Attachment
2
Potential
for
Efficiency
Improvements
At
Existing
Coal­
Fired
Power
Plants
The
average
efficiency
of
the
US
fleet
of
coal­
fired
power
plants
was
33%
in
2000,
which
equates
to
a
heat
rate
(
fuel
energy
needed
to
generate
one
kilowatt­
hour
of
power)
of
10,240
Btu/
kwh.
Possible
measures
to
improve
this
efficiency
have
been
investigated
in
recent
years
due
to
interest
in
fuel
savings,
as
well
as
in
the
greenhouse
gas
emissions
reduction
(
i.
e.,
CO2)
which
would
accompany
such
a
performance
improvement.
For
example,
see
Integrating
Consultancy
­
Efficiency
Standards
for
Power
Generation,
Australian
Greenhouse
Office,
Jan.
2000;
Review
of
Potential
Efficiency
Improvements
at
Coal­
Fired
Power
Plants,
Perrin
Quarles
Associates,
Inc.,
for
Clean
Air
Markets
Division,
USEPA,
April
2001;
Increasing
Electricity
Availability
From
Coal­
Fired
Generation
in
the
Near­
Term,
The
National
Coal
Council,
May
2001.
The
Department
of
Energy
is
currently
studying
such
measures.
Preliminary
results
from
the
DOE
effort
have
highlighted
several
promising
options,
as
indicated
in
the
table
below.

Table
2.1
Technology
%
Efficiency
Improvement
Concept
Turbine
reblading
5­
10%
Replace
existing
turbine
blade
sections
and
seals
with
more
efficient
computer­
based
designs.

Energy
management
1­
5%
Replace
current
power
plant
control
system
with
real­
time
performance
monitoring
and
adjustment
of
chemical
feed
rates,
air
flow,
steam
temperatures,
outage
maintenance
against
a
theoretical
model.

Intelligent
soot
blowing
1­
2%
Replace
current
soot
blowing
system
with
"
smart"
systems,
which
can
reduce
steam
use
for
soot
blowing
by
30%.

Distributed
Control
System
controls
0.5­
2%
Automate
manual
adjustment
of
air
registers,
burner
tilt,
fan
power,
etc.,
using
an
optimal
computer
"
smart"
system.

Generator
exciter
replacement
1­
2%
Replace
current
mechanical
exciter
with
a
more
efficient
solid­
state
system.

Condenser
enhancement
1%
Replace
condenser
with
a
larger
unit
to
reduce
back
pressure
and
make
steam
turbine
more
efficient.

Overall
improvments
8­
17%
(
Numbers
are
not
additive
due
to
some
overlapping
improvements,
and
possible
double­
counting.)
In
general,
these
improvements
all
improve
the
efficiency
of
steam
conversion
to
electricity,
or
reduce
parasitic
power
consumption
within
the
power
plant.
Such
improvements
in
efficiency
produce
more
power
for
the
same
amount
of
energy
consumed,
and
therefore
do
not
increase
emissions,
in
and
of
themselves.
However,
two
larger
scale
issues
must
also
be
considered.

First,
if
the
increase
in
efficiency
also
improves
the
unit's
economics,
then
the
unit
might
be
dispatched
(
used)
more,
while
other
units
are
used
less.
It
is
likely
that
voluntary
efficiency
improvements
would
improve
unit
economics.
To
further
evaluate
this
situation,
emission
changes
at
a
midwestern
power
pool
for
were
evaluated
for
a
range
of
hypothetical
efficiency
improvements
in
coal­
fired
units.
Four
hypothetical
NGCC
generating
units
were
"
added"
to
the
system
to
reflect
a
future
scenario
in
which
a
broader
mix
of
coal
and
gas­
fired
units
would
probably
exist.
The
analysis
found
that
only
a
very
large
change
in
efficiency
(
13.5%,
or
greater)
would
be
likely
to
change
the
relative
dispatching
of
coal
units
with
generically
different
units,
such
as
natural
gas­
fired
or
nuclear
units,
and
that
even
at
such
high
levels
of
efficiency
improvement,
the
net
effect
was
a
reduction
in
emissions
of
3­
4%.
(
10/
19/
2001
email
from
V.
Koritarov,
ANL,
to
D.
Carter,
USDOE)
It
would
be
reasonable
to
expect
some
shifting
in
dispatching
between
other
coal
units,
if
all
the
units
were
not
comparably
improved.
Such
a
change
could
result
in
an
increase
in
emissions
if
higher
emitting
units
were
the
subject
of
efficiency
improvements.
However,
given
emission
caps
for
sulfur
dioxide,
caps
in
certain
states
for
nitrogen
oxides,
and
a
tendency
by
power
plant
operators
to
achieve
optimal
performance
(
both
efficiency
and
emissions)
at
their
"
flagship"
units,
it
seems
much
more
likely
that
units
receiving
the
greatest
efficiency
upgrades
would
be
the
cleaner
units.
Under
these
circumstances,
efficiency
improvements
of
the
type
cited
above
would
reduce
emissions.
In
other
words,
if
one
expects
a
10%
overall
improvement
in
efficiency
at
coal
units,
efficiency
improvements
of
12%
might
occur
at
the
lowest
emission
units,
and
improvements
of
8%
at
the
highest
emitting
units.
This
behavior
would
lead
to
significant
emission
reductions
in
periods
of
"
off­
peak"
generation,
which
includes
the
major
portion
of
the
year.
However,
this
type
of
behavior
is
difficult
to
model
and
was
not
simulated
in
the
model
runs
by
ANL
or
EIA.

The
second
larger
scale
issue
is
that
of
demand
growth.
The
growth
in
electricity
demand
over
the
next
decade
is
projected
to
be
greater
than
expected
growth
in
electricity
production
due
to
efficiency
improvements
at
coal­
fired
power
plants.
Another
way
of
looking
at
this
is
that
almost
all
additional
generation
which
comes
from
efficiency
improvements
would
be
power
not
needed
from
new
generators.
Because
the
efficiency
improvements
are
at
existing
coal
units,
whereas
new
generation
over
the
next
decade
will
be
dominated
by
much
lower
emitting
natural
gas
combined
cycle
units,
one
might
suppose
that
the
efficiency
improvement
would
result
in
increased
emissions
overall.
However,
this
is
not
the
case.
As
long
as
the
increased
power
production
does
not
require
additional
coal
consumption
(
which
is
the
case
for
these
efficiency
improvements),
then
the
resulting
net
emissions
will
be
lower
than
the
total
emissions
for
"
unimproved"
coal
plants
and
"
super­
clean"
natural
gas
plants.

Most,
but
not
all
of
the
efficiency
improving
technologies
cited
in
Table
2.1
reflect
replacement
components.
This
is
important
to
note
because
some
NSR
policies
might
apply
differently
to
replacement
parts
versus
new
components
added
to
a
plant
for
the
sole
purpose
of
improving
efficiency.

Given
the
expanding
suite
of
efficiency
improving
technologies,
and
growing
interest
in
reducing
greenhouse
gas
emissions
through
efficiency
improvements,
it
is
reasonable
to
project
overall
efficiency
improvements,
in
the
absence
of
NSR
constraints,
as
large
as10­
15%.
Such
a
range
is
much
larger
than
conventional
wisdom,
which
is
perhaps
shaped
by
expectations
under
the
current
NSR
policy,
and
the
absence
of
efficiency
incentives
related
to
climate
change
concerns.
To
cover
a
broad
range
of
possible
improvements,
a
range
of
5%
to
15%
was
examined
using
the
NEMS
modeling
system.
APPENDIX
B
EVALUATION
OF
ROUTINE
MAINTENANCE
MODEL
SCENARIO
FOR
POWER
PLANTS
(
ENVIRONMENTAL
PROTECTION
AGENCY)
1
This
finding
is
described
in
detail
in
EPA's
June
13,
2002
New
Source
Review
Report
to
the
President.
EVALUATION
OF
ROUTINE
MAINTENANCE
MODEL
SCENARIO
FOR
POWER
PLANTS
Purpose:
This
analysis
uses
model
scenarios
to
evaluate
the
impact
that
the
changes
to
the
routine
maintenance
provisions
of
NSR
are
likely
have
on
emissions
from
the
power
generation
sector.

Methodology:
In
order
to
evaluate
the
impact
of
the
routine
maintenance
provisions,
EPA
considered
a
scenario
under
which
NSR
regulations
remained
in
place
and
a
range
of
scenarios
that
could
occur
if
NSR
did
not
exist.
The
first
scenario
is
intended
to
represent
the
existing
program,
which
the
EPA
has
found
impedes
or
results
in
cancellation
of
projects
that
maintain
and
improve
reliability,
availability,
and
efficiency
at
existing
power
plants.
1
The
second
range
of
scenarios
represents
companies
receive
flexibility
under
the
NSR
program
that
removes
many
of
these
impediments
.
As
part
of
this
analysis,
EPA
reviewed
three
key
variables:
change
in
SO2
emissions,
change
in
NOx
emissions
and
change
in
cost.

It
is
worth
noting
that
EPA
recently
promulgated
final
rules
governing
the
use
of
plantwide
applicability
limits
(
PALs),
and
Clean
Units.
Some
sources
with
in
the
electric
utility
generation
industry
may
take
advantage
of
these
changes.
However,
any
such
decision
will
be
based
on
case
specific
information
related
to
their
past
operating
levels,
current
levels
of
control
and
company's
specific
strategies
for
complying
with
NSR.
Therefore,
we
can
not
make
estimates
on
how
many
sources
may
take
advantage
of
PALs
and
Clean
Units.
To
the
extent
they
are
used
within
the
industry,
they
will
dampen
the
effects
shown
in
this
analysis
(
i.
e.,
estimated
decreases
and
increases
will
not
be
as
large.

This
analysis
was
performed
using
the
Integrated
Planning
Model
(
IPM).
IPM
is
a
linear
programming
model
that
EPA
uses
to
analyze
the
effect
of
various
environmental
policies
on
the
power
sector.
It
provides
forecasts
of
least­
cost
capacity
expansion,
electricity
dispatch
and
emission
control
strategies
for
meeting
energy
demand
and
environmental,
transmission,
dispatch
and
reliability
constraints.
EPA
has
used
it
to
analyze
many
environmental
policies
including
the
Phase
II
Acid
Rain
Nitrogen
Oxide
regulations
and
the
Nitrogen
Oxide
SIP
Call.
Analysis
can
be
performed
varying
multiple
constraints
such
as
availability
of
various
types
of
power
plants
(
e.
g.
coal­
fired,
nuclear,
gas­
fired
combined
cycle
units),
heat
rates
of
various
types
of
power
plants,
environmental
constraints
(
e.
g.
caps
on
emissions,
emission
rate
limitations).
More
detail
regarding
IPM
can
be
found
in
the
document
titled
"
Documentation
of
EPA
Modeling
Application
(
V.
2.1)
Using
the
Integrated
Planning
Model,
which
can
be
found
at:
http://
www.
epa.
gov/
airmarkets/
epa­
ipm/
index.
html.

Assumptions:
The
first
scenario,
referred
to
as
the
NSR
base
case,
approximates
utility
behavior
under
the
current
program,
where
the
EPA
has
found
that
companies
perform
limited
maintenance
on
coal
plants
because
of
concerns
about
NSR.
In
this
scenario,
it
was
assumed
that
the
performance
of
coal
units
would
deteriorate,
resulting
in
higher
heat
rates
and
lower
capacities.
EPA
did
not
assume
that
reduced
maintenance
resulted
in
a
change
in
maximum
potential
unit
availability.
This
is
because
over
the
last
20
years,
availability
of
coal­
fired
plants
has
increased
even
as
the
plants
have
aged.
This
is
due
in
large
part
to
improved
maintenance
practices.
For
instance
tests
to
inspect
boiler
tubes
have
been
continually
improving
(
see
"
Preventing
Boiler
Tube
Failures
with
EMAT's",
S.
P.
Clark
et
al,
"
EPRI
International
Conference
on
Boiler
Tube
Failures
and
HRSG
Tube
Failures
and
Inspects",
November
6­
8,
2001).
These
improved
preventive
maintenance
practices
allow
companies
to
replace
components
during
regularly
scheduled
outages
before
they
fail
rather
than
causing
unscheduled
outages
after
they
fail.
Although
we
did
not
assume
availability
decreases,
we
did
assume
that
continued
future
availability
increases
would
diminish,
and
any
remaining
increases
would
essentially
be
negated
by
deterioration
caused
by
limited
maintenance.

The
second
range
of
scenarios,
referred
to
as
increased
maintenance
cases
#
1
­
#
5
,
looks
at
a
range
of
scenario
for
what
might
happen
in
the
utility
sector
if
companies
were
provided
with
increased
flexibility
under
NSR
to
perform
maintenance.
This
would
result
in
lower
heat
rates,
higher
capacities
and/
or
higher
unit
availabilities
for
these
units.
Finally
EPA
looked
at
one
case
(
standard
base
case)
in
which
heat
rate,
capacity
and
unit
availability
did
not
change.
Table
1:
Key
modeling
assumptions
in
routine
maintenance
analysis
Winter
Availability
Summer
Availability
Heat
Rate
Change
Capacity
Change
NSR
Basecase
81.6%
89.8%
+
0.1%
per
year
­
0.1%
per
year
Increased
Maintenance
Case
#
1
85.0%
92.0%
­
0.1%
per
year
+
0.1%
per
year
Increased
Maintenance
Case
#
2
81.6%
89.8%
­
0.1%
per
year
+
0.1%
per
year
Increased
Maintenance
Case
#
3
85.0%
92.0%
­
1.6%
in
year
2005
and
beyond
+
1.6%
in
year
2005
and
beyond
Increased
Maintenance
Case
#
4
85.0%
92.0%
­
3.2%
in
year
2005
and
beyond
+
3.2%
in
year
2005
and
beyond
Increased
Maintenance
#
5
81.6%
89.8%
­
1.6%
in
year
2005
and
beyond
+
1.6%
in
year
2005
and
beyond
Standard
Base
Case
81.6%
89.8%
No
change
No
change
It
is
important
to
note
several
limitations
to
this
analysis.
First
this
analysis
only
considered
emission
regulations
that
are
currently
in
effect
(
e.
g.
the
NOx
SIP
Call
and
the
Title
IV
Acid
Rain
Provisions).
Future
environmental
regulations
such
as
emission
reduction
requirements
necessary
to
meet
the
fine
particulate
matter
standards
or
emission
reductions
under
multi­
pollutant
regulations
could
significantly
change
this
analysis.
Second,
the
analysis
assumed
the
operating
and
maintenance
costs
of
coal­
fired
units
was
the
same
for
units
performing
limited
maintenance
and
for
units
performing
increased
maintenance..
Since
the
most
significant
cost
associated
with
running
an
existing
power
plant
is
the
cost
of
fuel,
this
impact
is
probably
fairly
small.

Results:

Changes
in
SO2
Emissions,
NOx
emissions
and
cost
are
summarized
in
Tables
2,
3
and
4
below.

Table
2:
Changes
in
SO2
emissions
in
scenarios
considered
in
routine
maintenance
analysis.

2005
SO2
Emissions
(
tons)
2010
SO2
Emissions
(
tons)
2015
SO2
Emissions
(
tons)
2020
SO2
Emissions
(
tons)

NSR
Base­
case
10,168,230
9,713,684
9,101,622
9,103,275
Increased
Maintenance
Case
#
1
10,135,120
9,739,029
9,104,121
9,102,688
Increased
Maintenance
Case
#
2
10,186,660
9,701,112
9,099,363
9,099,271
Increased
Maintenance
Case
#
3
10,075,060
9,773,242
9,104,836
9,103,779
Increased
Maintenance
Case
#
4
10,009,250
9,813,664
9,105,429
9,104,396
Increased
Maintenance
#
5
10,079,510
9,764,971
9,099,923
9,100,361
Standard
Base
Case
10,168,520
9,712,499
9,100,264
9,100,680
As
shown
in
Table
2,
there
is
very
little
change
in
SO2
emissions
over
the
entire
time
period
studied
under
the
two
scenarios.
This
is
because
SO2
emissions
are
already
capped
nationally
under
the
Title
IV
Acid
Rain
Provisions.
Therefore
if
a
unit
decreases
its
emissions
to
make
room
under
its
PAL,
it
could
instead
sell
excess
allowances
to
another
unit.
However
because
emissions
can
also
be
shifted
temporally
by
banking
emission
allowances
to
be
used
in
a
future
year
there
can
be
significant
changes
in
emissions
for
a
specific
year.
While
temporal
distribution
of
emissions
did
not
change
much
over
time
in
the
NSR
cases
considered,
there
was
more
temporal
distribution
of
emissions
in
the
increased
maintenance
scenarios
considered.

Table
3:
Changes
in
NOx
emissions
in
scenarios
considered
under
routine
maintenance
scenarios.

2005
NOx
Emissions
(
tons)
2010
NOx
Emissions
(
tons)
2015
NOx
Emissions
(
tons)
2020
NOx
Emissions
(
tons)

NSR
Base­
case
4,279,362
4,285,400
4,338,461
4,375,486
Increased
Maintenance
Case
#
1
4,340,166
4,362,948
4,442,881
4,471,499
Increased
Maintenance
Case
#
2
4,276,550
4,283,081
4,327,979
4,362,859
Increased
Maintenance
Case
#
3
4,307,796
4,350,737
4,423,141
4,472,706
Increased
Maintenance
Case
#
4
4,276,172
4,334,671
4,412,340
4,460,041
Increased
Maintenance
#
5
4,259,170
4,271,294
4,324,992
4,363,930
Standard
Base
Case
4,277,407
4,285,423
4,332,209
4,360,044
Increasing
capacity
(
under
the
increased
maintenance
cases)
leads
to
increases
in
NOx
emissions.
When
comparing
increased
maintenance
cases
#
1
and
#
2
(
which
had
the
same
increases
in
efficiency,
but
different
changes
in
maximum
availability,
NOx
emissions
increase
by
an
average
of
almost
92,000
tons
per
year
over
the
time
period
analyzed.

It
appears
that
changing
heat
rates
and
capacities
has
the
opposite
effect
on
emissions..
NOx
emissions
actually
decrease
when
flexibility
under
NSR
allows
power
generation
companies
to
improve
efficiency
by
performing
increased
maintenance
if
maximum
availability
of
these
units
does
not
change
at
the
same
time.
For
instance
if
one
compares
two
scenarios
with
the
same
maximum
capacities:
NSR
Base­
case
,
increased
maintenance
case
#
2
and
the
standard
base
case,
average
emissions
are
about
7000
tons
per
year
higher
over
the
time
period
analyzed
in
NSR
Base­
case
where
heat
rates
are
higher
and
capacities
are
lower.
Looking
at
increased
maintenance
cases
#
3
and
#
4
shows
the
same
trend.
In
these
two
cases
maximum
availability
remains
constant,
but
heat
rates
are
lower
and
capacities
are
higher
in
increased
maintenance
case
#
4.
These
lower
heat
rates
and
higher
capacities
lead
to
emissions
that
are
on
average
nearly
18000
tons
per
year
less
in
increased
maintenance
case
#
4
than
in
increased
maintenance
case
#
5.

Another
point
to
note
is
that
EPA
also
looked
at
the
speed
in
which
the
improvements
to
the
units
were
made.
For
example
by
2020,
the
heat
rate
decrease
and
the
capacity
increase
was
the
same
in
both
increased
maintenance
case
#
2
and
increased
maintenance
case
#
5
were
the
same.
However
in
case
#
5,
those
changes
happened
in
one
step
in
2005,
in
case
#
2,
the
changes
happened
gradually.
When
the
changes
occurred
all
at
emissions
were
lower
in
the
early
years.
In
the
later
years,
when
the
total
magnitude
of
the
changes
was
more
similar
in
both
cases,
the
NOx
emissions
were
also
more
similar.

This
analysis
suggests
that
the
effect
that
changing
the
requirements
of
NSR
with
regards
to
routine
maintenance
will
have
on
emissions
is
dependent
upon
the
effect
that
it
will
have
on
maximum
unit
availabilities.
If
the
routine
maintenance
changes
increase
efficiency
and
plant
capacity
without
increasing
maximum
unit
availability,
this
analysis
suggests
that
the
changes
could
decrease
emissions.
The
amount
of
that
emission
decrease
would
depend
both
on
how
much
heat
rate
decreased
and
capacity
increased
and
how
quickly
these
changes
occurred.
The
greater
the
heat
rate
decrease
and
capacity
increase
and
the
more
quickly
the
changes
occurred,
the
greater
the
emission
reductions.
If
on
the
other
hand,
the
new
provisions
increase
maximum
unit
availabilities
this
analysis
suggests
that
the
changes
could
increase
emissions.

Changes
in
cost
are
summarized
in
Table
4
below.
Note
that
this
analysis
does
not
consider
changes
in
maintenance
costs,
it
only
assumes
changes
in
fuel
costs
and
changes
in
capital
costs
associated
with
new
generating
units
and
new
emission
control
equipment.
Therefore
it
probably
understates
the
cost
of
the
increased
maintenance
scenarios
and
understates
the
cost
of
the
NSR
Base­
case.

Table
4:
Total
cost
of
scenarios
considered
(
in
1999
dollars)

2005
Total
Cost
(
million
1999
dollars)
2010
Total
Cost
(
million
1999
dollars)
2015
Total
Cost
(
million
1999
dollars)
2020
Cost
(
million
1999
dollars)

NSR
Base­
case
76,187
80,934
88,921
95,819
Increased
Maintenance
Case
#
1
75,432
79,819
87,306
92,817
Increased
Maintenance
Case
#
2
76,088
80,290
87,861
93,781
Increased
Maintenance
Case
#
3
74,422
79,309
86,715
92,788
Increased
Maintenance
Case
#
4
73,740
78,250
85,898
91,932
Increased
Maintenance
#
5
75,164
79,782
87,600
93,784
Standard
Base
Case
76,149
80,572
88,404
94,588
For
more
detailed
results,
see
the
attached
IPM
run
summaries.
The
runs
are
listed
in
Table
5
below.

Table
5:
IPM
Runs
used
in
this
analysis
Scenario
IPM
Run
#

NSR
Base­
case
NSR­
13
Increased
Maintenance
Case
#
1
NSR­
8
Increased
Maintenance
Case
#
2
NSR­
11
Increased
Maintenance
Case
#
3
NSR­
14
Increased
Maintenance
Case
#
4
NSR­
15
Increased
Maintenance
#
5
NSR­
16
Standard
Base
Case
IPM2000s100d
