Standards
of
Performance
for
New
Stationary
Sources
and
Emissions
Guidelines
for
Existing
Sources:
Large
Municipal
Waste
Combustors
Summary
of
Public
Comments
and
Responses
for
December
19,
2005
Proposed
Rule
U.
S.
Environmental
Protection
Agency
OAQPS/
Sector
Policies
and
Programs
Division
Energy
Strategies
Group
Research
Triangle
Park,
NC
27711
April
25,
2006
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
ii
Disclaimer
This
report
has
been
reviewed
by
the
Sector
Policies
and
Programs
Division
of
the
Office
of
Air
Quality
Planning
and
Standards,
and
EPA,
and
approved
for
publication.
Mention
of
trade
names
or
commercial
products
is
not
intended
to
constitute
endorsement
or
recommendation
for
use.
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
iii
TABLE
OF
CONTENTS
Chapter
Page
LIST
OF
TABLES.................................................................................................................
iv
Chapter
1
Background
Information
........................................................................................
1
Chapter
2
Public
Comments
Received.....................................................................................
3
Chapter
3
Applicability
and
Compliance
Schedule.................................................................
5
3.1
Compliance
Schedule
..................................................................................................
5
3.2
Modified
Sources
........................................................................................................
5
3.3
CAA
112
Overlap........................................................................................................
6
3.4
40
CFR
Subparts
Db
and
E..........................................................................................
8
Chapter
4
MACT
Floor
and
MACT........................................................................................
8
4.1
General......................................................................................................................
10
4.2
Distinction
between
ESP
and
Fabric
Filter
Units........................................................
12
4.3
Pre­
Combustion
Measures
.........................................................................................
15
4.4
New
Combustor
Categories
.......................................................................................
17
Chapter
5
Data
Analysis
and
Emission
Limits
......................................................................
18
5.1
General......................................................................................................................
18
5.2
Data
Screening
..........................................................................................................
20
5.3
Data
Sources
.............................................................................................................
23
5.4
Statistical
Methods
....................................................................................................
25
5.5
Achievability
of
Emission
Limits
..............................................................................
31
5.6
Removal
Efficiency...................................................................................................
32
5.7
New
Control
Technologies
........................................................................................
35
5.8
Carbon
Monoxide
Limits...........................................................................................
37
5.9
Dioxin/
Furan
Limits
..................................................................................................
38
5.10
HCl
Limits
................................................................................................................
38
5.11
Mercury
Limits..........................................................................................................
39
5.12
Lead
Limits
...............................................................................................................
40
5.13
NOx
Limits
...............................................................................................................
41
5.14
SO2
Limits
.................................................................................................................
42
Chapter
6
Testing
and
Monitoring.........................................................................................
45
6.1
Schedule....................................................................................................................
45
6.2
Reduced
Testing
Frequency
for
Exceptionally
Well­
Operated
Units..........................
46
6.3
Alternative
CEM
Option............................................................................................
48
6.4
CEM
Validation
Requirements
..................................................................................
52
6.5
CEM
Data
Availability..............................................................................................
53
6.6
Harmonization
of
Monitoring
Requirements
between
40
CFR
60
and
75...................
55
6.7
Valid
Data
for
Compliance
Calculations
....................................................................
55
6.8
ACI
Pneumatic
Pressure
............................................................................................
56
6.9
Start­
up,
Shutdown
and
Malfunction
Provisions
........................................................
57
6.10
Relative
Accuracy
Test
Audits
(
RATA).....................................................................
58
6.11
Other
.........................................................................................................................
59
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
iv
TABLE
OF
CONTENTS
(
Continued)

Chapter
7
Cost
Impacts
..........................................................................................................
62
Chapter
8
Miscellaneous.........................................................................................................
62
8.1
Comment
Period........................................................................................................
67
8.2
State
Plans.................................................................................................................
67
8.3
Operator
Training
......................................................................................................
68
8.4
Renewable
Energy.....................................................................................................
69
8.5
Support
Other
Commenters
.......................................................................................
69
8.6
Revised
Emission
Limits                       ...
69
LIST
OF
TABLES
Table
Page
2­
1.
List
of
Commenters
for
the
Standards
of
Performance
for
New
Stationary
Sources
and
Emissions
Guidelines
for
Existing
Sources
..................................................
3
8­
1.
Final
Results
for
Computed
UCLs
and
LCLs
for
Stack
Test
and
CEM
Data
..................
70
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
1
Chapter
1
Background
Information
Section
129
of
the
Clean
Air
Act
(
CAA),
entitled
``
Solid
Waste
Combustion,''
requires
EPA
to
develop
and
adopt
NSPS
and
emission
guidelines
for
solid
waste
incineration
units
pursuant
to
CAA
sections
111
and
129.
Section
111(
b)
of
the
CAA
(
NSPS
program)
addresses
emissions
from
new
municipal
waste
combustor
(
MWC)
units
and
CAA
section
111(
d)

(
emission
guidelines
program)
addresses
emissions
from
existing
MWC
units.

In
December
1995,
EPA
adopted
NSPS
(
subpart
Eb)
and
emission
guidelines
(
subpart
Cb)
for
MWC
units
with
a
combustion
capacity
greater
than
250
tons
per
day.
These
MWC
units
are
referred
to
as
large
MWC
units.
Both
the
NSPS
and
emission
guidelines
require
compliance
with
emission
limitations
that
reflect
the
performance
of
maximum
achievable
control
technology
(
MACT).
The
NSPS
apply
to
new
MWC
units
after
the
effective
date
of
the
NSPS
(
June
19,
1996)
or
at
start­
up,
whichever
is
later.
The
emission
guidelines
apply
to
existing
MWC
units
and
required
compliance
by
December
2000.
These
retrofits
were
completed
on
time,
and
the
controls
installed
to
meet
the
required
emission
limitations
were
highly
effective
in
reducing
emissions
of
all
of
the
CAA
section
129
pollutants
emitted
by
large
MWC
units.

Section
129(
a)(
5)
of
the
CAA
requires
EPA
to
conduct
a
5­
year
review
of
the
NSPS
and
emissions
guidelines
and,
if
appropriate,
revise
the
NSPS
and
emission
guidelines.
The
EPA
completed
that
review,
and
issued
proposed
amendments
to
reflect
the
changes
EPA
believed
were
appropriate.

Thirty­
one
different
organizations
submitted
public
comments
on
the
proposed
amendments.
These
comments
came
from
wide
variety
of
sources,
consisting
mainly
of
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
2
government
agencies,
environmental
organizations,
industry
associations,
and
owner/
operators
of
affected
large
MWC
facilities.
All
of
the
comments
that
were
submitted
and
the
responses
to
the
comments
are
summarized
in
this
document.
This
summary
is
the
basis
for
the
revisions
made
to
the
rule
between
proposal
and
promulgation.
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
3
Chapter
2
Public
Comments
Received
The
EPA
received
comments
from
31
organizations
on
the
proposed
amendments
for
the
NSPS
and
emission
guidelines
for
large
MWC
units,
some
of
which
contained
attachments.
A
list
of
commenters,
their
affiliations,
and
the
EPA
docket
item
number
assigned
to
their
correspondence
is
provided
in
Table
2­
1.
To
achieve
an
organized
presentation,
EPA
has
grouped
the
comments
by
topic
in
Chapters
3
through
8
of
this
document.

TABLE
2­
1.
List
of
Commenters
for
the
Standards
of
Performance
for
New
Stationary
Sources
and
Emissions
Guidelines
for
Existing
Sources
(
Docket
ID
:
EPA­
HQ­
OAR­
2005­
0117)

EPA­
HQOAR
2005­
0117­
Commentera
Company
City/
State
0064
C.
Jones
City
&
County
of
Honolulu
Honolulu,
HI
0065.1
D.
Lue
Montenay
Power
Corp.
Miami,
FL
0068.1
W.
Smith
Pinellas
Co.
SW
Clearwater,
FL
0073.1
T.
Michaels
IWSA
Washington,
DC
0076
T.
Richter
Minnesota
RRA
Minneapolis,
MN
0077.1
T.
Porter
Wheelabrator
Hampton,
NH
0078.1
D.
Foerter
ICAC
Washington,
DC
0079.1
(
no
name)
MSI
Mechanical
Systems,
Incorporated
Sun
Prairie,
WI
0080.1
S.
Evans
CleanAir
Engineering
Palatine,
IL
0081.1
L.
Eberly
Xcel
Energy
Services
Minneapolis,
MN
0083.1
J.
Bilmes
Bristol
RRF
Operating
Committee
Bristol,
CT
0084.1
A.
Gobin
Connecticut
DEP
Hartford,
CT
0085.1
N.
McCann,
et
al.
IWSA
Municipal
Committee
and
Other
Public
Entities
Tampa,
FL,
et
al.
0086
J.
Brooks
Maine
Bureau
of
Air
Quality
Augusta,
ME
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
4
TABLE
2­
1.
List
of
Commenters
for
the
Standards
of
Performance
for
New
Stationary
Sources
and
Emissions
Guidelines
for
Existing
Sources
(
Docket
ID
:
EPA­
HQ­
OAR­
2005­
0117)

EPA­
HQOAR
2005­
0117­
Commentera
Company
City/
State
0087
J.
Lafreniere
Town
of
Rochester
Rochester,
NY
0088
D.
Taam
Spokane
Regional
Solid
Waste
System
Spokane,
WA
0089.1
J.
Skinner
SWANA
Silver
Spring,
MD
0090
T.
Cochran
US
Conference
of
Mayors
Washington,
DC
0092
W.
Ehrman
York
County
SWRA
York,
PA
0093
F.
Giordano
IWSA
Camden
Co.,
NJ
0094
G.
Laboa,
R.
Nunes,
and
B.
Morse
Rochester
Board
of
Selectmen
Rochester
0095
B.
Tippetts
La
Crosse
County
SW
La
Crosse,
WI
0096
T.
Hartman
Waste
System
Authority
of
Eastern
Montgomery
Co.
Norristown,
PA
0097
J.
Hadfield
Southeastern
Public
Service
Authority
Chesapeake,
VA
0098
B.
Mace
Camden
Co.
Energy
Recovery
Assoc.
Camden,
NJ
0099
M.
Brown
City
of
Savannah
Savannah,
GA
0100
W.
O'Sullivan
New
Jersey
DEP
Trenton,
NJ
0101
J.
Hadfield
Southeastern
Public
Service
Authority
Chesapeake,
VA
0102
A.
Rhoads
Onondaga
County,
NY
North
Syracuse,
NY
0103.1
Earthjustice
­
attachment
0104
D.
Donohoue
CARB
Sacramento,
CA
0106
D.
Foerter
ICAC
Washington,
DC
0107
D.
Foerter
ICAC
Washington,
DC
121
E.
Bennett
Montenay
Savannah
Operations,
Inc.
Savannah,
GA
122
G.
Varner
County
of
Charleston
Charleston,
SC
a
Commenters
listed
multiple
times
due
to
duplicate
comments
and
attachments
are
referred
to
in
Chapters
3
through
8
of
this
document
by
the
first
number
that
was
assigned
to
that
commenter
in
the
docket.
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
5
Chapter
3
Applicability
and
Compliance
Schedule
3.1
Compliance
Schedule
Comment:
Three
commenters
(
OAR­
2005­
0117­
0073.1,
OAR­
2005­
0117­
0065,
OAR­

2005­
0117­
0077)
proposed
a
five­
year
compliance
deadline,
from
the
date
of
publication
of
the
final
rule,
for
sources
that
choose
upgrade
existing
pollution
control
equipment.
The
commenter
noted
that
a
five­
year
deadline
was
provided
for
in
the
1995
rulemaking
and
suggested
that
this
timeframe
is
appropriate
for
facilities
to
raise
funds
and
acquire
additional
equipment.
One
commenter
(
OAR­
2005­
0117­
0073.1)
also
requested
a
two­
year
compliance
deadline
for
all
other
facilities.

Response:
EPA
agrees
with
the
commenter
and
revised
the
compliance
provisions
in
the
final
rule
to
include
two
alternatives.
First,
for
facilities
that
make
minor
changes,
EPA
retained
the
two­
year
compliance
deadline
as
proposed.
Secondly,
facilities
planning
substantial
upgrades
can
apply
to
the
Administrator
for
compliance
schedules
up
to
five
years.

3.2
Modified
Sources
Comment:
Two
commenters
(
OAR­
2005­
0117­
0073.1,
OAR­
2005­
0117­
0065)

requested
that
EPA
clarify
that
physical
or
operational
changes
made
to
an
existing
MWC
unit
primarily
for
the
purposes
of
complying
with
the
proposed
amendments
in
Subpart
Cb
will
not
classify
the
facility
as
a
modified
facility.
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
6
One
commenter
(
OAR­
2005­
0117­
0098)
requested
EPA
to
clarify
that
facilities
that
choose
to
make
partial
technology
upgrades
to
maintain
what
they
determine
to
be
an
adequate
margin
of
compliance,
not
be
qualified
as
a
new
source
under
the
MACT
rule.

Response:
EPA
agrees
that
any
change
made
to
an
MWC
unit
for
the
principal
purpose
of
complying
with
the
subpart
Cb,
40
CFR
part
60,
emission
guidelines
is
not
considered
to
be
a
modification
or
reconstruction.
The
principal
reason
this
occurs
is
the
air
pollution
control
system
is
not
part
of
affected
facility
regulated
by
the
rule.
Nevertheless,
the
preamble
to
the
final
amendments
contains
a
detailed
discussion
on
applicability
of
the
rules
and
a
discussion
of
the
fact
that
changes
made
to
an
MWC
principally
to
comply
with
the
MACT
limits
are
not
considered
a
modification
or
reconstruction.

3.3
CAA
112
Overlap
Comment:
One
commenter
(
OAR­
2005­
0117­
0073.1)
is
concerned
about
the
preamble
discussion
identifying
MWCs
as
a
source
category
"
subject
to
regulation"
for
purposes
of
section
112(
c)(
6)
with
respect
to
112(
c)(
6)
pollutants
that
include
polycyclic
organic
matter
(
POM)
and
polychlorinated
biphenyls
(
PCBs).
The
commenter
referenced
language
in
CAA
section
129
and
previous
EPA
documents
indicating
that
MWCs
are
to
be
regulated
under
section
129
rather
than
section
112.
The
commenter
requested
that
EPA
clarify
its
position
that
MWC
are
not
subject
to
section
112.
As
a
remedy,
the
commenter
suggests
that
EPA
either:
1)
Remove
discussion
of
section
112
from
the
preamble,
2)
clarify
that
section
129
sources
are
only
considered
"
subject
to
standards"
in
112(
c)(
6)
for
purposes
of
accounting
for
emission
reductions
achieved
since
implementation
of
112(
c)(
6);
or
3)
clarify
that
the
proposed
amendments
fully
satisfy
section
112
requirements
and
the
States
have
no
basis
for
additional
regulation
of
POMs
and
PCBs.

Another
commenter
(
OAR­
2005­
0117­
0092)
disagreed
with
the
language
in
the
April
1998
Federal
Register
that
stated
MWC
are
"
subject
to
regulation
for
the
purpose
of
CAA
section
112(
c)(
6)
with
respect
to
the
CAA
section
112(
c)(
6)
pollutants
that
MWC
units
emit"
(
70
FR
75356/
1).
The
commenter
contended
that
large
MWC
are
subject
to
standards
under
sections
129
and
111
and
cannot
also
be
subject
to
standards
under
112(
d)
or
requirements
under
112(
c)(
6).

Another
commenter
(
OAR­
2005­
0117­
0103.1)
requested
that
EPA
must
use
its
authority
under
Section
129(
a)(
4)
to
set
emission
floors
and
propose
standards
for
emissions
of
PCB
and
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
7
POM.
The
commenter
contended
that
the
large
MWC
source
category
accounts
for
greater
than
10
percent
of
the
aggregate
emissions
of
PCBs
and
POM,
and
therefore
EPA
must
set
emission
standards
for
these
pollutants
in
order
to
comply
with
the
requirements
contained
in
Sections
112
(
c)(
6)
and
112(
d)(
2).
The
commenter
refuted
the
claims
that
the
standards
for
PCB
and
POM
are
unnecessary
because
the
proposed
standards
for
other
pollutants
"
substantially"
reduce
the
emissions
of
PCB
and
POM.
The
commenter
contended
that
using
the
proposed
standards
for
other
pollutants
as
surrogates
for
PCB
and
POM
are
not
appropriate
in
this
instance,
and
the
commenter
added
that
EPA
did
not
claim
that
any
of
the
regulated
pollutants
serve
as
an
appropriate
surrogate
for
PCB
and
POM.
The
commenter
also
added
that
CAA
Section
112(
c)(
6)
requires
direct
numerical
emission
standards
for
bioaccumulative
air
pollutants,
and
does
not
provide
EPA
with
the
discretion
to
regulate
PCB
and
POM
through
surrogates.
Finally,

the
commenter
suggested
that
if
EPA
were
to
claim
the
proposed
standards
satisfy
CAA
112(
c)(
6)
as
surrogate
standards
for
PCB
and
POM,
the
proposed
standards
would
fail
a
three
prong
test
that
judicial
review
has
found
to
define
EPA's
discretion
to
develop
surrogate
standards.
The
commenter
cited
the
conditions
of
the
test
from
National
Lime
Ass'n
v.
EPA
233
F.
3d
625,
639,
and
emphasized
the
third
prong
of
the
test
which
states
that
PM
would
be
an
acceptable
surrogate
for
metals
if
"
PM
control
is
the
only
means
by
which
facilities
`
achieve'

reductions
in
HAP
metal
emissions."
The
commenter
refuted
that
combustion
controls
and
endof
stack
controls
are
the
only
means
to
achieve
reductions
in
PCB
and
POM.
The
commenter
identified
that
the
emissions
of
PCB
and
POM
are
affected
by
the
amount
of
these
pollutants,
as
well
as
the
amount
of
certain
chlorinated
compounds
and
other
plastics,
in
the
waste
stream.
The
commenter
concluded
that
EPA
does
not
have
the
authority
to
set
surrogate
standards
for
PCB
and
POM
and
requested
that
the
final
rule
incorporate
standards
for
these
two
pollutants.

Response:
EPA
did
not
propose,
and
is
not
promulgating,
numerical
emission
limits
for
PCBs
or
POMs
as
part
of
this
rulemaking.
The
section
112(
c)(
6)
discussion
was
included
in
the
preamble
to
the
proposed
amendments
to
provide
notice
that
EPA
intends
to
take
credit
for
reductions
in
emissions
of
PCBs
and
POMs
resulting
from
compliance
with
the
emission
limits
applicable
to
large
MWC
in
fulfilling
its
obligations
under
section
112(
c)(
6).

The
EPA
does
not
believe
that
either
section
129
or
section
112(
c)(
6)
require
it
to
establish
numerical
limits
for
PCB
and
POM
emissions
from
large
MWC
in
order
to
take
credit
for
reductions
in
emissions
of
those
pollutants
arising
from
compliance
with
the
large
MWC
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
8
NSPS
and
emission
guidelines
in
fulfilling
its
obligations
under
section
112(
c)(
6).
In
addition,

contrary
to
the
commenter's
contention,
EPA
did
not
purport
to
use
other
pollutants
emitted
by
large
MWC
as
surrogates
for
PCB
and
POM.
Rather,
EPA
was
simply
making
the
point
that
controls
installed
on
large
MWC
to
comply
with
applicable
emission
limits
also
effectively
minimize
emissions
of
PCB
and
POM
along
with
other
pollutants
specifically
required
to
be
regulated
under
section
129.
In
this
regard,
EPA
continues
to
believe
that
the
spray
dryer
scrubbing
systems
(
also
known
as
semi­
dry
scrubbing
systems)
employed
to
comply
with
the
emission
limits
in
the
rule
for
organic
pollutants
will
also
reduce
emissions
of
PCB
and
POM
from
large
MWC.
As
EPA
noted
in
the
December
19,
2005
proposal,
spray
dryer
scrubbing
systems
achieve
a
high
level
of
emissions
control
for
all
organic
compounds.
While
these
controls
are
used
primarily
to
comply
with
the
numerical
emission
limits
for
compounds
such
as
dioxin,
they
also
achieve
a
high
level
of
control
of
emissions
of
other
organic
compounds
including
PCB
and
POM.
EPA's
Municipal
Waste
Combustion
Study
 
Flue
Gas
Cleaning
Technology
(
EPA/
530/
SW­
021d;
table
4­
6)
lists
control
levels
exceeding
99
percent
in
most
cases
for
all
of
these
organic
compounds.
The
EPA
has
stack
test
data
confirming
these
projected
efficiencies.
At
proposal,
EPA
identified
three
tests
from
a
large
commercial
MWC
that,
in
addition
to
demonstrating
low
dioxin
emissions,
showed
PCB
and
POM
emissions
below
the
method
detection
limit.
Since
proposal,
EPA
has
obtained
results
from
an
additional
six
tests
of
a
large
commercial
MWC
in
which
PCB
and
POM
emissions
were
measured.
All
six
tests
show
low
dioxin
emissions
and
PCB
and
POM
emissions
at
levels
less
than
the
method
detection
limit.
This
data
confirms
that
spray
dryer
control
systems
effectively
control
a
wide
range
of
organic
emissions,
including
emissions
of
PCB
and
POM.

3.4
40
CFR
Subparts
Db
and
E
Comment:
Four
commenters
(
OAR­
2005­
0117­
0073.1,
OAR­
2005­
0117­
0088,
OAR­

2005­
0117­
0085,
OAR­
2005­
0117­
0077)
requested
that
facilities
subject
to
the
proposed
rule
be
exempted
from
certain
requirements
in
Subpart
Db.
Specifically,
the
commenter
requested
that
units
subject
to
the
proposed
rule
not
be
subject
to
the
PM
limits
in
Subpart
Db
(
which
are
less
stringent
than
the
MWC
MACT
limits)
or
daily
record
keeping
requirements
for
the
MSW
feed
rate
that
are
contained
in
40
CFR
60
Subpart
Db.
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
9
One
commenter
(
OAR­
2005­
0117­
0088)
noted
that
the
requirements
of
Subpart
Db
involve
additional
recordkeeping
without
any
corresponding
environmental
benefits.

Response:
EPA
agrees,
and
has
already
addressed
the
concern
of
the
commenters
regarding
the
applicability
with
Subpart
Db
(
Standards
of
Performance
for
Industrial­

Commercial­
Institutional
Steam
Generating
Units).
The
exemption
of
MWC
units
from
Subpart
Db
was
published
on
February
27,
2006
at
71FR9866.
The
exemption
for
large
MWC
units
was
inserted
into
the
applicability
section
of
Subpart
Db.

Comment:
Four
commenters
(
OAR­
2005­
0117­
0073.1,
OAR­
2005­
0117­
0088,
OAR­

2005­
0117­
0085,
OAR­
2005­
0117­
0077)
requested
that
facilities
subject
to
the
proposed
rule
be
exempted
from
certain
requirements
in
Subpart
E
that
have
been
superseded
by
the
MWC
MACT
standards.
Specifically,
the
commenters
requested
that
units
not
be
subject
to
the
particulate
matter
(
PM)
limits
in
Subpart
E
or
daily
charging
rate
monitoring
requirements
in
Subpart
E
because
the
emission
limits
in
the
MWC
MACT
rule
are
more
stringent
and
the
MACT
steam
flow
monitoring
requirement
is
a
better
indicator
of
MWC
load
and
operation
than
charge
rate
monitoring.
The
commenters
added
that
compliance
with
these
Subpart
E
requirements
are
included
in
many
title
V
permits
and
require
facilities
to
perform
unnecessary
reporting.

Response:
EPA
agrees
with
the
commenters
and,
in
the
final
amendments,
an
exemption
of
MWC
units
was
inserted
in
the
applicability
section
of
Subpart
E
(
Standards
of
Performance
for
Incinerators)
to
exempt
from
Subpart
E
facilities
subject
to
Subpart
Eb,
or
subject
to
large
MWC
emission
guidelines
(
Subpart
Cb)
via
a
state
plan
or
federal
plan
(
Subpart
FFF).
EPA
also
exempts
from
Subpart
E
small
MWC
units
subject
to
the
small
MWC
NSPS
(
Subpart
AAAA
for
the
same
reason
as
the
exemption
for
large
MWCs.
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
10
Chapter
4
MACT
Floor
and
MACT
4.1
General
Comment:
One
commenter
(
OAR­
2005­
0117­
0090)
supports
EPA's
decision
not
to
revisit
the
MACT
floor
or
perform
another
MACT
analysis.
The
commenter
contended
that
the
large
MWC
facilities
in
the
U.
S.
operate
with
modern
and
efficient
pollution
control.

Response:
EPA
acknowledges
the
commenter's
support.
EPA
has
discussed
the
rationale
for
why
the
MACT
floor
was
not
recalculated
in
the
preamble
to
the
final
amendments.

Comment:
One
commenter
(
OAR­
2005­
0117­
0065)
requested
that
EPA
not
seek
to
reduce
the
emission
limits
(
i.
e.
"
turn­
down")
every
time
the
5­
year
review
is
conducted.

Another
commenter
(
OAR­
2005­
0117­
0092)
contended
that
the
proposed
standards
are
inappropriate
and
noted
that
the
five
year
review
for
an
NSPS
is
required
only
if
appropriate.

The
commenter
contended
that
there
was
no
new
MACT
technology
established
or
a
different
set
of
operating
practices
implied
in
the
preamble
to
warrant
a
review.
The
commenter
added
that
the
MACT­
equipped
facilities
that
were
left
in
the
database
performed
better
than
had
been
projected
and
as
a
result
the
proposed
emission
standards
were
tightened.
The
commenter
cited
a
court
case,
Financial
Institutions
Retirement
Fund
v.
Office
of
Thrift
Supervision
to
prove
"
once
again
that
no
good
deed
goes
unpunished."
The
commenter
noted
that
by
applying
reduced
limits
to
existing
pollution
control
technologies,
the
proposed
rule
removes
any
incentive
for
facilities
to
maximize
emission
reduction
and
control.

Response:
As
required
by
the
CAA,
EPA
is
required
to
review
the
MWC
rule
every
5
years,
however
EPA
agrees
that
the
emission
limits
will
not
necessarily
be
reduced
upon
each
review.
In
this
action
EPA
provides
a
review
of
the
1995
emission
limits
for
large
MWC
units.

The
1995
rulemaking
was
based
on
engineering
judgment,
given
a
lack
of
long
term
test
data
available
at
the
time
of
the
rulemaking.
Given
the
compliance
test
data
and
long­
term
data
now
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
11
available,
EPA
concluded
it
was
appropriate
to
revise
the
emission
limits
to
align
them
with
the
actual
performance
of
MACT.

Comment:
One
commenter
(
OAR­
2005­
0117­
0103.1)
contested
that
EPA
did
not
revise
the
MACT
floor
in
accordance
with
the
requirements
of
CAA
Section
129(
a)(
5)
to
ensure
the
floor
reflects
the
actual
performance
of
the
best
performing
units.
The
commenter
identified
that
the
revised
standards
reflect
emission
levels
regarded
to
be
achievable
for
all
units
operating
an
ESP,
and
the
Cement
Kiln
Recycling
Assoc.
v.
EPA
case
was
cited
as
precedent
to
establish
that
EPA
cannot
establish
a
MACT
floor
by
setting
standards
that
are
achievable
for
all
units
using
a
chosen
control
technology.

The
commenter
emphasized
the
importance
of
following
the
guidelines
set
forth
in
CAA
129(
a)(
5)
to
correctly
revise
the
MACT
floor,
given
that
the
commenter
claims
that
the
original
MACT
floors,
established
in
the
1995
rulemaking,
were
not
set
to
reflect
the
actual
performance
of
the
relevant
best
performing
units.

Response:
As
discussed
in
detail
in
the
preamble
to
the
final
amendments,
EPA
does
not
believe
that
it
is
required
to
revise
the
MACT
floor
each
time
it
updates
the
NSPS
and
emission
guidelines.
Rather,
EPA
is
required
to
evaluate
the
capability
of
the
available
control
technology
and
to
update
the
NSPS
and
emission
guidelines
based
on
that
evaluation.

Comment:
One
commenter
(
OAR­
2005­
0117­
0103.1)
stated
that
they
were
petitioning
the
EPA
to
amend
the
1995
MWC
regulations.
The
commenter
cited
precedent
in
Kennecott
Utah
Copper
Corp
V.
Dept.
of
Interior
as
a
basis
for
their
right
to
petition,
and
claimed
several
parts
of
the
1995
MWC
to
be
in
violation
of
statutory
floor
provisions.
Specifically
the
commenter
petitioned
EPA
to
amend
the
1995
MWC
regulations
on
the
following
issues:
(
1)

bring
regulations
into
compliance
with
the
requirements
contained
in
CAA
Section
129(
a)(
2)
for
setting
MACT
floors;
(
2)
incorporate
emissions
monitoring
requirements;
(
3)
eliminate
the
distinction
between
MWC
units
based
on
type
of
air
pollution
control
device;
and,
(
4)
consider
the
maximum
degree
of
emission
reductions
that
is
achievable,
including
pre­
combustion
measures.

Response:
The
EPA
does
not
believe
that
it
is
either
necessary
or
appropriate
to
amend
the
1995
MWC
regulations.
The
time
for
challenging
those
regulations
has
long
since
passed
and
nothing
that
EPA
did
or
didn't
do
in
promulgating
the
1995
MWC
regulations
justifies
reopening
those
regulations
to
review
at
this
time.
The
EPA
fully
complied
with
the
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
12
requirements
in
CAA
section
129(
a)(
2)
in
establishing
the
relevant
MACT
floors
on
which
the
1995
MWC
regulations
are
based.
The
1995
MWC
regulations
included
all
appropriate
and
currently
demonstrated
emissions
monitoring
requirements.
All
of
the
standards
in
the
1995
MWC
regulations
that
were
based
on
particulate
matter
control
were
beyond­
the­
floor
standards.

As
such,
it
was
appropriate
for
EPA
to
consider
the
cost
of
replacing
existing
ESP
with
fabric
filters.
Based
on
its
evaluation,
EPA
concluded
that
the
cost
of
mandating
the
replacement
of
ESP
with
fabric
filters
could
not
be
justified.
Once
EPA
concluded
that
it
was
not
appropriate
to
require
that
all
ESP
be
replaced
with
fabric
filters,
it
acted
within
its
authority
in
establishing
separate
emission
limits
(
based
on
their
respective
performance)
for
MWC
equipped
with
ESP
and
MWC
equipped
with
fabric
filters.
Finally,
as
discussed
in
detail
in
the
preambles
and
background
information
documents
for
the
1991
and
1995
MWC
regulations
(
see
56
FR
5488
and
5514
and
60
FR
65387),
EPA
fully
considered
a
wide
range
of
control
measures,
including
pre­
combustion
measures
such
as
source
separation,
in
developing
the
1995
MWC
regulations
and
exercised
its
discretion
not
to
require
such
measures.
Nothing
more
was
required
in
this
regard.

4.2
Distinction
between
ESP
and
Fabric
Filter
Units
Comment:
One
commenter
(
OAR­
2005­
0117­
0073.1)
agreed
that
spray
dryer/
electrostatic
precipitator/
activated
carbon
injection/
selective
non­
catalytic
reduction
(
SD/
ESP/
ACI/
SNCR)
is
MACT
for
existing
facilities
subject
to
Subpart
Cb,
and
that
spray
dryer/
fabric
filter/
activated
carbon
injection/
selective
non­
catalytic
reduction
(
SD/
FF/
ACI/
SNCR)
is
MACT
for
MWC
units
subject
to
Subpart
Eb.
Furthermore,
the
commenter
agreed
with
EPA's
analysis
of
the
cost
to
replace
ESP
control
devices.

Response:
EPA
recognizes
the
commenter's
support
for
the
separation
of
MACT
for
existing
and
new
units
and
the
cost
analysis
for
replacing
ESP
devices.
EPA
continues
to
believe
that
these
two
control
technology
combinations
are
an
appropriate
approach
for
this
rulemaking.

Comment:
One
commenter
(
OAR­
2005­
0117­
0103.1)
requested
that
EPA
remove
any
distinction
between
units
based
on
control
device
type
(
i.
e.
units
with
ESP
vs.
units
without
an
ESP).
The
commenter
asserted
that
EPA
does
not
have
the
authority
to
make
distinctions
based
on
control
device
type
when
establishing
standards
under
Section
129(
a)(
5).
The
commenter
requested
that
although
EPA
has
proposed
that
all
existing
units
must
meet
the
21
ng/
dscm
limit
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
13
for
dioxin/
furan
by
2009,
the
EPA
should
not
wait
3
years
to
eliminate
the
control
device
distinction.

The
commenter
also
requested
that
the
final
rule
amendments
require
units
with
an
ESP
control
technology
to
be
replaced
by
a
fabric
filter
technology
to
ensure
that
the
emission
standards
reflect
maximum
achievable
control.

The
commenter
contended
that
the
current
basis
of
rejecting
the
requirement
for
ESP
replacement
is
unlawful.
The
commenter
suggested
that
EPA
incorrectly
analyzed
a
costemission
reduction
ratio,
instead
of
determining
whether
or
not
ESP
replacement
would:
1)

achieve
additional
reductions
in
emissions;
and
2)
be
achievable
by
a
majority
of
the
industry,

considering
cost.
The
commenter
added
that
although
EPA
argued
that
the
cost­
reduction
ratio
is
excessive,
it
also
claimed
the
cost
impacts
of
a
regulation
requiring
ESP
replacement
is
expected
to
be
insignificant
(
Docket
Item
OAR­
2005­
0117,
Item
0119).
The
commenter
also
identified
two
apparent
flaws
of
the
cost­
benefit
approach
used
by
EPA.
First
the
commenter
contended
that
the
$/
ton
costs
for
particulate
were
inflated
since
EPA
separated
reductions
of
total
particulate
and
fine
particulate.
Second,
the
commenter
identified
that
EPA
acted
arbitrarily
when
it
did
not
acknowledge
the
benefits
of
reductions
in
dioxins,
cadmium,
lead,
and
mercury,

which
are
achieved
through
ESP
replacement.

Response:
EPA
has
fully
considered
the
issue
raised
by
the
commenter
and
believes
that
it
is
appropriate
to
continue
to
allow
the
use
of
both
ESP
and
fabric
filters
for
compliance
with
this
rule.
The
1995
MWC
regulations
established
different
emission
limits
for
MWC
equipped
with
ESP
and
MWC
equipped
with
fabric
filters.
Both
of
these
limits
were
beyond­
the­
floor
limits.
As
such,
it
was
appropriate,
in
setting
the
limits,
for
EPA
to
consider
the
cost
of
replacing
ESP
with
fabric
filters,
to
conclude
that
the
cost
was
not
acceptable
and,
on
that
basis,
to
establish
different
emission
limits
based
on
the
relative
performance
of
the
two
control
devices.

In
conducting
its
review
of
the
1995
MWC
regulations,
EPA
has
again
concluded
that
the
cost
of
requiring
that
existing
ESP
be
replaced
with
fabric
filters
is
not
acceptable.
As
the
commenter
correctly
points
out,
as
set
forth
in
detail
in
the
preamble
to
the
December
19,
1995
proposal,
that
conclusion
is
based
on
EPA's
comparison
of
the
cost
of
replacing
the
ESP
with
the
additional
reduction
in
emissions
that
will
be
realized
as
a
result
of
the
change.
The
EPA
believes
that
this
cost
benefit
number,
rather
than
the
cost
of
replacement
to
an
individual
facility,
is
the
appropriate
metric
to
be
considered
in
determining
whether
the
cost
of
the
replacement
is
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
14
justified.
As
a
result,
the
conclusion
that
the
cost
of
replacement
is
inconsequential
is
not
at
odds
with
EPA's
determination
that
the
costs
associated
with
replacing
all
ESPs
with
fabric
filters
are
not
acceptable.

EPA
provides
a
measure
of
the
monetary
benefits
per
ton
of
PM
reduction
from
switching
from
an
ESP
to
a
fabric
filter
since
such
benefits
are
readily
available
to
the
Agency
as
seen
in
recently
completed
Regulatory
Impact
Analyses
(
e.
g.,
Industrial
Boilers/
Process
Heaters
NESHAP,
Clean
Air
Interstate
Rule,
Clean
Air
Mercury
Rule,
etc.).
Benefit
estimates
vary
by
the
fraction
of
PM
that
is
reduced.
It
is
typical
that
a
reduction
of
a
ton
of
fine
PM
has
more
monetized
benefit
compared
to
a
ton
of
coarse
PM
(
see
benefits
analyses
for
recently
completed
RIAs
mentioned
above).
The
Agency's
presentation
of
different
cost­
effectiveness
values
for
reducing
these
PM
fractions
is
to
allow
comparison
of
these
values
to
available
benefits
estimates
per
ton
for
these
fractions.
Reductions
of
both
fine
and
total
particulate
do
take
place,

and
the
total
capital
and
annualized
control
cost
for
the
switch
from
ESP
to
FF
is
the
same
regardless.
Replacement
of
ESPs
with
fabric
filters
will
yield
additional
reductions
in
dioxins,

cadmium,
lead
and
mercury
and
thus
additional
benefits,
but
the
Agency
did
not
have
a
methodology
to
estimate
the
monetized
benefits
of
reductions
of
these
pollutants
except
for
mercury.
EPA
did
not
provide
a
comparison
of
benefits
per
ton
of
mercury
reduction
since
no
modeling
of
mercury
reductions
sufficient
for
estimating
monetized
benefits
from
sources
such
as
large
MWCs
was
available.

The
economic
impact
analysis
of
the
switch
from
an
ESP
to
a
fabric
filter
for
21
large
MWC
units
is
meant
to
provide
an
assessment
of
the
impact
of
this
switch
upon
the
firms
owning
these
units.
In
this
instance,
the
economic
impacts
are
insignificant
from
the
standpoint
of
comparing
the
annualized
compliance
costs
to
the
revenues
of
the
parent
entity
(
private,

government,
and
community).
Significance
in
the
magnitude
of
this
impact
metric
is
different
than
significance
in
magnitude
from
a
cost
or
benefit
perspective.

Comment:
One
commenter
(
OAR­
2005­
0117­
0107)
contended
that
EPA
did
not
consider
upgrades
to
existing
equipment
as
an
option
for
enhancing
pollution
control
capability.

The
commenter
specifically
suggested
that
EPA
analyze
upgrading
existing
ESP
equipment.

Response:
EPA
disagrees
with
the
commenter.
The
emission
guidelines
represent
actual
performance
of
control
devices
on
existing
units
at
the
time
of
the
proposed
rulemaking.
The
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
15
emission
guidelines
do
not
include
changes
to
control
devices,
such
as
an
upgrade
to
an
ESP.

EPA
recognizes
that
States
or
individual
sources
retain
the
right
to
elect
more
stringent
emission
limits
that
could
require
an
upgrade
to
an
ESP.

Comment:
One
commenter
(
OAR­
2005­
0117­
0100)
requested
that
EPA
require
all
new
units
constructed
after
December
19,
2006,
be
equipped
with
fabric
filter
in
order
to
minimize
emissions
of
mercury,
fine
particulates,
dioxin,
and
other
heavy
metals.

Response:
The
new
source
applicability
date
of
Subpart
Eb
is
September
20,
1994,
and
the
1995
rulemaking
requires
new
MWC
units
constructed
after
that
date
to
meet
emission
limits
that
are
based
on
the
use
of
a
fabric
filter
control
system.
Thus,
the
outcome
advocated
by
the
commenter
has
already
been
realized.

4.3
Pre­
Combustion
Measures
Comment:
One
commenter
(
OAR­
2005­
0117­
0103.1)
contended
that
EPA
has
violated
CAA
129(
a)(
5)
by
not
attempting
to
ensure
that
the
proposed
standards
require
the
maximum
achievable
degree
of
reduction
in
emissions.
The
commenter
contended
that
EPA
must
set
final
standards
based
on
pre­
combustion
technologies
(
in
addition
to
add­
on
technologies)
unless
these
pre­
combustion
techniques
would
(
1)
not
yield
any
additional
reductions;
or
(
2)
would
not
be
achievable
as
outlined
in
CAA
129(
a)(
2).
The
commenter
added
that
the
docket
demonstrates
that
pre­
combustion
measures
are
achievable,
and
cited
Docket
#
A­
89­
08.
This
commenter
quoted
the
agency
as
concluding
that
"
the
overall
costs
of
the
materials
separation
requirements
will
be
negligible."
The
commenter
also
stated
that
EPA
identified
that
benefits
from
materials
separation
would
include:
" 
reduced
combustor
emissions"
of
several
pollutants.
The
commenter
argued
that
it
is
evident
removing
mercury
and
lead
from
the
waste
stream
would
reduce
emissions
of
these
metals.
They
also
claimed
removal
of
chlorinated
plastics
from
the
waste
stream
would
reduce
emissions
of
chlorinated
organic
pollutants
and
hydrogen
chloride
(
HCl).
The
commenter
concluded
that
because
the
docket
contains
references
that
indicate
that
pre­
combustion
measures
are
achievable
and
yield
emission
reductions,
these
measures
must
be
part
of
the
final
MWC
standards.

The
commenter
emphasized
the
importance
of
revising
the
proposed
emission
limits
to
reflect
the
maximum
degree
of
emission
reduction,
including
pre­
combustion
measures,
given
that
the
commenter
claimed
the
original
emission
standards,
created
in
the
1995
rulemaking,
did
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
16
not
meet
the
CAA
129(
a)(
2)
and
(
3)
requirements
for
maximum
degree
of
emissions
reduction
including
pre­
combustion
measures.

Response:
EPA
agrees
that
it
has
the
authority
to
consider
pre­
combustion
technologies,

including
source
separation,
in
establishing
numerical
emission
limits
for
large
MWC.
EPA
is
not,
however,
required
to
base
those
numerical
emission
limits
on
mandatory
waste
separation
requirements.
EPA
discussed
its
position
on
this
issue
in
the
1991
and
1995
MWC
rulemaking.

(
Please
see
56
FR
5496
and
Chapter
4,
Section
4.2.6,
of
EPA
document
EPA­
453/
R­
95­
0136
for
further
discussion
of
EPA's
position
regarding
pre­
combustion
and
materials
separation
during
the
1991
and
1995
MWC
rulemaking.)
In
essence,
EPA
concluded
that
the
question
of
whether
mandatory
source
separation
is
appropriate
is
best
answered
at
the
local
level,
not
in
the
context
of
a
national
rule
designed
to
minimize
air
emissions
from
the
combustion
of
municipal
waste.

This
conclusion
is
based
on
the
fact
that
reductions
in
air
emissions
are
site
specific.
That
is,
the
degree
by
which
source
separation
will
effectively
reduce
air
emissions
will
vary
from
one
MWC
to
another
depending
on
waste
stream
received.
The
EPA
believes
that
these
conclusions
continue
to
be
true.
The
EPA
also
points
out,
however,
that
waste­
separation
is
a
required
element
of
a
facility
siting
plan
for
large
MWC
subject
to
the
NSPS.
Furthermore,
it
has
been
EPA's
observation
that
most
existing
MWC
units
are
located
in
areas
that
currently
do
require
at
least
some
level
of
waste
separation
and
the
available
data
shows
that
areas
where
MWC
are
located
achieve
a
higher
waste
separation
level
than
the
national
average.
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
17
4.4
New
Combustor
Categories
Comment:
One
commenter
(
OAR­
2005­
0117­
0103.1)
disagreed
with
the
Agency's
method
to
establish
carbon
monoxide
(
CO)
standards
for
two
new
combustor
types:
spread
stoker
refuse
derived
fuel
(
RDF)­
fired/
100
percent
coal
capable
combustor
and
a
semisuspension
RDF­
fired/
wet
RDF
process
conversion
combustor.
The
commenter
contended
that
when
establishing
the
CO
standards
for
these
two
subcategories,
EPA
did
not
determine
the
relevant
best
performing
units
in
each
category
and
did
not
set
floors
for
CO
at
the
levels
achieved
by
such
units.
The
commenter
also
contended
that
EPA
did
not
determine
the
maximum
achievable
degree
of
reduction
in
CO.
The
commenter
stated
that
the
process
used
for
setting
the
CO
standard
on
these
new
combustors
is
unlawful
and
does
not
satisfy
the
provisions
for
establishing
a
standard
set
for
in
CAA
129(
a)(
2).

The
commenter
added
that
EPA
should
have
established
standards
for
other
pollutants
in
these
two
subcategories.
The
commenter
added
that
EPA
is
obligated,
under
CAA
129(
a)(
2),
to
ensure
that
emission
standards
to
reduce
the
relevant
pollutants
to
the
maximum
achievable
degree
are
set
for
each
subcategory.
The
commenter
concluded
that
EPA
acted
arbitrarily,
when
only
setting
CO
standards
for
these
two
new
combustor
subcategories.

Response:
EPA
has
considered
the
commenter's
issues
and
disagrees
with
the
commenter.
First,
EPA
identified
unique
combustor
designs
that
impact
the
CO
emissions
from
the
two
new
subcategories
of
combustors,
and
EPA
established
CO
emission
limits
based
on
the
data
available
for
each
unique
type
of
combustor.
For
the
other
eight
pollutants
covered
by
the
standards,
the
units
are
regulated
under
the
broader
subcategory
of
refuse­
derived
fuel
fired
combustors
as
is
done
for
all
combustor
types.

Comment:
One
commenter
(
OAR­
2005­
0117­
0073.1)
agreed
with
the
new
combustor
classifications
and
their
associated
CO
limits.

Response:
EPA
acknowledges
the
commenter's
support
for
the
new
combustor
classifications
and
associated
CO
limits.
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
18
Chapter
5
Data
Analysis
and
Emission
Limits
5.1
General
Comment:
Several
commenters
(
OAR­
2005­
0117­
0073.1,
OAR­
2005­
0077,
OAR­
2005­

0117­
0065,
OAR­
2005­
0117­
0089,
OAR­
2005­
0117­
0080,
OAR­
2005­
0117­
0098,
OAR­
2005­

0117­
0081,
OAR­
2005­
0117­
0092,
OAR­
2005­
0117­
0102,
OAR­
2005­
0117­
0094,
OAR­
2005­

0117­
0087,
OAR­
2005­
0117­
0064,
OAR­
2005­
0117­
0083,
OAR­
2005­
0117­
0088,
OAR­
2005­

0117­
0090,
OAR­
2005­
0117­
0085,
OAR­
2005­
0117­
0076,
OAR­
2005­
0117­
0095,
OAR­
2005­

0117­
0096,
OAR­
2005­
0117­
0084,
OAR­
2005­
0117­
0107,
OAR­
2005­
0117­
0097,
OAR­
2005­

0117­
0068,
OAR­
2005­
0117­
0100,
OAR­
2005­
0117­
0099)
had
questions
on
how
the
proposed
(
and
final)
limits
and
percent
reduction
requirements
were
developed.
The
commenters
had
concerns
about
documentation
of
the
years
included
in
the
dataset
and
the
statistical
analysis
used
to
support
the
limits.
Commenters
also
questioned
the
achievability
of
the
limits
over
the
long
term
and
whether
or
not
the
limits
accounted
for
an
adequate
compliance
margin.
Further,

the
commenters
had
questions
on
whether
or
not
the
limits
represented
the
practical
limits
of
certain
control
technologies.
The
detailed
comments
on
specific
pollutant
emission
limits
and
data
concerns
are
presented
in
the
remainder
of
this
chapter.

Response:
The
development
of
the
proposed
limits
is
presented
in
documents
that
were
contained
in
the
docket
at
time
of
proposal.
For
promulgation,
the
final
emission
limits
were
selected
using
a
different
three­
step
process.
The
first
step
was
to
develop
statistical
estimates.

The
second
step
was
to
consider
the
statistical
estimates
in
relation
to
current
performance
levels.

The
third
step
was
for
EPA
to
select
emission
limits.
Relative
to
the
first
step,
EPA
identified
an
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
19
appropriate
statistical
model,
defined
reasonable
assumptions,
and
applied
the
model
to
year
2000
compliance
data
for
all
MWC
units
with
the
relevant
control
technologies
to
estimate
the
peak
emission
rate
that
is
estimated
to
occur
from
time
to
time.
Next,
EPA
obtained
year
2001
to
2005
test
data
from
more
than
a
dozen
MWC
units.
These
data
were
compared
to
the
statistical
estimates
and
considered
in
relation
to
public
comments.
As
a
last
step,
EPA
selected
the
emission
limits
for
the
final
amendments.
The
statistical
analysis
used
for
the
final
standards
is
described
in
a
technical
memorandum
in
the
docket.
The
results
of
the
statistical
analysis
are
presented
in
the
Table
8­
1
at
the
end
of
this
document.

With
the
exception
of
lead
(
Pb)
and
mercury
(
Hg)
emissions
for
existing
MWCs
and
Hg
emissions
for
new
MWCs,
the
statistical
estimates
appeared
as
reasonable
prediction
of
future
MWC
performance.
The
estimated
emission
levels
were
rounded
up
to
the
next
multiple
of
five
and
adopted
as
a
final
limit.

For
the
lead
emission
limit
for
existing
MWCs,
a
different
approach
was
used
because
the
2001
to
2005
data
differed
from
the
statistical
estimate.
The
same
first
step
(
statistical
estimates)
and
second
step
(
review
of
2001­
2005
test
data)
was
followed,
but
the
variability
in
Pb
emissions
observed
following
2001
exceeded
the
statistical
estimates.
EPA
discounted
both
the
EPA
and
industry
statistical
estimates,
and
based
the
final
limit
on
a
review
of
the
test
data
and
consideration
of
public
comment
noting
high
variability
of
Pb
emissions,
and
selected
a
higher
Pb
emission
limit
of
400
milligrams
per
dry
standard
cubic
meter
(
mg/
dscm)
for
existing
MWCs.

This
deviation
from
estimates
also
occurred
for
the
mercury
emission
limits
for
existing
and
new
MWCs.
Again,
the
same
first
step
(
statistical
estimates)
and
second
step
(
review
of
2001
to
2005
data)
was
followed.
This
time
the
mercury
test
results
for
the
past
three
years
(
2003,
2004
and
2005)
for
all
MWCs
was
noticeably
less
than
the
statistical
estimate
for
Hg
or
the
current
Hg
standard
(
80
micrograms
(
ug)/
dscm).
To
understand
this
better,
EPA
reviewed
uncontrolled
Hg
emissions
data
from
a
number
of
MWCs
for
the
1995
to
2005
time
period
as
well
as
data
from
1991
when
the
Hg
control
database
was
developed.
In
1991,
Hg
emission
levels
of
650
ug/
dscm
were
not
uncommon.
The
original
standard
of
80
ug/
dscm
was
based
on
85
percent
control
of
Hg
levels
up
to
about
525
ug/
dscm
with
the
alternate
85
percent
reduction
applying
to
higher
concentrations.
The
new
data
showed
that
by
1995,
when
the
MACT
standards
were
adopted,
average
uncontrolled
Hg
emission
levels
were
about
250
ug/
dscm,
and,
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
20
by
2005,
the
level
was
reduced
by
about
50
percent
to
about
125
ug/
dscm.
The
result
of
application
of
85
percent
Hg
control
to
these
lower
Hg
inlet
levels
has
resulted
in
much
lower
Hg
outlet
levels,
as
demonstrated
by
the
test
data.
A
50
percent
reduction
in
inlet
Hg
levels
suggests
an
emission
limit
of
40
ug/
dscm
in
the
MACT
standards.
However,
both
public
comments
and
test
data
suggested
that
levels
less
than
30
ug/
dscm
are
achieved.
In
consideration
of
the
increased
use
of
Hg
continuous
emissions
monitoring
systems
(
CEMS)
and
the
higher
Hg
variability
that
may
be
observed
with
CEMS
use,
the
final
standards
were
set
at
50
ug/
dscm
for
both
existing
and
new
MWCs.

Comment:
One
commenter
(
OAR­
2005­
0117­
0103.1)
highlighted
several
improvements
in
the
proposed
standards
compared
to
the
standards
promulgated
during
the
1995
rulemaking.

The
commenter
noted
that
several
of
the
emission
limits
have
been
improved.

Response:
EPA
recognizes
the
commenter's
support
for
the
proposed
changes
to
the
emission
standards.
EPA
has
revised
the
emission
standards
since
proposal
and
continues
to
believe
that
these
standards
still
provide
an
improvement
to
the
standards
promulgated
in
the
1995
rulemaking.

5.2
Data
Screening
Comment:
Several
commenters
(
OAR­
2005­
0117­
0073.1,
OAR­
2005­
0117­
0089,
OAR­

2005­
0117­
0077,
OAR­
2005­
0117­
0080,
OAR­
2005­
0117­
0098,
OAR­
2005­
0117­
0081,
OAR­

2005­
0117­
0065,
OAR­
2005­
0117­
0092,
OAR­
2005­
0117­
0102,
OAR­
2005­
0117­
0094,
OAR­

2005­
0117­
0087,
OAR­
2005­
0117­
0064,
OAR­
2005­
0117­
0083,
OAR­
2005­
0117­
0088,
OAR­

2005­
0117­
0090)
contended
that
EPA
used
an
unprecedented
method
for
identifying
and
removing
outliers
in
the
emissions
test
database.

The
commenters
disagreed
with
the
outlier
screening
approach
for
facilities
with
more
than
one
similar
MWC
unit
and
argued
that
EPA
provided
no
support
or
explanation
for
eliminating
data
based
on
this
novel
screening
approach.
The
commenters
contended
that
the
removal
due
to
this
screening
will
bias
the
database
and
no
longer
represent
the
natural
variability
of
the
operating
conditions
at
a
MWC
facility.
The
commenters
added
that
the
exclusion
of
variability
from
the
dataset
resulted
in
emission
limits
that
are
not
achievable
in
practice
over
the
long
term,
considering
adverse
circumstances.
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
21
One
of
the
commenters
(
OAR­
2005­
0117­
0098)
requested
that
EPA
evaluate
whether
or
not
the
omitted
data
had
a
correlation
to
a
facility's
air
pollution
control
configuration
in
order
to
identify
in
the
proposed
guidelines
representative
facilities
with
a
certain
control
technology.

Another
commenter
(
OAR­
2005­
0117­
0083)
requested
that
EPA
justify
the
data
elimination
and
verify
that
the
eliminations
do
not
skew
the
results.

Several
commenters
(
OAR­
2005­
0117­
0073.1,
OAR­
2005­
0117­
0089,
OAR­
2005­
0117­

0077,
OAR­
2005­
0117­
0080)
cited
a
definition
of
outlier
from
an
EPA
guidance
document,

(
Document
No.
EPA­
QA/
G­
9).
The
commenters
added
that
all
of
the
data
submitted
for
the
MACT
analysis
were
obtained
from
actual
compliance
test
reports
that
were
representative
of
normal
operating
conditions.
As
a
result,
none
of
the
data
collected
fit
the
definition
of
"
outlier"

defined
in
the
EPA
guidance
document,
as
all
the
data
represented
emissions
under
conditions
which
can
reasonably
be
expected
to
recur.

One
commenter
(
OAR­
2005­
0117­
0089)
suggested
that
EPA
did
not
use
any
of
the
recommended
statistical
tests
in
the
guidance
document
for
identifying
outliers
and
suggested
that
EPA
re­
evaluate
the
proposed
emissions
limits
based
on
the
entire
dataset
of
operating
history.
One
commenter
pointed
out
that
standard
statistical
practice
considers
data
greater
than
three
standard
deviations
from
the
mean
to
be
an
outlier.
The
commenter
added
that
EPA
provided
no
justification
for
not
following
standard
statistical
practice.
The
commenter
also
added
that
the
screening
method
only
filtered
out
emission
data
on
the
high
side
of
the
mean
and
did
not
remove
testing
data
that
had
an
unusually
low
emission
rate.

Two
commenters
(
OAR­
2005­
0117­
0073.1,
OAR­
2005­
0117­
0080)
outlined
the
EPA
screening
approach,
where
a
range
of
emission
data
from
the
lowest
emitting
unit
and
the
highest
emitting
unit
is
compared
to
the
mean
range
of
the
MACT
floor
plus
one
standard
deviation.

Although
the
EPA
claimed
that
less
than
14
percent
of
the
data
were
excluded
during
screening,

the
commenter
identified,
based
on
the
pollutant,
between
6
and
20
percent
of
the
NSPS
data,

and
between
13
and
20
percent
of
the
emission
guideline
data
were
removed.
Another
one
of
the
commenters
(
OAR­
2005­
0117­
0092)
added
that
the
alleged
14
percent
of
eliminated
data
is
significant,
especially
since
it
was
primarily
high
emission
rates
that
were
excluded
as
a
result
of
the
screening.

One
commenter
(
OAR­
2005­
0117­
0080)
added
that
the
screening
test
appeared
to
be
designed
to
validate
EPA's
assumptions
about
how
MWCs
should
operate.
The
commenter
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
22
added
that
the
variable
nature
of
solid
waste
as
a
fuel
does
not
support
EPA's
claim
that
similar
units
at
a
single
facility
are
expected
to
have
similar
emissions.
The
commenter
cited
a
court
case,
Cement
Kiln
Recycling
Coalition
v.
EPA,
as
well
as
the
"
EPA
Guidance
document
for
Data
Quality
assessment"
to
demonstrate
that
actual
test
results
should
not
be
removed
as
outliers,
and
that
the
removal
of
outliers
should
be
done
with
extreme
caution.
The
commenter
cited
another
court
case,
Natural
Resources
Defense
Council
v.
Texaco
Refining
and
Marketing,

Inc.,
to
point
out
that
EPA
cannot
exclude
data
simply
because
they
are
outside
of
a
range
of
expected
emission
levels
while
still
being
able
to
justify
using
data
that
are
outside
the
range
of
expected
emission
levels
to
determine
compliance
with
a
standard.

One
commenter
(
OAR­
2005­
0117­
0081)
presented
a
situation
at
one
of
their
MWC
plants
where
two
similar
boilers,
with
identical
pollution
control
equipment
and
fuel
sources,

result
in
different
emissions
for
dioxin,
CO,
and
HCl.
The
commenter
provided
an
example
of
one
of
the
two
boilers
emitting
a
maximum
dioxin
emission
rate
equal
to
the
lowest
dioxin
emission
rate
of
the
other
boiler.
The
commenter
suggested
that
these
differences
between
like
units
are
not
unique
to
their
three
plants
and
the
commenter
suggested
that
EPA
use
all
available
data
for
all
units.

Another
commenter
(
OAR­
2005­
0117­
0088)
contended
that
EPA
incorrectly
assumed
that
similar
units
would
have
similar
emissions
characteristics.
On
the
contrary,
the
commenter
described
that
it
has
two
similar
MWC
units
at
its
facility,
and
noted
that
the
two
units
although
identical
in
equipment,
are
very
different
in
how
they
combust
municipal
solid
waste
(
MSW).

One
commenter
(
OAR­
2005­
0117­
0085)
added
that
EPA
did
not
cite
any
study
to
support
the
hypothesis
that
similar
combustor
types
at
the
same
facility
should
result
in
similar
emissions.

Finally,
one
commenter
(
OAR­
2005­
0117­
0073.1)
suggested
that
the
mean
and
standard
deviation
used
to
calculate
the
an
emission
limit
based
on
the
99%
confidence
limit
are
not
representative
of
the
actual
dataset,
due
to
the
removal
of
data
in
the
screening
analysis.

Response:
As
discussed
fully
in
the
preamble
to
the
final
amendments,
there
are
various
possible
approaches
to
analyzing
the
emissions
data.
Therefore,
to
address
commenters'

concerns,
EPA
has
updated
its
database
and
revised
its
methodology
to
conduct
another
analysis.

In
the
revised
analysis
EPA
has
not
screened
the
data
or
removed
any
data
points
from
the
data
set
for
units
with
the
relevant
control
technologies.
EPA
has
used
the
results
of
the
revised
analysis
as
an
aid
in
determining
the
emission
limits
in
the
final
amendments.
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
23
Comment:
One
commenter
(
OAR­
2005­
0117­
0092)
contended
that
the
eliminated
data
points
include
instances
where
the
proposed
emissions
limits
cannot
be
achieved.
The
commenter
also
stated
that
although
the
EPA
recognizes
the
highly
variable
and
heterogeneous
nature
of
municipal
solid
waste,
the
EPA
does
not
account
for
this
variability
in
its
database.

The
commenter
also
noted
that
trace
metals
can
vary
by
an
order
of
magnitude
prior
to
control,

based
on
the
composition
of
the
waste
stream.
The
commenter
cited
Docket
Item
(
OAR­
2005­

0117­
0048)
and
noted
that
although
their
facility
could
meet
the
proposed
emission
levels,
it
noted
that
several
of
their
semi­
annual
stack
tests
for
cadmium,
lead,
and
PM
would
not
meet
the
proposed
NSPS
PM
limits.

The
same
commenter
also
cited
two
court
cases,
National
Lime
Association
V.
EPA,
and
Northeast
Maryland
Waste
Disposal
Authority
v.
EPA
to
note
the
requirement
under
CAA
129(
a)(
2)
for
EPA
to
account
for
the
variability
of
the
best
performing
unit.
The
commenter
contended
that
by
rejecting
actual
operating
data
from
the
database,
EPA
did
not
meet
the
requirement
to
account
for
this
variability.
The
commenter
also
noted
that
EPA
did
not
provide
any
explanation
for
not
accounting
for
this
variability
over
long­
term
data.

Response:
As
discussed
in
other
responses
and
in
the
preamble
to
the
final
amendments,

EPA
has
revised
its
analysis
and
collected
and
considered
long­
term
emissions
data
in
developing
the
emission
limits
in
the
final
amendments.

5.3
Data
Sources
Comment:
Six
commenters
(
OAR­
2005­
0117­
0088,
OAR­
2005­
0117­
0089,
OAR­
2005­

0117­
0092,
OAR­
2005­
0117­
0076,
OAR­
2005­
0117­
0081,
OAR­
2005­
0117­
0095)
disagreed
with
the
EPA's
decision
to
base
the
proposed
standards
on
compliance
tests
from
the
year
2000
and
requested
EPA
to
use
all
available
data
before
establishing
revised
emission
limits
for
large
MWCs.
The
commenters
added
that
the
year
2000
performance
tests
do
not
reflect
long­
term
performance
of
MWCs
equipped
with
MACT.
Instead,
the
2000
test
data
represent
emissions
from
new
units
with
a
limited
operating
history.
One
commenter
(
OAR­
2005­
0117­
0095)

contented
that
this
data
does
not
represent
the
reduced
equipment
performance
resulting
from
wear
and
tear
over
time.
One
commenter
(
OAR­
2005­
0117­
0092)
contended
that
at
least
five
years
and,
in
some
instances,
as
much
as
nine
years
of
valid
and
comprehensive
data
for
large
MWC
was
readily
available
to
EPA.
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
24
Several
commenters
(
OAR­
2005­
0117­
0088,
OAR­
2005­
0117­
0089,
OAR­
2005­
0117­

0092,
OAR­
2005­
0117­
0076,
OAR­
2005­
0117­
0081)
suggested
that
EPA
should
analyze
stack
test
data
for
years
2000
to
2004
and
reconsider
the
emission
limits
in
the
proposed
rules.
The
commenter
contended
that
given
the
large
variability
in
fuel
make­
up
of
MSW,
a
larger
database
is
warranted
in
order
to
characterize
the
variability,
and
the
commenter
offered
to
provide
its
own
stack
test
data
for
units
with
MACT
equipment
to
EPA.

One
commenter
(
OAR­
2005­
0117­
0076)
added
that
multiple
years
of
test
data
have
been
provided
to
the
Minnesota
Pollution
Control
Agency
that
is
readily
available
to
EPA.
The
commenter
added
that
the
EPA
did
not
provide
a
basis
for
using
a
single
year
of
test
data.

While
one
commenter
(
OAR­
2005­
0117­
0085)
agreed
with
the
use
of
year
2000
compliance
data,
they
also
expressed
concern
for
MWCs
that
have
experienced
increased
emissions
for
one
or
more
pollutants
following
the
compliance
tests
in
2000.
This
commenter
points
out
an
analysis
by
commenter
OAR­
2005­
0117­
0073.1
that
appears
to
indicate
that
some
MWCs
may
have
difficulty
maintaining
continuous
compliance
with
the
proposed
emission
limits.

Response:
Although
EPA
notes
that
the
amount
of
data
used
in
the
proposed
amendments
was
substantial,
EPA
has
supplemented
its
original
data
set
with
test
data
from
2000
to
2005
for
more
than
a
dozen
MWC
units.
This
additional
data
addresses
the
commenters'

concern
of
equipment
performance
over
time.
Further,
the
data
supports
the
fact
that
the
revised
emission
limits
included
in
the
final
amendments
can
be
achieved
over
time.

Comment:
Three
commenters
(
OAR­
2005­
0117­
0085,
OAR­
2005­
0117­
0073.1,
OAR­

2005­
0117­
0077)
stated
that
EPA's
selection
of
year
2000
compliance
test
data
is
an
appropriate
database
from
which
to
derive
new
emission
limits
for
pollutants
measured
by
EPA
reference
methods,
as
long
as
EPA
uses
all
the
data
and
does
not
remove
any
valid
compliance
test
data.

Response:
EPA
agrees
with
the
commenter
that
year
2000
compliance
data
provides
for
a
good
base
year
for
the
rulemaking,
and
EPA
has
revised
its
statistical
analysis
to
use
all
of
the
year
2000
compliance
test
data
from
MWC
units
with
the
relevant
control
technologies
rather
than
removing
any
data.
However,
EPA
has
supplemented
its
dataset
with
2000
to
2005
data
from
over
a
dozen
sources
in
order
to
address
concerns
of
emission
variability
over
a
long
period
of
time.
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
25
Comment:
Two
commenters
(
OAR­
2005­
0117­
0065,
OAR­
2005­
0117­
0073.1)
noted
several
corrections
to
the
background
information
provided
by
EPA.
First,
the
facility
cited
in
York,
Pennsylvania,
does
not
have
SNCR,
but
it
has
SD,
FF,
and
ACI.
Second,
the
MWC
facility
in
Savannah,
Georgia,
is
a
mass
burn
rotary
waterwall
facility
with
SD/
FF/
ACI/
SNCR.

However
since
this
facility
was
not
classified
as
this
combustor
type
in
2000,
it
was
excluded
from
the
database.
The
commenters
also
corrected
the
control
device
category
for
two
units
at
McKay
Bay,
which
resulted
in
13
units
for
the
MACT
database
and
91
units
for
the
NSPS
database.

Response:
After
further
investigation,
EPA
agrees
with
these
corrections.
EPA
has
made
these
corrections
to
its
database
before
performing
a
reanalysis
of
the
data.

Comment:
For
HCl,
one
commenter
(
OAR­
2005­
0117­
0102)
suggested
that
EPA
consider
ten
years
of
stack
test
and
operating
data,
when
developing
the
HCl
limit.
They
concluded
the
data
analyzed
for
the
proposed
limit
inappropriate.

Response:
As
a
result
of
the
additional
analysis,
EPA
concluded
the
final
amendments
contain
the
existing
(
1995)
HCl
emission
limit
and
percent
reduction
requirements.
As
a
result,

the
commenter's
concern
regarding
the
achievability
of
the
HCl
limit
over
the
long
term
is
no
longer
an
issue.

Comment:
Another
commenter
(
OAR­
2005­
0117­
0081)
added
that
two
of
its
circulating
fluidized
bed
boilers
burning
RDF,
wood
waste,
and
railroad
ties
were
not
operating
in
2000,
and
that
data
from
these
units
should
have
been
included
when
establishing
the
proposed
emission
limits.

Response:
EPA
did
not
include
these
units
in
its
revised
analysis.
However,
EPA
has
discussed
the
final
limits
with
the
commenter
and
the
final
limits
have
addressed
the
concerns
the
commenter
had
regarding
the
elimination
of
these
two
data
points.

5.4
Statistical
Methods
Comment:
One
commenter
(
OAR­
2005­
0117­
0065)
requested
that
EPA
apply
a
seasonal
variability
factor
to
the
upper
confidence
limit
of
EPA's
analysis
for
continuously
monitored
pollutants.
The
commenter
suggested
that
EPA
analyze
a
full
year
of
CEM
data
for
one
facility
and
apply
the
seasonal
variability
factor
to
other
facilities.
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
26
Response:
EPA
disagrees
with
applying
a
seasonal
variability
factor
to
its
analysis.
The
revised
analysis
has
incorporated
data
from
2000
to
2005,
which
includes
seasonal
variability
over
time.
EPA
does
not
see
a
need
to
add
in
a
factor
for
seasonal
variability
in
the
revised
dataset.

Comment:
One
commenter
(
OAR­
2005­
0117­
0098)
suggested
that
the
proposed
guidelines
may
require
ESP
replacement
in
order
to
maintain
an
adequate
compliance
margin.

The
commenter
added
that
ESP
controls
tend
to
have
a
larger
variation
in
stack
test
emissions
for
lead
and
cadmium
than
was
accounted
for
in
the
screened
database.
The
commenter
requested
that
EPA
perform
a
separate
statistical
analysis
of
all
ESP
plants
in
order
to
determine
a
separate
class
of
emission
guidelines
for
ESP­
equipped
units.

Response:
As
EPA
has
noted
in
earlier
responses
and
in
the
preamble
to
the
final
amendments,
the
database
has
been
updated,
additional
long­
term
data
has
been
reviewed,
and
the
analysis
methodology
has
been
revised
in
establishing
the
emission
limits
in
the
final
amendments.
The
analysis
for
the
emission
guidelines
is
based
on
units
with
ESPs
and
resulted
in
revised
lead
and
cadmium
limits.

Comment:
One
commenter
(
OAR­
2005­
0117­
0080)
did
not
identify
any
evidence
that
EPA
used
established
statistical
methods
to
determine
if
separate
data
sets
(
i.
e.,
units
with
an
ESP
vs.
units
with
a
FF)
are
similar
enough
to
be
grouped
together.
The
commenter
added
that
standard
methods
exist
to
combine
data
sets
with
different
variability
where
a
pooled
variance
would
be
more
appropriate
than
a
standard
deviation
statistic.

Response:
The
commenter
suggested
another
statistical
procedure.
That
procedure
is
an
acceptable
alternative
approach
for
developing
estimates.
The
suggested
approach
was
not
used
by
EPA.

Comment:
One
commenter
(
OAR­
2005­
0117­
0080)
contended
that
EPA
incorrectly
analyzed
non­
normal
data
distributions.
First,
the
commenter
noted
that
EPA
did
not
correctly
transform
the
data
prior
to
screening
for
outliers.
Second,
EPA
did
not
use
any
established
statistical
tests
for
determining
normality.
The
commenter
provided
the
Kolomogorov­
Smirnov
and
Shapiro­
Wilks
tests
as
examples.
The
commenter
recognized
that
EPA
did
include
the
skewness
and
kurtosis
of
the
dataset
in
the
analysis.
However,
the
commenter
noted
that
particulate
data
from
fabric
filter
MWCs
are
classified
as
a
normal
distribution,
even
though
the
distribution
fails
the
skewness
and
kurtosis
tests.
The
commenter
also
added
that
EPA
log
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
27
normal
transformation
also
failed
the
skewness
and
kurtosis
tests
by
an
even
higher
margin.
The
commenter
contended
that
EPA
did
not
use
any
statistically
valid
technique
to
determine
the
normality
of
the
data.

One
commenter
(
OAR­
2005­
0117­
0080)
suggested
that
is
more
appropriate
to
use
a
onesided
tolerance
interval,
as
compared
to
the
confidence
interval
used
in
EPAs
analysis
of
stack
test
data,
for
determining
the
proposed
emission
limits.
The
commenter
added
that
the
confidence
interval
approach
results
in
a
range
that
will
contain
the
average
of
all
MWC
emission
data
ninety­
nine
percent
of
the
time,
under
ideal
conditions
of
identical
waste
stream,

operations
and
control
device
performance.
The
commenter
emphasized
that
the
confidence
interval
does
not
explain
how
likely
it
is
for
emissions
from
an
individual
unit
to
exceed
the
limit
at
the
upper
confidence
interval.
The
commenter
cited
two
reference
texts
for
further
details
on
tolerance
intervals.

One
commenter
(
OAR­
2005­
0117­
0080)
also
suggested
that
it
is
more
appropriate
to
use
a
one­
sided
tolerance
interval,
as
compared
to
the
confidence
interval
used
in
EPAs
analysis
of
CEM
data,
for
determining
the
proposed
emission
limits.
In
this
case,
where
unrelated
data
sets
were
combined,
the
commenter
suggested
that
the
confidence
interval
statistic
is
even
less
useful.

The
commenter
presented
an
example
of
CEM
data
for
mass
burn
water
wall
nitrogen
oxides
(
NOx)
data,
which
fails
the
skewness
and
kurtosis
tests,
but
is
still
analyzed
as
a
normal
distribution.
The
commenter
also
noted
that
the
log­
normal
transformation
also
fails
these
two
normality
tests,
but
is
still
analyzed
as
a
normal
distribution.
The
commenter
suggested
that
a
square
or
cube
transformation
would
have
brought
the
data
closer
to
a
normal
distribution
and
would
have
been
more
appropriate
for
the
analysis.
The
commenter
concluded
that
EPA
should
have
selected
more
appropriate
tools
for
the
analysis
and
referred
to
EPA
statistical
guidance
documents.

One
commenter
(
OAR­
2005­
0117­
0073.1)
presented
an
alternative
approach
for
analyzing
the
new
MACT
floor
and
what
the
resulting
limits
would
be.
The
commenter
agreed
with
the
EPA's
determinations
of
MACT
as
a
SD/
ESP/
ACI/
SNCR
for
the
emissions
guidelines,

and
SD/
FF/
ACI/
SNCR
for
the
NSPS,
and
the
commenter
kept
only
units
with
these
control
categories
for
further
analysis.
The
commenter
analyzed
all
of
the
year
2000
compliance
data
from
these
units
and
did
not
remove
any
data
points
in
a
screening
procedure.
The
mean,
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
28
standard
deviation,
and
confidence
intervals
at
99%
and
99.7%
were
calculated
for
each
MACT
database.
The
commenter
used
what
they
called
a
"
standard
T­
statistic",
because
they
did
not
understand
why
EPA
apparently
used
one
T­
statistic
for
SD/
ESP/
ACI/
SNCR
units
and
another
for
SD/
FF/
ACI/
SNCR
units.
The
commenter
provided
the
results
of
their
analysis
in
Attachments
6
and
7
of
their
comment.
The
CEM
and
stack
test
emission
limits
were
determined
from
the
calculated
99.7%
upper
confidence
limit
and
the
percent
reduction
limits
were
determined
from
the
99.7%
lower
confidence
limit.
The
commenter's
results
and
recommended
emission
limits
are
summarized
and
compared
with
EPA's
proposal
in
Attachment
12.

Another
commenter
(
OAR­
2005­
0117­
0085)
requested
that
EPA
consider
the
additional
analyses
that
are
submitted
by
OAR­
2005­
0117­
0073.1.
The
commenter
added
that
it
is
imperative
to
reconsider
these
analyses
to
establish
emission
limits
for
the
proposed
rule.

Response:
EPA
conducted
a
revised
statistical
analysis
for
the
promulgated
rule.
The
revised
analysis
addresses
the
commenter's
concerns.
Some
of
the
issues
raised
by
the
commenters
are
discussed
in
this
response,
however
EPA
refers
the
reader
to
see
the
memorandum
in
the
docket
entitled:
Analysis
of
Municipal
Waste
Combustor
(
MWC)
Stack
Test
and
Continuous
Emission
Monitoring
(
CEM)
Data
for
Multiple
Pollutants
in
Units
Equipped
with
SD/
FF/
CI/
SNCR
and
SD/
ESP/
CI/
SNCR
Control
Device
Types,
for
a
detailed
description
of
the
statistical
methods
used
in
the
data
analysis
to
support
the
final
rule.
In
the
revised
analysis
EPA
fully
examined
a
number
of
evaluation
measures
to
determine
normality,
and
applied
data
transformation
as
presented
in
detail
in
the
memorandum
mentioned
above.

Historically,
EPA
has
relied
on
the
distribution
of
test
data
submitted
by
industry
in
establishing
achievable
compliance
levels
for
MWC
units.
In
EPA's
analysis,
an
upper
(
or
lower)

confidence
or
a
critical
limit
(
UCL
or
LCL)
is
defined
as
the
emission
level
that
a
certain
percent
(
e.
g.,
99
percent)
of
the
test
data
would
fall
below
(
or
above)
based
on
the
distribution
of
the
dataset.
For
example,
if
the
data
distribution
is
normal,
the
UCL
(
or
LCL)
is
determined
by
the
critical
values
available
from
the
z
(
or
t­
distribution)
tables.
Note
that
the
term
upper
(
or
lower)

confidence
limits
as
computed
in
EPA's
analysis,
does
not
refer
to
the
upper
(
or
lower)

confidence
limit
one
would
compute
for
a
population
mean
based
on
a
sample.
The
commenter
suggests
yet
another
possible
method
to
use
tolerance
intervals,
as
opposed
to
confidence
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
29
intervals
to
analyze
the
data.
EPA
investigated
this
statistical
approach
and
the
results
are
contained
in
a
separate
memorandum
in
the
docket.

The
revised
analysis
uses
the
t­
statistic
for
both
the
SD/
ESP/
ACI/
SNCR
units
and
the
SD/
FF/
ACI/
SNCR
units.
The
Student's
t
distribution
is
an
approximation
of
the
normal
distribution
for
small
samples,
and
is
appropriate
for
the
size
of
the
data
sets,
since
the
dataset
includes
only
13
units
with
ESP­
based
control
systems.
The
previous
analysis
used
the
zstatistic
for
the
SD/
FF/
ACI/
SNCR
dataset,
which
has
a
larger
number
of
data
points.
The
distribution
characteristics
approach
those
of
the
normal
distribution
as
sample
size
increases.

Typically,
for
sample
sizes
greater
than
30,
the
t­
statistics
are
equivalent
to
the
z­
statistics
computed
from
the
normal
distribution.
The
use
of
either
statistic
(
z­
statistic
versus
t­
statistic)
in
the
computations
yields
the
same
result
for
the
units
with
FF­
based
controls.

EPA
continues
to
believe
that
the
use
of
a
99%
upper
confidence
limit
(
UCL)
is
appropriate
for
"
stack
test"
pollutants.
A
99.7%
UCL
is
excessive.
EPA
provides
further
reasoning
on
its
selection
of
this
confidence
limit
in
the
preamble
to
the
promulgated
rule.

EPA
has
partially
relied
on
the
revised
statistical
analysis
in
establishing
achievable
compliance
levels
for
MWC
units.
However,
statistical
methods
alone
have
never
dictated
the
agency's
decision
on
achievable
emission
limits.
Previous
responses
describe
EPA's
method
for
considering
the
statistical
analysis,
along
with
other
information,
to
select
the
final
emission
limits.

Finally,
EPA
has
taken
into
account
the
analysis
referenced
by
commenters
OAR­
2005­

0117­
0073.1
and
OAR­
2005­
0117­
0085.
The
results
of
EPA's
revised
analysis
using
the
updated
data
are
in
fact
equivalent
for
most
pollutants
with
the
results
of
the
referenced
analysis
submitted
by
commenter
OAR­
2005­
0117­
0073.1.

Comment:
One
commenter
(
OAR­
2005­
0117­
0080)
disagreed
with
combining
15
yearold
CEM
data
with
2000
MACT
data
in
order
to
establish
variability
factors.
The
commenter
suggested
that
EPA
should
have
developed
variability
factors
on
the
2000
MACT
data
alone.

Response:
EPA
has
removed
the
15­
year­
old
CEM
data
from
its
revised
analysis.
The
revised
analysis
considers
year
2000
data,
a
recent
supplemental
dataset
of
CEMS
data,
and
additional
long­
term
data
for
over
a
dozen
units
for
the
period
from
2000
to
2005.
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
30
Comment:
One
commenter
(
OAR­
2005­
0117­
0080)
requested
that
EPA
withdraw
the
proposal
until
a
scientifically
valid
data
analysis
has
been
conducted.

Response:
As
discussed
in
earlier
responses
and
in
the
preamble
to
the
final
amendments,
the
dataset
has
been
updated
and
expanded,
and
the
analysis
has
been
revised
for
the
final
amendments.
These
measures
adequately
address
many
of
the
concerns
over
the
scientific
validity
of
the
data
analysis
supporting
the
proposed
amendment
emission
limits.

Please
refer
to
the
revised
analysis
documentation
located
in
the
docket
to
the
amendments
for
a
detailed
description
of
statistical
methods
employed
to
support
the
selection
of
the
final
emission
limits.

Comment:
One
commenter
(
OAR­
2005­
0117­
0080)
requested
that
EPA
provide
better
documentation
of
their
data
and
specifically
asked
EPA
to
provide
access
to
the
actual
spreadsheets
used,
or
provide
full
and
complete
calculations
as
opposed
to
Excel
formulas.

Response:
EPA
elected
to
submit
data
to
the
docket
in
.
pdf
format
for
convenience.
The
size
of
the
particular
file
was
420
kb
in
Excel
®
format
and
was
reduced
to
119
kb
in
.
pdf
format.

Additionally,
the
Federal
Docket
Management
System
(
FDMS)
accepts
data
in
many
formats
including
Excel
®
,
but
converts
all
files
to
.
pdf
format
before
entering
the
information
into
the
docket.
The
EPA
submission
of
data
in
.
pdf
format
simply
saved
a
step
in
the
file
conversion
process
for
uploading
the
data
for
the
FDMS
website.
The
data
in
.
pdf
format
is
fully
satisfactory
for
review
by
the
public.
Additionally,
if
a
spreadsheet
of
the
data
is
needed
 
the
data
in
.
pdf
format
can
be
copied
to
create
a
new
spreadsheet.

Comment:
Four
commenters
(
OAR­
2005­
0117­
0073.1,
OAR­
2005­
0117­
0094,
OAR­

2005­
0117­
0087,
OAR­
2005­
0117­
0064)
contended
that
the
99th
percentile
used
by
EPA
in
its
analysis
does
not
incorporate
an
adequate
compliance
margin.
One
commenter
(
OAR­
2005­

0117­
0073.1)
supported
the
use
of
a
99.7
percentile
for
deriving
an
emission
limit
for
pollutants
measured
by
either
CEM
or
manual
stack
test
methods,
stating
that
a
99.7
confidence
interval
is
a
more
appropriate
statistic
to
use
for
calculating
enforceable
emission
limits
that
incorporate
an
adequate
compliance
margin.
The
commenter
recognized
EPA's
basis
for
selecting
this
probability
is
one
noncompliant
event
per
year,
which
works
out
to
a
99.726
percentile
for
a
daily
average
on
CEM
pollutants.
However,
the
commenter
suggested
that
the
manual
test
method
is
based
on
a
12­
hour
compliance
period,
half
as
long
as
the
daily
average
period
for
CEM
pollutant.
The
resulting
probability
of
a
noncompliant
event
would
be
99.863
for
a
12
hour
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
31
sample
period.
Therefore,
they
disagreed
with
EPA's
use
of
the
99.0
percentile
for
stack
test
pollutants.

Response:
In
the
preamble
to
the
final
amendments,
EPA's
rationale
for
selecting
the
99th
percentile
is
discussed
in
detail.
EPA
refers
the
commenters
to
the
preamble
discussion
to
respond
to
this
comment.

5.5
Achievability
of
Emission
Limits
Comment:
Two
commenters
(
OAR­
2005­
0117­
0073.1,
OAR­
2005­
0117­
0077)
declared
that
the
proposed
emission
limits
are
not
achievable
under
the
worst
foreseeable
circumstances
and
the
commenter
states
legal
precedence
supporting
previous
interpretations
of
the
phrase
`
emission
control
that
is
achieved
in
practice'
to
mean
`
achieved
under
the
worst
foreseeable
circumstances.'
The
commenters
cited
cases
from
1980
and
2000
where
the
Court
explained
that
standards
must
be
capable
of
being
met
under
adverse
conditions
that
can
reasonably
be
expected
to
recur.
The
commenter
argued
that
emission
limits
must
reflect
a
range
of
operating
and
maintenance
conditions,
waste
loads,
and
emissions
and
control
technology
performance.

Provided
that
the
review
of
the
MACT
standard
is
based
on
actual
emission
data,
the
commenter
requested
that
all
of
the
emissions
data
it
submitted
to
EPA
be
considered
in
determining
the
achievable
emission
limits
in
the
proposed
rule.

Response:
EPA
has
revised
the
emission
limits
for
the
final
rule
and
concluded,
based
on
the
data
used
in
the
revised
analysis,
that
these
revised
limits
are
achievable
over
long
periods
of
time
and
under
the
worst
foreseeable
circumstances.

Comment:
One
commenter
(
OAR­
2005­
0117­
0065)
requested
that
EPA
set
the
proposed
limits
for
Subpart
Eb
based
on
the
achievable
limits
at
a
single
facility
using
the
level
of
technology
described
in
the
rule.
While
the
commenter
does
not
operate
any
new
MWC
facilities,
they
note
that
none
of
their
existing
facilities
produce
emissions
near
the
levels
proposed
for
new
units.
The
commenter
is
concerned
that
there
are
no
guarantees
from
equipment
manufacturers
that
their
own
facility's
existing
control
equipment
can
meet
the
proposed
limits.
In
the
event
that
EPA
will
not
consider
basing
the
limit
on
the
emission
level
achieved
by
the
technology
at
a
single
facility,
the
commenter
requested
that
the
proposed
limits
be
delayed
until
engineering
and
design
companies
of
pollution
control
equipment
can
confirm,

with
an
adequate
compliance
margin,
that
the
proposed
limits
are
achievable.
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
32
Response:
The
commenter
was
concerned
with
guarantees
from
control
device
equipment
manufacturers
to
meet
the
proposed
standards
and
EPA
believes
it
has
addressed
their
concerns
in
the
final
amendment
limits.
EPA
has
revised
its
statistical
analysis,
analyzed
additional
longterm
data,
and
considered
public
comments
in
developing
revised
emission
limits,
as
explained
in
detail
in
other
responses.
All
of
the
data
analyzed
indicate
that
new
MWC
units
with
SD/
FF/
ACI/
SNCR
are
capable
of
meeting
the
emission
limits
included
in
the
final
amendments.

Comment:
Two
commenters
(
OAR­
2005­
0117­
0073.1,
OAR­
2005­
0117­
0065)

requested
that
EPA
consider
a
compliance
margin
that
properly
reflects
the
variations
in
emissions
resulting
from
the
heterogeneous
nature
of
municipal
solid
waste.
The
commenters
noted
that
MACT
limits
must
consider
the
expected
variability
from
normal
operations,
and
a
variety
of
materials
comprising
the
solid
waste
fuel
is
one
example
of
expected
variability.
One
commenter
(
OAR­
2005­
0117­
0073.1)
cited
a
2004
precedent
case
of
allowing
for
wide
variability
in
emissions
when
establishing
a
MACT
standard
for
vinyl
chloride
to
support
its
request
to
allow
for
variability
resulting
from
waste
mixtures.
One
commenter
(
OAR­
2005­

0117­
0065)
added
that
the
variability
in
emissions
makes
it
necessary
to
have
a
margin
to
obtain
performance
guarantees
for
air
pollution
control
equipment
when
the
equipment
is
being
purchased.
The
commenter
added
that
without
this
compliance
margin,
it
is
difficult
to
ensure
continual
compliance.
Yet
another
commenter
(
OAR­
2005­
0117­
0085)
added
that
EPA
may
not
have
provided
an
adequate
margin
for
compliance
under
all
reasonably
foreseeable
operating
conditions
in
establishing
the
proposed
emission
limits.

Response:
EPA
agrees
that
the
emissions
limits
should
consider
emissions
variability
and
EPA
has
taken
several
steps
to
correct
for
this
consideration
in
the
final
rule.
The
data
screening
procedure
has
been
removed,
which
allows
for
more
emissions
variability
in
the
data
set,
and
the
original
dataset
has
been
supplemented
with
2000
to
2005
data
from
over
12
MWC
units
in
order
to
analyze
emission
variability
over
a
longer
period
of
time.
Please
see
the
preamble
to
the
final
amendments
for
a
more
detailed
discussion
of
measures
taken
to
ensure
that
the
final
limits
address
emission
variability.

5.6
Removal
Efficiency
Comment:
Three
commenters
(
OAR­
2005­
0117­
0094,
OAR­
2005­
0117­
0064,
OAR­

2005­
0117­
0087)
proposed
that
EPA
consider
limiting
factors
inherent
in
the
design
and
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
33
processes
of
air
pollution
control
technologies
before
applying
statistics
to
the
data
to
obtain
percent
removal
efficiency
limits.

Response:
The
final
amendments
retain
the
existing
(
1995)
percent
reduction
requirements
for
HCl
and
sulfur
dioxide
(
SO2).
Given
that
there
are
no
changes
made
to
these
removal
efficiencies,
EPA
has
addressed
the
concerns
from
commenters
on
reaching
the
practical
limits
of
acid
gas
control
technologies.

Comment:
One
commenter
(
OAR­
2005­
0117­
0073.1)
disagreed
with
the
use
of
a
lower
confidence
interval
(
LCL)
to
determine
the
removal
efficiency
in
the
proposed
amendments
for
Hg,
HCl,
and
SO2.
The
commenter
suggested
that
the
LCL
does
not
adequately
consider
that
a
removal
efficiency
condition
is
a
minimum
performance
level.

The
same
commenter
suggested
that
EPA
use
the
confidence
interval,
at
a
probability
of
99.726,
for
determining
the
minimum
percent
removal
achieved
in
practice
for
the
amendments
in
the
final
rule.
The
commenter
added
that
the
use
of
one
standard
statistic
provides
for
a
consistent
rational
in
calculating
both
emission
limits
for
percent
reduction
and
stack
test
and
CEM
limits.
The
commenter
suggested
that
the
percent
reduction
limit
be
calculated
by:
[
mean
 
STD
DEV
*
t]
since
the
unit
must
operate
above
the
percent
reduction
limit,
whereas
the
emission
limit
should
be
calculated
by
[
mean
+
STD
DEV
*
t],
consistent
with
the
EPA
analysis,

since
the
limit
is
a
value
not
to
be
exceeded.

Response:
As
noted
before
for
HCl
and
SO2,
the
final
amendments
retain
the
existing
(
1995)
removal
efficiency
requirements
for
HCl
and
SO2.
EPA
has
also
opted
to
retain
the
existing
removal
efficiency
from
the
1995
rulemaking
for
Hg
in
the
final
amendments.
The
statistical
analysis
conducted
for
the
proposed
amendments
was
not
used
as
the
basis
for
the
final
percent
reduction
limits.

Comment:
Two
commenters
(
OAR­
2005­
0117­
0073.1,
OAR­
2005­
0117­
0077)
objected
to
the
procedures
used
by
EPA
to
determine
percent
reduction
requirements
for
HCl,
and
concluded
that
the
current
95%
removal
efficiency
requirement
for
HCl
should
be
maintained
in
the
final
amendments
to
the
rule.
Specifically,
the
commenters
noted
that
the
97
and
98%

removal
efficiencies
for
HCl,
on
existing
and
new
units,
respectively,
does
not
consider
the
need
for
a
1
to
2%
compliance
margin,
the
inherent
limitations
of
the
removal
process
itself,
and
errors
inherent
in
the
manual
stack
test
method.
The
commenter
added
that
existing
facilities
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
34
meeting
the
95%
removal
efficiency
actually
operate
at
1
to
2%
above
this
requirement
to
ensure
compliance.

One
of
these
commenters
(
OAR­
2005­
0117­
0073.1)
noted
that
58
percent
of
the
SD/
FF/
ACI/
SNCR
units
in
the
NSPS
database
would
not
have
met
the
proposed
98%
reduction
requirement
for
HCl
with
a
1
to
2
percent
compliance
margin.
The
commenter
added
that,
when
such
a
high
percentage
of
the
units
in
the
dataset
cannot
meet
the
proposed
limits
with
an
adequate
compliance
margin,
the
proposed
emissions
standards
are
inappropriate.

Further,
the
same
commenter
(
OAR­
2005­
0117­
0073.1)
submitted
a
long­
term
dataset
(
for
years
1994
to
2001)
from
a
waste­
to­
energy
facility
in
Lee
County,
FL
that
was
recognized
as
a
model
MACT
plant.
Upon
analyzing
this
data,
the
commenter
found
that
the
long
term
results
did
not
show
continuous
compliance
with
the
proposed
97%
reduction
for
Subpart
Cb
nor
the
proposed
98%
reduction
for
Subpart
Eb.
The
commenter
suggested
that
a
lack
of
continuous
compliance
in
the
long­
term
supports
the
claim
that
a
facility
will
require
a
1
to
2%
compliance
margin
to
assure
that
a
removal
efficiency
is
being
met.
The
commenter
also
suggested
that
the
long­
term
data
from
Lee
County
demonstrates
a
limit
on
the
performance
of
a
semidry
scrubbing
process
in
the
range
of
95
to
97%.

Two
commenters
(
OAR­
2005­
0117­
0073.1,
OAR­
2005­
0117­
0096)
suggested
that
extending
this
same
compliance
margin
to
the
new
98%
reduction
standard
requires
the
removal
efficiency
to
be
run
between
99
and
100%,
which
is
unrealistic
given
the
limits
on
the
removal
efficiencies
of
the
process.

Response:
EPA
agrees
with
not
increasing
the
removal
efficiencies
to
unrealistic
levels,

and
it
recognizes
the
need
for
a
compliance
margin.
As
noted
previously,
EPA
is
retaining
all
of
the
percent
reduction
requirements
from
the
existing
rule,
which
adequately
addresses
the
concerns
of
these
commenters.

Comment:
One
commenter
(
OAR­
2005­
0117­
0102)
suggested
that
EPA
establish
the
97%
removal
efficiency
for
HCl
as
an
emission
reduction
goal
rather
than
an
enforceable
limit.

The
commenter
contended
that
EPA
did
not
consider
additional
impacts
on
facility
operations,

costs
and
the
environment
from
increased
ash
disposal,
that
result
from
the
proposed
HCl
removal
efficiency.
The
commenter
suggested
that
EPA
should
avoid
regulation
that
increases
waste
streams
to
landfills.
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
35
Response:
Under
the
regulation,
EPA
is
required
to
establish
an
enforceable
limit
and
cannot
establish
a
"
goal"
as
recommended
by
the
commenter.
However,
EPA
is
retaining
all
of
the
percent
reduction
requirements
from
the
existing
rule,
which
addresses
the
concerns
of
the
commenter.

5.7
New
Control
Technologies
Comment:
One
commenter
(
OAR­
2005­
0117­
0107)
requested
that
EPA
consider
emission
limits
for
all
pollutants
that
reflect
U.
S.
and
global
experience
in
the
final
rule
amendments.
The
commenter
suggested
that
often
times
control
technologies
are
transferable
between
source
categories
and
transferable
between
similar
sources
in
foreign
countries
and
the
United
States.
The
commenter
also
suggested
that
limited
construction
of
new
MWC
units
since
the
1995
rulemaking
has
skewed
the
top
performing
12
percent
of
units
towards
older
control
technology
applications
with
respect
to
more
recent
global
air
pollution
control
technology
experience.
The
commenter
contended
that
the
more
advanced
globally
demonstrated
control
techniques
are
particularly
relevant
for
development
emission
limits
for
MWCs
since
the
proposed
standards
relied
on
the
use
of
data
from
a
small
set
of
control
systems
that
were
installed
more
than
a
decade
ago
and
this
data
set
does
not
indicate
the
recent
advances
achieved
by
emissions
control
technologies.
The
commenter
referenced
a
paper
on
good
combustion
practices
that
was
used
in
response
to
the
EPA
MACT
rulemaking.

The
commenter
also
requested
that
EPA
compare
the
performance
of
the
catalytic
dioxin
emissions
control
systems
in
Europe
to
the
proposed
limits
for
new
and
existing
MWCs
in
the
US.
The
commenter
presents
a
table
representing
43
units
firing
MSW,
RDF,
industrial
waste,

biomass
waste
and
pharmaceutical
wastes
to
show
the
experience
of
this
control
technology.

The
table
includes
two
installations
of
combustion
units
in
the
US
that
combust
hazardous
liquid
waste.
The
remaining
catalytic
dioxin
installations
were
located
in
Asia
and
Europe.
The
commenter
cited
performance
specifications
of
this
technology
to
yield
greater
than
99.9%

destruction
of
dioxins
and
furans
down
to
0.1
nanograms
toxic
equivalents
per
normal
cubic
meter
(
ng
TEQ/
Nm3).
The
commenter
referenced
a
paper
to
support
the
performance
of
the
catalytic
dioxin
destruction
technologies.
The
commenter
added
that
one
type
of
this
dioxin
removal
technology
has
proven
to
be
effective
at
very
low
temperatures,
which
would
allow
this
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
36
technology
to
be
easily
retrofitted
at
the
end
of
the
flue
gas
treatment
system,
just
prior
to
the
stack.

Response:
EPA
has
considered
the
commenter's
issues
and
it
disagrees
with
the
commenter.
The
final
rule
is
based
upon
a
technical
review
of
the
performance
of
existing
large
MWC
units
in
the
United
States.

Comment:
One
commenter
(
OAR­
2005­
0117­
0107)
requested
that
EPA
consider
the
performance
of
new
selective
non­
catalytic
reduction
(
SNCR)
installations
as
well
as
selective
catalytic
reduction
(
SCR)
technologies
that
achieve
emission
rates
lower
than
150
ppm
NOx.

The
commenter
suggested
that
recent
advances
in
SNCR
technologies,
including
chemical
injection
hardware,
wider
applicable
temperature
ranges,
and
advanced
process
control
have
resulted
in
cost
effective
NOx
reductions
less
than
100
ppm
at
MWCs.
The
commenter
recognized
that
new
SNCR
installations
are
typically
more
effective
than
retrofit
applications.

However,
the
commenter
also
presented
an
example
from
four
different
MWC
units
in
Pennsylvania
and
Florida
to
that
were
permitted
to
meet
NOx
emissions
between
110
and
150
ppm
with
retrofit
SNCR
control
systems.

With
respect
to
SCR
systems,
the
commenter
presented
an
experience
list
of
SCR­
control
upgrades
performed
on
German
MWC
units.
The
commenter
added
that
several
of
the
MWC
units
have
been
in
operation
for
over
17
years.
The
commenter
suggested
that
SCR
technology
can
also
be
retrofitted
and
operated
at
low
temperatures.
The
commenter
added
that
many
MWC
plants
in
Europe
and
Asia
operate
a
combined
SCR
and
dioxin/
furan
control
system.
The
commenter
added
that
this
combination
is
efficient
and
possible
because
the
same
catalytic
reactor
can
be
used
for
both
the
SCR
reaction
and
the
catalytic
dioxin
destruction.

Response:
EPA
has
considered
the
commenter's
issues
and
it
disagrees
with
the
commenter.
The
MWC
units
with
SNCR
retrofits
in
Pennsylvania
and
Florida
mentioned
by
the
commenter
are
permitted
with
these
NOx
values,
however
the
MWC
units
are
still
being
built,

and
therefore
were
not
included
in
EPA's
review
of
actual
data
from
existing
large
MWC
units.

To
the
best
of
EPA's
knowledge
and
available
data
during
the
rulemaking,
no
units
are
operating
subject
to
emission
limits
less
than
150
ppmv.
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
37
5.8
Carbon
Monoxide
Limits
Comment:
One
commenter
(
OAR­
2005­
0117­
0084)
requested
that
EPA
review
all
units
that
would
belong
to
the
new
proposed
combustor
categories
before
finalizing
the
CO
emission
standards
for
these
two
new
subcategories.
The
commenter
added
that
EPA
did
not
justify
why
these
subcategories
were
established
or
why
these
units
have
an
elevated
CO
emission
standard.

The
commenter
contended
that
there
are
three
spreader
stoker
refuse­
derived
fuel­
fired
combustor/
100
percent
coal­
capable
that
meet
the
existing
CO
standard.
The
commenter
also
added
that
the
facility
for
which
EPA
based
its
standard
may
not
be
optimally
designed
or
operated
to
limit
CO
emissions.

Response:
EPA
recognizes
that
the
description
of
the
subcategory
in
the
proposal
was
too
general
and
has
obtained
additional
details
on
the
unique
design
of
the
units
that
were
intended
to
be
covered
by
this
new
subcategory.
The
subcategory
description
(
definition)
has
been
revised
to
reflect
the
unique
design
and
to
exclude
the
units
cited
by
the
commenter.
The
MWCs
units
at
the
MWC
plant
analyzed
now
represent
the
entire
combustor
subcategory.

Comment:
One
commenter
(
OAR­
2005­
0117­
0073.1)
disagreed
with
the
proposed
geometric
mean
over
the
24­
hour
period
for
CO
limits
for
fluidized
bed
mixed
fuel
units.
The
commenter
requested
that
the
24­
hour
arithmetic
average
for
CO
be
retained
for
all
combustor
types
because
they
were
unaware
of
any
reason
why
a
geometric
mean
was
used
for
only
this
category.

Response:
EPA
has
resolved
this
issue
with
the
commenter.
The
CO
limit
for
fluidized
bed
mixed
fuel
units
was
promulgated
in
an
earlier
rulemaking
and,
because
it
was
not
part
of
this
proposal,
it
has
not
been
changed.

Comment:
One
commenter
(
OAR­
2005­
0117­
0097)
disagreed
with
the
proposed
change
to
the
averaging
time
for
CO
emissions
for
Spreader
stoker
refuse­
derived
fuel­
fired
combustor/
l00
percent
coal
capable
units
to
go
to
a
24­
hour
block
average,
geometric
mean.
The
commenter
estimated
that
its
cost
would
increase
in
order
to
reprogram
the
logic
controllers
and
CEM
software,
and
the
commenter
did
not
see
any
benefit,
compared
to
the
arithmetic
mean,
in
changing
this
reporting
requirement.

Response:
EPA
has
resolved
this
issue
with
the
commenter.
The
reasons
for
the
revised
CO
standards
are
documented
in
multiple
memoranda
that
are
included
in
the
docket,
and
EPA
maintains,
based
on
its
analyses,
the
proposed
CO
limits
are
appropriate.
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
38
5.9
Dioxin/
Furan
Limits
Comment:
One
commenter
(
OAR­
2005­
0117­
0095)
requested
that
EPA
consider
one
of
the
following
options
when
establishing
the
final
rule:
(
i)
retain
the
current
dioxin
criteria
for
existing
MWC
units,
(
ii)
apply
the
new
dioxin
criteria
to
new
MWC
units,
or
(
iii)
develop
a
subcategory
of
units
for
plants
that
co­
fire
RDF
and
biomass.
The
commenter
stated
that
the
MWC
facility
at
French
Island,
Wisconsin,
is
the
only
operating
fluidized­
bed,
refractory
wall
MWC
in
the
country,
and
the
performance
capabilities
are
affected
by
the
excess
air
required
to
operate
the
fluidized
bed.
The
commenter
added
that
the
fuel
mix
of
RDF
and
wood
requires
a
different
set
of
operating
conditions
and
requests
that
the
proposed
rule
recognize
a
different
set
of
operating
parameters
(
such
as
temperature
and
excess
oxygen).
The
commenter
also
added
that
the
rule
should
provide
for
an
allowance
in
the
operating
conditions
during
periods
when
a
RDF
and
wood
co­
fired
unit
burns
wood
waste
alone.
The
commenter
was
concerned
that
they
would
have
to
modify
the
MWC
at
significant
expense
in
order
to
meet
the
proposed
emission
criteria.

Response:
EPA
has
clarified
the
concerns
of
the
commenter,
and
has
verified
that
the
final
amendments
address
their
concerns.
The
final
dioxin
standard
addresses
these
concerns.

5.10
HCl
Limits
Comment:
One
commenter
(
OAR­
2005­
0117­
0068)
contended
that
the
existing
HCl
limit
reflects
MACT
with
adequate
attention
to
emission
variability.
The
commenter
added
that
if
EPA
maintains
the
proposed
26
parts
per
million
dry
volume
(
ppmdv)
limit,
then
the
alternative
HCl
removal
efficiency
should
be
maintained
at
95%.

The
commenter
presented
three
years
of
HCl
data
from
its
Pinellas
county
Resource
Recovery
Facility
to
demonstrate
that
the
SD/
FF
acid
gas
control
configuration
meets
the
existing
95%
removal
efficiency
requirement
for
HCl,
but
it
did
not
meet
the
proposed
97%

removal
efficiency
in
2003.
The
commenter
also
added
that
the
facility
met
the
proposed
26
ppmdv
HCl
limit
with
a
reduced
compliance
margin
of
15%.

Another
commenter
(
OAR­
2005­
0117­
0096)
requested
that
EPA
provide
additional
technical
data
and
discussions
to
confirm
the
basis
for
the
proposed
HCl
emission
limits.
The
commenter
added
that
many
of
the
MWC
units
do
not
have
HCl
CEM
data
to
reflect
the
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
39
variability
in
HCl
emissions,
and
that
if
the
proposed
HCl
limits
are
based
on
stack
test
results,

the
commenter
contended
this
data
provided
only
a
limited
sampling
of
HCl
emissions.
Further,

the
commenter
suggested
that
although
increased
reagent
may
slightly
increase
removal
efficiencies
for
HCl,
this
extra
reagent
will
not
be
able
to
ensure
compliance
during
spikes
in
HCl.

Response:
As
discussed
in
earlier
responses,
EPA
has
retained
all
of
the
existing
(
1995)

limits
and
percent
reduction
requirements
for
HCl
in
the
final
amendments.
As
a
result,
EPA
has
addressed
the
commenters'
concerns
about
meeting
the
limits
for
HCl.

5.11
Mercury
Limits
Comment:
Two
commenters
(
OAR­
2005­
0117­
0100,
OAR­
2005­
0117­
0086)

recommended
further
tightening
of
the
mercury
emission
limits.
One
commenter
(
OAR­
2005­

0117­
0100)
recommended
that
EPA
tighten
the
standards
for
mercury
to
28
micrograms
per
cubic
meter
corrected
to
7
percent
oxygen
(
O2)
or
at
least
95%
removal
efficiency
for
both
Subparts
Cb
and
Eb.
The
commenter
emphasized
that
it
has
over
10
years
of
experience
in
regulating
MWC
emissions
from
13
MWC
combustors
(
7
with
fabric
filters
and
6
with
ESPs).

Further,
the
commenter
identified
that
all
of
the
units
in
its
jurisdiction
are
meeting
the
95%

removal
efficiency
regardless
of
whether
the
unit
has
an
ESP
or
fabric
filter
control.
The
commenter
added
that
although
their
State
rules
were
revised
in
2004
to
require
the
mercury
standards
proposed
by
the
commenter,
none
of
the
facilities
had
to
go
through
major
modifications
to
achieve
the
revised
emission
standard
of
28
micrograms
per
standard
cubic
meter
or
at
least
95%
reduction.
In
fact,
the
commenter
presented
data
to
show
that
all
of
the
facilities
in
the
jurisdiction
are
operating
between
5
and
10
micrograms
per
cubic
meter
corrected
to
7
percent
O2.
The
commenter
attached
a
summary
of
the
MWC
test
data
from
the
jurisdiction.

Another
commenter
(
OAR­
2005­
0117­
0086)
recommended
that
EPA
promulgate
the
emission
standards
for
Hg
to
be
28
micrograms/
dscm
or
at
least
85%
removal
efficiency.
The
commenter
also
recommended
that
EPA
limit
annual
mercury
emissions
to
35
lb/
year
for
each
unit.
They
explained
that
these
limits
are
consistent
with
their
State
rules
and
have
been
met
by
the
4
large
MWCs
in
their
State.
The
commenter
supported
the
reduced
mercury
requirements
by
presenting
the
negative
effects
of
mercury
on
water
pollution
and
healthcare
costs
in
their
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
40
state.
In
addition,
the
commenter
presented
a
list
of
procedures
the
state
has
used
to
reduce
mercury
pollution.
However,
the
commenter
cited
a
NESCAUM
study
that
concluded
that
the
state's
mercury
problems
stems
from
airborne
transport
of
Hg.
The
commenter
also
noted
that
MWC
are
the
second
largest
contributor
to
Hg
emissions
and
the
commenter
concluded
that
the
more
stringent
reductions
in
the
federal
rules
for
mercury
are
required
to
reduce
impacts
from
airborne
transport.

Response:
Based
on
a
review
of
the
updated
database,
EPA
reduced
the
Hg
limits
for
mercury
downward
from
the
limits
presented
in
the
proposed
amendments.

Comment:
One
commenter
(
OAR­
2005­
0117­
0095)
requested
that
the
proposed
standards
consider
the
unique
circumstances
of
its
MWC
facility
including
its
combustor
type,

fuel
mix,
and
RDF
processing.
The
commenter
explained
that
the
RDF
fuel
burned
at
its
facility
has
been
processed
to
remove
most
of
the
heavy
metals.
As
a
result
of
this
RDF
processing,
the
commenter
suggested
that
activated
carbon
injection
is
not
required
in
order
to
control
mercury
emissions.

Response:
EPA
emphasizes
that
MACT
standards
do
not
require
a
certain
technology
in
order
to
meet
the
mercury
limit.
The
emission
limits
are
based
on
application
of
a
certain
technology,
but
the
rules
only
require
that
the
facility
complies
with
the
emission
limit.
If
a
facility
can
meet
the
limit
without
ACI,
then
the
facility
is
still
in
compliance
with
the
mercury
limit.

5.12
Lead
Limits
Comment:
One
commenter
(
OAR­
2005­
0117­
0065)
requested
that
EPA
maintain
the
existing
limits
for
lead
for
ESP
facilities.
The
commenter
added
that
the
only
method
for
reducing
lead
emissions
at
facilities
where
the
proposed
limit
cannot
be
achieved
is
by
replacing
the
ESP
with
a
fabric
filter.

Response:
The
revised
final
amendment
limits
for
lead
resolve
the
compliance
concerns
that
the
commenter
had
with
proposed
amendment
standards.
As
described
in
another
response
and
in
the
preamble,
the
lead
limit
selected
for
the
final
amendments
is
higher
than
in
the
proposed
amendments.
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
41
5.13
NOx
Limits
Comment:
One
commenter
(
OAR­
2005­
0117­
0099)
requested
that
EPA
make
a
change
to
incorporate
the
uniqueness
of
the
mass
burn
tumbling
tile
combustor
at
the
Savannah
Resource
Recovery
Facility.
One
commenter
(
OAR­
2005­
0117­
0121)
provided
NOx
emissions
for
this
facility
for
the
entire
year
of
2005.
This
commenter
also
described
the
changes
made
to
the
retrofit
SNCR
system
and
the
results
for
test
runs
at
the
facility
once
changes
were
made
to
its
SNCR
system.
The
test
results
indicated
that
the
facility
would
have
been
out
of
compliance
with
the
158
ppmdv
proposed
NOx
limits.
The
commenter
added
that
the
since
the
tests
were
conducted
in
the
winter,
in
the
absence
of
yard
waste,
the
uncontrolled
NOx
levels
are
lower
than
typical
summer
levels.

One
of
the
commenters
(
OAR­
2005­
0117­
0099)
requested
that
EPA
revise
the
proposed
NOx
limit
for
mass
burn
rotary
waterwall
combustors
to
205
ppmdv
as
opposed
to
the
proposed
158
ppmdv.
Alternately,
the
commenter
requested
that
if
the
proposed
NOx
limit
is
not
changed,

then
EPA
should
change
the
classification
of
the
unit
at
the
facility,
which
was
classified
as
a
"
mass
burn
rotary
waterwall
combustor,"
to
a
"
mass
burn
tumbling
tile
combustor"
with
respect
to
NOx
and
CO
emissions,
since
this
is
the
only
Large
MWC
facility
of
this
type
in
the
country.

The
commenter
added
that
a
NOx
emission
limit
of
205
ppmdv
and
a
CO
limit
of
250
ppmdv
be
applied
to
this
new
combustor
category.
The
commenter
added
that
205
ppmdv
is
their
current
enforceable
NOx
limit.

One
commenter
(
OAR­
2005­
0117­
0065)
requested
that
EPA
not
amend
the
NOx
limit
for
mass
burn
rotary
waterwall
facilities
until
methods
are
available
to
control
the
negative
impacts
of
increased
reagent
feed
rates.
The
commenter
added
that
the
increased
ammonia
or
urea
feed
rates
will
reduce
equipment
life
and
reliability
as
well
as
increases
the
risk
of
a
visible
ammonia
slip.
The
commenter
suggested
that
it
would
require
long
term
testing
to
determine
the
NOx
emission
rates
required
to
minimize
the
negative
impacts
on
equipment.

On
the
contrary,
if
EPA
proceeds
with
the
proposed
NOx
limits,
the
commenter
requested
that
EPA
not
set
a
limit
below
205
ppm,
which
currently
applies
to
mass
burn
rotary
waterwall
facilities.

One
commenter
(
OAR­
2005­
0117­
0073.1)
concluded
that
the
emission
guidelines
for
mass
burn
rotary
waterwall
combustors
(
MB/
RC)
and
RDF
units
reflect
MACT
for
NOx
and
should
remain
unchanged.
The
commenter
based
this
conclusion
on
a
revised
database
that
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
42
included
the
Savannah
plant
in
the
MB/
RC
category
and
other
data
that
were
inappropriately
removed
in
EPA's
screening
process
(
see
separate
comment
on
screening
process).
The
commenter
determined
that
an
appropriate
limit,
based
on
99.7%
UCL
would
be
207
ppmdv
for
MB/
RC
combustors
if
the
Savannah
plant
and
the
other
units
were
included
in
the
analysis,
and
that
the
99.7%
UCL
for
the
RDF
data
was
279
ppmdv,
indicating
that
the
current
limit
of
250
should
not
be
lowered.

Response:
As
a
result
of
EPA's
revised
analysis,
the
final
amendment
emission
limit
for
NOx
for
this
combustor
type
is
set
at
210
ppmdv.
This
limit
addresses
the
concerns
of
the
commenters.

Comment:
One
commenter
(
OAR­
2005­
0117­
0092)
requested
that
EPA
clarify
and
confirm
its
previous
position
that
some
low
NOx
combustors
do
not
require
SNCR
technology
to
meet
the
NOx
limits.
The
commenter
stated
that
previous
Federal
Register
notices
from
the
1991
and
1995
rulemakings
contain
this
clarifying
language.

Response:
EPA
continues
to
agree
that
SNCR
technology
is
not
required
in
order
to
meet
the
NOx
limits
in
the
final
rule.
If
a
facility
can
meet
the
NOx
limits
by
another
means,
then
the
facility
remains
in
compliance
with
the
NOx
standard.
EPA
believes
this
response
is
sufficient
to
address
the
commenter's
concern
without
including
a
statement
in
the
preamble
to
the
final
amendments.

5.14
SO2
Limits
Comment:
One
commenter
(
OAR­
2005­
0117­
0068)
contended
that
the
existing
SO2
limit
reflects
MACT
with
adequate
attention
to
emission
variability.
The
commenter
presented
two
years
of
CEM
data
from
its
Pinellas
county
Resource
Recovery
Facility
to
demonstrate
that
the
proposed
SO2
limits
would
cause
this
facility
to
significantly
increase
the
number
of
SO2
exceedances.

Two
commenters
(
OAR­
2005­
0117­
0073.1,
OAR­
2005­
0117­
0077)
proposed
a
revised
standard
for
SO2
percent
reduction
and
CEM
emission
limits
in
the
emission
guidelines
and
the
NSPS
and
disagreed
with
the
proposed
emission
standard
for
SO2
percent
reduction
limits.
One
commenter
(
OAR­
2005­
0117­
0073.1)
cited
Attachment
6
to
their
comment
to
show
that
eight
out
of
the
13
units
in
the
SD/
ESP/
ACI/
SNCR
EPA
emissions
test
database
would
not
have
met
the
proposed
80%
reduction
requirement
for
SO2
with
a
2
percent
compliance
margin.
The
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
43
commenter
added
that
when
such
a
high
percentage
of
the
units
in
the
dataset
cannot
meet
the
proposed
limits
with
an
adequate
compliance
margin,
the
proposed
emissions
standards
are
inappropriate.

Another
commenter
(
OAR­
2005­
0117­
0077)
noted
that
EPA
analyzed
and
combined
data
from
two
entirely
separate
and
independent
databases
that
represented
a
timeframe
of
1988
to
2000.
The
commenter
contends
that
the
mixing
of
data
generated
from
10­
11
years
prior
to
the
compliance
data
with
the
recent
compliance
data
will
not
correctly
predict
achievable
performance
levels.

Existing
MACT
Standard
The
two
commenters
(
OAR­
2005­
0117­
0073.1,
OAR­
2005­
0117­
0077)
based
the
proposed
changes
for
the
emission
guidelines
on
the
99.726
percentile
of
one
year
of
CEM
data
from
the
Millbury,
Massachusetts
large
mass
burn
water
wall
facility,
which
has
SDA/
ESP/
ACI/
SNCR
emission
controls
and
a
state­
of­
the
art
CEM.
The
CEM
data
analyzed
by
the
commenter
was
24­
hour
block
geometric
mean
for
inlet
and
outlet
concentration
(
ppm
@
7%

O2)
and
percent
removal
efficiencies,
and
contained
over
16,000
hourly
averages.
The
commenter
noted
that
older
data
from
the
Millbury
facility
were
used
to
provide
the
variability
factor
in
the
proposed
limits.

One
commenter
(
OAR­
2005­
0117­
0077)
referenced
their
attached
CEM
data
and
cited
an
electronic
spreadsheet
of
the
same
data
that
was
submitted
with
OAR­
2005­
0117­
0073.1
comments
and
analysis.
The
result
of
the
analysis
of
the
Millbury
facility
resulted
in
mean
performance
levels
that
are
nearly
twice
as
high
as
the
levels
identified
in
EPA's
database.

Further
the
commenter
noted
that
the
variability
factor
for
the
data
is
1.5
times
the
variability
factor
used
in
EPA's
analysis.
Both
commenters
(
OAR­
2005­
0117­
0073.1,
OAR­
2005­
0117­

0077)
proposed
to
revise
the
limit
for
SO2
on
existing
MACT
units
to
29
ppm
or
75%
removal
based
on
24
hour
geometric
means.

NSPS
Standard
The
commenters
(
OAR­
2005­
0117­
0073.1,
OAR­
2005­
0117­
0077)
based
the
proposed
revised
standard
for
the
NSPS
on
one
year
of
SO2
CEM
data
from
Westchester,
New
York
facility,
which
is
also
a
large
mass
burn
water
wall
unit.
The
facility
is
equipped
with
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
44
SD/
FF/
ACI/
SNCR,
and
it
is
considered
to
have
a
modern
SD/
FF
system
and
a
state
of
the
art
CEM.
The
CEM
data
analyzed
by
the
commenter
was
24­
hour
block
geometric
means
for
inlet
and
outlet
concentration
(
ppm
@
7%
O2)
and
percent
removal
efficiencies,
and
contained
over
23,000
hourly
averages.

One
commenter
(
OAR­
2005­
0117­
0077)
presented
the
results
of
the
analysis
on
the
Westchester
facility.
The
mean
performance
levels
were
2­
3
times
as
high
as
the
levels
identified
in
EPA's
database.
Further,
the
commenter
noted
that
the
variability
factor
for
the
data
is
15
ppm
higher
than
the
variability
factor
used
in
EPA's
analysis.
Both
commenters
(
OAR­
2005­
0117­
0073.1,
OAR­
2005­
0117­
0077)
proposed
to
revise
the
limit
for
SO2
on
new
MWC
to
26
ppm
or
80%
removal
based
on
24
hour
geometric
means.

The
commenters
(
OAR­
2005­
0117­
0073.1,
OAR­
2005­
0117­
0077)
suggested
that
the
Millbury
and
Westchester
databases
are
more
rigorous
in
terms
of
the
number
of
data
points
and
the
level
of
data
quality,
when
compared
to
the
databases
used
by
EPA
in
the
development
of
the
proposed
SO2
standards.
The
commenters
contend
that
it
is
more
appropriate
to
determine
the
SO2
emission
limits
in
the
emission
guidelines
and
NSPS
from
these
robust
datasets.

Response:
In
the
final
rule,
EPA
has
maintained
the
current
standards
in
the
1995
rule
for
SO2,
rather
than
the
levels
in
the
proposed
amendments.
EPA's
revised
statistical
analysis
does
not
support
any
changes
to
the
SO2
standard
from
the
1995
level.
Furthermore,
EPA
considered
the
commenter's
analysis
of
over
16,600
hourly
averages
of
year
2005
CEMS
data
from
the
two
units
at
Millbury
and
over
23,400
hourly
averages
of
year
2006
CEMS
data
from
the
three
units
at
the
Westchester
facilities,
which
did
not
demonstrate
that
the
proposed
SO2
levels
could
be
continuously
achieved
over
time.
The
commenter's
analysis
of
CEMS
data
from
these
two
facilities
was
consistent
with
EPA's
decision
to
retain
the
1995
emission
limits
for
SO2.
EPA
agrees
that
the
proposed
standards
were
pushing
the
practical
limits
of
SO2
control
devices
and
EPA
recognizes
the
need
to
account
for
a
compliance
margin
in
the
standard.
EPA
believes
the
final
rule
addresses
the
commenters'
concerns
on
the
proposed
amendments'
SO2
standards.
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
45
Chapter
6
Testing
and
Monitoring
6.1
Schedule
Comment:
Six
commenters
(
OAR­
2005­
0117­
0073.1,
OAR­
2005­
0117­
0077,
OAR­

2005­
0117­
0080,
OAR­
2005­
0117­
0076,
OAR­
2005­
0117­
0081,
OAR­
2005­
0117­
0068)

supported
the
flexibility
provided
for
by
EPA
on
scheduling
annual
compliance
tests
no
less
than
9
months
and
no
more
than
15
apart.
However,
the
commenter
noted
that
in
the
proposal,
the
time
interval
for
mercury
testing
remained
at
no
more
than
12
months.
They
requested
that
the
9
to
15
month
range
be
expanded
to
all
pollutants
subject
to
annual
stack
testing.
The
commenter
added
that
stack
testing
for
all
pollutants
is
typically
scheduled
concurrently,
and
in
order
to
take
advantage
of
this
flexibility,
all
pollutants
must
be
eligible
for
the
9
to
15
month
range
for
testing
frequency.

Response:
EPA
agrees
with
the
commenter,
and
it
meant
to
apply
the
flexibility
for
scheduling
compliance
tests
across
all
pollutants
when
it
proposed
the
rule.
It
was
an
oversight
in
the
proposed
rule
language
that
it
did
not
change
the
12
months
to
15
months
for
mercury.

EPA
has
corrected
the
final
rule
to
allow
for
this
range
on
scheduling
compliance
tests
for
all
pollutants.

Comment:
One
commenter
(
OAR­
2005­
0117­
0084)
disagreed
with
the
proposed
changes
to
annual
testing
frequency
and
does
not
think
this
change
will
support
the
goal
to
add
flexibility
for
weather
or
plant
outages.
Although
the
commenter
disagreed
with
the
proposed
change,
they
suggested
that
the
proposed
language
in
40
CFR
60.58b(
d)(
2)(
ix))
contains
an
inconsistency,
which
lists
that
mercury
be
tested
between
9
and
12
months
from
the
previous
test,
whereas
the
other
test
schedules
contain
the
between
9
and
15
months
language.
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
46
Response:
EPA
has
considered
the
commenter's
issues,
and
it
disagrees
with
the
commenter.
Owners
and
operators
of
MWC
facilities
support
the
additional
flexibility
and
EPA
has
retained
this
9
to
15
month
range
in
the
final
rule.
However,
to
limit
potential
abuse
of
this
flexibility,
the
final
amendments
also
require
that
5
tests
be
performed
every
5
calendar
years.

EPA
discussed
the
mercury
inconsistency
in
a
prior
response.

6.2
Reduced
Testing
Frequency
for
Exceptionally
Well­
Operated
Units
Comment:
Three
commenters
(
OAR­
2005­
0117­
0073.1,
OAR­
2005­
0117­
0077,
OAR­

2005­
0117­
0080)
disagreed
with
replacing
the
testing
waiver
for
dioxin
testing
requirements
with
a
waiver
that
would
require
a
facility
to
meet
exceptionally
low
emissions
for
both
dioxin
and
mercury
in
order
to
qualify
for
a
testing
waiver.
The
commenters
requested
that
the
current
dioxin
waiver
be
retained
and
that
EPA
develop
a
similar,
but
separate,
testing
waiver
for
facilities
with
low
mercury
emissions.
The
commenters
stated
that
mercury
and
dioxin
emissions
are
independent
of
one
another.
The
commenters
emphasized
that
mercury
emissions
are
dependent
on
the
mercury
content
in
the
solid
waste
stream,
where
dioxin
emissions
are
dependent
on
good
combustion
practices,
where
low
dioxin
emissions
can
serve
as
a
surrogate
to
represent
a
continuously
well­
operated
and
 
maintained
unit.
The
commenters
cited
that
an
incoming
solid
waste
load
with
high
mercury
content
could
create
a
spike
in
mercury
emissions,

even
if
this
facility
was
practicing
good
combustions
practices
to
control
dioxin
emissions.
As
a
result
of
the
mercury
spike,
the
proposed
revisions
would
require
that
a
facility
test
for
both
dioxin
and
mercury.
The
commenters
estimate
that
the
revised
testing
waiver
would
cost
a
facility
between
$
15,000
and
$
30,000
per
year.

One
commenter
(
OAR­
2005­
0117­
0103.1)
supported
the
proposed
requirement
for
sources
to
meet
both
mercury
and
dioxin
criteria
in
order
to
qualify
for
the
testing
waiver
as
an
"
exceptionally
well­
operated"
source.

Response:
EPA
has
reviewed
the
issues
from
these
commenters
and
it
agrees
that
mercury
emissions
are
independent
of
dioxin.
Further,
EPA
recognizes
that
mercury
can
be
impacted
by
other
factors.
As
a
result,
the
mercury
testing
component
has
been
removed
from
the
reduced
testing
waiver
in
the
final
rule
amendments.
A
State
can
still
choose
to
couple
the
reduced
testing
waiver
requirement
if
they
wish
to
do
so.
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
47
Comment:
One
commenter
(
OAR­
2005­
0117­
0068)
requested
that
EPA
revise
the
proposed
language
in
40
CFR
60.38b(
b)
to
clarify
that
the
mercury
testing
frequency
is
also
reduced
for
exceptionally
well
controlled
MWC
units.

Response:
EPA
disagrees
with
the
commenter.
EPA
has
removed
the
mercury
component
from
the
reduced
testing
waiver
and
has
retained
the
existing
reduced
testing
waiver
for
dioxin
requirements
in
the
final
amendments.
Given
that
mercury
is
no
longer
part
of
the
waiver,
EPA
sees
no
reason
to
reduce
the
testing
frequency
for
mercury.

Comment:
One
commenter
(
OAR­
2005­
0117­
0100)
requested
that
EPA
maintain
the
proposed
concurrently
low
emissions
of
mercury
and
dioxin/
furan
requirement,
with
one
exception,
in
order
to
qualify
for
the
testing
waiver.
The
commenter
suggested
that
facilities
be
subject
to
14
µ
g/
dscm
instead
of
the
proposed
requirement
of
25
µ
g/
dscm.
The
commenter
added
that
the
reduced
mercury
requirement
provides
an
incentive
to
optimize
the
carbon
injection
system
and
other
operating
parameters
on
the
MWC.
The
commenter
referenced
N.
J.
A.
C.
for
a
similar
testing
waiver
using
the
14
µ
g/
dscm.

The
commenter
also
suggested
that
the
testing
waiver
provision,
with
the
commenter's
recommended
revised
mercury
limit
of
14
µ
g/
dscm,
be
incorporated
in
Subpart
Cb.

Response:
EPA
has
decided
to
uncouple
the
mercury
requirements
from
the
reduced
dioxin
testing
requirements
that
were
listed
in
the
proposed
amendments.
Based
on
the
mercury
data
analyzed
in
the
final
rulemaking,
EPA
observed
a
mean
mercury
level
of
10
ug/
dscm
for
fabric
filter­
equipped
MWC
units,
which
is
very
close
to
the
recommended
14
µ
g/
dscm
suggested
by
the
commenter
for
the
testing
waiver.
The
presence
of
such
low
mean
performance
levels
indicates
that
there
is
already
ample
incentive
to
optimize
operating
parameters,
including
ACI
feed
rate,
at
a
MWC
facility
without
adding
a
mercury
requirement
to
the
reduced
testing
waiver.

Comment:
One
commenter
(
OAR­
2005­
0117­
0068)
requested
that
EPA
add
language
to
both
Subparts
Cb
and
Eb
to
allow
for
the
reduced
testing
frequency
for
PM,
cadmium,
and
lead
at
exceptionally
well
controlled
MWC
units.
The
commenter
noted
that
these
three
pollutants
are
not
related
to
the
ACI
system
and
would
therefore
be
unaffected
by
the
proposed
reduced
testing
waiver.

The
commenter
suggested
some
limits
to
define
exceptionally
well
controlled
units
for
each
of
these
pollutants.
The
commenter
proposed
that
emissions
of
PM
must
be
no
more
than
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
48
12
mg/
dscm,
cadmium
emissions
must
not
exceed
15
µ
g/
dscm
and
lead
emissions
must
not
exceed
125
µ
g/
dscm,
all
of
these
proposed
limits
are
corrected
to
7%
O2
in
order
to
qualify
as
an
exceptionally
well
controlled
unit.

The
commenter
also
suggested
emission
limits
for
the
three
pollutants
that
could
provide
allowance
for
a
MWC
owner
or
operator
to
test
one
unit
at
a
facility
as
opposed
to
each
and
every
unit.
These
emission
suggested
by
the
commenter
are
less
than
or
equal
to
12
mg/
dscm
for
PM,
less
than
or
equal
to
1.75
µ
g/
dscm
for
cadmium
and
less
than
or
equal
to
42
µ
g/
dscm
for
lead.
The
commenter
added
that
all
performance
tests
over
a
2­
year
period
must
indicate
these
levels
in
order
to
qualify
for
the
single
unit
testing.

Response:
As
discussed
in
earlier
responses,
EPA
is
not
allowing
for
a
reduced
testing
frequency
for
any
pollutants
other
than
dioxin
in
the
final
amendments.

6.3
Alternative
CEM
Option
Comment:
Two
commenters
(
OAR­
2005­
0117­
0084,
OAR­
2005­
0117­
0088)
agreed
with
incorporating
alternative
CEM
monitoring
requirements
for
PM,
HCl,
Hg,
and
multi­
metals
as
an
alternative
to
annual
stack
testing.
One
of
these
commenters
(
OAR­
2005­
0117­
0084)
also
urged
EPA
to
update
the
corresponding
Performance
Specifications
for
these
technologies.
The
other
commenter
(
OAR­
2005­
0117­
0088)
added
that
it
is
important
to
provide
the
facilities
with
their
own
discretion
to
adopt
this
technology
given
the
limited
track
records
of
these
monitoring
devices
at
the
time
these
amendments
are
proposed.

One
commenter
(
OAR­
2005­
0117­
0080)
suggested
that
changing
the
method
of
compliance
from
stack
testing
to
CEMS
may
change
the
stringency
of
the
proposed
limit.
The
commenter
added
that
CEM
data
detects
many
spikes
in
the
data
that
are
not
picked
up
during
a
stack
test.
The
commenter
recognized
that
EPA
tried
to
account
for
these
differences
using
separate
analysis
techniques
for
stack
test
data
and
CEM
data,
however
the
commenter
noted
that
no
similar
separate
analysis
was
done
when
changing
to
the
alternate
method
of
compliance.

The
same
commenter
also
cited
two
court
cases
to
support
the
need
to
consider
the
impact
of
offering
an
alternative
monitoring
method.
Portland
Cement
Assn.
v.
Ruckelshaus
noted
that
when
there
is
a
significant
difference
in
techniques
used
to
determine
compliance
with
the
standard,
serious
questions
are
raised
on
the
validity
of
the
standard.
Additionally,
in
National
Lime
Ass'n
v.
EPA
it
was
determined
that
EPA
must
provide
ample
consideration
to
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
49
the
possible
impact
on
emissions
from
variations
in
operations.
The
commenter
requested
that
EPA
analyze
if
the
proposed
standards
are
achievable
when
measured
with
a
CEM.

Seven
commenters
(
OAR­
2005­
0117­
0073.1,
OAR­
2005­
0117­
0102,
OAR­
2005­
0117­

0083,
OAR­
2005­
0117­
0085,
OAR­
2005­
0117­
0077,
OAR­
2005­
0117­
0080,
OAR­
2005­
0117­

0092)
disagreed
with
the
inclusion
of
alternative
CEM
options
for
particulate,
multi­
metals,
HCl,

mercury
and
dioxins
in
the
proposed
amendments
until
each
of
these
monitoring
technologies
have
been
validated
on
MWCs.
One
of
these
commenters
(
OAR­
2005­
0117­
0073.1)
proposed
that,
if
a
facility
wants
to
use
a
CEMS,
it
must
wait
until
the
desired
monitoring
technology
has
been
validated.
The
facility
can
then
submit
a
request
for
an
alternative
monitoring
plan
allowed
for
in
the
general
provisions
at
40
CFR
60.13(
i).

One
commenter
(
OAR­
2005­
0117­
0107)
discussed
the
increased
presence
of,
as
well
as
the
benefits
of,
mercury
CEMS.
The
commenter
suggested
that
mercury
CEMS
are
expected
to
see
large
growth
as
a
result
of
emissions
measurement
requirements
under
the
Clean
Air
Mercury
Rule.
The
commenter
also
cited
an
article
to
demonstrate
the
recent
increase
in
contracts
for
mercury
controls,
of
which
the
commenter
suggested
will
be
accompanied
by
mercury
CEMS.
Additionally,
the
commenter
stated
that
mercury
CEMS
provides
for
real­
time
information
that
can
serve
as
a
valuable
process
control
tool.
This
commenter
also
urged
EPA
to
adopt
PM
CEMS
as
a
result
of
an
increased
level
of
comfort
with
the
performance
of
this
monitoring
instrument.

One
commenter
(
OAR­
2005­
0117­
0079.1)
supported
the
optional
provision
for
PM
CEMS
in
place
of
EPA
Method
5
testing.
The
commenter
provided
detail
on
several
installations
and
certifications
of
PM
CEMS
in
the
United
States
on
several
sources
including:

electric
utilities,
recovery
boilers,
and
hazardous
waste
incinerators.

Several
of
the
commenters
(
OAR­
2005­
0117­
0073.1,
OAR­
2005­
0117­
0102,
OAR­

2005­
0117­
0085,
OAR­
2005­
0117­
0080)
were
concerned
that,
if
this
equipment
is
incorporated
as
an
alternative
in
the
rule,
then
regulators
will
call
for
these
CEMS
to
become
mandatory
prior
to
the
equipment
getting
validated.
The
commenters
emphasized
that
EPA
should
not
imply
their
approval
to
the
States
interpreting
this
rule
without
first
validating
these
CEM
monitoring
technologies.
One
commenter
(
OAR­
2005­
0117­
0083)
requested
that
EPA
provide
more
detail
on
certification
and
maintenance
requirements
for
new
monitoring
instruments.
This
commenter
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
50
also
requested
that
EPA
detail
how
these
alternative
CEM
requirements
relate
to
existing
and
future
permit
and
reporting
requirements.

One
commenter
(
OAR­
2005­
0117­
0073.1)
suggested
two
additional
reasons
that
alternative
CEM
requirements
should
not
be
included
in
the
final
amendments.
First,
the
commenter
stated
that
section
129(
c)
does
not
allow
for
facilities
to
use
a
monitoring
method
that
has
not
been
validated
on
MWCs.
Second,
the
commenter
contended
that
EPA
did
not
provide
data
or
sufficient
rationale
for
deciding
to
include
an
alternative
CEM
requirement.
The
commenter
added
that
a
reference
to
the
use
of
CEM
equipment
in
Europe
does
not
address
the
implications
of
violating
the
Clean
Air
Act
if
a
CEM
fails
to
meet
availability
requirements.

Three
commenters
(
OAR­
2005­
0117­
0094,
OAR­
2005­
0117­
0087,
OAR­
2005­
0117­

0064)
disagreed
with
allowing
an
alternative
PM
CEM
requirement
to
be
part
of
the
rule
until
a
PM
CEM
technology
has
been
verified.
The
commenter
added
that,
by
incorporating
this
alternative
provision,
State
agencies
may
assume
that
the
technology
has
met
the
verification
requirements
of
CAA
section
129(
c).

One
commenter
(
OAR­
2005­
0117­
0092)
requested
that
EPA
clearly
state
in
the
final
regulations
that
none
of
the
CEMS
have
been
validated
for
use
at
MWC
facilities
(
with
the
possible
exception
of
HCl
CEM).
The
commenter
made
this
request
with
the
concern
that
State
regulators
may
otherwise
assume
these
technologies
have
been
validated
and
as
a
result,
require
monitoring
to
be
done
with
a
CEM.

One
commenter
(
OAR­
2005­
0117­
0103.1)
requested
that
alternate
CEM
provisions
be
replaced
with
mandatory
requirements
to
install
CEM.
The
commenter
added
that
EPA
must
determine
what
types
of
monitoring
requirements
are
necessary
to
protect
the
public
health
and
the
environment.
The
commenter
contended
that
the
standards
developed
as
part
of
the
1995
rulemaking
process
did
not
establish
any
emissions
monitoring
requirements
for
a
majority
of
the
performance
standards
and
only
included
parametric
monitoring
requirements.
The
commenter
added
that
the
proposed
standards
need
to
include
mandatory
monitoring
requirements
in
order
to
bring
the
MWC
standards
into
compliance
with
CAA
129(
c)
and
further
concluded
that
CEM
systems
are
the
only
type
of
monitoring
requirements
that
can
provide
adequate
data
to
ensure
compliance
with
emission
standards
protective
of
public
health
and
the
environment
They
argued
that
the
language
of
CAA
section
129(
c)
does
not
give
EPA
the
authority
to
establish
only
parameter
monitoring
requirements
for
some
pollutants.
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
51
Response:
EPA
has
considered
the
commenters'
issues
and
supports
the
inclusion
of
CEM
as
an
alternative
to
stack
testing
for
PM
and
Hg.
EPA
emphasizes
that
CEM
equipment
for
these
pollutants
is
not
mandatory.
Please
see
the
preamble
for
further
discussion
on
incorporating
the
option
of
CEM
for
PM
and
Hg
in
the
final
amendments.

Comment:
One
commenter
(
OAR­
2005­
0117­
0107)
recommended
that
EPA
require
CEMs
for
monitoring
HCl.
The
commenter
contended
that
there
are
15
MWC
operating
in
the
U.
S.
with
continuous
HCl
monitors.
Further,
the
commenter
added
that
three
states,

Pennsylvania,
Maryland,
and
New
Jersey,
all
require
HCl
CEMS
on
new
units.
Additionally,

Pennsylvania
requires
HCl
CEMS
on
existing
units.
The
commenter
emphasized
that
the
costs
for
installing
a
HCl
CEMs
would
be
limited
to
the
cost
of
purchasing
and
maintaining
a
detector,

since
CEMS
monitoring
and
reporting
are
already
required
for
NOx,
SO2,
and
CO.

Response:
EPA
has
considered
the
issues
raised
by
the
commenter
and
disagrees
that
HCl
CEMS
should
be
included
in
the
final
amendments.
Without
EPA's
publication
of
performance
specifications
it
is
premature.
As
the
commenter
points
out,
several
States
are
currently
requiring
HCl
CEMs
on
existing
units.
EPA
acknowledges
that
any
State,
region,
or
individual
facility
retains
the
right
to
elect
or
require
the
use
of
HCl
CEMS.

Comment:
The
commenter
(
OAR­
2005­
0117­
0092)
stated
that
EPA
provided
no
information
on
the
use
of
PM
CEM
on
German
MWCs
in
the
docket.
The
commenter
also
added
that
the
widespread
use
of
CEMs
on
fossil­
fueled
units
is
not
relevant
to
MWC
which
fire
a
fuel
with
comparatively
higher
composition
variability.

Response:
EPA
has
considered
the
commenter's
issues
and
believes
that
the
use
of
CEMS
on
fossil
fuel
units
is
technically
relevant
to
their
performance
on
MWC
units.

Furthermore,
as
noted
in
the
preamble,
CEMS
data
for
a
variety
of
pollutants,
including
PM,

from
a
German
MWC
are
available
via
the
internet.
Finally,
these
CEMS
are
being
included
in
the
rule
as
an
alternative,
rather
than
a
requirement.

Comment:
Two
commenters
(
OAR­
2005­
0117­
0079.1,
OAR­
2005­
0117­
0084,)

suggested
that
EPA
drop
the
opacity
monitoring
requirement
for
MWC
units
that
use
a
PM
CEMS.
One
commenter
(
OAR­
2005­
0117­
0079.1)
suggested
several
shortcomings
of
opacity
monitoring
and
noted
that
the
variations
in
waste
stream
make­
up
create
particulate
emissions
where
the
size,
shape,
and
density
of
particles
change
over
time.
These
variations
create
opacity
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
52
readings
that
do
not
correlate
with
particulate
emissions.
The
same
commenter
also
indicated
that,
given
the
recent
installations
of
PM
CEMS
as
well
as
the
promulgated
performance
specifications
and
quality
assurance
and
control
procedures,
PM
CEMS
should
replace
the
continuous
opacity
monitoring
requirement.

One
commenter
(
OAR­
2005­
0117­
0107)
suggested
that
EPA
drop
the
opacity
requirement
when
a
PM
CEM
is
installed
and
opacity
is
typically
less
than
5
percent.
The
commenter
added
that
ASTM
D6216­
03
states
that
the
design
specifications
are
not
applicable
for
demonstrating
compliance
with
opacity
limits
of
less
than
10%.
The
commenter
encouraged
the
continued
use
of
any
existing
opacity
monitor
to
establish
a
correlation
between
opacity
and
PM
emissions
such
that
the
PM
data
could
demonstrate
that
the
opacity
limit
is
not
being
exceeded.

One
commenter
(
OAR­
2005­
0117­
0084)
also
suggested
that
bag
leak
detectors
(
triboelectric
monitors)
could
also
serve
in
lieu
of
opacity
monitors
if
EPA
developed
corresponding
performance
specifications.

Response:
EPA
has
considered
the
commenters'
noted
shortcomings
of
opacity
monitors
and
EPA
has
changed
the
final
amendments
to
provide
the
option
to
use
a
PM
CEMS
for
continual
compliance
with
the
PM
limit
in
lieu
of
monitoring
opacity
with
an
opacity
monitor
or
doing
manual
stack
tests.
Furthermore,
EPA
does
not
believe
it
is
necessary
to
limit
this
optional
use
of
a
PM
CEMS
to
situations
where
an
opacity
threshold
of
5%
is
met.
EPA
believes
the
requirement
for
an
opacity
monitor
and
manual
stack
tests
should
be
dropped
whenever
a
unit
chooses
to
install
a
PM
CEMS.
As
a
PM
CEMS
provides
direct
measurement
of
emissions
rather
than
a
parametric
value,
bag
leak
detectors
are
unnecessary.

6.4
CEM
Validation
Requirements
Comment:
Two
commenters
(
OAR­
2005­
0117­
0073.1,
OAR­
2005­
0117­
0077)

requested
that
EPA
clarify
the
procedures
to
address
partial
compliance
averaging
periods
particularly
for
periods
of
start­
up,
shutdown,
and
control
device
malfunction.
The
commenters
requested
that
the
final
amendments
include
a
revised
monitoring
requirement
stating
that
hourly
averages
are
not
to
be
calculated
for
hours
with
less
than
30
minutes
of
valid
data.

The
commenters
also
requested
that
EPA
define
data
validation
criteria
for
compliance
averages
to
be
a
75
percent
validation
requirement
for
the
CO,
NOx,
and
SO2
emissions
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
53
standards.
For
example,
three
hours
of
operating
time
to
validate
the
four
hour
averaging
period
for
CO
and
18
hours
of
operating
time
to
validate
the
24
hours
SO2
and
NOx
standards.

Response:
Commenters
raised
questions
on
the
proper
use
of
CEMS
data.
One
must
always
remember
that
under
the
regulations
(
1)
determining
CEMS
data
availability
to
judge
CEMS
performance
and
(
2)
the
use
of
CEMS
data
to
calculate
emissions
to
determine
compliance,
are
two
different
things.
For
CEMS
data
availability
determination,
each
hour
of
data
must
have
at
least
two
data
points
and
this
must
be
done
for
at
least
90%
of
operating
hours
per
calendar
quarter
and
at
least
95
%
of
the
operating
hours
per
calendar
year.
The
previous
requirement
in
the
1995
rule
to
obtain
data
for
75%
of
the
operating
hours
per
day
has
been
dropped.
The
CEMS
data
requirements
apply
for
all
pollutants
monitored
using
CEMS.
There
is
no
change
in
use
of
CEMS
data
to
calculate
emissions.
All
valid
continuous
emissions
monitoring
data
must
be
used
for
calculating
the
24­
hour
average
emission
levels
required
by
the
rule,
even
if
the
minimum
data
availability
requirements
are
not
met.

6.5
CEM
Data
Availability
Comment:
One
commenter
(
OAR­
2005­
0117­
0084)
supported
the
proposed
increase
in
CEM
availability.
The
commenter
cited
their
own
study,
which
analyzed
3
years
of
data
for
all
sources
required
to
operate
CEMS
in
their
State,
and
found
that
the
CEM
95%
availability
requirement
could
be
met
97%
of
the
time
for
all
sources.
Further,
they
added
that
the
availability
rate
was
higher
for
MWC
sources.
The
commenter
provided
a
caveat
for
this
study
with
a
note
that
the
CEM
did
not
include
monitoring
data
for
PM,
Hg,
or
HCl
CEMS.

One
commenter
(
OAR­
2005­
0117­
0103.1)
supported
the
increased
CEM
availability
requirements.
The
commenter
added
that
increased
availability
will
provide
the
public
and
permitting
authorities
more
adequate
information
with
which
to
evaluate
emissions
and
compliance
from
MWCs.

Nine
commenters
(
OAR­
2005­
0117­
0073.1,
OAR­
2005­
0117­
0088,
OAR­
2005­
0117­

0096,
OAR­
2005­
0117­
0089,
OAR­
2005­
0117­
0081,
OAR­
2005­
0117­
0077,
OAR­
2005­
0117­

0080,
OAR­
2005­
0117­
0095,
OAR­
2005­
0117­
0092)
disagreed
with
the
proposed
95%

requirement
for
CEM
availability.
The
commenters
contended
that
raising
the
availability
from
90
to
95%
will
require
facilities
to
install
redundant
monitoring
equipment
in
order
to
avoid
violating
CEM
availability.
The
commenters
added
that
the
existing
availability
requirements
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
54
already
provide
for
incentive
to
operate
and
maintain
a
high
level
of
CEM
availability,
and
that
further
increases
in
the
availability
will
result
in
increased
operating
costs
without
any
environmental
benefit.

One
commenter
(
OAR­
2005­
0117­
0089)
suggested
that
the
basis
for
proposing
a
95%

availability
requirement
relied
on
an
inappropriately
low
amount
of
calendar
quarterly
availability
information
(
2
percent
or
72
out
of
a
possible
4008
calendar
quarters).
Another
commenter
(
OAR­
2005­
0117­
0081)
suggested
that
EPA
should
have
based
the
availability
on
several
or
all
of
the
large
MWC
in
order
to
determine
the
proposed
availability
requirement.

One
commenter
(
OAR­
2005­
0117­
0096)
requested
that
EPA
provide
a
more
detailed
discussion
and
presentation
of
the
basis
for
the
proposed
increase
in
data
availability
requirements.

One
commenter
(
OAR­
2005­
0117­
0092)
noted
that
EPA
does
not
suggest
in
its
proposal
that
the
current
90%
availability
requirement
is
yielding
insufficient
data
on
MWCs.
Moreover,

the
commenter
stated
that
EPA's
basis
for
increasing
the
availability
relies
on
one
year
of
data
from
a
single
plant,
which
does
not
meet
the
"
achieved
in
practice"
standard
contained
in
CAA
129(
a)(
2).
The
commenter
also
faulted
the
analysis
contained
in
Docket
Item
OAR­
2005­
0117­

0013,
since
this
analysis
does
not
consider
days
when
a
monitor
fails
to
meet
the
75%
daily
requirement.

Two
commenters
(
OAR­
2005­
0117­
0073.1,
OAR­
2005­
0117­
0096)
suggested
that
EPA
analyze
the
costs
of
requiring
a
facility
to
install
back­
up
CEM
equipment
to
meet
the
proposed
availability.
One
commenter
(
OAR­
2005­
0117­
0073.1)
added
that,
if
EPA
still
sees
fit
to
increase
the
availability
after
these
costs
are
considered,
EPA
should
maintain
the
quarterly
requirements
for
hours
per
day
and
hours
per
quarter
and
only
increase
the
availability
to
95%
of
the
number
of
operating
days
per
year.
The
commenter
suggested
that
these
revised
availability
requirements
allow
for
a
facility
to
remain
in
compliance
in
the
event
of
CEM
equipment
failure.

One
commenter
(
OAR­
2005­
0117­
0083)
noted
that
the
expense
of
CEMS
equipment
to
support
the
higher
availability
rate
is
inappropriate,
given
that
these
monitoring
increases
to
not
demonstrate
a
benefit
to
public
health
or
the
environment.

Response:
EPA
has
considered
the
commenters'
concerns
and
has
adopted
a
90
percent
CEMS
availability
requirement
per
calendar
quarter
and
a
95
percent
availability
requirement
per
calendar
year
as
suggested
by
some
commenters.
EPA
has
eliminated
the
75
percent
daily
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
55
requirement
because
the
90
and
95
percent
requirements
are
sufficient
to
ensure
good
data
availability.
It
was
not
EPA
intent
to
require
sources
to
install
back­
up
monitoring
equipment
in
order
to
meet
the
proposed
95
percent
availability
requirement.
However,
EPA
sees
no
reason
why
CEMS
should
not
be
able
to
meet
the
data
availability
of
the
final
amendments,
and
the
95
percent
annual
availability
requirements
assures
that
demonstrated
performance
of
CEMS
is
reflected
in
the
regulation.
EPA
has
included
a
more
complete
discussion
of
CEMS
availability
in
the
preamble.

6.6
Harmonization
of
Monitoring
Requirements
between
40
CFR
60
and
75
Comment:
Two
commenters
(
OAR­
2005­
0117­
0073.1,
OAR­
2005­
0117­
0077)

requested
that
EPA
harmonize
the
monitoring
requirements
for
40
CFR
part
60
and
40
CFR
part
75.
Specifically,
the
commenters
requested
that
the
definitions
of
operating
hours
and
CEMS
data
as
proposed
in
the
Electric
Utility
source
category
be
included
in
the
final
amendments.

The
commenters
also
cited
the
harmonization
of
the
monitoring
requirements
in
the
recently
proposed
Industrial­
Commercial­
Institutional
Electric
Utility
Steam
Generating
units
rulemaking.

Response:
EPA
agrees
with
the
commenters
about
the
advantages
of
harmonizing
these
two
separate
and
sometimes
overlapping
monitoring
requirements
for
certain
applications.

Following
proposal
of
the
two
utility
rules
mentioned
in
the
comment,
EPA
determined
that
the
language
proposed
to
resolve
these
issues
was
overly
broad
and
would
lead
to
unintended
consequences
when
applied
to
situations
other
than
the
utility
or
MWC
applications
or
to
low
emissions
rate
situations.
EPA
is
reviewing
alternative
language
changes,
and
EPA
expects
to
insert
language
to
achieve
harmonization
at
a
later
time.

6.7
Valid
Data
for
Compliance
Calculations
Comment:
Two
commenters
(
OAR­
2005­
0117­
0073.1,
OAR­
2005­
0117­
0077)
noted
conflicted
language
between
the
general
provisions
and
60.58b(
i)
on
the
validation
procedures
for
MWC
operating
data
such
as
carbon
federate,
steam
flow,
and
baghouse
inlet
temperature.

The
commenters
added
that
because
of
this
conflicting
language,
no
validation
procedures
are
defined
for
these
operating
parameters.
The
commenters
suggested
that
EPA
define
how
much
of
data
is
necessary
for
these
parameters
to
make
up
a
valid
one­
hour
average
and
recommended
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
56
that
EPA
require
operating
data
to
be
collected
at
a
minimum
of
75%
of
the
hour
in
order
to
validate
these
process
parameters.

Response:
The
commenter
did
not
submit
any
data
to
address
the
issue.
EPA
will
investigate
the
issue
and
will
address
this
concern
in
a
separate
notice.

6.8
ACI
Pneumatic
Pressure
Comment:
One
commenter
(
OAR­
2005­
0117­
0103.1)
supported
the
proposed
requirement
to
monitor
pneumatic
injection
pressure
at
the
location
where
the
activated
carbon
is
injected
into
the
flue
gas.
The
commenter
added
that
this
measure
would
provide
more
certainty
that
the
pollution
control
devices
are
working
correctly,
and
that
MWC
are
meeting
the
appropriate
emission
limits.
The
commenter
also
supported
an
8­
hour
block
average
for
measuring
the
activated
carbon
injection
rate.

A
second
commenter
(
OAR­
2005­
0117­
0100)
also
supported
the
proposed
amendment
to
measure
the
pneumatic
injection
pressure
at
the
location
where
the
activated
carbon
is
injected
into
the
flue
gas.
The
commenter
added
that
injector
nozzle
clogging
will
reduce
the
performance
of
emission
control,
and
that
the
proposed
metric
will
help
improve
the
annual
control
efficiency.

Three
commenters
(
OAR­
2005­
0117­
0073.1,
OAR­
2005­
0117­
0077,
OAR­
2005­
0117­

0092)
disagreed
with
the
proposed
amendment
to
require
monitoring
of
the
pneumatic
pressure
on
activated
carbon
injection
(
ACI)
systems
and
stated
this
type
of
monitoring
is
unnecessary.

The
commenters
contended
that
the
current
methods
for
monitoring
carbon
feed
rates
with
a
carbon
feeder
speed
or
direct
weight
loss
feeders
provide
for
adequate
monitoring
method.
The
commenters
added
that
carbon
injection
systems
already
have
built­
in
diagnostics
to
notify
facility
operators
if
there
is
a
problem
with
the
carbon
feed.
One
commenter
(
OAR­
2005­
0117­

0092)
added
that
its
facility
monitors
the
operation
of
the
carbon
feed
screw
in
8­
hour
block
averages.

Response:
As
a
result
from
the
response
from
commenters,
EPA
recognizes
that
there
are
multiple
methods
to
monitor
and
verify
carbon
feed
rates,
one
of
which
is
the
method
it
proposed
to
monitor
­
ACI
pneumatic
pressure.
However,
several
existing
units
have
other
types
of
monitoring
procedures
in
place
for
monitoring
carbon
feed,
which
EPA
believes
are
also
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
57
acceptable.
EPA
has
amended
the
final
rule
to
require
monitoring
of
ACI
pneumatic
pressure
or
other
carbon
injection
system
operational
indicator.

Comment:
One
commenter
(
OAR­
2005­
0117­
0096)
requested
that
EPA
clarify
the
pneumatic
injection
reporting
and
recordkeeping
requirements.
The
commenter
added
that
the
proposed
rule
language
requires
monitoring
at
"
the
location
where
the
activated
carbon
is
injected
into
the
flue
gas"
which
seems
to
define
a
situation
where
a
single
orifice
pressure
is
monitored
at
the
point
of
injection.
The
commenter
contended
that
the
injection
system
typically
has
a
monitor
that
is
set
to
alarm
an
operator
of
either
a
low
eductor
suction
pressure
or
a
high
blower
discharge
pressure.
The
commenter
also
suggested
that
on­
site
recordkeeping
of
alarms
without
routine
reporting
are
adequate
to
ensure
that
facilities
maintain
their
ACI
system.

Response:
As
noted
in
a
previous
comment,
EPA
has
provided
the
allowance
of
ACI
pneumatic
pressure
and
updated
the
corresponding
reporting
and
recordkeeping
requirements
in
the
rule.

6.9
Start­
up,
Shutdown
and
Malfunction
(
SSM)
Provisions
Comment:
Five
commenters
(
OAR­
2005­
0117­
0073.1,
OAR­
2005­
0117­
0088,
OAR­

2005­
0117­
0076,
OAR­
2005­
0117­
0085,
OAR­
2005­
0117­
0077)
requested
that
EPA
make
the
provisions
governing
CEM
data
during
periods
of
start­
up,
shutdown,
or
malfunction
consistent
with
the
language
in
the
small
MWC
rule
such
that
SSM
data
is
not
used
in
the
calculation
of
multi­
hour
averages
during
these
time
periods,
as
long
as
the
time
period
does
not
exceed
3
hours.
The
commenter
added
that
the
SSM
data
will
be
reported
as
"
excluded"
data
in
the
semiannual
reports
submitted
in
accordance
with
40
CFR
60.59b(
g)(
1)(
v).

One
commenter
(
OAR­
2005­
0117­
0073.1)
requested
that
EPA
extend
the
data
exclusion
period
from
3
to
15
hours
during
certain
periods
of
startup,
shutdown,
or
malfunction
for
CO
data,
similar
to
the
15­
hour
extension
provided
for
NOx
and
SO2
emissions.
The
commenter
noted
that
the
two
major
causes
of
malfunctions
are
loss
of
water
seal
or
combustion
air
control.

During
these
periods
there
are
high
levels
of
excess
O2,
and
the
NOx
and
SO2
data
corrected
to
7%
O2
will
grossly
exceed
the
limits
for
these
two
pollutants,
and
the
commenter
suggested
that
a
15­
hour
period
is
a
more
appropriate
timeframe
for
emissions
to
return
to
normal
levels.

The
commenter
also
requested
that
the
final
rule
amendments
include
a
diluent
cap
for
O2
and
CO2
of
14%
and
5%,
respectively,
similar
to
the
diluent
cap
granted
to
40
CFR
Part
60
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
58
Subpart
GG
sources.
The
commenter
emphasized
that
the
diluent
cap
would
prevent
the
overstatement
of
emissions
during
periods
of
start­
up,
shutdown,
and
malfunction
when
O2
or
CO2
exceeds
the
14%
and
5%
cap
values.

Response:
EPA
agrees
with
the
commenters'
suggestion
to
make
provisions
for
CEM
data
handling
during
periods
of
start­
up,
shutdown,
or
malfunction,
and
the
rule
language
in
the
final
amendments
has
been
updated.

Additionally,
the
final
amendments
include
an
extension
from
3
to
15
hours
for
data
exclusion
period
for
CO
emissions.
EPA
recognizes
that
the
15­
hour
time
frame
is
more
appropriate
for
readjusting
normal
facility
operations
and
CO
emissions.
This
time
frame
is
the
same
as
the
time
frame
included
for
NOx
and
SO2
emissions.

Finally,
EPA
agrees
with
using
a
diluent
cap
for
O2
and
CO2
during
periods
of
a
CO
malfunction
when
the
O2
level
exceeds
14
percent
or
the
CO2
level
exceeds
5
percent.
EPA
has
added
language
to
support
the
use
of
a
diluent
cap
during
CO
malfunctions
in
the
final
amendments.
However,
EPA
disagrees
with
the
use
of
a
diluent
cap
during
periods
of
start­
up
and
shutdown
not
related
to
CO
malfunction.

6.10
Relative
Accuracy
Test
Audits
(
RATA)

Comment:
Two
commenters
(
OAR­
2005­
0117­
0073.1,
OAR­
2005­
0117­
0077)
proposed
additional
language
to
clarify
that
the
five
ppmdv
RA
criteria
for
CO
includes
the
mean
difference
between
the
RM
and
the
CEMS
without
the
2.5%
confidence
limit.
The
commenters
suggested
that
the
following
language
be
included
in
60.85(
b)(
i)(
3)(
ii):
"
For
affected
facilities
subject
to
the
100
ppmdv
CO
standard,
the
alternative
relative
accuracy
criterion
of
5
ppmdv
is
calculated
as
the
absolute
value
of
the
average
difference
between
the
RM
and
CEMS."

Response:
EPA
agrees
with
the
commenters.
The
Relative
Accuracy
criterion
for
CO
CEMS
should
be
changed
as
indicated.
In
the
final
amendments
60.58b
(
i)
(
3)
(
ii)
is
being
changed
to
add
the
following
sentence:

For
affected
facilities
subject
to
the
100
ppmdv
CO
standard,
the
relative
accuracy
criterion
of
5
ppmdv
is
calculated
as
the
absolute
value
of
the
mean
difference
between
the
reference
method
and
continuous
emissions
monitoring
system.

Comment:
Three
commenters
(
OAR­
2005­
0117­
0073.1,
OAR­
2005­
0117­
0077,
OAR­

2005­
0117­
0080)
proposed
additional
language
to
the
final
rule
amendments
to
allow
for
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
59
situations
where
low
inlet
SO2
concentrations
fail
to
meet
the
20%
relative
accuracy
standard.

The
commenters
suggested
that
the
following
language
be
included:
"
The
RA
of
the
CEMS
must
be
no
greater
than
20
percent
of
the
mean
value
of
the
RM
test
data
in
terms
of
the
units
of
the
emission
standard,
or
5
ppmdv
mean
difference,
whichever
is
greater."

Response:
EPA
agrees
with
the
commenter.
The
Relative
Accuracy
criterion
for
inlet
SO2
CEMS
should
be
changed
as
indicated.
In
the
final
amendments,
60.58b
(
e)
(
12)
is
being
changed
to
read
as
follows:

The
continuous
emission
monitoring
system
shall
be
operated
according
to
Performance
Specification
2
in
appendix
B
of
this
part.
For
sources
that
have
actual
inlet
emissions
less
than
100
parts
per
million
dry
volume,
the
relative
accuracy
criterion
for
inlet
sulfur
dioxide
continuous
emission
monitoring
systems
should
be
no
greater
than
20
percent
of
the
mean
value
of
the
reference
method
test
data
in
terms
of
the
units
of
the
emission
standard,
or
5
parts
per
million
dry
volume
absolute
value
of
the
mean
difference
between
the
reference
method
and
the
continuous
emission
monitoring
systems,
whichever
is
greater.

6.11
Other
Comment:
One
commenter
(
OAR­
2005­
0117­
0084)
requested
that
EPA
add
ammonia
to
the
list
of
pollutants
that
MWCs
must
monitor.
The
commenter
recommended
that
CEM
be
used
to
monitor
ammonia
emissions.
In
light
of
a
CEM
monitor,
the
commenter
requested
a
minimum
of
annual
stack
testing
concurrent
with
parametric
monitoring
to
ensure
optimal
use
of
the
reagent.
The
commenter
is
concerned
with
ammonia
emissions
because
SNCR
was
identified
as
the
MACT
for
NOx
and
ammonia
is
a
significant
by­
product
of
the
SNCR
process.

The
commenter
added
that
ammonia
is
a
pre­
cursor
to
PM
and
noted
that
it
is
important
to
reduce
ammonia
emissions
in
order
to
meet
federal
reductions
for
PM10
and
PM
2.5.

Response:
The
NOx
performance
levels
specified
by
the
MWCs
rules
were
established
at
an
SNCR
performance
level
that
assures
ammonia
emissions
(
ammonia
slip)
would
be
at
a
minimum.
At
such
levels
ammonia
testing
was
judged
unnecessary.
At
higher
NOx
performance
levels
ammonia
testing
would
become
more
meaningful.
If
a
source
or
a
State
agency
wished
to
require
ammonia
testing,
they
can
do
so.
EPA
has
two
methods
for
ammonia.
One
is
EPA
Method
320,
using
FTIR
as
the
analytical
instrument.
To
date
this
has
not
been
used
for
MWCs.

The
other
method
is
conditional
method
CTM­
027.
CTM­
027
has
been
used
for
ammonia
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
60
testing
at
an
electric
utility
plant
but
this
method
has
not
been
proposed.
Although
EPA
PS­
15
is
the
performance
specification
using
the
FTIR
technology
and
has
the
potential
basis
for
a
performance
specification
for
ammonia
CEMS,
it
has
not
been
used
on
MWCs.

Comment:
Three
commenters
(
OAR­
2005­
0117­
0073.1,
OAR­
2005­
0117­
0077,
OAR­

2005­
0117­
0080)
supported
the
revised
Hg
testing
requirements
to
allow
for
optimizations
of
mercury
control
operating
parameters
by
waiving
operating
parameter
limits
during
the
mercury
performance
test
and
during
the
2
weeks
preceding
the
mercury
performance
test.

One
commenter
(
OAR­
2005­
0117­
0068)
requested
that
EPA
revise
the
proposed
language
in
40
CFR
60.58b(
m)(
2)(
i)
to
clarify
that
the
control
operating
parameter
limits
are
waived
for
two
(
2)
weeks
prior
to
the
annual
performance
test
for
both
dioxin/
furan
and
mercury.
The
commenter
was
concerned
that
they
could
not
find
any
regulatory
language
to
support
the
statement
in
the
proposal
that
recognized
that
the
waiver
of
control
operating
parameter
limits
was
already
done
for
dioxin.

Response:
EPA
has
inspected
the
rule
language
mentioned
by
the
commenter
and
agrees
that
the
language
must
be
adjusted
to
waive
the
control
operating
parameter
limits
in
order
to
be
consistent
with
the
statement
in
the
preamble
and
with
EPA
intent.
This
correction
has
been
made
in
the
final
amendments.

Comment:
One
commenter
(
OAR­
2005­
0117­
0092)
requested
that
EPA
modify
the
language
in
60.53b(
b)(
1)
to
conform
to
section
60.53b(
c)(
1).
The
commenter
added
that
currently
60.53b(
b)(
1)
requires
prior
approval
from
EPA
or
a
State
Agency
to
get
a
waiver
for
load
limits
during
annual
dioxin/
furan
and
mercury
testing,
whereas,
60.53b(
c)(
1)
provides
for
an
automatic
waiver
of
PM
control
device
temperature
limits
during
annual
dioxin/
furan
and
mercury
performance
testing.
The
commenter
suggested
this
was
an
oversight
of
EPA.

Response:
EPA
has
considered
this
issue
and
it
agrees
with
the
commenter.
In
the
final
amendments,
EPA
has
required
that
the
facility
notify
the
Administrator
two
weeks
before
the
optimization
tests
will
be
run
for
dioxin
and
mercury.
However,
the
Administrator
is
not
required
to
approve
a
waiver
of
load
limits
during
these
periods
of
dioxin/
furan
and
mercury
testing.

Comment:
One
commenter
(
OAR­
2005­
0117­
0092)
requested
that
EPA
clarify
that
a
surrogate
use
of
a
facility­
specific
parameter,
such
as
average
carbon
mass
feed
rate
or
ACI
operating
parameters,
is
optional,
and
is
not
required
to
be
applied
to
all
units
at
the
site.
The
commenter
added
that
maximum
and
minimum
operating
parameters
for
Hg
control
can
be
unit
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
61
specific,
and
thus
it
would
not
be
appropriate
to
require
the
use
a
facility­
wide
parameter
on
each
individual
unit.

Response:
EPA
agrees
that
a
facility
retains
the
right
to
decide
to
use
either
a
facilitywide
parameter
or
a
unit­
specific
parameter.
Although
EPA
has
allowed
for
the
option
of
using
a
facility­
wide
parameter,
which
was
intended
to
provide
flexibility,
a
facility
can
still
choose
to
use
unit­
specific
parameters.
EPA
has
left
this
choice
up
to
each
affected
facility.

Comment:
Two
commenters
(
OAR­
2005­
0117­
0073.1,
OAR­
2005­
0117­
0077)

requested
that
EPA
expand
the
list
of
optimized
parametric
monitoring
data
to
include
MWC
unit
loads
and
particulate
control
device
(
PCD)
inlet
temperatures.
The
commenter
added
that
these
parameters
tested
at
one
unit
should
be
applied
to
other
similar
units
if
a
facility
is
eligible
for
the
reduced
dioxin
test
program
as
a
result
of
exceptionally
low
dioxin/
furan
emissions.

Response:
EPA
has
reviewed
the
proposed
regulatory
text
and
it
agrees
that
the
list
of
optimized
parameters
should
be
expanded
when
a
facility
becomes
eligible
for
the
reduced
dioxin
testing
program.
It
was
EPA
intent
to
include
load,
PCD
inlet
temperature,
and
ACI
rate
in
the
list
of
optimized
parametric
monitoring
data
under
the
reduced
dioxin
testing
program.

EPA
has
revised
the
text
in
the
final
amendments
to
include
these
three
parameters.

Comment:
One
commenter
(
OAR­
2005­
0117­
0084)
suggested
that
EPA
consider
including
a
revised
Method
202
in
order
to
quantify
the
total
particulate
emissions.
The
commenter
added
this
method
is
more
appropriate
given
the
trend
toward
regulating
fine
particulate
and
Method
5
does
not
require
reporting
of
the
back­
half
catch
of
condensable
particulate.

Response:
The
data
that
the
particulate
standard
was
based
upon
did
not
include
the
condensible
portion
of
the
sampling
train
and
was
gathered
by
Method
5.
Therefore
use
of
Method
202
did
not
have
public
notice
and
comment
during
proposal.
Therefore
EPA
cannot
consider
using
Method
202.
Again,
a
State
or
MWC
owner
can
always
conduct
additional
testing.
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
62
Chapter
7
Cost
Impacts
Comment:
Twelve
commenters
(
OAR­
2005­
0117­
0088,
OAR­
2005­
0117­
0097,
OAR­

2005­
0117­
0094,
OAR­
2005­
0117­
0087,
OAR­
2005­
0117­
0064,
OAR­
2005­
0117­
0065,
OAR­

2005­
0117­
0076,
OAR­
2005­
0117­
0081,
OAR­
2005­
0117­
0073.1,
OAR­
2005­
0117­
0089,

OAR­
2005­
0117­
0085,
OAR­
2005­
0117­
0077)
disagreed
with
both
of
EPA's
assertions
that:
1)

the
proposed
amendments
will
not
increase
capital
or
operating
costs
for
existing
MWC
units,

and
2)
that
the
proposed
emission
standards
can
be
achieved
with
the
same
pollution
control
technology
used
to
achieve
the
existing
limits.

One
commenter
(
OAR­
2005­
0117­
0095)
suggested
that
the
limited
data
upon
which
EPA
proposed
the
new
performance
standard
underestimated
the
cost
impact
of
the
standard.
The
commenter
suggested
that
cost
to
comply
with
the
rule
is
likely
to
be
in
the
millions
of
dollars.

The
commenter
added
that
if
it
is
required
to
meet
the
proposed
standards,
as
written,
the
standards
would
impose
significant
economic
stress
to
the
facility.

Four
commenters
(
OAR­
2005­
0117­
0088,
OAR­
2005­
0117­
0094,
OAR­
2005­
0117­

0087,
OAR­
2005­
0117­
0085)
suggested
that,
as
a
result
of
the
tightened
limits,
the
margin
of
compliance
will
be
reduced
and
facility
operators
will
have
to
modify
operations
and
costs
to
reduce
their
risk
of
noncompliance.
The
commenter
suggested
that
the
costs
will
increase
due
to
an
increase
in
lime,
carbon,
urea/
ammonia
usage
and
increased
maintenance
requirements.

Four
other
commenters
(
OAR­
2005­
0117­
0097,
OAR­
2005­
0117­
0064,
OAR­
2005­

0117­
0094,
OAR­
2005­
0117­
0087)
suggested
that
the
revised
SO2
and
NOx
limits
will
cause
an
increase
in
costs
in
order
to
comply
with
the
limits
in
the
proposed
rule.
One
commenter
(
OAR
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
63
2005­
0117­
0097)
added
that
the
operating
costs
for
SO2
removal
will
increase
by
an
estimated
$
40,000
annually
as
a
result
of
increased
lime
and
water
usage
to
maintain
the
increased
lime
slurry
feed
rate.
The
same
commenter
also
anticipated
increased
maintenance
costs
for
the
lime
slurry
system
as
a
result
of
the
increased
lime
feed
rate.

One
commenter
(
OAR­
2005­
0117­
0097)
suggested
that
EPA
maintain
the
existing
NOx
emission
limit.
The
commenter
added
that
although
the
facility
did
not
require
a
SNCR
to
meet
the
1995
MACT
NOx
limit
of
250
ppm,
this
facility
would
require
an
SNCR
retrofit
to
meet
the
proposed
219
ppm
NOx
limit.
The
commenter
recognized
EPA's
proposal
to
allow
for
an
increase
in
CO
emissions
from
the
spreader
stoker
RDF­
fired/
100
percent
coal
capable
combustor
in
order
to
lower
NOx
emissions.
However,
the
commenter
contended
that
the
proposed
CO
limits
are
based
on
levels
necessary
to
achieve
the
1995
NOx
limits
and
not
the
CO
levels
that
would
occur
if
the
MWC
must
achieve
the
proposed
NOx
limits.
The
commenter
estimated
that
the
capital
costs
to
retrofit
the
four
units
at
the
facility
to
meet
the
proposed
limits
would
be
$
2
to
3
million,
and
the
annual
operating
costs
are
estimated
to
be
an
additional
$
250,000.
Another
commenter
(
OAR­
2005­
0117­
0076)
contended
that
existing
facilities
meeting
the
existing
NOx
standards
without
ACI
or
SNCR
will
have
to
install
new
control
equipment
in
order
to
meet
the
proposed
NOx
standard.
The
commenter
added
concerns
about
equipment
costs
and
delivery
by
the
2009
compliance
data
due
to
many
industries
competing
for
the
same
types
of
pollution
control
equipment
as
a
result
of
other
federal
regulations.
The
commenter
added
that
EPA
did
not
provide
a
basis
for
justifying
a
more
stringent
proposed
NOx
standard.

One
commenter
(
OAR­
2005­
0117­
0081)
determined
that
all
three
of
its
large
MWC
plants
will
not
be
able
to
continuously
meet
the
proposed
standards
without
retrofitting
pollution
control
equipment,
or
ramping
up
operating
costs
to
meet
higher
removal
efficiencies.

Specifically,
the
commenter
added
that
modifications
required
to
meet
the
proposed
NOx
limits
will
increase
capital
costs
at
four
of
their
MWC
units.
The
commenter
also
added
that
operating
costs
will
increase
to
comply
with
the
proposed
SO2,
HCl,
and
NOx
limits.
The
commenter
also
added
that
the
increased
use
of
lime
and
ammonia
will
increase
emissions
of
ammonia
and
solid
waste
loads.
Further
the
commenter
contended
that
EPA
does
not
present
the
environmental
benefits,
or
impacts
on
energy
associated
with
the
proposed
limits.
The
commenter
concluded
that
EPA
must
develop
a
better
basis
for
justifying
these
added
costs.
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
64
One
commenter
(
OAR­
2005­
0117­
0073.1)
discussed
the
significant
costs
that
came
as
part
of
the
1995
MWC
MACT
standards,
and
suggested
that
the
additional
costs
tied
to
the
2005
proposed
amendments
would
impede
the
viability
of
waste­
to­
energy
and
result
in
marginal
improvements
in
emissions
reductions
from
MWCs.
The
commenter
stressed
the
importance
of
waste­
to­
energy
facilities
in
light
of
the
nation's
need
for
renewable
sources
of
energy
and
discouraged
standards
that
would
make
these
facilities
economically
out
of
reach.

One
commenter
(
OAR­
2005­
0117­
0073.1)
stated
that
devices
such
as
a
spray
dryer
will
have
to
increase
operating
parameters
to
meet
the
higher
removal
efficiencies,
with
a
compliance
margin,
in
the
proposed
standards.
The
commenter
added
that
many
plants
may
have
already
reached
the
design
and
operating
limits
of
their
lime
injection
system
or
the
process
limits
of
their
spray
dryer
absorber.
The
commenter
added
that
the
design
of
the
control
system
and
ash
handling
system
might
be
undersized
to
handle
increased
amounts
of
lime
in
the
control
system
and
associated
increased
capture
of
particulates
in
the
ESP
or
FF.
The
commenter
also
added
that
increased
pollutant
loading
will
impact
the
operating
costs
in
several
ways
including:

shortened
bag
life
for
fabric
filters,
increased
cleaning
frequency,
increased
lime
use,
and
increased
disposal
costs.

Two
commenters
(
OAR­
2005­
0117­
0073.1,
OAR­
2005­
0117­
0077)
emphasized
that,
for
removal
efficiencies
exceeding
95%,
the
lime
feed
rates
become
less
efficient,
removing
only
small
increments
of
pollutant
in
return
for
large
increases
in
lime
feed.
Both
commenters
estimated
the
increased
lime
feed
rate
at
one
plant
to
cost
an
additional
$
220,000
in
order
to
meet
the
new
SO2
and
HCl
standards
with
a
compliance
margin.
In
addition
to
these
costs,
the
commenters
suggested
that
the
increased
lime
feed
rates
would
impact
air,
water,
and
energy
requirements
to
support
the
control
system
as
well
as
adverse
impacts
to
the
environment
resulting
from
an
increased
trucking
of
lime
and
disposal
of
particulate.

Aside
from
operating
costs,
one
commenter
(
OAR­
2005­
0117­
0073.1)
suggested
that
the
capital
components
of
a
control
system
would
also
be
impacted.
The
commenter
added
that
if
a
control
system
at
a
MWC
were
designed
to
meet
the
proposed
limits,
the
facility
would
have
to
install
larger
spray
dryer
absorbers
and
pumping
systems
in
addition
to
larger
ESP
devices
to
handle
the
increased
lime
feed
rates.
Given
that
existing
units
were
designed
to
meet
the
standards
promulgated
in
1995,
the
commenter
stated
that
these
increased
lime
feed
rates
will
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
65
increase
the
particulate
matter
input
to
the
ESP
and
could
affect
the
ESP's
particulate
collection
efficiency
and
impede
compliance
with
the
standards
for
metals,
dioxin
and
particulate.

Several
commenters
(
OAR­
2005­
0117­
0089,
OAR­
2005­
0117­
0085,
OAR­
2005­
0117­

0122)
provided
an
example
of
an
additional
cost
to
consider
associated
with
contractual
agreements
between
MWC
operation
and
maintenance
contractors
and
municipalities,
such
as
the
"
change
in
law"
provision.
The
costs
that
result
from
an
increased
risk
of
non
compliance,

and
the
costs
involved
in
renegotiating
the
contract
fees,
will
impact
the
total
cost
to
the
communities
that
support
the
publicly­
owned
large
MWC.

All
of
these
commenters
concluded
that
the
proposed
emission
rates
will
result
in
increased
costs
due
to
capital,
operating,
and
compliance
risk
expenses.
The
commenter
requested
that
EPA
consider
the
cost
and
environmental
impacts
associated
with
constructing
or
modifying
control
systems
to
meet
the
new
limits.

Two
commenters
(
OAR­
2005­
0117­
0096,
OAR­
2005­
0117­
0083)
requested
that
EPA
provide
additional
technical
data
to
support
the
claim
that
the
proposed
emission
limits
can
be
achieved
using
the
same
air
pollution
control
technology
that
was
used
to
comply
with
the
current
emission
limits.
The
commenters
are
concerned
that
the
proposed
limits
require
facilities
to
maintain
tight
compliance
margins,
which
will
lead
to
increased
costs.

One
commenter
(
OAR­
2005­
0117­
0065)
informed
EPA
that
a
facility
it
operates
in
Charleston,
with
a
SD/
ACI/
SNCR/
ESP
control
will
not
be
able
to
meet
the
proposed
limit
for
lead
and
will
not
be
able
to
continuously
meet
other
pollutant
limits
without
replacing
its
ESP
with
a
fabric
filter.
The
commenter
suggested
that
this
replacement
will
cost
$
10
million.

Response:
EPA
has
considered
the
commenters'
concerns
and
contends
that
its
revised
standards
will
not
significantly
increase
capital
or
operating
costs
at
existing
MWC
units.

Further,
EPA
believes
its
revised
analysis,
using
a
more
robust
database,
demonstrates
that
the
standards
contained
in
the
final
rule
can
be
achieved
with
the
same
pollution
control
technology
used
to
achieve
the
existing
limits
with
an
adequate
compliance
margin.
These
adjustments
to
the
final
standards
have
addressed
some
of
the
commenters'
concerns
with
respect
to
added
cost
from
additional
compliance
risk
and
contractual
agreements.
For
example,
the
SO2
and
HCl
limits
in
the
final
amendments
are
the
same
as
in
the
1995
MWC
rules
and
some
NOx
limits
are
also
higher
than
in
the
proposed
amendments,
so
the
commenters'
concerns
that
there
would
be
added
costs
to
meet
the
lower
emission
limits
for
these
pollutants
are
no
longer
applicable.
EPA
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
66
has
promulgated
the
final
amendments
under
the
assumption
that
these
amendments
will
not
cause
any
significant
economic
impact.
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
67
Chapter
8
Miscellaneous
8.1
Comment
Period
Comment:
Nine
commenters
(
OAR­
2005­
0117­
0073.1,
OAR­
2005­
0117­
0094,
OAR­

2005­
0117­
0087,
OAR­
2005­
0117­
0064,
OAR­
2005­
0117­
0089,
OAR­
2005­
0117­
0076,
OAR­

2005­
0117­
0085,
OAR­
2005­
0117­
0077,
OAR­
2005­
0117­
0088)
declared
that
the
review
and
comment
period
for
the
proposed
standard
is
unreasonable.
One
commenter
(
OAR­
2005­
0117­

0073.1)
suggested
that
the
shortened
period
of
45
days
was
in
reality
shorter
due
to
the
fact
that
this
review
period
fell
over
several
major
religious
and
federal
holidays.
The
commenter
also
added
that
the
issues
raised
by
the
proposed
standards
are
very
data
intensive,
which
requires
stakeholders
to
evaluate
a
large
database
containing
emission
information.
Three
commenters
(
OAR­
2005­
0117­
0094,
OAR­
2005­
0117­
0087,
OAR­
2005­
0117­
0064)
added
that
the
period
was
not
sufficient
for
individual
municipalities
to
prepare
detailed
comments.

Response:
EPA
recognizes
the
shortened
length
of
time
provided
for
public
comment.

However,
given
the
schedule
required
in
order
to
meet
a
Court
order
to
promulgate
this
standard,

EPA
is
unable
to
provide
a
longer
amount
of
time
for
public
comment.

8.2
State
Plans
Comment:
One
commenter
(
OAR­
2005­
0117­
0104)
requested
EPA
to
develop
a
streamlined
procedure
for
efficiently
updating
the
State
Plan.
The
commenter
suggested
that
the
permitting
authority
modify
title
V
or
local
air
district
permits
with
the
proposed
emission
limits,

and
then
submit
the
permits
to
EPA
as
a
State
Plan
update.
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
68
Response:
The
procedure
for
updating
the
State
Plan
is
outside
the
scope
of
this
specific
rulemaking.
However,
EPA
will
consider
these
comments
when
guidance
is
provided
on
the
procedures
for
updating
the
State
plans.

8.3
Operator
Training
Comment:
Three
commenters
(
OAR­
2005­
0117­
0065,
OAR­
2005­
0117­
0073.1,
OAR­

2005­
0117­
0077))
requested
that
the
training
requirement
of
40
CFR
60.54b(
d)
be
removed
from
the
emission
guidelines
and
NSPS.
The
commenters
added
that
the
training
provision
is
a
subset
of
the
ASME
QRO
certification
program
and
is
unnecessary.
One
commenter
(
OAR­
2005­

0117­
0065)
added
that
certain
sections
of
the
1993
version
of
the
manual
are
outdated
and
less
useful
to
operators
in
the
present.

The
commenters
believe
the
training
requirements
have
become
obsolete
and
have
been
replaced
by
ASME
QRO
or
state
certification
programs.
The
commenters
suggested
that
the
intent
of
the
requirements
in
40
CFR
60.54(
d)
was
to
bridge
the
gap
in
training
during
the
initial
implementation
period
of
MACT
until
operators
could
obtain
ASME
or
State
certification,
and
now
the
EPA
training
requirements
have
become
obsolete
and
redundant.

Two
commenters
(
OAR­
2005­
0117­
0073.1,
OAR­
2005­
0117­
0077)
suggested
that
a
provisionally
certified
operator
who
is
newly
promoted
or
recently
transferred
to
a
fully
certified
operator/
shift
supervisor
be
able
to
serve
up
to
six
months
without
notification
before
taking
the
ASME
QRE
certification
exam.
The
commenter
added
that
if
the
period
of
temporary
service
exceeds
six
months,
the
facility
will
submit
notifications
and
progress
reports
to
EPA.

Response:
Section
129
of
the
CAA
requires
us
to
retain
the
training
requirements
set
forth
in
40
CFR
60.54(
d)
and
therefore
EPA
is
unable
to
remove
these
requirements
even
if
there
is
overlap
with
the
ASME
QRO
certification
program.
As
noted
in
the
preamble
to
the
proposed
rule,
the
ASME
was
in
the
midst
of
updating
the
1993
version
of
the
manual.
This
updated
manual
(
QRO­
1­
2005)
was
published
on
February
8th
2006.
However,
because
of
the
late
publication
date
of
the
QRO
and
the
fact
that
new
provisions
affect
operators
of
smaller
MWC
units,
EPA
has
not
incorporated
the
new
QRO
procedures
into
the
final
MWC
rule
at
this
time.

EPA
agrees
with
the
commenters'
suggestion
to
allow
for
a
provisionally
certified
operator
to
serve
without
notification
prior
to
taking
the
ASME
QRE
exam
for
up
to
six
months.

In
order
to
allow
for
flexibility
in
operator
and
testing
schedules
this
allowance
has
been
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
69
included
in
the
final
amendments
and
also
required
the
source
to
submit
notifications
and
progress
reports
to
EPA
when
the
period
of
a
provisional
operator
extends
past
six
months.

8.4
Renewable
Energy
Comment:
Four
commenters
(
OAR­
2005­
0117­
0094,
OAR­
2005­
0117­
0087,
OAR­

2005­
0117­
0076,
OAR­
2005­
0117­
0064)
requested
that
EPA
consider
that
Large
MWC
provide
a
form
of
renewable
energy
and
reduce
greenhouse
gases.
One
commenter
added
that
it
would
be
unfortunate
to
adopt
proposed
standards
that
limit
the
likelihood
of
construction
of
new
facilities
or
impede
the
operations
at
existing
facilities.

Response:
In
developing
the
revised
standards,
it
is
not
EPA
intent
to
discourage
this
type
of
renewable
energy.
EPA
believes
the
final
standards
provide
for
new
facilities
to
enter
this
energy
market
while
ensuring
the
safety
of
human
health
and
the
environment.

8.5
Support
Other
Commenters
Comment:
Ten
commenters
fully
supported
the
issues
raised
and
suggestions
provided
by
commenter
OAR­
2005­
0117­
0073.1.

Response:
Please
see
the
individual
comments
raised
by
commenter
OAR­
2005­
0117­

0073.1
for
EPA
responses.

Comment:
One
commenter
fully
supported
the
issues
raised
and
suggestions
provided
by
SWANA
(
OAR­
2005­
0117­
0089.1).

Response:
Please
see
the
individual
comments
raised
by
commenter
OAR­
2005­
0117­

0089.1
for
EPA
responses.

8.6
Revised
Emission
Limits
Below
are
the
revised
numerical
emission
limits
that
are
included
in
the
revised
final
rule.

Further
discussion
on
the
analysis
associated
with
developing
these
limits
is
provided
in
Chapter
5
of
this
document.
0154­
03­
002\
Boiler
MACT
Recon
­
comment­
response.
doc
70
Table
8­
1
Final
Results
for
Computed
UCLs
and
LCLs
for
Stack
Test
and
CEM
Datac
Upper
Confidence
Limit
(
UCL)

Pollutant
95%
98%
99%
99.726%
a
SD/
FF/
CI/
SNCR
Units
CDD/
CDF
(
ng/
dscm)
13.0
22.5
32.5
NA
PM
(
mg/
dscm)
11.9
15.7
18.6
NA
Cd
(
ug/
dscm)
3.4
5.0
6.5
NA
Pb
(
ug/
dscm)
56.5
95.2
135.4
NA
Hg
(
ug/
dscm)
35.1
55.5
75.7
NA
Hg
Reduction
(%)
b
77.8
74.5
72.4
NA
HCl
(
ppmV)
23.5
28.4
32.0
NA
HCl
Reduction
(%)
b
96.4
95.8
95.5
NA
SO2
(
ppmV)
27.9
46.3
65.2
115.9
SO2
Reduction
(%)
b
80.1
76.7
74.3
70.1
NOx
(
ppmV)
196.4
201.0
203.9
208.6
SD/
ESP/
CI/
SNCR
Units
CDD/
CDF
(
ng/
dscm)
14.5
23.3
32.8
NA
PM
(
mg/
dscm)
18.4
21.4
23.5
NA
Cd
(
ug/
dscm)
24.6
28.8
31.9
NA
Pb
(
ug/
dscm)
262.7
304.8
335.5
NA
Hg
(
ug/
dscm)
62.7
73.0
80.5
NA
Hg
Reduction
(%)
b
69.2
64.6
61.2
NA
HCl
(
ppmV)
23.8
26.8
29.1
NA
HCl
Reduction
(%)
b
95.8
95.3
94.9
NA
SO2
(
ppmV)
26.5
30.9
34.1
40.0
SO2
Reduction
(%)
b
70.0
65.6
62.5
56.6
NOx
(
ppmV)
192.8
204.2
212.5
227.8
a
­­
99.726
value
if
based
on
one
exceedance
in
365
days
(
1
­
1/
365
=
0.99726).
b
­­
%
reduction
values
are
based
on
LCL
values
instead
of
UCL.
c
­­
the
concept
of
upper
confidence
limits
(
UCL)
is
discussed
in
the
promulgation
preamble.
