MEMORANDUM
Date:
August
1,
2002
Subject:
Analysis
of
Test
Data
for
Columbia
Falls
Aluminum
Company
(
CFAC)

From:
Steve
Fruh,
PPSG/
ESD/
OAQPS/
EPA
and
Marvin
Branscome,
RTI
To:
Docket
ID
No.
OAR­
2002­
0031
The
attached
report
is
an
analysis
of
test
data
for
polycyclic
organic
matter
(
POM)
as
measured
by
EPA
Method
315
for
the
vertical
stud
Soderberg
two
(
VSS2)
subcategory
of
potlines.
These
data
were
collected
at
CFAC
in
Columbia
Falls,
Montana
from
testing
in
1996
through
2000
by
Amtest
Air
Quality,
LLC.
1
C.
A.
Laughner,
(
Browning,
Kaleczyc,
Berry
&
Hoven,
P.
C.
for
Columbia
Falls
Aluminum)
to
C.
T.
Whitman,
U.
S.
EPA.
Request
for
Extension
of
POM
Emission
Limit
for
Existing
VSS2
Aluminum
Potlines
and
Rulemaking
Request
to
Revise
40
CFR
63.843(
a)(
2)(
iii).
May
10,
2001.
Docket
Item
VII­
A­
1in
Docket
ID
No.
OAR­
2002­
0031.

1
ANALYSIS
OF
POLYCYCLIC
ORGANIC
MATTER
(
POM)
DATA
FOR
VSS2
POTLINES
I.
Purpose
The
purpose
of
this
report
is
to
analyze
POM
data
for
the
VSS2
subcategory
and
determine
the
average
emission
limitation
achieved
by
the
top
five
performing
potlines.
These
data
were
submitted
by
Columbia
Falls
Aluminum
Company
(
CFAC)
with
a
request
that
the
POM
emission
limits
be
revised.
1
The
testing
was
performed
to
determine
fugitive
or
secondary
emissions
that
escape
the
primary
capture
and
control
system.

II.
Approach
The
technical
approach
is
the
same
as
that
used
to
develop
the
original
emission
limits
 
determine
the
upper
99th
percentile
of
performance
for
the
top
five
performing
potlines
for
secondary
emissions
and
add
the
small
contribution
from
the
primary
control
system.
The
99th
percentile
represents
a
level
that
has
been
demonstrated
as
achievable
on
a
continuing
basis
and
accounts
for
the
normal
variability
in
performance
over
time.
For
the
original
analysis,
data
were
available
for
only
two
potlines.
Consequently,
it
was
not
possible
at
that
time
to
determine
the
control
level
achieved
by
the
top
five
potlines.
The
additional
data
significantly
expand
the
database
available
to
evaluate
the
level
of
control
achieved
and
now
includes
data
for
the
top
five
potlines.

III.
Results
The
analysis
evaluated
all
POM
data
collected
by
Amtest
for
VSS2
potlines
from
1996
to
2000
(
a
total
of
88
runs).
The
results
are
summarized
in
Table
1.

TABLE
1.
TOP
FIVE
PERFORMING
VSS2
POTLINES
(
1996
to
2000)

Potline
Average
POM
(
lb/
ton)
Number
of
runs
4
2.40
24
9
2.82
27
5
3.33
14
1
3.41
12
7
3.94
11
Average
of
top
five
3.18
88
The
99th
percentile
is
calculated
from:
2
99th
percentile
=
average
+
z
*
standard
deviation
+
primary
control
system
emissions
where
average
=
overall
average
of
the
five
potlines
(
3.18
lb/
ton),
z
=
2.326
for
the
99th
percentile,
standard
deviation
=
1.03
lb/
ton
based
on
the
averages
of
three
runs,
and
primary
control
system
emissions
=
0.1
lb/
ton.

Substituting
into
the
equation
yields
a
value
of
5.7
lb/
ton:

99th
percentile
=
3.18
+
2.326
*
1.03
+
0.1
=
5.7
lb/
ton.

The
emission
limits
for
emission
averaging
are
calculated
by
dividing
the
standard
deviation
by
the
square
root
of
the
number
of
potlines.
This
approach
accounts
for
the
reduction
in
variability
from
emission
averaging
and
results
in
lower
emission
limits
as
the
number
of
potlines
in
the
average
increases.
The
results
are
given
in
Table
2
and
were
calculated
from:

Emission
averaging
limit
=
3.18
+
2.326*
1.03/(
number
of
lines)
½
+
0.1
TABLE
2.
LIMITS
FOR
EMISSIONS
AVERAGING
Number
of
potlines
POM
limits
(
lb/
ton)

1
5.7
2
5.0
3
4.7
4
4.5
5
4.4
6
4.3
7
4.2
8
4.1
Table
3
presents
the
test
results
for
each
run
and
the
three­
run
averages
for
each
potline.
3
TABLE
3.
RESULTS
FOR
EACH
RUN
(
1996­
2000)

Date
Potroom
POM
(
lb/
ton)
Each
run
Average
of
3
runs
Nov­
99
1
2.43
Jan­
00
1
2.50
Feb­
00
1
2.10
2.34
Mar­
00
1
3.30
Apr­
00
1
1.90
May­
00
1
1.80
2.33
Jun­
00
1
2.90
Jul­
00
1
3.4
Aug­
00
1
6.00
4.1
Sep­
00
1
4.60
Oct­
00
1
1.70
Nov­
00
1
8.30
4.87
Average
for
Potline
1
3.41
Oct­
96
4
2.09
Oct­
96
4
1.49
Oct­
96
4
1.27
1.62
Dec­
96
4
1.27
Dec­
96
4
1.63
Dec­
96
4
1.96
1.62
Mar­
97
4
2.29
Mar­
97
4
4.56
Mar­
97
4
3.57
3.47
Jul­
97
4
3.19
Jul­
97
4
2.63
Jul­
97
4
4.13
3.32
Aug­
97
4
2.09
Aug­
97
4
2.10
Aug­
97
4
2.40
2.2
Nov­
99
4
0.51
Jan­
00
4
1.30
Feb­
00
4
0.90
0.90
Mar­
00
4
3.40
Apr­
00
4
2.20
May­
00
4
3.70
3.1
Jun­
00
4
3.40
Jul­
00
4
3.80
Aug­
00
4
1.60
2.93
Average
for
Potline
4
2.40
Date
Potroom
POM
(
lb/
ton)
Each
run
Average
of
3
runs
4
Nov­
99
5
2.56
Jan­
00
5
2.10
Feb­
00
5
2.10
2.25
Mar­
00
5
3.20
Apr­
00
5
3.90
May­
00
5
3.60
3.57
Jun­
00
5
3.80
Jul­
00
5
2.40
Aug­
00
5
8.80
5
Sep­
00
5
2.00
Oct­
00
5
2.20
Nov­
00
5
2.80
2.33
Dec­
00
5
3.00
Dec­
00
5
4.10
3.3
Average
for
Potline
5
3.33
Jan­
00
7
4.00
Feb­
00
7
3.20
Mar­
00
7
4.40
3.87
Apr­
00
7
2.70
Jun­
00
7
3.70
Jul­
00
7
5.00
3.8
Aug­
00
7
4.30
Sep­
00
7
5.20
Oct­
00
7
2.70
4.07
Nov­
00
7
5.70
Dec­
00
7
2.40
3.6
Average
for
Potline
7
3.94
Oct­
96
9
2.34
Oct­
96
9
1.52
Oct­
96
9
1.40
1.75
Dec­
96
9
2.13
Dec­
96
9
2.06
Dec­
96
9
2.34
2.18
Mar­
97
9
2.09
Mar­
97
9
3.39
Mar­
97
9
2.97
2.82
Jul­
97
9
1.81
Jul­
97
9
2.35
Jul­
97
9
3.40
2.52
Aug­
97
9
2.86
Date
Potroom
POM
(
lb/
ton)
Each
run
Average
of
3
runs
5
Aug­
97
9
2.51
Aug­
97
9
1.41
2.26
Jan­
00
9
3.50
Feb­
00
9
1.50
Mar­
00
9
2.40
2.47
Apr­
00
9
2.00
May­
00
9
3.80
Jun­
00
9
3.60
3.13
Jul­
00
9
2.80
Aug­
00
9
8.70
Sep­
00
9
3.70
5.07
Oct­
00
9
1.40
Nov­
00
9
3.80
Dec­
00
9
4.30
3.17
Average
for
Potline
9
2.82
Standard
deviation
(
averages
of
3
runs)
1.03
