
1
Copies
of
the
correspondence
are
available
in
the
docket.
A
complete
listing
is
provided
at
the
end
of
this
memo
under
"
Recent
Correspondence."

1
MEMORANDUM
Date:
August
27,
2003
To:
Docket
ID
Nos.
OAR­
2002­
0034
and
A­
2000­
56
(
Iron
and
Steel
Foundries)

From:
Kevin
Cavender,
EPA/
OAQPS
Subject:
Summary
of
Recent
Information
Provided
by
Foundries
Industry
We
recently
received
several
emails
from
Iron
and
Steel
Foundry
representatives.
1
The
materials
were
provided
well
after
the
close
of
the
public
comment
period,
and
only
days
prior
to
the
consent
decree
deadline
for
having
the
final
rule
signed
by
the
Administrator.
We
have
not
had
sufficient
time
to
adequately
review
the
information.

The
additional
information
provided
in
these
emails
pertain
to
the
costs
and
emission
reductions
associated
with
the
proposed
PM
limit
for
cupola
furnaces.
All
of
the
information
can
be
broken
down
into
the
following
four
areas
1)
the
expected
PM
emission
reduction,
2)
the
expected
annual
costs,
3)
the
expected
capital
costs,
and
4)
options
that
would
result
in
an
emission
limit
that
well
operated
wet
scrubbers
could
meet.

While
we
do
not
have
sufficient
time
to
adequately
review
all
of
the
material
that
the
industry
has
recently
provided,
based
on
our
brief
review
of
the
information
it
appears
that
these
messages
provide
no
new
information
that
would
materially
change
our
analysis
and
decisions.
The
following
discussion
points
out
some
of
the
key
points
raised
and
our
initial
review
of
the
information
provided.

Emission
Reduction
Estimates
Our
initial
estimate
of
the
PM
reduction
that
will
be
achieved
by
the
cupola
limit
was
1,290
tpy.
Industry
has
provided
a
number
of
estimates
of
the
PM
reduction
that
will
be
achieved
by
the
cupola
limit,
ranging
from
291
tpy
to
1,150
tpy.
Their
most
recent
estimates
indicate
a
PM
emission
reduction
of
850
tpy.

°
This
most
recent
estimate
is
a
variation
of
the
estimate
they
had
made
earlier.
In
this
estimate,
they
use
actual
emissions
data
where
it
is
available,
and
"
emission
factors"
for
those
plants
where
emissions
data
is
not
available.
°
We
agree
with
industries
approach
for
using
actual
emissions
data
for
foundry­
specific
emission
estimates.
2
°
We
disagree
with
the
emission
factors
they
use,
because
it
does
not
reflect
the
average
performance
of
scrubbers
and
baghouses.
°
We
revised
our
emission
estimates
using
facility­
specific
data
for
which
EPA
has
had
a
chance
to
review
the
source
test
data.
°
Our
revised
emission
estimate
is
1,100
tpy
using
source
test
data
that
has
been
submitted
to
EPA.
°
The
main
difference
between
our
current
estimate
and
AFS's
estimate
is
that
the
industry
is
using
stack
test
data
for
the
GM
Defiance
foundry
that
has
not
been
submitted
to
EPA.
°
We
are
reluctant
to
accept
the
new
data
as
it
is
very
low
in
comparison
to
other
wet
scrubbers,
and
no
supporting
information
was
provided
to
support
the
data.
°
Using
the
additional
site­
specific
emission
factors
as
presented
in
the
Alliance
spreadsheet
(
e.
g.,
for
GM
Defiance
foundry
and
others
that
we
do
not
have
substantiating
source
test
reports),
we
estimate
nationwide
impacts
of
880
tpy.

We
now
estimate
the
actual
PM
emission
reductions
are
between
880
and
1,100
tpy.
Because
we
can
not
substantiate
the
site
specific
emissions
estimates
for
the
GM
Defiance
foundry,
our
best
estimate
is
that
PM
will
be
reduced
by
1,100
tpy.

Annualized
Cost
Estimates
Industry
has
provided
significantly
higher
estimates
of
the
annualized
costs
of
the
requirement.
They
base
their
annual
cost
estimates
on
the
assumption
that
annual
operating
costs
of
the
baghouse
and
wet
scrubber
are
equivalent,
and
as
such,
the
total
annualized
cost
is
simply
calculated
based
on
the
annualized
capital
cost.

We
strongly
disagree
with
the
assumption
that
the
operating
costs
are
equivalent
between
a
baghouse
and
a
wet
scrubber.

°
Their
review
of
our
power
estimates
incorrectly
assumes
we
were
estimating
total
power
for
the
two
systems
when
in
fact
we
were
estimating
the
incremental
power
requirements.
°
They
attempt
to
use
measured
power
consumption
to
demonstrate
that
our
estimates
are
wrong.
But
in
fact
they
show
that
we
may
be
grossly
underestimating
the
power
savings
for
baghouses.
Our
procedures
for
this
system
would
estimates
a
savings
of
100KW,
where
their
measurements
show
a
savings
of
270KW,
2.7
times
what
we
are
estimating.
°
Their
second
comparison
of
the
design
horsepower
for
the
baghouse/
gas
cooler
versus
the
measured
horsepower
of
the
wet
scrubber
is
inappropriate.
The
average
power
usage
for
a
baghouse/
cooler
will
be
much
less
than
the
"
design"
power
connection
requirement.
°
Actual
data
from
foundries
having
converted
to
baghouses
system
show
an
actual
cost
savings.
°
GM
has
recently
provided
a
detailed
assessment
of
the
cost
savings
associated
with
retrofitting
their
system
to
a
wet
scrubber.
The
vendor
analysis
suggests
for
their
system
the
baghouse/
gas
cooler
will
save
them
millions
of
dollars
per
year
when
power
and
coke
savings
are
considered.
°
Numerous
foundries
have
undergone
the
retrofit
for
purely
economic
reasons,
which
provides
further
support
that
the
retrofit
results
in
actual
savings
over
the
continued
use
of
a
wet
scrubber
.
3
Capital
Cost
Estimates
Industry
representatives
suggest
that
we
underestimate
the
capital
costs
of
the
retrofit,
and
that
we
have
not
included
the
capital
cost
associated
with
the
gas
cooler
needed
to
cool
the
gas
to
a
level
that
a
baghouse
can
handle.

°
Our
estimates
are
for
a
traditional
vertical
baghouse
and
gas
cooler
that
is
typical
of
the
systems
from
which
the
existing
source
0.006
gr/
dscf
MACT
limit
are
based
on.
°
Our
cost
estimates
compare
very
well
with
the
actual
costs
incurred
by
facilities
that
have
installed
these
types
of
systems.
°
All
recent
baghouse
costs
submitted
by
industry
have
been
horizontal
baghouse
systems
that
far
exceed
the
baghouse
design
requirements
to
meet
a
0.006
gr/
dscf.
These
horizontal
baghouses
designs
have
all
tested
at
less
than
0.001
gr/
dscf.
°
The
horizontal
baghouses
are
designed
to
be
able
to
achieve
a
much
lower
air
to
cloth
ratio,
and
as
such
are
much
larger
and
more
expensive
than
the
traditional
designs
used
at
foundries.
°
While
the
horizontal
baghouses
and
have
a
higher
capital
cost,
based
on
discussions
with
owners
and
vendors,
operating
costs
are
lower
for
these
systems
than
for
traditional
designs,
which
seems
to
make
these
systems
attractive
for
the
retrofit.

Alternatives
Suggested
by
the
Industry
The
industry
has
suggested
a
number
alternatives
to
the
current
proposed
MACT
limit
for
cupolas.
These
suggestions
include:

°
Subcategorizing
based
on
control
technoloy
°
Subcategorizing
based
on
clean
scrap
°
Developing
a
total
HAP
emission
limit
for
the
furnace
°
Allow
for
emissions
averaging
These
alternatives
are
not
new,
and
we
have
evaluated
each
of
these
options.
However,
we
have
not
found
a
way
to
implement
these
options
in
a
way
that
comports
with
the
Clean
Air
Act
with
the
available
information
and
still
achieves
the
industry's
goal
of
reducing
the
burden
on
facilities
with
cupola
furnaces
equipped
with
a
wet
scrubber.

Recent
Correspondence
(
available
in
the
docket)

1.
P.
Maciejewski,
GM,
to
K.
Cavender,
EPA:
ESD:
MG,
August
22,
2003,
GM
cost
data
(
4
attachments).

2.
G.
Mosher,
American
Foundrymen's
Society,
to
K.
Cavender,
EPA:
ESD:
MG,
August
25,
2003,
Information
for
ongoing
discussions
on
subcategorization
with
attachment.

3.
G.
Mosher,
American
Foundrymen's
Society,
to
K.
Cavender,
EPA:
ESD:
MG,
August
25,
2003,
Why
AFS
estimate
is
more
accurate
(
2
attachments).
4
4.
Jeff
Goudzwaard,
NFCO,
to
K.
Cavender,
EPA:
ESD:
MG,
August
25,
2003,
Objecting
to
EPA's
energy
analysis
with
attachment.

5.
R.
Ross,
Warner
Norcross
&
Judd,
to
K.
Cavender,
EPA:
ESD:
MG,
August
25,
2003,
Information
regarding
cost
implications
of
the
foundry
MACT
(
2
attachments).

6.
G.
Logan,
Ford,
to
K.
Cavender,
EPA:
ESD:
MG,
August
26,
2003,
Ford
cupola
retrofit
costs
with
attachment.

7.
G.
Mosher,
American
Foundrymen's
Society,
to
K.
Cavender,
EPA:
ESD:
MG,
August
26,
2003,
EPA
has
underestimated
PM
reductions
from
a
0.03
gr/
dscf
wet
scrubber
subcategory
(
2
attachments).
