manufacturers Association (AAMA) and the Association of
International Automobile Manufacturers (AIAM)  greatly enhanced
the EPA's database on off-cycle emissions.  The manufacturer's
program included 26 late model vehicles representing 7 vehicle
manufacturers.   Testing included real-time measurement of engine-
out and tailpipe emissions,  as well as various engine parameters.
The second-by-second data were helpful in understanding off-cycle
                                20

-------
emissions. This program also provided a unique opportunity  to
look at a potential strategy for controlling offcycle emission.
A subset of 15 vehicles went through a second phase of testing
after each vehicle's  calibration was changed to eliminate
commanded enrichment.  Much was learned in the comparison of
emissions from the production and "stoich" (no commanded
enrichment)  configurations of these vehicles.  The  vehicle
manufacturers' program began shortly after the completion of EPA
testing and was completed by the Spring of 1994.
   The  program's  emphasis  was  the  testing  of  potential  control  •
cycles.  Testing was conducted on REP05, ARB02, FTP, and HL07
cycles; very little testing was done on the REM01 due to the fact
that it was principally thought of as an emission assessment
cycle(to assess the amount of  emissions generated in use), not a
control cycle  (to control  in-use emission as part of a test
procedure).   Unfortunately, this omission makes it impossible to
do a full in-us£._emission assessment with the manufacturer data,
and thus, a direct comparison cannot be made with EPA's results
on the difference between in-use and FTP emissions.  However, two
 t
comparisons can be made.  First, the average results from the two
programs over the REP05 and ARB02 cycles are compared to see if
the levels of off-cycle emissions are in general agreement.
Second, to look at the increase between offcycle and hot,
stabilized FTP emissions,  a comparison can be made of the
difference of REP05 and hot stabilized FTP emissions from the two
test programs.
   Table  1-7 provides  a  summary  of the  results  from EPA's  and the
vehicle manufacturers'  test programs.   The average emissions by
vehicle type are pretty consistent of the two programs.  For
LDVs,  average NMHC emissions are slightly higher for the
manufacturers test program, while EPA testing showed  higher
average CO emissions and a slightly larger difference between
REP05 and FTP emissions.  Average NOx emissions were
substantially higher for the vehicle manufacturers' tested
 s
vehicles, as was the REP05 and FTP difference.
                                21

-------
   A comparison  of  light-  duty  truck  results  shows  somewhat
larger average "differences between the two programs.  Over the
REP05 and ARB02 cycles, NMHC emissions are pretty similar for the
two programs; however, EPA testing showed much lower  emissions
for hot FTP driving and as a consequence, the REP05 and  FTP
difference is higher for EPA testing than manufacturer testing.
The two programs had similar average CO emissions for LDTs.   In
contrast, NOx emissions were much higher  for the vehicles tested
by the manufacturers, as was the difference between REP05 and FTP
emissions.   Only the manufacturer test program tested heavy
light duty trucks.
   On average, the  two programs showed  consistent   emission
results.  The largest discrepancy was found for NOx emissions, in
which case EPA vehicles had substantially lower emissions.  The
consistency of the two programs' average emissions gives support
to EPA's emission assessment based on the 8 vehicles tested by
EPA.          	

3.2.2.1      Load-  adjusted Testing
     In addition to the testing discussed above,  a subset of
vehicles were tested after making adjustments to the dynamometer
load settings.  These adjustments were made to vehicles which
fell more than 1.5 seconds behind'the speed-time trace of the
REP05, HL07,  or ARB02 cycles..  The auto manufacturers were
concerned that portions of the high speed/load cycles were too
aggressive and "unrepresentative" for some vehicles,  such as
lowar performance vehicles.   The adjustment was intended  to
allow the vehicle to follow the trace within the 1.5 second
tolerance band.   Two of the three dynamometer load coefficients
were candidates for adjustment.   If the out of tolerance event
occurred at speeds less than 50 mph,  the inertia,  or A
coefficient,  was adjusted.  For such events with speeds above 50
mph, the aerodynamic drag, or C coefficient was adjusted.  For
some vehicles and some cycles,  adjustments were necessary on both
coefficients.  The load adjustment was applied to the entire
                                22

-------
driving cycle, not just the portions where the out of tolerance
event occurred.  Thirteen of the 28 vehicles in the
manufacturers/ test program required an adjustment on at least
one cycle. The EPA program saw only one vehicle with'the need for
a load adjustment and that was on the HL07 cycle.
   In  practice,  the  determination  of  the percent  adjustment  was
an iterative and an imprecise process.   The adjustment was
successful  in allowing the vehicle to meet the 1.5 second
tolerance band.  It is impossible, however,  to evaluate whether
the'specific adjustments were the appropriate amount or whether
they were excessive.  Further, adjusting the inertia for the
entire cycle decreases the severity for the entire cycle, while
only a very small portion of the cycle  "needed" the adjustment.

3.2.3 GARB Test Program
   The nature  and scope  CARB  tested paralleled  the  EPA  and
AAMA\AIAM programs.   One significant difference was CARB's use of
a twin-roll dynamometer instead of a large, single-roll
dynamometer.  EPA has limited its analysis to data collected
using the single-roll dynamometer;  future testing by CARB will
be on a single-roll dynamometer and EPA will consider such data,
as is appropriate.

3.3  Emissions associated with Road Grade
   Road grades  affect  on emissions results  from the increased
engine load above that which is associated with the driving
behavior alone.  The higher mass"flow associated with the
increased load will produce higher emission for all three
pollutants.  The extra load associated with road grade can also
increase the frequency or extend the duration of enrichment
resulting in large increases in CO,  and to a lesser extent HC.
While EPA lacked the in-use information necessary to conduct a
full emission assessment,  emission testing with simulated road
grade was conducted on several vehicles to get a sense of the
emissions increase associated with road grade.  EPA tested 3
                                23

-------
vehicles over the in-use driving cycles with a 2 Percent road
grade added by means of increasing the inertia load of the
dynamometer.
   The  HC  increase was  consistent  across  the  three  vehicles
averaging only .04 grams/mile.  The CO increase avenged 3.2
grams/mile, with significant differences in the level of increase
among the vehicles tested.  The vehicles averaged a 0.19 g/mile
NOx increase, with most of the increase accounted for by one
vehicle, the Crown Victoria.

Section 4.   Controlling Emissions from non-FTP Driving  Behavior
   High load events  (hard  accelerations), and high  speed and
transient driving behavior were all.components of non-LA4 driving
behavior which showed the potential for large emissions increases
relative to FTP,  controlled emissions.  The results of the test
program established a need to control these emissions.  This
section considers alternative methods of controls and the
feasible levels of emission control.

4.1  Causes of Emission Increases
   There are several  causes  for exhaust emissions that resulted
from high speed and load,  and transient,  non-FTP driving
behavior,  as found over the ARB02 and REP05 cycles.  Commanded
and transient enrichments had a significant impact on HC and CO
emissions; while high combustion chamber temperatures resulting
from high speed and load operation, and poor catalyst conversion
efficiency levels,   caused high NOx emissions.

4.1.1  Commanded Enrichment
   Commanded enrichment is any extra-  fuel, beyond what is
necessary to maintain a stoichiometric air-fuel ratio, that is
deliberately delivered to the engine via a command from the
engine calibration through the electronic engine control system
to the electronic fuel injection system.   It is analogous to the
acceleration enrichment that was used for carburetored fuel
                               24

-------
systems.  Commanded enrichment is typically used whenever the
engine is under high loads, such-as those that occur during hard
accelerations or pulling a loaded trailer.  The extra fuel
provides the engine with a  power gain and is also used to cool
the engine and catalyst.
   Emission  data  from  the  EPA/CARB/AAMA/AIAM  cooperative  test
program indicates that hydrocarbon  (HC) and carbon monoxide (CO)
emissions are very sensitive to commanded enrichment.  Engine-out
CO emissions increase as the air-fuel ratio is richened from
stoichiometric levels, due to the lack of oxygen available to
complete the combustion process to C02.  Engine-out HC emissions
are simply unburned fuel that result from wall quenching,
deceleration misfire,  rich operation, and by hiding in combustion
chamber and piston wall crevices.  As hydrogen reactions are
favored over carbon reactions and tend to continue to occur even
in a rich environment, engine-out HC emissions show relatively
little sensitivity to air-fuel ratio as compared to CO emissions.
Catalyst conversion efficiency levels for both HC and CO
emissions are very sensitive to air-fuel ratio.  In a rich
environment, the lack of oxygen causes the oxidation of HC and CO
into C02 and water vapor,  to drop off very quickly, causing
catalyst conversion efficiency to be reduced, especially for CO.
Air-fuel ratios during commanded enrichment events can be as rich
as 11.7:1,  compared to the normal stoichiometric A/F level of
approximately 14.7:1.            .

4.1,1.1  Impact of Commanded Enrichment
   Figure 1-11 provides  an  example of  the  impact of a commanded
enrichment event on CO emissions.  For an 8-second segment of an
acceleration event ,  the figure compares CO emission and the A/F
ratio for vehicle 304  (Oldsmobile 98), tested in the production
configuration and in the stoich  (no commanded enrichment
configuration.   The commanded enrichment event lasted
approximately eight seconds and changed the air-fuel ratio from
14.7:1 to 13.0:1, and increased maximum engine-out CO about an
                               25

-------
order in magnitude.
     As  part of the manufacturers test  program,  a subset of
approximately  15 vehicles were tested with stoichiometric
(referred to as  "stoich")   calibrations as well as with  the
original production  calibrations.   The manufacturers eliminated
the commanded  enrichment  strategies for the stoich calibrations,
but made no attempt  to reduce any other enrichments, such  as
starting or transient,  or to optimize spark timing or other
strategies as  a  result of eliminating commanded enrichment.10
In fact, some  of the vehicles still ran slightly rich under high
loads with the stoich calibrations.  Therefore, it should  be  kept
in mind that while the stoich calibrations demonstrate the
reductions in  HC and CO emissions that can be achieved, by
eliminating commanded enrichment,  they have not been optimized
for overall emission control or impacts on driveability  and
performance.   Thus,   without directly proving feasibility, they
do demonstrate -the approximate improvement in high speed and  load
off-cycle HC and CO  emissions that  can be achieved.
   The effects  of commanded enrichment  on  CO  and HC emissions is
further illustrated  in table 1-8.   Table 1-8 shows the HC  and CO
tailpipe emissions   results of eight vehicles over an
acceleration event on the ARB02 cycle with the production  and
stoich calibrations.  The  production calibrations use commanded
enrichment during this acceleration whereas commanded enrichment
has been removed for the  stoich calibrations.  By comparing the
emissions generated  during this  acceleration for those
calibrations with and without commanded-enrichment, the  impact  of
commanded enrichment  can  be clearly demonstrated.  This
acceleration was chosen because it  is one of the most aggressive
accelerations  found  on -the cycle and almost every vehicle  in  the
     10Transient enrichmentis used to compensate for lean spikes that
typically accompany sudden throttle opening or momentary accelerations that
occur during microtransient .operation.  "Starting Enrichment" is used during
cold and hot engine start-up.   It is required to overcome poor atomization of
fuel droplets that occur during extreme ambient conditions.

                                26

-------
test program  went  into commanded enrichment when  operated over
it11.   The  average  duration of the commanded enrichment events
for these  vehicles over this acceleration was  8.3  seconds with an
average change  in  air-fuel ratio of 14.61:1 to 12.45:1.
                             Table 1-8
  Impact of commanded enrichment  on  HC and CO tailpipe emissions
          Initial acceleration of hill 19 of ARB02 cycle
Vehicle

Escort
Supreme
GPrix
Olds 98
Seville
Saturn
Metro
GrandAm
Average
HC (g/sec)
Prod
0.390
0.250
0.275..
0.146
0.643
0.268
0.411
0.364
0.343
Stoich
0.048
0.008
0.020
0.027
0.029
0.020
0.110
0.032
0.037
Increase
0.342
0.242
0.255
0.119
0.614
0.248
0.031
0.332
x 0.306
CO (g/sec)
Prod
30.58
20.78 -
16.38
14.83
28.76
22.55
19.71
22.02
21.95
Stoich
2.67
0.39
0,43
2.60
0.74
1.50
4.12
3.98
2.05
Increas
e
27.91
20.39
15.95
12.23
28.02
21.05
15.59
18.04
19.89
   For HC emissions, the average  tailpipe levels were 0.037 g/sec
with the stoich  calibration (no commanded enrichment)  and 0.343
g/sec with the production  calibration  (commanded  enrichment),
with an average  increase of 0.306 g/sec.  The average  CO tailpipe
emission levels  were  2.05  g/sec with the stoich calibration and
21.95 g/sec with the  production calibration, for  an  average
      The Mercedes 420 SEL does not use commanded enrichment.  Mercedes is
currently the only  manufacturer that produces  some vehicle models that do not
utilize commanded enrichment."       '
                                27

-------
increase of 19.89 g/sec.   The increase in HC and CO tailpipe
emissions for these vehicles due to commanded enrichment, is
about one order of magnitude, 1000%.
                                    /
4.1.1.2  Comparison of Production and Stoich Bag Results
   Figure  1-12  compares  the  production  and  stolen HC  and  CO
emission results for the REP05 cycle.   The stoichiometric
calibration reduces CO emissions dramatically; HC emissions
decrease in almost all cases.  The stoich.  calibration tends to
increase NOx  emissions.  This potential trade-off between CO/ HC
emission reduction and NOx control is an important issue which is
explored more fully in section 5.

4.1.2  Transient Fuel Control
   EPA  and manufacturers  have long  believed that slight changes
in throttle movement can impact HC and CO emissions.   Data from
the EPA/CARB/AAMA/AIAM test programs support this theory for HC
and CO emissions during non-FTP driving operation.   There is not
one clear explanation for the increase in emissions.   The data
does point to one possible cause:  rich spikes in the air-fuel
ratio.    These rich spikes do not appear to be caused by
commanded enrichment since they were observed in results from
both, production and stoich calibrations.   Rather,  they seem to
occur for two different reasons,  either from a series of short,
abrupt throttle openings that happen during rapid throttle
movement or from moderate to heavy deceleration events.
   /
4.1.2.1  Throttle Movement
   The  rich air-fuel  ratio spikes that  occur from short,  abrupt
throttle openings during periods of significant throttle
movement,  appear to be related to enrichment strategies  .  When
the throttle is initially opened, there is a lag between the air
entering the cylinder and the fuel being injected into the
cylinder.   The air enters the cylinders instantaneously whereas
the injection of the fuel cannot occur until the fuel control
                                28

-------
system  (i.e., control module)  senses that the throttle has been
opened;  calculates how much fuel is necessary;  and sends the
proper voltage signals to the injectors to release the fuel.
This results in a momentary period of enleanment or a lean spike.
A very common calibration strategy for minimizing the length and
depth of these lean spikes is to use extra fuel, known as
transient enrichment, as soon as the control system senses that
the throttle is being opened.  Unfortunately, it is common for
the transient enrichment calibration to use too much fuel and
result in an unwanted rich spike.
   A good example  of  transient enrichment  is  shown by  the
Mercedes 420 SEL;  a vehicle which, was included in both the EPA
and vehicle manufacturer testing.  A unique feature of this
Mercedes is that it .does  not use any commanded enrichment
strategies, yet it demonstrated an emissions  sensitivity to
variability in the amount  of throttle movement from one test to
another.   Over__a given cycle,  a driver that drove with more
fluctuation in the throttle  generated higher emissions than a
driver with less throttle variation ,  or a more "steady  foot."
Table 1-9 presents emissions results and throttle movement
measures for three runs over the REM01 cycle, using 3 different
drivers.   The change in throttle (DTP)  was calculated as the
change in the measured throttle position from one second to the
next.  The sum of  the DTP'shows driver C to be the smoothest
driver.  This driver also had the lowest HC and CO emissions.
The results suggest that there are some vehicles which show  HC
and/CO emission sensitivity to the amount of  throttle movement.

4.1.2.2  Throttle  Movement Modelling
   As a follow-up  to  the  analysis of the Mercedes', EPA used
data from  the vehicle manufacturers test program to analyze the
emission impact of throttle movement.  Only  stoich tests were
used in order to eliminate the impact of commanded enrichment,
and the test data  were limited to  the two non-FTP cycles:  ARB02
and REP05 (note:   the manufacturers test program did not include
                               29

-------
testing on REM01).   A simple tailpipe emissions regression model
was developed to look at the determinants of the emission
differences.    After controlling for  vehicle and cycle, the
model looked at impact of  throttle variation, measured as the
sum of the one-second change in throttle (DTP).   As shown in
table 1-10, DTP is statistically significant  (factor,) at the 10%
level  for both HC and CO  emissions.
   The movement  of  the  throttle  is  the vehicle's  response to  the
driver's demands to achieve or maintain a particular speed.  In
following a driving schedule, the changes in throttle are
associated with changes in the driving behavior; a speed-based
measure that relates well with the throttle change is the sum of
the change in power (dpwrsum).   Table 1-10 suggests that dpwrsum
shows a similar 10% significance for explaining marginal
variation in HC and CO emissions.

                           Table 1-10
           MargiTTal Effects of Change in  Throttle  (DTP)   ' ,
             and Change  in Specific  Power  (DPWRSUM)
                 Stoich Tests, ARE  and REP  Cycles
                 with no  load adjustment (n=58)

DTP



DPWRSU
M



Pollutant

CO
HC
NOX

CO
HC
NOX
' Full
model R2

0.875
0.883
0.980

0.873
0.882
0.980
Coefficient
(xlOOO)

13.044
0.315
-0.064

3.016
0.076
0.014
t
statistic

1.90
2.68
-0.18

.1.74
2.55
0.16
prob . >t

0.064
0.011
0.855

0.089
0.015
0.876
                                30

-------
4.1.2.3  Heavy Deceleration Enrichment
   Increases  in  HC and CO emissions  during heavy deceleration
events is due to the potentially large slugs of raw fuel that are
drawn into the combustion chamber during quick closures of the
throttle that usually occur during sudden deceleration events.
As fuel flows through the intake system into the combustion
chamber, a fuel boundary layer is formed,  where' liquid fuel is
"stored" along the surfaces of the intake manifold, port area,
combustion chamber, and cylinder walls.  The thickness of the
fuel boundary layer is inversely proportional to the manifold
vacuum.  When a vehicle suddenly decelerates, the manifold vacuum
decreases dramatically in response to closure of the throttle
blade.  This results in the simultaneous drop of air to very low
levels, due to the throttle closing, and a surge of fuel being
drawn off the intake and combustion surfaces, due to the increase
in manifold vacuum.  For most vehicles equipped with port fuel
injection (PFI)  systems,  the amount of extra fuel drawn is
usually small.   However, the extra fuel that is drawn into the
combustion chamber stays in a liquid form and doesn't properly
mix with the rest of the fuel-air mixture and is passed through
the engine in a raw uncombusted state, raising engine-out HC
emissions.
   This  phenomenon was  the most apparent for  vehicles equipped
with throttle-body fuel injection (TBI) during heavy
decelerations.  This is most likely due to the fact that more
fue^i is "stored" along the intake manifold walls for TBI vehicles
than for PFI vehicles since the fuel injector is located in the
throttle-body where air and fuel are combined,  and then must flow
along the length of the intake runners to the combustion
chambers.  For PFI vehicles, the fuel injector is located in the
intake-manifold runner as close to the valves in the combustion
chamber as possible.  When the fuel is injected into the
combustion chamber, there is considerably less intake runner
surface for fuel to adhere to,  thus, -less liquid fuel is drawn
                               31

-------
into the combustion chamber.  While this phenomenon was the most
prevalent with TBI vehicles, there were some PFI vehicles that
also experienced increases in HC emissions due to heavy
decelerations.  This is probably.a result of poor calibration
strategy, due to either a lack of the proper strategy needed to
anticipate when a heavy deceleration will occur, or poor
calibration technique of the existing deceleration strategies.
      Most  of  the  heavy decelerations  that  caused  HC  increases
over the various cycles were aggressive, i.e., instantaneous
throttle closing combined with braking for several seconds, that
typically occurred at the end of a hill.  However, the test data
revealed that even relatively short, but abrupt throttle closings
that occurred during the middle of a hill,  not in an attempt to
stop the vehicle but rather as a result of excessive or even
erratic throttle behavior on the part of the test driver to
maintain the driving trace, can cause similar HC increases.
   The  increases—in  HC  and  CO  emissions  due to enrichment
occurring during heavy deceleration events were small.  This is
most likely due to the fact that even though air-fuel ratios were
very rich  (approximately 13:1) during heavy deceleration events,
the exhaust mass flow is very low, thus, only generating
relatively low emission levels.

4.1.3  High Combustion Temperatures
    The  NOx emissions  resulting  from  the ARB02  and REP05  cycles
were significantly higher than warm, stabilized FTP levels.  This
is due to the fact that NOx emissions are temperature sensitive.
 Excessively high temperatures in the combustion chamber cause
free oxygen and nitrogen from the air-fuel.mixture to combine and
create NOx.-  During high speed and load operation, more fuel and
air are burned in the combustion chamber.   As the amount of
combustion increases, temperatures also increase.   The
temperatures in the combustion chamber are considerably higher
during high speed and load operation than during typical FTP
operation,  thus causing the  significant increase in engine-out
                               32

-------
NOx emissions.
   Figure 1-13 illustrates the increase in NOx emissions from bag
.2 of  the  FTP  to  the  REP05  cycle.   Bag  2  emission  results were
chosen, rather than  overall  FTP emission  results,  because bag 2
results are warm,  stabilized results without  any start-up
emissions, similar to REP05  cycle  results.
   As discussed in Section 4.1.1.,   commanded enrichment is used
to reduce combustion temperatures.  During stoichiometric
operation, combustion is much more complete than during  rich or
lean  operation,  where there  is an  excess  of fuel or air.   The
more  complete the  combustion process is,  the  more  heat that is
given off as  a byproduct.  Commanded enrichment reduces
combustion temperatures from those observed at stoichiometry
-because the extra  fuel in  the air/fuel  mixture, above
stoichiometry, is  not combusted and dampens or absorbs the heat
from  the  combustion  process,  keeping combustion temperatures
lower than what—would occur  during stoichiometric  operation.
Thus  the  removal or  reduction of commanded enrichment causes an
increase  in combustion temperatures and consequentially, engine-
out NOx emissions.   The effects of removing commanded enrichment
over  the  ARB02 and REP05 cycles is illustrated in  Table  1-11.
                                33

-------
                            Table 1-11
       Impact on engine-out NOx emissions from removal of
        commanded enrichment over ARB02 and REP05 cycles
Vehicle

Escort
Cruiser
Seville
Supreme
Olds 98
Saturn
Average
ARB02 Cycle
Prod
4.83
2.03
3.11
3.89
4.02
2.16
3.34
Stoich
5.61
2.68
3.68
4.37
4.28
2.41
3.84
Increase
0.78
0.65
0.57
0.48
0.26
0.25
0.50
REP05 Cycle
Prod
4.11
1.85
2.56
3.20
3.67
1.92
2.89
Stoich
4.74
2.22
2.69
3.54
3.81
2.18
3.19
Increase
0.63
0.37
0.13
0.34
0.14
0..26
0.30
   Table  1-11  shows  that  for  both,  the  ARB02  and REP05.cycles,
that as commanded enrichment1 was removed, engine-out NOx levels
Increased.  The average increase over the ARB02 cycle was 0.50
g/mi, while the average-increase for the REP05 cycle was 0.30
g/mi.  The difference in the average engine-out increase between
the two cycles is most likely due to the fact, that the ARB02
cycle contains more aggressive acceleration events than the REP05
cycle.                -                   '

4.1.4  NOx Catalyst Conversion Efficiency
       In addition to increased combustion temperatures, data
frorf the test programs indicated  that high NOx emission levels
also  appear to be related to not having tight enough air-fuel
ratio control (i.e., numerous rich and lean spikes, and a lean
bias for the fuel control system during this type of operation).
Analyses done by EPA showed that vehicles from the manufacturer
test.program that had high NOx emissions during high speed and
load operation,  also had erratic NOx catalyst conversion
                                34

-------
efficiency levels12.   Those  vehicles  that  had low NOx emissions
during high speed and load  operation,  had almost  continuous  high
catalyst conversion efficiency  levels.  Further  examination
revealed that the vehicles  that had  erratic  conversion efficiency
levels also had erratic air-fuel ratio levels,, with  numerous rich
and lean spikes, while the  vehicles  with  high  conversion
efficiency had.very tight air-fuel ratio  control.    In addition,
the vehicles with high catalyst conversion efficiency levels
appeared to use a rich bias; their air-fuel  ratios averaged
around 14.4:1 to 14.5:1, whereas the vehicles  with the erratic
conversion efficiency seemed to have more of a lean  biased or
unbiased fuel control strategy with  the average  air-fuel  ratio
                                                 ;
centering around 14.6:1 to  14.7:1.
   Vehicles  201  (Escort)  and 306 (Custom Cruiser) both used a
control strategy known as lean-on-cruise  (LOG).   The  purpose of
this strategy is to enhance fuel economy  by  operating at  an  air-
fuel ratio of approximately 17:1 during steady-state  cruises.
For both of these vehicles,' LOG only  occurred during relatively
high speed cruises.  Table  1-llb shows the time  spent in  LOG and
the grams/second emissions  resulting from LOG  over the FTP and
REP05 cycles for both vehicles with  production calibrations.   The
Custom Cruiser spent no time in LOG  during the FTP,  but averaged
351 seconds of LOG operation over the  REP05  cycle, and the Escort.
averaged 34.5 seconds.on the FTP and 358.5 seconds during the
REP05 cycle.  Over the REP05 cycle,  the contribution  of LOG
operation to total NOx emissions was 52.5% for the Custom
Cruiser and 71.5% for the Escort.
     12Memorandum from Ted Trimble to John German, titled "Nox emissions on
REP05",  dated April 8,  1994
                                35

-------
                           Table 1-llb
         Lean-On-Cruise Operation for Vehicles 201 & 306
            with Production Calibrations   (grams/sec.)
Veh

201

306

Test

FTP
REP05
FTP
REP05
Lean-On -Cruise
HC
.004
.130
0
.585
CO
.59
.58
0
6.10
NOx
.92
24.74
0
7.28
Total
HC
2.73
2.70
n/a
3 .81
CO
32.32
156.3
n/a
203.5
NOx
' 5.83
34.72
n/a
13.82
% of Total .
HC
4.0
5.0
0
15.0
CO
1.5
0.4
0
3 .0
NOx
15.5
71.5
0
52.5
Time
(sec
)
34.5
358.
5
0
351.
0
4.2  Approaches to Compliance Testing for Non-FTP Driving
     Behavior
   In  examining~alternative  control  methods,  EPA's basic  premise
is that the better the representation of in-use driving the
better the control of in-ude emissions.  While this premise has
to be balanced by practical considerations such as cost and
feasibility,  the importance of trying to accurately reflect
actual in-use conditions cannot be lost.

4.2.1.  Air-Fuel control cycle
   The HL07 cycle  is  an  engineered cycle designed to drive  a
vehicle through a series of high acceleration/load events
covering a range of severity and speeds.  The cycle also includes
a two and half minute cruise at 65 mph.  Prior to starting the
test program manufacturers'  suggested that a cycle like the HL07
could be used to  develop a air-fuel based emission standard.  In
general, this approach would eliminate commanded enrichment by
      •\
establishing an air-fuel band around stoichiometry.   This option
would place standards on the duration or magnitude of deviations
from stoichiometry.,  measured over the short HL07 cycle.   Such an
                                36

-------
approach would likely eliminate much of in-use,. commanded
enrichment, thus greatly reducing CO emissions, and also
achieving HC reductions.  The ability to control offcycle NOx and
HC emissions caused by poor transient fuel control  is less
certain.   A very tight band would be necessary to ensure NOx
control,, while this may not be entirely appropriate or feasible.
However, drawbacks to this approach include the following: the
lack of control for microtransient enrichment; lack of a suitable
methodology for achieving NOx control; difficulties in devising
an A/F standard for vehicles operating on alternative fuels, like
diesel or compressed natural gas (CNG); and reduced manufacturer
flexibility in designing a control strategy.  This option
effectively mandates a control system strategy, while an emission
performance standard provides manufacturers the flexibility to
determine,  case-by-case, the most cost effective way to achieve
the desired emissions result.

4.2.2.  Representative cycle
   A  second method  for controlling non-FTP  emissions,  a
representative cycle, stands in sharp  contrast to the HL07
cycle.  As a control cycle representing the full range of non-FTP
operation,  it can be argued that a representative cycle is the
best method for ensuring that the emission control achieved in
testing will fully translate to in-use emission control.
   The  principal difficulty .in  implementing.such  an approach is
that a such cycle must try to represent speed, acceleration, the
interaction of speed and acceleration, as well as the change in
accelerations.  All these variables lead to the need for a very
long cycle.  EPA developed REP05 expressly to represent the range
of non-FTP operation, and as a consequence the cycle is 1400
seconds (over 23 minutes)  in duration.  A large fraction of the
representative cycle will be high speed cruise operation, since
this is the predominant mode of non-FTP operation.  The extra
testing and facility time to include all the cruise operation is
hard to justify of a cost/benefit.   It is reasonable to assume
                                37

-------
that control of emissions during high speed cruise operation can
be achieved without having to match its exact in-use
representation.

4.2.3  High speed/load transient control cycle (US06)
  A third  control  approach  involves  a  hybrid  cycle  that   shares
characteristics of both the air-fuel control approach and the
representative cycle approach.  The new cycle, US06, is 600
seconds in duration and is comprised of segments of CARS's ARB02
cycle and EPA's REP05 cycle.  Similar to the air-fuel control
method, this method targets specific high emission,  non-FTP
operation.  And like the representative cycle,  the US06 is based
on actual segments of in-use driving.
  Through  a concerted, coordinated effort,  staff  from the two
agencies (with helpful manufacturers' comments)  developed the
US06 based largely on a  review of  the second-by second
emissions over ±Jae REP05 and ARB02 cycles, from  the vehicle
manufacturer's test program.  From the two driving cycles, staff
identified segments which they felt would provide control
emissions from aggressive driving and transient operation.  The
US06 includes the range of non-FTP driving operation including
all of the more severe acceleration events and includes
representative high speed cruise operation,  while reducing the
cycle time to 10 minutes.    The US06 cycle is shown in figure 1-
14.         •

4.2*4  Justification for selecting US06 as preferred option
  US06  is  EPA's preferred method   for  establishing  control of
emissions from non-LA4 driving behavior.  The US06 covers  the
range of non-LA4 driving, while targeting severe,  high emission
events.   Because the driving.modes generating the highest
emissions differed widely across vehicles, it is very important
to include a'variety of high load events representing actual
aggressive driving behavior.  In addition,.the US06 cycle
achieves the objectives of both EPA and GARB,  thus eliminating
                                38

-------
issues or costs associated with the respective agencies having
two different control  An important CARD objective is to make
sure outer bounds of in-use aggressive driving is represented and
controlled; this is achieved with the inclusion of the ARB02
high-speed micro-trip. A second, ARB02 high-speed microtrip was
rejected due to an extended, high-speed acceleration which might
result in.excessive catalyst temperatures in vehicles which are
controlling commanded enrichment.  Thus, the US06 provides for
control of short -durat ion commanded enrichment events associated
with aggressive driving.  As discussed in the feasibility section
which follows,  the duration of commanded enrichment control needs
to be limited due to catalyst temperature concerns.  EPA's
analysis of catalyst temperature data from the manufacturer's
test program concluded that the ARB02 high speed microtrip used
in US06 provides for a reasonable duration of control.

4.2.4.1  Road Grade
   The  severe driving  events .contained within US06  also  help
provide for some control over  the emission impact of road grade.
As discussed in section 2.-2, the extra load placed on the engine
by road grade is analogous to the load from an acceleration, and
thus, if sufficient,  can result in a sharp emission increases as
a result of commanded enrichment. The in-use frequency and
duration of commanded enrichment events are a function of the
combined effects of driving behavior, road grade, and vehicle
loading.  To the extent that US06 contains driving which is more
aggressive than that called for by a representative cycle,  this
"safety margin" provides for some control of commanded enrichment
resulting from road grade or vehicle loading.  For example,
commanded enrichment events associated with aggressive driving on
a road grade (such as an entrance ramp)  would be controlled--up
to the^ control  duration established by US06.

4.2.5  US06 adjustments
   As a  control  cycle,  the US06  is  appropriate for  a  large
                                39

-------
fraction of light-duty vehicles and light-duty trucks.  However,
inasmuch as the cycle is intended to control emissions during
severe, high speed and load operation,   the severe operation
characterized by the cycle may  exceed some vehicles'
capabilities.  Section 2.2 discussed differences in in-use
driving behavior as a function of vehicle performance,
transmission type,  and vehicle type.  These differences need to
be considered in judging the appropriateness of the control cycle
for. all vehicles.

4.2.5.1  Vehicle Performance and Transmission Type
   In-use driving patterns  data  indicate  that  for manual
transmission vehicles, high and low performance vehicles
(performance based on W/P)  are driven differently than the broad
range of mid-performance vehicles (see section 2.2.1.).  High
performance vehicles were driven in a more aggressive manner,
while the low performance vehicles were typically driven  less
aggressively.   For automatic transmission vehicles, it appears
that low performance vehicles are driven less aggressively than
middle and high performance vehicles.  The US06 cycle is a hybrid
of the REP05 and ARB02 cycles, and it is intended to be
represent the vehicle fleet as a whole.  The portions of the
REP05 cycle used in the US06 cycle are identified below in table
1-12, along with a description of the vehicle which actually
generated the driving segment.  The W/P for the vehicles'  which
comprise the REP05 cycle cover the full .performance spectrum; one
segment, R5,  came from a high performance vehicle.   It is not
possible to identify the vehicle's which generated the driving
                                k
for the Los Angeles chase car data.
                                40

-------
                            Table  1-12
Segment
Rl
R2
R3
R4
R5
R6
R7
In
US06
yes
no
yes
no . •
yes
no
yes
Veh.
#
B163
B467
B287
B419
S224
B368
B389
Description
1988 Honda Accord
Lxi
1986 Honda Accord
1979 Chevrolet Monte
Carlo
1980 Buick Regal
1988 Ford
Thunderbird
1988 Hyundai Excel
1982 Toyota Corolla
Transmissi
on
manual
automatic
automatic
automatic
manual
automatic
automatic
W/P
25.00
28.75
27.78
25.83
19.33
36.76
37.50
The in-use driving survey results suggest that it would be
unrepresentative~~and inappropriate to require low performance
vehicles to drive portions of the US06, such as Rl and R5,
without adjustments.

4.2.5.2  HLDTs
   Evidence  from  the  in-use  driving  survey data points  to  the
need for adjustments over the US06 cycle for HLDTs. The case for
HLDTs is analogous to that for high and low performance vehicles,
The US06 and predecessors were developed to represent non-FTP
driving behavior for the fleet as a whole.    As discussed in
    »
section 2.2.2,  HLDT driving behavior appears to be different
than the rest of  the light- duty  vehicles' behavior  .
Determining the appropriate testing for HLDTs is complicated by
the current method of testing HLDTs at adjusted load vehicle
weight  (1/2 payload).   The driving segments found in US06 are
from LDVs, and the load they were subjected to while exhibiting
such behavior in not known.   It can be assumed,  however, that
these vehicles were not loaded to the extent that  HLDTs are
                               41

-------
loaded when tested at 1/2 payload.

4.2.5.3  Adjustment approach
 1  The  tailoring of  the  cycle   to  meet  the  needs  of  each
individual vehicle is impractical.    As is the case for the  FTP
driving schedule, there will always be some vehicles for which
the  US06 will be "easier" than it is for  others.  However, it
is desirable to have the US06  broadly representative for all
vehicles.  To achieve this objective, EPA feels it is appropriate
to make adjustments for low performance vehicles and HLDTs  (to
reduce the cycle's severity where it appears overly severe).
Again,  the adjustments are to make the cycle more representative
of actual vehicle operation and to ensure that the emission
control demonstrated on the test procedure results in in-use
emission reductions.
   One  approach  would  involve making  adjustments  to  the  test
procedure, by modifying  the. actual US06 driving cycle.   This
would lead to a proliferation of cycles and would greatly
increase the cost and complexity of  testing.      An alternative
to modifying the speed-time trace is to make adjustments to the
dynamometer inertia settings.   An adjustments to.the inertia load
can have the same effect as a modification to the cycle, as the
engine can't tell the difference between an inertia load and an
acceleration load.  This approach was tested during the emission
test programs with general success (see section 3.2.2.1)  EPA
believes adjustments to dynamometer load settings is a reasonable
and (practical method for handling the need for modifications to
the cycle.  Section 7 proposes refinements- to the dynamometer
load adjustment approach used in the emission test program.   The
proposed adjustments to low  performance vehicles are intended to
bring these vehicles toward the mid-performance vehicles.   The
HLDT adjustments serve to reduce the severity of the cycle to be
more consistent with their in-use operation.

4.3     Potential  Strategies for Controlling Emissions from Non-
                               42

-------
        FTP  Driving Behavior
   The  Agency has  determined  that  significant  reductions  in
emissions, resulting from aggressive driving and microtransient
operation, can be achieved by improved fuel control through
optimization of engine calibrations and some slight hardware
modifications.  Significant reductions in HC and CO emissions can
be gained by reducing or eliminating commanded enrichment and
optimizing transient fuel control strategies to'better maintain
stoichiometric air-fuel ratio operation.  In order to realize
these reductions,   some vehicles may have to also switch from
synchronous-fire port fuel injection systems to sequential-fire
port fuel injection systems.   Large reductions in NOx emissions
can be achieved by improving NOx catalyst conversion efficiency
levels through tight closed-loop fuel system control of an air-
fuel ratio with a slightly rich bias during high speed and load
operation.
   Emissions  could be  further reduced by using "drive-by-wire"
technology and going to larger catalysts with higher noble metal
loadings.  However, these would be very costly and the level of
emission reductions would be  small compared to what can be
achieved through the calibration optimization discussed above.
Therefore,  EPA feels that, the emission benefits that could be
gained by using "drive-by-wire" systems and larger catalysts are
not sufficient to require vehicle manufacturers to incur the
costs of using such control strategies.

4.3-1  Improved Fuel Control  Through Calibration

4.3.1.1  Commanded Enrichment
   As previously discussed'in Section 4.2.3, the vehicle
operation simulated by the US06 cycle is considerably more
aggre&eive than that found on the FTP.   For example, the Geo
Metro equipped with a 1.0 liter three cylinder engine, had an
average throttle opening of only 7.5% over the FTP with a maximum
throttle opening of 42.4%.  On the US06 cycle, the Metro had an
                                43

-------
average throttle opening of 22.0% with a maximum opening of 100%.
 The Metro was the lowest performing vehicle in the test program
with a weight-to-power ratio of  38.6 (2125 ETW/55hp),  and yet it
never needed to exceed a throttle opening of 50% over the FTP.
                                t
   Because  of  the  relatively  low power requirements  on  the  FTP,
especially when considering accelerations, only a few vehicles in
the test program ever went into commanded enrichment, over the
FTP.  This greatly contrasts with the observations for the US06
cycle.  All of the vehicles in the test program, except the
Mercedes,  went into commanded enrichment during the US06 cycle.
The actual level of enrichment, or the amount of additional fuel
added, during commanded enrichment events varies from vehicle to
vehicle.  The amount of commanded enrichment necessary, and the
strategy for when it is utilized,  is dependent on vehicle design
constraints and calibration refinement.     For some high
performance vehicles the additional power generated by commanded
enrichment may no.t be as important as the cooling effect the
extra fuel has on the engine and catalyst.  Typically,  these
vehicles have more concern over engine and catalyst durability
because of the higher exhaust throughput generated by the large
engines, and their frequent use of small close-coupled warm-up
catalysts.      Smaller low performance vehicles may have a
greater need for the extra power that is generated by commanded
enrichment.  These vehicles often have small displacement,  in-
line 3 and 4 cylinder,  high revving engines that typically run
hotter than larger 6, 8, and 10 cylinder V-type engines.
Because of this, a common strategy among manufacturers is to keep
engine temperatures down by using commanded enrichment during
high load operation.
   In  addition to  optimizing  commanded enrichment  for specific
engine/vehicle configuration constraints, there is the impact of
calibration sophistication.  Unfortunately, not all development
engineers who calibrate the engine and emission control systems,
have the same level of skill or experience.   It has been
suggested by various manufacturers that the occasional
                               44

-------
discrepancy in test data or occurrence  of an unexplained
phenomenon may be the result of  poor calibration technique.
    Fuel  control  calibrations  are  currently  not  calibrated  to
control exhaust emissions that occur during heavy load operation.
There are several reasons for this: 1) Commanded enrichment
calibrations have been intended to meet specific performance and
durability criteria, such as enhancing power and/or to cool the
catalyst; 2) Lack of sufficient high load operation over the FTP
necessary to engage commanded enrichment: 3) No emission
standards for off-cycle emissions; and 4) The occasional lack of
calibration sophistication.  As previously demonstrated, the
commanded enrichment events that occur during  the US06 cycle
have significant effects on HC and CO emissions.  The vast  '
majority of vehicles sold in the United States utilize some level
of commanded enrichment.   It is therefore apparent that the
relationship between commanded enrichment and exhaust emissions
from heavy load__operation, have not been an area of focus for
manufacturers.  The results from the various vehicles tested with
stoich calibrations over the US06 cycle, demonstrate the large
reduction in HC and CO emissions that can be obtained by reducing
or eliminating commanded enrichment.  It is apparent then,  that
one of the most important control strategies that will have to be
implemented by manufacturers in order to comply with proposed HC
and CO emission levels, will be the reduction or, in some cases,
the elimination of commanded enrichment.  The potential effects
of reducing or eliminating commanded enrichment on vehicle
performance and engine and catalyst durability is discussed in
section 5.

4.3.1.2  Transient Enrichment
  Another potential  strategy  for  reducing HC  and CO  emissions
during- non-FTP driving behavior is to maintain tight
stoichiometric air-fuel ratio control during transient operation,
or more specifically, during rapid throttle movement where there
is a series of short, abrupt throttle openings.    This could be
                                45

-------
achieved by optimizing transient enrichment calibrations.   As
discussed in Section 4.1.2.1, transient enrichment is used to
compensate for brief periods of enleanment that occur immediately
following throttle opening due to time lags in the fuel control
system.  By optimizing transient enrichment calibrations such
that the amount of enrichment is.minimized and  throttle opening
is better anticipated, HC and CO emissions resulting from
transient operation should be greatly reduced.
   Unlike  the  reduction or  elimination of  commanded enrichment,
this strategy will not be necessary for all vehicles.  Several
vehicles from the test program, such as 303 (Pontiac Grand Prix)
and 305 (Cadillac Seville)  had very tight air-fuel ratio control
and low HC and CO emissions over all of the high speed and load
cycles.  The tight air-fuel ratio control and low emissions
indicate that transient enrichment is not a problem for these
vehicles.   Therefore, EPA believes optimization of transient
enrichment calibrations can be readily applied  to all vehicles
that have high non-FTP HC and CO emissions resulting from poor
transient fuel control.

4.3.1.3  Heavy Deceleration Enrichment
   As previously mentioned,  the  emission  impact  of  heavy
deceleration enrichment on HC and CO emissions are minimal.
However, there still is some merit in discussing the elimination
of rich air-fiiel ratio  spikes  that occur  during heavy deceleration
events as a potential control strategy for the reduction of HC
and.'CO emissions.   As discussed in-Section 4.1.2.3, the
enrichment that causes rich air-fuel ratio spikes during
deceleration events is not the result of programmed enrichments
like commanded or transient.  Rather, it is the result of poor
transient fuel control strategies, and the natural phenomenon of
fuel being stored on the surfaces of the intake manifold wall,
valves, etc.,  and then being drawn into the engine during sudden
throttle closure.   However, the later cause is the most
                                46

-------
prominent.  Therefore, calibration enhancements will only be part
of the potential control strategy.   The most viable calibration
technique that can be used to control heavy deceleration
enrichment is a control strategy known as decel fuel shut-off.
During moderate to heavy decelerations, fuel is shut off, thus
greatly reducing HC and CO emissions.  This strategy has been
used on numerous vehicle models by various manufacturers for
several years.  It has been typically used to eliminate the heavy
deceleration event rich spikes for increased fuel economy.
Vehicles 304  (Oldsmobile 98)  and 314 (Pontiac Grand-Am) were both
equipped with.decel fuel shut-off and never experienced any rich
spikes due to heavy decelerations during any of the high speed
and load cycles.

4.3.1.4  NOx Catalyst Conversion Efficiency
   There  are  several potential  strategies  for  controlling NOx
emissions from non-FTP driving behavior.  An analysis by EPA
examining NOx catalyst conversion efficiency levels, engine-out
NOx emissions,  and tailpipe NOx emissions, appear to indicate
out that control of engine-out NOx levels were not as significant
in reducing tailpipe emissions as maximizing NOx conversion
efficiency levels over the various high speed and load cycles.
The two vehicles with the highest tailpipe emissions, 306 (Custom
Cruiser)  and 305 (Seville),  also had some of the lowest engine-
out levels.  Consequentially, these two vehicles also had the
some of the lowest conversion efficiency levels at 62.8% and
74.8%, respectively,  over the REP05 cycle.
   Therefore,  one of the main control strategies  available for
controlling non-FTP NOx emissions is to raise NOx catalyst
conversion efficiency levels during high speed and load
operation.  The  above mentioned analysis suggests that this can
be accomplished by incorporating very tight fuel control around a
slightly rich biased air-fuel ratio (14.5:1-14.6:1).  At least
four vehicles from the manufacturers test program, 314  (Grand-
Am) ,  401  (Civic),  601 (Mirage), and 801 (Camry) experienced
                               47

-------
catalyst efficiency levels over 95% and had very low tailpipe NOx
emission levels over the REP05 cycle.  All four vehicles used a
slight rich bias and had very tight fuel control.  However, it is
possible that the use of a rich air-fuel ratio bias could cause
CO emissions to increase.  This issue has not been fully
evaluated by EPA.
  Another control strategy, the elimination of the lean-on-cruise
fuel strategy, will have less impact on NOx emissions for the
vast majority of vehicles since only a small percentage of
vehicles utilize it.  Lean-on-cruise is typically used during
high speed cruise operation to improve fuel economy.   During the
65 mph cruise at the end of the HL07 cycle, the air-fuel ratio
was maintained at approximately 17:1.  Corresponding to this
enleanment,  the rate of NOx emission generation rises
dramatically.  Therefore, the elimination of this strategy will
cause a significant reduction in NOx emissions.
                i                 <•
  An  additional...control  strategy would  be  to  concentrate  on
lowering engine-out NOx emissions through the use of exhaust gas
recirculation (EGR) and reducing spark advance.  Both of these
strategies have been used extensively for years throughout the
auto industry as a means to reduce engine-out NOx emissions by
                                  i
lowering combustion temperatures.   These strategies would still
be very beneficial in lowering NOx emission levels and EPA
expects that they will used by most manufacturers.  However,
based on the above discussions, it appears that solely
concentrating on lowering engine-out NOx levels through the use
of $GR or reducing spark advance,  may not be as effective of a .
control strategy as improving catalyst conversion efficiency
levels.

4.3.2   Improved Fuel Control  Through.Sequential-Fire Port Fuel
        Injection
    The ability  to  fire each fuel injector  individually  rather
than simultaneously firing several injectors>  as in synchronous-
fire systems,  allows even greater control of the air-fuel
                               48

-------
mixture  for each  cylinder.  Each cylinder  is assured  of getting
the complete  fuel  injector pulsewidth-worth of  fuel.   For
synchronous-fire  systems, if  two injectors are  fired
simultaneously, only one of the injectors  is being  fired into  a
cylinder where the  intake valve(s) is open and  ready.  The other
cylinder is at a  different point in the firing  cycle  and may not
have the intake valve(s) fully open yet, thus fuel  is  injected
onto the valve instead of into the cylinder.  While some of the
fuel may get  into  the cylinder, it's not the  amount  of fuel that
was intended  to go  in, nor is the same as what  went into the
other cylinder.    This can also cause an excess of  fuel to
"puddle" on the valve so that when the valve is fully  open, this
excess fuel is drawn in with  the injected  fuel, causing the
cylinder to receive too much  fuel.
   The  type of fuel injection system used by a  vehicle is  moot
when the potential  control strategy is the reduction  or
elimination of .commanded enrichment.  However,  for  tighter air-
fuel ratio control, which is  one of the main potential control
strategies available for controlling all three  pollutants, the
difference between  fuel injection systems can be important.    The
type of  fuel  injection system used by a vehicle can help reduce
the vehicle's sensitivity to  throttle movement  and  improve the
NOx catalyst  conversion efficiency.  Systems that utilize
throttle-body  (TBI) fuel injection systems will typically have
'more difficulty maintaining tighter air-fuel ratio'control
because  of poorer  distribution of the air-fuel  mixture and the
greater  lag in the  air-fuel mixture arriving to the combustion
chamber, than those vehicles  equipped with PFI  systems.  However,
as demonstrated by  the Mercedes, even a PFI system  can still have
poor air-fuel ratio control during-rapid throttle movement.  As
discussed above,  sequential-fire PFI systems can offer fuel
control  advantages  that may be necessary for the level of air-
fuel control  required to reduce HC and CO emissions and increase
NOx catalyst  conversion efficiency levels for those vehicles that
are unable to accomplish this through calibration strategies and
                                49

-------
synchronous-fire PFI systems.
   The  use  of  PFI  fuel  control  systems  (synchronous  or
sequential) should also eliminate the majority of HC emissions
that result from unburned HC's ingested by the engine during
moderate-to-heavy deceleration events.
   Manufacturers have indicated that  the  current  direction  for
fuel injection systems throughout the industry is to have all
vehicles equipped with sequential-fire PFI systems.   EPA believes
that prior to implementation of this regulation  (1998 model year)
TBI fuel systems will be eliminated and there will be few
synchronous-fire PFI systems.
   EPA  feels that  the main  strategies that will be required for
controlling emissions from non-FTP driving will be calibration-
related.  The incidence of vehicle models needing to go to
sequential-fire PFI systems in order to control non-FTP emissions
should be low.  This is a moot point since the vast  majority of
the vehicle flest will already be equipped with sequential-fire
PFI systems by the time this regulation takes effect.

4.3.3  Drive-by-Wire Systems (Electronic Throttle Control)
   The  term "drive-by-wire"  refers  to an  electronic  throttle
control system.  For most current vehicles,  the accelerator pedal
is connected to the throttle blade by a metal linkage.or cable.
The  engine control module measures throttle movement by means of
a variable resistor, known as the throttle position sensor (TPS),
located on the shaft of the throttle blade.   Unfortunately, there
is gome time lag between when the driver moves the accelerator
pedal and the control module receives the information and is able
to process it into necessary fuel or spark levels.  Drive-by-wire
systems eliminate the physical connection between the accelerator
pedal and the throttle blade.  Instead, the accelerator pedal is
electronically connected directly to the control module.  The
throttle blade is operated by ah electronic servo motor and the
accelerator pedal has a variable resistor connected to it,
similar to the TPS.  When the accelerator pedal is moved, the
                                50

-------
variable resistor on the pedal sends a voltage signal, indicating
the driver's desired throttle opening, to the control module.  In
turn, the control module processes the signal and  sends a
voltage signal to the servo motor on the throttle blade and opens
the throttle accordingly.  This greatly improves the ability to
anticipate when the throttle will be opened or closed and allow
for optimization of fuel to address rich and lean spikes that
typically occur during throttle movement.
   This  technology  was  originally developed  for  vehicles
utilizing all-wheel drive, as a means for controlling wheel
slippage.  This technology currently exists on a few relatively
expensive luxury vehicles.  The cost of utilizing such a system
would be considerable due to the additional and complex hardware
that is required.  EPA believes that the vast majority of
emission reductions that can be gained from better air-fuel
control can be achieved by optimization of engine calibrations
along with sequential-fire PFI'systems.  The additional level of
emission reduction that could be achieved by using drive-by wire
systems is small.  The cost of  using drive-by-wire systems
outweigh the emission benefits to.be gained by utilizing drive-
by-wire. EPA does not believe that drive-by-wire technology will
be necessary to comply with proposed HC emission levels.
Therefore, to achieve the level of reduction in emissions desired
by the Agency, EPA does not believe that drive-by-wire systems
will be necessary.

4.3*4  Improved Catalysts
   Analysis  of the  test  data  indicates  that  the  use  of  larger
catalysts with higher noble metal loading would not further
reduce HC emissions from the levels achievable from the reduction
or elimination of commanded enrichment; the optimization of
transient enrichment; and use of sequential-fire PFI systems.
However, further reductions in CO and NOx emissions could be
realized for some vehicles by using larger catalysts and higher
noble metal loading due to catalyst breakthrough that can occur
                                51

-------
at high speeds.  Catalyst breakthrough is the inability of the
catalyst to oxidize HC and CO, or reduce NOx in the face of high
exhaust mass flow, despite the likely presence of  sufficient
oxygen to sustain catalysis.  No catalyst breakthrough was
observed for HC control.
   The  increase in CO  and NOx  emissions  from  high  speed  catalyst
breakthrough is very small compared to the increases resulting
from commanded enrichment and poor NOx catalyst conversion
efficiency levels.  The only vehicles that would likely
experience breakthrough are those vehicles with large
displacement engines that would have a very high exhaust mass
flow rate during high speed operation.  Increasing catalyst size
and noble metal loading is expensive.  The additional reductions
in CO and NOx emissions beyond what can be achieved by
eliminating commanded enrichment; optimizing transient
enrichment; increasing NOx catalyst conversion efficiency levels;
and using sequential-fire PFI systems, are very small and would
not be cost effective due to the high cost of larger sized
catalysts and high noble metal loading.

4.3.5  Reasonable Conclusion
   As discussed in Section 4^1,  the  vast majority  of HC, CO,  and
NOx emissions that occur during high speed and load non-FTP
driving behavior result from commanded enrichment; poor transient
fuel control; and inadequate NOx catalyst conversion efficiency
levels.  The potential control strategies for these causes in
emission increases, as discussed in the above sections,  should
reduce emissions substantially.  The two other potential control
strategies; drive-by-wire and larger catalysts,  are very costly
and would only result in a very small additional reduction in
emissions.   Therefore, EPA feels that the achievable levels of
emission control should be based on the reduction or -elimination
of commanded enrichment; optimization of transient fuel control;
and the improvement of NOx catalyst conversion efficiency levels.
                               52

-------
4.4  Level of Control
   Section 4.3  identified recalibrations  as  the  emission control
strategy EPA believes will be the most cost effective.  Thus,
recalibration was the emission control strategy assumed in
determining the appropriate levels of control and the most cost-
effective control program for aggressive driving.  The discussion
below assumes the US06 as the control cycle, thus, the discussion
on appropriate emission levels  are specific to'the US06 cycle.

   EPA's  conceptual  approach  to  establishing the appropriate
level of control is based on the premise that it is reasonable
and feasible to expect off-cycle air/fuel calibration .and the
associated emissions to be consistent with calibration and
emissions found on  FTP.  The level of control that these
strategies can achieve over aggressive driving is pollutant
specific, reflecting the different impact on the pollutant levels
of factors like__e_ngine load and air/fuel calibration.    As
discussed above, this level of control can be attained by the
elimination of commanded enrichment and improved air-fuel
control.   EPA accepts that there are exceptions, such as high
load events of an extended nature--towing a trailer up a long
grade.  Trying to control emissions to FTP levels under these
severe, infrequent,  events would likely lead to the undesirable
effect of  catalyst deterioration.

4.4.1  HC Control
   I/i  establishing HC  tailpipe emission levels for US06,  EPA
began with the hot,  stabilized emission levels on the FTP; this
corresponds to bag 2 of the FTP, as bags 1 and 3 include start-up
emissions.  The US06 emission data were simulated by splicing
together the appropriate seconds of emission data from, the US06
"parent"  cycles, REP05 and ARB02. (Both of the latter cycles were
also tested in a hot stabilized condition.)   In the following
tables, emission results for both -production and stoich
configurations  are presented for US06,  while for  bag 2 of the
                                53

-------
FTP and the full FTP, only the production emissions are shown, as
a baseline.
   EPA believes  that  because  engine-out  hydrocarbon  emissions
vary directly with load on the engine, it should be possible to
achieve comparable per-mile HC emissions over two cycles of
comparable average load, as long as similar catalyst conversion
efficiencies are maintained.   Comparisons of the fuel economy
results from the US06 and bag 2 of the FTP indicate that the US06
cycle, although clearly more aggressive in the speed and
acceleration of its individual events, actually has a slightly
lower average load (on a per mile basis) than bag 2 of the FTP.
These results are presented in figure l-15a.  In nearly all cases
the US06 fuel economy (miles per gallon) is above that for bag 2
of the FTP.  The average load equivalency established by the fuel
economy results is supported by a comparison of engine-out
emissions.  Figure l-15a compares engine-out HC for  bag 2 of the
FTP and US06, and the full FTP.   The data indicate that in both
the production and stoich calibration, US06 engine-out emissions
were lower than the vehicle's corresponding bag 2 emissions.
For tailpipe emissions,  figure 1-16  suggests that while
production calibrations show large differences between the US06
and the FTP cycles, vehicles using the stoich calibration had
US06 emissions which were comparable to FTP bag 2 emission with
the exception of 312(Saturn).  The Saturn uses  throttle body
fuel injection,  a relatively old technology, and demonstrated
poor transient fuel control and thus EPA feels it should be
discounted.  Overall, EPA feels it is reasonable to expect that
with proper calibration to maintain good catalyst conversion
efficiency, US06 HC levels can be controlled to the equivalent of
bag 2,  FTP  levels.

4.4.2  CO Control
   Tailpipe CO emissions  in the production  calibration are
extremely high on the US06,  as shown in figure 1-17. With stoich
calibrations, five vehicles had the same or lower emissions on
                               54

-------
the US06 compared to the full FTP, while three vehicles were .
higher.  In the stoich calibration,  only vehicle 305(Seville)
had US06 emissions that were lower than FTP bag 2 emissions.  The
Seville was one of the high performance vehicles in the test
program and rarely got into high throttle openings even on the
US06, and thus, spent little time in commanded enrichment.
   The  sensitivity of  CO  emissions to  the drive  cycle  is  further
illustrated in figure 1-18.  In the stoich calibration, the
relationship between engine-out US06 CO emissions and FTP
emissions is similar to the corresponding tailpipe relationships;
the same vehicles tend to be higher or lower than the FTP.  This
indicates that catalyst conversion efficiency plays a minor role
in the difference between HC and CO emission response on the
US06; rather,  most of the effect is due to the extreme
sensitivity of engine-out CO emissions to any minor excursion in
air/fuel ratio.  Such  excursions are likely the result of
incomplete contjrol of commanded enrichment or transient
enrichment.  These results lead EPA to conclude that holding US06
CO emissions to full FTP levels instead of bag 2 FTP levels is
more appropriate.  At this level it still will achieve a large
reduction in offcycle. CO emissions without forcing expensive
control systems,  such as drive-by-wire.

4.4.3.   NOx Control
    In  contrast to engine-out HC which increases  linearly with
load, engine out NOx increases exponentially.  While the average
load on the US06 is similar to bag. 2 of the FTP, the high
instantaneous loads during the hard: accelerations of US06 result
in a nonlinear increase in engine-out NOx emissions.  Figure 1-19
shows large differences between US06  NOx emissions and the FTP
emissions--bag 2 as well as the full FTP.   For nearly all
vehicles on every test, the NOx emissions during for the stoich
tests are higher than those for the production tests.   As
discussed earlier, the elimination of commanded enrichment has
the undesirable effect of increasing NOx emissions due to higher
                               55

-------
engine temperatures
   The  comparison  of  tailpipe NOx  emissions   suggests   a  problem
in trying to control US06 emissions to bag 2, FTP levels (figure
1-20).   In the production configuration,  among the 16 LDVs, only
two vehicles (314 and 401)  had US06 emissions .below the bag 2,
FTP levels; three other  vehicle were below the full FTP levels.
In addition, the  stoich NOx levels ran higher than the
production levels for every vehicle except the Saturn and this
vehicle had high HC emissions,  as discussed above. However,  that
fact that with the stoich calibration, tailpipe NOx increased
proportionally far more than engine-out NOx on every vehicle
suggests that mediocre NOx conversion efficiency accounts for
most of the NOx increase between production and stoich, and it
indicates that most vehicles are not calibrated for optimal NOx
catalyst conversion efficiency.   Evaluation of the catalyst
conversion efficiency impacts,  for each vehicle tested, indicated
that the vehicles with large drops in NOx conversion efficiency
in stoich configuration also had fuel control which allowed
significant lean A/F episodes.
   Calibration  changes  can greatly reduce  the engine out  NOx;  but
the EPA believes that above results suggest that level is some
margin above the levels found on bag of the FTP.  Figure 1-21
presents the hypothetical NOx conversion efficiency that would be
required on US06 in order to bring US06 NOx emissions down to
full FTP  levels.  While these rates of efficiencies are high,
there is evidence to suggest they are achievable.   The required
conversion rates are, in fact,  less than  or equivalent to the
actual NOx conversion efficiency achieved on the FTP,  bag 2 for
10 of 16 vehicles in the production configuration.    In
addition, although it is not shown on the graph, the.Grand Am in
production calibration achieved an overall US06 NOx conversion
efficiency of 96.4%,  well above the efficiency required to bring
US06 emissions down to full FTP levels.  In fact,  there we're 5
vehicles with actual NOx conversion efficiencies greater than 95
percent.   On this basis, EPA believes that with adequate
                                56

-------
attention to A/F control, manufacturers can attain NOx conversion
efficiencies during US06 operation that are on par with levels
over the bag 2 of the FTP, and in the process control NOx
emissions to the levels found on the full FTP.
    One  issue  EPA has  not  fully  evaluated  is  possible
correlations between NOx and CO emissions.  It is likely that the
optimal control of NOx emissions would involve slight rich
biasing of the A/F ration to improve NOx conversion efficiency,
and/or limited amounts of commanded enrichment to control engine-
out NOx emission levels.  Either strategy, if needed,  could
increase CO emissions.
   In  summary,  EPA believes  that  US06  emissions can be greatly
reduced by manufacturer recalibration to eliminate or greatly
reduce commanded enrichment, and to tighten A/F control to
increase HC arid NOx conversion efficiency.  HC emissions on US06
can be reduced  to the same level as a vehicle's FTP,  bag 2
emissions.  For—CO and NOx,  the US06 control level is the full
FTP emission level.   No explicit gram per mile  emission levels
are presented here.   Rather, as discussed in the preamble, these
vehicle-specific control levels are converted into a numerical
standard as part of a composite standard.
Section 5.  Technological  Feasibility
   This  section  discusses EPA's  assessment  of  the  effects  on
vehicle performance and durability of using these feasible
technological controls to comply with the proposed emission
levels over the US06 cycle.  Specifically,  this section will
focus on the impact-on performance and driveability,  and catalyst
and engine durability.

5.1  Impact on Performance
   Automotive manufacturers have indicated  that  in today's
society, automotive performance has become very important..  The
motoring public has come to expect a certain level of performance

                                57

-------
out of their vehicles.  Vehicles today have the greatest
combination of fuel economy and power in history.  Fuel economy
is at an all time high, while 0-60 mph times for the  average
vehicle continues to decrease.  There are a number of factors
that contribute to this: improvements in aerodynamics;  lighter
materials; better tire designs; advanced electronic engine
management with multiport fuel injection; electronic four speed
transmissions; improved, more sophisticated engine calibrations;
and more efficient, higher output engines, to name just a few.
   The  mix of  vehicle  types  today is  more  diverse than in  the
past.  Coupes, sedans, 2-seaters, station wagons, convertibles,
recreational vehicles, sport utilities,  mini-vans, full size
vans, small pick-up trucks,  and large pick-up trucks are just
some of numerous categories of motor vehicles available on the
market.  Manufacturers have indicated that there are two common
requirements that the motoring public demands from all of, these
different vehicle types: good fuel economy and vehicle
performance. One of the main reasons that vehicles have been able
to achieve high fuel economy and good vehicle performance is that
over the past decade,  engine displacements have been slowly
getting smaller while still maintaining excellent power.  There
are more four and six cylinder engines than ever, and the number
of eight and ten cylinder engines has been slowly increasing from
an all-time low,  although .their displacements have dramatically
decreased over the past 10-20 years.
   Two  factors  that  are essential  for good vehicle performance
are^> power and driveability.  Power is the ability of the engine
to perform work,  while driveability is how well the engine
performs that work.  Power is defined as the engine's ability to
perform work in a given time. For example/ the rate at which a
vehicle is able to accelerate at, or the ability to pull a
trailer up a mountain.pass,  are functions of power.   Driveability
is defined as how well a vehicle operates and performs; i.e.,  how
well the engine starts, how smooth the vehicle accelerates,  a
steady imperceptible idle, whether the engine  stalls, or if the
                                58  '

-------
vehicle surges, hesitates, or stumbles.  It's possible to have
good power with poor driveabity and to also have the opposite,
poor power with good driveability.   The two are not directly
related.

5.1.1   Power
   One  of  the  most  important  factors  in vehicle  performance  is
engine power output.  Power is the ability of the engine to
perform work.   The best examples of power are the rate at which a
vehicle is able to accelerate at,  and/or  the ability to pull or
carry heavy loads,  such as a pulling a camper or carrying a load
of lumber in the bed of a pick-up truck.   In addition to power
output, there are many other factors that contribute to overall
vehicle performance:  aerodynamic drag, transmission type, gear
ratio,  final drive ratio, vehicle weight, and tire size and type.
Another important factor in vehicle performance is the vehicle
weight-to-power^ratio (W/P).   Light vehicles with engines that
have high power outputs have far better vehicle performance than
heavy vehicles equipped with low power output engines.
   Even though there  are  a number of  factors  that  influence
performance, this discussion will focus on power output.  Not
only is power output one of the most important factors in vehicle
performance, as stated above, but it is also the factor that is
the most directly affected by trying to .control emissions over
the US06 cycle.  Vehicle manufacturers have been concerned that
in order to control emissions, especially HC and CO, over a high
load/high speed cycle,  it would mean the elimination of any
commanded enrichment which is typically used during heavy load
operation, such as aggressive accelerations or pulling trailers,
for increases in power output and for catalyst and engine
cooling.  They have even expressed concern that any loss in
power,, especially for smaller vehicles with high^W/P ratios,
would result in unsatisfactory and even dangerous vehicle
performance and may require the replacement of small
displacement,  fuel efficient engines with larger displacement
                                59

-------
four and six cylinder engines that could have poorer fuel
economy.  Because of these concerns, EPA feels that the issue of
power loss due to the reduction or elimination of commanded
enrichment, is a very important issue for the feasibility of
technological control of the  proposed emission levels over the
US06 cycle.
   There are  a  number of  factors  that  contribute  to  engine  power
output.  Engine design parameters, such as combustion chamber
type, compression ratio,  cylinder bore and stroke, and exhaust
and intake manifold configurations, are essential to maximum
power output.  Another important factor is the management of
specific engine control systems,  such as the spark timing and
metering of air and fuel.  The metering of air and fuel, or more
specifically, the control of the air-fuel ratio,   has the
greatest effect on tailpipe exhaust emissions and is also the
most affected by operating over the US06 cycle.
   Data from  the—test program  indicates  that,  over the  FTP,
approximately 96%13 of all fuel control operation  for most
vehicles occurs at a stoichiometric air-fuel ratio.
Stoichiometry is typically  in the range of 14.6:1 to 14.7:1.
Only about  0.2% of total fuel control operation over the FTP is
commanded enrichment. The frequency of commanded enrichment over
the FTP was extremely low; in fact, the vast majority of vehicles
never went into commanded enrichment over the FTP.  For those
rare occasions where a vehicle did go into commanded enrichment
     i
on the  FTP,  the enrichment duration was very short, on the order
of ^-2 seconds. Air-fuel ratios as rich as 12:1 were observed for
some vehicles,  but only for a very short period.  Over the US06
cycle, the average air-fuel ratio was very similar to that for
the FTP, however, the rich excursions resulting from commanded
enrichment were much more frequent and of a longer duration.
Approximately 92% of all fuel control operation over the US06
cycle was at Stoichiometry,  while approximately 3.5% was
                                60

-------
commanded enrichment14.   The average air-fuel ratio for a
commanded enrichment event  over  the US06 cycle was approximately
12.5:1 with some vehicles having events as rich as 11.5:1.  The
average duration of commanded  enrichment events over the US06
cycle was 5.015  seconds.
    Information gathered from literature and the manufacturers
have indicated, that for most  engines,  maximum horsepower occurs
at an air-fuel  ratio rich of stoichiometry (typically between
12:1 to 13:1).   It should  be  noted that the extra power gained
with enrichment used to  be  much  more than they are today, due to
fuel distribution problems.  The use of PFI  has virtually
eliminated fuel distribution problems,  allowing higher power
levels without  the need  for extra fuel.  The air-fuel ratios
experienced by  vehicles  in  the test program over US06 typically
feel between  12:1 and 13:1.    In discussions with EPA,  several
vehicle manufacturers indicated  that operating the engine at a
steady stoichiojnetric air-fuel ratio,  rather than at the air-fuel
ratio for maximum power.of  approximately 12:1 to 13:1, during an
extended acceleration or during  heavy engine load operation
(i.e., pulling  a trailer up a  hill)  would result in a 3%-10% loss
in horsepower depending  on  the engine and vehicle application.
   The  impact  of a'  3-10% loss in power  is- relative to  the
specific vehicle model -and  the horsepower of the particular
engine in that model.  For  example,  if the Dodge Viper,  with a
rated power of  400hp arid a  weight-to-power ratio of 8.5,
experienced a 3-10% loss in power,  it would result in a loss of
12 tjo 40hp, leaving the  vehicle  with a rated power of 360 to
388hp and weight-to-power ratios of 9.4-8.8.   While this may
cause some concern to the marketing people,  the performance of
the vehicle should still'be excellent and not cause any safety
     "Memorandum from Phil Enns to John German, dated December 6, 1993 and
titled "Enrichment Event Analysis."

     15Memorandum from Phil Enns to John German, dated December 6, 1993 and
titled "Enrichment Event Analysis."

                                61

-------
concerns.  On the other hand, the Geo Metro has a power-rating of
55hp with a weight-to-power ratio of 38.6.  A loss of 3-10% in
power would result in a loss of 2.0 to 5.5hp,  leaving the vehicle
with a rated power of 49.5 to 53hp and weight-to-power ratios of
42.9-40.1. It's much less clear .whether this type of power
reduction for this vehicle would result in unsatisfactory
performance or even be a safety issue.
   Unfortunately,  EPA was  unable  to  measure  the power  loss  that
resulted from removing commanded enrichment from the production
calibrations to the stoich calibrations.  Due to logistical and
timing constraints, neither EPA nor the manufacturers were able
to operate the vehicles with stoich calibrations on the road to
determine the effect eliminating commanded enrichment had on
vehicle performance.   Therefore, in an attempt to find a way to
evaluate the effect that eliminating commanded enrichment had on
power output and vehicle performance, EPA compared the
differences in .wide open throttle (WOT) times for the production
and stoich calibrations over various accelerations from the US06
cycle.   The general approach was to see whether or not any
significant differences in WOT times occurred for the vehicles
when operating with the stoich calibrations, which had no
commanded enrichment, compared with the  production calibrations
that had commanded enrichment.  If WOT times were significantly
different, then an argument could be made that the elimination of
commanded enrichment  would indeed influence vehicle performance.
However, if the difference in times were small,  it could be
argued that the impact on vehicle performance was minimal and any
reduction in power output should be insignificant for most
vehicles.  It should also be noted that the stoich calibrations
simply had the commanded enrichment strategies removed, while
other fuel and spark strategies were left  alone. They were not
optimized for performance or emissions and therefore,  do not
represent optimized production-level calibrations with commanded
enrichment eliminated. Thus the results should represent a
scenario worse than what might end up in production.
                                62

-------
   EPA  realizes  that  this method  for  analyzing  effects  on  vehicle
performance may be overly simplistic and does not convincingly
prove or disprove that the feasible technological controls
proposed by EPA to comply with the proposed US06 cycle emission
levels will not affect vehicle performance.  However,  without any
other mechanism to accurately measure power loss or quantify
performance effects,  EPA believes that this approach will provide
a general understanding as to whether or any negative effects on
vehicle performance will tend to occur as a result of complying
with the proposed US06 cycle emission levels.
     EPA evaluated the WOT times for 11 vehicles that had been
tested with both  production and stoich calibrations.   The
maximum continuous WOT time that occurred during any particular
acceleration and the total amount of WOT time that occurred over
the cycle for each vehicle was examined.  While total WOT time
can be a good indicator for the frequency of WOT for a given
vehicle, it is .hard to make any assessments as to the impact on
vehicle performance.   For example, 20 seconds of total WOT time
could consist of two WOT events lasting 10 seconds each or 20 one
second WOT events spread throughout the cycle.   Maximum
continuous WOT time is better for assessing effects on
performance since that represents the worst case scenario for
each vehicle.  For the purpose of this analysis, WOT was defined
as any throttle opening greater than 90%.
   Table 1-13  showf tB»  iTCttMW* from production to  stoich
calibrations in mvartfCmaxima* continuous and total WOT times
for/all 11 vehicles over the US06 cycle.  These increases
demonstrate the effect of eliminating or reducing commanded
enrichment on WOT times.
                                63

-------
                            Table  1-13
           Average  WOT Time  Increase  From Production to
            Stoich  Calibrations  for  Individual  Vehicles
Vehicle
Suburban
Metro
CIO P/U
Supreme
Grand-Am
Sonoma
Escort
Grand
Prix
Olds 98
Seville
F250 P/U
Average
Max
Prod
9
6.5
4
3
5.5
1.5
7
2.5
1-rS
3
3
4.2
Max
Stoich
10
7.5
5
4
5.5
1.5
8
3.5
2
0
3.5
4.6
Max
Increase
1.0
1.0
1.0
1.0
0.0
. 0.0
1.0
1.0
0.5
-3.0
0.5
0.4
Total
Prod
26.0
31.5
16.5
6.5
11.5
1.5
36.0
8.0
5.5
4.0
14.0
14.6
Total
Stoich
38.5
38.0
22.0
11.5
16.0
2.0
35.5
5.5
3.0
0.0
9.0
16.5
Total
Increase
12.5
7.5
5.5
5.0
4.5
0.5
-0.5
-2.5
-2.5
-4.0
-5.0
1.9
     The average increase in maximum continuous WOT time for all
of the vehicles was 0.4 seconds, while the average increase in
total WOT time was 1.9 seconds. .These increases appear to be
minimal.  EPA.feels that such minimal increases would seem to
illustrate that any losses in power output and consequentially
vehicle performance,  would be negligible.
        Three of the vehicles had no increase in maximum
continuous WOT time and only marginal increases in total WOT
time.  In fact, the Cadillac Seville had a decrease in WOT time.
This brings up an interesting observation that was made on all 11
vehicles.  The way the vehicle was driven over the cycle had an
impact on WOT times.   For example, on the initial acceleration of
                                64

-------
the third hill of the cycle, the Seville had an of average of 3.0
seconds continuous WOT operation in the production calibration
but never went above 80% throttle with the stoich calibration
over the same acceleration.  To illustrate this behavior, figure
1-22 presents the drive trace data for one of production tests. '
It appears that the driver got behind the speed trace and had to
go WOT for 3.0 seconds to catch back up with the trace.  On one
of the stoich tests(figure 1-23), in contrast, although the
driver got behind the trace, the speed was .still within the speed
tolerances, and it appears that the driver didn't attempt to
catch back up. Thus, the vehicle  did not go into WOT.  This
particular observation seems to distort the results for the
Seville to some extent, but not entirely. Chances are good that
if the driver had behaved the same way for all of the stoich
tests as they did for the production tests, the WOT time results
would have been very similar.  Thus for this vehicle, the effect
on vehicle performance over the cycle seems to be negligible.
      The Escort,  Grand Prix, Olds 98, and F250 P/U all had total
WOT times that decreased.  Again, examination of the driving
trace data for the individual tests revealed that,  for all of
these vehicles, there were situations, where on certain
accelerations, the driver attempted to follow the trace very
closely and would occasionally go into WOT for a period of time,
whereas on other occasions over the same accelerations, the
driver could avoid WOT operation by not following the trace as
closely but still staying within the speed tolerances.  From the
perspective of affecting power output and vehicle performance,
these observations suggest that driver behavior was more
important than any power change from production to stoich
calibrations.  This would seem to .further enforce the position
that the difference in1performance between the production and
stoich^ calibrations is minimal and the effect on vehicle
performance is negligible.
     The vehicles with the highest WOT time increases were the
Chevy Suburban and Geo Metro.  The Suburban had an average
                                65 ^

-------
maximum continuous WOT time increase of 1.0 second and a total
WOT time increase of 12.5 seconds.  Examination of the drivers
traces indicated that  the same driving, phenomenon that occurred
for the Seville also occurred for the Suburban.  There were
numerous instances where the driver attempted to follow the trace
very closely and went into WOT, and cases where the driver didn't
follow the trace as closely but still remained within the speed
tolerances and subsequential didn't go into WOT.
     The Geo Metro had an average maximum continuous WOT time
increase of 1.0 second and a total WOT time increase of 7.5
seconds.  The Metro's WOT performance also seemed to be
influenced by how it was driven over the cycle, but not nearly as
much as the rest of the vehicles.  The Metro was equipped with a
manual transmission which would typically give it an advantage in
performance over a similar vehicle equipped with an automatic
transmission.  The Metro had a W/P ratio of 38.6 and was the
worst case vehicle for W/P ratio in the test program.  As
discussed in section 3, EPA will allow load adjustments for  low
W/P vehicles so that they will not be penalized by the
aggressiveness of the cycle that it is more apparent for these
type of vehicles.  The Metro was tested in both the production
and stoich calibrations with load adjustments made.  Reductions
were made to the inertia weight and aerodynamic drag coefficient.
When retested over the US06 cycle in the production and stoich
calibrations with the load adjustments, the Metro actually had a
reduction in maximum continuous WOT time of 0.5 second and a
reduction in total WOT time of.1.0 second.  The average maximum
continuous and total WOT times were 5.0 seconds and 13 seconds,
respectively, for the production calibration and 4.5 seconds and
12 seconds for the stoich calibrations.  It would appear that
these differences also support the assumption that the
differences in WOT time are insignificant and that the effect on
power loss and vehicle performance is negligible.
       Another factor that may have influenced WOT times was
automatic transmission shift schedules.  Comments received by EPA
                                66

-------
test drivers and test engineers were that automatic transmission
shift schedules for several vehicles were obviously not
calibrated for the US06 cycle because they shifted either too
early or too late and made it difficult for the vehicles to
maintain the speed trace.  This further illustrates that the way
the vehicle was operated over the cycle had a greater impact on
WOT times than the elimination of commanded enrichment.
     As stated earlier, the real concern about losses in vehicle
performance resulting from control strategies necessary to comply
with proposed US06 cycle emission levels have centered around
lower performance vehicles.  Moderate to high powered vehicles
seem to have a sufficient
combination of power and available gearing that proposed control
strategies such as the reduction or elimination of commanded
enrichment  should have negligible effects on vehicle
performance. The results of EPA's analysis on WOT time
comparisons seem£ to indicate that even for low powered vehicles
the effects on performance should be negligible,  especially since
they will be designed to reflect operation that occurs over the
FTP and the lowest performance vehicles will have load
adjustments over the cycle that should put them on an even par
with the moderate performance vehicles.
     EPA believes that some vehicles may inevitably suffer some
losses in engine power and vehicle performance as a result of
complying with proposed US06 cycle emission levels. However, EPA
expects that for the vast majority of vehicles, the loss in power
and /vehicle performance will be negligible.

5.1.2     Driveability
     There are .numerous definitions o-f driveability used
throughout the automotive industry. EPA views driveability as how
          u
well a-vehicle, operates.  How well it starts, idles,  accelerates,
decelerates, and cruises.  Does it have spark knock,  surge,  or
any hesitations when accelerating? Does the idle quality
deteriorate when the rear defroster is on or can the driver feel
                                67

-------
the air -conditioning compressor engaging?  All of these are part
of driveability.  Driveability, unlike vehicle performance, is
very subjective. There are very few measurements of driveability
that can be quantified like rated horsepower, torque, or 0-60mph
times.    Because of this, it is very hard to assess the effect
that proposed feasible technologies and control strategies for
complying with the proposed US06 cycle emission levels will have
on driveability. - Adding to this complication is the fact that
the only operation of the test vehicles occurred on the chassis
dynamometer (referred to as the rolls) during emission tests.  No
vehicles were operated on the road and evaluated for driveability
issues.  It is very difficult to make driveability evaluations on
the rolls. The somewhat harsh ride that occurs on the rolls tends
to mask many driveability problems. Neither EPA nor manufacturer
test personnel were prepared to evaluate driveability.  Drivers
and test operators typically report any unusual vehicle behavior.
The manufacturers, never reported any driveability problems to
EPA.
     All of the proposed control strategies and technologies  for
complying with the proposed emission levels should further
enhance driveability.  Control strategies, such as tighter air-
fuel ratio control over the majority of off-cycle high speed and
load operation,  and technologies, such as PFI systems,
sequential-fire PFI systems,  or drive-by-wire systems, are all
geared towards improved open and closed-loop operation and should
be valuable assets in improving driveability.  None of these
'strategies or technologies should have adverse effects on
driveability.   The only item that could potentially be an issue
                           /
is the reduction or elimination of commanded enrichment.
However, EPA believes that the removal of commanded enrichment is
a performance issue rather than a driveability issue.
     EPA feels that the effect of complying with proposed US06
cycle emission levels' on driveability should be minimal. Although
no evaluation on the subject was performed by EPA or the
                                    9
manufacturers, good engineering judgement supports EPA's
                                68

-------
assumption. The control strategies and technologies that should
be necessary to comply with the proposed emission levels should
help rather than hinder driveability.  The lack of any obvious
driveability concerns- reported during the test program indicates
that operation with the stoich calibrations,  which were not
optimized for driveability, had no blatant driveability problems.
5.2. Impact on Durability
     One of the greatest concerns over the control strategies and
technologies analyzed as feasible  to comply with the proposed
US06 cycle emission levels, is the impact on catalyst and engine
durability.

5.2.1     Catalyst Durability
     The discussion on catalyst durability will focus on the
following areas:           .
•    Catalyst thermal degradation
•    Production catalyst temperatures over the US06 -cycle
•    Stoich catalyst temperatures over the US06 cycle
•    Temperature differences between production and stoich
     calibrations
•    Summary
5.2/1.1   Catalyst Thermal Degradation
     The catalytic converter is the single most important part of
any vehicle's emission control system.  Because it is the last
element in the emission control system, it provides the final
means to decrease the level of undesirable tailpipe emissions.
The catalyst's function is to initiate two different types of
chemical conversions, oxidation and reduction.  The products of
incomplete combustion, i.e. hydrocarbons and carbon monoxide, are
oxidized into carbon dioxide and water vapor,  and the nitrogen
                                69

-------
oxides are reduced to molecular nitrogen and other products
depending on the reducing agent.  Catalyst deterioration due to
thermal exposure can cause these two types of conversion to occur
with reduced efficiency, allowing more pollutants generated
upstream to be emitted into the atmosphere.
     A major cause of catalyst deterioration is thermal
degradation, which results from excessive catalyst temperature.
Prolonged exposure to excessive temperatures results in washcoat
surface area loss and/or sintering of the noble metals.  Extreme
cases  can lead to  monolith meltdown.   During stoichiometric
operation,  the ratio of reducing agents (i.e.,   HC, CO, H,  etc.)
and oxygen is optimum for promoting oxidation of HC and CO and
the reduction of NOx.  However, these reactions can be thought of
as simply finishing the combustion process in the .catalyst.
These processes burn up the HC and CO, which dramatically raises
the catalyst temperature.  Under normal FTP type operation, the
mass flow rate of emittants entering the catalyst is fairly low,
thus the exotherm (or heat generated) doesn't typically raise
catalyst temperatures to high levels.  Under high load
accelerations, the mass flow rate of the emittants is much
greater, which has two effects:  The temperature of the exhaust
gases entering the catalyst tend to be higher,  and the higher
mass of emittants increases the exotherm.   This combination
produces higher.temperatures  that could  potentially damage the
catalyst.   A common strategy for avoiding these high
temperatures, is to use a rich strategy during these high load
accelerations.  The additional fuel acts as a heat sink in the
   r
engine, lowering exhaust temperatures, and restricts the amount
of oxygen available for oxidation and lowers the exotherm,  thus
lowering the catalyst temperature.
     It is generally accepted throughout the automotive industry
that vehicle operation which results in catalyst temperatures
below 900 C will not result in thermal degradation of catalyst
conversion efficiency for catalysts containing  platinum (Pt)  and
                                70

-------
rhodium  (Rh)  .16    For palladium  (Pd) catalyst,  this  temperature
may be as high as  950°C.    Complete failure of  the catalyst  could
be expected when the  cera'mic substrate reaches  1093  C.

5.2.1.2.  Production  Catalyst Temperatures
     Table 1-14 shows the average of the maximum catalyst
temperatures,  time  at temperatures greater than 816  C  (1500  F),
and continuous time at temperatures greater than 816 C,  for  13  of
the vehicles  tested in the production and stoich calibrations
over the ARB02, REP05,  and US06 cycles.  These vehicles  represent
a fairly wide  range of vehicles;  9 passenger cars, 2 light-heavy-
duty trucks,  and 2  light-duty trucks.
     16C.D.  Tyree,  "Emission Levels and Catalyst Temperatures as a Function of
Ignition Misfire," SAE Technical Paper  92098
                                71

-------
                            Table 1-14
              US06 Production Catalyst Temperatures
Vehicle
Escort
Olds ,98
GranAm
Metro
Camry
F250
CIO
Supreme
Saturn
Gran
Prix
Cruiser
Sonoma
Seville
Suburban
Avg
Type
LDV
LDV
LDV
LDV
LDV
LHDT
LOT
LDV
LDV
LDV
LDV
LDT
LDV
LHDT

Time @
> 816
C
(sees)
70
. 24
19
6
3
0
0
0
0
0
0
0
0
0
8.7
Cont.
Time
(sees)
20.5
23
19
6
3
0
0
0
0
0
0
0
0
0
5.1
Time @
> 872
C
(sees)
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Cont .
Time
(sees)
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Max
Temp
(C)
864
843
832
825
817
783
773
766
753
739
732
697
692
670
770
Cycle
REP05
US06
US06
ARE 02
US06
ARB02
US06
US06
ARB02
ARB02
US06
US06
ARB02
ARB02

     The average of the maximum catalyst temperatures  was 770°C.
The range of temperatures was 194°C.   Five vehicles, the Escort,
Olds 98, Grand-Am, Camry, and Metro-experienced temperatures over
816° C  (1500° F).     The Escort  and Camry were the only vehicles
equipped with close-coupled three-way catalysts.   The Escort had
                                72

-------
a single three-way catalyst,  while the Camry had a light-off
catalyst followed by a three-way underfloor catalyst. .The rest
of the catalyst temperatures were for three-way underfloor
catalysts that were not close-coupled.
     It is important to examine the maximum catalyst temperatures
over all three cycles:  ARB02; REP05; and US06,   because all
three of these cycles consist of actual in-use driving events.
Any catalyst temperature experienced over these cycles could also
occur in-use.    Therefore,  it can be assumed that the catalysts
for these vehicles are  already designed to withstand such
temperatures.

5.2.1.3.  Stoich Catalyst Temperatures
     Table 1-15 displays the same information as table 1-14 for
the stoich  catalyst temperatures.
                               73

-------
                           Table  1-15
                US06  Stoich Catalyst  Temperatures
Veh
Escort
GranAm
Olds 98
Metro
F250
Cruiser
Supreme
CIO
Saturn
Gran
Prix
Seville
Sonoma
Suburban
Avg
Type
LDV
LDV
LDV
LDV
LHDT
LDV
LDV
LOT
LDV
LDV
LDV
LOT
LHDT

Time @
> 816
C
(sees)
136.5
58.5
37
18
30
0
0
0
. 0
0
0
a
0
21.5
Cont .
Time
(sees)
47.5
19.5 .
24.5
11
26
0
0
0
0
0
0
0
0
9.8
Time @
> 872
C
(sees)
48.5
1.5
0
0
0
0
0
0
0
0
0
; o
0
3.8
Cont.
Time
(sees)
20.5
1
0
0
0
0
,0
0
0
0
0
0
0
1.6
Max
Temp
(C)
920
872
851
843
826
820
812
788
782
776
732
708
686
801
Cycle
US06
US06
US06
US06
US06
US06
US06
USD 6
US06
US06
US06
US06
US06

   t
     The average of the maximum catalyst temperatures with the
stoich calibration  was 801° C.     .Seven vehicles  had maximum
catalyst temperatures exceeding 800°C.  The'range of temperatures
was 234°C,  ranging from  the  Escort  which had the highest
maximum temperature at 920°C to the Suburban which had the lowest
at 686°C.   The Escort  was the only vehicle to exceed the 900°C
                                74

-------
threshold.   Figure l-23b shows that for one test,  the Escort
spent as much as 11 consecutive seconds over 900°C as a result of
the acceleration in the middle of the third hill of the US06
cycle that simulates a high speed passing maneuver.
     All of the maximum stoich catalyst temperatures found in
table 1-15 are those that occurred over the US06 cycle.   This is
because manufacturers will only be required to design their
catalysts to withstand temperatures that occur over this cycle
rather than temperatures that could result from any in-use
accelerations or loads more severe than those occurring on the
US06 cycle.  Manufacturers will try to optimize the amount of
enrichment they will be able to use for catalyst cooling by
determining how long they will have to operate at or near
stoichiometry during the WOT conditions that will be implemented
by the US06 cycle.    For example, if the most aggressive
acceleration on the cycle requires a vehicle to operate at WOT
for five seconds^  then that is the  amount of time that they
will have to operate at stoichiometry.  For any WOT operation
tha.t occurs in-use that is greater than five seconds, the
manufacturer will be allowed to use some cooling enrichment for
however much WOT operation occurs after the five seconds has
surpassed.

5.2.1.4.  Temperature Differences Between Production and Stoich
          Calibrations
     Table 1-16 illustrates the increase in the average of the
maximum catalyst temperatures for the vehicles when tested with
the production and stoich calibrations.
                                75

-------
                            Table 1-16
       Increase  in average  of  maximum catalyst  temperatures
            between production and stoich  calibrations
Vehicle
Cruiser
CIO
Escort
Supreme
F250
Saturn
Grand -Am
Metro
Seville
Grand
Prix
Olds 98
Suburban
Sonoma
Average
Type
LDV
LDT
LDV
LDV
LHDT •
LDV
LDV
LDV
LDV
LDV""
LDV
LHDT
LDT

W/P
26.4
23.8
31.3
22.0
25.0
30.9
21.0
38.6
15.7
19.4
23. 5 '
28.6
17.3

Stoich
temp
820
788
920
812
826
782
872 .
843
732
776
851 .
686
708
801
Prod
temp
732 .
773
864
766
783
753
832
825
692
739
843
670
697
767
Increas
e
88
• 15
56
46
43
29
40
18
40
37
8
18
11
34
     The average increase in maximum catalyst temperature between
the production and stoich calibrations was 34°C.  The largest
increase was 88°C for the Cruiser, while the lowest increase was
for 8°C for the Olds 98.   The increase in catalyst temperature
between production and stoich calibrations is very important.
The average and maximum catalyst temperatures experienced over
the US06 cycle are higher than those for the FTP.   This is not
any surprise since the loads and speeds for the US06 cycle are
much higher than those found on the FTP.  While there is an
obvious increase in catalyst temperatures between the cycles, the
                                76

-------
increased temperatures are not a concern, since this type of
                       #
operation exists in the real world and vehicle manufacturers have
had to design their catalyst systems to withstand this type of
operation.  However, in order to comply with proposed emission
levels over the US06 cycle, manufacturers will have to reduce or
even eliminate the extra fuel from commanded enrichment that they
have traditionally relied on to keep catalyst temperatures down.
As the above table shows,  without commanded enrichment, catalyst
temperatures increase beyond the production levels which utilized
commanded enrichment.
     If the increase in temperature is too large,  there is a
strong possibility that the catalyst could suffer thermal
degradation and catalytic conversion efficiency would deteriorate
at a faster rate.  In addition to 'the increase in catalyst
temperatures resulting from the elimination of commanded
enrichment, the level  of the maximum temperature for the
production calibration is important.   For example,  the maximum
production temperature for the Escort is 864°C.  With an average
increase of 56°C, the  maximum stoich temperature is 920°C, which
is over the 900°C threshold that has been traditionally
acknowledged as the upper limit temperature before thermal
degradation starts to  occur.  But, for the Cruiser, which had an
increase in temperature of 88°G and a maximum production
temperature of 732°C,  the maximum stoich temperature is only
820°C,  which is well below the 900°C threshold.   As previously
stated, the average increase in maximum catalyst temperature for
all-of the vehicles was 34°C while the average maximum catalyst
temperature in the stoich calibration was 801°C.    While the
maximum stoich catalyst temperatures and the increases in maximum
catalyst temperature between production and stoich calibrations
are high,  they are lower than what the Agency thought they would
be at the onset of the test programs.  An interesting fact is
that all of the  stoich  catalyst  temperatures for the rest of
the vehicles were lower than the Escort's production temperature.
Since all of the other catalyst temperatures were for "underfloor
                                77

-------
catalysts," that is, catalysts located downstream  of the exhaust
manifold,  rather than right next to it,  this would seem to
indicate that catalyst technology exists that should allow most
underfloor catalysts to withstand the temperature increases that
would result from eliminating commanded enrichment.

5.2.1.5  Summary
      The evidence from the test program indicates that catalyst
temperatures will rise during high load operation as a result of
reducing or eliminating commanded enrichment.  Obviously any
increase in catalyst temperatures beyond current design
tolerances  are a concern to both EPA and industry.  However, EPA
believes that the catalyst temperatures experienced by the
vehicles tested in the stoich configuration are generally low
enough that thermal degradation should not be a concern.

5.2.1.5.1  Underfloor Catalysts
     For underfloor catalysts, the average maximum stoich
catalyst temperatures experienced over the US06 cycle  was 791°C.
The vehicle with the highest maximum stoich underfloor catalyst
temperature was the Olds 98 with a temperature of 851°C.
However, this was only 8°C higher than it's maximum production
temperature! Several vehicles, such as the Suburban (686°C)  and
Seville  (732°C) ,  had temperatures considerably lower than the
average.  EPA believes that these temperatures are low enough
that thermal degradation should not be a concern.  As previously
mentioned,  11 out of 13 vehicles equipped only with underfloor
catalysts,  had stoich temperatures less than the Escort's (close-
coupled catalyst) production temperatures.   It is apparent that
the technology exists to make underfloor catalysts more thermally
durable.  One potential technology that could be used to improve
underfloor  (and close-coupled) catalyst thermal resistance would
be to use the noble metal palladium rather than platinum, since
                               78

-------
palladium is more thermally  durable  than platinum.
                                                   17
5.2.1.5.2  Close-Coupled Catalysts
     Close-coupled catalyst  temperatures  were measured and
recorded for only two vehicles  ;  the  Escort  and Camry.   The
Camry was tested in the production  configuration only, while the
Escort was tested in both  the stoich  and  production
configurations.  The Escort  had a maximum production temperature
of 864°C, while the Camry  had a maximum production temperature of
817°C.    The Escort had the highest   production temperature of
the 13 vehicles.  The Camry  was not only  lower than the Escort,
but it was also lower than three of the vehicles equipped with
underfloor catalysts only.   The Escort was equipped with a single
close-coupled three-way catalyst that was located as close to the
exhaust manifold as possible.   The  Camry  was equipped with a
single close-coupled three-way  light-off  catalyst followed by a
single three-way__underfloor  catalyst.   Over  the US06 cycle, the
Camry operated considerably  more rich than'the Escort.  This may
                                                  t
be responsible for the differences  in the close-coupled catalyst
temperatures between the Escort and' Camry.
      Figures 1-24 and 1-25  compare the production and stoich
catalyst temperature profiles of the  Escort  with the Saturn over
the REP05 and ARB02 cycles.  The Saturn is compared with the
Escort because it used an  underfloor  catalyst and had the same
engine displacement as the Escort and was also similar in weight.
These figures clearly illustrate the  higher  temperatures
experienced by the close-coupled Escort.   In fact,  the close-
coupled Escort catalyst saw  higher  catalyst  temperatures than all
of the other vehicles, including the  Camry.   The  maximum
catalyst temperature experienced over the US06 cycle for the
Escort in the stoich configuration  was 920°C.   Over this cycle,
     17J.C.  Summers, W.B. Williamson, and J.A. Scaparo,  "The Role of
Durability and Evaluation Conditions on the Performance of Pt/Rh and Pd/Rh
Automotive Catalysts," SAE Technical Paper 900495

                                79

-------
the Escort averaged a total of 10  seconds over 900°C.   All of
this time was continuous and occurred during the high speed
passing maneuver found in the middle of hill. 3.    EPA feels that
the length of duration and  maximum catalyst temperatures
experienced by the Escort could result in some catalyst thermal
degradation.   Based on just this one vehicle, EPA cannot claim
that all close-coupled catalyst applications will experience
catalyst thermal degradation during high load operation resulting
from the reduction or elimination of commanded enrichment,
especially since the production catalyst temperatures for the
Camry's close-coupled light-off catalyst was 47°C lower than the
Escorts temperature.   However, EPA also has no reason to believe
that the close-coupled  catalyst  found on the Escort does not
represent typical close-coupled catalyst technology,  and that
it's catalyst temperature profiles should not be fairly
representative of  other close-coupled catalyst applications.
Therefore, EPA .acknowledges that for close-coupled catalyst
applications, there is the possibility that the reduction or
elimination of commanded enrichment could cause some thermal
degradation of the catalyst during high load operation.
     However, EPA believes that in-use driving modes that would
result in elevated catalyst temperatures such that thermal
degradation would occur,  will be so .infrequent that the
occurrence of thermal.damage should be minimal.   Based on in-use
survey data on enrichment activity, aggressive driving accounts
for less than 2 percent of in-use operation.  Because commanded
enrichment can be used in-use for catalyst cooling during any
accelerations or high load conditions that occur beyond the
maximum acceleration event found on the US06 cycle, maximum
catalyst temperatures should not be any higher in-use than over
the US06 cycle.   EPA is in the process .of assessing the loss of
catalyst conversion efficiency over a typical vehicle's life as a
result of increased temperature exposure resulting from the
reduction or elimination of commanded enrichment.  A similar
methodology exists in Section IX, -part E of the Preamble, where a
                                80

-------
projection of  conversion efficiency losses was made for the
insulation of catalysts.  Based on these projected losses,  and
the preliminary work being done for losses due to the reduction
of commanded enrichment using the same methodology,  EPA expects
that losses in catalyst conversion efficiency over useful life
resulting from the reduction or elimination of commanded
enrichment should be very low.
     Given that the temperature increases associated with control
at this level are within design limits,  and. the deterioration
impacts of the increases should be low,  the Agency believes that
additional catalyst system modifications solely to address
catalyst deterioration will be unnecessary.

5.2.2.  Engine Durability
     Actual engine temperatures were not recorded as part of the
test'program.  Instead, exhaust temperatures were measured as a
surrogate.  Vehicle manufacturers have indicated that material
temperature restraints for exhaust manifolds, exhaust valves,
turbo chargers, and oxygen sensors range from 720°C to 850°C
depending on the item and the material used.   Maximum exhaust
temperatures for vehicles tested with stoich "calibrations over
the US06 cycle ranged from 641°C to 816°C.   The average maximum
temperature for all of these vehicles was 705°C.  Only two
vehicles out of 16 had maximum exhaust temperatures greater than
800°C.  They were the Ford F250 pick-up truck and the Ford
Escort.  The maximum temperatures  for the F250 and  Escort were
816%C and 856°C,  respectively.  The escort utilized a control
strategy known as lean-on-cruise that  controls the air-fuel
ratio at very lean levels during high speed cruises as a means of
fuel economy.  The maximum production  exhaust temperatures for
the Escort were 805°C.  Examinations of exhaust temperature
profiles for this vehicle over hill 3 of the US06 cycle, clearly
show that during the lean-on-cruise operation, the exhaust
temperatures raised significantly.  For the stoich calibrations,
this strategy made the situation worse.   Immediately following a
                                81

-------
high speed cruise portion of hill 3, where lean-on-cruise is
active, the vehicle must accelerate through an aggressive  high
speed acceleration.  The exhaust temperature is already high and
still rising during the lean-on-cruise operation, when the
vehicle suddenly accelerates without the benefit of any commanded
enrichment.  The combination of rising high exhaust temperatures
                                                               /
from the lean-on-cruise operation and the sudden large engine
load,  resulting from the aggressive  acceleration, at a
stoichiometric air-fuel ratio caused the exhaust temperature to
rise to 856°C.
     The proposed NOx emission standards for the US06 cycle will
prohibit the use of lean-on-cruise strategies.  Therefore, the
exhaust temperatures experienced by the Escort are not  "
representative of what can be expected as a result of complying
with the proposed emission levels, thus that it why the Escort
was not included in the range of maximum temperatures for the
stoich vehicles^jnentioned above.
     There is no explanation for the relatively high exhaust
temperatures experienced by the F250 pick-up truck.  However, its
temperatures fall right into the middle of the range listed by
the manufacturers and it's possible that the material temperature
restraints for this particular model is higher than 805°C.

     EPA feels that for the vast majority of vehicles, exhaust
temperatures resulting from control strategies necessary to
comply with'proposed US06 cycle emission levels, will not have
any/negative impacts on engine durability.

5.3  Excessive Temperatures and Road Grade
     Section 4.2 discusses the-way in which the US06 cycle
implicitly controls for the effect of road grade on emissions.
The duration  of control for road grade is determined by the
duration of commanded enrichment control over the US06.  Thus,
emission control will be  limited for extended road grade, or
other extended high-load events  (such as certain acceleration
                                82

-------
events while trailer towing).   EPA had concluded that it is not
feasible to control enrichment beyond the US06 duration due to
temperature concerns associated with stoichiometric control of
the air-fuel mixture during high load events  (see section 5.2.1).
EPA believes the infrequent nature of these extended high load
events does not justify expanding control to address these events
given the potential for significant catalyst degradation and/or
the increased hardware cost.

5.4.  Projected Vehicle Modifications

5.4.1.    Modifications Necessary to Comply with Proposed
          Standards
     To comply with the proposed US06 cycle standards, nearly all
manufacturers will need to optimize engine calibrations.  These
optimization will involve the reduction or elimination of
commanded enricJunent during heavy accelerations and WOT found on
the US06 cycle, improvements to air-fuel control strategies such
that air-fuel ratio is tightly controlled in order to reduce
transient emissions,  and increasing  catalyst NOx conversion
efficiency levels through tighter air-fuel control during high
speed operation.
     In addition to the above control strategy modifications,
some vehicles may also have to make hardware modifications or
additions.  Some vehicles may have to incorporate sequential-fire
port fuel injection systems..  Currently, there are several
different types of fuel injection systems being used in
production.  The current trend is a fairly even mixture of
different types of port fuel injection systems.  There are
synchronous-fire, synchronous double-fire, and sequential-fire
systems.  The primary difference is that synchronous-fire systems
fire multiple injectors simultaneously,  whereas sequential-fire
systems fire each injector separately.  This guarantees that each
cylinder is getting the proper amount of fuel at the right time.
There are also some vehicles that still use throttle-body fuel
                                83

-------
injection systems.   Vehicle  manufacturers have indicated to EPA
that throttle-body and synchronous-fire injection PFI systems are
being phased-out and the vast majority of vehicles will already
use sequential-fire injection PFI systems by the time this rule
is effective.
     EPA feels that modifications to control strategies
(calibrations) should be all that's  necessary for  the majority
of vehicles.  For a small percentage of vehicles, a combination
of control strategy modifications along with sequential-fire port
fuel injection, may be required to comply with the proposed
emission standards.

5.4.2.    Modifications Necessary to Offset Performance and
          Durability Impacts
     The Agency feels that  there should be no modifications
necessaryxto offset performance and durability impacts resulting
from complying. _with the proposed emission standards.   EPA
believes that any impacts on vehicle performance will be minimal
and will not require any hardware modifications.

Section 6.      US06 Test Procedures

6.1  Preconditioning
     The US06 driving schedule and-the associated  level of
emission control were developed  for a test in which the vehicle
in a hot,  stabilized condition.  Thus, sufficient driving is
required prior to the test to achieve the hot, stabilized
condition.   EPA believes running a vehicle on the 505  driving
cycle (bag 1 of the FTP)  or the new start cycle  (ST02) is
sufficient if the prior soak period'  is 2 hours or less.   If
the soak period in longer, a complete  LA4 driving cycle is
necessary.    To address potential adaptive memory concerns, EPA
believes it is appropriate to drive the vehicle on the US06
driving cycle, using  certification fuel, prior to the actual
certification emission test.
                                84

-------
6.2  Sequencing
                                         /
     The only sequencing requirement for the US06 cycle is the
preconditioning discussed  above.  The US06 cycle can run in
conjunction with the air conditioning test  requirement.

6.3 'Dynamometer Procedure
     The characteristics of the US06 driving cycle require the
cycle be tested on a single-roll, large diameter dynamometer, or
its equivalent.  EPA believes a large-capacity constant volume
sampler(CVS) will be necessary for properly testing vehicles on
the US06.   The FTP test program used a  large-capacity CVS   (700
cfm)  during the FTP test programs.   This was necessary due to
the "high-volume exhaust flow produced by the larger-displacement
vehicles as they were tested on the driving cycles which served
as  the basis for the US06.

6.3  Manual transmission shift points
     The US06 cycle and the FTP driving cycle represent very
    w
different types of driving behavior, thus, it is not reasonable
to,expect that manual transmission shift points will be identical
for the two cycles.  The higher rates of acceleration found on
the US06  requires a fairly aggressive shift pattern (shifting at
high RPM) in order to properly follow the driving schedule.
Further, it is expected that these shift points will be vehicle-
specific and it would be inappropriate to prescribe pre-
determined shift points.  EPA is proposing to allow the vehicle
manufacturer to determine appropriate shift points on a case by
case basis.  The shift points should be consistent with the
recommendations specified in the owner's manual with respect to
the maximum RPM for each gear.  In general, EPA will allow
manufacturers to specify upshift points, but downshifting will
not be- permitted unless the vehicle is unable to stay within the
driving tolerance on the speed trace.in the existing gear.

6.4  Adjustments for vehicle performance
                                85

-------
     As discussed in section 4.3,  adjustments to US06 for
differences in vehicle performance are necessary.  EPA proposes
to account for vehicle performance differences through -
adjustments to the dynamometer inertia load.

6.4.1  Performance Criteria
     The objective of the performance criteria is to classify
vehicles into three vehicle performance categories: high, medium,
and low.  The boundaries of these categories are established from
the in-use driving survey data, separately for automatic and
manual transmission vehicles. As appropriate, dynamometer
adjustments will  be made for the vehicles in the high and low
performance categories.  Two alternatives measures of vehicle
performance were considered by EPA: the ratio of vehicle weight
to peak horsepower, and the number of seconds for a vehicle to
accelerate from 0 to 60 mph.

6.4.1.1  W/P measure
     To date,   all  analysis of iri-use data has used the ratio of
vehicle weight to peak horsepower.(W/P) as a proxy of vehicle
performance.  This measure provides a good indication of vehicle
performance; however, it fails to take in account several
factors.  The W/P measure fails, to account for differences in
torque curves, aerodynamic design,   and the performance
difference between manual and automatic transmissions.  The
separate treatment of automatic and manual transmissions
addresses this last concern.

6.4.1.2  Measure based on 0 to 60 acceleration time
     A vehicle's 0 to 60 acceleration time is a direct measure of
performance and as such, it is a preferred over W/P, an indirect
measure.   The use of such a measure takes into account all the
vehicle variables which were unaccounted for by the W/P criteria,-.
however, there are a number of practical concerns with using 0 to
60 times.
                                86

-------
1. High and low performance cut-off points based on 0-60 times
   need to be established from the in-use driving survey data.
   This will require going back to the in-use driving survey
   database and trying to establish 0 to 60 times for each of
   the vehicles.   This will be a time consuming task and the
   consistency and accuracy of the historical data is likely to
   be a problem.
2. A new test procedure will need to be developed to obtain 0 to
   60 times for new vehicles; these results will need to
   correlate with the existing in-use data.
3. The 0 to 60 times do not take into account vehicle
   performance for speeds between 60 and 80 mph, an area of
   vehicle operation characterized on the US06.

While EPA feels the use .of  0 to 60 times is the best  vehicle
performance measure,  the lack of information and the technical
difficulties identified above preclude EPA from proposing it as
the preferred option.

6.4.2.  Performance categories
   The vehicle performance categories  are established  from the
in-use survey data.  Analyses in section.2.1 identified a
correlation between aggressive driving  and  vehicle performance;
however,  the analysis did not identify specific boundaries„for
low,  middle, and high performance vehicles.  To establish  these
categories, EPA went back to the in-use survey da,ta for Baltimore
and/Spokane to look for natural breakpoints in aggressive driving
based on the W/P measure.  For each* vehicle, two measures of
aggressive driving were used in the analysis:  the fraction of
time spent above power values of 200  (roughly equal to non-FTP
operation), and fraction of time above power values of 300(very
aggressive operation).    The high performance category was
defined by W/P,  starting at WP <20, and the category was expanded
by increments of  one;   the low performance category was fixed at
W/P >32.
                                87

-------
   For  manual  and  automatic  transmission vehicles  separately,
table 1-17 compares the two measures of aggressive driving for
the alternative high performance categories.  Among the automatic
transmission vehicles, the high 'performance vehicles  were driven
less aggressively than the mid-performance for all possible high
performance categories.  In contrast, for manual transmission
vehicles, there appears to be distinct differences between the
high and mid-performance vehicles,  and the data show a WP of 21
to be the upper value for the  high performance category. These
results are based on a very limited number of vehicles for the
high performance category, and as mentioned earlier, EPA plans to
add the data from the Atlanta in-use driving survey  to the
Spokane and Baltimore in-use survey database.
   The  in-use  survey  data  provides  a  good  sample of  mid-  and  low
performance vehicles.  Table 1-18 compares alternative thresholds
for low performance vehicles.  For manual transmission vehicles,
there is very little difference between middle and low
performance categories for the fraction of time above 200 until
you get to the W/P >34, thus EPA feels the most appropriate is
W/P >34, the corresponding value for automatic transmission
vehicles in W/P >31,  as the data shows a sharp drop-off in the
fraction of operation above 200 going from the W/P >30 to the W/P
>31 category.   Thus,  the low performance cutoff point for manual
transmission vehicles in W/P of 31, while automatic transmission
vehicles would have a cutoff point of 34 W/P.

 6.4.2  Adjustment Calculation
   For  testing purposes, a vehicle's  W/P will  be calculated   as
the ratio of  its estimated test weight (ETW) and its maximum
rated horsepower.   The adjustment calculation is best shown by
example.  If a manual transmission vehicle has a W/P value of 40,
then the adjustment would be 34/40 or 0.85.  The dynamometer
inertia load would be set to 85 percent of vehicle weight,
thereby implicitly bringing the 40 W/P vehicle down to a 34 W/P
vehicle.   EPA proposes an adjustment  cap of 50%,  although EPA
                                88

-------
expects the 50% cap to  impact  few,  if   any,  low performance
vehicles.
   EPA proposes to apply the load adjustment only during the high
load portions  of US06.   Figure  1-26  illustrates the segments of
US06 subject to  adjustment; these points  represent 43  seconds or
7 percent of the cycle.  This  contrasts with the earlier
application of load adjustment  carried  out in the vehicle
manufacturers  test program.  In that program the entire cycle was
modified, in large part  because dynamic adjustment was not a
viable option  at the time.  EPA feels dynamic adjustment .. is more
consistent with the objective  of modifying only the high load
portions of the cycle which are not  appropriate for some
vehicles.  Special programming  is required to have the
dynamometer provide a dynamic  adjustment of load,  and further
test work is planned to  identify any feasibility issues

6.4.3  Requirement for high performance vehicles
   The  limited  in-use  data  indicate high performance, manual
transmission  vehicles are driven more  aggressively than the rest
of  fleet and as such, the US06vmay  not appropriate.  Also,  the
testing of several high  performance, automatic transmission
vehicles, indicated that tine US 0-6'-= '.may not  be sufficiently
aggressive to force 'these. Vehicles." to '"wide open throttle (WOT)
operation.  EPA "believes '"it-vi&^rfeilessary to; 'ensure some WOT
                   '.,'•'•_• •'• !w"- '•/"•"""^•ipff:.^.. **••'•. '••':  ,.'•.••.'
emission control Jbrc;al^|:vehi€le§f::'^          high performance
                       ''
vehicles.  Thus, fo^-fen'ee'Ve&les, Manuacturers :would be
required to provide""' a .demonstration  of.' Stoichiometric A/F control
for wide open throttle events of two seconds  or  less.   EPA
proposes high performance cutoff -point- of  18  W/P, based on  a
conservative evaluation of the in-use data.

6.5. "Adjustments for Heavy, Light-duty Trucks
   For  the  US06,  HLDTs  will  be tested at curb' weight + 300  Ibs
instead of adjusted loaded vehicle weight (  (1/2  payload) .
HLDTs will continue to be tested at  1/2 payload  during the FTP
                                89

-------
driving schedule.  By bringing the load down to curb + 300 for
the US06,  HLDTs' would be tested in the same manner as LDVs.
This is more in line with the underlying assumptions used in
developing US06.  US06 was developed to represent a broad
spectrum of driving behavior in both cars and trucks.  As most of
the data used in developing this was from cars and smaller
trucks--probably lightly loaded-- it does not seem appropriate to
require heavy,  light-duty trucks to drive the same cycle with the
additional load.

6.6  Driving Schedule Tolerances
   The discussion  in  section  4.1.2.2 demonstrated  the emission
sensitivity of vehicles to how they are driven over a prescribed
driving schedule,  such as US06.   The emissions for some vehicles
are particularly sensitive to  the amount of throttle variation.
  Given that a principal objective of the US06 is to provide a
more representative driving cycle, it is important that the test
procedure ensures that the minor speed deviations characteristic
of  in-use operation, and the resulting throttle movement,  are
preserved.  The current FTP regulations (section 86.115-78 b)
require that "the driver should attempt to follow the target
schedule as closely as possible."  To accomplish this,  the
regulation specifies a speed tolerance band and associated
requirements for a valid test. • The regulations also suggest that
minimum throttle action sfiould be used to maintain the proper
speed-time relationship.  EPA believes that the speed tolerance
band does not ensure that the microtransient components of the
driving schedule are preserved, and the current•minimum throttle
action language in the regulations exacerbates the problem of
microtransient representation.
   A throttle-based  measure  which establishes  an  acceptable
range of  throttle variation would appear to be most desirable;
however, such  a measure is not practical.  The change in
throttle required to follow a driving schedule  will be
determined in part the performance of the vehicle.  If  two
                                90

-------
vehicles were  driven identically over a driving schedule, it is
to be expected that a low performance  vehicle would show greater
throttle variation than a high performance vehicle.  In
addition, differences in throttle design can result in physical
differences in the effect on the engine for  identical throttle
angles.
   EPA proposes  an additional  trace  tolerance  criteria  for all
FTP driving cycles using the speed-based measure, the sum of
change in specific power (DPWRSUM).   EPA's analysis in section
4.1.2.2  showed this measure to correlate with change in throttle
as well as emissions.  Unlike a throttle-based measure, DPWRSUM
is independent of the physical characteristics of the vehicle and
there exists a unique value corresponding to the nominal driving
schedule.  A test run which exactly matches the nominal driving
schedule, and thus matches the microtransient behavior of the
driving schedule, would have a sum of change in power equal to
the nominal DPWRS.UM.  Tests runs in which the DPWRSUM is less
than the nominal value indicate that the exact trace was not
maintained.  Test run where the DPWRSUM is greater than the
nominal suggest excessive changes.in power, and most likely,
excessive throttle action  EPA proposes that DPWRSUM be
calculated for each emission test.   A test with a DPWRSUM value
greater than the nominal DPWRSUM value would be invalid.  EPA's
analysis for establishing a.  lower DPWRSUM threshold is
incomplete and we will•seek input - on ;a  appropriate method for
doing so.                '•..-••'
                                91

-------
Figure 1-1 a.  FTP Driving Cycle "LA4"|
Q.
E
  90
  80
  70
  60
  50
0)
  40
  30
  20
  10
       0    100  200  300   400   500   600
700   800   900  1000  1100  1200  1300  1400  1500  1600
  Time (seconds)

-------
Figure I -1.  Frequency Distribution of high power operation for Automatic Transmission Vehicles
             0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
                                   Midpoint of Percentage of Time above 300
       High power operation = time above specific power of 300 mphA2/sec
1.0

-------
Figure 1-2. Frequency Distribution of high power operation for Manual Transmission Vehicles
             0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
                                    Midpoint of Percentage of Time above 300
       High power operation = time above specific power of 300 mphA2/sec

-------
 Figure 1-5.  Start Driving  Cycle |
   90
   80
   70
   60
•§.  so
T3
ID
0)
   30
   20
   10
        0    100   200   300   400   500   600
700   800  900   1000  1100  1200  1300   1400  1500   1600

  Time (seconds)

-------
 Figure 1-6a.  REP05 Driving Cycle]
.
Q.



T3
0)
0)
   80
   70
   60
   50
   40
   30
   20
   10
        0    100   200   300   400   500  600
700   800   900   1000  1100

  Time (seconds)
1200  1300  1400  1500  1600

-------
Figure 1-6b.  Remnant Driving Cycle]
  90
  80
  70
  60
  50
Q>
Q>
  30
  20
  10
       0   100   200   300  400  500   600
700   800   900  1000  1100   1200  1300  1400  1500  1600

  Time (seconds)

-------
Figure 1-7.  ARB02 Driving Cycle]
   90
   80
   70
   60
 ex
TJ
0)
0>
   50
   30
   20
    10
    0 -
         0    100  200  300   400   500   600   700  800  900   1000 1100  1200  1300  1400  1500  1600

                                            Time (seconds)

-------
Figure 1-8.  HL07 Driving Cycle|
   90
   80
   70
   60
   50  -
Q.



TJ
O>
Q)

W  40
   30  -
   20
   10  -
f\l
        0   100   200   300   400   500   600   700   800   900   1000   1100  1200  1300  1400  1500   1600

                                             Time (seconds)

-------
Figure 1-9.  Unweighted NMHC and NOx emissions
 0.3
               NMHC
NOx
Figure 1-10. Unweighted CO emissions
 6 -
                                                     ! • LA4
                                                     I
                                                     ! E3 Start
                                                     I
                                                     j • Remnant

                                                     I ED REP05

-------
       '
       Mean Vehicle WGT/HP vs POWER - MAN'UAL TRANS
              Upper and Lower Deciles define aggressiveness classes
CTJ
CD
                                                       400
LEAST AGGRESSIVE


NORMAL
             !*'

MOST AGGRESSIVE

-------
«^i. ' r
      Mean Vehicle WGT/HP vs POWER - AUTOMATIC TRANS
               Upper and Lower Deciles define aggressiveness classes
                                * * *******
                                    **,
                                      ,11
                             • LEAST AGGRESSIVt

                             A NORMAL

                             I | MOST AGGRESSIVE
                                                         II.
   20
     0
100
 200
Power
300
400

-------
                                FIGURE
   8
o
O
O
   0
135
               Engine-Out CO with and without Commanded Enrichment
                                      Olds 98
                  137
CE = Commanded Enrichment |
                                      	G--	--0

                                        A   -.— _^.
139             141
    Time (sec)
-	«>--

  143
                    CO w/CE   ® CO wo/CE  A A/F w/CE   \ \ A/F wo/CE
                                                                           16
                                                                           12


                                                                           10


                                                                           8


                                                                           6


                                                                           4
                                                 .2
                                                 CO
145

-------
Figure 1-12 Comparison of Emissions for vehicles in Production and Stoichiometric configura
                Tailpipe HC Emissions:REP05
                      p=production s=stoichiometric
  •s.

  n
   1 -
Illllll
      Illlll
      201p 201s 302p 302s 303p 303s 304p 304s 305p 305s 306p 306s 312p 312s 313p 313s 314p 314s LDV LDV
                          Vehicle #
                 Tailpipe CO Emissions:REP05
                      p=production s=stoichiometric
    40
    30 -
    20
    10 -
            111
1
Hi
111
       201 p 201s 302p 302s 303p 303$ 304p 304s 305p 305s 306p 306s 312p 312s 313p 313s 314p 314s LDV LDV
                           Vehicle #
                Tailpipe NOx Emissions:REP05
                       p=production s=stoichiometric
    6 -
  JS)

  E
  I 4

  2
  O)
    2 -
                       mil
       20lp 201s 302p 302s 303p 303s 304p 304s 305p 305s 306p 306s 312p 312s 313p 313s 314p 314s LDV LDV
                           Vehicle #

-------
Engine Out NOx Emissions
       Bag 2 of FTP vs REP05
201    302     303
       304    305    306
            Vehicle #

      • Bag 2 FTP m REP05  ~|
313    314

-------
Figure 1-14.  US06 Driving Cycle]
   90
   80
   60
   50
   40
W
   30
   20
   10
        0    100   200   300   400   500  600   700  800  900  1000  1100  1200  1300  1400  1500  1600
                                           Time (seconds)

-------
Figure 1-15a.  Fuel Economy \
   40
   35
c
:i  30
(0
O)
0>
Q.
£ 25
o
§
in
0) 20
    15
    10  ;  i     i     i     ii     i    i '    i     i     i     i     r    i     i     i     i
        ^01  202  203  302   303  304  305   306  312   313  314  401  601  801   802  901
                                           Vehicle
                                                                                          ,,, US06-Pr
                                                                                          • USOG-St.
                                                                                          i i FTP-Bag 2
                                                                                          • FTP-full

-------
Figure 1-15.  Engine-out HC emissions
   10
E
In
O
    8
    0
                                                                                    US06-Pr
                                                                                    usoe-st.
                                                                                    FTP-Bag 2
                                                                                    FTP-fuIl
 I  	|  	|	  I. .  ..I	I. ..__!._. ..J ..   I    ..l..._J	I.-..  I     I     I     I
201  202  203  302  303  304  305  306   312   313   314  401  601   801   802  901
                                  Vehicle

-------
Figure 1-16. Tailpipe HC emissions  jj
   0.5
   0.4
   0-3
 o
 <
O
   0.2
   0.1
     0
 I   	I  ..  .I.   . .I   _  I   ..  I.   .I     I     I   ...I 	J	L	1 ..I     I     I
201   202  203  302  303   304   305   306   312   313   314  401  601  801  802  901.
                                   Vehicle
                                                                                       US06-Pr
                                                                                       US06-St.
                                                                                     I  FTP-Bag2
                                                                                       FTP-full

-------
Figure 1-17. Tailpipe CO emissions jj
   30
   20
1
1»
O
O
   10
    0
US06-Pr
US06-St.
FTP-Bag 2
FTP-full
        201  202  203  302  303  304  305  306  312  313  314  401  601   801   802  901
                                          Vehicle

-------
Figure 1-18. Engine-out CO emissions
   40
   30
   20
O
O
   10
    0
 .1     .1 .    I ....  1... .  I	 I.	I .....  I    I     I..	_i	I	I     I     I    I
201   202   203  302  303  304  305  306  312  313   314  401  601   801  802  901
                                  Vehicle
                                                                                   US06-Pr
                                                                                   US06-St
                                                                                   FTP-Bag 2
                                                                                   FTP-full

-------
Figure 1-19. Engine-out NOx emissions
   8
§4
0)
X
O
   0
._..]_	...I	L	I		I	 ...I....	I	 I  ...  I.    .U__._..l	_.  J-	I.... I.    I     I
 201  202  203   302  303   304   305   306   312  313  314  401   601   801  802  901
                                     Vehicle
                                                                                       O US06-Pr
                                                                                       • US06-S!.
                                                                                       l,l FTP-Bag 2
                                                                                       • FTP-full

-------
Figure 1-20.  Tailpipe NOx emissions |
   2.5
     2  -
    1.5
 /5
x
i
    0.5
     0
          [3
                           (;/i         iyi-
._.!...__	j	J	.....i	L	I.... .....i	i.  ..  i...  ....i.._.	i—	...i	 'r'.   11,1    i     'i1
201   202  203  302  303   304  305  306  312   313  314  401  601   801   802  901
                                      Vehicle
                                                                                            US06-Pr
                                                                                            US06-St.
                                                                                            FTP-Bag 2
                                                                                            FTP-full

-------
Figure 1-21.  Theoretical US06 NOx efficiency to match FTP NOx levels
                  compared to Bag 2 FTP efficiency
   110
   100
0>
'o
£
LLJ
O
x
o
    90
    80
    70
          I 	I ..   I	I. .. . .1. .. .1    I .    I    I    I	I	I     I    I     I    I

        201  202  203  302  303  304  305  306  312  313  314  401  601  801  802  901

                                        Vehicle
o US06-Pr

• USOG-St.

I j FTP-Bag 2

-------
         Figure 1-22.   Driving Trace for Hill3(Parti)  of the  US06, Cadillac  Seville,
         Test 305P2M3U
70 n	
                                                observed

                                                lower  limit

                                                nominal
                                         —- — upper  limit

                                           •   throttle  > 90
   133
138
143           148
       Time in seconds
153
158
163

-------
           Figure 1-23. Driving Trace for Hi113(Parti) of the OS06, Cadillac Seville,
           Test 305S2M3D
  70 nr
  60
  50 - -
  40 --
•o
9
e30
  20
  10 --
                              "" lower limit

                                nominal

                                upper limit

                                throttle > 90
     133
138
143           148
       Time in seconds
153
158
163

-------
   Figure 1-23b Stoichiometric US06 Cycle Catalyst Temperature I
      200  r	
      150
   Q.



   -a
   
-------
  800
  700
  600
TJ
C
o
£  500
o  400

-------
   800
   700
1  600
o
£  500
o  400
0>
g-  300
   200
   100
     0
       i
       I
            538
                     Catalyst Temperature Profiles
                              Ford Escort & GM Saturn
                    593
649      704      760
     Temperalure C
815
871
       Escort Production i!J Saturn Production  • Escort Stoich
                                                        Saturn Stoich
I   ARB02 Cycle
Average of two tests

-------
Figure 1 -26.  US06 dynamometer adjustment points
     90 -i-

     80 -•

     70 -•

     60 -•
JS
Q.
£ 50 -
c
"Q
O* Af\ m
1
O.

30 "

20 -

10 -

•







•

•






I i
i
/
















r
T
i

I \
f 1
i i
f ! ,

f |
i
1
, V,
/
p 	 • 	 •" 	 1

p 	 Nominal Speed 1
* . • ' \
* Dynamometer Adjustment 1
1 * a * ft
i AM
1 n II i
\ 'I H1
1 " i
f ii
II
y '
i i — i 	 1 — 	 i, 	 .„ 	 1 	 — i 	 U 	 — 	 i
?,
T |

T I
J 1
" i
I
i' 1
II
|

,|











-J
50     100    150    200    250     300
                                 Seconds
                                                         350
400
450
500
550
600

-------
Table 1-2
Measures of Aggressive Driving by transmission type

Fraction of time above 200
Mean
Maximum
Standard deviation
'Transmission Type:
jAutomatic
i
0.208
i 1.819
0.352
1
Fraction of time above 300
Mean
Maximum
Standard deviation
i
! 0.014
i 0.373
; 0.049
i
Manual

0.354 ;
3.215:
0.58!


0.03 i
0.765 i
0.1221

Weight to Power Ratio i I i
Mean
Minimum
! 29.37 i 31.311
i 16.67,
19.33!
Maximum
Standard deviation
! 38.46
j 4.15
38.96
4.79

-------
Table 1-3 Description of vehicles in FTP Test Program

Vehicle                  Displacement     Weight/Power
                         (Liters)          (ETW/Net HP)
1992 Ford Crown Victoria
1991 Honda Accord
1992 Dodge Dakota
1991 GMC Sonoma
1993 Dodge Intrepid
1993 Mercedes 400SEL
1992 VW Golf
1993 Saturn SL
4.6
2.2
5.2
2.8
3.3
4.2
1.8
1.9
21.05
27.00
17.39
27.00
24.21
16.67
27.50
32.35
Table 1-4 In-use weighting factors

                        Fraction of total distance   Fraction of total time
                              (grams/mile)           (grams/minute)
FTP
  Bag 1                           0.48                   0.368
       2                           0.52                   0.632
In-use
  Start                           0.24                   0.296
  Remnant                       0.48                   0.581
  Non-FTP high speed            0.264                  0.105
  Non-FTP high accel             0.016                 , OX) 18

-------
Table 1-5. Comparison of weighted FTP and in-use emissions (hot stabilized, grams/mile).

Vehicle
Ford Crown Victoria
Honda Accord
Dodge Dakota
CMC Sonoma
Dodge Intrepid
Mercedes 400SEL
VWGolf
Saturn SL
Average
NMHC
FTP
0.03
0.03
0.03
0.04
0.08
0.02
0.05
0.04
0.04
In-use
0.07
0.05
0.08
0.15
0.137
0.05
0.08
0.07
0.08
CO
FTP
2.9
1.7
2.2
2.6
0.7
0.5
1.1
0.9
1.6
In-use
4.9
3.6
5.1
9.7
2.9
1.7
2.9
4.2
4.4
NOx
FTP
0.42
0.20
0.18
0.22
0.33
0.07
0.03
0.07
0.19
In-use:
0.57
0.17
0.24
0.44
0.54
0.07
0.08
0.07
0.27
Table 1-6 In-use driving modes contribution to emission increase (percent of total increase).
Driving mode
Start
Remnant
REP05
Total
NMHC
30%
34%
36%
100%
CO
17%
25%
58%
100%
NOx
23%
46%
31%
100%

-------
Table 1-7 Comparison of EPA and Manufacturer Test Results
Vehicle type
and pollutant
LDV~
NMHC
CO
NOx
LLDT
NMHC
CO
NOx
LDV + LLDT average
NMHC
CO
NOx
HLDT
NMHC
CO
NOx
I FTP, hot stabilized
I EPA I MANUF

0.04
1.31
0.19

0.03
2.39
0.20

0.04
1.58
0.19





0.05
1.11
0.26

0.13
2.53
0.40

0.08
1.54
0.30

0.28
5.84
1.88
REPO5
EPA I MANUF

0.09
6.33
0.30

0.12
10.19
0.24

0.10
7.29
0.28





0.10
5.85
0.47

0.15
9.18
0.72

0.12
6.86
0.55

0.39
22.24
2.66
ARB02
EPA [ MANUF

0.16
12.19
0.40

0.20
15.39
0.26
1
0.17
13.10
0.36





0.17
11.43
0.48

0.27
15.63
0.80

0.20
12.71
0.58

0.56
33.26
2.80
REP05 minus FTP
EPA [ MANUF

0.05
5.02
0.11

0.08
7.79
0.04

0.06
5.71
0.09





0.05
4.74
0.21

0.02
6.65
0.33

0.04
5.32
0.25

0.11
16.40
0.78


-------
Table 1-9 Emissions and throttle measures for the Mercedes on REM01 cycle
Start
Remnant
fest#
3629
3626
2628

3629
3626
2628
Driver Id
C
B
A

C
B
L A
HC (g/mi)
0.056
0.081
0.302

0.029
0.066
0.095
CO (g/mi)
0.82
1.85
2.16

0.64
2.10
3.90
NOx (g/mi)
0.028
0.022
0.023

0.057
0.025
0.061
Sum of
DTP
490
739
740

2357
, 3208
*~ 3329
DTP >3 %
0.0%
13.2%
13.6%

0.0%
10.2%
12.4%
DTP > 5 %
0.0%
7.0%
8.5%

0.0%
7.1%
7.8%
DTP> 10%
0.0%
4.0%
3.3%

0.0%
3.7%
3.3%

-------
Table 1-17 Alternative W/P Thresholds for High Performance Category


Manual
High .
Middle
Low

Automatic
High
Middle
Low

'Percent above power=200 [| Percent i
W/P <20

2.162
0.336
0.310


0.160
0.257
0.068

W/P <21

2.109
0.276
0.310


0.160
0.257
0.068

W/P <22 I W/P <23 I W/P <20

1.636
0.262
0.310


0.188
0.259
0.068


1.377
0.249
0.310


0.205
0.259
0.068


0.765
0.016
0.020


0.008
0.019
0.001

W/P <21

0.469
0.010
0.020


I 0.008
0.019
0.001

above power=300
W/P<22 | W/P<23
-
0.313
0.010
0020


0.016
0.019
0.001


0.241
6.010
0020


0.012
0.019
0.001
.
Table 1-18 Alternative W/P Thresholds for Low Performance Category

Category and transmission type
Manual
High
Middle
Low

Automatic
High
Middle
Low
'Percent above power=200
W/P>30 I W/P>31

2.109
0.403
0.248


0.160
0.263
0.149

2.109
0.314
0.278


0.160
0.268
0.082
W/P>32 | W/P>33

2.109
0.276
0.310


0.160
0.257
0.068

2.109
0.288
0.301


0.160
0.240
0.065

W/P>34

2.109
0.344
0.171


0.160
0.241
0.023
Percent above power
W/P >30

0.4686
0.0161
0.0147


0.0077
0.0188
0.0095
W/P >31

0.4686
0.0122
0.0172


0.0077
0.0209
0.0006
W/P >32

0.4686
0.0101
0.0201


0.0077
0.0190
0.0008
=300
W/P >33

0.4686
0.0086
0.0257


0.0077
0.0173
0.0009

W/P>34

I 0.4686
0.0209
0.0012

00077
0.0169
0.0000

-------