Note to:	Files

From:		Doug Carter, DOE/FE-26

Subject:	Logic for aAnalysis of NSR rule impacts

Date:		August 62111, 2003

Introduction

Both EPA and DOE/EIA have modeled potential impacts of allowing
efficiency upgrades on coal-fired power plants, and results of these
simulations were included in the draft Regulatory Impact Analysis (RIA)
published by EPA.  The major differences in the two analyses were that
they used different simulation models (NEMS vs IPM), and they evaluated
somewhat different assumptions for how the change at the power plant
might manifest in operation (i.e., assumptions on efficiency, capacity,
and availability changes).  After further additional consideration, this
note explains the basis for believing that additional assumptions should
be evaluatedDOE conducted further analysis, using assumptions differing
from both of the earlier analyses.

Discussion

A change in the definition of “routine maintenance, repair, and
replacement” (RMRR) would likely remove a barrier to improvements in
power plant efficiency (or heat rate), which is generally perceived as a
positive outcome, with probable emission reduction implications.  Such
changes could also affect availability, as they mightif companies repair
or replace components which are likely sources of future breakdowns, and
capacity, as they mightif repairs enable a unit to increase its
electrical output with or without an increase in fuel consumption. 
Increases in availability and capacity would likely lead to an emissions
increase, if applied uniformly across all coal units.  Hence, from an
emissions perspective, there is a tension between pure efficiency
improvements, and availability or capacity improvements.

Availability changes are perhaps the easiest to consider with available
historic data.  A substantial body of data accumulated by the North
American Electric Reliability Council (NERC) shows that, over the past
two decades, reliability at coal-fired power plants has improved, but at
a declining rate in recent years.  The data in Figure 1 suggest that
future improvements in availability will be modest compared to
improvements in the 1980’s, for example.  A cumulative future
improvement of 1% seems like a most likely scenario, but an improvement
of 2% may be possible.  These two values are recommended for
assessment.evaluated in this analysis.

Possible efficiency improvements have been discussed adequatelywere
discussed in the draft RIA.  They apply almost equally to the “boiler
side” of the power plant and the “turbine side”.  Improvements of
up to 15% seem feasible with known or emerging technologies.  Earlier
modeling by DOE evaluated rates of 5%, 10%, and 15%.  Because
improvements of 15% were essentially extensions of the trends
established by the lower efficiency improvement assumptions, it is
recommended thatthis analysis focuses on improvements of 5% and 10% be
assessed.  A greater efficiency improvement is still deemed reasonable,
given the ongoing voluntary programs to minimize carbon dioxide
emissions as well as the potential for fuel savings and lower overall
operational costs, but the emission implications of such efficiency
improvements can be inferred from simulations of efficiency improvements
of 5% and 10%.

The significant change in this set of simulations by DOE, compared to
DOE’s earlier analysis, is the inclusion of an assumed improvement in
capacity, where efficiency is improved.  (EPA’s simulations, included
in the draft RIA, incorporated capacity increases, but DOE’s did not.)
 DOE discussed this issue with knowledgeable individuals involved in
power generation and reached the following conclusions:

Most older power plants in the U.S. were designed with excess capacity
in the turbine unit, compared to the boiler systems.  Newer plants in
the U.S. are designed in a more balanced manner.  Typical steam turbines
in the 1960’s, for example, would have the ability to generate about
10% more electricity than the boiler could support with steam
production, if the turbine were in optimum condition. 

 

Many existing unitsboilers operate outside their original design
specifications, due to use of lower sulfur coal, application of low-NOx
burners, or other changes which have left the boilers with insufficient
auxiliary equipment, such as pulverizers, to maintain original design
steam production.  These changes Such changes at power plants have
impacted the boiler-side of the power plant more than the steam
turbine-side.

Steam turbines tend to degrade gradually over time.  In terms of
efficiency and capacity, the most important changes appear to be
increased steam leakage through seals (thereby avoiding rotating turbine
blades), and blade erosion, which can change the angle of impact of
steam on blades and reduce the force imparted by the steam on the
blades.  In terms of reliability, turbine blades can suffer from
corrosion and metal fatigue, both of which can lead to blade breakage,
which can lead to serious injury and equipment damage.

To the extent that most steam turbines are thought to have greater
capacity than the associated boiler can support, it is reasonable to
conclude that a turbine efficiency upgrade would extend capacity only to
the extent that the boiler was also upgraded.  The capacity of a given
boiler relative to its associated steam turbine is a site-specific
relationship, and probably varies over time.  However, it is clear that
if a boiler’s capacity is either greater than or equal to that of its
associated steam turbine, any capacity increases in the boiler (which
might accompany a boiler efficiency improvement) cannot be exploited
unless the turbine increases in capacity as well.  Since relatively
equivalent efficiency improvements are envisioned for both the boiler
and turbine systems at a power plant, the net effect on capacity (for
improving both boilers and turbines) should be about one-half the
improvement in efficiency.This phenomenon suggests that overall capacity
improvements resulting from equipment replacement will be somewhat
smaller than overall efficiency improvements.  At this point, there is
insufficient information to determine the exact difference.  Earlier
simulations by DOE assumed no improvement in capacity from efficiency
improvements, and earlier EPA scenarios assumed capacity improvements
equaled efficiency improvements.  The simulations which are described in
the current analysis make an assumption in between the earlier analyses:
 that capacity improvements will equal one-half the assumed efficiency
improvement.  This is the assumption recommended for assessment.

Scenario Matrix

Table 1 identifies assumptions used in the current computer analysis. 
Assessments of efficiency improvements of 15% are not necessary because
these results can be inferred from the 5% and 10% simulations, and
earlier runs.

Table 1.

Simulation Run	Heat Rate & Capacity Increase	Availability Increase

Base	0	0

S5.1	5% & 2.5%	1%

S5.2	5% & 2.5%	2%

S10.1	10% & 5%	1%

S10.2	10% & 5%	2%





Figure   SEQ Figure \* ARABIC  1   SEQ Figure \* ARABIC   

Results

The revised assumptions led to modest changes from the projected
emission baselines.  Figures 2 and 3 show the projections for emissions
of nitrogen oxides (NOx) and sulfur dioxide (SO2).  The efficiency,
capacity, and availability improvements for each simulation are
indicated in the legends on the figures.  For example, “NOx 5-2.5-1”
is a scenario for NOx emissions in which efficiency increases 5%,
capacity increases 2.5%, and availability increases by 1%.  For SO2, it
was surprising that emissions changed at all, given the national limit,
or “cap”, on power plant SO2.  The small projected changes are due
to possible “banking” behavior, in which source owners control more
in some years and less in others to minimize anticipated control costs. 
Over time, all of these scenarios net out to the same cumulative SO2
emission reductions.  For the scenarios simulated, NOx emissions are
consistently generally less than baseline emissions, although the
reduction is quite minor for the simulations assuming only 5%
improvements in efficiencyheat rate., and in four of the years
projected, national emissions are slightly above baseline for one
scenario.  Table 2 presents NOx results for selected years.

Figure 4 shows projected emissions of mercury under all of the scenarios
modeled.  Emissions under a more flexible NSR policy are consistently
lower than baseline emissions, with the exception of certain years in
the case for 5% efficiency, 2.5% capacity, and 2% availability
improvement.

In general, for the scenarios including 10% improvements in heat rates,
non-capped pollutant emissions were reduced by about 2-5%, relative to
the base case (current NSR policy).  For scenarios including 5%
improvements in heat rates, non-capped emissions were reduced by 0.5 –
2%.  For the lesser set of efficiency improvement simulations, there
were certain combinations of assumptions which led to small emission
increases in specific years.

Table 2.  NOx emission changes for each scenario.

Scenario	Assumed % change in	NOx Emissions, million tons per year

	Heat Rate	Capacity	Availability	2005	2010	2015	2020	2025

Baseline	0	0	0	       3.60 	       3.93 	       3.98 	       4.05 	     
 4.11 

S 5.2.1	-5	2.5	1	       3.59 	       3.88 	       3.93 	       4.01 	   
   4.08 

S 5.2.2	-5	2.5	2	       3.60 	       3.90 	       3.98 	       4.05 	   
   4.12 

S 10.5.1	-10	5	1	       3.55 	       3.75 	       3.83 	       3.91 	   
   3.97 

S 10.5.2	-10	5	2	       3.56 	       3.78 	       3.87 	       3.95 	   
   4.02 



Overall cost savings can be estimated by making a conservative estimate
that the cost of improvements total $100 per kilowatt of capacity (this
is substantially greater than preliminary estimates developed by DOE).  
Cumulative projected savings, assuming this level of investment in
replaced components, are presented in Figure 7.   These totals range
from about $10 billion to $100 billion through the year 2025,
considering only the change in electricity prices.  Secondary benefits,
not quantified, would likely evolve from reduced consumption of natural
gas.  In the scenarios simulated, natural gas use for power production
typically declined by 5 -14%.  Such a reduction would reduce the price
of natural gas in all sectors, leading to additional consumer savings. 

Figure 2  SEQ Figure \* ARABIC   

Figure 3  SEQ Figure \* ARABIC   

Figure 4  SEQ Figure \* ARABIC   

Figure 5

Related Considerations

The projections in this analysis do not include consideration of:

Heat rate (efficiency) improvements exceeding 10%

Improvements occurring to a subset of coal units, such as units above a
certain size or less than a certain age

Potential improvements to similar technologies not burning coal

In general, all of these considerations would tend to change the extent
of predicted changes, but not the direction.  Additionally, although
heat rate improvements as low as 5%, and availability improvements as
high as 2% are deemed possible scenarios, a scenario of 10% heat rate
improvement, 5% capacity improvement, and 1% availability improvement is
considered a central tendency for the industry under a more flexible NSR
policy.this combination of assumptions is considered less likely than
other combinations examined.

Conclusions

Future emissions from power plants, aggregated at a national level, are
projected to decrease under a broad set of scenariosthe range of
scenarios examined in this analysis, which are predicated on an NSR
policy whichthat encourages equipment replacement that leadings to
higher efficiency.  With minimal (5%) improvements in efficiency,
emissions decrease 0-1%.  With efficiency improvements of 10%, emissions
decrease 2-5%.  In addition to reductions in emissions, substantial
savings in the cost of electricity are possible through these
improvements, totaling between $10 billion and $100 billion (in 2001
dollars) by the year 2025.Under a “most likely” scenario, emissions
of pollutants which are not “capped” by regulation would decrease by
roughly 2-5%.

Note to:	Files

From:		Doug Carter, DOE/FE-26

Subject:	Analysis of NSR rule impacts

Date:		August 21, 2003

Introduction

Both EPA and DOE/EIA have modeled potential impacts of allowing
efficiency upgrades on coal-fired power plants, and results of these
simulations were included in the draft Regulatory Impact Analysis (RIA)
published by EPA.  The major differences in the two analyses were that
they used different simulation models (NEMS vs IPM), and they evaluated
somewhat different assumptions for how the change at the power plant
might manifest in operation (i.e., assumptions on efficiency, capacity,
and availability changes).  After additional consideration, DOE
conducted further analysis, using assumptions differing from both of the
earlier analyses.

Discussion

A change in the definition of “routine maintenance, repair, and
replacement” (RMRR) would likely remove a barrier to improvements in
power plant efficiency (or heat rate), which is generally perceived as a
positive outcome, with probable emission reduction implications.  Such
changes could also affect availability, if companies repair or replace
components which are likely sources of future breakdowns, and capacity,
if repairs enable a unit to increase its electrical output with or
without an increase in fuel consumption.  Increases in availability and
capacity would likely lead to an emissions increase, if applied
uniformly across all coal units.  Hence, from an emissions perspective,
there is a tension between pure efficiency improvements, and
availability or capacity improvements.

Availability changes are perhaps the easiest to consider with historic
data.  A substantial body of data accumulated by the North American
Electric Reliability Council (NERC) shows that, over the past two
decades, reliability at coal-fired power plants has improved, but at a
declining rate in recent years.  The data in Figure 1 suggest that
future improvements in availability will be modest compared to
improvements in the 1980’s, for example.  A cumulative future
improvement of 1% seems like a most likely scenario, but an improvement
of 2% may be possible.  These two values are evaluated in this analysis.

Possible efficiency improvements were discussed in the draft RIA.  They
apply almost equally to the “boiler side” of the power plant and the
“turbine side”.  Improvements of up to 15% seem feasible with known
or emerging technologies.  Earlier modeling by DOE evaluated rates of
5%, 10%, and 15%.  Because improvements of 15% were essentially
extensions of the trends established by the lower efficiency improvement
assumptions, this analysis focuses on improvements of 5% and 10%.  A
greater efficiency improvement is still deemed reasonable, given the
ongoing voluntary programs to minimize carbon dioxide emissions as well
as the potential for fuel savings and lower overall operational costs,
but the emission implications of such efficiency improvements can be
inferred from simulations of efficiency improvements of 5% and 10%.

The significant change in this set of simulations by DOE, compared to
DOE’s earlier analysis, is the inclusion of an assumed improvement in
capacity, where efficiency is improved.  (EPA’s simulations, included
in the draft RIA, incorporated capacity increases, but DOE’s did not.)
 DOE discussed this issue with knowledgeable individuals involved in
power generation and reached the following conclusions:

Most older power plants in the U.S. were designed with excess capacity
in the turbine unit, compared to the boiler systems.  Newer plants in
the U.S. are designed in a more balanced manner.  Typical steam turbines
in the 1960’s, for example, would have the ability to generate about
10% more electricity than the boiler could support with steam
production, if the turbine were in optimum condition. 

Many existing boilers operate outside their original design
specifications, due to use of lower sulfur coal, application of low-NOx
burners, or other changes which have left the boilers with insufficient
auxiliary equipment, such as pulverizers, to maintain original design
steam production.  Such changes at power plants have impacted the
boiler-side of the power plant more than the steam turbine-side.

Steam turbines tend to degrade gradually over time.  In terms of
efficiency and capacity, the most important changes appear to be
increased steam leakage through seals (thereby avoiding rotating turbine
blades), and blade erosion, which can change the angle of impact of
steam on blades and reduce the force imparted by the steam on the
blades.  In terms of reliability, turbine blades can suffer from
corrosion and metal fatigue, both of which can lead to blade breakage,
which can lead to serious injury and equipment damage.

The capacity of a given boiler relative to its associated steam turbine
is a site-specific relationship, and probably varies over time. 
However, it is clear that if a boiler’s capacity is either greater
than or equal to that of its associated steam turbine, any capacity
increases in the boiler (which might accompany a boiler efficiency
improvement) cannot be exploited unless the turbine increases in
capacity as well.  This phenomenon suggests that overall capacity
improvements resulting from equipment replacement will be somewhat
smaller than overall efficiency improvements.  At this point, there is
insufficient information to determine the exact difference.  Earlier
simulations by DOE assumed no improvement in capacity from efficiency
improvements, and earlier EPA scenarios assumed capacity improvements
equaled efficiency improvements.  The simulations which are described in
the current analysis make an assumption in between the earlier analyses:
 that capacity improvements will equal one-half the assumed efficiency
improvement.  

Scenario Matrix

Table 1 identifies assumptions used in the current computer analysis. 
Assessments of efficiency improvements of 15% are not necessary because
these results can be inferred from the 5% and 10% simulations, and
earlier runs.

Table 1.

Simulation Run	Heat Rate & Capacity Increase	Availability Increase

Base	0	0

S5.1	5% & 2.5%	1%

S5.2	5% & 2.5%	2%

S10.1	10% & 5%	1%

S10.2	10% & 5%	2%



Figure   SEQ Figure \* ARABIC  1 

Results

The revised assumptions led to modest changes from the projected
emission baselines.  Figures 2 and 3 show the projections for emissions
of nitrogen oxides (NOx) and sulfur dioxide (SO2).  The efficiency,
capacity, and availability improvements for each simulation are
indicated in the legends on the figures.  For example, “NOx 5-2.5-1”
is a scenario for NOx emissions in which efficiency increases 5%,
capacity increases 2.5%, and availability increases by 1%.  For SO2, it
was surprising that emissions changed at all, given the national limit,
or “cap”, on power plant SO2.  The small projected changes are due
to possible “banking” behavior, in which source owners control more
in some years and less in others to minimize anticipated control costs. 
Over time, all of these scenarios net out to the same cumulative SO2
emission reductions.  For the scenarios simulated, NOx emissions are
generally less than baseline emissions, although the reduction is quite
minor for the simulations assuming only 5% improvements in efficiency,
and in four of the years projected, national emissions are slightly
above baseline for one scenario.  Table 2 presents NOx results for
selected years.

Figure 4 shows projected emissions of mercury under all of the scenarios
modeled.  Emissions under a more flexible NSR policy are consistently
lower than baseline emissions, with the exception of certain years in
the case for 5% efficiency, 2.5% capacity, and 2% availability
improvement.

In general, for the scenarios including 10% improvements in heat rates,
non-capped pollutant emissions were reduced by about 2-5%, relative to
the base case (current NSR policy).  For scenarios including 5%
improvements in heat rates, non-capped emissions were reduced by 0.5 –
2%.  For the lesser set of efficiency improvement simulations, there
were certain combinations of assumptions which led to small emission
increases in specific years.

Table 2.  NOx emission changes for each scenario.

Scenario	Assumed % change in	NOx Emissions, million tons per year

	Heat Rate	Capacity	Availability	2005	2010	2015	2020	2025

Baseline	0	0	0	       3.60 	       3.93 	       3.98 	       4.05 	     
 4.11 

S 5.2.1	-5	2.5	1	       3.59 	       3.88 	       3.93 	       4.01 	   
   4.08 

S 5.2.2	-5	2.5	2	       3.60 	       3.90 	       3.98 	       4.05 	   
   4.12 

S 10.5.1	-10	5	1	       3.55 	       3.75 	       3.83 	       3.91 	   
   3.97 

S 10.5.2	-10	5	2	       3.56 	       3.78 	       3.87 	       3.95 	   
   4.02 



Overall cost savings can be estimated by making a conservative estimate
that the cost of improvements total $100 per kilowatt of capacity (this
is substantially greater than preliminary estimates developed by DOE).  
Cumulative projected savings, assuming this level of investment in
replaced components, are presented in Figure 7.   These totals range
from about $10 billion to $100 billion through the year 2025,
considering only the change in electricity prices.  Secondary benefits,
not quantified, would likely evolve from reduced consumption of natural
gas.  In the scenarios simulated, natural gas use for power production
typically declined by 5 -14%.  Such a reduction would reduce the price
of natural gas in all sectors, leading to additional consumer savings. 

Figure 2

Figure 3

Figure 4

Figure 5

Related Considerations

The projections in this analysis do not include consideration of:

Heat rate (efficiency) improvements exceeding 10%

Improvements occurring to a subset of coal units, such as units above a
certain size or less than a certain age

Potential improvements to similar technologies not burning coal

In general, all of these considerations would tend to change the extent
of predicted changes, but not the direction.  Additionally, although
heat rate improvements as low as 5%, and availability improvements as
high as 2% are deemed possible scenarios, this combination of
assumptions is considered less likely than other combinations examined.

Conclusions

Future emissions from power plants, aggregated at a national level, are
projected to decrease under the range of scenarios examined in this
analysis, which are predicated on an NSR policy that encourages
equipment replacement leading to higher efficiency.  With minimal (5%)
improvements in efficiency, emissions decrease 0-1%.  With efficiency
improvements of 10%, emissions decrease 2-5%.  In addition to reductions
in emissions, substantial savings in the cost of electricity are
possible through these improvements, totaling between $10 billion and
$100 billion (in 2001 dollars) by the year 2025. 

 Telecommunication between D. Carter, DOE, and R. Moates, TVA, July 29,
2003.  Telecommunication between D. Smith, DOE, and J. Murphy, SAIC,
July 29, 2003.

 Telecommunication between D. Carter, DOE, and R. Moates, TVA, July 29,
2003.  Telecommunication between D. Smith, DOE, and J. Murphy, SAIC,
July 29, 2003.

