Edmond_
Toy@
omb.
eop.
gov
To:
Mary
Kissell/
RTP/
USEPA/
US@
EPA
cc:
Arthur_
G._
Fraas@
omb.
eop.
gov,
William_
Nickerson@
omb.
eop.
gov
Subject:
Re:
follow­
up
11/
06/
03
11:
56
AM
Hi
Mary
Tom,

Thanks
for
sending
us
those
write­
ups...
I'm
still
going
through
them.

In
the
meantime,
I
wanted
to
get
back
to
you
on
the
issue
of
averaging
with
PBCO.
Based
on
your
example,
it
is
still
not
clear
why
averaging
between
regulated
units
should
not
be
allowed.

|­­­­­­­­­­­­­­­­­­­­­­­­­­+­­­­­­­­­­­­­­+­­­­­­­­­­­­­­|
|
|
Press
|
Strand
dryer
|
|­­­­­­­­­­­­­­­­­­­­­­­­­­+­­­­­­­­­­­­­­+­­­­­­­­­­­­­­|
|
Production
volume
|
100,000
MSF|
65,000
ODT|
|­­­­­­­­­­­­­­­­­­­­­­­­­­+­­­­­­­­­­­­­­+­­­­­­­­­­­­­­|
|
Baseline
emission
rate
|
3.0
lb/
MSF|
1.8
lb/
ODT|
|
without
any
intervention
|
|
|
|
(
business
as
usual,
no
|
|
|
|
P2,
no
add­
on
controls,
|
|
|
|
etc).
|
|
|
|­­­­­­­­­­­­­­­­­­­­­­­­­­+­­­­­­­­­­­­­­+­­­­­­­­­­­­­­|
|
PBCO
emission
standard
|
0.3
lb/
MSF|
0.18
lb/
ODT|
|
(
90%
reduction
from
|
|
|
|
baseline)
|
|
|
|­­­­­­­­­­­­­­­­­­­­­­­­­­+­­­­­­­­­­­­­­+­­­­­­­­­­­­­­|
|
Aggregate
emissions
after|
30,000
lbs.|
11,700
lbs.|
|
meeting
PBCO
|
|
|
|
(
std*
production)
|
|
|
|­­­­­­­­­­­­­­­­­­­­­­­­­­+­­­­­­­­­­­­­­+­­­­­­­­­­­­­­|
|
Total
aggregate
emissions|
41,700
lbs.
|
|
|­­­­­­­­­­­­­­­­­­­­­­­­­­+­­­­­­­­­­­­­­+­­­­­­­­­­­­­­|
|
|
|
|
|­­­­­­­­­­­­­­­­­­­­­­­­­­+­­­­­­­­­­­­­­+­­­­­­­­­­­­­­|
|
With
averaging
between
|
|
|
|
regulated
units
|
|
|
|­­­­­­­­­­­­­­­­­­­­­­­­­­+­­­­­­­­­­­­­­+­­­­­­­­­­­­­­|
|
Controlled
emission
rate
|
0.25
lb/
MSF|
0.2
lb/
ODT|
|
(
e.
g.,
innovative
P2
|
|
|
|
measures)
|
|
|
|­­­­­­­­­­­­­­­­­­­­­­­­­­+­­­­­­­­­­­­­­+­­­­­­­­­­­­­­|
|
Aggregate
emissions
|
25,000
lbs.|
13,000
lbs.|
|
(
controlled
emissions
|
|
|
|
rate
*
production)
|
|
|
|­­­­­­­­­­­­­­­­­­­­­­­­­­+­­­­­­­­­­­­­­+­­­­­­­­­­­­­­|
|
Total
aggregate
emissions|
38,000
lbs.
|
|
|­­­­­­­­­­­­­­­­­­­­­­­­­­+­­­­­­­­­­­­­­+­­­­­­­­­­­­­­|

With
averaging
between
regulated
units,
the
firm
has
the
incentive
to
innovate
and
search
for
cost­
effective
ways
to
reduce
emissions.
In
this
example,
the
firm
finds
that
it
is
cheaper
to
concentrate
its
efforts
on
emissions
from
its
press,
overcomplying
with
the
PBCO
while
undercomplying
with
the
PBCO
for
the
strand
dryer.
The
example
shows
that
the
total
aggregate
emissions
are
lower
than
without
averaging
(
38,000
vs.
41,700
lbs).
This
reduction
in
total
aggregate
emissions
may
not
actually
occur,
since
the
firm
has
no
incentive
to
undercomply
in
the
aggregate,
so
the
firm
may
decide
to
control
its
strand
dryer
to
a
lesser
extent
(
e.
g.,
instead
of
0.2
lb/
ODT,
it
could
emit
0.257
lb/
ODT,
thereby
using
up
all
of
the
"
credits"
generated
by
the
press
and
leading
to
a
total
aggregate
emissions
level
of
41,700
lbs).

Why
shouldn't
the
firm
be
allowed
to
average
across
regulated
units
(
if
it
is
cost­
effective
to
do
so)
if
the
total
aggregate
emissions
from
regulated
units
are
at
least
as
low
compared
to
the
case
with
no
averaging?

We
should
talk
about
this
more.
Also,
there
are
other
types
of
averaging
that
we
may
need
to
discuss
(
averaging
between
both
regulated
and
non­
regulated
units
within
PBCO...
by
non­
regulated
I
mean
those
without
MACT
floors;
also
what
can
generate
credits
in
the
EA
approach­­
which
Art
brought
up
when
you
came
to
brief
us).

­
Edmond
Kissell.
Mary@
epamail.
epa.
gov
11/
03/
2003
03:
10:
23
PM
To:
Edmond
Toy/
OMB/
EOP
cc:
See
the
distribution
list
at
the
bottom
of
this
message
Subject:
Re:
follow­
up
Edmond,

You
had
asked
some
questions
and
here
are
our
responses.
Hope
these
are
helpful
to
you
and
cover
all
the
follow­
up
we
had
from
our
two
meetings
and
one
telephone
conversation.
Mary
Tom,
(
919)
541­
4516
1.
RTO
and
RCO.
Regenerative
thermal
oxidizers
(
RTO)
combust
gas
and
recover
some
energy.
RTOs
incorporate
ceramic
beds
at
the
inlet
and
outlet
of
the
combustion
chamber
for
up
to
95
percent
energy
recovery.
The
primary
difference
between
catalytic
oxidation
systems
and
thermal
oxidation
systems
is
that
catalytic
systems
use
a
catalyst
to
increase
the
rate
of
the
combustion
reaction
by
allowing
oxidation
to
occur
at
lower
operating
temperatures.

2.
Cost
savings
in
an
emissions
averaging
standard
that
requires
add­
on
controls
to
generate
emission
credits.
Emissions
averaging
(
EA)
in
the
PCWP
MACT
has
real
cost­
saving
opportunities.
There
are
four
ways
EA
can
be
used
to
reduce
costs­­(
1)
control
only
a
portion
of
a
regulated
unit
with
MACT
controls
(
e.
g.,
with
onsite
incineration
or
with
a
smaller
RTO/
RCO/
BIO);
(
2)
control
a
lower
flow,
more
concentrated
unregulated
process
unit
(
e.
g.,
blender,
former,
atmospheric
refiner);
(
3)
control
an
unregulated
process
unit
that
is
easier
to
control
(
lower
moisture,
lower
particulate);
and,
(
4)
use
a
non­
MACT
control
device
(
e.
g.,
the
AAT
adsorption/
recovery
device)
that
achieves
less
than
90
percent
control.
The
first
two
options
save
money
because
you
are
controlling
a
lower
flow
rate
stream
and
lower
flow
rate
means
the
facility
can
use
a
smaller
(
and
presumably
less
costly)
add­
on
control.
Or
the
facility
may
be
able
to
route
the
lower
flow
stream
to
an
onsite
combustion
unit
(
which
are
usually
limited
in
the
volume
of
emissions
they
can
treat).
The
third
option
saves
money
because
lower
flow
is
not
the
only
factor
affecting
control
costs;
for
example,
it's
cheaper
to
control
a
stream
with
low
or
no
particulate
than
it
is
to
control
one
that
has
a
lot
of
particulate
(
e.
g.,
direct­
fired
OSB
dryer)
since
you
may
only
need
the
HAP
control
device
(
e.
g.,
RTO)
and
not
the
up­
front
particulate
controls
(
e.
g.
WESP).
The
same
is
true
regarding
moisture
content­­
a
lower
moisture
content
stream
is
more
cost­
effective
to
incinerate.
Also,
EA
would
allow
sources
to
use
a
less
effective
control
device
and
some
of
these
device
may
be
cheaper.

3.
Production­
based
compliance
option
(
PBCO)
and
generating
emissions
credits.
We
had
discussed
whether
a
process
unit
that
meets
a
PBCO
limit
and
is
under
the
limit
could
generate
credits
and
use
those
credits
to
achieve
the
PBCO
for
another
process
unit.
Allowing
a
facility
to
apply
the
unused
portion
of
a
PBCO
from
one
process
unit
to
a
second
process
unit,
so
that
the
second
process
unit
also
achieve
PBCO
is
less
than
MACT.
I
do
not
recommend
this.
Below
is
an
example.

Press
PBCO:
0.30
lb/
MSF
3/
4"
Example
press
processes
100,000
MSF
3/
4"
per
year
and
emits
0.25
lb/
MSF
3/
4"
Difference
in
press
PBCO
and
emissions
is
0.30
­
0.25
=
0.05
lb/
MSF
3/
4"
Difference
x
throughput
(
PBCO
credit)
is
100,000
x
0.05
=
5,000
lb/
yr
Strand
dryer
PBCO:
0.18
lb/
ODT
Example
strand
dryer
processes
65,000
ODT/
yr
and
emits
0.20
lb/
ODT
Uncontrolled
strand
dryer
emissions:
65,000
x
0.20
=
13,000
lb/
yr
Emissions
allowed
under
PBCO:
65,000
x
0.18
=
11,700
lb/
yr
Reduction
to
get
under
PBCO:
13,000
­
11,700
=
1,300
lb/
yr
Reduction
by
90%
as
required
by
MACT:
0.9
x
13,000
=
11,700
lb/
yr
The
1,300
lb/
yr
PBCO
debit
generated
by
the
strand
dryer
could
be
offset
by
the
5,000
lb/
yr
press
PBCO
credit;
however,
MACT
would
have
required
11,700
lb/
yr
reduction
for
the
strand
dryer
emissions
instead
of
1,300
lb/
yr
reduction.

4.
Model
plant
parameters
used
in
EPA
risk
assessments.
Attached
are
two
memos
documenting
how
we
estimated
baseline
organic
HAP
emissions
and
stack
parameters
for
PCWP
facilities.
These
estimates
were
used
as
inputs
to
EPA's
rough
risk
analysis.
In
short,
the
baseline
emissions
were
based
on
emission
factors
(
lb
HAP/
production
rate)
developed
from
emission
test
data.
Emission
factors
were
developed
for
each
HAP
and
for
different
process
unit
types.
The
emission
factors
were
multiplied
by
process
unit
production
rate
to
estimate
emissions
for
each
process
unit.
Facility
emissions
were
determined
by
summing
emissions
estimated
for
each
process
unit
at
the
facility.
Information
of
the
numbers
and
types
of
process
units
and
their
respective
production
rates
was
collected
through
our
1998
survey.
We
were
not
conservative
with
our
baseline
emission
estimates;
we
simply
performed
the
calculations
using
what
the
data
showed.

The
stack
parameters
were
based
on
far
less
information
than
the
baseline
emission
estimates
and
incorporate
engineering
judgement.
We
did
not
collect
data
on
stack
parameters
or
distance
to
the
fenceline
in
our
survey
because
we
never
envisioned
needing
this
information.
Typical
process
unit
stack
heights,
diameters,
and
exit
gas
velocities
were
compiled
from
several
title
V
permits
and
emission
test
reports.
If
data
were
not
available
for
one
type
of
process
unit,
data
were
used
either
from
a
similar
process
unit
or
from
an
average
of
similar
process
units
retrieved
from
title
V
permits
or
emission
test
reports.
The
result
was
different
stack
parameters
for
different
types
of
process
units.
For
the
risk
analysis,
we
assumed
two
stacks
with
one
stack
representing
one­
third
of
the
emissions
from
the
facility
and
having
stack
parameters
similar
to
PCWP
press
stack
parameters.
The
other
stack
was
assigned
two­
thirds
of
the
facility
emissions
and
"
dryer"
stack
parameters.
The
model
"
dryer"
parameters
do
not
consider
whether
the
dryer
is
an
OSB
rotary
dryer,
MDF
tube
dryer,
or
veneer
dryer.
The
two­
stack
assumption
for
the
rough
risk
analysis
may
be
very
different
from
reality.
In
reality,
PCWP
plants
have
more
than
two
stacks
and
the
stack
parameters
can
vary
from
stack
to
stack.
For
example,
veneer
dryers
have
multiple
stacks
per
dryer,
with
multiple
veneer
dryers
per
facility
in
addition
to
the
press.
We
did
not
attempt
to
make
conservative
assumptions
with
regard
to
stack
parameters
as
inputs
for
the
rough
risk
analysis.
We
used
the
typical
stack
parameters.
These
stack
parameter
assumptions
used
to
create
the
model
plant
are
good
enough
for
generating
impacts,
but
are
too
rough
to
be
the
basis
of
a
facility­
specific
low­
risk
determination.

Also,
in
looking
into
this
information
for
you,
we
realized
that
we
had
an
input
error
for
the
stack
diameters.
The
model
was
run
with
a
stack
diameter
of
2.5
feet
and
should
have
been
5
ft
for
presses
and
6
ft
for
dryers.
We're
uncertain
how
this
will
affect
the
impacts.
We
are
re­
running
the
model
to
see
how
it
changes
the
results.
There
was
also
a
minor
discrepancy
with
temperature
(
10o
Kelvin
difference),
which
we
don't
think
will
change
the
results.

5.
HAP
in
EPA
risk
assessments.
See
attached
table
for
comparison.

(
See
attached
file:
Baseline.
wpd)(
See
attached
file:
Stack
Parameter
Documentation.
wpd)(
See
attached
file:
PCWP
HAPs
110303.
wpd)

Edmond_
Toy@
omb.
eop.
gov
To:
Mary
Kissell/
RTP/
USEPA/
US@
EPA
cc:
Subject:
follow­
up
10/
29/
03
04:
20
PM
Hi
Mary,

From
our
conversation
on
Monday,
there
were
two
items
still
outstanding.
First,
you
mentioned
that
you
would
check
into
the
various
parameters
used
in
the
risk
assessment,
eg,
do
the
assumed
stack
heights
represent
the
average?
median?
Second,
a
write­
up
that
describes
the
emissions
data
(
from
where,
how
often
are
different
data
sources
used,
are
they
measured
levels,
how
representative,
etc).

Also,
could
you
list
the
HAP
that
were
included
in
the
updated
risk
assessment
(
the
ones
not
included
in
the
analysis
at
proposal)?

Please
let
me
know.

Thanks.

­
Edmond
Message
Copied
To:_____________________________________________________________

thrift.
mike@
epamail.
epa.
gov
Chamberlin.
Johnb@
epamail.
epa.
gov
Throwe.
Scott@
epamail.
epa.
gov
Craig.
Jeneva@
epamail.
epa.
gov
Pagano.
Dennis@
epamail.
epa.
gov
Durkee.
Stan@
epamail.
epa.
gov
(
See
attached
file:
pic32142.
pcx)
