TO:
Bill
Maxwell,
U.
S.
Environmental
Protection
Agency,
OAQPS
(
C439­
01)

FROM:
Jeffrey
Cole,
RTI
International
DATE:
December
2003
SUBJECT:
MACT
Floor
Analysis
for
Coal­
and
Oil­
Fired
Electric
Utility
Steam
Generating
Units
National
Emission
Standards
for
Hazardous
Air
Pollutants
This
memorandum
describes
the
development
process
for
the
Maximum
Achievable
Control
Technology
(
MACT)
floor
for
the
coal­
and
oil­
fired
electric
utility
steam
generating
units
under
the
National
Emission
Standards
for
Hazardous
Air
Pollutants
(
NESHAP).
The
memorandum
presents
the
methodology
used
to
develop
the
MACT
floor,
the
assumptions
used
for
the
analysis,
the
data
sources,
and
the
resulting
MACT
floor
determination
for
new
and
existing
sources.

MEMORANDUM
OUTLINE
1.0
Introduction
2.0
Background
Information
2.1
Statutory
and
Regulatory
Requirements
2.2
Data
Sources
2.3
Affected
Source
Definition
2.4
Description
of
Industry
Characteristics
3.0
Subcategorization
Analysis
3.1
Coal
and
Oil
Subcategories
3.2
Subcategorization
within
Coal­
fired
Units
3.3
Subcategorization
within
Oil­
fired
Units
3.4
Subcategorization
Options
Considered
3.5
Subcategorization
Scheme
for
Existing
Units
3.6
Subcategorization
Scheme
for
New
Units
4.0
Evaluation
of
MACT
Floor
Performance
for
Existing
Units
4.1
Pollution
Prevention
Alternatives
4.2
Regulatory
Approach
4.3
Control
Technology
Performance
Analysis
5.0
Emission
Limitation
Determination
for
Existing
Units
2
5.1
Emission
Limitation
Format
5.2
Variability
Issues
5.2.1
General
Discussion
of
Variability
in
Data
5.2.2
Strategy
to
Address
Variability
for
Hg
5.2.3
Strategy
to
Address
Variability
for
Ni
5.3
Emission
Limitation
Calculations
5.3.1
Mercury
Emission
Limitation
Calculation
5.3.2
Nickel
Emission
Limitation
Calculation
6.0
Evaluation
of
MACT
Floor
Performance
for
New
Units
6.1
Pollution
Prevention
Alternatives
for
MACT
for
New
Units
6.2
Control
Technology
Performance
Evaluations
7.0
Emission
Limitation
Determination
for
New
Coal­
and
Oil­
fired
Units
7.1
Emission
limitation
format
for
New
Coal­
and
Oil­
fired
Units
7.2
Variability
Issues
7.2.1
General
7.2.2
Strategy
to
Address
Variability
for
Hg
7.2.3
Strategy
to
address
variability
for
Ni
7.3
Emission
Limitation
Calculations
for
New
Units
7.3.1
Mercury
Emission
Limitation
Calculation
for
New
Units
7.3.4
Nickel
Emission
Limitation
Calculation
for
New
Units
8.0
Other
issues
8.1
Blended
Coals
8.2
Dual­
fired
units
8.3
Cogeneration
Units
9.0
References
10.0
Appendices
10.1
List
of
Acronyms
10.2
Data
and
Emission
Limitation
Calculations
10.3
Mercury
Speciation
Analysis
by
Coal
Rank
1.0
INTRODUCTION
Over
the
span
consisting
approximately
the
last
10
years,
EPA
along
with
the
Department
of
Energy
(
DOE)
and
industry
stakeholders
have
been
researching
information
and
gathering
and
analyzing
data
for
development
of
the
MACT
standard
for
electric
utility
steam
generating
units.

The
process
has
culminated
in
several
reports
and
publications
by
EPA
and
others.
This
memorandum
serves
to
provide
an
overview
of
the
process
as
it
relates
to
development
of
the
MACT
floor
for
the
standard.
The
docket
for
the
standard
development
contains
studies,
data,

reports
and
memoranda
that
support
and
provide
basis
for
the
discussion
below.
The
EPA
has
determined
that
the
MACT
standard
will
only
address
mercury
(
Hg)
from
coal­
fired
units
and
3
nickel
(
Ni)
from
oil­
fired
units.
For
the
sake
of
simplicity,
the
term
"
electric
utility
steam
generating
unit,"
as
defined
in
Clean
Air
Act
(
CAA)
section
112(
a)(
8),
will
be
referred
to
as
"
unit"
(
i.
e.
referring
to
either
coal­
fired
or
oil­
fired)
in
this
memorandum.
In
addition,
the
acronym
"
HAP"
as
used
in
discussion
refers
to
only
Hg
and/
or
Ni
as
is
appropriate
for
the
use
of
the
term
in
context.
For
ease
of
reference,
Appendix
10.1
contains
a
list
of
acronyms
used
throughout
the
memorandum.

2.0
BACKGROUND
INFORMATION
2.1
Statutory
and
Regulatory
Requirements
Section
112(
a)(
8)
of
the
Act
defines
an
"
electric
utility
steam­
generating
unit"
as
"
any
fossil­
fuel­
fired
combustion
unit
of
more
than
25
megawatts
electric
(
MWe)
that
serves
a
generator
that
produces
electricity
for
sale."
A
unit
that
cogenerates
steam
and
electricity
and
supplies
more
than
one­
third
of
its
potential
electric
output
capacity
and
more
than
25
MWe
output
to
any
utility
power
distribution
system
for
sale
is
also
considered
an
electric
utility
steam­
generating
unit.

All
standards
established
pursuant
to
section
112(
d)
of
the
Act
must
reflect
the
maximum
degree
of
HAP
emission
reduction
achievable
by
the
industry
source
category.
For
existing
sources,
MACT
cannot
be
less
stringent
than
the
average
emission
reduction
achieved
by
the
bestperforming
12
percent
of
sources
for
a
category
or
subcategory
with
30
or
more
sources
for
which
the
Administrator
has
data.
The
term
"
average,"
as
it
pertains
to
MACT
floor
determinations
for
existing
sources
and
described
in
section
112(
d)(
3)
of
the
Act,
is
not
defined
in
the
statute.
In
a
Federal
Register
notice
published
on
June
6,
1994
(
59
FR
29196),
the
EPA
announced
its
conclusion
that
Congress
intended
"
average"
as
used
in
section
112(
d)(
3)
to
mean
a
measure
of
mean,
median,
mode,
or
some
other
measure
of
central
tendency.
The
EPA
concluded
that
it
retains
substantial
discretion
within
the
statutory
framework
to
set
MACT
floors
at
appropriate
levels,
and
that
it
construes
the
word
"
average"
(
as
used
in
section
112(
d)(
3))
to
authorize
the
EPA
to
use
any
reasonable
method,
in
a
particular
factual
context,
of
determining
the
central
tendency
of
a
data
set.
For
new
sources,
the
Act
requires
that
MACT
be
based
on
the
degree
of
emissions
reductions
achieved
in
practice
by
the
best­
controlled
similar
source.

These
minimum
stringency
levels
are
often
referred
to
as
the
"
MACT
floor."
The
MACT
4
floor
is
based
on
any
combination
of
measures
or
techniques
that
are
ascertained
to
have
contributed
to
the
level
of
emission
reductions
demonstrated
by
the
best
unit(
s)
(
e.
g.,
pollution
prevention
alternatives,
capture
and
control
technologies,
operational
limitations,
or
work
practices).
The
MACT
standard
for
the
source
category
would
require
all
new
and
existing
sources
in
the
source
category
to
achieve
the
corresponding
"
floor"
level
of
performance
on
a
continuous
basis.
Because
the
MACT
represents
the
level
of
reduction
in
HAP
emissions
that
is
actually
demonstrated
by
the
best­
controlled
source(
s)
in
the
subcategory,
EPA
may
not
consider
cost
and
other
impacts
in
determining
the
standard.

The
following
sections
describe
the
process
used
by
EPA
to
determine
the
MACT
floor
for
new
and
existing
units
in
the
coal­
and
oil­
fired
electric
utility
source
category.

2.2
Data
Sources
Various
sources
of
data
were
used
in
the
MACT
floor
analysis
for
coal­
fired
and
oil­
fired
electric
utility
units.
For
the
coal­
fired
units,
EPA's
1999
Electric
Utility/
Information
Collection
Effort
(
EU/
ICE)
Part
II
configuration
database
was
used
to
characterize
the
number
and
types
of
existing
boilers,
the
types
of
fuels
burned,
the
capacity
of
the
boilers,
the
types
of
existing
add­
on
control
technologies,
and
the
locations
of
these
facilities.
This
database
includes
information
on
1,143
units
(
in
1999).
The
1,143
units
were
located
at
a
total
of
461
facilities.
The
EU/
ICE
Part
II
coal
analysis
database
was
used
to
characterize
the
mercury
content,
chlorine
content,
and
other
fuel
constituents
from
all
coal­
fired
electric
utility
plants
for
an
entire
year
(
1999).
This
database
contains
a
minimum
of
three
coal
analyses
per
month
per
plant,
as
well
as
detailed
fuel
usage
data
on
every
operating
coal­
fired
electric
utility
boiler
in
1999.
The
EPA
used
the
EU/
ICE
Part
III
stack
emission
database,
which
contained
the
results
of
all
usable
EU/
ICE
speciated
mercury
emissions
stack
test
reports.
The
EU/
ICE
Part
III
stack
emission
database
contains
data
from
80
stack
emissions
tests
at
79
units
(
one
unit
was
tested
twice
at
different
times).
The
EPA
also
used
data
from
the
U.
S.
Department
of
Energy
(
DOE)
Energy
Information
Administration
(
EIA)
Form
EIA­
767
(
1999)
"
Steam­
Electric
Plant
Operation
and
Design
Report"
to
obtain
data
for
use
in
estimating
MACT
impacts
and
costs.

To
obtain
data
for
the
oil­
fired
units
to
characterize
the
number
and
types
of
existing
boilers,
the
types
of
fuels
burned,
the
capacity
of
the
boilers,
the
types
of
existing
add­
on
control
technologies,
estimating
MACT
impacts
and
costs,
and
the
locations
of
these
facilities,
the
5
DOE/
EIA
Form
EIA­
767
(
2001)
"
Steam­
Electric
Plant
Operation
and
Design
Report"
was
used.

This
analysis
of
this
database
includes
information
on
218
oil­
fired
units
(
in
2001).
To
characterize
emissions
from
oil­
fired
units,
stack
test
data
from
the
EPRI
PISCES
emissions
testing
study
were
used.
All
of
these
databases
are
available
on
EPA's
Electric
Utility
Steam
Generating
Units
Section
112
Rule
Making
Web
site
(
http://
www.
epa.
gov/
ttn/
atw/
combust/
utiltox/
utoxpg.
html),
as
well
as
in
EPA
legacy
docket
A­
92­
55.
Other
data
sources
used
during
the
MACT
floor
analysis
were
published
emission
test
results
and
regulatory
permit
information
that
pertain
to
coal­
and
oil­
fired
electric
utility
units
and
from
various
state
air
pollution
control
agencies.

2.3
Affected
Source
Definition
An
affected
source
under
MACT
is
defined
as
the
equipment
or
collection
of
equipment
or
practices
to
which
the
MACT
standard
limitations
or
control
technology
is
applicable.
The
EPA
evaluated
the
effect
of
several
affected
source
definitions
on
development
of
the
standard.
The
CAA
defined
the
source
category
to
be
a
electric
utility
steam
generating
unit.
This
definition
could
include
all
furnace/
boiler
combinations
at
an
electric
utility
plant
or
could
refer
to
only
one
boiler
furnace
configuration.
The
basic
tenet
of
defining
"
affected
source"
for
the
purposes
of
standard
development
is
to
provide
a
specific
identity
to
the
regulated
source
of
emissions
to
a
degree
that
is
reasonable
and
practical,
while
accomplishing
the
goals
of
the
standard.
In
particular,
the
affected
source
definition
should
be
specific
enough
to
avoid
overly
burdening
the
industry
by
regulating
fugitive
or
incidental
emissions
at
a
facility.
The
EPA
determined
that
the
affected
source
would
be
either
an
individual
coal­
or
oil­
fired
unit
or
a
group
of
units,
particularly
where
units
are
commonly
controlled.
An
individual
unit
consists
of
the
combination
of
a
furnace
firing
a
boiler
used
to
produce
steam,
which
in
turn
is
used
for
a
steam­
electric
generator
that
produces
electrical
energy
for
sale.
This
affected
source
definition
is
based
on
information
provided
to
EPA
that
indicates
that
the
configurations
of
electric
utility
units
at
facilities
are
diverse
and
that
the
intent
of
the
standard
can
be
realized
at
either
the
individual
or
group
level
of
identification
of
the
sources.

The
EPA
determined
that
further
definition
of
the
affected
source
was
necessary
with
regard
to
units
that
burn
multiple
types
of
fuels.
The
EPA
determined
that:

°
If
a
unit
burns
coal
(
either
as
a
primary
fuel
or
as
a
secondary
fuel),
or
any
combination
of
6
coal
with
any
other
fuel,
the
unit
is
considered
to
be
coal­
fired
under
the
standard.

°
If
a
unit
burns
oil
only,
or
oil
in
combination
with
natural
gas
(
except
as
noted
below),
the
unit
is
considered
to
be
oil­
fired
under
the
standard.

°
If
a
new
or
existing
unit
burns
natural
gas
exclusively
or
natural
gas
in
combination
with
oil
whereas
the
oil
constitutes
less
than
two
percent
of
the
unit's
annual
fuel
consumption,

the
unit
is
considered
to
be
natural
gas­
fired
and
would
be
exempt
from
the
standard.

2.4
Description
of
Industry
Characteristics
The
EPA
conducted
a
thorough
analysis
of
all
the
data
sources
mentioned
above
to
gain
an
understanding
of
the
coal­
and
oil­
fired
electric
utility
industries'
process
configurations,
fuel
characteristics,
and
Hg
and
Ni
emissions.
The
following
section
contains
an
overview
of
the
industry
characteristics
pertinent
to
determination
of
the
MACT
floor
level
of
performance
and
development
of
the
MACT
floor
basis.

2.4.1
Fuel
of
Coal­
fired
Units.
For
the
purpose
of
the
development
of
the
MACT
standard,
coal
is
defined
as
all
solid
fossil
fuels
classified
as
anthracite,
bituminous,
subbituminous,

or
lignite
by
American
Society
for
Testing
and
Materials
(
ASTM)
Designation
D388­
77,
90,
91,

95,
or
98a.
The
ASTM
standard
classifies
coals
by
rank,
a
term
which
relates
to
the
carbon
content
of
the
coal
and
other
related
parameters
such
as
volatile­
matter
content,
heating
value,

and
agglomerating
properties.
The
higher
heating
value
(
HHV)
of
coal
is
measured
as
the
gross
calorific
value,
reported
in
British
thermal
units
per
pound
(
Btu/
lb).
The
heating
value
of
coal
increases
with
increasing
coal
rank.
The
youngest,
or
lowest
rank,
coals
are
termed
lignite.

Lignites
have
the
lowest
heating
value
of
the
coals
typically
used
in
power
plants.
Their
moisture
content
can
be
as
high
as
30
percent,
but
their
volatile
content
is
also
high;
consequently,
they
ignite
easily.
Next
in
rank
are
subbituminous
coals,
which
also
have
a
relatively
high
moisture
content,
typically
ranging
from
15
to
30
percent.
Subbituminous
coals
also
are
high
in
volatile
matter
content
and
ignite
easily.
Their
heating
value
is
generally
in
between
that
of
the
lignites
and
the
bituminous
coals.
Bituminous
coals
are
next
in
rank,
with
higher
heating
values
and
lower
moisture
and
volatile
content
than
the
subbituminous
and
lignite
coals.
Anthracites
are
the
highest
rank
coals.
The
coal­
fired
electric
utility
industry
combusts
the
following
coal
ranks,
presented
in
decreasing
HHV
order:
anthracite,
bituminous,
subbituminous,
lignite,
and
coal
refuse
(
i.
e.,
7
anthracite
coal
refuse
[
culm],
bituminous
coal
refuse
[
gob],
and
subbituminous
coal
refuse).

Because
of
the
difficulty
in
obtaining
and
igniting
anthracite,
only
a
single
electric
utility
boiler
in
the
United
States
currently
(
1999)
burns
anthracite
as
its
only
fuel.
Because
bituminous
coal
is
the
most
similar
coal
to
anthracite
coal
based
on
coal
physical
characteristics
(
ash
content,
sulfur
content,
HHV),
anthracite
coal
is
considered
to
be
equivalent
to
bituminous
coal
for
the
purposes
of
the
MACT
development
process
and,
thus,
the
anthracite­
fired
unit
is
considered
a
bituminous­
fired
unit.
Although
there
is
overlap
in
some
of
the
ASTM
classification
properties,

the
ASTM
method
of
classifying
coals
by
rank
generally
is
successful
in
identifying
some
common
core
characteristics
that
have
implications
for
power
plant
design
and
operation.

The
rank
of
coal
to
be
burned
has
a
significant
impact
on
overall
plant
design.
The
goal
of
the
plant
designer
is
to
arrange
boiler
components
(
furnace,
superheater,
reheater,
boiler
bank,

economizer,
and
air
heater)
to
provide
the
rated
steam
flow,
maximize
thermal
efficiency,
and
minimize
cost.
Engineering
calculations
are
used
to
determine
the
optimum
positioning
and
sizing
of
these
components,
which
cool
the
flue
gas
and
generate
the
superheated
steam.
The
accuracy
of
the
parameters
specified
by
the
owner/
operators
is
critical
to
designing
and
building
an
optimal
plant.
The
rank
of
coal
burned
also
has
significant
impact
on
the
design
and
operation
of
the
emission
control
equipment
(
e.
g.,
ash
resistivity
impact
on
ESP
performance).

One
of
the
most
important
factors
in
modern
electric
utility
boiler
design
involves
the
range
of
coal
ranks
to
be
fired,
which
determines
the
design
specifications
and
overall
arrangement
of
boiler
components.
Coal
rank
is
so
important
that
plant
designers
and
manufacturers
require
a
complete
list
of
all
coal
ranks
presently
available
or
planned
for
future
use,
along
with
their
complete
chemical
and
ash
analyses,
so
that
the
engineers
can
properly
design
and
specify
plant
equipment.
The
various
coal
characteristics
(
e.
g.,
how
hard
the
coal
is
to
pulverize,
how
high
its
ash
content,
the
chemical
content
of
the
coal,
how
big
the
boiler
has
to
be
to
adequately
utilize
the
heat
content,
etc.),
all
impact
on
boiler
design
from
the
pulverizer
through
the
boiler
to
the
final
steam
tubes.
For
a
boiler
to
operate
efficiently,
it
is
critical
to
recognize
the
differences
in
coals
and
make
the
necessary
modifications
in
boiler
components
during
design
to
provide
optimum
conditions
for
efficient
combustion.
As
would
be
expected,

coal­
fired
units
are
designed
and
constructed
with
different
process
configurations
partially
because
of
the
constraints
placed
on
the
initial
design
of
the
unit
by
the
fuel
to
be
used.
8
Accordingly,
these
site­
specific
constraints
dictate
the
process
equipment
selected,
the
component
order,
the
materials
of
construction,
and
the
operating
conditions.

The
EPA
found
that
a
portion
of
the
coal­
fired
units
burn
more
than
one
rank
of
coal.

Approximately
23
percent
of
coal­
fired
electric
utility
steam
generating
units
either
(
1)
co­
fire
two
or
more
ranks
of
coal
(
with
or
without
other
fuels)
in
the
same
boiler,
or
(
2)
fire
two
or
more
ranks
of
coal
(
with
or
without
other
fuels)
in
the
same
boiler
at
different
times.
2
This
coal
"
blending"
is
done
generally
for
one
of
three
reasons:
to
achieve
sulfur
dioxide
(
SO
2)
emission
compliance
with
CAA
Title
IV
provisions,
to
prevent
excessive
slagging
by
improving
the
heat
content
of
a
lower
grade
coal,
or
for
economic
reasons
(
i.
e.,
coal
rank
price
and
availability).

However,
these
blended
coals,
although
of
different
rank,
have
similar
properties.
That
is,

because
of
the
overlap
in
various
characteristics
in
the
ASTM
definitions
of
coal
rank,
certain
bituminous
and
subbituminous
coals
(
for
example)
exhibit
similar
handling
and
combustion
properties.
Plant
designers
and
operators
have
learned
to
accommodate
these
blends
in
certain
circumstances
without
significant
impact
on
plant
operation
or
control.
The
majority
of
coal­
fired
units
in
the
United
States
firing
multiple
coals
fire
bituminous
and
subbituminous
coals,
either
through
direct
blending
or
through
independently
combusting
each
coal
at
some
period
during
the
year.
In
the
United
States,
the
number
of
units
that
burn
a
majority
of
bituminous
coal
in
their
mix
(
9
percent
of
all
units
in
the
United
States)
is
nearly
double
the
number
of
units
that
burn
a
majority
of
subbituminous
coal
(
5
percent
of
all
units
in
the
United
States).
Also,
some
units
co­
fire
subbituminous
and
lignite
coals.

The
flue
gases
resulting
from
the
combustion
of
these
different
coal
ranks
can
exhibit
different
Hg
emissions
characteristics.
These
Hg
emissions
characteristics
consist
of
varying
percentages
of
the
three
relevant
forms
(
or
species)
of
Hg
(
particulate­
bound,
oxidized
[
ionic],

and
elemental)
that
make
up
the
total
Hg
in
the
flue
gas.

Analysis
of
available
source
test
data
and
Hg
in
coal
data
shows
that
combustion
of
bituminous
coal
results
in
Hg
emissions
that
are
composed
of
relatively
more
oxidized
Hg
compared
to
the
other
coal
ranks.
Combustion
of
bituminous
coal
produces
the
most
particulate­
bound
Hg
of
any
of
the
three
major
coal
ranks
combusted.
Combustion
of
subbituminous
coal
results
in
emissions
that
are
composed
of
relatively
more
Hg0
(
compared
to
bituminous
coal),
with
little
particulate­
bound
Hg
(
less
than
half
that
of
bituminous
coal
9
emissions).
Combustion
of
lignite
coal
also
results
in
emissions
that
are
composed
of
relatively
more
Hg0
(
compared
to
bituminous
coal)
with
little
particulate­
bound
Hg
(
also
less
than
half
that
of
bituminous
coal
emissions).
Available
data
indicate
that
emissions
from
the
combustion
of
coal
refuse
tend
to
speciate
almost
entirely
to
particulate­
bound
Hg
(
greater
than
99
percent
for
both
units
tested
in
the
EU/
ICE).
With
few
exceptions,
particulate­
bound
Hg
can
be
removed
with
PM
controls,
oxidized
Hg
can
be
removed
with
wet
SO
2
controls
(
flue
gas
desulfurization
[
FGD]

scrubbers),
but
Hg0
usually
shows
little
to
no
removal
with
any
existing
conventional
type
of
air
pollution
control
device
(
APCD)
such
as
ESP
or
FF
units
used
on
utility
boilers.
Appendix
10.3
of
this
document
provides
the
data
and
results
of
an
analysis
of
coal­
fired
units
with
regard
to
speciation
in
Hg
across
coal
ranks.

Other
types
of
fuel
are
blended
with
coal
for
a
variety
of
unit­
specific
needs.
The
two
most
common
"
supplementary
fuels"
in
the
coal­
fired
industry
are
petroleum
coke
and
tire­
derived
fuel
(
TDF).
These
supplementary
fuels
are
generally
blended
with
a
much
larger
percentage
of
coal
before
combustion.
If
a
unit
were
to
burn
one,
or
a
combination
of
these
supplementary
fuels
exclusively,
it
would
not
be
subject
to
the
coal­
and
oil­
fired
electric
utility
NESHAP.
To
our
knowledge,
oil
is
used
only
during
start­
up
of
coal­
fired
units
and
is
not
a
"
supplementary"
fuel
for
these
units.

2.4.2
Boiler
Firing
Configurations
Used
in
the
Coal­
fired
Electric
Utility
Industry.

There
are
five
basic
types
of
coal
combustion
processes
used
in
the
coal­
fired
electric
utility
industry.
These
are
conventional­
fired
boilers,
stoker­
fired
boilers,
cyclone­
fired
boilers,
IGCC
units,
and
fluidized
bed
combustors
(
FBC).

Conventional
boilers,
also
known
as
pulverized
coal
(
PC)
boilers,
have
a
number
of
firing
configurations
based
on
their
burner
placement.
The
basic
characteristic
that
all
conventional
boilers
have
in
common
is
that
they
inject
PC
and
primary
air
through
a
burner
where
ignition
of
the
PC
occurs,
which
in
turn
creates
an
individual
flame.
Conventional
boilers
fire
through
many
such
burners
mounted
in
the
furnace
walls.

In
stoker­
fired
boilers,
fuel
is
deposited
on
a
moving
or
stationary
grate
or
spread
mechanically
or
pneumatically
from
points
usually
10
to
20
feet
above
the
grate.
The
process
utilizes
both
the
combustion
of
fine
coal
powder
in
air
and
the
combustion
of
larger
particles
that
10
fall
and
burn
in
the
fuel
bed
on
the
grate.

Cyclone­
fired
boilers
use
several
water­
cooled
horizontal
burners
that
produce
high­
temperature
flames
that
circulate
in
a
cyclonic
pattern.
The
burner
design
and
placement
cause
the
coal
ash
to
become
a
molten
slag
that
is
collected
below
the
furnace.

Fluidized
bed
combustors
combust
coal
in
a
bed
of
inert
material
(
e.
g.,
sand,
silica,

alumina,
or
ash)
and/
or
a
sorbent
such
as
limestone
that
is
suspended
through
the
action
of
primary
combustion
air
distributed
below
the
combustor
floor.
"
Fluidized"
refers
to
the
state
of
the
bed
of
material
(
coal
and
inert
material
[
or
sorbent])
as
gas
passes
through
the
bed.
As
the
gas
flow
rate
is
increased,
the
force
on
the
fuel
particles
becomes
just
sufficient
to
cause
buoyancy.
The
gas
cushion
between
the
solids
allows
the
particles
to
move
freely,
giving
the
bed
a
liquid­
like
(
or
fluidized)
characteristic.

Integrated­
coal
gasification
combined
cycle
units
are
specialized
units
in
which
coal
is
first
converted
into
synthetic
coal
gas.
In
this
conversion
process,
the
carbon
in
the
coal
reacts
with
water
to
produce
hydrogen
gas
and
carbon
monoxide
(
CO).
The
synthetic
coal
gas
(
syngas)
is
then
combusted
in
a
combustion
turbine
which
drives
an
electric
generator.
Hot
gases
from
the
combustion
turbine
then
pass
through
a
waste
heat
boiler
to
produce
steam.
This
steam
is
fed
to
a
steam
turbine
connected
to
a
second
electric
generator.

2.4.3
APCD
Used
to
Control
Coal­
fired
Emissions.
Coal­
fired
electric
utility
units
are
controlled
by
a
varied
group
of
APCD
depending
on
individual
fuel
characteristics
and
design
considerations.
The
following
discussion
describes
those
possible
configurations.

a.
PM
Controls.
The
two
major
types
of
PM
APCD
used
in
the
coal­
and
oil­
fired
electric
utility
industry
are
ESP
and
FF.
Particulate
scrubbers
are
used
on
a
limited
number
(
25)

of
coal­
fired
units
in
the
United
States
and
mechanical
APCD
(
multiclones)
are
used
on
41
of
218
oil­
fired
units
and
only
one
coal­
fired
unit.
2
Electrostatic
precipitators
are
the
most
frequently
used
PM
control
devices
on
coal­
and
oil­
fired
electric
utility
units.
They
operate
by
imparting
an
electrical
charge
to
incoming
particles,

then
attracting
the
particles
to
oppositely
charged
plates
for
collection.
The
collected
particles
are
periodically
dislodged
in
sheets
or
agglomerates
by
rapping
the
plates.
Particle
removal
in
an
ESP
depends
largely
on
the
electrical
resistivity
of
the
particles
being
collected.
There
are
two
basic
configurations
of
ESP,
cold­
side
and
hot­
side.
Cold­
side
ESP
are
located
after
the
unit's
air
11
preheater
while
hot­
side
ESP
are
located
before
the
unit's
air
preheater.

Fabric
filters
are
inherently
efficient
and
are
effective
when
high­
efficiency
PM
collection
is
required.
Fabric
filters
collect
PM
by
placing
a
fabric
barrier
in
the
flue
gas
path.
Gas
passes
freely
through
the
fabric,
but
particles
are
trapped
and
retained
for
periodic
removal.

Particulate
scrubbers
operate
by
shattering
streams
of
water
into
small
droplets
that
collide
with
and
trap
PM
contained
in
the
flue
gas
or
by
forcing
the
flue
gas
into
intimate
contact
with
water
films.
The
particle­
laden
droplets
or
water
films
coalesce
and
are
collected
in
a
sump
at
the
bottom
of
the
scrubber.

Mechanical
collectors
are
generally
in
the
form
of
groups
of
cylinders
with
conical
bottoms
(
multicyclones).
Particles
in
the
entering
gas
stream
are
hurled
to
the
outside
of
the
cylinder
by
centrifugal
force
and
are
discharged
at
the
bottom
of
the
cone.

b.
SO
2
Controls.
The
two
major
types
of
SO
2
APCD
used
in
the
coal­
fired
electric
utility
industry
are
known
as
wet
scrubbers
and
dry
scrubbers.

In
a
wet
scrubber,
flue
gas
containing
SO
2
is
brought
into
contact
with
a
alkali
sorbent­
water
slurry.
The
most
often
used
sorbent
is
limestone.
The
SO
2
is
absorbed
into
the
slurry
and
reacts
with
alkali
sorbent
to
form
an
insoluble
sludge.
The
sludge
is
usually
disposed
of
in
a
pond
specifically
constructed
for
the
purpose.

In
a
dry
scrubber
(
known
as
an
spray
dryer
adsorber
[
SDA]),
flue
gas
at
the
air
preheater
outlet
is
contacted
with
fine
spray
droplets
of
hydrated
lime
slurry
in
a
spray
dryer
vessel.
The
SO
2
is
absorbed
in
the
slurry
and
reacts
with
the
hydrated
lime
reagent
to
form
solid
calcium
sulfite
and
calcium
sulfate,
as
in
a
wet
lime
scrubber.
The
water
is
evaporated
by
the
heat
of
the
flue
gas.
The
dried
solids
are
entrained
in
the
flue
gas,
along
with
fly
ash,
and
are
collected
in
a
PM
collection
device
(
FF
or
ESP).

2.4.4
Fuel
for
Oil­
fired
Units.
The
EPA
analyzed
the
data
available
on
the
fuel,
process,

emission
profiles,
and
APCD
for
oil­
fired
units
at
existing
affected
sources.
The
ASTM
classifies
oils
by
"
grade,"
a
term
which
relates
to
the
amount
of
refinement
that
the
oil
undergoes.
The
level
of
refinement
directly
affects
the
metallic
HAP
and
carbon
content
of
the
oil
and
other
related
parameters
such
as
sulfur
content,
heating
value,
and
specific
gravity.
The
heating
value
of
oil
is
measured
as
the
gross
calorific
value,
reported
in
British
thermal
units
per
gallon
(
Btu/
gal),
and
increases
with
increasing
oil
grade.
The
most
refined
oil
used
by
the
oil­
fired
12
electric
utility
industry
is
amber
in
color
and
known
as
distillate
oil
(
also
known
as
medium
domestic
fuel
oil).
The
least
refined
oil
used
by
the
oil­
fired
electric
utility
industry
is
black
in
color
and
known
as
residual
oil
(
also
known
as
Bunker
C
oil).
By
comparison,
distillate
oil
is
lower
in
metallic
HAP,
sulfur,
ash
content,
and
heating
value
but
higher
in
carbon
content
than
residual
oil.
Only
a
handful
of
boilers
(
8
of
218)
fire
distillate
fuel
oil
exclusively.
However,

28
out
of
218
boilers
fire
distillate
oil
and
residual
oil
in
the
same
boiler
(
either
simultaneously
or
at
separate
times).
To
EPA's
knowledge,
number
1,
3,
4,
and
5
fuel
oils
are
not
used
currently
in
the
oil­
fired
electric
utility
industry.
4
The
type
of
oil
to
be
burned
has
little
impact
on
overall
plant
design.
The
goal
of
the
plant
designer
is
to
make
sure
the
plant
can
handle
the
different
viscosities
of
oil
(
and
natural
gas
if
applicable)
that
the
boiler
is
likely
to
combust.
For
example,
because
of
its
viscosity,
residual
oil
must
be
heated
to
make
it
flow
(
i.
e.,
heated
storage
tanks
and
heated
fuel
supply
lines).

An
oil­
fired
electric
utility
boiler
combusts
oil
exclusively,
or
combusts
oil
at
certain
times
of
the
year
and
natural
gas
at
other
times
(
not
simultaneously).
The
choice
of
when
to
combust
oil
exclusively
or
the
blend
of
oil
and
natural
gas
at
a
single
boiler
is
usually
based
on
economics
or
fuel
availability
(
including
seasonal
availability).
Additionally,
the
blended
unit
could
also
potentially
burn
a
blend
of
distillate
and
residual
oil.

2.4.5
Boiler
Firing
Configuration
Used
in
the
Oil
 
fired
Electric
Utility
Industry.

There
is
only
one
basic
type
of
oil
combustion
process
used
in
the
oil­
fired
electric
utility
industry,

known
as
a
conventional­
fired
boiler.
Conventional­
fired
boilers
have
a
number
of
firing
configurations
based
on
their
burner
placement.
The
basic
characteristic
that
all
conventionalfired
boilers
have
in
common
is
that
they
inject
oil
and
primary
air
through
a
burner
where
ignition
of
the
oil
occurs,
which
in
turn
creates
an
individual
flame.
Conventional­
fired
boilers
fire
through
many
such
burners
mounted
in
the
furnace
walls.

2.4.6
APCD
Used
to
Control
Oil
 
fired
Emissions.
Only
79
of
the
218
oil­
fired
units
mentioned
above
have
any
APCD
controlling
their
emissions.
Uncontrolled
units
represented
the
largest
portion
of
the
oil­
fired
units
(
64
percent).
Electrostatic
precipitators
were
used
on
38
units,
which
constitute
17
percent
of
the
oil­
fired
units
in
the
oil­
fired
industry.
Mechanical
controls
(
cyclones
and
multiclones)
were
used
on
41
units,
which
constitute
19
percent
of
the
oil­
fired
units.
Three
units
have
both
mechanical
controls
and
ESP.
These
three
units
with
both
13
controls
were
included
in
the
ESP
equipped
units
count
above.
4
3.0
SUBCATEGORIZATION
ANALYSIS
The
definition
of
affected
source
for
this
source
category
includes
a
wide
range
of
regulated
units
with
varying
process
configurations
and
emission
profile
characteristics.
In
order
to
develop
the
MACT
standard,
EPA
must
consider
the
variation
in
the
sources
within
the
category
to
determine
if
any
variations
between
the
sources
would
be
significant
enough
to
warrant
subcategorization.
The
subcategorization
of
sources
within
the
source
category
is
necessary
when
sources
exhibit
differences
in
operation,
design,
size,
or
raw
materials
used
(
etc.)

that
would
limit
the
feasibility
of
developing
standards
that
equitably
address
the
entire
population
of
sources.
The
EPA
must
provide
a
standard
that
is
based
on
emissions
reductions
that
can
be
achieved
by
all
sources
with
technology
demonstrated
to
be
available
and
effective
for
those
sources.
Therefore,
it
was
necessary
for
EPA
to
determine
the
appropriate
level
of
subcategorization
for
the
coal­
and
oil­
fired
units.
The
criteria
used
by
EPA
in
evaluating
differences
in
sources
for
this
standard
included
the
fuel
used,
the
process
design
or
operation
of
the
unit,
variations
in
emissions
profiles
from
the
source,
and
differences
in
application
of
control
technology
(
APCD
or
work
practices).

3.1
Coal
and
Oil
Subcategories
For
the
coal­
and
oil­
fired
electric
utility
steam
generating
unit
source
category,
the
affected
sources
exhibited
obvious
and
significant
variations
with
regard
to
these
criteria.
The
most
prominent
dissimilarity
was
that
between
coal­
and
oil­
fired
units.
Coal­
and
oil­
fired
units
have
vastly
different
emission
characteristics
due
to
their
fuel
sources.
The
electric
utility
industry
generally
uses
coal­
fired
units
as
base­
loaded
units
(
i.
e.,
the
units
are
designed
to
run
continuously
except
for
maintenance
intervals).
Oil­
fired
units
are
generally
used
as
"
peaking"
units
(
i.
e.,
the
units
are
operated
when
extra
electrical
power
is
needed).
Coal
combustion
produces
higher
emission
levels
of
metals,
halogenated
inorganic
compounds,
and
organic
compounds
than
a
comparably
sized
oil­
fired
unit,
with
the
exception
of
emission
of
Ni
compounds.
For
these
reasons,
EPA
divided
the
affected
sources
into
the
initial
subcategories
of
coal­
and
oil­
fired
units.

Additional
evaluation
of
the
data
were
then
conducted
to
ascertain
if
further
subcategorization
within
coal­
fired
or
within
oil­
fired
units
was
warranted.
14
3.2
Subcategorization
within
Coal­
fired
Units
After
examining
a
number
of
possible
subcategorization
options,
EPA
identified
three
basic
ways
to
subcategorize
coal­
fired
electric
utility
steam
generating
units.

°
A
no
subcategorization
scheme
which
would
treat
all
coal
ranks
as
one,
with
the
MACT
floor
developed
using
all
of
the
coal­
fired
unit
data.

°
Subcategorization
by
coal
rank
which
would
address
the
differences
in
the
characteristics
of
the
Hg
emissions
(
i.
e.,
speciation
of
Hg),
the
resulting
ability
to
control
Hg,
and
the
various
design
and
control
constraints
resulting
from
the
various
coal
ranks.

°
Subcategorization
by
process
type
which
would
address
potential
emissions
differences
produced
by
process
variations,
which
in
turn
lead
to
corresponding
differences
in
the
nature
of
emissions
and
the
technical
feasibility
of
applying
emission
control
techniques.

To
determine
the
appropriate
subcategorization
approach,
EPA
evaluated
fuel,
process,

and
control
technology
to
determine
which
aspect
determined
the
better
performance
by
the
top
units
and
found
that
the
data
did
not
identify
any
common
attribute
that
could
be
credited
with
the
demonstrated
better
performance.
The
EPA
found
that
each
of
the
best­
performing
units
had
a
combination
of
factors
that
was
the
basis
for
the
better
performance.
The
factors
that
were
identified
to
contribute
to
the
better
performance
were
more
closely
fuel­
dependent
than
either
APCD
or
process­
dependent.
The
dependency
on
fuel
as
a
controlling
factor
was
particularly
prominent
with
regard
to
Hg
emissions.
A
top­
performing
unit
may
have
lower
Hg
emissions
because
a
combination
of
events
took
place
(
e.
g.,
the
coal
may
have
been
of
a
lower
Hg
content,

and/
or
the
Hg
may
have
been
primarily
speciated
to
a
particulate
form,
and/
or
the
unit
may
have
been
controlled
for
PM
using
a
FF).
In
this
case,
the
lower
level
of
the
Hg
in
the
coal
and
its
speciation
form
were
the
controlling
factors
in
the
better
performance.
Conversely,
if
the
Hg
level
was
higher
in
the
coal
and/
or
the
Hg
speciated
to
another
form,
the
demonstrated
performance
would
not
be
as
good
even
if
controlled
by
the
same
control
device.
The
data
also
indicated
that
both
factors,
the
Hg
in
the
coal
and
the
speciated
form
of
the
Hg,
are
dependent
on
coal
rank
(
or
even
coal
seam
within
a
rank).

Based
on
this
information,
EPA
then
analyzed
the
available
data
to
determine
which
coal
ranks
were
burned,
and
why,
to
ascertain
if
changing
coal
rank
would
be
a
conceivable
control
strategy.
The
EPA
found
that
the
characteristics
of
the
coal
rank
to
be
burned
were
the
driving
15
factors
in
how
a
coal­
fired
unit
was
designed.
Further,
the
choice
of
coal
ranks
to
be
burned
for
a
given
unit
is
based
on
economic
issues,
including
availability
of
the
coal
within
the
region
or
locale.
A
number
of
coal­
fired
units,
including
all
known
lignite­
fired
units,
are
"
mine
mouth"
(
or
near
mine­
mouth)
operations
(
i.
e.,
the
unit
is
constructed
on
or
near
the
coal
mine
itself,
with
coal
transport
often
being
done
by
conveyor
directly
from
the
mine)
and
many
do
not
have
the
infrastructure
in
place
(
e.
g.,
interstate
rail
lines)
to
import
other
ranks
of
coal
in
quantity
sufficient
to
replace
all
lignite
coal
combusted.
Additionally,
many
plants
have
long­
term
contract
obligations
to
burn
low­
sulfur­
content
coal
to
achieve
compliance
with
SO
2
standards.
This
coal
is
delivered
to
the
plants
by
large
unit­
trains
from
very
long
distances.
Thus,
a
standard
based
on
"
no
subcategorization"
could
be
unachievable
for
such
units
at
such
plants
with
fixed
fuel
delivery
(
by
physical
location
or
by
contractual
agreement)
requirements.
The
EPA
also
found
that
substitution
of
coal
rank,
in
most
cases,
would
require
significant
modification
or
retooling
of
a
unit,
which
would
indicate
a
valid
difference
in
the
design/
operation
of
the
units.
For
these
reasons,
EPA
decided
that
subcategorization
of
coal­
fired
units
based
on
coal
rank
(
fuel
type)
was
warranted.

Although
conventional­,
stoker­,
and
cyclone­
fired
boilers
use
different
firing
techniques,

the
Hg
emissions
characteristics
of
these
boilers
are
similar
(
given
that
common
ranks
of
coal
are
fired)
and,
therefore,
the
units
can
be
grouped
together.
Although
these
units
fire
a
variety
of
coal
ranks,
they
have,
to
date,
only
combusted
coal
refuse
in
lesser
amounts
as
a
secondary
fuel
source.

Based
on
their
unique
firing
designs,
FBC
units
employ
a
fundamentally
different
process
for
combusting
coal
from
that
employed
by
conventional­,
stoker­,
or
cyclone­
fired
boilers.

Fluidized­
bed
combustors
are
capable
of
combusting
many
coal
ranks,
including
coal
refuse.
For
these
reasons,
FBC
units
can
be
considered
a
distinct
type
of
boiler.
However,
the
Hg
emissions
test
data
results
for
FBC
units
were
not
substantially
different
from
those
at
similarly
fueled
conventionally
fired
units
with
similar
emission
levels,
either
in
mass
of
emissions
or
in
emissions
characteristics.
Therefore,
EPA
does
not
believe
subcategorization
of
FBC
units
is
necessary.

Integrated
gasification
combined
cycle
units
combust
a
synthetic
coal
gas.
No
coal
is
directly
combusted
in
the
unit
during
operation
(
although
a
coal­
derived
fuel
is
fired),
and
thus,

IGCC
units
are
a
distinct
class
or
type
of
boiler.
16
3.3
Subcategorization
within
Oil­
fired
Units
The
EPA
analyzed
the
data
available
on
the
fuel,
process,
emission
profiles,
and
APCD
for
oil­
fired
units
at
existing
affected
sources.
The
data
available
to
EPA
on
oil­
fired
units
indicated
that
there
is
very
little
variation
in
the
process
or
control
technologies
used
in
the
industry.

3.4
Subcategorization
Options
Considered
The
EPA
had
determined
that
subcategorizing
coal­
fired
units
into
five
subcategories;

bituminous
coal,
subbituminous
coal,
lignite
coal,
coal
refuse,
and
IGCC
is
the
most
obvious
subcategorization
scheme.
Another
possible
option
is
to
subcategorize
coal­
fired
units
into
four
subcategories
(
bituminous
and
subbituminous
coals,
lignite
coal,
coal
refuse,
and
IGCC).
This
second
option
is
claimed
by
some
industry
sources
to
allow
greater
fuel
choice
flexibility.
As
mentioned
previously,
approximately
23
percent
of
the
coal­
fired
units
in
1999
fired
a
blend
of
coal
ranks
or
coals
and
other
fuels.
2
The
majority
of
blended
coal­
fired
units
in
the
United
States
combust
a
blended
coal
composed
of
bituminous
and
subbituminous
coal,
either
through
direct
blending
or
through
independently
combusting
each
coal
at
some
period
during
the
year.
A
standard
that
would
subcategorize
bituminous
and
subbituminous
together
would
allow
easier
emissions
permitting
and
flexibility
because
most
blended
coal­
fired
units
do
not
keep
the
ratio
of
the
coals
blended
constant.

3.5
Subcategorization
of
Existing
Units
Based
on
the
above
information,
EPA
believes
the
coal­
fired
units
at
existing
affected
sources
should
be
subcategorized
into
five
subcategories
based
on
a
combination
of
coal
rank
and
process
type:
bituminous
(
including
anthracite),
subbituminous,
lignite,
coal
refuse
(
which
includes
anthracite,
bituminous,
and
subbituminous
coal
refuse),
and
IGCC
(
coal
syngas).

Because
few
units
fire
anthracite
coal,
EPA
chose
to
combine
it
with
bituminous
coal
for
the
purposes
of
this
standard
development.
Because
petroleum
coke
and
TDF
do
not
meet
the
definition
of
a
fossil
fuel,
EPA
does
not
believe
that
they
should
be
given
their
own
subcategories.

As
mentioned
above,
the
data
available
to
EPA
on
oil­
fired
units
indicated
that
there
is
very
little
variation
in
the
process
or
control
technologies
used
in
the
industry.
Also,
because
units
burning
greater
than
or
equal
to
98
percent
natural
gas
(
based
on
the
annual
Btus
contributed
by
all
fuels
burned)
would
not
be
subject
to
the
standard
and
units
burning
distillate
oil,
exclusively,
would
be
exempt
from
compliance
requirements
of
the
standard,
EPA
does
not
17
think
that
natural
gas
and
distillate
oil
should
be
given
their
own
subcategories.
Therefore,
EPA
found
no
criteria
that
would
warrant
further
subcategorization
within
existing
oil­
fired
units.

3.6
Subcategorization
of
New
Units
With
regard
to
new
sources,
EPA
has
no
data
that
indicate
that
the
rationale
for
subcategorization
for
existing
sources
would
not
be
applicable
to
new
units.
New
units
constructed
on
the
same
site
as
existing
units
could
still
be
restricted,
at
least
in
concept,
to
the
same
physical
constraints
(
e.
g.,
coal
handling
and
processing,
access
to
interstate
rail
lines)
as
are
the
co­
located
existing
units.
Further,
the
variability
of
Hg
content
within
a
coal
rank
and
within
specific
coal
seams
would
preclude
the
ability
to
find
a
consistent
source
of
low­
Hg
coal.
For
these
reasons,
EPA
believes
that
the
subcategorization
scheme
for
new
coal­
and
oil­
fired
units
should
be
the
same
as
for
the
existing
units.

4.0
EVALUATION
OF
MACT
FLOOR
PERFORMANCE
FOR
EXISTING
UNITS
Once
the
sources
were
subcategorized,
EPA
then
evaluated
each
subcategory
to
determine
the
best
performing
units
and
further,
the
attributes
by
which
the
best
performing
units
demonstrate
the
better
performance.
The
following
sections
summarize
the
evaluation
to
determine
the
MACT
floor
units
and
of
the
attributes
which
could
contribute
to
better
performance.

4.1
Pollution
Prevention
Alternatives
Pursuant
to
current
EPA
policy,
the
development
of
all
MACT
standards
must
consider,

as
a
potential
MACT
control
strategy,
any
pollution
prevention
techniques
that
could
reduce
or
eliminate
the
pollutants
of
concern
from
being
produced
by
the
process.
During
development
of
the
electric
utility
standard
and
other
combustion­
related
rules,
EPA
considered
several
pollution
prevention
techniques
as
alternatives
to
application
of
add­
on
pollution
control
technology.
Each
of
the
measures
considered
are
"
pre­
combustion"
techniques
that
would
address
formation
of
HAP
prior
to
the
fuel
being
combusted
in
the
furnace.
The
measures
evaluated
include
fuel
substitution
or
treatment,
combustion
process
changes,
and
work
practices,
all
of
which
could
potentially
increase
combustion
efficiency
and
decrease
production
of
HAP
from
the
combustion
process.

4.1.1
Fuel
Substitution
or
Treatment.
The
fossil
fuels
used
in
the
electric
utility
18
industry
consist
of
primary
fossil
fuels
such
as
coal,
oil,
and
natural
gas.
In
addition
there
are
several
supplementary
fuels
(
as
mentioned
earlier)
used
to
enhance
the
combustion
process.
The
Administrator
has
previously
determined
that
HAP
emissions
from
the
burning
of
natural
gas
are
not
significant
and,
therefore,
has
determined
that
natural
gas­
fired
units
are
not
included
in
this
standard
development.
It
would
follow
that
since
the
HAP
emissions
from
natural
gas
are
low,
it
would
be
a
desirable
alternative
to
substitute
natural
gas
for
the
coal
and
oil
currently
burned
in
the
industry.
If
a
unit
could
not
switch
to
natural
gas,
then
perhaps
it
would
be
possible
to
decrease
HAP
produced
by
burning
a
"
better"
type
of
coal
or
oil
(
based
on
HAP
content
in
the
fuel
to
start
with
or
the
behavior
of
the
fuel
during
the
combustion
process).
The
EPA
first
considered
the
feasibility
of
fuel
substitution
from
several
perspectives:
(
1)
switching
to
natural
gas;
(
2)
switching
to
other
fuels
in
the
same
subcategory
(
e.
g.,
a
"
lower"
Hg
or
other
HAP
content
bituminous
coal,
or
distillate
oil
instead
of
residual
oil);
or
(
3)
switching
to
fuels
used
in
another
subcategory
(
e.
g.,
firing
bituminous
coal
instead
of
lignite
coal).
The
EPA
considered
several
aspects
of
fuel
switching
in
evaluating
these
alternatives.
These
aspects
included
whether
switching
fuels
would
actually
achieve
lower
HAP
emissions,
whether
fuel
switching
could
be
technically
achieved
considering
the
process
design
characteristics
of
the
units,
and
the
availability
of
various
types
of
fuel.

For
both
coal­
fired
and
oil­
fired
units,
the
first
alternative
considered
was
switching
to
natural
gas.
Based
on
all
data
available
to
EPA,
and
as
was
published
in
the
Utility
RTC,
HAP
emissions
from
the
burning
of
natural
gas
are
significantly
lower
than
that
of
either
coal
or
oil
as
a
fuel.
Although
some
coal­
fired
units
utilize
natural
gas
as
a
start­
up
fuel,
the
design
and
configuration
of
the
furnace
unit
would
not
support
the
burning
of
natural
gas
easily.
The
primary
burner
configuration
of
coal
fired
units
is
designed
to
accommodate
the
large
coal
loads
common
for
production
units.
The
burner
configuration
for
the
start­
up
warming
using
natural
gas
would
only
accommodate
small
fuel
loads
and
is
not
sufficient
for
production.
Some
oil­
fired
units
burn
natural
gas
instead
of
oil
on
a
seasonal
basis,
a
practice
which
is
primarily
economically
driven
and
based
on
the
availability
and
price
of
oil.
In
most
cases,
the
design
characteristics
of
the
primary
burner
configuration
of
the
oil­
fired
units
would
allow
use
of
natural
gas
as
a
primary
fuel
The
major
limiting
factor
with
regard
to
mandating
use
of
natural
gas
instead
of
coal
or
oil
is
the
availability
of
natural
gas
for
a
given
unit.
Natural
gas
pipelines
are
not
available
in
all
regions
of
19
the
United
States.
In
addition,
even
where
pipelines
provide
access
to
natural
gas,
supplies
of
natural
gas
may
not
be
available
in
adequate
quantities
for
utilities
to
maintain
capacity
production
when
necessary.
For
example,
it
is
common
practice
in
large
metropolitan
areas
during
winter
months
(
or
periods
of
peak
demand)
to
prioritize
natural
gas
usage
for
residential
areas
before
industrial
areas.
Requiring
an
EPA­
regulated
utility
unit
to
switch
to
natural
gas
would
place
an
even
greater
strain
on
natural
gas
resources,
and,
in
some
circumstances,
the
change
would
interfere
with
ability
to
run
at
full
capacity.
For
these
reasons,
EPA
decided
that
mandating
switching
to
natural
gas
is
not
an
appropriate
alternative
for
a
MACT
control
strategy
for
existing
coal­
or
oil­
fired
units.

Another
alternative
in
fuel
substitution
would
involve
the
use
of
a
better
(
e.
g.,

lower­
Hg­
containing)
seam
of
coal
within
a
subcategory,
or
switching
between
subcategories
for
coal­
fired
units,
or
switching
from
one
type
of
oil
to
another
(
i.
e.,
residual
oil
to
distillate
oil).

The
issue
related
to
switching
between
coals
involves
the
disparity
of
HAP
constituents
in
different
seams
of
coal
and
the
disparity
of
HAP
emissions
from
different
seams
of
coal.
The
data
indicate
that,
although
one
seam
may
have
less
Hg
than
another,
it
may
be
higher
in
another
HAP.

Further,
as
discussed
previously,
the
amount
of
Hg
in
coal
is
not
the
only
factor
influencing
its
control.
The
speciation
of
Hg
in
the
flue
gas
is
another
characteristic
that
differs
between
seams
of
coal.
The
data
show
that
although
a
coal
may
have
a
lower
Hg
loading
in
the
coal,
the
Hg
emissions
may
be
more
difficult
to
control
if
that
seam
of
coal
tends
to
speciate
Hg
to
an
elemental
form.
The
EPA
reviewed
coal
data
from
the
EU/
ICE
coal
content
and
found
a
wide
range
of
HAP
constituents
in
the
coal;
however,
the
data
does
not
support
identification
of
the
"
best"
seam,
or
rank,
of
coal
on
which
to
base
such
a
requirement.
Further,
the
HAP
constituent
loading
of
different
coal
ranks
and/
or
seams
of
coal
tends
to
be
similar
for
coals
from
the
same
region,
although
that
was
not
universal.
Therefore,
even
if
a
"
better/
best"
seam
or
rank
of
coal
could
be
identified,
changing
to
a
specific
or
different
rank
or
seam
of
coal
would
essentially
determine
the
area
or
even
mine
from
which
the
coal
could
be
produced.
The
fuel
substitution
issue
then
becomes
dependent
on
the
regional
differences
in
coal
characteristics
and
the
subsequent
feasibility
of
placing
a
burden
on
units
that
are
located
further
from
the
"
better/
best"

seams,
and
even
more
importantly,
the
extent
of
the
coal
deposits
and
the
ongoing
availability
of
that
particular
coal.
The
EPA
believes
that
the
intent
of
the
Act
was
to
develop
standards
that
20
were
consistent
across
the
industry
and
avoid
actions
that
create
regional
disparities
between
units
or
place
an
unreasonable
burden
on
any
local
natural
resource.

Another
perceived
use
of
alternate
ranks
or
seams
of
coal
could
be
to
use
clean
coal.
The
term
"
clean
coal"
generally
refers
to
a
fuel
that
is
lower
in
sulfur
content.
Data
gathered
by
EPA
indicate
that
within
specific
coal
ranks,
the
HAP
content
can
vary
significantly
and
that
lower
sulfur
content
does
not
necessarily
mean
lower
HAP
content.
In
some
cases,
it
was
found
that
low­
sulfur­
coal
may
actually
result
in
an
increase
in
emissions
of
some
HAP,
including
Hg.
In
addition,
EPA
has
determined
(
as
stated
earlier)
that
the
existing
utility
units
were
designed
based
on
the
availability
of
certain
coal
ranks
and
found
that,
in
some
instances,
the
units
were
actually
built
co­
located
with
a
particular
coal
source.
(
i.
e.,
many
lignite­
fired
units).

The
EPA
has
determined
that
coal
ranking
and
subsequent
system
design
characteristics
are
issues
that
are
formidable
enough
to
warrant
subcategorization
within
the
coal­
fired
units.
A
unit
may
require
extensive
changes
to
the
fuel
handling
and
feeding
system
(
e.
g.,
a
stoker
using
bituminous
coal
as
fuel
would
need
to
be
redesigned
[
i.
e.,
retooled])
in
order
to
burn
a
different
rank
of
coal.
Additionally,
existing
burners
and
combustion
chamber
designs
are
generally
not
capable
of
handling
different
fuel
types
and
generally
cannot
accommodate
increases
or
decreases
in
the
fuel
volume
and
shape.
Design
changes
to
allow
different
fuel
use
may,
in
some
cases,

reduce
the
capacity
and
efficiency
of
the
unit.
Reduced
efficiency
may
result
in
less
complete
combustion
and,
thus,
an
increase
in
organic
HAP
emissions.

4.1.2
Process
changes.
Process
changes
would
be
considered
a
pollution
prevention
alternative
when
a
change
to
the
process
that
emits
the
HAP
could
be
made
to
reduce
or
eliminate
the
HAP
emissions.
Process
changes
for
the
electric
utility
process
might
include
changes
to
furnace
or
boiler
design;
changes
in
fuel
storage;
changes
to
handling
and
feeding
systems;
or
changes
to
burner
or
component
configuration.
The
HAP
of
concern
for
this
standard
include
Hg
and
from
coal­
fired
units
and
Ni
from
oil­
fired
units.
The
EPA
found
that
both
Hg
and
Ni
emissions
are
primarily
a
result
of
the
combustion
process
itself
and
that
the
loading
and
type
of
HAP
emissions
are
more
dependent
upon
the
composition
of
the
fuel
and,
to
a
lesser
extent,
the
combustion
process.
Consequently,
process
changes
(
i.
e.,
changes
to
unit
design,
configuration,

or
operation)
would
have
very
little
effect
in
reducing
these
type
HAP
emissions.
Further,
EPA
did
not
identify
any
process
changes
that
would
have
an
effect
on
reducing
Hg
or
Ni
emissions
21
from
the
combustion
process.
Therefore,
EPA
determined
that
process
changes
would
not
be
an
appropriate
criteria
for
identifying
the
MACT
floor
level
of
control
for
existing
or
new
coal­
or
oil­
fired
units.

4.1.3
Work
Practices.
Work
practices
for
combustion
sources
are
those
practices
that
would
promote
and
support
efficient
combustion
and
are
also
known
as
"
Good
Combustion
Practices"
(
GCP)
in
the
industry.
Good
combustion
practices
are
dependent
on
the
specific
type
of
equipment
utilized
and
fuel
input
to
the
combustion
device.
Operations­
based
GCP
include
documented
operating
procedures,
operating
logs/
record
keeping,
operator
training,
documented
maintenance,
inspection
and
overhaul
procedures.
Good
combustion
practices
with
regard
to
fuels
include
fuel
quality
(
analysis),
and
fuel
handling.
The
EPA's
research
was
unable
to
identify
any
uniform
requirements
or
set
of
work
practices
that
would
meaningfully
reflect
the
use
of
GCP
or
that
could
be
meaningfully
implemented
across
any
subcategory
of
units,
particularly
with
regard
to
Hg
or
Ni
emissions.

In
general,
electric
utility
units
are
designed
for
efficient
combustion.
Facilities
have
an
economic
incentive
to
ensure
that
fuel
is
not
wasted
and
that
the
combustion
device
operates
properly
and
is
appropriately
maintained.
Therefore,
EPA
decided
that
establishing
combustion
practice
requirements
as
a
part
of
the
MACT
floor
for
existing
or
new
coal­
or
oil­
fired
units
would
not
be
necessary
or
useful.

4.2
Regulatory
Approach
The
EPA's
policy
for
MACT
standard
development
is
to
allow
as
much
flexibility
as
possible
for
the
regulated
industry
to
develop
and
implement
new
and
more
effective
control
technologies.
Therefore,
EPA
strategy
remains
focused
on
developing
standards
that
provide
a
target
for
achievement
rather
than
technology­
based
requirements,
particularly
where
existing
technologies
are
not
proven
as
effective
in
addressing
the
HAP
of
concern.
Accordingly,
EPA
decided
to
address
MACT
development
for
Hg
from
coal­
fired
units
and
Ni
from
oil­
fired
units
using
an
emission
limitation­
based
approach,
as
opposed
to
a
control­
equipment­
based
approach.

The
selection
of
emission
limitations
as
the
format
for
the
standard
provides
flexibility
for
the
regulated
community
in
that
a
regulated
source
may
choose
any
control
technology
or
technique
to
meet
the
emission
limit
that
suits
the
unit
or
units,
rather
than
requiring
each
unit
to
use
a
prescribed
method
that
may
not
be
appropriate
or
most
cost­
effective.
This
flexibility
is
22
particularly
relevant
for
coal­
fired
electric
utility
steam
generating
units
due
to
the
potential
for
widely
varying
emission
profiles
and
the
need
for
owners/
operators
to
be
able
to
employ
control
devices
that
are
best
suited
for
their
particular
emission
characteristics.

In
order
to
develop
an
emission
limitation
for
Hg
and
Ni
that
accurately
reflects
the
MACT
floor
level
of
performance,
EPA
evaluated
the
top
performing
units
based
on
the
stack
test
data
and
the
control
technologies
employed
by
the
best
performers.
The
EPA
first
examined
the
population
database
of
existing
sources
and
divided
the
units
according
to
the
subcategorization
scheme
described
above;
first
by
coal­
and
oil­
fired,
then
by
the
five
subcategories
of
coal­
fired
units.
The
EPA
then
examined
the
stack
test
data
to
determine
the
individual
numerical
mean
of
the
stack
test
results
and
ranked
the
units
from
lowest
to
highest
within
each
subcategory
for
each
regulated
HAP
(
or
surrogate).
The
EPA
then
determined
which
units
represented
the
top
12
percent
(
or
equivalent)
of
the
units
for
which
EPA
had
test
data
for
each
subcategory
(
based
on
the
lowest
stack
test
mean
emission
rate).
The
EPA
then
evaluated
those
units
to
determine
what
criteria
could
be
credited
with
the
better
performance
and
how
that
could
be
translated
into
the
MACT
floor
level
of
performance.
The
sections
below
describe
the
evaluation
of
the
better
performing
units
for
purposes
of
deciding
how
the
MACT
standard
should
be
developed
for
Hg
from
coal­
fired
units
and
Ni
from
oil­
fired
units.

4.3
Control
Technology
Performance
Analysis
The
MACT
floor
determination
must
be
based
on
demonstrated
performance.
The
first
test
for
EPA
to
determine
a
basis
for
performance
is
to
determine
what
control
technology
is
commonly
used
and
is
effective
in
controlling
the
pollutants
of
concern.
In
this
case,
EPA
used
existing
industry
information
and
test
data
from
EU/
ICE
as
well
as
results
and
findings
of
the
Utility
RTC
to
evaluate
control
technology
performance
as
it
relates
to
Hg
and
Ni
and
its
potential
use
as
MACT
floor
level
of
control.

4.3.1
Control
equipment
performance
with
regard
to
Hg.
The
EPA
initiated
the
evaluation
of
the
units
within
each
subcategory
by
ranking
them
from
lowest
to
highest
based
on
emission
rates
representing
the
outlet
Hg
concentration
of
the
stack
tests.
The
better
performing
units
were
controlled
by
either
FF
or
ESP
units,
with
FF
being
the
predominant
technology
used
in
the
top
performing
units.
Evaluation
of
the
test
data
also
indicated
that
FF
and
ESP
technology
were
also
used
at
some
of
the
worst
performing
units.
The
effectiveness
of
the
FF,
ESP,
and
23
other
technology
used
was
inconsistent,
even
within
a
subcategory
of
coal.
Further,
the
evaluation
of
the
test
report
data
indicated
that
no
specific
control
technology
or
combination
of
technologies
could
be
credited
with
the
better
performance;
however,
the
evaluation
indicated
that
FF
technology
did
provide
a
degree
of
removal
of
Hg
and
that
ESP
units
also
provided
a
degree
of
removal,
although
to
a
less
consistent
and
lower
degree
than
FF­
equipped
units.
Over
the
last
several
years,
EPA
and
other
organizations
have
conducted
a
significant
amount
of
research
with
regard
to
control
of
Hg
from
combustion
processes.
The
outcome
of
the
research
indicates
that
FF
and
ESP
control
technologies
are
effective
for
the
control
of
Hg
in
flue
gas
streams;
however,

the
effectiveness
is
more
dependent
on
the
Hg
loading
and
Hg
speciation
in
the
flue
gas
than
on
the
control
technology
applied.
The
demonstrated
performance
of
the
units
further
supported
this
conclusion.
The
information
available
indicated
that
since
FF
and
ESP
technologies
were
designed
for
particulate
control,
Hg
presented
in
particulate­
bound
form
was
readily
addressed
by
both
technologies.

This
phenomenon
was
further
evaluated
using
the
entire
database
of
coal­
fired
units
to
determine
if
the
variations
in
the
control
device
performances
could
be
correlated
to
the
speciated
form
of
the
mercury
presented
to
the
APCD.
This
evaluation
encompassed
an
evaluation
of
existing
coal­
fired
units
from
EU/
ICE
data
that
provided
Hg
speciation
data,
Hg
in
coal
data,
and
pre­
and
post­
last­
control
unit
emissions
test
data.
The
data
indicated
that
where
Hg
was
presented
to
the
control
device
in
particulate­
bound
form,
both
control
devices
provided
a
degree
of
control,
with
FF
generally
performing
better
than
ESP.
Where
Hg
was
presented
to
the
control
device
in
an
elemental
form,
the
performance
of
the
various
control
devices
was
highly
variable.

Test
data
indicate
that
both
the
type
and
the
proportion
of
speciated
Hg
presented
to
an
APCD
are
not
consistent
across
units;
however,
as
stated
above,
the
data
do
indicate
that
PM
controls
are
reasonably
effective
where
particulate­
bound
Hg
is
present.
The
variation
of
the
proportions
of
speciated
Hg
within
the
flue
gas
between
units
provided
further
explanation
for
the
observed
removal
characteristics
for
different
units
using
the
same
control
technology.
Using
the
EU/
ICE
coal
data,
EPA
analyzed
the
Hg
speciation
of
the
different
coal
ranks
and
found
that
certain
coal
ranks
tend
to
speciate
to
a
predominantly
similar
proportion
of
speciated
forms
of
Hg,
thus
further
supporting
the
rationale
for
the
subcategorization
of
coal­
fired
units
by
coal
rank.

The
EPA
determined
that
although
variable,
FF
and
ESP
control
technologies
were
24
reasonable
and
viable
technologies
on
which
to
base
the
MACT
floor
level
of
control.
The
EPA
then
evaluated
performance
of
the
various
FF­
and
ESP­
equipped
units
to
determine
what
criteria
would
most
effectively
reflect
the
performance.
The
EPA
considered
using
the
percent
removal
efficiency
of
the
control
device,
the
percent
reduction
of
Hg
from
coal
to
emissions,
and
outlet
concentration
as
viable
criteria
to
demonstrate
performance
of
the
technology.

The
EPA
first
evaluated
percent
removal
efficiency
as
the
performance
criteria
on
which
to
base
the
floor
performance;
however,
the
use
of
the
criteria
proved
problematic.
The
EU/
ICE
Hg
data
were
based
on
stack
test
data
developed
by
testing
before
and
after
the
last
control
device
at
each
utility
unit
tested.
The
emissions
were
measured
in
mg
of
Hg
per
volume
of
test
solution
used
in
the
Ontario­
Hydro
method.
Using
the
duct
or
stack
flue­
gas
flow
volume
and
the
heat
input
to
the
unit
being
tested,
the
measured
quantity
of
Hg
was
converted
and
reported
in
units
of
lb/
TBtu.
In
reviewing
the
data,
EPA
found
that
the
inlet
measurement
showed
deficiencies
due
to
the
flow
rate
and
short
duct
runs
available
for
testing
before
the
control
device
and
that
these
values
were
suspect
as
reliable
representations
of
actual
inlet
concentrations.
The
EPA
determined
that
without
reliable
inlet
concentration
data,
calculation
of
percent
removal
efficiency
based
on
the
data
would
provide
potentially
inaccurate
removal
values.
As
a
result,

EPA
decided
that
percent
removal
efficiency
would
not
be
an
appropriate
criteria
for
MACT
floor
development
due
to
insufficient
data
being
available
to
accurately
determine
the
values.
The
EPA
did
determine,
however,
that
the
outlet
concentration
data
that
were
derived
from
the
stack
tests
were
reliable
based
on
the
method
used
and
the
fact
that
only
one
measurement
was
needed
for
the
determination
of
the
value.

The
next
approach
EPA
evaluated
was
determining
the
percent
reduction
of
Hg
demonstrated
by
the
best
performing
units
and
using
that
value
as
the
MACT
floor
performance
level.
The
percent
reduction
value
would
represent
the
amount
of
Hg
reduction
that
the
unit
accomplished
based
on
the
Hg
in
the
coal
to
the
stack
outlet
concentration.
This
approach
would
also
incorporate
EPA's
desire
to
promote,
and
give
credit
for,
coal
preparation
that
removes
Hg
before
firing
(
i.
e.,
coal
washing).
In
order
to
use
the
percent
reduction
value
as
the
criteria
for
performance,
the
operator
would
be
required
to
track
Hg
concentrations
in
the
coal
from
receipt
to
the
stack.
Tracking
and
recordkeeping
of
Hg
concentrations
in
coal
is
not
currently
conducted
in
the
industry.
Therefore,
the
issue
presented
a
logistical
concern
as
to
what
would
be
involved
25
and
how
a
such
tracking
method
could
be
uniformly
and
equitably
regulated
by
the
rule.

Therefore,
EPA
determined
that
the
tracking
of
Hg
in
the
coal
would
be
unworkable
from
the
regulatory
perspective.
Further,
EPA
concluded
that
without
the
ability
to
give
credit
for
Hg
removal
prior
to
firing,
the
percent
reduction
criteria
would
not
be
a
desirable
criteria
on
which
to
base
MACT
floor
performance.

4.3.2
Control
equipment
performance
with
regard
to
Ni.
The
EPA
examined
available
test
data
and
found
that
units
equipped
with
ESP
units
(
for
PM
control)
can
effectively
reduce
Ni.
The
controls
currently
in
use
on
electric
utility
oil­
fired
units
to
address
PM
were
installed
as
a
result
of
requirements
to
address
criteria
pollutants
under
other
regulations.
The
data
available
to
EPA
indicate
that
the
Ni
is
present
in
flue
gas
streams
in
varying
concentrations,

yet
mostly
in
particulate
form.
The
Utility
RTC
emissions
test
data
support
the
conclusion
that
the
same
control
techniques
used
to
control
the
fly­
ash
PM
will
also
indiscriminately
control
Ni
and
that
the
effective
removal
of
PM
indicates
removal
of
Ni,
for
a
given
control
device.

Therefore,
EPA
believes
that
ESP
technology
represents
the
MACT
floor
for
Ni.
The
EPA
has
determined
that
the
emission
limitation
for
the
oil­
fired
units
should
reflect
the
performance
that
would
be
expected
over
time
for
a
well
designed
and
operated
ESP
unit
PM
removal
technology.

The
EPA
determined
that
the
better
performing
units
within
the
database
were
all
equipped
with
ESP
units.

4.3.3
Conclusion.
As
a
result
of
these
evaluations,
EPA
determined
that
the
most
credible
data
element
available
that
quantified
the
better
performance
of
the
top
units
would
be
the
outlet
concentration
as
provided
in
the
stack
test
reports
(
translated
into
emission
rate).
In
order
to
use
these
data
elements
as
the
criteria
representing
the
MACT
floor
level
of
control
or
performance,
EPA
would
need
to
develop
an
emission
limitation
for
Hg
and
Ni
based
on
the
stack
test
result
values
that
would
be
representative
of
the
average
performance
of
the
top
12
percent
of
the
units
in
each
subcategory
on
an
ongoing
basis.

5.0
EMISSION
LIMITATION
DETERMINATION
FOR
EXISTING
UNITS
The
EPA
evaluated
several
format
options
for
the
limits
including
the
formats
used
for
previous
combustion
rules,
formats
representing
standard
practice
within
the
industry
with
regard
to
data
tracked
and
reported,
and
formats
suggested
by
industry
and
stakeholder
groups.

The
options
evaluated
included
emission
limitation,
percent
reduction,
and
outlet
26
concentration
formats.
The
emission
limitation
option
can
be
described
as
a
not­
to­
exceed
numerical
value
expressed
as
a
rate.
The
emission
limitation
would
be
derived
by
determining
the
mass
of
HAP
emissions
that
represents
the
average
HAP
reduction
demonstrated
by
the
top
performing
units.
The
rate
component
of
the
limitation
would
include
some
input­
or
outputbased
parameter
that
is
representative
of
the
industry.
The
percent
reduction
format
is
a
value
presented
in
the
form
of
a
percentage
that
represents
either
the
percent
reduction
of
HAP
demonstrated
by
the
top­
performing
units
based
on
the
efficiency
of
the
control
equipment
and/
or
the
use
of
mass
balance
calculations
where
control
equipment
efficiency
is
not
applicable.
Finally,

the
outlet
concentration
format
presents
a
numerical
value
in
the
form
of
a
concentration
(
mass/
volume)
that
would
be
a
not­
to­
exceed
value
and
would
be
derived
in
the
same
manner
as
the
mass
component
of
the
emission
limitation.

Where
an
emission
limitation
is
used,
EPA
must
also
determine
what
basis
will
be
used
as
the
rate
characteristic.
For
the
electric
utility
industry,
the
input­
based
characteristic
would
be
heat
or
power
input
to
the
unit
in
order
to
generate
steam.
The
output­
based
characteristic
would
be
the
amount
of
heat
or
power
(
electricity)
generated.
Finally,
EPA
must
also
consider
whether
it
is
appropriate
to
base
the
emission
limitation
on
the
gross
amount
of
heat/
power
generated
by
the
system
or
the
net
amount
of
heat/
power
that
is
available
for
sale
(
less
the
heat/
power
used
for
internal
purposes).

For
development
of
the
MACT
standard,
EPA
determined
that
an
emission
limitation
is
the
appropriate
format
to
be
used
based
on
considerations
with
regard
to
available
data,

compliance
options,
and
consistency
with
other
combustion
rules.
The
percent
reduction
option
was
not
considered
appropriate
because,
as
stated
earlier,
there
was
no
control
technology
identified
that
was
consistent
within
any
subcategory
that
could
be
used
as
the
preferred
control
technology
on
which
to
base
a
reduction
requirement.
The
EPA
also
considered
using
outlet
concentration
as
an
alternative
format;
however,
although
this
format
was
consistent
with
other
Federal
and
many
State
combustion
source
regulations
and
allowed
easy
comparison
between
requirements,
the
format
did
not
promote
pollution
prevention
and
has
become
inconsistent
with
many
of
the
newer
regulations.

5.1
Emission
Limitation
Format
An
emission
limitation
format
can
be
either
input­
based
or
output­
based
(
as
discussed
27
above).
The
use
of
an
input­
based
standard
(
lb/
TBtu)
has
several
advantages:
(
1)
it
is
consistent
with
the
majority
of
historical
Agency
electric
utility
rulemakings;
(
2)
it
would
not
need
to
be
adjusted
for
energy
requirements
for
auxiliaries
such
as
emission
control
equipment;
and
(
3)
it
does
not
need
to
take
into
account
the
baseline
efficiency
of
the
boiler/
furnace.

An
output­
based
standard
would
have
the
following
advantages:
(
1)
it
provides
incentive
for
efficiency
upgrades
(
i.
e.,
an
output­
based
standard
would
be
preferable
for
promoting
energy
efficiency
in
electric
utility
steam
generating
facilities);
(
2)
it
is
consistent
with
recent
Agency
rulemakings
(
e.
g.,
nitrogen
oxides
[
NO
x]
new
source
performance
standards
[
NSPS]
revision)
and
some
State
actions;
and
(
3)
it
would
not
cause
an
undue
compliance
burden
to
the
industry.

The
EPA
has
found
considerable
support
for
both
an
input­
based
and
an
output­
based
standard
for
emission
limits
for
electric
utility
units.
The
EPA
concludes
that
both
types
of
format
have
merit
and
has
decided
that
both
an
input­
based
and
an
output­
based
standard
would
be
appropriate
for
the
standard.

With
regard
to
cogeneration
units,
to
comply
with
an
output­
based
standard,
the
energy
content
of
the
process
steam
must
also
be
considered
in
determining
the
energy
output
when
determining
the
emission
rate.
The
EPA
has
determined
that
existing
plant
monitoring
and
energy
calculation
curves
are
available
and
can
be
easily
programmed
to
determine
the
steam's
equivalent
electrical
energy
component.
This
component
can
then
can
be
added
to
the
plant's
actual
gross
electrical
output
to
arrive
at
the
plant's
total
gross
energy
output.

The
EPA
considered
two
possible
output­
based
formats:
(
1)
mass
of
HAP
emitted
per
gross
boiler
steam
output
(
lb
HAP/
TBtu
heat
output),
and
(
2)
mass
of
HAP
emitted
per
net
energy
output
(
lb
HAP/
MWh).
An
output­
based
standard
in
lb/
MWh
gross
would
be
consistent
with
recent
Agency
rulemakings
and
some
State
actions.
The
option
of
lb
HAP/
TBtu
steam
output
accounts
only
for
boiler
efficiency,
ignores
both
the
turbine
cycle
efficiency
and
the
effects
of
energy
consumption
internal
to
the
plant,
and
provides
minimal
opportunities
for
promoting
energy
efficiency
at
the
units.
The
EPA
has
found
that
the
second
output­
based
format
option
of
lb
HAP/
MWh
is
preferable
as
it
accounts
for
all
aspects
of
efficiency
and
provides
opportunity
for
promoting
energy
efficiency
for
the
units.

The
format
of
lb/
MWh
can
be
measured
in
two
ways:
net
and
gross
energy
output.
The
net
plant
energy
output
provides
the
owners/
operators
with
all
possible
opportunities
for
28
promoting
energy
efficiency
and
can
easily
accommodate
both
electrical
and
thermal
(
process
steam)
outputs.
The
disadvantage
of
a
net
plant
energy
output
is
that
implementation
could
require
significant
and
costly
additional
monitoring
and
reporting
systems
because
the
energy
output
that
is
used
for
internal
components
(
and
not
sent
to
the
grid)
cannot
be
accounted
for
by
simply
installing
another
meter.
The
gross
plant
energy
output,
on
the
other
hand,
represents
the
energy
generated
before
any
internal
energy
consumption
and
losses
are
considered.
Standards
based
on
this
format
do
not
have
the
disadvantages
of
the
net­
based
format
mentioned
above.

Based
on
this
analysis,
the
format
based
on
mass
of
HAP
emissions
per
gross
plant
energy
output
is
most
desirable.
Because
electrical
output
at
all
power
plants
is
typically
measured
directly
in
MWe,
a
format
in
"
lb/
MWh
gross"
is
most
appropriate.

Because
all
data
provided
to
EPA
throughout
the
development
of
the
standard
were
in
the
format
of
lb/
TBtu
heat
input,
EPA
chose
to
apply
a
conversion
factor
to
convert
the
input­
based
emission
limitation
to
the
output­
based
HAP
limitations.
The
conversion
factor
was
based
on
the
baseline
net
efficiency
of
the
unit.
The
efficiency
of
electric
utility
steam
generating
unit
is
usually
expressed
in
terms
of
heat
rate,
where
efficiency
of
a
steam
generating
plant
is
referred
to
as
net
efficiency.
The
EPA
believes
that
an
output­
based
MACT
emission
limitation
format
of
lb
HAP/
MWh
gross
is
appropriate
and
that
the
net
efficiency
value
can
be
used
to
calculate
the
output­
based
emission
limit.
Most
existing
electric
utility
steam
generating
units
fall
in
the
range
of
24
to
35
percent
efficiency.
12
The
EPA
therefore
decided
to
use
32
percent
as
the
baseline
efficiency
for
existing
coal­
and
oil­
fired
units.
Because
new
coal­
and
oil­
fired
units
are
assumed
to
be
built
for
maximum
efficiency,
EPA
believes
it
was
appropriate
to
apply
the
35
percent
efficiency
in
conversion
of
the
new
unit
emission
limitations
to
the
output­
based
limitations.
The
conversion
factors
used
were:

°
Conversion
factor
for
mass/
1012
Btu
to
mass/
MWh
(
32%
combustion
efficiency;
the
mass
can
be
either
Hg
or
Ni)

(
TBtu/
1,000,000,000,000
Btu)
*
(
3.414
Btu/
Wh)
*
(
1,000,000
Wh/
MWh)
*
(
1/
0.32)
=

10.7
x
10­
6
TBtu/
MWh
°
Conversion
factor
for
mass/
1012
Btu
to
mass/
MWh
(
35%
combustion
efficiency;
the
mass
can
be
either
Hg
or
Ni)
29
(
TBtu/
1,000,000,000,000
Btu)
*
(
3.414
Btu/
Wh)
*
(
1,000,000
Wh/
MWh)
*
(
1/
0.35)
=

9.8
x
10­
6
TBtu/
MWh
5.2
Variability
Issues
5.2.1
General
discussion
on
variability
in
data.
Although
EPA
is
confident
that
the
data
available
are
representative
of
the
industry,
it
is
evident
that
the
test
report
data
exhibit
a
significant
degree
of
variability,
even
within
a
given
subcategory.
The
EPA
decided
it
was
necessary
to
develop
a
methodology
to
address
the
multiple
sources
of
the
observed
variability
in
order
to
assure
that
an
emission
limitation
value
could
be
derived
that
would
be
achievable.
The
origins
of
variability
and
approaches
available
for
addressing
the
apparent
variability
found
in
the
test
data
are
described
below.

5.2.1.1
Origins
of
variability
in
the
data.
Variability
is
inherent
whenever
measurements
are
made
or
whenever
mechanical
processes
operate.
The
variability
in
the
emission
test
data
may
arise
from
one
or
more
of
the
following
areas:
(
1)
the
emission
test
method(
s);
(
2)
the
analytical
method(
s);
(
3)
the
design
of
the
unit
and
control
device(
s);
(
4)
the
operation
of
the
unit
and
control
device(
s);
and
(
5)
the
amount
of
the
constituent
being
tested
in
the
fuel.

Test
and
analytical
method
variability
can
be
quantified
by
statistical
analysis
of
the
results
of
a
series
of
tests.
The
results
can
be
analyzed
to
establish
confidence
intervals
within
which
the
true
value
of
a
test
result
is
presumed
to
lie.
Confidence
intervals
can
be
estimated
for
multiple­
run
series
of
tests
based
on
the
differences
found
from
one
test
run
to
the
next,
with
only
the
upper
confidence
interval
having
meaning
(
signifying
the
chance
of
the
standard
being
exceeded).

When
testing
is
done
at
more
than
one
unit,
similar
confidence
intervals
can
be
established
to
account
for
the
variability
from
unit
to
unit.
One
can
combine
the
test­
to­
test
and
unit­
to­
unit
variability
into
a
single
factor
that
can
be
applied
to
reported
test
values
to
give
an
upper
limit
for
the
likely
true
value.
One
can
also
estimate
the
combined
factor
for
any
desired
confidence
level.

Testing
for
a
short
time
may
not
reveal
the
range
of
emissions
that
would
be
found
over
extended
time
periods.
Normal
changes
in
operating
conditions
or
in
fuel
characteristics
may
affect
emission
levels.
For
example,
an
increase
in
the
Hg
content
of
the
fuel
being
fired
in
a
unit
30
may
tend
to
increase
the
Hg
emission
rate
from
the
associated
stack.
Mercury
emissions
rates
may
also
change
with
unit
loads.
As
load
changes,
so
does
gas
flow
rate
through
APCD
downstream
from
the
unit.
Changes
in
gas
flow
rate
may
affect
APCD
effectiveness.

5.2.1.2
Available
methods
to
address
and
incorporate
variability.
Variability
may
be
addressed
in
a
number
of
ways,
depending
on
the
circumstances
existing
within
the
source
category.
For
example,
different
test
run
results
can
be
analyzed
statistically
to
arrive
at
an
upper
limit
that
represents
the
highest
likely
value
for
each
test
to
be
used
in
setting
emission
limits.
The
poorest­
performing
(
worst­
case)
unit
in
the
top
12
percent
of
each
subcategory
can
be
reviewed
to
determine
the
causes
of
poor
performance
with
a
factor
then
assigned
that
can
be
applied
to
each
of
the
test
runs.
These
offsets
would
give
emission
values
that
would
not
likely
be
exceeded
over
long
time
periods.
Considering
only
control
devices
used
by
sources
in
the
top
12
percent,

control
device
performance
can
also
be
examined
to
determine
likely
emission
reductions
for
different
devices
operating
on
different
units
firing
different
fuels.
The
range
in
emission
reductions
could
be
used
to
set
upper
limits
of
expected
control
performance;
then
these
limits
could
be
used,
as
above,
to
set
emission
limitations
for
each
subcategory.
Correlations
between
constituents
of
concern
and
other,
perhaps
more
easily
measured,
constituents
can
be
used
to
develop
algorithms
that
incorporate
variability.

The
EPA
found
that
there
are
two
fundamentally
different
approaches
to
incorporating
variability
into
a
rule:
(
1)
including
variability
in
the
MACT
floor
calculation,
or
(
2)
including
variability
in
the
compliance
method.
Addressing
variability
in
the
MACT
floor
calculation
requires
that
all
of
the
origins
of
variability
be
assessed
and
quantified
into
factors
that
can
be
applied
into
the
emission
limitation
calculations
for
each
subcategory's
floor.
Each
unit
used
for
floor
calculations
is
assumed
to
operate
such
that
its
measured
emission
rate
is
increased
by
the
amount
of
variability
found
from
statistical
analysis,
worst­
case
analysis,
and
control
device
performance
analysis.
Each
unit
in
the
top
12
percent
of
its
subcategory
would
be
adjusted
to
reflect
the
uncertainty
associated
with
the
various
origins
of
variability,
and
the
average
emission
rate
for
these
units
would
be
used
as
the
floor
emission
limitation.

Addressing
variability
in
the
compliance
method
would
involve
allowing
an
averaging
time
for
compliance
that
would
accommodate
variations
in
pollutant
emissions
over
time.
For
example,
averaging
over
a
month
or
a
year
of
data
will
provide
opportunity
for
variations
in
the
31
amount
of
a
constituent
in
the
fuel
to
be
accommodated
without
exceeding
the
emission
limitation.

In
trying
to
address
the
apparent
sources
of
variability
in
the
emissions
test
data,
EPA
tried
to
obtain
data
that
reflected
as
many
different
plant
configurations
as
would
be
found
in
the
entire
industry
profile
and
conducted
tests
at
units
believed
to
be
representative
of
those
within
the
source
category.
The
tests
and
measurements,
typically
a
three­
run
series
of
manual
samples
taken
over
1
or
2
days
of
testing,
are
limited
by
the
emission
test
method's
accuracy
and
precision,
by
the
short
duration
of
the
test,
and
by
differences
from
one
run
to
the
next
and
one
unit
to
the
next.
Based
on
these
limitations
on
the
test
data,
EPA
has
decided
to
use
both
of
the
approaches
described
above
for
addressing
test
data
variability.

5.2.1
Strategy
to
address
variability
for
Hg.
Studies
available
to
EPA
indicated
that
the
variability
of
Hg
emissions
from
coal­
fired
units,
both
instantaneous
and
over
time,
is
significantly
influenced
by
the
variability
in
the
chemical
composition
and
properties
of
the
coal
as
burned
(
i.
e.,
differences
in
Hg
content,
chlorine
content,
and
heat
content
of
coal).
The
differing
physical
and
chemical
properties
of
Hg­
containing
compounds
in
the
flue
gas
result
in
significant
differences
in
the
feasibility
and
effectiveness
of
controls
for
removing
the
compounds
from
flue
gas.
Thus,
which
Hg
compounds
are
present
in
the
flue
gas
impacts
the
amount
of
Hg
that
will
be
captured
by
control
devices
and
how
much
Hg
will
be
released
in
stack
emissions.
The
studies
indicated
that
the
chlorine
content
of
the
coal
has
a
significant
impact
on
which
Hg
compounds
are
contained
in
the
flue
gas
stream
and,
even
more
importantly,
can
be
used
as
a
key
indicator
of
the
type
of
Hg
compound
that
will
be
present
in
flue
gas.
The
EPA
found
that,
when
combined
with
other
relevant
data
such
as
coal
Hg
content,
the
chlorine
content
of
coal
can
be
used
to
predict
Hg
emissions.

The
data
results
from
the
multivariable
study11
lend
support
to
the
significance
of
chlorine
content
of
coal
to
Hg
emissions
controllability.
The
higher
the
chlorine:
mercury
ratio,
the
more
likely
the
formation
of
mercuric
chloride
(
ionic
or
oxidized
Hg)
that
is
more
readily
captured
by
existing
control
devices.
This
chlorine:
mercury
ratio
is
independent
of
the
coal
rank
as
an
indicator
of
Hg
controllability.
In
sum,
the
coal
chlorine
content
is
one
of
the
primary
determinants
of
which
Hg­
containing
compounds
will
be
present,
and
in
what
amounts,
in
the
flue
gas
of
an
individual
utility
unit.
32
The
EPA
determined
that
the
stack
tests
in
the
EU/
ICE
database
alone
are
insufficient
to
estimate
the
effect
of
fuel
variability
over
time
on
the
emissions
of
the
best
performing
facilities.

However,
the
EU/
ICE
database
contains
extensive
data
on
variation
in
coal
composition
recorded
over
the
course
of
a
year.
The
EPA
developed
a
methodology
to
link
fuel
composition
data
to
Hg
emissions
in
order
have
a
better
estimate
of
Hg
emissions,
and
subsequently,
the
controllability
of
the
emissions
over
time.
The
methodology
is
described
below.

The
units
in
each
of
the
five
subcategories
were
sorted
in
ascending
order
of
stack­
tested
Hg
emission
factors,
measured
in
units
of
lb/
TBtu
(
as
adjusted
by
a
method
that
normalizes
Hg
emissions
to
coal
heat
content
[
F­
factor
Adjustment]).
Accordingly,
the
top
performing
units
of
each
subcategory
were
selected
for
further
analysis.

To
link
fuel
composition
data
to
Hg
emissions
data,
correlation
equations
were
developed
to
represent
the
relationship
between
Hg
removal
fraction
and
chlorine
concentration
for
each
of
the
control
configurations
used
by
the
best
performing
units.
The
steps
used
to
develop
these
correlation
equations
are
set
forth
below.

The
control
configuration
of
each
of
the
best
performing
units
was
identified.
The
Hg
removal
fraction
and
test
coal
chlorine
concentrations
were
obtained
from
the
EU/
ICE
database
for
each
of
the
units
in
the
database
that
have
one
of
the
identified
control
configurations.
Finally,

a
correlation
equation
was
derived
for
each
identified
control
configuration
by
fitting
the
following
mathematical
expression
to
the
Hg
removal
fractions
and
corresponding
chlorine
concentrations
obtained
from
the
EU/
ICE
stack
test
database.

In
the
selection
of
the
format
of
the
correlation
equation,
care
was
taken
that
the
mathematical
expression
accurately
reflected
the
physical
and
chemical
process
by
which
chlorine
contributes
to
the
controllability
of
stack
Hg
emissions.
The
correlation
equation
is
based
on
the
assumption
that
the
rate
of
conversion
of
Hg
to
mercury
chloride
is
proportional
to
the
chlorine
concentration
in
the
coal,
irrespective
of
coal
rank.
With
this
expression,
the
maximum
removal
fraction
is
limited
to
1,
because
the
exponent
term
is
always
nonnegative,
regardless
of
the
chlorine
concentration.
This
corresponds
to
the
real­
world
limitation
that
no
more
than
100
percent
of
the
Hg
in
flue
gas
can
be
removed
(
i.
e.,
there
cannot
be
negative
Hg
emissions).
As
the
coal
chlorine
concentration
drops
to
zero,
the
Hg
removal
fraction
does
not
of
necessity
approach
zero
because
some
Hg
removal
may
be
achieved
without
reaction
with
chlorine.
The
33
purpose
of
deriving
a
correlation
equation
for
each
control
configuration
used
by
the
top
performing
units
was
to
provide
a
numerical
means
of
predicting
the
fraction
of
Hg
removed
for
the
best
performing
sources
over
the
entire
range
of
fuel
variability
experienced
over
the
course
of
a
year.
Correlation
equations
were
derived
for
each
control
configuration,
but
were
only
used
to
predict
Hg
removal
if
they
were
found
to
have
acceptable
explanatory
power.

To
determine
whether
the
explanatory
power
of
each
correlation
equation
warranted
its
use
on
a
larger
range
of
EU/
ICE
coal
composition
data,
each
correlation
equation
was
validated
against
the
EU/
ICE
stack
test
data.
For
each
of
the
test
chlorine
concentrations
in
the
EU/
ICE
stack
test
database,
the
Hg
removal
fraction
was
calculated
by
using
the
correlation
equation
with
parameters
selected
to
give
the
best
fit
to
the
data.
A
correlation
coefficient
was
then
calculated
to
evaluate
the
accuracy
of
the
fit.

For
each
of
the
best
performing
units,
unit­
specific
coal
composition
data
for
a
one­
year
period
were
extracted
from
the
EU/
ICE
database
to
find
the
coal
heat
content,
Hg
content,
and
chlorine
content.
For
each
set
of
coal
composition
data
from
the
EU/
ICE
database,
the
controlled
Hg
emissions
were
calculated
by
multiplying
uncontrolled
Hg
emissions
by
(
1
 
Hg
removal
fraction).
For
each
of
the
best­
performing
sources,
this
process
was
repeated
for
each
set
of
measured
coal
composition
values,
yielding
a
range
of
Hg
emission
levels
for
that
unit
over
time.

The
test
coal
composition
data
from
the
EU/
ICE
database
(
heat
and
Hg
content)
was
used
to
calculate
the
uncontrolled
Hg
emission
level.
The
Hg
removal
fraction
was
calculated
in
one
of
the
following
two
ways:

(
1)
Where
the
correlation
equation
was
found
to
have
sufficient
explanatory
power,
it
was
used
to
estimate
the
Hg
removal
fraction
based
on
coal
chlorine
composition
data
from
the
EU/
ICE
database.
This
approach
accounted
for
variations
in
the
Hg,
chlorine,
and
heat
content
of
fuel.

(
2)
Where
the
correlation
equation
was
a
poor
fit,
the
Hg
removal
fraction
was
based
on
the
average
Hg
removal
fraction
observed
in
the
EU/
ICE
stack
tests
of
that
unit.
This
latter
approach
yielded
a
constant
removal
fraction
based
upon
the
source
test,
and
had
the
effect
of
reducing
the
variability
of
predicted
Hg
emissions.
Under
this
approach,
the
measured
impact
of
fuel
variability
was
limited
to
the
effect
of
variations
in
Hg
and
heat
content,
while
variations
in
chlorine
concentration
were
not
explicitly
considered.
34
For
each
of
the
best
performing
units,
the
calculated
Hg
emissions,
calculated
in
accordance
with
the
procedures
outlined
above,
were
then
sorted
from
smallest
to
largest
to
obtain
a
cumulative
frequency
distribution
(
CFD).
The
97.5th
percentile
value
of
this
distribution
(
i.
e.,
an
emission
rate
that
is
expected
to
be
exceeded
only
2.5
percent
of
the
time)
was
determined
to
represent
the
operation
of
the
unit
under
worst
conditions.

The
EPA
decided
to
account
for
unit­
to­
unit
variability
by
calculating
a
97.5
percent
upper
confidence
level
for
the
mean
by
use
of
the
t­
statistic.
This
adjustment
reflects
the
fact
that
the
top
performing
sources
in
the
data
base
do
not
represent
the
full
population
of
the
best
performing
12
percent
of
coal­
fired
utility
units
.10
Although
fuel
variability
is
a
principal
cause
of
emission
variability,
other
factors
also
play
a
role
in
contributing
to
variability
in
Hg
emissions.
Analysis
of
fuel
variability
accounts
for
some,

but
not
all,
of
the
variability
in
the
stack
testing
of
each
unit
that
comprises
the
EU/
ICE
database.

Other
drivers
of
variability
in
the
test
results,
such
as
measurement
error,
are
not
included
in
the
analysis.
Intermittent
maintenance
events,
which
themselves
can
contribute
to
short­
term
increases
in
Hg
emissions,
also
are
not
considered.
In
addition,
the
stack
testing
on
which
this
assessment
is
based
places
artificial
limitations
on
the
variability
of
its
results.
Testing
was
performed
with
plants
operating
at
full
and
constant
load
and
without
ongoing
maintenance
activities.
Actual
operation
requires
load­
following
in
addition
to
intermittent
maintenance
activities.
Insofar
as
the
methodology
discussed
herein
does
not
incorporate
these
effects,
its
results
are
likely
to
underestimate
the
reasonable
worst­
case
emissions
of
the
best
performing
facilities.
For
these
and
other
reasons,
EPA
believes
a
12­
month
rolling
averaging
period
would
be
appropriate
for
the
standard.

5.2.3
Strategy
to
address
variability
for
Ni.
The
data
used
to
determine
the
Ni
emission
limitation
consisted
of
stack
test
reports
from
the
DOE/
EIA4
effort.
These
emissions
rates
were
adjusted
for
test­
to­
test
run
variability
using
the
coefficient
of
variation
(
standard
deviation
of
the
data
set
divided
by
the
mean
of
the
data
set)
and
then
were
adjusted
for
unit­

tounit
variation
using
a
student
T­
statistic
to
derive
the
97.5
percentile
confidence
interval.

5.3
Emission
Limitation
Calculations
In
order
to
determine
the
MACT
floor
emission
limitations
for
existing
units,
EPA
examined
the
population
database
of
existing
sources.
Available
emissions
test
data
were
divided
35
according
to
the
subcategorization
scheme
described
above;
first
coal­
and
oil­
fired,
then
the
five
subcategories
of
coal­
fired
units.
The
EPA
examined
the
existing
emissions
test
data
to
determine
the
individual
numerical
average
of
the
test
results
from
the
best­
performing
12
percent
(
or
equivalent)
of
each
subcategory
for
each
regulated
HAP
(
or
surrogate).
The
EPA
then
applied
variability
factors
as
described
above
to
derive
the
MACT
floor
limits.
All
test
data
were
provided
to
EPA
in
an
input­
based
format
(
lb
Hg/
TBtu).
Therefore,
EPA
conducted
all
MACT
floor
calculations
using
the
input­
based
format
and
then
converted
the
input­
based
format
into
an
output­
based
format
(
lb
HAP/
MWh)
as
a
compliance
option,
according
to
the
approach
described
in
section
5.1
above.
Appendix
10.2
of
this
document
provides
the
detail
spreadsheets
listing
the
data
used
and
calculations
for
determination
of
the
variability
factors
and
the
emissions
limitation
values.

5.3.1
Mercury
Emission
Limitation
Calculations.
The
EPA
calculated
the
emission
limitation
for
Hg
for
the
subcategories
of
bituminous­
fired,
subbituminous­
fired,
lignite­
fired,

IGCC,
and
coal
refuse­
fired
units
as
follows.

For
bituminous­
fired
units,
EPA
had
data
from
32
units.
Because
this
subcategory
(
i.
e.,

nationwide
population)
included
more
than
30
units,
EPA
determined
that
the
top
12
percent
of
the
units
in
the
subcategory
would
be
composed
of
12
percent
of
the
number
of
units
for
which
EPA
had
data
(
i.
e,
4
units).
The
EPA
determined
the
top
four
units
from
a
ranking
of
units
based
on
their
emission
rates
from
the
stack
test
reports.
The
emission
rates
from
these
units
ranged
from
0.1062
lb/
TBtu
to
0.1316
lb/
TBtu,
with
a
mean
of
0.1180
lb/
TBtu.
After
applying
variability
as
described
above
and
rounding
to
3
significant
figures,
EPA
determined
the
input­
based
emission
limitation
to
be
1.97
lb/
TBtu.
Using
the
conversion
described
in
section
5.1
above
(
and
based
on
32
percent
net
efficiency),
the
input­
based
emission
limitation
of
1.97
lb/
TBtu
was
converted
to
21.0
x
10­
6
lb/
MWh
as
the
output­
based
emission
limitation.

For
subbituminous­
fired
units,
EPA
had
data
from
32
units.
Because
this
subcategory
(
i.
e.,
nationwide
population)
included
more
than
30
units,
EPA
determined
that
the
top
12
percent
of
the
units
in
the
subcategory
would
be
composed
of
12
percent
of
the
units
for
which
EPA
had
test
data
(
i.
e.,
4
units).
The
EPA
determined
the
top
units
from
the
ranking
of
the
units
based
on
their
emission
rates
from
the
stack
test
reports.
The
emission
rates
from
these
units
ranged
from
0.4606
lb/
TBtu
to
1.207
lb/
TBtu,
with
a
mean
of
0.7638
lb/
TBtu.
After
applying
variability
as
36
described
above
and
rounding
to
3
significant
figures,
EPA
determined
the
input­
based
emission
limitation
to
be
5.77
lb/
TBtu.
Using
the
conversion
described
in
Section
5.1
above
(
and
based
on
32
percent
net
efficiency),
the
input­
based
emission
limitation
of
5.77
lb/
TBtu
was
converted
to
61.6
x
10­
6
lb/
MWh
as
the
output­
based
emission
limitation.

For
lignite­
fired
units,
EPA
had
data
from
12
units.
Because
this
subcategory
(
i.
e.,

nationwide
population)
consisted
of
fewer
than
30
units
(
in
1999),
EPA
determined
that
the
top
performers
must
include
the
top
5
units.
The
emission
rates
from
these
units
ranged
from
3.977
lb/
TBtu
to
6.902
lb/
TBtu,
with
a
mean
of
5.032
lb/
TBtu.
After
applying
variability
as
described
above
and
rounding
to
3
significant
figures,
EPA
determined
the
input­
based
emission
limitation
to
be
9.24
lb/
TBtu.
Using
the
conversion
described
in
section
5.1
above
(
and
based
on
32
percent
net
efficiency),
the
input­
based
emission
limitation
of
9.24
lb/
TBtu
was
converted
to
98.6
x
10­
6
lb/
MWh
as
the
output­
based
emission
limitation.

For
IGCC
units,
EPA
had
data
on
two
units.
Because
this
subcategory
(
i.
e.,
nationwide
population)
included
less
than
30
units,
EPA
determined
that
all
available
units
would
be
included
and
were
ranked
based
on
their
emission
rates
from
the
stack
test
reports.
The
emission
rates
from
these
units
ranged
from
5.334
lb/
TBtu
to
5.471
lb/
TBtu,
with
a
mean
of
5.403
lb/
TBtu.
The
EPA
applied
the
variability
factors
and,
with
rounding
to
3
significant
figures,
determined
the
IGCC
input­
based
emission
limitation
to
be
18.7
lb/
TBtu.
Using
the
conversion
described
in
section
5.1
above
(
and
based
on
32
percent
net
efficiency),
the
input­
based
emission
limitation
of
18.7
lb/
TBtu
was
converted
to
200
x
10­
6
lb/
MWh
as
the
output­
based
emission
limitation.

For
coal
refuse­
fired
units,
EPA
had
data
from
two
units.
Because
this
subcategory
(
i.
e.,

nationwide
population)
included
fewer
than
30
units,
EPA
used
all
units
for
the
calculation
based
on
their
emission
rates
from
the
stack
test
reports.
The
emission
rates
from
these
units
ranged
from
0.0816
lb/
TBtu
to
0.0936
lb/
TBtu,
with
a
mean
of
0.0876
lb/
TBtu.
The
EPA
applied
the
variability
factors
as
described
above
and
with
rounding
to
3
significant
digits,
determined
the
input­
based
emission
limitation
to
be
0.385
lb/
TBtu.
Using
the
conversion
described
in
section
5.1
above
(
and
based
on
32
percent
net
efficiency),
the
input­
based
emission
limitation
of
0.385
lb/
TBtu
was
converted
to
4.11
x
10­
6
lb/
MWh
as
the
output­
based
emission
limitation.

Table
1
below
summarizes
the
emission
limitations
for
existing
coal­
fired
units.
37
TABLE
1.
Hg
EMISSION
LIMITS
FOR
EXISTING
COAL­
FIRED
UNITS
Unit
Type
Hg
(
lb/
TBtu)
Hg
(
10­
6
lb/
MWh)

Bituminous­
fired
1.97
21.0
Subbituminous­
fired
5.77
61.6
Lignite­
fired
9.24
98.6
IGCC
unit
18.7
200
Coal
refuse­
fired
0.385
4.11
The
EPA
believes
that
the
Hg
emissions
limitations
derived
above,
using
the
test
data
with
application
of
appropriate
variability,
provided
a
reasonable
estimate
of
actual
performance
of
the
MACT
floor
unit
on
an
ongoing
basis.

5.3.2
Nickel
Emission
Limitation
Calculation.
The
emission
limit
for
Ni
from
existing
oil­
fired
units
was
determined
by
analyzing
the
emissions
data
available.
The
data
were
obtained
from
the
Utility
RTC.
The
EPA
examined
available
test
data
and
found
that
ESP­
equipped
units
can
effectively
reduce
Ni.
The
Utility
RTC
emissions
test
data
support
the
conclusion
that
the
same
control
techniques
used
to
control
the
fly­
ash
PM
will
also
indiscriminately
control
Ni
and
that
the
effective
removal
of
PM
indicates
removal
of
Ni,
for
a
given
control
device.
Therefore,

EPA
believes
that
ESP
technology
represents
the
MACT
floor
for
Ni
removal.
The
EPA
has
determined
that
the
emission
limitation
for
the
oil­
fired
units
should
reflect
the
performance
that
would
be
expected
over
time
for
a
well
designed
and
operated
ESP
unit
PM
removal
technology.

The
EPA
determined
the
value
of
the
Ni
emission
limitation
by
ranking
the
stack
test
Ni
emission
rates
of
the
17
units
for
which
EPA
had
data.
The
top
12
percent
of
the
units,
or
2
units,

were
ESP­
controlled
and
the
range
of
emission
rates
was
29.97
lb/
TBtu
to
357.16
lb/
TBtu
with
a
mean
of
125.06
lb/
TBtu.
After
applying
variability
as
described
above
and
rounding
to
2
significant
figures,
EPA
determined
the
input­
based
emission
limitation
to
be
210
lb/
TBtu.
The
output­
based
Ni
emission
limitation
was
determined
to
be
0.002
lb/
MWh
after
conversion
using
32
percent
net
efficiency.
The
EPA
believes
that
these
Ni
emission
limits
are
a
reasonable
estimate
of
the
actual
performance
of
the
MACT
floor
unit
on
an
ongoing
basis.
38
6.0
EVALUATION
OF
MACT
FLOOR
PERFORMANCE
FOR
NEW
UNITS
In
order
to
develop
a
MACT
standard
for
new
coal­
and
oil­
fired
units,
EPA
used
the
same
data
described
above
for
existing
sources.
The
MACT
floor
for
new
sources
must
reflect
the
level
of
control
demonstrated
by
the
best
performing
similar
source.
Therefore,
EPA
evaluated
the
existing
data
to
determine
the
best
unit
on
which
to
base
the
emission
limitation
for
new
units.

6.1
Pollution
Prevention
Alternatives
In
developing
a
MACT
strategy
for
new
units,
EPA
considered
several
prevention
measures
as
an
alternative
to
HAP
control
technology.
These
measures
were
the
same
precombustion
techniques
evaluated
for
existing
units,
which
included
fuel
substitution,
process
changes,
and
work
practices.

The
feasibility
of
mandating
which
fossil
fuel
should
be
burned
was
evaluated
from
several
perspectives:
(
1)
mandating
"
perceived
better"
fuels
from
the
same
subcategory
(
e.
g.,
a
lower
Hg
content
bituminous
coal);
(
2)
mandating
a
fuel
from
another
subcategory
(
e.
g.,
firing
bituminous
coal
instead
of
lignite
coal);
or
(
3)
mandating
the
use
of
natural
gas.
The
EPA
recognizes
that
an
owner/
operator,
in
designing
a
new
unit,
would
be
able
to
choose
a
perceived
better
coal
rank
(
between
subcategories)
or
a
perceived
better
coal
seam
within
a
rank
(
within
the
subcategory)

based
on
known
issues
of
HAP
and
other
pollutant
control
and
would
be
able
design
the
new
unit
to
that
fuel's
characteristics.
However,
the
economics
of
fuel
availability
would
still
be
a
determining
factor
as
to
what
fuel
was
chosen,
particularly
with
regard
to
new
units
co­
located
with
existing
units.

With
regard
to
a
possible
EPA
requirement
for
new
sources
to
burn
natural
gas,
EPA
believes
that
availability
and
economics
again
would
determine
whether
a
source
would
chose
to
burn
natural
gas
and
that
such
a
requirement
would
be
unduly
restrictive
given
the
owner/
operator's
inability
to
control
access
to,
or
availability
of,
natural
gas.
For
these
reasons,

EPA
decided
that
mandated
fuel
type
is
not
an
appropriate
criterion
for
identifying
the
MACT
level
of
control
for
new
coal­
fired
units.

With
regard
to
process
design
alternatives
and
GCP,
EPA
believes,
as
discussed
above
in
section
4.1
for
existing
sources,
the
industry
has
a
strong
economic
incentive
to
pursue
improvement
in
combustion
and
plant
efficiencies
and
that
the
trends
in
design
and
technology
39
development
will
continue
in
the
direction
of
improvement
in
efficiencies
such
that
imposition
of
regulatory
incentives
based
on
the
existing
knowledge
base
would
be
not
only
unnecessary
but
potentially
restrictive.
Therefore,
as
with
existing
units,
EPA
determined
that
pre­
combustion
techniques
were
not
a
viable
regulatory
strategy
for
the
MACT
standard
for
new
coal­
or
oil­
fired
units.

6.2
Control
Technology
Performance
Evaluations
Once
EPA
determined
that
pollution
prevention
alternatives
would
not
be
appropriate
for
the
new
coal­
or
oil­
fired
MACT
development,
EPA
then
evaluated
the
options
to
develop
the
standard
for
new
units
based
on
the
control
technology
used
by
the
top
performing
unit
(
i.
e.,

equipment
based),
on
the
level
of
emission
reduction
that
the
top
unit
in
each
subcategory
demonstrated,
or
a
combination
of
both.

With
regard
to
Hg
and
Ni
emissions
from
new
units,
EPA
believes
that
the
character
and
levels
of
Hg
and
Ni
emitted
by
new
coal­
and
oil­
fired
units
will
be
similar
to
those
emitted
by
existing
coal­
and
oil­
fired
units
because
the
source
of
these
pollutants
is
primarily
the
fuel
and,
to
a
limited
extent,
the
combustion
process.
The
EPA
has
no
data
or
information
that
indicated
that
these
characteristics
would
change
in
the
future,
particularly
because
EPA
anticipates
the
use
of
primarily
the
same
fossil
fuel
sources
for
new
units
as
are
being
used
for
existing
units.

The
EPA
is
aware
that
the
industry
has
the
ability
during
the
designing
of
new
units
to
choose
a
fuel
that
would
minimize
Hg
or
Ni
emissions
production
and
recognizes
that
the
MACT
standard
for
new
units
should,
to
the
extent
possible,
encourage
the
industry
in
that
direction.

The
type,
grades,
and
ranks
of
fossil
fuel
available
for
future
use
in
new
units
will
not
likely
change,
and
the
availability
and
economics
of
the
fuel
choice
for
these
units
will
likely
still
be
a
dominating
factor
in
the
design
of
new
units.
However,
future
technology
may
allow
for
better
efficiencies
in
the
units
and,
potentially,
the
use
of
a
wider
range
of
fossil
fuels
for
a
given
locale
or
region.
The
EPA
used
the
same
data
available
for
existing
units
which
provided
an
evaluation
of
the
Hg
control
performance
of
various
emission
control
technologies
that
are
either
currently
in
use
on
coal­
fired
units
(
designed
for
pollutants
other
than
Hg)
or
that
could
be
applied
to
such
units
for
Hg
control.
According
to
the
data
available
to
EPA,
none
of
the
existing
control
systems
were
specifically
designed
to
remove
Hg
or
Ni;
however,
most
of
the
controls
removed
these
pollutants
to
some
degree.
In
reviewing
these
data
with
regard
to
new
units,
EPA
found
no
40
control
technology
to
be
available
for
specifically
addressing
Hg
for
either
coal­
or
Ni
for
oil­
fired
units,
however,
existing
units
were
achieving
a
level
of
control
using
the
current
PM
removal
technologies
such
as
FF
and
ESP
units.

7.0
EMISSION
LIMITATION
DETERMINATION
FOR
NEW
UNITS
As
was
discussed
in
MACT
floor
development
for
existing
sources,
EPA
is
confident
that
the
test
data
available
were
representative
of
the
industry;
however,
EPA
did
believe
that
some
adjustments
were
justified
in
light
of
the
variability
in
test
method
and
in
HAP­
in­
fuel
that
was
discussed
previously
with
regard
to
existing
units.
Although
it
was
necessary
to
address
the
variability
issues,
the
use
of
one
data
set
(
i.
e.,
the
best
unit
vs.
a
number
of
top
units)
negated
the
applicability
of
the
unit­
to­
unit
variability
issue.
Otherwise,
the
variability
issues
were
addressed
in
the
same
manner
as
was
discussed
above
for
existing
units.

The
MACT
for
new
units
is
based
on
the
emission
level
achieved
by
the
best­
performing
similar
source
in
each
subcategory.
In
order
to
develop
an
emission
limitation
for
new
coal­
and
oil­
fired
units,
EPA
ranked
the
existing
coal­
and
oil­
fired
units
from
lowest
to
highest
within
each
subcategory
based
on
Hg
emission
rates
from
the
stack
test
data.
The
EPA
then
selected
the
numerical
performance
value
from
the
best­
performing
unit
(
or
equivalent).
Because
test
data
were
provided
to
EPA
based
on
an
input­
based
format
(
lb/
TBtu),
EPA
conducted
the
emission
limitation
calculations
using
the
input­
based
format
and
then
converted
the
input­
based
format
into
an
output­
based
format
(
lb/
MWh)
according
to
the
approach
described
in
section
5.1
above.

7.1
Emission
Limitation
format
One
of
EPA's
major
policy
strategies
is
to
encourage
energy
efficiency
and
pollution
prevention
in
the
development
of
new
standards.
Therefore,
EPA
determined
that
the
format
for
the
new
units
under
the
standard
should
be
based
solely
on
an
output­
based
format
(
lb/
MWh)
in
order
to
encourage
and
reward
efficiency
in
the
operation
for
new
units.

7.2
Variability
Issues
7.2.1
General.
Because
the
emission
limitations
for
new
units
are
based
on
the
same
data
as
existing
units,
the
same
variability
issues
as
described
in
section
5.2
above
were
of
concern.

The
following
sections
describe
how
EPA
addressed
the
variability
for
development
of
emission
limitations
for
new
units.
41
7.2.2
Strategy
for
addressing
variability
for
Hg.
The
evaluation
of
the
data
(
see
section
5.2.2
above)
for
existing
units
provided
a
ranking
of
data
that
had
been
adjusted
for
fuel
and
test
method
variability.
The
EPA
decided
that
the
rate
of
the
best
performing
unit
from
this
ranking
was
the
appropriate
value
for
the
new
unit.

7.2.3
Strategy
for
addressing
variability
for
Ni.
The
variability
and
uncertainty
were
addressed
in
the
same
manner
as
for
existing
oil­
fired
units.
The
data
from
existing
units
was
evaluated
and
appropriate
test
method
variability
was
applied
using
the
coefficient
of
variation
method
described
above
in
section
5.2.3
above.
The
best­
performing
unit
was
chosen
and
that
value
was
used
for
the
emission
limitation.

7.3
Emission
Limitations
Calculations
for
New
Units
The
emission
limit
for
Hg
emissions
from
new
coal­
fired
units
was
determined
by
analyzing
the
available
Hg
emissions
data
in
each
subcategory.
The
data
were
obtained
from
the
EU/
ICE
and
included
data
for
Hg
emissions
and
mercury­
in­
coal
data
from
all
coal­
fired
units
for
calendar
year
1999.
The
MACT
emission
limitation
calculation
was
based
on
the
performance
of
the
top
unit
in
the
individual
subcategories
of
bituminous
coal,
subbituminous
coal,
lignite
coal,

coal
refuse,
and
IGCC
(
coal
gas).

7.3.1
Mercury
Emission
Limitation
Calculations
for
New
Units.
For
bituminous­
fired
units,
the
best
controlled
unit
was
controlled
with
FF,
and
the
Hg
emissions
factor
was
0.132
lb/
TBtu.
This
value
was
adjusted
for
variability
as
described
above,
and
converted
to
the
output­
based
format
as
discussed
in
section
5.1
above
(
using
35
percent
efficiency
factor).

Consequently,
the
output­
based
Hg
emissions
limitation
for
new
bituminous­
fired
units
was
determined
to
be
5.99
x
10­
6
lb/
MWh.

For
subbituminous­
fired
units,
the
best
controlled
unit
was
also
controlled
with
a
FF,
and
the
Hg
emissions
factor
was
0.6633
lb/
TBtu.
This
value
was
adjusted
for
variability
as
described
above
and
converted
to
the
output­
based
value
(
using
the
35
percent
efficiency
factor).
The
output­
based
Hg
emissions
limitation
for
new
subbituminous­
fired
units
was
determined
to
be
19.6
x
10­
6
lb/
MWh.

For
lignite­
fired
units,
the
best
controlled
unit
was
controlled
with
a
ESP,
and
the
Hg
emissions
factor
was
6.902
lb/
TBtu.
This
value
was
adjusted
for
variability
as
described
above
and
was
converted
to
the
output­
based
value
(
using
the
35
percent
efficiency
factor).
The
42
output­
based
Hg
emissions
limitation
for
new
lignite­
fired
units
was
determined
to
be
62.0
x
10­
6
lb/
MWh.

For
IGCC
units,
the
best
controlled
unit
was
uncontrolled,
and
the
Hg
emissions
factor
was
5.471
lb/
TBtu.
This
value
was
adjusted
for
variability
as
described
above
and
converted
using
the
35
percent
efficiency
factor,
for
an
output­
based
Hg
emissions
limitation
for
new
IGCC
units
of
200
x
10­
6
lb/
MWh.
However,
EPA
believes
that
a
90
percent
reduction
in
Hg
emissions
is
possible
from
new
IGCC
units
based
on
the
use
of
carbon
bed
technology.
The
EPA
believes
that
a
90
percent
Hg
reduction
by
a
beyond­
the­
floor
level
of
control
for
new
IGCC
units
is
achievable.
13
Consequently,
the
output­
based
Hg
emissions
limitation
for
new
lignite­
fired
units
was
determined
to
be
20.0
x
10­
6
lb/
MWh
(
90
percent
of
the
new
unit
limit
determined
above).

For
coal
refuse­
fired
units,
the
best
controlled
unit
was
controlled
with
a
FF,
and
the
Hg
emissions
factor
was
0.118
lb/
TBtu.
This
value
was
adjusted
for
variability
as
described
above,

and
converted
using
the
35
percent
efficiency
factor.
The
output­
based
Hg
emissions
limitation
for
new
coal
refuse­
fired
units
was
determined
to
be
1.16
x
10­
6
lb/
MWh.

Table
2
below
summarizes
the
Hg
emissions
limitations
from
new
coal­
fired
units.
43
TABLE
2.
Hg
EMISSION
LIMITS
FOR
NEW
COAL­
FIRED
UNITS
Unit
Type
Hg
(
10­
6
lb/
MWh)

Bituminous
5.99
Subbituminous
19.6
Lignite
62.0
IGCC
20.0
Coal
refuse
1.16
7.3.2
Nickel
Emissions
Limitation
Calculations
for
New
Units.
The
emission
limit
for
Ni
for
new
oil­
fired
units
was
determined
by
analyzing
the
same
emissions
data
available
for
existing
units.
The
data
were
obtained
from
the
Utility
RTC.
The
EPA
examined
available
test
data
and
found
that
ESP­
equipped
units
can
effectively
reduce
Ni.
The
Ni
emissions
data
mean
concentration
from
the
best­
controlled
oil­
fired
unit
was
used
to
determine
the
emissions
limitation
for
new
oil­
fired
units.
The
best
oil­
fired
unit
Ni
emissions
value
from
the
stack
test
data
was
0.0046
lb/
TBtu.
This
emissions
factor
was
then
adjusted
for
uncertainty
by
applying
variability
factors
as
described
above
for
existing
units,
with
a
resulting
input­
based
Ni
emission
limit
of
76
lb/
TBtu.
The
EPA
then
converted
the
input­
based
value
using
the
rationale
described
in
section
5.1
above
(
using
the
35
percent
net
efficiency
factor).
The
resulting
Ni
emissions
limitation
for
new
oil­
fired
units
is
0.0007
lb/
MWh.
The
EPA
believes
that
this
limitation
is
a
reasonable
estimate
of
actual
unit
performance
of
the
MACT
floor
unit
in
this
case.

8.0
OTHER
ISSUES
The
EPA
identified
several
issues
that
must
be
addressed
in
the
standard
with
regard
to
the
blending
of
fuels
which
fall
into
separate
subcategories
(
in
the
case
of
coal­
fired
units)
and
blending
of
fuels
which
EPA
has
determined
are
exempt
from
the
standard
(
in
the
case
of
oil­
fired
units).
The
EPA
determined
that
these
blending
of
fuels
did
not
warrant
separate
subcategorization
but
did
pose
an
issue
with
regard
to
compliance
with
any
proposed
or
final
rule.
44
Cogeneration
units
also
posed
an
issue
in
that
not
all
power
(
or
energy)
generated
by
the
unit
is
transferred
to
the
grid,
making
use
of
the
output­
based
format
problematic.
The
paragraphs
below
describe
EPA's
position
on
how
to
handle
these
issues.

8.1
Blended
Coals
The
EPA
recognizes
that
many
electric
utility
units
burn
more
than
one
rank
of
coal,
either
at
the
same
time
(
i.
e.,
blending)
or
at
separate
times
during
a
year
(
i.
e.,
seasonally).
Further,
EPA
is
aware
that
several
units
burn
a
supplementary
fuel
(
e.
g.,
petroleum
coke,
TDF)
in
addition
to
a
primary
coal
fuel.
The
EPA
recognizes
this
practice
and
acknowledges
the
effect
that
coal
blending
(
or
use
of
supplementary
fuels)
will
have
on
Hg
emissions.
Because
this
standard
is
not
applicable
to
the
non­
regulated
supplementary
fuels,
the
standard
does
not
provide
an
emission
limitation
for
those
fuels.
The
EPA
believes
that
the
most
appropriate
means
to
address
the
blending
scenarios
is
through
the
compliance
demonstration.

The
EPA
has
identified
several
blending
scenarios
that
might
occur
in
the
industry;

blending
two
or
more
ranks
of
coal,
blending
one
rank
of
coal
with
a
supplementary
(
non­
regulated
fuel),
or
blending
multiple
ranks
of
coal
with
a
supplementary
(
non­
regulated)
fuel.

There
are
two
potential
methods
for
addressing
the
blending
scenarios
where
two
or
more
ranks
of
coal
are
fired.
One
approach
would
be
to
classify
a
unit
based
on
the
predominant
coal
it
burns.
For
example,
if
90
percent
of
the
coal
burned
for
the
compliance
period
were
bituminous
coal,
the
unit
would
be
classified
as
bituminous
and
would
have
to
meet
the
Hg
emission
limitations
for
bituminous
coals.
A
second,
more
equitable
approach
would
be
to
develop
a
weighted
Hg
emission
limit
based
on
the
proportion
of
energy
output
(
in
Btu)
contributed
by
each
coal
rank
burned
during
the
compliance
period
and
the
coal's
subcategory
Hg
emission
limitation.

The
weighted
emission
limit
would,
in
effect,
be
a
blended
emission
limitation
based
on
the
Hg
emission
limitations
of
the
subcategories
of
the
coals
burned.

The
other
scenarios
discussed
above
involve
blending
a
regulated
fuel
(
coal,
oil,
coal
refuse,
or
coal
gas)
with
a
supplementary,
non­
regulated
fuel
(
e.
g.,
petroleum
coke,
TDF).
The
application
of
the
same
methods
would
be
appropriate
for
units
that
burn
a
regulated
fuel
with
supplementary,
non­
regulated
fuels;
however,
there
would
be
no
adjustment
to
the
Hg
emission
limitation
with
regard
to
the
supplementary,
non­
regulated
fuel.
45
For
example,
where
the
predominant
fuel
determines
which
emission
limitation
would
apply,
the
compliance
calculation
would
include
the
energy
output
(
Btu)
of
all
fuels
burned
(
including
the
supplementary
fuel);
the
emissions
considered
would
include
all
Hg
emissions
measured
by
the
CEMS;
and
the
unit
would
comply
with
the
emission
limitation
associated
with
the
subcategory
of
the
predominant
fuel.
Under
the
other
method,
a
weighted
Hg
emission
limitation
would
be
developed
based
on
the
proportions
of
energy
output
(
Btu)
contributed
by
only
the
regulated
fuels.
For
example,
if
the
unit
burned
bituminous,
subbituminous,
and
petroleum
coke
during
the
compliance
period,
and
40
percent
of
the
Btu
output
was
attributable
to
the
bituminous,
40
percent
of
the
Btu
output
was
attributable
to
the
subbituminous,
and
20
percent
of
the
Btu
output
was
attributable
to
the
petroleum
coke,
the
blended
Hg
emissions
limitation
would
be
based
on
the
bituminous
and
subbituminous
emission
limitations
in
a
50/
50
ratio.
The
compliance
calculation
would
include
the
energy
output
(
Btu)
of
all
fuels
burned
(
including
the
supplementary
fuel),
the
emissions
considered
would
include
all
Hg
emissions
measured
by
the
CEMS,
and
the
unit
would
comply
with
the
blended
Hg
emission
limitation.

The
EPA
recognizes
that
new
electric
utility
units
may
still
be
designed
to
burn
more
than
one
rank
of
coal,
either
at
the
same
time
(
i.
e.,
blending)
or
at
separate
times
during
a
period
of
time
(
i.
e.,
seasonally).
The
EPA
finds
no
reason
to
address
blended
coals
differently
for
new
units
than
it
did
for
existing
units.
Therefore,
the
method
of
addressing
blended
coals
with
regard
to
the
Hg
emission
limit
calculation
will
remain
the
same
for
new
units
as
is
prudent
for
existing
units.
Further,
EPA
believes
that
consistency
in
the
compliance
method
would
be
appropriate,

because
many
utility
owners/
operators
will
at
some
point
be
addressing
compliance
for
both
new
and
existing
units
at
the
same
facility.

8.2
Dual­
fired
Units
The
EPA
is
aware
that
an
oil­
fired
unit
may
fire
oil
at
certain
times
of
the
year
and
natural
gas
at
other
times,
as
well
as
blends
of
residual
oil
and
distillate
oil.
This
blending
of
fuels
is
conducted
for
many
reasons,
most
of
which
are
economically
driven
with
regard
to
the
availability
of
fuels
and
the
price,
and
may
be
seasonal
in
nature.

The
EPA
believes
that
units
that
burn
distillate
oil
exclusively
should
be
exempted
from
the
requirements
of
the
standard
and
natural
gas­
fired
units
are
excluded
from
the
definition
of
a
covered
source
by
the
Administrator.
Therefore,
the
requirements
of
the
standard
apply
to
units
46
that
fire
residual
oil
in
any
proportion
with
another
oil
and
to
units
that
fire
residual
oil
at
98
percent
or
greater
of
their
annual
fuel
consumption,
where
the
supplementary
fuel
is
natural
gas.

The
EPA
believes
that
a
cutoff
of
two
percent
fuel
oil­
firing
would
separate
those
units
that
are
"
fundamentally"
natural
gas­
fired
but,
for
startup
or
other
operational
needs,
burn
fuel
oil.
The
blending
scenarios
that
might
occur
for
oil­
fired
units
include
the
co­
firing
of
residual
oil
with
distillate
oil
and
the
firing
of
residual
oil
and
natural
gas
at
different
times.

The
unit
that
burns
residual
oil
exclusively
would
be
required
to
meet
the
oil­
fired
Ni
emission
limitation.
For
units
that
burn
exclusively
distillate
oil,
the
unit
would
be
exempted
from
meeting
the
Ni
emission
limitation.
For
units
that
blend
residual
oil
with
distillate
oil,
the
unit
would
be
required
to
meet
the
Ni
emission
limitation,
and
would
include
all
Ni
and
Btus
or
megawatt
hours
generated
from
the
use
of
the
distillate
oil
in
the
compliance
demonstration
calculation.
Likewise,
a
unit
that
burns
residual
oil
during
certain
periods
and
natural
gas
during
certain
periods
would
include
the
natural
gas­
fired
contributions
(
Ni
and
Btu
or
megawatt
hours)

in
the
compliance
calculation.

Although
EPA
has
not
identified
any
other
supplementary
fuels
burned
in
the
oil­
fired
industry,
we
are
aware
that
such
a
scenario
may
exist
or
might
occur
in
the
future.
The
EPA
intends
that
where
any
supplementary
fuel
is
co­
fired
with
residual
oil,
the
Ni
and
the
Btus
or
megawatt
hours
contributed
by
the
supplementary
fuel
be
accounted
for
in
the
compliance
calculation
and
that
the
unit
be
required
to
meet
the
Ni
emission
limit
for
existing
oil­
fired
units.

The
EPA
is
aware
that
new
oil­
fired
units
may
be
designed
and
built
to
fire
the
combination
of
oil
and
natural
gas,
as
are
existing
units.
The
EPA
believes
that
the
reasons
for
not
burning
natural
gas
exclusively
will
continue
to
be
based
on
economics
or
availability
of
fuel
(
i.
e.,
seasonal
considerations).
Therefore,
EPA
intends
to
treat
new
oil­
fired
units
that
burn
a
combination
of
oil
and
natural
gas
in
the
same
manner
as
existing
units
for
compliance.

8.3
Cogeneration
Units
A
cogeneration
facility
that
sells
excess
steam
or
electricity
to
any
utility
power
distribution
system
equal
to
less
than
one­
third
of
its
potential
electric
output
capacity
and/
or
less
than
or
equal
to
25
MWe
is
considered
to
be
either
an
industrial,
commercial,
or
institutional
boiler.
However,
a
cogeneration
facility
that
meets
the
above
definition
of
an
electric
utility
steam
generating
unit
during
any
portion
of
a
year
would
be
subject
to
the
standard.
47
For
cogeneration
units,
steam
is
also
generated
for
process
use.
The
energy
content
of
this
process
steam
must
also
be
considered
in
determining
compliance
with
the
output­
based
standard.
This
consideration
is
accomplished
by
taking
the
net
efficiency
of
a
cogeneration
unit
into
account.
Under
a
Federal
Energy
Regulatory
Commission
regulation,
the
efficiency
of
cogeneration
units
is
determined
from
the
useful
power
output
plus
one
half
the
useful
thermal
output
(
18
CFR
292.205).
To
determine
the
process
steam
energy
contribution
to
net
plant
output,
a
50
percent
credit
of
the
process
steam
heat
is
necessary.

Therefore,
owners/
operators
of
cogeneration
units
would
need
to
monitor
the
portion
of
their
net
plant
output
that
is
process
steam
so
that
they
can
take
the
50
percent
credit
of
the
energy
portion
of
their
process
steam
net
output.
For
example,
a
cogeneration
unit
measures
its
net
electrical
output
over
a
compliance
period,
as
30,000
MWh.
During
the
same
period
the
unit
burns
coal
that
provides
750
billion
Btu
input
to
its
furnace/
boiler,
and
emits
0.2
lb
Hg.
Using
equivalents
found
in
40
CFR
60
for
electric
utilities
(
i.
e.,
250
million
Btu/
hr
input
to
a
boiler
is
equivalent
to
73
MWe
input
to
the
boiler;
73
MWe
input
to
the
boiler
is
equivalent
to
25
MWe
output
from
the
boiler;
therefore,
250
million
Btu
input
to
the
boiler
is
equivalent
to
25
MWe
output
from
the
boiler)
the
50
percent
credit
could
be
found
as
follows.
The
net
output
calculation
would
be
750
billion
Btu
x
(
25
MWe
output/
250
million
Btu/
hr
input)
=
75,000
MWh
equivalent
electrical
output
from
the
boiler
over
the
compliance
period.
Of
this
amount,
30,000
MWh
was
produced
as
electricity
sent
to
the
grid,
leaving
45,000
MWh
as
the
energy
converted
to
steam
for
process
use.
Half
of
this
amount
is
22,500
MWh.
The
unit's
Hg
CEMS
records
a
total
of
0.2
lb
Hg
over
the
same
compliance
period.
The
adjusted
Hg
emission
rate
is
then:
0.2
lb
Hg/(
30,000
MWh
+
22,500
MWh)
=
3.8
x
10­
6
lb
Hg/
MWh.
Cogeneration
units
would
have
to
account
for
the
process
steam
portion
of
their
emissions
in
the
same
manner
for
Ni
emissions,
if
applicable,
as
well.
48
9.0
REFERENCES
1.
U.
S.
EPA.
"
Study
of
Hazardous
Air
Pollutant
Emissions
from
Electric
Utility
Steam
Generating
Units
­­
Final
Report
to
Congress,
Volume
1."
EPA­
453/
R­
98­
004a.
February
1998.

2.
U.
S.
EPA.
1999
Electric
Utility/
Information
Collection
Effort
(
EU/
ICE).
http://
www.
epa.
gov/
ttn/
atw/
combust/
utiltox/
utoxpg.
html
3.
Department
of
Energy/
Energy
Information
Administration
(
DOE/
EIA).
"
Steam­
Electric
Plant
Operation
and
Design
Report
Database
based
on
EIA
Form,
EIA­
767
(
1999)."

4.
DOE/
EIA.
"
Steam­
Electric
Plant
Operation
and
Design
Report
Database
based
on
EIA
Form,
EIA­
767
(
2001)."

5.
American
Society
for
Testing
and
Materials
(
ASTM).
"
D388­
77,
90,
91,
95,
or
98a,
Standard
Specification
for
Classification
of
Coals
by
Rank
Designation."
1998.

6.
U.
S.
EPA.
"
Electric
Utility
Steam
Generating
Units
Section
112
Rule
Making"
web
site:
http://
www.
epa.
gov/
ttn/
atw/
combust/
utiltox/
utoxpg.
html
7.
Singer,
J.
G.
(
ed.).
Combustion,
Fossil
Power.
Fourth
Edition.
Combustion
Engineering,
Inc.,
Windsor,
CT.
1991.
p.
1­
12.

8.
Federal
Advisory
Committee
Act
(
FACA).
Clean
Air
Act
Advisory
Committee
(
CAAAC).
Permits,
New
Source
Review,
and
Toxics
Subcommittee.
"
Recommendations
for
the
Utility
Air
Toxics
MACT,
Final
Working
Group
Report."
October
2002.
http://
www.
epa.
gov/
ttn/
atw/
combust/
utiltox/
utoxpg.
html
9.
DOE/
EIA.
"
Challenges
of
Electric
Power
Industry
Restructuring
for
Fuel
Suppliers,
Chapter
4.
Impacts
of
Electric
Power
Industry
Restructuring
on
Crude­
Oil­
Derived
Fuels"
pp.
64­
65.
DOE/
EIA­
0623,
Distribution
Category
UC­
950.
September
1998.

10.
Maxwell,
W.
H.,
U.
S.
EPA/
OAQPS/
ESD/
CG.
"
Analysis
of
variability
in
determining
MACT
floor
for
coal­
fired
electric
utility
steam
generating
units."
Docket
A­
92­
55,
Item
No.
II­
B­
8.

11.
Multi
variable
Method
to
Estimate
the
Mercury
Emissions
of
the
Best­
Performing
Coal­
Fired
Utility
Units
Under
the
Most
Adverse
Circumstances
Which
Can
Reasonably
be
Expected
to
Recur,
West
Associates,
submitted
at
the
March
4,
2003
Clean
Air
Act
Advisory
Committee
Mercury
MACT
Working
Group
meeting.

12.
Memorandum
and
attachment
from
W.
Maxwell,
EPA/
CG,
to
Utility
MACT
Project
Files,
December
2003.
Power
plant
efficiency
table.
49
13.
Rutkowski,
M.
G.,
M.
G.
Klett,
and
R.
C.
Maxwell.
"
The
Cost
of
Mercury
Removal
from
Coal­
Based
IGCC
Relative
to
a
PC
Plant."
Gasification
Technologies
2002
Symposium,
October
27­
30,
2002,
San
Francisco,
CA.
Docket
A­
92­
55,
Item
No.
II­
I­
23.
50
10.0
APPENDICES
51
APPENDIX
10.1
LIST
OF
ACRONYMS
APCD­
Air
pollution
control
device
ASTM
­
American
Society
for
Testing
and
Material
Btu
­
British
Thermal
Units
CAA
­
Clean
Air
Act
CEMS
­
Continuous
emissions
monitoring
system
CFD
­
Cumulative
frequency
distribution
CO
­
Carbon
monoxide
DOE/
EIA
­
Department
of
Energy,
Energy
Information
Administration
EPA
­
Environmental
Protection
Agency
EPRI
PISCES
­
Electric
Power
Research
Institute,
PISCES
Study
ESP
­
Electro­
static
precipitator
EU/
ICE
­
Electric
Utility/
Information
Collection
Effort
FBC
­
Fluidized
bed
combustor
FF
­
Fabric
filters
FGD­
Fluidized
gas
desulfurization
FR­
Federal
Register
gal
­
Gallon
GCP
­
Good
combustion
practices
HAP
­
Hazardous
air
pollutants
Hg
­
Mercury
HHV
­
Higher
heat
value
IGCC
­
Integrated
gasification
combined
cycle
lb
­
Pound
MACT
­
Maximum
achievable
control
technology
MWe
­
Megawatt
electricity
MWh
­
Megawatt
hour
NESHAP­
National
Emissions
Standards
for
Hazardous
Air
Pollutants
52
Ni
­
Nickel
NO
x
­
Nitrogen
oxides
NSPS
­
New
Source
Performance
Standards
PC
­
Pulverized
coal
PM
­
Particulate
matter
RTC
­
Electric
Utility
Report
to
Congress
SDA
­
Spray
dryer
adsorber
SO
2
­
Sulfur
dioxide
Syngas
­
Synthetic
coal
gas
TBtu
­
Trillion
British
thermal
units
TDF
­
Tire­
derived
fuel
53
APPENDIX
10.2
Data
and
Emission
Limitation
Calculations
See
Excel
Spreadsheet:
MACT
Floor
Data.
xls
54
APPENDIX
10.3
Mercury
Speciation
Analysis
by
Coal
Rank
See
Excel
Spreadsheet:
Hg
Speciation
by
fuel.
xls
