&EPA
United States
Environmental Protection
Agency
Office of Transportation EPA420-D-04-003
and Air Quality June 2004
Draft Technical Support
Document: In-Use Testing for
Heavy-Duty Diesel Engines and
Vehicles
-------
EPA420-D-04-003
June 2004
Draft Technical Support Document:
In-Use Testing for Heavy-Duty Diesel
Engines and Vehicles
Assessment and Standards Division
Office of Transportation and Air Quality
U.S. Environmental Protection Agency
NOTICE
This Technical Report does not necessarily represent final EPA decisions or positions.
It is intended to present technical analysis of issues using data that are currently available.
The purpose in the release of such reports is to facilitate an exchange of
technical information and to inform the public of technical developments.
-------
Table of Contents
CHAPTER 1: Economic Assessment 1
1.1 Overview 1
1.2 Cost Components 2
1.2.1 Fixed Costs 2
1.2.2 Variable Costs 3
1.2.2.1 General Description of Testing Scenarios 4
1.2.2.2 Variable Cost Per Test 7
1.2.2.2.1 Direct Labor 7
1.2.2.2.2 Labor Overhead 8
1.2.2.2.3 Other Direct Costs 8
1.2.2.2.4 Repeat Tests 8
1.2.2.2.4 General and Administrative Overhead 8
1.2.2.2.5 Summary of Variable Cost per Test 8
1.2.2.3 Variable Cost Per Engine Family 8
1.2.2.3.1 Vehicle Incentives 9
1.2.2.3.2 Direct Labor 9
1.2.2.3.3 Travel 10
1.2.2.3.4 Labor Overhead 11
1.2.2.3.5 General and Administrative Overhead 11
1.2.2.3.6 Summary of Variable Cost per Engine Family 11
1.3 Costs of the Proposed Program 11
-------
1.3.1 Variable Costs by Level of Testing Intensity 12
1.3.2 Total Annual Costs by Level of Testing Intensity 12
1.3.3. Total Annual Cost Point Estimate 12
1.3.4. Total Costs Over 5 and 30 Years 13
CHAPTER 2: On-Vehicle Portable Emissions Measurement Technology Review 28
2.1 Overview 28
2.2 Measurement Technologies 28
CHAPTER 3: New In-Use Testing Instrument Measurement Allowance 36
3.1 Revi ew of NTE zone 36
3.2 Review of existing NTE allowances 37
3.3 Review of other proposed NTE allowances 38
3.4 Discussion of the two (2) proposed NTE measurement allowances 38
APPENDIX A: Examples of Determining the Number of Engine Families to be Tested 46
11
-------
CHAPTER 1
Economic Assessment
This chapter contains our economic analysis of the potential costs associated with the
proposal to implement a manufacturer-run, in-use NTE testing program for heavy-duty diesel
engines and vehicles. The eventual cost of the program is dependent on several key variables.
One of these is the number of vehicles tested under the Phase 1 and 2 testing schemes. This, of
course, depends on how many vehicles pass, or fail, the vehicle pass criteria at various points
under the tiered testing procedures. Also important is just how each manufacturer will chose to
design and conduct the test program, how many portable emission measurement systems
(PEMS) will be purchased, and the availability of test vehicles. Obviously, it is difficult to
project these variables for an all new program. However, based on our experience with in-use
emissions testing, including the development and use of a portable emission measurement device
for compliance testing, we have identified a set of reasonable testing scenarios that allow us to
estimate the potential costs associated with the proposed program.
This chapter is divided into several parts. Section 1.1 contains a brief outline of the
methodology used to estimate the associated costs is presented. Section 1.2 develops the fixed
and variable cost components associated with the program. Section 1.3 summarizes the
component costs into estimates of the overall cost of the program. All costs are estimated in
2003 dollars.
1.1 Overview
All costs are divided into either fixed or variable cost components. Fixed costs are
associated with the direct expense of purchasing the requisite portable emission measurement
system (PEMS) units. Variable costs depend primarily on the number of families and vehicles
tested. They include the direct costs for vehicle recruiting, labor for on-site testing, instrument
calibration and maintenance, travel, data analysis, and reporting expenses. Variable costs also
include indirect costs associated with overhead and general and administrative (G&A) expenses.
To explore the range of possible costs, we developed two testing scenarios that differ in the
relative availability of test vehicles (i.e., how difficult it is to access, instrument, and test
vehicles at a job site), and the type and amount of travel required to conduct the test campaign
(e.g., overnight versus local travel). We also assessed a range of costs associated with different
testing intensities under Phase 1 or Phase 2 of the program (i.e., the amount or number of vehicle
tests that might occur under the two phases of testing). Finally, we combined this information to
show a range of possible costs and a single point estimate by assuming a specific mix of the
above testing variables. The results are presented for all heavy-duty engine manufacturers as
annualized costs, total costs for the first five years of the program, and costs over 30 years.
-------
1.2 Cost Components
1.2.1 Fixed Costs
Fixed costs for this program are primarily associated with buying PEMS units. Some of the
fixed cost components have significant uncertainties associated with them. Portable
measurement devices are already commercially available that can measure all the gaseous
pollutants required by the proposed program. These systems cost approximately $70,000 per
unit. Based on our experience and investment in developing portable particulate matter (PM)
measurement technology, we estimate that systems capable of measuring both gaseous and PM
emissions will cost an additional $30,000 per unit.1 Therefore, the capital cost for portable
measurement devices that measure both gaseous and PM emissions is approximately $100,000
per unit.
As described in Chapter 2 of this document, we assume that engine manufacturers will
initially purchase PEMS units with the capability to measure gaseous pollutants in time to
coincide with the initiation of the pilot program at the start of the 2005 calendar year. Add-on,
modular devices that measure PM emissions will be purchased later in 2005 as this technology
becomes available. Regardless of the purchasing strategy, we assume all equipment purchases
occur at the beginning of 2005 in order to simplify the analysis.
We estimate that portable emission measurement devices have a product life cycle of five
years for the purposes of the proposed program. After that time they are assumed to be replaced
with brand new equipment. Also, we assume there is no salvage value for units that may remain
in service for other less rigorous or less important duties after five years, although this could
occur in some instances. Alternatively, some manufacturers may chose to replace or rebuild
component parts of a PEMS unit rather than replace the entire unit after five years. To the extent
this may occur, we assume such a maintenance strategy will cost approximately the same as the
replacement strategy. The annualized cost of a single PEMS unit can now be calculated by
using the above values and assuming a typical capital recovery rate of seven percent per annum.
The result is an annualized cost of $24,390 per unit.
The total annualized fixed cost for the program depends on how many PEMS units each
manufacturer will purchase, the fraction of time the equipment is used for the in-use testing
program, and the number of manufacturers subject to the requirements. Each of these cost
components is addressed separately below.
We assume that each manufacturer will purchase two units. We chose this number to
illustrate the average equipment cost of the program recognizing that two units is adequate to
perform more than the needed amount of tests for even the largest manufacturer if its program is
1 Chapter 2 contains additional information on the status and development of portable
emission measurement systems.
-------
designed so that testing can be conducted in an orderly, efficient manner. We recognize that
manufacturers with a limited number of engine families may need only one unit. Conversely,
manufacturers with a large number of families may prefer additional units. However, we think
that two units is adequate for the average manufacturer.
We also expect that manufacturers will likely want a spare unit in the field beyond the two
PEMS that may be in service, to prevent disrupting the test program if a serious, unserviceable
problem arises with one of the primary systems. We expect that the spare unit will be taken from
other PEMS units the manufacturer has purchased for separate engine or emission control
technology development work. Rather than trying to estimate the cost of such a practice, we are
making an offsetting assumption that the PEMS units purchased for the proposed in-use testing
program are also sometimes used for other purposes, e.g., engine and emission control
technology development. We anticipate that these units would otherwise often be idle, because
the intensity of the test requirements associated with the in-use testing program will result in
significant periods of downtime for these units.
Our final assumption in estimating the total annual fixed cost of the proposed program is that
14 engine manufacturers will participate in the program. This is the number of companies that
certified heavy-duty diesel engines in the 2003 model year. The manufacturers are shown
below.
Caterpillar, Inc.
Cummins Engine Company, Inc.
DaimlerChrysler AG
Detroit Diesel Corporation
General Engine Products
General Motos Corporation
Hino Motors, Ltd.
International Truck and Engine Corporation
Isuzu Motors Limited
Mack
Mitsubishi Motors Corporation
Nissan Diesel Motor Co., Ltd.
Scania CV AB
Volvo Powertrain Corporation
Using the above information, the total annualized fixed cost of the proposed program
$682,890, as shown in Table 1-1.
1.2.2 Variable Costs
Variable costs are grouped into two broad categories: cost per vehicle test and cost per
family. This approach allows us to more easily account for tests that must be repeated on the
same vehicle in order to obtain a valid test result. Repeat testing can occur in the laboratory due
-------
to equipment or vehicle malfunctions, and operator error. We expect that similar problems may
occur in field testing, and assume that these issues can generally be remedied at the testing
location. Further, a vehicle may be tested a second day under the terms of the proposed program
if less than three hours of non-idle operation are recorded during the first "shift day" of testing.
Obviously, multiple tests on the same vehicle do not directly affect other costs associated with
testing an engine family, e.g., vehicle recruiting. Therefore, these costs are estimated separately
in our analysis.
Also, as noted earlier, many of the costs of the proposed program vary by the relative
availability of test vehicles (i.e., how difficult it is to access, instrument, and test vehicles at a job
site), and the type and amount of travel required to conduct the test campaign (e.g., overnight
versus local travel). In order to reasonably bracket these cost elements, we constructed two
testing scenarios that differ in the above characteristics. These scenarios are based on our direct
experience in conducting in-use testing of heavy-duty trucks with portable emission
measurement systems under the 1998 consent decrees, our continued development of portable
measurement systems, and a recently awarded EPA contract to conduct a large scale, in-use
testing pilot program in Kansas City, Missouri for passenger cars (USEPA 2003). The two
testing scenarios are described in the following section by first identifying some of the key
elements shared by both scenarios and then presenting the specifics of each scenario separately.
1.2.2.1 General Description of Testing Scenarios
Our testing scenarios share a number of key assumptions which we believe provide a
reasonable representation of how manufacturers are likely to design and conduct testing under
the proposed program. Alternatively, if an engine manufacturer decides to contract for testing
services, we expect the service provider to similarly design and conduct the testing campaign.
One of our most basic methodological assumptions is that field testing will usually consist of
a multi-day campaign where a minimum of five vehicles are tested. This number was chosen for
several reasons. First, it captures the type of back-to-back vehicle testing likely to be employed
in order to facilitate efficient testing. Second, it represents the minimum number of test vehicles
for Phase 1 testing under the proposed program. Third, and finally, it simplifies the analysis
when evaluating the potential costs of higher testing intensities associated with the maximum
number of vehicles that may be tested under Phase 1 (10 total vehicles) and Phase 2 (20 total
vehicles). These later testing levels are simple multiples the Phase 1 minimum testing scenario,
i.e., two times or four times, respectively. Other key assumptions are described below.
- Vehicle recruiting and pre-screening of prospective test vehicles will be done by
telephone or other means prior to the test site visit.
- All portable measurement systems will be inspected and pre-calibrated at the
manufacturer's facility prior to deployment in the field.
- Field testing will be conducted by two people. One is an engineer and the other a
qualified technician. Both are capable of installing and operating the portable
measurement systems, screening vehicles for OBD trouble codes and MIL
indications, performing maintenance on the portable systems, etc. The technician is
-------
also capable of performing all required inspections of the vehicle's mechanical,
electrical, and emission control systems; as well as performing allowable
maintenance and the setting of adjustable engine parameters as required.
- The test engineer and technician will coordinate their activities to optimize their
productivity. For example, the engineer may acquire and enter a vehicle's history
and vital statistics into an electronic database concurrently while the technician
performs vehicle inspection and allowable maintenance.
- Test vehicles for an engine family are obtained from two independent sources, as
required by the proposed regulation. It is assumed that the sources are located
relatively close to each other to minimize travel distances between test sites.
- The test sites will be in relatively close proximity to a manufacturer's technical
center, test center, maintenance facility, or other similar location to minimize
personnel travel and field logistics.
- Test vehicles will depart and return to the same location the same day. Further, a
"shift day" is approximately eight hours in duration.
- Field personnel have access to the test vehicles and a service location before and after
the shift day as needed to install and remove the portable measurement devices.
Special arrangements with the vehicle owner/operator may be necessary.
- Two portable emission measurement systems will be deployed simultaneously during
a test site visit when possible, i.e., two vehicles will be tested at the same time.
- All necessary tools, spare parts, testing supplies, office supplies, etc. will be taken to
or otherwise supplied at the site of testing to avoid unnecessary delays.
One of the key simplifying assumptions from above is that the test vehicles will be away
from their home base for a single day. It is also possible that test vehicles may be operated on
multi-day driving routes, i.e., long-haul operation. We do not think this will be a standard
practice, but may happen from time to time. To the extent it does occur, the cost of such tests
may be somewhat higher than reflected in this analysis. For example, the field logistics and site
visits may be more complex. Also, even though a PEMS unit is capable of operating unattended
for extended times, it may require a suitably sized gas bottle for the Flame lonization Detector
(FID), which would increase the cost of the test. We believe it is unlikely that the PEMS unit
would need to be calibrated or otherwise maintained by a technician during a multi-day test,
although this has not been demonstrated in actual field testing. Finally, in the event that fuel or
self-contained standby battery power is expended, the unit will simply shut itself off. We
assume that upon returning home, the system would checked, calibrated, and the emissions data
downloaded as described next.
Each of the test scenarios also share some of the same core on-site work activities. These
categories are described below in their approximate order of occurrence.
- Vehicle History/Documentation. Obtaining all relevant information not available
prior to the field visit or verifying the accuracy of information previously obtained.
- Vehicle Set-to-Specification. Inspecting the vehicle, performing allowable
maintenance, and setting all adjustable engine control parameters.
-------
- PEMS Installation. Install the portable measurement system onboard the test vehicle.
Warm the instrument to operating temperature, verify proper operation, perform final
calibrations and span instruments, etc.
- PEMS Data Acquisition. Download the measurement data set; perform on-site data
verification and initial quality assurance; and record and store all other relevant test
information.
- PEMS Removal. Verify proper operation, perform post-test calibration, and remove
system from the test vehicle.
- Miscellaneous Time. Non-routine labor for repairing or replacing parts of the
portable emission measurement systems or test vehicles, obtaining supplies not at the
test site, etc.
Now that the overall components of the testing scenarios have been identified, the specific
scenarios will be presented.
Scenario 1 -Easy Vehicle Access and Local Travel
In this scenario the owner/operator of the test vehicles make them readily available on an "as
needed" basis so that the testing personnel can execute the test program in an expeditious
manner. This means, for example, there will be little or no "idle" time while at the job site
waiting for a vehicle to become available for either installing or removing the portable
measurement system. Also, the test sites are located close enough to the manufacturer's facility,
or employees base of operation, so they can commute to and from the job each day. This avoids
overnight travel expenses. This scenario reflects how most of EPA's in-use testing is conducted
under the consent decrees.
Scenario 2 - Limited Vehicle Access and Overnight Travel
This scenario reflects a less time efficient case than Scenario 1. Test vehicles are not readily
available and the testing technician and engineer must sometimes work around a test vehicle's
normal daily work shift. This means that the work day for the testing personnel includes the
time the vehicle is being driven over its normal work route. In these instances, we made a worst
case assumption that the testing personnel remain "idle" on the job site, with this time charged to
the in-use testing program. In reality, we expect that the engineer and technician may perform
work that could be billed to other assignments for all or part of this time. Also, the test sites are
located far enough away from the manufacturer's facility, or employees home base, that a single
round trip to and from the job site and overnight travel is required. However, the sites are still
close enough to one another that travel between the two locations is not restrictive or prohibitive.
-------
1.2.2.2 Variable Cost Per Test
As described above, the cost to perform an individual vehicle test is based on a Phase 1
testing scheme where a minimum of five vehicles must be tested. Test costs consist of direct
labor, labor overhead, other direct costs, and general and administrative overhead. Each of these
cost components is described below.
1.2.2.2.1 Direct Labor
The cost of direct labor for each scenario is estimated by applying an hourly compensation
rate by labor category to the various activities associated with the field testing campaign. Tables
1-2 and 1-3 display the work flow for Scenario 1 and 2, respectively, broken down by activity,
labor type, and number of hours spent preforming each activity. These depictions reflect the
assumptions, work activities, and optimization of the work flow as described in Section 1.2.2.1.
The time spent in the various work tasks are estimated based on EPA's direct experience in
conducting in-use testing with portable emission measurement devices and on the recently
awarded Kansas City, Missouri test program.
The test campaign for Scenario 1 is completed in four days. It is assumed that the testing
personnel return to the manufacturer's facility at the end of Day 4 for to work on other
assignments. The total direct labor is 27 hours for the engineer and 27 hours for the technician.
Scenario 2 is completed in five days. The lack of vehicle flexibility leads to long work days in
this scenario. The total direct labor is 55 and 56 hours for the engineer and technician,
respectively.
We developed an hourly estimate of employee compensation from information published by
the Bureau of Labor Statistics, Office of Compensation and Working Conditions, Employer
Costs for Employee Compensation (BLS 2003). Table 1-4 shows our estimate of $31 and $28
per hour for an engineer and technician, respectively. These hourly compensation rates are "total
compensation," and include wages and salaries, paid leave, supplemental pay, and insurance. By
comparison, these labor rates compare well with the contract labor costs we incur in conducting
our in-use testing under the consent decrees. Finally, we assume that labor exceeding 40 hours
per week is paid as overtime, i.e., 1.5 times the normal hourly rate.
For convenience, Table 1-4 also shows an hourly compensation rate for a manager using the
same literature source as described above. This labor category will be used in Section 1.2.2.3,
where variable costs are discussed.
The resulting direct labor cost per test can now be calculated based on the above labor hours
and hourly compensation rates. As shown in Table 1-5, the resulting per test cost is $319 and
$757 for Scenario 1 and 2, respectively.
-------
1.2.2.2.2 Labor Overhead
We assume that all direct labor is burdened at 100 percent of the total compensation rate.
For simplicity, overtime pay is also burdened at this same overhead rate.
1.2.2.2.3 Other Direct Costs
A number of other costs not related to labor that are "consumed" during in-use testing
include office supplies such as office supplies, DVDs, calibration gases, and fuel for the flame
ionization detector (FID). Based on our experience with using portable measurement systems,
we estimate that calibration gases will cost about $75 per test and all other supplies will cost
about $25 per test. Therefore, we estimate that a total of $100 per test for other direct costs.
1.2.2.2.4 Repeat Tests
Some in-use tests will be voided due to operator error and test equipment malfunctions.
Other tests will be repeated if less than three hours of non-idle vehicle operation are recorded
during the first day of testing. At our National Vehicle Fuel and Emissions Laboratory in Ann
Arbor, Michigan, we experienced a test void rate for laboratory-based, non-research testing of
approximately four percent over the last two years. We expect a somewhat higher void rate for
field testing. Also, as noted, some tests will be repeated do to the three hour non-idle
requirement. Overall, we assume a combined repeat test rate of 10 percent for this analysis.
1.2.2.2.5 General and Administrative Overhead
Certain costs are incurred for common or joint objectives and therefore cannot be
identified specifically with a particular project or activity. We assume these general and
administrative costs to be 6.5 percent of all other costs.
1.2.2.2.6 Summary of Variable Cost per Test
Table 1-6 summaries the various direct cost elements described above. As shown, the
resulting total variable costs per test are $865 and $1,891 for Scenario 1 and 2, respectively.
1.2.2.3 Variable Cost Per Engine Family
As with the previous section, the cost per engine family is based on a Phase 1 testing scheme
where a minimum of five vehicles must be tested. This overall cost is composed of a number of
individual expenses such as paying the test vehicle's owner/operator an incentive, recruitment,
travel, instrument pre-calibration, data analysis, and reporting. Each of the cost elements are
described below.
-------
1.2.2.3.1 Vehicle Incentives
We generally offer a vehicle's owner an incentive in the form of a government bond and free
vehicle repairs as part of our in-use test programs. Sometimes the owner cooperates without
such an incentive, as most often occurs in our in-use testing under the consent decrees. For the
purposes of this analysis, we assume that a cash incentive of $150 per vehicle will be paid to the
owner by the engine manufacturer. This is the average cost of the incentive, with some owners
being offered more some less, and some cooperating without an incentive. Therefore, the total
incentive for an engine family tested under the Phase 1 minimum requirements is $750.
1.2.2.3.2 Direct Labor
Each engine family will incur costs in three principle labor categories: vehicle recruitment,
instrument pre-calibration, and data analysis and reporting. We expect that manufacturers will
rely heavily on their existing customer relationships to recruit appropriate test vehicles from
fleets or individual owners. Alternatively, they will create new lines of communication with
their customers. A significant amount of pre-screening and vehicle history will also be
associated with vehicle recruitment. We assume that with a heavy emphasis on existing
customer relationships and data bases, recruiting the requisite five test vehicles will average
about $300 per engine family.
Prior to being deployed in the field, each portable measurement system will be carefully
examined at the manufacturer's facility to ensure proper operation. Based on our experience
with portable emission measurement systems, we estimate that pre-calibrating each unit will
require 0.5 and 1.5 hours of engineer and technician time, respectively. Using the total
compensation rates previously described in Tablel-4, this would cost $58 per unit. Since it is
assumed that testing will be conducted using two portable systems, the total direct labor cost for
pre-calibrating the instruments is $116 per engine family.
The last category of direct labor per engine family is primarily for final data analysis and
quality assurance (beyond that which is conducted in the field), reporting results, and archiving
information. We assume that engine manufacturers will develop a number of automated
methods to perform many of these functions to minimize labor requirements. Our direct labor
estimates are basically taken from another EPA report that was prepared to support a new pilot
program aimed at developing new in-use data collection methods for nonroad diesel-powered
equipment (USEPA 2004). That program will also collect, analyze, and report emissions data
using portable emission measurement systems. For the purposes of this analysis, we doubled the
time per test for managerial oversight, since the original estimate was developed to reflect an
emission factor style program, while the proposed in-use program has more important
compliance implications.
Table 1-7 presents the estimated labor hours for each data analysis and reporting activity, the
cost per test, and the cost per engine family. The cost per test is based on a labor rate of $31 per
hour for an engineer and $47 per hour for a manager. These labor classifications and
-------
compensation rates were previously discussed in Section 1.2.2.2 and presented in Table 1-4. The
total cost of post-data analysis and reporting is estimated to be $728 per engine family.
The resulting total direct labor for the three categories described above is $1,144 per engine
family.
1.2.2.3.3 Travel
Our basic assumptions regarding travel needs are described in Section 1.2.2.1. The costs of
travel are divided into direct labor, vehicle costs, and per diem expenses for room and board.
For Scenario 1, we assume the test sites are located close enough to the manufacturer's facility,
or employees base of operation, so they can commute to and from the job site each day. This
avoids overnight travel expenses. More specifically, we assume that commuting at the
beginning and end of each work day is 30 minutes each way, two trips occur per day, and there
are four testing days for a total of four hours. There is also 30 minutes of travel time between
test sites on Day 2 and Day 3 (Table 1-2). Based on these assumptions, the total travel time is
five hours. Using the total compensation rates previously presented in Table 1-4 of $31 and $28
per hour for an engineer and technician, respectively, the travel-related direct labor cost is$236
for daily commuting and $59 for inter-site travel, or a total of $295 per engine family.
The vehicle expenses associated with Scenario 1 are estimated by using an assumed average
speed of 50 miles per hour and a mileage fee of $0.40 per mile. The mileage cost is based on a
federal government reimbursement rate for privately-owned vehicles of $0.365 per mile in 2003,
with an increase of approximately 10 percent to account for the use of a large van or small truck.
We assume such a vehicle will be required to transport the portable systems and supplies, and
provide a small work space if needed. Using the commuting time of 4 hours, 50 miles per hours,
and $0.40 per mile the total travel related vehicle cost is $80 per engine family. Combining this
with the travel-related labor of $295, the total travel cost for Scenario 1 is $375 per engine
family.
For Scenario 2, we assume the test sites are located far enough away from the manufacturer's
facility, or employees home base, that a single round trip to and from the job site and overnight
travel is required. More specifically, we assume that commuting at the beginning and end of the
work week is four hours one way for a total of eight hours. Again, using the total compensation
rates previously presented in Table 1-4 of $31 and $28 per hour for an engineer and technician,
respectively, the travel-related direct labor cost is $472 per engine family.
The vehicle expenses associated with Scenario 2 consist of the round trip travel to the testing
locations described above, and daily travel to and from the test sites as well as itinerant travel,
e.g., lunch and dinner. The round trip travel expense associated with the eight hours of
commuting time is estimated by using an assumed average speed of 50 miles per hour and a
mileage fee of $0.40 per mile. The result is $160. The itinerant travel distances are assumed to
be 30 miles for Day 1 and 45 miles each day for the next three days, or a total of 165 miles.
Using the assumed vehicle reimbursement fee, this amounts to $66.
10
-------
This scenario requires overnight travel. There will be approximately four days of full per
diem expenses and a partial day of meals on the fifth day (Table 1-3). We assume it costs $100
per night for lodging and $40 per day for meals. This results in per diem expenses of $600 for
the full week. Combining this with the travel-related labor of $472 and the vehicle cost of $66
from above, the total travel cost for Scenario 2 is $1,138 per engine family.
1.2.2.3.4 Labor Overhead
We assume that all direct labor is burdened at 100 percent of the total compensation rate.
1.2.2.3.5 General and Administrative Overhead
We assume general and administrative expenses to be 6.5 percent of all other costs.
1.2.2.3.6 Summary of Variable Cost per Engine Family
Table 1-8 summaries the various cost elements discussed above. As shown, the resulting
total variable costs per engine family are $3,635 and $4,447 for Scenario 1 and 2, respectively.
1.3 Costs of the Proposed Program
Now that the basic fixed and variable cost inputs for each of the two scenarios have been
developed, we will use that information to identify a range of total annual costs for the proposed
program. This range reflects the two testing scenarios, as described in Section 1.2.2.1, and three
different levels of testing intensity that may occur under the Phase 1 and 2 requirements, which
are described below. We will also develop a single point estimate of the proposed program's
annual cost. Finally, we will use this point estimate to present total costs for the first five years
of the program, and costs over 30 years. These costs are presented for the entire industry.
The first level of testing intensity is the minimum number of vehicles that must be tested
under Phase 1 of the proposed program to demonstrate if a designated engine family passes the
NTE criteria, i.e., five vehicle tests per family. This is the basis upon which the variable cost
components were developed in Section 2.1, and is referred to as Phase 1 minimum. The second
level of testing is the maximum number of vehicles that could be required under Phase 1, i.e., 10
vehicle tests per family. This is referred to as Phase 1 maximum. The third level of testing is a
worst case where a manufacturer must complete Phase 2 testing for an engine family. At its
maximum, Phase 2 requires up to 20 vehicle tests per family. This is referred to as Phase 2
maximum.
Overall, our methodology for estimating the costs associated with the three testing levels is
simple. We assume that a manufacturer will complete each level of testing in discreet steps. For
example, after completing Phase 1 minimum, the test results for the engine family will be
thoroughly evaluated at the manufacturer's technical center. If one or more of the vehicles do
not pass the testing criteria, the manufacturer is assumed to return to the field to continue testing
11
-------
five more vehicles, i.e., Phase 1 maximum. For the purposes of this analysis, this means that the
variable cost of Phase 1 maximum testing is twice the cost of performing Phase 1 minimum
testing. Similarly, the variable cost of Phase 2 maximum, i.e., 20 vehicles, is twice the cost of
Phase 1 maximum, i.e., 10 vehicles. The fixed cost of testing is constant for each of the three
testing intensities, since the cost of purchasing the portable measurement systems does not
change with the number of tests performed.
1.3.1 Variable Costs by Level of Testing Intensity
As noted above, fixed costs do not change by the number of tests performed, although
variable costs do vary by testing intensity. Therefore, the first step in estimating the range of
annual costs is to determine variable cost per family for each of the testing levels. This is
presented in Table 1-9 for the two test scenarios.
The next step is to find the range of annual costs for all manufacturers, i.e., all engine
families. Under the proposed program, we may generally select up to 25 percent of an engine
manufacturer' families for testing each year. In the 2003 model year, there were 95 heavy-duty
diesel engine families certified. Hence, we may select up to 24 engine families per year. Using
this value, the resulting range of annual costs for all manufacturers is shown in Table 1-10.
1.3.2 Total Annual Costs by Level of Testing Intensity
Table 1-11 summarizes the fixed and variable cost estimates and presents the total annual
cost for each test scenario and testing level. The low end of the range is about $874 thousand per
year for Phase 1 minimum, and the high end of the range is $2.02 million for Phase 2 maximum.
1.3.3. Total Annual Cost Point Estimate
Our point estimate assumes the overall program will reflect the average of the two scenarios
and weighted at 90 percent of the Phase 1 minimum average cost and 10 percent of the Phase 2
maximum average cost. This reflects our belief that most of the engine families will be designed
and built in full conformance with the applicable NTE standards. But also that the program will
identify some level of potential nonconformance. Table 1-12 summarizes the Phase 1 minimum
and Phase 2 maximum costs from the previous table for convenience and presents our point
estimate of the total annual cost for all manufacturers. The point estimate is $1.02 million per
year.
12
-------
1.3.4. Total Costs Over 5 and 30 Years
We developed an estimate of the total program costs over both 5 and 30 years using the
annual point estimate costs from Table 1-13 and a discount rate of seven percent per annum. As
shown, the 5 year cost is about $7.82 million and the 30 year cost is about $33.5 million.
13
-------
Chapter 1 References
1. U.S. Environmental Protection Agency. 2003. Characterizing Exhaust Emissions from
Light-Duty Gasoline Vehicles in the Kansas City Metropolitan Area. ERG. EPA Contract
Number GS-10F-0036K. Office of Transportation and Air Quality, Assessment and Standards
Division, Ann Arbor, Michigan. Awarded February 2004.
2. Bureau of Labor Statistics. 2003. Employer Costs for Employee Compensation-June 2003.
USDL: 03-446. U.S. Department of Labor, Washington, D.C.
3. U.S. Environmental Protection Agency. 2004. Mobile Source Emission Factors:
Populations, Usage and Emissions of Diesel Nonroad Equipment in EPA Region 7. Agency
Form Number 0619.11, Supporting Statement, Part A. Office of Transportation and Air Quality.
March 2004. EPA Edocket No. OAR-2003-0225-0003.
14
-------
Table 1-1. Total Annualized Fixed Costs1
Cost per
PEMS Unit
($)
100,000
Annualized cost
of PEMS Unit
($)
24,390
Annual Cost per
Manufacturer
($)
48,780
Number of
Manufacturers
(#)
14
Total
Annual Cost
($)
682,890
2003 dollars.
15
-------
Table 1-2. Scenario 1 - Easy Vehicle Access and Local Travel
Day 1
Activity
Travel
(Location 1)
VI, V22
History
VI, V2
Set-to-Spec
VI, V2
Install
Warm
Misc. Time
Travel
(Home)
Totals
Hrs
0.5
2
2
3
1
0.5
7
7
Labor
Type
B1
E
T
B
B
B
T
E
Day 2
Activity
Travel
(Location 1)
VI, V2 Data
Acquisition
VI, V2
Remove
V3
History
V3
Set-to-Spec
V3
Install
Warm
Misc. Time
Travel
(Home)
Hrs.
0.5
1
1.5
1
1
1.5
1
0.5
7
7
Labor
Type
B
B
B
E
T
B
B
B
T
E
Day 3
Activity
Travel
(Location 1)
V3Data
Acquisition
V3
Remove
Travel
(Location 2)
V4, V5
History
V4, V5
Set-to-Spec
V4, V5
Install
Warm
Misc. Time
Travel
(Home)
Hrs
0.5
0.5
.75
0.5
2
2
3
1
0.5
9
9
Labor
Type
B
B
B
B
E
T
B
B
B
T
E
Day 4
Activity
Travel
(Location 2)
V4,V5 Data
Acquisition
V4, V5
Remove
Misc. Time
Travel
(Home)
Hrs
0.5
1
1.5
.25
0.5
4
4
Labor
Type
B
B
B
B
B
T
E
Day 5
Activity
Not
Applicable
Hrs
Labor
Type
1 T=Technician, E=Engineer, B=Both.
2 V = Vehicle (identifier).
16
-------
Table 1-3. Scenario 2 - Limited Vehicle Access and Overnight Travel
Day 1
Activity
Travel
VI, V22
History
VI, V2 Set-
to-Spec
VI, V2
Install
Misc. Time
Totals
Hrs
4
2
2
3
1
10
10
Labor
Type
B1
E
T
B
B
T
E
Day 2
Activity
VI, V2
Warm-Up
Shift Wait
Time
V3 Set-to-
Spec
V3 History
VI, V2 Data
Acquisition
VI, V2
Remove
V3 Install
Misc. Time
Hrs
1
8
1
1
1
1.5
1.5
1
15
14
Labor
Type
T
B
T
E
B
B
B
B
T
E
Day 3
Activity
V4 Set-to-
Spec
V4 History
V4 Install
V3 Warm
Shift Wait
Time
V3, V4 Data
Acquisition
V3, V4
Remove
Misc. Time
Hrs
1
1
1.5
8
1
1.5
1
14
14
Labor
Type
T
E
B
B
B
B
B
T
E
Day 4
Activity
V5 Set-to-
Spec
V 5 History
V5 Install
Shift Wait
Time
V5 Remove
V5Data
Acquisition
Misc. Time
Hrs
1
1
1.5
8
.75
.5
1
13
13
Labor
Type
T
E
B
B
B
B
B
T
E
Day 5
Activity
Travel
Hrs
4
4
4
Labor
Type
B
T
E
1 T=Technician, E=Engineer, B=Both
2 V = Vehicle (identifier).
17
-------
Table 1-4. Labor Compensation Rates
BLS Category1
Technical3
Precision Production, Craft, and
Repair4
Executive, administrative, and
managerial5
In-Use Testing Category
Engineer
Technician
Manager
Total Compensation2
($/Hour)
31.00
28.00
47.00
1 BLS 2003.
2 Total compensation includes wages and salaries, paid leave, supplemental pay, and insurance. Rounded to the nearest whole dollar. June
2003 dollars.
3 Table 11, Private industry, goods-producing and service-producing industries, by occupational group; All workers, goods-producing
industries; White-collar occupations; Professional specialty and technical.
4 Table 11, Private industry, goods-producing and service-producing industries, by occupational group; All workers, goods-producing
industries; Blue-collar occupations; Precision production, craft, and repair.
5 Table 11, Private industry, goods-producing and service-producing industries, by occupational group; All workers, goods-producing
industries; White-collar occupations; Executive, administrative, and managerial.
18
-------
Table 1-5. Direct Labor Costs Per Vehicle Testl
(Based on Phase 1 Minimum Five Vehicle Tests per Engine Family)2
Scenario
1
2
Labor Rate
Type
Regular
Regular
Overtime2
S2 Total
Hours Per Family (5 tests)
Engineer
27
40
15
—
Technician
27
40
16
—
Hourly Compensation
($ per hour)
Engineer
31
31
46
—
Technician
28
28
46
—
Total Cost
($/5 tests)
1593
2360
1426
3786
Cost Per
Test
($/test)
319
472
285
757
1 June 2003 dollars.
2 Based on Phase 1 testing five vehicles.
3 Overtime paid for work exceeding 40 hours/week at 1.5 times the regular pay rate.
19
-------
Table 1-6. Variable Costs Per Test Vehicle1
($/test)
Scenario
1
2
Direct
Labor2
319
757
Labor
Overhead3
319
757
Other Direct
Costs4
100
100
Voided
Tests5
74
161
General and
Administrative6
53
115
Total
865
1,891
1 2003 dollars.
2 See Table 1-5.
3 100 percent of direct labor.
4 General supplies, PEMS maintenance, calibration gases, FID fuel, etc.
5 Assumes 10 percent of tests are void (i.e., 0.10 * (direct labor, labor overhead, and other direct costs)).
6 6.5 percent of all costs.
20
-------
Table 1-7. Post-Test Data Analysis and Reporting Variable Cost Per Family1
Activity
Q A Measure-
ments
Load
Database
Analysis,
Write Report,
Archive
Total
Hours/Test
(hrs)2
Manager
0.056
0.056
0.056
0.168
Engineer
3.0
0.278
1.167
4.445
Cost/Test
($)
Manager
2.63
2.63
2.63
7.90
Engineer
93.00
8.62
36.18
137.80
Cost Per Engine Family
($/5 Tests)
Manager
13
13
13
39
Engineer
465
43
181
689
Total
478
56
194
728
1 2003 dollars.
2 See USEPA 2004.
21
-------
Table 1-8. Summary of Variable Costs Per Engine Family1
($/family)
Scenario
1
2
Direct Labor
1,144
1,144
Labor
Overheard2
1,144
1,144
Incentive
750
750
Travel
375
1138
General and
Administrative3
222
271
Total
3,635
4,447
1 2003 dollars.
2 100% of direct labor.
3 6.5% of all costs.
22
-------
Table 1-9. Summary of Variable Costs Per Engine Family by Level of Testing Intensity1
($)
Scenario
1
2
Phase 1 Minimum2
Vehicle
Testing4
4,325
9,455
Engine
Family
3,635
4,447
Total
7,960
13,902
Phase 1 Maximum3
Vehicle
Testing4
8,650
18,910
Engine
Family
3,635
4,447
Total
12,285
23,357
Phase 2 Maximum4
Vehicle
Testing4
17,300
37,820
Engine
Family
3,635
4,447
Total
20,935
42,267
1 2003 dollars.
2 Phase 1 minimum = 5 test vehicles.
3 Phase 1 maximum = 10 test vehicles.
4 Phase 2 maximum = 20 test vehicles.
5 Cost per vehicle from Table 1-6 multiplied by the number of vehicles tested.
23
-------
Table 1-10. Total Annual Variable Costs for All Manufacturers by Level of Testing Intensity1
($)
Scenario
1
2
Phase 1 Minimum2
Cost
Per
Engine
Family
7,960
13,900
#
Families
Per
Year5
24
24
Total
191,040
333,600
Phase 1 Maximum3
Cost
Per
Engine
Family
15,920
27,800
#
Families
Per
Year
24
24
Total
382,080
667,200
Phase 2 Maximum4
Cost
Per
Engine
Family
31,840
55,600
#
Families
Per
Year
24
24
Total
764,160
1,334,400
1 2003 dollars.
2 Phase 1 requires that a minimum of 5 test vehicles.
3 The maximum number of vehicles tested in Phase 1 is 10.
4 The maximum number of vehicles tested through Phase 2 is 20.
5 25% of a total of 95 engine families certified in the 2003 model year.
24
-------
Table 1-11. Total Annual Costs for All Manufacturers by Level of Testing1
Intensity
($ Thousands)
Scenario
1
2
Phase 1 Minimum2
Fixed
Cost
683
683
Variable
Cost
191
334
Total
874
1,017
Phase 1 Maximum3
Fixed
Cost
683
683
Variable
Cost
382
667
Total
1,065
1,350
Phase 2 Maximum4
Fixed
Cost
683
683
Variable
Cost
764
1,334
Total
1,447
2,017
1 2003 dollars.
2 Phase 1 minimum = 5 vehicle tests.
3 Phase 1 maximum = 10 vehicle tests.
4 Phase 2 maximum = 20 vehicle tests.
25
-------
Table 1-12. Total Annual Cost Point Estimate for All Manufacturers1
($ Thousands)
Scenario
1
2
Average3
Phase 1 Minimum
Fixed
Cost
683
683
683
Variable
Cost
191
334
263
Total
874
1,017
945
Phase 2 Maximum
Fixed
Cost
683
683
683
Variable
Cost
764
1,334
1,049
Total
1,447
2,017
1,732
Point Estimate2
Fixed
Cost
683
683
683
Variable
Cost
248
434
341
Total
931
1,117
1,024
1 2003 dollars.
2 Assumes a 90/10 split between Phase 1 minimum and Phase 2 maximum, respectively.
3 Assumes a 50/50 split between Scenarios 1 and 2.
26
-------
Table 1-13. Total Program Cost Over 5 and 30 Years1
(Based on Point Cost Estimate)
($)
Year
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
30 Year NPV in 2004
IstSYearNPVin
2004
Annualized Fixed
Costs
682,890
682,890
682,890
682,890
682,890
682,890
682,890
682,890
682,890
682,890
682,890
682,890
682,890
682,890
682,890
682,890
682,890
682,890
682,890
682,890
682,890
682,890
682,890
682,890
682,890
682,890
682,890
682,890
682,890
682,890
13,384,945
3,127,436
Annual Variable
Costs
,024,000
,024,000
,024,000
,024,000
,024,000
,024,000
,024,000
,024,000
,024,000
,024,000
,024,000
,024,000
,024,000
,024,000
,024,000
,024,000
,024,000
,024,000
,024,000
,024,000
,024,000
,024,000
,024,000
,024,000
,024,000
,024,000
,024,000
,024,000
,024,000
,024,000
20,070,852
4,689,620
Total Annual
Costs
,706,890
,706,890
,706,890
,706,890
,706,890
,706,890
,706,890
,706,890
,706,890
,706,890
,706,890
,706,890
,706,890
,706,890
,706,890
,706,890
,706,890
,706,890
,706,890
,706,890
,706,890
,706,890
,706,890
,706,890
,706,890
,706,890
,706,890
,706,890
,706,890
,706,890
33,455,797
7,817,056
1 2003 dollars.
27
-------
CHAPTER 2
On-Vehicle Portable Emissions Measurement Technology Review
2.1 Overview
With respect to measurement equipment, we already have equipment readily available to measure
gaseous emissions on-vehicle using the test procedures proposed for this program. Additionally, we
think it is possible that PM emissions measurement equipment will be available for the start of the
2005/2006 pilot program.
In setting the NTE standard we have already taken into account the variation in emissions due to
varying engine operation and ambient conditions. In addition, in this proposal, we have taken into
account the measurement tolerances of on-vehicle measurement systems.
Given the very active interest in portable measurement equipment by EPA, the California Air
Resources Board, and the automotive industry, and given the available lead time, we believe that
measurement equipment will be widely available so that this proposed program will be fully
implemented for all regulated emissions-including PM-by 2007. For the 2005-2006 pilot program,
gaseous emissions measurement equipment is already available for use at the 2005-2006 gaseous
emissions levels. Complete portable systems that measure PM emissions will take additional time
before they are field ready. Based on the current availability of key measurement technologies and
ongoing work to develop the requisite PM sampling hardware, we are confident these systems will
be widely available by 2007, and that they may be available in time for use in the pilot program
beginning in 2005. This section discusses these measurement technologies and summarizes research
results.
2.2 Measurement Technologies
We expect that several complete systems will be available for use in the proposed in-use testing
program that will be capable of performing the measurements needed to determine whether or not a
vehicle passes an on-vehicle emissions test. At a minimum, any such measurement system must
include individual analyzers and sensors that can quantify the following parameters:
1. Regulated emissions concentrations in exhaust:
a. Oxides of nitrogen, NOx.
b. Carbon monoxide, CO (and carbon dioxide CO2).
c. Non-methane hydrocarbons, NMHC.
d. Particulate mass, PM.
2. Exhaust flow rate.
3. Engine operation:
a. Speed.
b. Torque.
28
-------
c. Coolant temperature and intake manifold temperature and pressure.
4. Ambient conditions:
a. Temperature.
b. Dewpoint.
c. Altitude.
In this section we describe the measurement technologies that we expect to be used to quantify these
parameters. If these technologies are properly applied, we believe that they are acceptable for
measuring emissions on-vehicle. Note too that we also allow for the use of alternate technologies
according to §1065.10. Note that this provision is proposed to be amended as part of another
proposed rulemaking that is a companion to this notice.
1. Regulated emissions concentrations in exhaust. Emissions concentrations need to be measured to
determine brake-specific emissions.
a. NOx measurement technology. We typically accept NOx measured as the sum of NO and
NO2 since conventional engines and aftertreatment systems do not emit significant amounts of other
NOx species. NO may be measured either by a chemiluminescence detector (CLD), a non-
dispersive ultra-violet (NDUV) detector, or a zirconia oxide (ZrO2) sensor. NO2 may be converted
catalytically to NO and detected by a CLD, or it may be detected directly via NDUV or ZrO2. We
believe that CLDs are not likely to be used on-vehicle due to their compressed gas requirements, and
they might likely be sensitive to vehicle vibration. NDUV and ZrO2, on the other hand, are already
available as components of complete on-vehicle emissions measurement systems, and they already
have been performing well in on-vehicle applications. For example, a recent study by the California
Air Resources Board (CARB) indicated that for 27 heavy-duty diesel chassis dynamometer tests, an
NDUV-based on-vehicle system reported NOx emissions within 4.6 % of the current NOx standard,
as compared to laboratory measurements.(l) We are currently studying NDUV analyzer
performance with a 2002 light-heavy duty diesel (LHDD) on a chassis dynamometer. Our results so
far indicate that the NDUV-based system reported NOx within 3.1 % of our laboratory, prior to a
5000-mile cross-country road test. After running the NDUV-based system for the entire road
test-with no failures, the vehicle was returned to the laboratory, and the NDUV-based system
reported NOx emissions within 3.9 % of our laboratory. The manufacturer of the NDUV-based
system has also indicated that several engine manufacturers have evaluated their system, and their
results from 11 HDDE FTP tests indicate Nox emissions were reported within 4.4 % of the current
standard in engine manufacturer's laboratories. (1) ZrO2 sensors are expected to be used on-vehicle
not only for Nox emission measurements, but also for feedback control of NOx aftertreatment
systems.(2) The ZrO2 sensor needed for aftertreatment control is a component originally designed
and developed for gasoline powered vehicles (in this case lean-burn gasoline vehicles) that are
already well developed and can be applied with confidence in long life for NOx adsorber based
diesel emission control. The ZrO2 sensor is an evolutionary technology based largely on the current
Oxygen (O2) sensor technology developed for gasoline three-way catalyst based systems. Oxygen
sensors have proven to be extremely reliable and long lived in passenger car applications, which see
significantly higher temperatures than would normally be encountered on a diesel engine. Diesel
engines do have one characteristic that makes the application of ZrO2 sensors more difficult. Soot
29
-------
in diesel exhaust can cause fouling of the ZrO2 sensor damaging its performance. However this
issue can be addressed through the application of a catalyzed diesel particulate filter (CDPF) in front
of the sensor. The CDPF then provides a protection for the sensor from PM while not hindering its
operation. Since we expect NOx measurement only downstream of a CDPF in each of the potential
technology scenarios we have considered, this solution to the issue of PM sooting is already
addressed.
b. CO (and CO2) measurement technology. Since we first regulated CO, non-dispersive infra-red
(NDIR) detector technology has been used for measuring CO and CO2 in laboratory applications.
Many laboratory NDIR analyzers have a moving part called a chopper-wheel to pulse infra-red light
through the exhaust gas sample. The pulsing light is used to alternately detect the CO and CO2
concentrations and then the dark-current of the NDIR detector. This is done to maintain accuracy,
but the moving chopper-wheel is not durable under the high vibration environment of on-vehicle
testing. However, new NDIR analyzers have been commercialized that electrically switch the infra-
red light source on and off. These new NDIR analyzers are already available in complete on-
vehicle emissions measurement systems, and they have been performing well in on-vehicle
applications. For example, a recent study by the California Air Resources Board (CARB) indicated
that for 27 heavy-duty diesel chassis dynamometer tests, an NDIR-based on-vehicle system reported
CO emissions within 0.7 % of the current CO standard (2.1 % for CO2), as compared to laboratory
measurements.(1) We are currently studying NDIR analyzer performance with a 2002 light-heavy
duty diesel (LHDD) on a chassis dynamometer. Our results so far indicate that the NDIR-based
system reported CO within 1.0 % of our laboratory (1.1% for CO2), prior to a 5000-mile cross-
country road test. After running the NDIR-based system for the entire road test-with no failures, the
vehicle was returned to the laboratory, and the NDIR-based system reported CO emissions within
7.1 % of our laboratory (2.2 % of the current CO standard) and 4.7 % for CO2. The manufacturer of
the NDIR-based system has also indicated that several engine manufacturers have evaluated their
system, and their results from 11 HDDE FTP tests indicate CO emissions were reported within 0.5
% of the current standard in engine manufacturer's laboratories (3.55 % for CO2).(1)
c. NMHC measurement technology. The flame ionization detector (FID) has been the measurement
technology of choice for hydrocarbon measurements since the late 1950s. The FID has been used as
a detector in liquid and gas chromatography systems for individual hydrocarbon speciation and
quantification. It is used because of its response to a broad range of hydrocarbons, its inherent
stability and its remarkably linear response from very high levels to very low levels of hydrocarbons.
Because the FID responds to a very broad range of hydrocarbons, we chose to set our initial
hydrocarbon standard based on the FID response to the total hydrocarbons (THC) in engine exhaust.
Later, by allowing for the subtraction of methane (CH4) from THC, we set non-methane
hydrocarbon (NMHC) standards based on the FID's response to all non-methane hydrocarbons in
engine exhaust. Because the FID has a range of response factors for all of the hydrocarbons that it
detects, and because the mixture of hydrocarbon species in engine exhaust changes as a function of
engine operation, fuel, and aftertreatment systems, the FID's response to NMHC in engine exhaust
is characteristic to hydrocarbon measurement via FID technology alone. This makes it almost
impossible for other hydrocarbon detector technology to equivalently detect engine exhaust NMHC.
30
-------
Fortunately, FIDs have been adapted for on-vehicle use. These new FID analyzers are already
available in complete on-vehicle emissions measurement systems, and they have been performing
well in on-vehicle applications. For example, a recent study by the California Air Resources Board
(CARB) indicated that for 27 heavy-duty diesel chassis dynamometer tests, a FID-based on-vehicle
system reported THC emissions within 2.8 % of the current NMHC standard, as compared to
laboratory measurements.(l) We are currently studying FID analyzer performance with a 2002
light-heavy duty diesel (LHDD) on a chassis dynamometer. Our results so far indicate that the FID-
based system reported THC within 7.8 % of our laboratory after running the FID-based system for a
5000 mile road test-with no failures, (2.4 % of the current NMHC standard). The manufacturer of
the FID-based system has also indicated that several engine manufacturers have evaluated their
system, and their results from 11 HDDE FTP tests indicate THC emissions were reported within 1.3
% of the current NMHC standard in engine manufacturer's laboratories.(1)
d. PM measurement technology. PM measurement has been traditionally conducted by depositing
diluted exhaust PM on a sample filter and then weighing the filter in a PM measurement laboratory
before and after testing to determine the net mass gain due to PM.
This technique has been applied to on-vehicle testing by one on-vehicle emissions measurement
system manufacturer. This system was tested in the laboratory by the California Air Resources
Board (CARB) and the results were compared to those from chassis dynamometer testing.(3)
Thirty-three tests were run on two different heavy-duty trucks, and one of the trucks was equipped
with a PM trap. The 33 emissions results were collected over five different test cycles for each
truck. For the current-technology truck, on-average, the on-vehicle system reported PM results
within 0.6 % of the current standard, which will also remain in effect during the 2005-2006 pilot
program, when compared to laboratory results. For the trap-equipped truck, on-average, the on-
vehicle system reported PM results within 38 % of the 2007 standard, when compared to the
laboratory. However, because the trap-equipped truck was emitting PM at only 44 % of the 2007
standard (according to laboratory results), the 38 % error of the on-vehicle system would not have
caused any false indication of a failure. Furthermore, neither the laboratory nor the on-vehicle
system were equipped to sample PM according to our specifications for measuring PM from engines
that meet the 2007 PM standards. These specifications were tailored to reduce variability in this
type of PM measurement.
These filter-based results demonstrate that accurate on-vehicle PM measurement technology is
already available for the current level of PM emissions, and it demonstrates that proportional
sampling of PM on-vehicle is available today. However, we do not expect filter-based methods to
be used for conducting NTE tests in the field. This is because for NTE testing, PM emissions must
be quantified for several individual NTE events, which would require many filters and a means to
switch these filters in an automated way. No such system is available or in development to our
knowledge, and we believe that such an automated system might be cumbersome on-vehicle.
We are currently evaluating more automated technologies for quantifying PM mass on-board. The
underlying automated measurement technologies detect the inertia of particulate mass (PM) by
31
-------
accelerating it via vibration, rather than detecting its weight due to the acceleration of Earth's
gravity. The inertial measurement technologies include the Tapered Element Oscillating
Microbalance (TEOM) and the Quartz Crystal Microbalance (QCM). Since these technologies are
compact, they are suitable for on-vehicle applications. And since they impart greater acceleration
upon PM versus Earth's gravity, they are more sensitive than their laboratory microbalance cousins.
They also eliminate the need to transport PM sample filters to a PM measurement laboratory for pre-
and post-weighing. Researchers at West Virginia University have used a heavy-duty diesel engine
to compare QCM and TEOM measurement devices, which can be purchased today, versus engine
dynamometer testing in the laboratory. They showed that for seven repeats of EPA's heavy-duty
FTP, the TEOM and QCM devices can quantify PM within 5 % of a traditional microbalance at
current emissions standards.(4)
Although these results demonstrate that this automated on-vehicle PM measurement devices exist
today, more work is needed to demonstrate their accuracy in the lab and in the field. Further,
conducting NTE tests in the field poses additional challenge . Namely, quantification of PM over
sampling intervals as short as 30 seconds has yet to be demonstrated for engines certified to the
current PM emission standards, or those that are expected to meet the more stringent standards
beginning in 2007. At least initially, the technological challenge represented by such a short
sampling interval may require sampling times beyond the minimum 30 second NTE event.
Additionally, because PM equilibration is required before and after each NTE event, the time it takes
to equilibrate PM sample after an NTE event might prevent capturing all or part of the next NTE
occurrence, if that subsequent event closely follows the previous one. Even if these limitations are
not resolved, however, they can be accommodated in the design of the proposed in-use testing
program. NTE sampling times longer than the 30 second minimum, or not capturing a portion of a
valid NTE sampling event, simply means that some potential NTE samples will not be detectable.
In addition to the measurement technologies described above, work is continuing to miniaturize on-
board proportional sampling devices, and to develop suitable exhaust sampling techniques and
hardware. We believe this work will lead to completely integrated on-board PM measurement
systems that will be available for use in the proposed in-use testing program sometime in the 2006-
2007 time frame.
Also, we are currently investigating the sources of error in the laboratory PM measurement of trap-
equipped engines. When we initially compared the TEOM and QCM versus laboratory PM
measurements from trap-equipped engines, we discovered that the laboratory results were very
sensitive to sampling conditions. This is due to the fact that PM from a trap is mostly semi-volatile
matter, such as high-molecular weight hydrocarbons and dilute sulfuric acid. These PM constituents
can exist either as a gas or as PM, depending upon dilution conditions, pressures, temperatures, and
PM collection media. It is important to note that the TEOM uses a different type of media than the
lab, and the QCM uses a platinum substrate to collect PM. Within our current specifications for
sampling post-trap PM, two different acceptable laboratory filter media give results where one is
four times that of the other.(5) We believe that this difference is from a combination of gaseous
hydrocarbons adsorbing onto one filter, while PM hydrocarbons are stripped off of the other filter.
32
-------
We are currently supporting a Coordinating Research Council study to resolve these laboratory PM
measurement issues, and we expect that results from this study, which should be available before the
end of 2004, will allow us to more accurately compare the TEOM and QCM to laboratory
measurements of trap-equipped diesels. Based on the results of this study, we will select a single
filter material specification for laboratory PM measurement. We expect that such a specification
will resolve most of the current issues with post-trap PM measurement in the laboratory.
Furthermore, based on this study, we will likely specify an on-vehicle PM sampling dilution rate and
ratio, along with the on-vehicle equilibration pressure, temperature and humidity for PM samples.
By specifying these sampling conditions, we can help assure that PM measurement on-vehicle will
be sufficiently equivalent to laboratory PM measurements-even at our most stringent PM standard.
2. Exhaust flow rate. In a CVS laboratory the entire volume of engine exhaust is diluted and then
measured. Since this is impractical for on-vehicle emissions measurement, the raw exhaust flow rate
must be measured. We are aware of four available technologies for on-vehicle exhaust flow
measurement. One has been developed and patented by us, and it is based on an averaging Pitot
tube (Patent No. 6,148,656). This technology is available because we have licensed the technology
to two on-vehicle emissions measurement system manufacturers. Another technology uses a hot-
wire anemometer to measure the flow of ambient air induced by a sub-sonic venturi placed in the
raw exhaust. A third technology uses a heated hot-wire anemometer directly in the exhaust. A
fourth technology measures a known proportion of raw exhaust flow via the laboratory CVS
technique. It's proportionality is maintained with the total raw exhaust flow by balancing certain
partial flow pressures with the exhaust tailpipe and ambient pressures. All of these techniques have
been demonstrated to be within 5 % of the true exhaust flow, and two of these techniques were used
to measure the flows required to achieve the gaseous and PM measurement results indicated in the
previous section (1,3).
3. Engine operation. Certain engine parameters are required to calculate emissions or to determine
whether or not an engine is operating in the NTE zone. Other parameters are used to determine if an
EGR-equipped engine is sufficiently warmed-up for NTE testing. These parameters may be
measured directly using the technologies described below. However, if the engine manufacturer
determines that an engine's Electronic Control Module (ECM) accurately quantifies these
parameters, the manufacturer may rely on ECM values for these parameters.
a. Speed. Engine crankshaft speed is required to determine whether or not an engine is
operating within the NTE zone. Engine speed also may be used to determine engine power for
emissions calculations. We have used magnetic flux detectors attached to the housing of an engine's
belt-driven alternator to measure engine speed. Other on-vehicle emissions measurement system
manufacturers detect alternator voltage ripple frequency. These signals are calibrated to actual
engine speed during each engine installation with a portable reference tachometer.
b. Torque. Engine torque is required to determine whether or not an engine is operating
within the NTE zone. Engine torque also may be used to determine engine power for emissions
calculations. Engine torque may be measured directly by installing and calibrating a strain gage on
the drive shaft. We also allow torque determination using fuel flow, as calculated via carbon-
33
-------
balance, in combination with engine speed and an estimated brake-specific fuel consumption. For
details, refer to §1065.650, which is proposed in a companion NPRM to this notice.
c. Coolant temperature, intake manifold temperature, and intake manifold pressure. These
three parameters are used to determine whether or not an EGR-equipped engine is sufficiently
warmed-up for NTE testing. These can be measured with standard thermocouples and automotive
pressure transducers, which can be mounted into coolant and intake system caps or plugs.
4. Ambient conditions. Ambient conditions are used to calculate emissions or to determine if
ambient conditions are within limits for NTE testing. These parameters may be measured directly
using the technologies described below, or if the engine manufacturer determines that an engine's
Electronic Control Module (ECM) accurately quantifies these parameters, the manufacturer may rely
on ECM values for these parameters.
a. Temperature. We have used thermistor-based and thermocouple-based ambient temperature
sensors for this purpose. Either technology is sufficient for this temperature measurement. These
sensors are rugged because they are commonly used in remote weather station applications, however
these sensors must be shielded from heat from the sun and heat from the engine to achieve accurate
ambient temperature readings.
b. Dewpoint. We have used thin-film capacitor-based ambient dewpoint sensors for this
purpose. This technology is sufficient for this dewpoint measurement. These sensors are rugged
because they are commonly used in remote weather station applications.
c. Altitude. We have used Global Positioning System (GPS) technology to measure altitude.
We have used this technology cross-country as part of a complete on-vehicle emissions
measurement system, and it measured altitude accurately.
34
-------
Chapter 2 References
1. "On the Road to Clear Skies: Semtech-D For On-highway Heavy-duty Diesel Applications."
SAE Government/Industry Meeting May 13, 2003, Andrew Reading, Sensors Incorporated.
2. "Control of Air Pollution from New Motor Vehicles: Heavy-Duty Engine and Vehicle Standards
and Highway Diesel Fuel Sulfur Control Requirements", Regulatory Impact Analysis, "Heavy-Duty
Standards / Diesel Fuel (EPA420-R-00-026)", Chapter III: Emissions Standards Feasability, pp. III-
56 -111-57., January 18, 2002, U.S. Environmental Protection Agency.
3. "Validation of the RAVEM Ride-along Vehicle Emissions Measurement System.", Final Report
June 19, 2001, Christopher Weaver, Engine, Fuels, and Emissions Engineering.
4. "Real-Time Particle Characterization of Diesel and Gasoline Particulate Mass", Booker, Dr.
David R., Gautam, Prof. Mridual, Carder, Daniel K, Gautam Seema, ETH Conference on
Nanoparticle Measurement, Zurich, August 2001.
5. "Parti culate Mass Measurements of Heavy-duty Diesel Engine Exhaust Usine 2007 CVS PM
Sampling Parallel to QCM and TEOM.", Final Report No. 08.06129 September 30, 2003, Imad A.
Khalek, Ph.D., Southwest Research Institute.
35
-------
CHAPTER 3
New In-Use Testing Instrument Measurement Allowance
Before discussing the basis for the new measurement allowance that we are proposing for on-vehicle
emissions measurements, it is instructive to review the restricted engine operation that the NTE zone
covers, the list of other NTE allowances that we already have finalized, and other allowances that we
propose elsewhere in this notice.
3.1 Revi e w of NTE zone
On October 6, 2000, we published Not-To-Exceed (NTE) rules and regulations for heavy-duty diesel
engines ( 65 Fed. Reg. 59896, Oct. 6, 2000); effective for engines starting with model year 2007.
These regulations were revised on January 18, 2001 consistent with our promulgation of more
stringent base standards for heavy-duty diesel engines (66 Fed. Reg. 5001, January 18, 2001).
Briefly, the NTE provisions specify brake-specific averaging periods as short as 30 seconds, and
under these provisions testing is restricted to a limited region of engine operation. Namely, when all
of the following conditions are simultaneously met for at least 30 seconds, an NTE event is
generated. Note however, that if an aftertreatment system were to regenerate during this time, the
minimum time under which all of these conditions must be met would increase to at least twice the
regeneration interval:
1. Engine speed must be greater than 15% above idle speed.
2. Engine torque must be greater than or equal to 30% of maximum torque.
3. Engine power must be greater than or equal to 30% of maximum power.
4. Vehicle altitude must be less than or equal to 5500 feet.
5. Ambient temperature must be less than or equal to 100 degrees F at sea level to 86 degrees
F at 5500 feet.
6. Brake-specific fuel consumption (BSFC) must be less than or equal to 105 % of the
minimum BSFC if an engine is not coupled to a multi-speed manual or automatic
transmission.
7. Engine operation must be outside of any manufacturer petitioned exclusion zone.
8. Engine operation must be outside of any NTE region in which a manufacturer states that
less than 5% of in-use time will be spent.
9. For EGR-equipped engines, the intake manifold temperature must be greater than or equal
to 86-1 OOF, depending upon intake manifold pressure.
10. For EGR-equipped engines, the engine coolant temperature must be greater than or equal
to 125-140 degrees F, depending on intake manifold pressure.
11. Engine aftertreatment systems' temperature must be greater than or equal to 250 degrees
C.
36
-------
3.2 Review of existing NTE allowances
As part of these rules, we also finalized several compliance allowances with respect to meeting the
Not-To-Exceed (NTE) standard. At that time we did not finalize any NTE allowances that were to
specifically account for any differences between the quality of laboratory measurements and on-
vehicle measurements. Note that we did have finalized the following NTE allowances:
1. We allowed for the use of the family emissions limit (FEL) to which an engine was certified,
rather than using actual emissions standard as the NTE standard. This allowance accounts for any
differences in the engine's certified emissions and the actual emissions standard when comparing
on-vehicle emissions to laboratory certified emissions.
2. We allowed for NTE multipliers of 1.25x and 1.5x times the engine's certified emissions,
depending upon the level of the engine's certified emissions compared to the standard. This
multiplier allowance accounts for any differences in engine operation and/or ambient conditions
when comparing on-vehicle emissions to laboratory certified emissions.
3. We allowed for rounding of the compliance limit after multiplying the engine's certified
emissions by 1.25x or 1.5x. Therefore, when an engine is certified at 0.01 g/hp-hr PM, this rounding
effectively increases the 1.5x multiplier to 2.Ox. This allowance creates an NTE compliance limit
with the same number of significant figures as the emissions standard. This allows for a pass/fail
determination with the same number significant figures when comparing on-vehicle emissions to
laboratory certified emissions.
4. For NOx emissions during the initial years of the new base standards, we allowed an additive
allowance of 0.10, 0.15, or 0.20 g/hp-hr following sale to the ultimate purchaser, based on an
engine's model year, its certified emissions, and vehicle accumulated mileage. Therefore, this
additive allowance effectively increases the 1.5x multiplier to 2.Ox, 2.25x and 2.50x for NOx. This
additive allowance accounts for any differences in the performance of the first an in-use-aged NOx
emissions control systems (including the engine and aftertreatment systems) when comparing their
on-vehicle emissions to laboratory certified emissions.
5. For PM emissions during the initial years of the new base standards, we allowed an additive
allowance of 0.01 g/hp-hr following sale to the ultimate purchaser, based on an engine's model year
and its certified emissions. Therefore, this additive allowance, combined with the effect of rounding
in (3) effectively increases the 1.5x multiplier to 3.Ox for PM. This additive allowance accounts for
any differences in the performance of the first an in-use-aged PM emissions control systems
(including the engine and aftertreatment systems) when comparing their on-vehicle emissions to
laboratory certified emissions.
37
-------
3.3 Review of other proposed NTE allowances
As discussed in other documents for this rulemaking we also propose additional NTE allowances for
demonstrating compliance with this program These include the following allowances:
1. Elsewhere in this notice we are proposing to allow 10 % of the measured NTE events to exceed
the NTE standard-that is, they can exceed the NTE standard after applying all of the allowances
discussed above. For model years 2007 through 2009, we are proposing to restrict these NTE
exceedences to 2x the NTE standard, after applying all of the allowances discussed above. In the
case of NOx emissions certified at or below 0.50 g/hp-hr-for model years 2007 through 2009-we
are also proposing to allow these NTE exceedences to occur up to a threshold of 2.00 g/hp-hr. After
model year 2009, we are proposing no upper limit to these NTE exceedences. In other words, for
model years 2007 through 2009, we would allow 10 % of the NTE events to have an effective NOx
multiplier of up to 5x andlOx, depending upon an engine's certified NOx emissions. For PM, the
effective NTE multiplier would be 6x.
2. Elsewhere in this notice we are proposing that we would consider an engine family to have met
the requirements of this program even though a fraction of engines within that engine family failed
to meet the vehicle-pass criteria.
3.4 Discussion of the two (2) proposed NTE measurement allowances
In addition to all of the NTE allowances discussed above, we also propose two additional allowances
to account for the quality of the measurements we expect on-vehicle. We are proposing these two
allowances in order to account for the difference between the minimum quality of the measurements
that we expect on-vehicle and the minimum quality of measurements that we expect in a laboratory.
For all standards where we expect laboratory-based engine control and measurement (i.e. FTP,
SET), we have already accounted for measurement quality within the numeric value of the emissions
standards. That is, in setting the numeric emissions standards, we have considered a nominal
"compliance margin" that engine manufacturers are likely to assume when developing an engine to
meet our standards. This margin typically consists of the variability of results we expect due to
engine production tolerances, test cycle repeatability, ambient conditions, and measurement error.
Therefore, if we were ever to consider an allowance for laboratory measurement error, we would, in
turn, have to similarly consider increasing the stringency of the laboratory standard in order to
maintain the same overall stringency of the standard.
Since our NTE standards are based upon multipliers of our laboratory-based standards, the NTE
standards already include our consideration of laboratory measurement error. On the other hand, the
NTE standards do not yet include any additional allowance that accounts for the difference in the
minimum measurement quality that we expect for on-vehicle measurements versus the minimum
measurement quality that we expect for laboratory-based measurements.
38
-------
Because we are proposing to adopt Part 1065 for on-vehicle test procedures in this notice, and we
are proposing to adopt Part 1065 for laboratory test procedures in a companion notice, below is a
table of all of the measurement accuracy and repeatability differences that we expect between on-
vehicle measurements and laboratory measurements:
39
-------
Table 3-1: Comparison of expected on-vehicle and laboratory measurement accuracy and repeatability
Measurement
Engine speed transducer
Engine torque (orBSFC)
General pressure transducer
(not a component of another
instrument)
Barometer
General temperature sensor
(not a component of another
instrument)
General dewpoint sensor
Exhaust rate meter
Constituent concentration,
continuous analyzer
Accuracy, repeatability, and noise are determined with the same collected data, according to §1065.305.
bpt. refers to a single point at the average value expected at the standard—the reference value used in §1065.305.
°max. refers to the maximum value expected at the standard over a complete test; not the maximum of the instrument's range.
Quantity
CJ
T
P
p
barom
T
* dew
n
X
On-vehicle
Accuracy3
pt.b or max.0
5.0%ofpt. or
1.0% of max.
6.0%ofpt. or
3 % of max.
5.0%ofpt. or
2.0% of max.
250 Pa
1.0%ofpt. or
3°C
3°C
5.0%ofpt. or
3.0% of max.
2.5%ofpt.
Laboratory
Accuracy3
pt.b or max.0
2.0%ofpt. or
0.5 % of max.
2.0%ofpt. or
1 % of max.
1.0%ofpt. or
0.5% of max.
50 Pa
0.5%ofpt. or
3°C
1°C
2.5%ofpt. or
1.5% of max.
1.0%ofpt.
On-vehicle
Repeatability3
pt.b or max.0
2.0%ofpt.
2 % of max.
2.0%ofpt.
0.5% of max.
200 Pa
0.5%ofpt. or
2°C
1°C
2.0%ofpt.
1.0%ofpt.
Laboratory
Repeatability3
pt.b or max.0
1.0%ofpt.
0.5 % of max.
0.5%ofpt.
0.25% of max.
25 Pa
0.25% of pt. or
1°C
0.5 °C
1.0%ofpt.
0.5%ofpt.
To evaluate the combined effect of the accuracy and repeatability specifications in Table 3-1, we used the definition of Limit of
Quantification (LOQ)(1), which is the sum of the accuracy plus 2 times the repeatability. In order to evaluate LOQ independent of engine
40
-------
specifics, we used the percent of point values where percent of point and percent of maximum
tolerances were given.
Table 3-2 Combined effect of the accuracy and repeatability specifications
Measurement Quantity On-vehicle Laboratory
LOQ LOQ
Engine speed transducer a> 9 % 4 %
Engine torque (or efuel T 10 % 3 %
e.g. BSFC)
General pressure P 9% 2%
transducer
(not a component of
another instrument)
Barometer Pbarom 650 Pa 100 Pa
General temperature T 2 % 1 %
sensor
(not a component of
another instrument)
General dewpoint sensor Tdew 5 °C 2 °C
Exhaust rate meter n 9% 4.5%
Constituent x 4.5 % 2.0 %
concentration,
continuous analyzer
In order to calculate the combined measurement error of these individual measurements, we
propagated the respective errors according to the emissions calculation formulas used in each case.
Because on-vehicle measurements essentially require raw exhaust continuous sampling and
laboratory measurements allow for raw exhaust continuous sampling, we compared both cases for
raw exhaust continuous sampling. Because the measurements of pressure, temperature, and
dewpoint have a relatively minor affect on emissions calculations in comparison to engine speed,
torque, emissions concentration and exhaust flow rate, we did not propagate the errors of pressure,
temperature, or dewpoint in either case. The formula for brake-specific emissions is:
m
e= —
W
where e is brake-specific emissions, m is total mass, and W is total work over a test interval.
For raw exhaust continuous sampling, the formula for total mass, m is:
41
-------
N
m = M\,nixiAt
2=1
where M is molar mass, N is the number of samples, ht is the exhaust rate, xt is the emissions
concentration, and At is the time interval between samples.
For laboratory testing the formula for total work is:
2=1
where coi is engine speed and Tt is engine torque.
For on-vehicle testing the formula for total work is:
N
MC E (XC02i + XCOi + XTHCi } "A*
w = —^
efUe!WC
where Mc is the molar mass of Carbon, TV is the number of samples, xc02i is the raw exhaust
concentration of carbon dioxide, xCOi is the raw exhaust concentration of carbon monoxide, xTHCi is
the raw exhaust concentration of total hydrocarbons, nt is the exhaust rate, At is the time interval
between samples, efuel is brake-specific fuel consumption, and wc is the mass fraction of carbon in
fuel.
To calculate the relative uncertainties, we first assume that all measured values and constants are set
to unity or one (1). This simplifies the equations without eliminating any uncertainty that needs to
be propagated, and it allows for one formula for error propagation regardless of the mathematical
operations of addition, multiplication, or division, which are the only operations in the formulas.
Namely, error is propagated by addition in quadrature, or in other words, the square root of the sum
of the squares of the error.
42
-------
For the on-vehicle case the brake-specific formula is:
z=l
Note that because nt is in the numerator and the denominator, it only has to be linearly proportional
to flow, not absolutely accurate or repeatable. Since there is no difference in the linearity
requirements between on-vehicle testing and laboratory testing, «. contributes no net error for
propagation in this case. At this point the summation signs can also be removed to further simplify
the formula. This will also be done for the laboratory case. Therefore the on-vehicle formula
becomes:
on—vehicle
I X^f, ,. ~T X^r-,^ ~T~ Xt
^C02i ' -"-COz ' -*THCi
Treating the laboratory case similarly:
hx
laboratory
0)T
Therefore, the on-vehicle uncertainty can be calculated as follows:
Oon-vehicle = JLOQ 2 + LOQ? + LOQ 2
°' on-vehicle = A/10%2 + 4.5%2 + 4.5%2
Van-vehicle =11-85%
Similarly, the laboratory uncertainty can be calculated as follows:
> laboratory = V^^^w ^ ^^x f ^^"^a + LOQT
O laboratory = V4.5%2 + 2%2 + 4%2 + 3%2
0'laboratory =7.02%
43
-------
Then the difference between these two values should be the allowance given for measurement error:
Con-VeMcle '^laboratory = 11.85% - 7.02% = 4.83% 5.00%
To apply this error allowance to a measured value as a multiplier, subtract this value from 1.00:
Error Allowance Multiplier = 1.00 - 0.05 = 0.95 or 95%.
This multiplier, 95 %, is the first proposed allowance, and we propose to apply it to every calculated
brake-specific emissions value for each and every individual NTE event.
44
-------
Chapter 3 References
1. S AE Technical Paper 2002-01-2711. Why Limit of Detection (LOD) Value is Not an Appropriate
Specification for Automotive Emissions Analyzers, Akard M., Tsurumi K., Oestergaard K. Horiba
Instruments.
45
-------
APPENDIX A: Examples of Determining the Number of Engine Families to be
Tested
This appendix contains a few examples showing how many engine families EPA may
designate for testing each year under the proposed in-use, manufacturer-run program. More
specifically, they illustrate how we would calculate the maximum annual number of engine families
for more complex cases where the proposed four-year average annual cap and 25 percent per year
limit might apply to a manufacturer with four or more engine families, all of which have annual
production volumes more than 1,500 units.
The multi-step methodology is identical regardless of number of engine families a
manufacturer has in these cases. The steps are discussed below and illustrated in Tables A-l
through A-3.
Step 1. For the calendar year in which we are testing (the evaluation year), identify the
number of engine families produced in the model year corresponding to that calendar year and
each of the three preceding model years.
Step 2. Divide by 4 the number of engine families produced in the model year corresponding
to the evaluation year and round the result to the nearest whole number using the rounding
convention specified by the National Institute of Standards and Technology (NIST 1995).2
The result is the 25 percent annual limit on the number of engine families that potentially may
be designated for testing in the evaluation year.
Step 3. Sum the engine families identified in Step 1 to determine the total four-year engine
family production.
Step 4. Divide by 4 the total four-year engine family production from Step 3 and round the
result to the nearest whole number using the rounding convention specified by the National
Institute of Standards and Technology (NIST 1995). The result is the four-year average cap
on the number of engine families that may be designated for testing in the evaluation year.
Step 5. Subtract the number of engine families we have required to be tested under this
program over the past three years from the four-year average cap. The result is the four-year
average annual cap.
Step 6. Select the lower of the rounded four-year average annual cap from Step 5 and the
rounded 25percent annual limit from Step 2. The result is the maximum number of engine
families that may be designated for testing in the evaluation year.
2 Under the rounding convention contained in this reference, when the first digit
discarded is exactly five, the last digit retained should be rounded upward if it is an odd number,
but no adjustment made if it is an even number.
46
-------
As noted, Tables A-l through A-3 illustrate the procedure described above. They also show
that the actual number of engines which may be designated in any year never exceeds the four-year
average annual cap, and may be less for some years, especially for the initial model year of testing,
i.e., 2007.
47
-------
Table A-l. Example of Engine Family Selection for Six or Seven Families Per Year
Inputs
Model Year
# Certified Families
4- Year Total Families (3
preceding years + current
year)
4- Year Ave. Cap (simple)
4-YearAve. Cap (NIST
811 rounded)
25% Annual Family Limit
(simple)
25% Annual Family Limit
(NIST 811 rounded)
04
6
05
6
06
6
07
6
24
6.00
6
1.50
2
08
6
24
6.00
6
1.50
2
09
7
25
6.25
6
1.75
2
10
7
26
6.5
6
1.75
2
11
7
27
6.75
7
1.75
2
12
7
28
7.00
7
1.75
2
13
7
28
7.00
7
1.75
2
14
7
28
7.00
7
1.75
2
15
6
27
6.75
7
1.50
2
Calculations
4 -Year Ave. Cap (from
above)
25% Annual Family Limit
(from above)
# Tests Allowed/Year
(4 Year Cap Applied?)
#Tested/4 -Year Ave.
Annual Cap
(% Tested)
6
2
2
(no)
2/6
(33)
6
2
2
(no)
4/6
(67)
6
2
2
(no)
6/6
(100)
6
2
0
(yes)
6/6
(100)
7
2
2
(no)
6/7
(86)
7
2
2
(no)
6/7
(86)
7
2
2
(no)
6/7
(86)
7
2
1
(yes)
111
(100)
7
2
2
(no)
111
(100)
48
-------
Table A-2. Example of Engine Family Selection for 10 Families Per Year
Inputs
Model Year
# Certified Families
4-Year Total Families (3
preceding years + current
year)
4-Year Ave. Cap (simple)
4-Year Ave. Cap (NIST
811 rounded)
25% Annual Family
Limit (simple)
25% Annual Family
Limit (NIST 811
rounded)
04
10
05
10
06
10
Calculations
4-Year Ave. Cap (from
above)
25% Annual Family
Limit (from above)
# Tests Allowed/Year
(4-Year Cap Applied?)
# Tested/4-Year Ave.
Annual Cap
(% Tested)
07
10
40
10
10
2.5
2
08
10
40
10
10
2.5
2
09
10
40
10
10
2.5
2
10
10
40
10
10
2.5
2
11
10
40
10
10
2.5
2
12
10
40
10
10
2.5
2
13
10
40
10
10
2.5
2
14
10
40
10
10
2.5
2
15
10
40
10
10
2.5
2
10
2
2
(no)
2/10
(20)
10
2
2
(no)
4/10
(40)
10
2
2
(no)
6/10
(60)
10
2
2
(no)
8/10
(80)
10
2
2
(no)
8/10
(80)
10
2
2
(no)
8/10
(80)
10
2
2
(no)
810
(80)
10
2
2
(no)
8/10
(80)
10
2
2
(no)
8/10
(80)
49
-------
Table A-3. Example of Engine Family Selection for 9, 10, or 14 Families Per Year
Inputs
Model Year
# Certified Families/Year
4- Year Total Families (3
preceding years + current
year)
4- Year Ave. Cap (simple)
4-YearAve. Cap (NIST
811 rounded)
25% Annual Family Limit
(simple)
25% Annual Family Limit
(NIST 811 rounded)
04
9
05
9
06
9
07
9
36
9.00
9
2.25
2
08
9
36
9.00
9
2.25
2
09
10
37
9.25
9
2.50
2
10
10
38
9.50
10
2.50
2
11
10
39
9.75
10
2.50
2
12
10
40
10.00
10
2.50
2
13
14
44
11.00
11
3.50
4
14
14
48
12.00
12
3.50
4
15
14
52
13.00
13
3.50
4
Calculations
4- Year Ave. Cap (from
above)
25% Annual Family Limit
(from above)
# Tests Allowed/Year
(4-Year Cap Applied?)
# Tested/4-Year Ave.
Annual Cap
(% Tested)
9
2
2
(no)
2/9
(22)
9
2
2
(no)
4/9
(44)
9
2
2
(no)
6/9
(67)
10
2
2
(no)
8/10
(80)
10
2
2
(no)
8/10
(80)
10
2
2
(no)
8/10
(80)
11
4
4
(no)
10/11
(91)
12
4
4
(no)
12/12
(100)
13
4
o
J
(yes)
13/13
(100)
50
-------
Appendix References
1. National Institute of Standards and Technology. 1995. Guide for the Use of the International
System of Units (SI), NIST Special Publication 811, 1995 Edition. U. S. Department of Commerce.
51
------- |