To
EMA
February
14,
2005
Measurement
Accuracy
Margin
for
the
Heavy­
Duty
In­
Use
Testing
Final
Rule
EPA
Staff
Working
Document
Background:
The
settlement
outline
states
that
new
measurement
accuracy
margins
for
the
heavy­
duty
in­
use
testing
(
HDIUT)
program
will
be
jointly
determined
by
EPA,
CARB,
and
the
engine
manufacturers.
The
purpose
of
these
margins
is
to
account
for
the
uncertainty
associated
with
portable
emissions
sampling
systems
incremental
to
laboratory
performance.
EPA
maintains
that
the
Final
Rule
needs
to
include
"
hard­
wired"
accuracy
margins
for
all
pollutants.
These
are
at
a
minimum
needed
for
the
"
pilot
programs."
EPA
has
indicated
that
it
would
provide
"
liberal"
values
for
these
interim
accuracy
margins
and
manufacturers
have
indicated
their
agreement
with
these
margins.
1
Further,
in
meetings
on
this
subject,
manufacturers
have
indicated
that
more
data
is
needed
to
establish
the
accuracy
margins
for
the
2007
fully
enforceable
program
for
gaseous
emissions
and
a
2008
fully
enforceable
program
for
particulate
matter
(
PM).
EPA
also
believes
that
more
data
would
be
helpful
in
determining
the
final
accuracy
margins
and
is
willing
to
engage
in
a
program
to
develop
that
information.
This
paper
identifies
how
the
data­
driven
accuracy
margins
can
be
developed
and
how
they
would
be
implemented
by
a
direct
final
rulemaking.
It
also
describes
the
consequences
of
failing
to
complete
the
program
to
define
data­
driven
accuracy
margins
in
time
to
start
either
of
the
enforceable
programs.

Program
Development:
EPA
will
develop
the
basic
scope
and
objectives
of
the
program
to
derive
the
data­
driven
margins
with
the
participation
of
the
engine
manufacturers.
It
would
be
highly
desirable
for
manufacturers
to
be
objectively
involved
in
the
design
of
the
program
given
the
scope
and
importance
of
the
project.
(
EPA
expects
that
each
manufacturer
would
have
one
representative
to
assist
in
developing
the
program.)
The
programmatic
objectives,
approach,
relative
roles
of
EPA
and
the
engine
manufacturers,
and
any
other
points
of
agreement
necessary
for
a
successful
program
would
be
reflected
in
a
Memorandum
of
Understanding
(
MOU)
or
similar
document.

More
specifically,
the
MOU
would
address
the
development
of
both
gaseous
and
PM
accuracy
margins.
For
gaseous
emissions,
the
document
would
explicitly
detail
the
objectives,
testing,
data
analysis,
and
commitments
from
each
party
including
cost
sharing,
payments­
in­
kind,
or
other
possible
contributions.
2
The
MOU
would
contain
a
hard
funding
limit,
i.
e.,
not
to
exceed
value,
for
the
gaseous
emissions
test
program.
For
PM,
the
document
would
specify
that
the
parties
agree
to
develop
a
detailed
testing
and
data
analysis
plan
(
such
as
is
now
underway
for
the
gaseous
pollutants)
within
two
months
of
signing
the
MOU.
The
document
would
also
explicitly
detail
for
PM
as
many
of
the
program
objectives,
scope,
and
commitments
from
each
party
as
possible,
including
cost
sharing,
payments­
in­
kind,
or
other
possible
contributions.
Furthermore,
the
MOU
would
specify
a
PM
test
plan
similar
in
approach
and
level
of
testing
as
that
used
to
develop
the
gaseous
emission
accuracy
margins.
The
PM
portion
of
the
MOU
would
also
contain
a
hard
funding
limit,
i.
e.,
not
to
exceed
value,
for
the
test
program.
Finally,
the
document
would
contain
an
agreement
that
the
parties
would
not
contest
the
data­
driven
1
The
additive
interim
accuracy
margins
(
g/
bhp­
hr)
are
as
follows:
NMHC
=
0.17;
NOx
=
0.5;
CO
=
0.60;
and
PM
=
0.10.
2
As
currently
envisioned
for
the
gaseous
test
program,
the
primary
contribution
by
all
parties
will
be
money.
The
total
dollar
cost
would
be
split
equally
between
industry
and
government.
Manufacturers
will
also
contribute
test
engines,
which
will
be
valued
as
part
of
their
monetary
contribution.
To
EMA
February
14,
2005
2
gaseous
and
PM
results
so
long
as
they
are
derived
from
the
agreed
upon
test
program
and
the
application
of
the
agreed
upon
analytical
methods.

Objective,
Approach,
and
Accommodations:
The
margins
will
compensate
for
the
difference
between
in­
use
emission
measurement
accuracy
and
laboratory
emission
measurement
accuracy.
Individual
margins
will
be
established
for
NOx,
NMHC,
CO,
and
PM.
This
program
is
not
intended
to
determine
potential
accuracy
allowances
for
laboratory
testing
at
the
levels
of
the
2007/
2010
emission
standards.
The
margins
determined
by
this
program
will
be
established
by
a
rigorous
in­
field
and
laboratory
test
program
that
assesses,
at
a
minimum,
lab
to
PEMS
emission
differences
and
accounts
for
in­
use
environmental
effects
on
PEMS
units.
The
overall
scope,
experimental
design,
and
analytical
methodology
for
determining
each
accuracy
margin
would
be
agreed
upon
well
in
advance
of
initiating
the
test
program.
The
test
programs
for
gaseous
and
PM
pollutants
would
ideally
be
integrated
and
sequenced
into
the
overall
test
plan
to
provide
final
results
in
time
for
the
fully
enforceable
programs
described
above.
Nonetheless,
in
no
case
should
work
on
developing
the
gaseous
accuracy
margins
be
delayed
while
a
work
plan
for
the
PM
test
program
is
being
completed.
It
is
imperative
that
the
test
program
for
gaseous
emissions
is
initiated
and
completed
quickly
to
avoid
delaying
the
enforceable
program
that
begins
in
2007.

All
testing
and
subsequent
data
analysis
will
be
overseen
by
EPA
and
coordinated
with
the
engine
manufacturers.
It
is
anticipated
that
most,
if
not
all,
of
the
testing
and
data
analysis
will
be
performed
under
contract
by
an
agreed
upon
third
party
or
parties.
Furthermore,
it
is
anticipated
that
the
experimental
designs
will
be
submitted
to
a
statistician
to
ensure
that
the
information
generated
is
sufficient
and
appropriate
for
developing
the
accuracy
margins,
and
to
establish
an
algorithm
for
calculating
the
margins.
This
overall
approach
will
provide
assurances
that
the
experimental
designs
are
acceptable
to
all
participants
and
is
intended
to
prevent
multiple
interpretations
of
the
data­
driven
results
so
that
the
margins
can
be
implemented
by
direct
final
rulemaking.
It
will
also
help
ensure
that
the
program
will
be
completed
on
schedule
and
budget.
(
Although
EPA
intends
to
implement
data­
driven
accuracy
margins
by
direct
final
rule,
the
Agency
cannot
guarantee
that
the
test
program
results
would
end
up
in
a
final
rule
if
significant
adverse
comments
are
received.)

Finally,
EPA
understands
that
participation
by
manufacturers
in
the
test
programs
will
result
in
some
burden
to
those
companies,
but
will
produce
useful
information.
Therefore,
if
manufacturers
agree
to
participate
in
the
test
programs,
EPA
would
reduce
the
manufacturers'
testing
burden
under
the
original
pilot
program
(
2005­
2006
for
gaseous
and
2006­
2007
for
PM)
to
five
vehicles
per
designated
engine
family.

Schedule:
In
order
to
provide
adequate
time
to
promulgate
the
data­
driven
accuracy
margins,
the
test
programs
will
generally
need
to
conform
to
the
following
schedule.
First,
the
MOU
containing
the
experimental
design,
analytical
methods,
and
resource
commitments
for
determining
the
accuracy
margins
should
be
signed
by
senior
management
of
all
parties
no
later
than
March
15,
2005.
Second,
the
subsequent
detailed
test
and
data
analysis
plan
for
PM
should
be
signed
by
senior
management
of
all
parties
no
later
than
May
15,
2005.
Third,
the
test
programs
should
be
completed
and
final
accuracy
margins
available
by
September
30,
2006
for
gaseous
emissions
and
by
September
30,
2007
for
PM.
To
EMA
February
14,
2005
3
Regulatory
Provision:
If
it
is
determined
that
the
test
programs
as
described
above
are
an
appropriate
mechanism
for
determining
final
accuracy
margins,
the
basic
programmatic
objectives
and
approach,
and
the
relative
roles
of
the
EPA
and
the
engine
manufacturers
would
be
reflected
in
the
HDIUT
final
rule.
The
key
provisions
of
the
MOU,
such
as
the
contributions
and
deliverables
for
each
party,
would
be
referenced
and
to
the
degree
possible
contained
in
the
preamble
of
the
rule.
Certain
key
provisions
would
also
be
reflected
in
the
regulatory
language
as
a
means
to
provide
certainty
to
all
parties.
These
provisions
would
include
items
such
as
the
pilot
program
provisions
and,
as
appropriate,
the
consequences
of
failing
to
complete
the
accuracy
margin
test
programs
as
specified
below.
The
accuracy
margin
algorithm
may
also
be
included
in
the
regulatory
text.

The
MOU
will
address
a
test
program
that
is
designed
to
provide
the
data
(
and
associated
data
analysis)
needed
to
derive
the
final
data­
driven
accuracy
margins
by
September
30,
2006
for
gaseous
emissions
and
September
30,
2007
for
PM.
The
MOU
will
specify
key
interim
deliverable
milestones.
It
will
also
specify
that
the
final
data­
driven
accuracy
margins
will
be
documented
in
a
final
report
from
the
contractor
as
soon
as
practicable,
but
in
no
case
later
December
1,
2006
for
gaseous
and
December
1,
2007
for
PM.
All
parties
will
agree
to
use
these
accuracy
margin
values
for
program
planning
and
implementation
of
the
respective
enforceable
program.
The
data­
driven
accuracy
margins
will
be
proposed
by
EPA
in
a
direct
final
rule
as
soon
as
possible
after
the
final
values
and
documentation
are
available.
The
MOU
will
acknowledge
that
publishing
the
direct
final
rule
with
the
data­
driven
margins,
or
if
necessary
publishing
a
final
rule
based
on
the
data­
driven
test
results
due
to
public
comment,
would
fulfill
EPA's
obligation
under
the
settlement
agreement.

If
a
failure
to
obtain
the
final
accuracy
values
and
documentation
from
the
contractor
by
the
December
1
date
is
the
fault
of
the
manufacturers,
and
the
delay
is
less
than
a
total
of
3
months,
the
enforceable
program
will
be
delayed
by
the
same
amount
of
time
it
takes
to
complete
and
finalize
the
final
contractor
report,
i.
e.,
a
one­
to­
one
correspondence.
3
If
the
final
values
and
documentation
are
delayed
beyond
3
months,
the
enforceable
program
will
go
into
abeyance
for
the
calendar
year
(
2007
for
gaseous
emissions
or
2008
for
PM).
However,
the
number
of
engine
families
that
would
otherwise
have
been
designated
for
testing
if
the
enforceable
program
was
not
delayed
will
be
accumulated
and
may
be
designated
for
testing
when
the
enforceable
program
is
initiated
without
counting
toward
the
allowable
annual
cap
on
the
number
of
engine
families
that
may
be
designated
for
in­
use
testing
in
that
year.
The
normal
18­
month
period
for
testing
and
reporting
would
be
expanded
to
24
months
for
such
"
carryover"
engine
families.
If
necessary,
this
cycle
will
be
repeated
until
the
final
accuracy
margins
are
identified
and
documented
in
a
final
report.
4
A
manufacturers'
failure
may
be
demonstrated
by
missing
one
or
more
of
the
critical
milestones
as
follows:

­­
Torque/
BSFC
v.
speed/
load
and
AECD
maps
delivered
to
contractor
at
time
of
contract
signature
3
The
anticipated
June
30
date
for
designating
engine
families,
which
initiates
the
18­
month
testing/
reporting
period,
would
be
delayed.
4
If
the
cycle
is
repeated,
an
additional
six
months
will
be
added
to
the
testing/
reporting
period
for
each
year
the
enforceable
program
is
delayed.
To
EMA
February
14,
2005
4
­­
Contracts
funded
and
signed
within
60
days
after
MOU
signature
­­
Engines
delivered
to
contractor
within
90
days
after
MOU
signature
A
failure
by
any
individual
manufacturer
that
leads
to
missing
one
or
more
of
the
critical
milestones
described
above
would
be
regarded
as
a
failure
by
all
manufacturers.

If
a
failure
to
obtain
the
final
accuracy
values
and
documentation
from
the
contractor
by
the
December
1
date
is
the
fault
of
EPA
or
a
party
other
than
the
manufacturers,
and
the
delay
is
less
than
a
total
of
3
months,
the
enforceable
program
will
be
delayed
by
the
same
amount
of
time
it
takes
to
complete
and
finalize
the
final
contractor
report,
i.
e.,
a
one­
to­
one
correspondence.
5
If
the
final
values
and
documentation
are
delayed
beyond
3
months,
the
Phase
1
and
2
pilot
programs
will
be
implemented
for
that
year
using
the
interim
accuracy
margins.
If
necessary,
this
cycle
will
be
repeated
until
the
final
accuracy
margins
are
identified
and
documented
in
a
final
report.

Regardless
of
the
party
at
fault,
if
the
enforceable
program
is
delayed,
2007
and
subsequent
model
year
engines,
as
appropriate,
may
be
selected
and
subject
to
enforcement
testing
(
for
gaseous
as
well
as
PM
emissions)
once
the
accuracy
margins
are
established
and
the
enforcement
program
begins.
Such
engines
would
be
counted
toward
the
total
number
of
engine
families
that
may
be
designated
for
in­
use
testing
in
any
calendar
year.
In
the
case
of
a
manufacturers'
failure,
the
total
number
of
engine
families
would
be
equal
to
the
engine
families
that
would
otherwise
have
been
designated
for
testing
if
the
enforceable
program
was
not
delayed,
plus
those
designated
for
testing
under
the
annual
cap
on
the
number
of
engine
families
that
may
be
designated
for
in­
use
testing
when
the
enforceable
program
begins.
In
the
case
of
a
failure
by
EPA
or
a
party
other
than
the
manufacturers,
the
total
number
of
engine
families
would
be
equal
to
the
annual
cap
on
the
number
of
engine
families
that
may
be
designated
for
in­
use
testing
in
the
calendar
year
when
the
enforceable
program
begins.

If
fundamental,
irresolvable
technical
problems
are
identified
relative
to
the
PM
portable
emission
measurement
systems,
the
PM
pilot
program
or
enforceable
program
goes
into
abeyance
until
such
time
as
suitable
emission
measurement
devices
are
identified
or
the
problems
otherwise
resolved.
Again,
if
the
enforceable
program
is
delayed,
the
applicable
2007
and
subsequent
model
year
engines
may
be
selected
and
subject
to
enforcement
testing
(
for
gaseous
as
well
as
PM
emissions)
once
the
accuracy
margins
are
established
and
the
enforcement
program
begins.
Such
engines
would
be
counted
toward
the
annual
cap
on
the
number
of
engine
families
that
may
be
designated
for
in­
use
testing
in
any
calendar
year.

5
The
anticipated
June
30
date
for
designating
engine
families,
which
initiates
the
18­
month
testing/
reporting
period,
would
be
delayed.
