EPA-450/3-78-002
THE ABILITY
OF ELECTRIC UTILITIES
WITH FGD TO MEET
ENERGY DEMANDS
by
Dr. E.P. Hamilton, III
Radian Corporation
8500 Shoal Creek
Austin, Texas 78766
Contract No. 68-02-2608
EPA Project Officer: Kenneth R. Durkee
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air and Waste Management
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
January 1978
-------
This report is issued by the Environmental Protection Agency to report
technical data of interest to a limited number of readers. Copies are
available free of charge to Federal employees, current contractors and
grantees, and nonprofit organizations - in limited quantities - from the
Library Services Office (MD-35) , U.S. Environmental Protection Agency,
Research Triangle Park, North Carolina 27711; or, for a fee, from the
National Technical Information Service, 5285 Port Royal Road, Springfield,
Virginia 22161.
This report was furnished to the Environmental Protection Agency
by Radian Corporation, 8500 Shoal Creek, Austin, Texas 78766, in
fulfillment of Contract No. 68-02-2608. The contents of this report
ai-e reproduced herein as received from Radian Corporation. The
opinions, findings, and conclusions expressed are those of the author
and not necessarily those of the Environmental Protection Agency.
Mention of company or product names is not to be considered as an
endorsement by the Environmental Protection Agency.
Publication No. EPA-450/3-.78-002
11
-------
ACKNOWLEDGEMENT S
Recognition should be given to the following Radian
personnel who were partners in this research effort and also
co-authored this report:
Dr. H. J. Williamson
Dr. J. B. Riggs
Miss T. J. Anderson
and to Mr. Kenneth R. Woodard, who acted as co-project officer
with Mr. Durkee.
We also wish to convey our sincere appreciation to
the following people who provided data or aided in the data-
gathering for this study:
Mr. Walter Brown, Executive Vice President, National
Electric Reliability Council
Mr. Jerry Weiser, Edison Electric Institute
Mr. Jerry Albert, East Central Area Reliability
Coordination Agreement
Mr. Grady Smith, Vice President, Southern Company
Services
Mr. Harold Tynan, Executive Secretary, Electric
Reliability Council of Texas
Mr. Floyd Curran, Mid-America Interpool Network
Mr. Bill Hulsey, Executive Director, Southwestern
Power Pool
Mr. Dennis Eyre, Administrative Manager, Western
Systems Coordinating Council
Mr. Elof Soderberg, Chief Engineer, Lower Colorado
River Authority
Mr. R. L. Hancock, Director, City of Austin Utilities
Mr. Tom Sweatman, Chief Engineer, Texas Public
Utilities Commission
Mr. Dennis Haverlaugh, Texas Air Control Board
Mr. Norton Savage, Federal Power Commission
Mr. Joe Flood, Federal Power Commission
We also convey our thanks to the National Electric
Reliability Council for their granting permission to reproduce
certain copyrighted material in this report.
-------
CONTENTS
Acknowlec
Contents-
Figures--
Tables
Summary- -
1.
2.
3.
4.
Igements
Introduction ;
Conclusions "
Description of Problem
Determination of Effect of FGD on Future
Generation Requirements
4.1 Determination of Generation Forecasts
4.2 Demand Characteristics :
4.3 Determination of Mix of Base-Load and
Intermediate-Load New Coal Units
4.4 Outage Rates
4.5 FGD Energy Penalties
4.6 Methodology for Calculation of Incremental
T.nss-n-F-T.nad Probabilities
111
IV
vi
vorii
X
1
8
15
18
18
23
30
36
39
41
4.6.1 Calculation of Distribution of
Available
Unit
Power at a Particular
4.6.2 Loss-of-Load Probability for a
Single Generating Unit
4.6.3 Expected Effect of FGD on the
Reliability of a Single Unit
4.6.4 Incremental Loss-of-Load Probability
for a Power System
4.6.5 Expected Effect of FGD on System
Reliability
Evaluation and Interpretation of Results
5.1 Effect of FGD on Additional Capacity
Requirements
5.1.1 Effect of FGD on Reliability - Units
On-line Prior to 1986
41
43
44
46
51
54
55
55
IV
-------
CONTENTS--Continued
References
Appendices
5.1.2 Effect of FGD on Reliability -
.2
. 3
5.4
Units On-line Between 1985 and 2000 --
Evaluation of Interchange Constraints
Evaluation of Utility Reserves
Sensitivity of Results to Projected Demand
and Generation Mix Statistics
57
fi ^
D J
7 A
/ o
Q7
y I
irn
A. Calculation of Incremental Loss-of-Load
Probability for New Coal Generation in a
Power System- 105
B. Estimated Load Factor Calculations by Region 121
C. Incremental Loss-of-Load Probabilities for
New Coal Only by Region in 2000 - Test Case
Results ' 131
-------
FIGURES
Number Pagt
1 Division of U.S. into NERC Reliability
Regions < 3
2 Division of U.S. into NERC Reliability
Regions '15
3 Typical Weekly Load Curve 24
4 Typical Weekly Load Curve Apportioned by
Prime Mover Type 25
5 Expected Load Duration Curve for New
Base-Load Coal Unit-' 28
6 Expected Load Duration Curve for New
Intermediate-Load Coal Unit 29
7 1990 Expected Load Duration Curve with
Typical Apportionment b y Prime Mover Type
for a Portion of NPCC (Region 6) 31
8 1990 Expected Load Duration Curve with
Typical Apportionment by Prime Mover Type
for a Portion of SERC (Region 7) 32
9 Non-Simultaneous Emergency Transfer
Capabilities (MWe) - WSCC- 68.
10 Non-Simultaneous Emergency Transfer
Capabilities (MWe) - ERGOT 69
11 Non-Simultaneous Emergency Transfer
Capabilities (MWe) - MARCA 70
12 Non-Simultaneous Emergency Transfer
Capabilities (MWe) - SPP 72
13 Non-Simultaneous Emergency Transfer
Capabilities (MWe) - MAIN-- 73
14 Non-Simultaneous Emergency Transfer
Capabilities (MWe) - SERC : 74
15 Non-Simultaneous Emergency Transfer
Capabilities (MWe) - NPCC : .75
16 Non-Simultaneous Emergency Transfer
Capabilities (MWe) - ECAR 76
17 Non-Simultaneous Emergency Transfer
Capabilities (MWe) - MAAC ' 77
18 Reserve Requirements as a Function of
Unit Size 81
vx
-------
FIGURESContinued
Number FaSe
19 NERC Large Unit Construction Forecast - U.S.
through 1986 84
20 NERC Reserve Forecast for Contiguous U. S. 85
21 NERC Reserve Forecast for Region 9 - WSCC 86
22 NERC Reserve Forecast for Region 5 - MARCA : 88
23 NERC Reserve Forecast for Region 8 - SPP 89
24 NERC Reserve Forecast for Region 7 - SERC 90
25 NERC Reserve Forecast for Region 4 - MAIN 91
26 NERC Reserve Forecast for Region 2 - ERGOT 92
27 NERC Reserve Forecast for Region 1 - ECAR 93
28 NERC Reserve Forecast for Region 3 - MAAC 94
29 NERC Reserve Forecast for Region 6 - NPCC 95
vx i
-------
TABLES
Number . Page
1 FGD System Configurations Studied 5
2 Mean Additional Capacity Requirements in
Percent of New Coal Capacity to Offset the
Reliability Effects of FGD 9
3 Total FGD-Related Additional Capacity
Requirement by Region in 2000 (MWe) 11
4 Total FGD-Related Additional Capacity
Requirements by Region in 2000 - 600 MWe Coal
Units 12
5 Expected Additional Generation Requirements
Due only to FGD Energy Penalties by Region - 2000- 13
6 Expected Generation to be Added (by Reliability
Region) Prior to 1986 19
7 Expected Generation to be Added (by Reliability
Region) 1986-2000 20
8 Estimated Mix of New Base and Intermediate
Load Coal Units in 2000 35
9 National Outage Rates for Large, Mature Coal-Fired
Electric Generating Stations - 390-599 MWe Units-- 37
10 National Outage Rates for Large, Mature Coal-Fired
Electric Generation Stations - 600+ MWe Units 38
11 Expected Energy Penalties from FGD Use 40 ,
12 Effect of FGD on the Reliability of a Single
Generating Unit 45
13 Mean Additional Capacity Requirements in Percent
of New Coal Capacity to Offset the Reliability
Effects of FGD- 52
14 Estimated Scrubber Additions before 1986 56
15 Additional Capacity Requirements in MWe to Offset
FGD Reliability Only in 2000 59
16 Expected Additional Generation Requirements Due
Only to FGD Energy Use Penalties by Region in ,
2000 60
17 Total FGD-Related Additional Capacity Requirement
by Region in 2000 - MWE 61
18 Total FGD-Related Additional Capacity Requirements
by Region in 2000.- 600 MWe Coal Units 62
V113.
-------
TABLES--Continued
Number Page
19 Additional Generation Required from the Use of
Scrubbers as a Proportion of Total New Coal
Capacity 101
20 Incremental Loss-of-Load Probabilities (ILOLP)
for Worst Case (5 Active Modules/No Spares;
r = 0.8) Load Factor Sensitivity Test 101
21 Incremental Loss-of-Load Probabilities (ILOLP)
for Best Case (5 Active Modules/1 Spare; r = 0.9)
Load Factor Sensitivity Test 102
3.X
-------
SUMMARY
A more stringent New Source Performance Standard (NSPS)
for S02 emissions is presently being considered. The implementa-
tion of this NSPS will require utilities to employ flue gas de-
sulfurization (FGD) or equivalent S02' removal techniques on all
new coal-fired boilers. The addition of FGD systems to all new
coal-fired generating units will reduce a utility's ability to meet
consumer demand because of reduced unit and system reliability and
increased in-plant energy consumption. The FGD system can thus
affect both generating unit and utility system adequacy. A re-
duction in individual utility system reliability may affect power
pool reliability as well. Therefore, it is important to determine
the overall impact of the widespread implementation of FGD on
electric utilities.
This study evaluated the overall impact of FGD systems
on U.S. electric reliability and adequacy through 2000. Coal-
fired units on-line before 1986 and between 1985 and 2000 were
considered separately for each of the nine National Electric
Reliability Council (NERC) reliability regions. Different
generation mixes and typical demand characteristics for each
region were calculated. With this information, an assessment of
the ability of each region to meet power demand (and to maintain
reasonable and typical reserves) as a power pool with and with-
out FGD was made. Several different FGD module configurations
and expected module availabilities were considered. It was
assumed that a generating unit would be required to be temporarily
derated upon FGD module failure if a spare module were not avail-
able. Bulk power interchange capabilities which might be used in
the event of such FGD-induced outages or reductions in capacity
were also evaluated, as were expected system reserves.
-------
This study concluded that the proposed NSPS would have
little effect on system reliability and adequacy prior to 1985
because of long lead times required for the construction of new
coal-fired generating units.
However, for the period between 1985 and 2000 it was
found that the proposed standard would have a significant effect
on system reliability and adequacy.
For four different FGD modular availabilities and con-
figurations, nationwide.additional capacity requirements in 2000
due to the impact of FGD availability and energy requirements
were found to range from approximately 38,000 megawatts to 105,000
megawatts.
The additional capacity requirement due to FGD is the
result of two factors: (1) the increase in in-plant power con-
sumption required to operate the FGD units (an "energy penalty")
and (2) the increase in capacity needed to offset a decrease in
^system reliability. The increase in capacity to offset the
energy -penalty associated with FGD is approximately 28,000 mega-
=»=-=
watts for the 1985-2000 period; the increase associated with
reliability ranges from approximately 9,000 megawatts to 77,000
megawatts. These estimates do not include the.effects of any FGD
systems on the additional capacity required. Total new coal
capacity that is to go on-line between 1985 and 2000 was projected
by Teknekron to be approximately 534,000 megawatts. It should be
noted that the upper limit of the range qf these,estimates
associated with reliability, 77,000 additional megawatts, assumes
an FGD modular availability of 80%, no redundance in FGD modules
to provide backup capability during FGD module outages, nor any
other mitigating factors.
-------
This report was submitted in fulfillment of Contract
68-02-2608 by Radian Corporation under the sponsorship of the
U.S. Environmental Protection Agency. This report covers the
period January 1, 1978, to February 28, 1978, and work was com-
pleted as of February 28, 1978.
XI i
-------
SECTION 1
INTRODUCTION
This report presents the results of work performed by
Radian Corporation of Austin, Texas, for the Office of Air
Quality Planning and Standards, Emission Standards and Engineer-
ing Division, of the United States Environmental Protection
Agency. This project assessed the impact of flue gas desulfuri-
zation (FGD) systems on the overall reliability .and expected
generation requirements of electric power generating systems in
the United States through the year 2000.
Power system reliability is directly affected by the
availability.of generating units. Since FGD affects this unit
availability, it also affects system reliability. For the purposes
of this report, the following definitions of reliability and
availability will be used:
Reliability - the ability of a power system to provide
continuity of service19. For generating units alone
(neglecting the effects of the transmission and
distribution networks), reliability can be measured
as the probability that demand exceeds available
capacity*. In this context, this probability can
be described by the loss-of-load probability
(LOLP), which is a well-known and accepted index
of reliability in the utility industry.
Availability - the amount of time a generating unit is
capable of operating at a particular power output (also
called-operating availability). Numerically, avail-
ability is expressed as the ratio of the number of
hours the unit is available to the number of hours in
Technically, this particular measure of reliability as given is
equivalent to (1.0- probability that service is continuous).
-------
the period7. Reliability in terms of LOLP or (1.0 -
LOLP) can be directly calculated from availability
using probabilistic relationships (see Appendix A).
Consequently, the impact of FGD on power system reli-
ability can be estimated by using probabilistic methods and by
considering any other pertinent factors. The methodology
to estimate this impact includes consideiration of the following
five items :
1. A determination of the projected generation types;
quantities, and typical demand characteristics for
new coal-fired generating units. (See Sections
4.1, 4.2, and 4.3.)
2. An assessment of the ability of each power pool
(or reliability council) to meet power demands in
2000 with and without scrubbers through the
calculation of LOLP. (Scrubber configura-
tions and availabilities were assumed for
this study.) (See Sections 4.6.4 and 4.6.5.)
3. A determination of the effects, if any, of scrubbers
on the ability to maintain reasonable and typical
reserves, and a calculation of the amount of any
additional capacity which must be added to offset
any scrubber impacts. (See Section 5.1 and 5.3.)
4. An investigation of nationwide reserves as to gen-
eration types and amounts. (See Section 5.3.)
5. An assessment of bulk interchange capabilities
which might be used in the event of outages
caused by FGD. (See Section 5.2.)
-------
This study addressed the problem of large-scale im-
plementation of FGD. on both a regional and national basis. The
wide-spread practice of interconnecting electric utilities
("power pooling") has led to the division of the country into
nine major power pools. These pools and their respective ser-
vice areas are .shown in Figure 1. With the formation of the
1 - ECAR - Eaat Central Area Reliability Coordination Agreement
2 - ESCOT - Electric Reliability Council of Texas
3 - MAAC - Mid-Atlantic Area Council
i* - MAIS - Mid-America Interpool Setwork
5 - MAS.CA - Mid-Continent Area Reliability Coordination Agreement
6 - HPCC - Northeast Power Coordinating Council
7 - SESC - Southeastern Electric Reliability Council
3 ,- SPP - Southwestern Power Pool
9 - WSCC - Western Systems Coordinating Council
Figure 1. Division of U.S. into NERC Reliability Regions
3
-------
National Electric Reliability Council (NERC), these nine power
pools became reliability regions which report to NERC on an
individual basis. Each pool has its own particular mix of
generation, load or demand characteristics, operating practices,
reserve requirements, and interchange agreements. The effects
of FGD considered in this study are thus addressed for each of
the nine reliability regions and for the nation as a whole.
Almost all existing commercial applications of flue
gas desulfurization on coal-fired boilers use either the lime
process or the limestone process and almost all the data
available at this time are for these processes. This study
utilized estimates or assumptions of performance for limestone
processes and it assumed that (1) all new FGD units, or scrubbers,
are of the limestone type, and that (2) each scrubber would have
an energy requirement necessary to achieve 90 percent removal
efficiency.9 The five scrubber configurations shown in Table 1
were studied. It was assumed that all new coal units for each
case had FGD systems which were configured identically; i.e.,
only module throughput capacity varied depending on the size of
the generating unit. While other modular configurations are
certainly possible for various unit sizes, only these five were
considered to allow for ease in computation. Furthermore,
although smaller generating units (less than 600 megawatts)
might require fewer scrubber modules, it was found that these
units most probably will be a small portion of total generating
capacity by the year 2000. Hence, it is expected that any
variations in scrubber configurations among these smaller units
would not significantly affect the results of this study. In
addition, this study did not address either the validity of
available data, the desirability of the limestone process, or
the technical feasibility of achieving various removal
efficiencies or various FGD module availability levels.
-------
.Table 1. FGD SYSTEM CONFIGURATIONS STUDIED
Case
1
(Base Case)
2"
3
4
5
Number
Scrubbers
per
Generating
Unit
0
1
1
1
1
Number
Modules
per
Scrubber
5
5
5
5
Number
Spares
Modules
0
1
0
1
Availability
per
Module
'907=
90%
80%
80%
Outage rate = 100% - Availability
Power plant outage data (exclusive of scrubbers) were
taken from Edison Electric Institute (EEI) data accumulated
over the period 1972 - 1974 (3 years) for large mature units
built before 1971. Only units which burned coal exclusively
were considered, and total outage rates (forced plus scheduled
full and partial outage) were considered. These data covered
units which were grouped into two categories.- 390-599 MWe*
capacity and capacity greater than 600 MWe. Consequently, this
study considered only those units in these two size ranges;
however, the vast majority of planned,-generating facilities are
in these, two groups.
* Electric megawatt output at the bus bar; also called MW by
the industry. In this report MWe will be used to avoid con-
fusion with thermal boiler input megawatts (MWt).
-------
These generating units were also grouped according to
probable duty. Typical power system practice has been to use
large new units for either intermediate (cycling) or base-load
duty. Load duration curves were developed to reflect before-
the-fact desired capacity factors of 70 percent for base-load
units and 55 percent for intermediate-load units. Anticipated
load duration coupled with expected additions by type and per-
cent capacity (coal, oil, nuclear, hydro, etc.) were used to
determine a reasonable expected mix of base-load and inter-
mediate-load coal units for each reliability region. This mix
was used to determine (1) overall ability to meet load and
(2) expected reserve generation types. Sensitivity of this
analysis was also estimated.
Because this study was done to provide data for the
evaluation of a revised New Source Performance Standard for
coal-fired power plants, two future time periods were con-
sidered. Because of the long lead times (7-10 years) required
for the construction of new coal-fired power plants, all units
scheduled for operation prior to 1986 were evaluated separately
in terms of any existing and/or potential FGD equipment. It was
found that any NSPS enacted today probably would not affect most
of these units. However, it was assumed that all coal-fired
units on-line between 1985 and 2000 would need FGD systems to
meet the revised NSPS.
Because of the short time constraints under which this
study was done, aggregate total reliability in terms of.the sum
of all types of generation which exist in each region (nuclear,
hydro, gas turbines, etc.) could not be evaluated. The same
demand curves for the two unit types considered were used for
all nine geographical regions. Different curves were used, of
course, for base and for intermediate units. The primary
-------
justification for using the same curves for all regions was
that region-to-region differences in demand per generating unit
were probably smaller than the uncertainty regarding the demand
curves which would apply more than a decade in the future.
Nevertheless, although the same unit demand curves were used for
all regions, the total demand varied by region. This was
accounted for in the analysis. Furthermore, it is felt that
correlations exist between demand at the same time at different
generating units in the same region. The demands at different
units, for example, are affected simultaneously by diurnal
variations and by area-wide weather changes. Treatment of these
correlations, however, was beyond the scope of this study. The
omission of the correlations affects the calculated incremental
loss-of-load probabilities to some extent, as is discussed in
Section 4.6.4 and Appendix A. It is believed, however, that
omission of the correlations does not have a large effect on the
decrease in power system reliability due to the.use of scrubbers.
Finally, in this study, only new coal units were analyzed, and
only that portion of the total power load they were expected to
carry was assessed. New coal units were the only type of unit
which were considered to require FGD systems. While a more
thorough analysis would certainly have been beneficial, it is
believed that the results presented in this study are sufficient
to estimate the effect of FGD systems on power system reliability
.and adequacy. The results can provide insight into the effects
'of FGD on different types of units with different duty cycles,
and they provide estimates of additional generation required to
offset the effects of implementing FGD systems on new coal units.
-------
SECTION 2
CONCLUSIONS
A detailed description of the results of this study
is contained in Section 5. From these results, the following
conclusions concerning a revised NSPS requiring FGD for all new
coal-fired electric generating units may be drawn:
1. Before 1985, the small numbers of committed FGD
systems required by a revised NSPS will have
very little effect on system reliability and
adequacy. Any revised NSPS enacted at the present
time will primarily affect new coal-fired units
anticipated to come on-line after 1985.
2. It was found that (1) FGD unavailability and (2)
energy penalties due to FGD cause reductions in
available system capacity leading to additional
capacity requirements. The mean percent reductions
in capacity caused only by FGD unavailability for
the four scrubber configurations in this study are
shown in Table 2. These mean percent reductions
were found to be virtually insensitive to changes
in mix of intermediate-load or base-load units or
load factors (Section 5.4). Estimated energy
penalties caused by increased in-plant energy
consumption for limestone scrubbers with 90 per-
cent removal were reported to be on the order of
3.4 to 3.8 percent of generating unit capacity9.
8
-------
t
Table 2. MEAN ADDITIONAL CAPACITY REQUIREMENTS IN PERCENT OF NEW
COAL CAPACITY TO OFFSET THE RELIABILITY EFFECTS OF FGD
Case
l(Base)
2
3
4
5
No. No. Availability
Modules Spares Per Module
0
5 0 90%
5 1 90%
5 0 80%
5 , 1 80%
Mean
; Additional
Generation Requirement
0
4.5%
1 . 2%
9.9%
4.4%
(Best)
(Worst)
a Expressed as a percent of new coal capacity without FGD. Does
not include boiler/turbine/generator availability effects.
3. The amount of additional capacity required due to
FGD was found to be very sensitive to scrubber
availability and modular configuration (i.e.,
number of modules, spares, etc.). Consequently,
if the values assumed for module availability were
too high or too low, the impact of FGD would either
be amplified or mitigated. Similarly, if the number
of modules and/or spares changed, or if bypass
were allowed, the estimated additional capacity
requirements would change.
4. It was found that the effects of FGD on reliability,
system reserves, and in-plant energy consumption
most probably would be offset by the addition of
more new units rather than by oversizing units
-------
which are already planned since system reliability
is degraded, not enhanced, by enlarging individual
sizes of a constant number of units. These new
units will probably be new coal units. ,
Assuming no spare FGD modules and a module avail-
ability of 80 percent (the worst case studied),
approximately 105,000 MWe would be required
nationally in 2000 to offset reduced system reli-
ability and increased in-plant power consumption
caused by FGD. The best case studied (one spare
module and 90 percent availability for each module)
requires about 37,500 additional MWe. The other
cases studied (90 percent, no spare; 80 percent,
1 spare) require about 63,000 MWe of additional
capacity. Additional results for each reliability
region are summarized in Tables 3, 4, and 5. As
can be seen from these tables, the impact on
individual regions varies considerably. These
generation estimates include mean boiler/turbine/
generator availability effects but do not include.
effects of any FGD on additional units.
It was found that, independent of any revised NSPS,
the electric utilities may encounter problems in
maintaining adequate reserves in the future due to
delayed construction of new plants. If future
reserves are marginal or inadequate, widespread
implementation of FGD would tend to compound this
problem. This impact results from boiler derating
when the unit is without bypass or available spare
scrubber modules and an FGD module or modules
fail.
10
-------
Table 3. TOTAL FGD-RELATED ADDITIONAL CAPACITY
REQUIREMENT BY REGION IN 2000a (MWe)
Region
1-ECAR
2 -ERGOT
3-MAAC
4-MAIN
5-MARCA
6-NPCC
7-SERC
8-SPP
9-WSCC
U.S. TOTAL
Case I
(base)
o
0
0
0
0
0
0
0
0
0
Case 2 ,
(5 70(390%) d
13,815
7,540
3,900
6,070
2,810
3,400
9,440
8,120
8,085
63,180
Case 3
(5/1(990%)
8,320
4,390
2,350
3,660
1,635
2,050
5,685
4,730
4,710
37,530
Case 4°
(5/0@80%)
22,805
12,690
6,435
10,025
4,725
5,615
15,580
13,665
13,610
105,150
Case 5
(5/1(980%)
13,650
7,440
3,850
6,000
2,770
3,360
9,325
8,015
7,980
62,390
alncludes reliability effects and energy penalties
Best Case
°Worst Case
(Number active modules /number spares (§ % modular availability)
11
-------
Table 4. TOTAL FGD-RELATED ADDITIONAL CAPACITY
REQUIREMENTS BY REGION IN 2000a -
600 MWe COAL UNITS
Region
1-ECAR
2 -ERGOT
3-MAAC
4-MAIN
5-MARCA
6-NPCC
7-SERC
8-SPP
9-WSCC
U.S. TOTAL
Case 1
(base)
0
0
0
0
0
0
0
0
0
0
Case 2
(5/0@90%)C
23
13
7
10
5
6
16
14
14
108
Case- 3a
(5/1(390%)
14
7
4
6
3
4
10
8
8
64
Case 4
(5/0@80%)
38
21
11
17
8
10
26
23
23
177
Case 5
(5/l@8070)
23
13
7
10
5
6
16
14
14
108
Includes reliability effects and energy penalties.
Best Case
°Worst Case
(Number active modules/number spares (§ % modular availability)
12
-------
Table 5. EXPECTED ADDITIONAL GENERATION REQUIREMENTS
DUE ONLY TO FGD ENERGY PENALTIES BY REGION - 2000
Maximum Additional Generation Requirements
Region Due to FGD Penalty - MWe
1-ECAR -6,325
2-ERCOT 3,245
3-MAAC 1,785
4-MAIN 2, 780
5-MARCA 1,210
6-NPCC 1,555
7-SERC 4,320
8-SPP 3,495
9-WSCC .3,480
U.S. TOTAL 28,200
alncludes only effects of new coal unit boiler/turbine/generator
availability. This assumes that the additional generation has
an average availability equal to that of a new coal unit.
13
-------
Because of a variety of technical and institutional
constraints on the interchange of power among
different regions, it was determined that bulk
power shipments probably could not significantly
reduce the additional capacity requirements caused
by the widespread implementation of FGD.
It was found that these effects of FGD on system
reliability and adequacy could be mitigated in a ;
number of ways. For example, redundancy or main-
taining spare modules can significantly reduce the
need to increase capacity to offset reduced reli-
ability. The use of alternative coal utilization
technologies which avoid or reduce the need for flue
gas cleanup is another obvious method for mitigating \
impacts. Finally, the adoption of a policy permitting
the bypass of scrubber modules on outage status when
reserve capacity is critically low or under other
temporary emergency conditions could also reduce
significantly the need to increase generating
capacity.
14
-------
SECTION 3
DESCRIPTION OF PROBLEM
The main objective of this study was to evaluate the
effect of FGD systems on the reliability and adequacy of electric
utility systems. The first step was to determine the incremental
reduction in system reliability due solely to scrubbers on new
coal plants on a regional basis. The nine NERC reliability
regions considered are shown in"Figure 2. Because of time con-
straints for this study, it was impossible to compute.total
system loss-of-load probability based on the interaction of all
1 - ECA& - Ease Central Area Reliability Coordination Agreement
2 - ERCOT - Electric Reliability Council of Texas
3 - MAAC - Mid-Atlantic Area Council
4 - MAIN - Mid-America Intsrpool Network
' 5 - MARCA - Mid-Continent Area Reliability Coordination Agreement
6 - HPCC - Hortheast Power Coordinating Council
7 - SERC - Southeastern Electric Reliability Council
3 - SPP - Southwestern Power Pool
9 - WSCC - Western Systems Coordinating Council
Figure 2. Division of U.S. -into NERC Reliability Regions
15
-------
forms of generation (coal, oil, nuclear, l^dro, etc.)- Instead,
this study assumed that the generation forecasts utilized re--
presented that capacity necessary to assure an acceptable system
loss-of-load probability without FGD on all new coal units, and
that the generation which had been planned or forecasted left
appropriate reserve margins. In this context, then, only new
coal units were considered in the calculation of incremental
loss-of-load probabilities (ILOLP) for new coal with and without
FGD. For the cases including FGD, additional generating require-
ments necessary to maintain the same incremental loss-of-load
probability as without FGD were computed for cases considering
the various FGD modular availabilities and configurations. This
calculation, therefore, assumed that none of the existing non-
new-coal capacity could be used to offset any effects of FGD. It
was also assumed that base and intermediates demand characteristics
for the new coal units were similar in each power pool. The
estimated mix of base- and intermediate-load new coal units was
then used to determine the aggregate demand for each region used
in the ILOLP calculations. The additional generating requirements
for the cases with FGD, determined from these calculations, were
then combined with expected additions in capacity required to off-
set energy penalties associated with FGD to give net additional
generating requirements due to FGD in each region. Since these
ILOLP calculations involved only new coal units on systems with
other forms of generation, reserve policies were investigated to
determine whether or not systems might be overbuilt or underbuilt
in the future, i.e., whether the assumptions above might be in-
fluenced by other factors. Constraints on emergency interchange
were also investigated. In addition, the effect of FGD on the
reliability of a single generating unit was also assessed.
The remainder of this report addresses the above points.
Section 4 discusses the determination of the effect of FGD on reli-
ability, adequacy and capacity based on generation mix, outage
16
-------
rates, demand characteristics, and energy penalties. Section 5
evaluates the results on reliability from Section 4, and
discusses the net effect of FGD on generation with respect to
additional capacity required, constraints on interchange and
reserve policies. The appendices contain a detailed mathematical
description of the probability calculations, a determination of
generation .mix for each region, and test case incremental loss-
of-load" probabilities.
17
-------
SECTION 4
DETERMINATION OF EFFECT OF FGD ON
FUTURE GENERATION REQUIREMENTS
To estimate accurately the effect on FGD on reliability
several intermediate calculations were made. The mix of power
generation was first determined for each power pool. Evaluation
of individual demand characteristics, outage rates, and energy
penalties for each area was completed next. The overall effect
of FGD was then evaluated for an individual unit and for all new
coal in each region. The effect of FGD on utility reliability and
adequacy was estimated separately for units put on-line before
1986 and for those on-line between 1985 and 2000.
4.1
Evaluation of Generation Forecasts
Because of differences in fuel resources, demand and
operating policy, each of the nine power pools generates power
using different proportions of various unit types. Since this
study assumed that each pool or region operates in a somewhat
autonomous fashion, it was decided that the effect of FGD could
be best estimated on a regional basis. Data developed by
Teknekron and supplied by EPA1 provided estimates of the number
and location of individual units planned for construction through
2010 and their expected or assumed* sizes in MWe. These data were
analyzed and plants were grouped by reliability region served so
that the expected mix of generation for each region could be
determined. Tables 6 and 7 show the mix for each region determined
from these Teknekron data for units put in service before 1986 and
from 1985 to 2000.
*Teknekron assumed sizes of 600 MWe for each new coal unit and
1200 MWe for each new nuclear unit if no other data were
available.
18'
-------
^
i °> «
I ^
o i!
! 2 C
O
II
-H
0 .
J-}
S|
\D
00 .-1
f-4 I ±J en ^s1 o O ^^ CM os o
en en r** m ^D xo en ^
O *T "-O f ) O O O r**4 P*>
CM .(*( . t CM "4 *H
-4OOO*^cnOsp-(r*
r^- «T en o P^ ^i CM c*) ^H
CM r-t i ( ^
m in CM CM r*» o vc o\ en
1-4 en en »r »-! CM
cnO-H^r^HOtneM'O
t .-
V «-l
»H cn
os so m en CM co r*«* < *T
COfi»-(OOsCOvTeMO
co-3'»-!mOP»-mOr»»
^T i (cMCM^Hf^^ren«T
H <
-------
CO
OS
u
el
§
S
g
,2
9
U
O
I Wl
!A .3
<* 3
ooooooo
ooooooooo
ooooooooo
ooooooooo
CN * 00
U
i
|
g
tn a
A o2
1 «
a a
Si^
I 1
^ 3
i i i t
20
-------
Several conclusions can be drawn from these forecasts.
From the unit size data which were reported and not assumed, it
was found that after 1985, utilities will probably concentrate
on the construction of units which are more standardized in
terms of size; this standardization process is presently underway
in at least one region.5 The types and sizes .of these units are
not uniformly distributed,among regions. It can be seen, however,
that the number of nuclear units is anticipated to increase. The
power pools installing large numbers of nuclear units may use fewer
coal-fired base load units unless demand is growing extremely fast.
In the 1986-2000 analysis, small coal units (<600 MWe), as well as
oil and gas units, show a sharp decline in their importance as
prime movers. The maximum percentage of total MWe contributed by
these smaller coal units is 19 percent after 1985, as opposed to
47 percent in earlier years. No oil or gas units are expected to
be built after 1985.
These Teknekron data were compared to Federal Power
mission (FPC) estimates for 1990 as contained in the 19-70 National
Power Survey (NPS)2 and as extrapolated to 2000 using NPS-derived
growth rates. While some differences did exist between the
Teknekron and NPS data, a combination of both was deemed suffi-
ciently accurate for forecasting purposes and was thus used to
compute the expected mix of generation in each area (Appendix B)
in the following manner:
All new generation estimates (1976-2000) for coal,
oil, gas, and nuclear from Teknekron Data .
Estimates for old units (pre-1976) remaining on
each system, for other units (hydro, pumped storage,
gas, turbine, diesel), and for old unit retirements
from NPS
Load factor data from NPS
21
-------
Many of the units anticipated in this forecast are still in early
planning stages, and thus, are subjected to operational and con-
tractual delays. Should these delays occur, actual generation
could vary from that used in this analysis.
22
-------
4.2 Demand Characteristics
Generating unit duty can be assessed from operational
and design viewpoints based upon expected demand as determined by
the utility's dispatching policy. Figures 3 and 4 show typical
weekly power system load (or demand), curves. An investigation of
similar curves furnished by various utilities for this study
showed that, while there are some seasonal and temporal varia-
tions, these figures can be viewed as being representative.
Figure 3 has load divided as to type: base-load, intermediate-
load, and peak-load. Figure 4 shows the same curve with a typical
division of generation responsibility by type of prime mover.
Economics dictate this division of responsibility; units with the
lowest operating costs (with some exceptions, e.g., some hydro
plants) are loaded to capacity (or, possibly, to their lowest
individual heat rate points) first, and more expensive forms of
generation remain in peaking or intermediate services. Also, as
.newer, more efficient steam units come on-line, older units which
may have been base-loaded become peakers or intermediate-service
units. Nuclear units are, almost without exception, base-loaded.
Since new coal units are generally the most economical fossil-
fueled steam units on most power systems, and since peaking
capacity is and will be made up almost exclusively of old steam
units, hydro units (including pumped storage) and internal com-
bustion peakers, e.g., diesels and gas turbines, it remains that
almost all new coal units will be designed for base-load and
intermediate-load duty. An investigation of coal units under
construction bears this point out - not a single unit was for
peaking and only a few were for intermediate load.3'1* It is
expected that this situation will continue to exist in the- future,
although the relative percentage of new coal units for inter-
mediate-load duty will increase as a greater proportion of nuclear
units (base-load) come on-line. Therefore, based on these find-
ings, this study has assumed that all new coal units will be
designed for either base-load or intermediate-load duty.
23
-------
U)
05
in
DC
13
O
Q
<
O
i£
LJJ
LU
OJ x-v
> 4-1
hi -fi
T3 fx,
oj PL.
O O
rH O
>i O
H C
13
0)
<1) -U
M O.
H <;
jo spuBsnoqx -
24
WSISA.S
-------
TYPE OF GENERATION
Gas Turbine
Ptimped-Storage Hydroelectric
Conventional Hydroelectric
Nuclear or Fossil-Fueled Steam-Electric
-.'X
Pumping Energy Supplied by Nuclear or Fossil-Fueled Steam-Electric Plants 2
Figure 4. Typical weekly load curve apportioned
by prime mover type.
(Reprinted from 1970 NFS; no copyright)
25
-------
Since generating units of a single type, e.g., base-
load, have similar controls and economies of operation, and since
most utilities dispatch and commit their generation in similar
fashion, it follows that hourly demands on units of the same type
will be somewhat similar. This point is especially true with .
regard to base units, which are usually run at nearly constant
power outputs except when economics of operation dictate that ;
load be reduced to avoid shutting intermediate units down.
Furthermore, new coal units will remain classified as base-load
or intermediate-load for a large percentage of their lifetimes
for two reasons. First, most base-load units require fairly
extensive modifications to their steam flow and control systems
to allow them to cycle (or follow load) mo
-------
curves do not directly show exact diurnal variation in demand,
they are derived directly from the diurnally varying curves and,
therefore, can be considered as representative of unit loading
which would be attempted for each unit type. These curves re-
present the percentage of time that demand on the unit equals or
exceeds a certain value. They are obtained figuratively by placing
each hourly demand value in descending order by magnitude and
plotting the results. Mathmetically, these curves are computed
from the probability densities of load versus probability of
occurrence (load level versus number of hours at that level).
This study has used the data in this latter form for .computational
purposes as in Appendix A. While the exact form of the actual
desired unit loading curves may vary from day-to-day and unit-to-
unit, the approximate curves used in this study represent what
might be reasonably expected of these two different unit types.
These curves also can be described using capacity factors
o-r load factors. Capacity factors represent the mean loading
by capacity; load factors, mean loading divided by peak
loading. The curves in Figures 5 and 6 thus represent capacity
factors in the neighborhoods of 70 percent for base-load new coal
units and 55 percent for intermediate-load new coal units. For
the purposes of this study, it was assumed that peak unit loading
would equal unit capacity at some point during the period under
investigation and thus that capacity factor would be equal or
nearly equal to load factor foreach unit. This assumption allowed
for the approximation of total load on new coal units as a function
of unit type (base or intermediate) only, and it eliminated the
otherwise necessary requirement of considering each new coal unit
individually. This approximation of total load on new coal units
was, therefore, made by determining the expected mix of base-load
and intermediate-load new coal units from existing load factor data,
summing the demand densities in percent of capacity for all units,
27
-------
-r
10
T
20
30
4Q 50 60
% OF TIME
70'
90 100
Figure 5- Expected load "duration curve for new
base-load coal unit-.
28
02-2S91-1
-------
10 20 30 40 50 SO 70 80 90 100
Figure 6. Expected load duration curve for new
intermediate-load coal unit.
29
02-2592-1
-------
and scaling by the sum of average unit sizes in MWe. The
determination of the mix of base-load and intermediate-load
new coal units for each region will now be discussed.
4.3
Determination of Mix of Base-Load and Intermediate-
Load New Coal Units
As was outlined in the previous section, the mix of
generating units by size and prime mover type was estimated for
each region. It was assumed that the mix of duty types (base
and intermediate) could be applied uniformily over new coal units
in both size ranges studied (390-599 MWe a,nd 600+ MWe) . This
mix was obtained by estimating the expected load factor for new
coal units and assuming that it and the capacity factor were
reasonably close in magnitude. Estimated system load factors
for 1990 were obtained from the 1970 NFS and assumed to be the
same, when rounded to .the nearest integer, for 2000. Total
system capacity was estimated using NFS and Teknekron data, and
this capacity was grouped as old units existing in 1976, pre-
1986 units by type and size, and post-1985 units by type and size.
Peaking capacity (hydro, pumped, storage, gas turbine, etc.)
was estimated from NFS data. System load duration allocation,
as in Figure 7 for a portion of NPCC and Figure 8 for a portion
of SERC, was used if available. Load factors were generally
assigned in the following manner:
old units*
pre-1986 steam units
all nuclear units
- system load factor
- system load factor or inter-
mediate load factor, if
estimated (e.g., NPCC)
- 70% load factor (base-load)
all internal combustion - 20% load factor (peaking)
all hydro and pumped
storage - 40% load factor (intermediate
and peaking)
Does riot include hydro,
*Units in service prior to 1978.
pumped storage, or internal combustion units.
30
-------
\
NORTHEAST REGIONAL ADVISORY COMMITTEE
COORDINATED STUDY AREA 8
TYPICAL LOADING OF ESTIMATED 1990 PEAK WEEK LOAD DURATION CURVE
124 -p 59.75 TOTAL NET GENERATING CAPACITY
120
55-
110-
50-
100-
UJ
U
90-
80-
70
60
50
40
30
20
48.1
-45-
LEGEND
EJijigS-jj PUMPED-STORAGE
KSv^SNj CONVENTIONAL HYDRO.
FOSSIL STEAM
Iw'-C'r^J?! NUCLEAR STEAM
I | PUMPING ENERGY
100
PE3CENT OF TIME
Figure 7. 1990 expected load duration curve with
typical apportionment by prime mover type for a
portion of NPCC (Region 6)
(Reprinted from 1970 NFS; no copyright)
31
-------
SOUTHEAST REGIONAL ADVISORY COMMITTEE
COORDINATED STUDY AREA FOR THE SOUTHERN COMPANY SYSTEM
TYPICAL LOADING OF ESTIMATED I99O PEAK WEEK LOAD DURATION CURVE
120-
110-
4O SO 6O
PERCENT OF TIME
IOO
Figure 8. 1990 expected load duration curve with
typical apportionment by prime mover type for a
portion of SERC (Region 7)
(Reprinted from 1970 NFS; no copyright)'
32
-------
In the case of NPCC and WSCC, the last three items above were
grouped together and given a net 50% load factor to account for
significant amounts of bas^-load .hydro available in those regions
Net new coal load factors were then estimated by
scaling each load factor by percentage of capacity, subtracting
the sum from the expected system load factor, and rescaling by
the ratio of system to new coal capacity, or
LF
c
S
NC
(1)
where
LFg = estimated system load factor
LFNC = new coal estimated load factor
LF. = estimated load factor for generation type i
Cg = expected total system capacity
Cj^ = expected new coal (1985-2000) capacity
G£ - expected capacity of generation type i
The percent mix of base-load and intermediate-load new coal
units was then estimated from the base-load and intermediate
load capacity factors
LFNC^:0.55 (~LL + 0.7 SL " ' (2)
33
-------
and from
CTT
%IL = J*± x 100 ' (3)
%BL = - x 100 . , (4)
where
GTT = expected capacity, intermediate- load
J.JL
C,,T = expected capacity, base- load
rSLi
= expected new coal capacity
and where it was assumed that peak load on each unit equaled
capacity at some time during that unit's lifetime. These
results were then checked based on NFS system makeup data and
general knowledge of regional problems and modified slightly
if necessary'. For example, base-load capacity was decreased
slightly in WSCC to reflect uncertainties in water availability
for hydro. In several other areas, slight: corrections were made
to account for expected uncertainties in nuclear plant construc-
tion. In general, no modifications in excess of 10 percent were
made. The resulting expected mix of new base and intermediate
load new coal units for each region is shown in Table 8; calcula-
tion results for each region are given in Appendix B.
These data were used as inputs for the determination
of incremental loss-of-load probabilities (ILOLP) . Each base
unit was assumed to have the previously mentioned base- load
demand characteristic; each intermediate unit; the intermediate-
load demand characteristic. The percentage of base and inter-
mediate units and the average unit size for each prime mover
34
-------
Tattle 8. ESTIMATED MIX OF NEW BASE AND
INTEEMEDIATE LOAD COAL UNITS IN 2000
Region "L Base % Intermediate
1 - ECAR 90 10
2 - ERGOT 10 90
3 - MAAC 10 90
4 - MAIN 70 30
5 - MARCA 70 30
6 - NPCC 0 100
7 - SERC 30 70
8 - SPP 0 100
9 - WSCC 90 10
35
-------
size group (390-599 MWe and 600+ MWe) were given for each region
as were the independent variables. Outage rate probability '.
densities by prime mover size group, as discussed in the next
section, were the only other variables needed for the calcula- ;
tion of ILOLP.
4.4 Outage Rates
The outage rates for coal-fired boilers that were
used in this study are listed in Tables 9 and 10. These data
were obtained from an Edison Electric Institute report5 on trends
of large mature fossil fuel units in operation prior to 1971. '
Planned and forced outages of coal-fired power plants were com-
piled for two size classifications: 390 to 599 MWe and 600+ MWe.
The outage rates presented in these tables were
determined from the number of hours of operation at each electric
.production rate, and are averages for the years 1972 through 1974.
They cover both partial and full outages. Other sources investi-
gated7'8 tended to uphold both the data used and the methods for
collecting it.
The total amount of forced outage as reported in about
twice that due to planned outage. Almost all the planned out-
age results in total outage while the forced outage is divided
between total outage and partial outages of 60 to 99.9 percent
of rated capacity.
The load factor (mean load scaled by dividing by the
peak load) of the 300 to 599 MWe units is 67.9 percent. The
600+ MWe units have a load factor of 69.2 percent. This indicates
that most of these units were probably base-load units, although
some smaller units may have seen intermediate duty. These re-
sults compared favorably with the assumptions concerning unit :
36
-------
Table 9- NATIONAL OUTAGE RATES FOR LARGE, MATURE COAL-FIRED
ELECTRIC GENERATING STATIONS - 390-599 MWe UNITS
Probability of Probability of
Percent of Capacity . Planned Outage a - Forced Outage-
0 -1329 .1357
.1 to 19.9 .0010 .0016
20 to 29.9 0 .0091
30 to 39.9 .0001 - .0041
40 to 49.9 .0001 .0038
50 to 59.9 .0001 .0176
60 to 69.9 .0007 -.0153
70 to 79.9 .0015 .0389
80 to 89.9 .0050 .0774
90 to 99.9 .0097 .0480
Probability of 100% capacity is .4974
^an AY^ifWitX...T_i67,8.l^,__Standard deviation = .4305,,
a Planned full or partial outage rate.
Forced full or partial outage rate.
37
-------
Table 10. NATIONAL OUTAGE FATES FOR LARGE, MATURE COAL-FIRED
ELECTRIC GENERATING STATIONS - 600+ MWe UNITS
Probability of a Probability of
Percent of Capacity Planned Outages Forced Outage13'
0 .1633 .0971
.1 to 19.9 .0001 .0007
20 to 29.9 .0001 - .0003
30 to 39.9 .0002 .0018
40 to 49.9 .0016 ' .0054
50 to 59.9 .0008 .0164
60 to 69.9 .0005 .0144
70 to 79.9 .0008 .0344
80 to 89.9 .0021 .1072
90 to 99.9 " .0033 .0550
Probability of 100% capacity is .4945
Mean availability = .6924. Standard deviation = .4239..'
a Planned full or partial outage rate.
b Forced full or partial outage rate.
38
-------
duty and load factor which were made in this study (Section 4.2).
Since outage rate distribution depends somewhat on load factor,
it is expected that future units will probably have outage rates
which are somewhat similar to these.
4.5 FGD Energy Penalties
Power consumption of FGD systems results in the effective
derating of coal-fired power plants, since power which could be
used to meet load must be consumed in-plant to operate the FGD
system. This power consumption is referred to as the FGD energy
penalty and is reported as an additional generating capacity re-
quirement necessary to provide rated unit power plus FGD require-
ments. Table 11 shows the energy penalties for each of the nine
regions for 25, 100, 500, and 1000 MWe power plants. The energy
penalties for the 500 MWe stations were used in this study since
almost all of the new power plants are expected or assumed to be
between 400 and 600 MWe.
The data in Table 11 are from an EPA study9 on energy
penalties for various FGD systems. They represent energy
penalties resulting from 90 percent sulfur removal using lime-
stone scrubbers. The energy penalties given represent the power
requirements to'produce 90 percent removal. The major types of
coal that will probably be burned within each region in 2000
were also estimated to determine the energy penalties, since
sulfur content affects FGD operation. It was assumed that ECAR
(Region 1) and MAIN (Region 4) would use medium sulfur coal from
the Midwest instead of the low sulfur coal from the West due to
the effect of a revised NSPS. The remainder of the regions were
assumed to use local coals or to adhere to present coal purchase
practices. This assumption is based on estimated transportation
requirements, historical use data from the 1970 NPS, and regional
forecast data from several other reference.10"13
39
-------
Table 11 . EXPECTED ENERGY PENALTIES FROM FGD USE
, _ - 7o of Additional Generating Capacity
Expected Coal Required to Operate the FGD System a
Region Type __^1
25 MWe 100 MWe 500 MWe 1QQO MWe
1-ECAR HV-MS local coal 3.7 3.7
2-ERCOT MV-LS/LV-LS 3.3 3.3
3-MAAC HV-MS 3.7 3.7
4-MAIN HV-MS local coal 3.7 3.7
5-MARCA HV-LS 3.3 3.3
6-NPCC HV-MS 3.7 3.7
7-SERC HV-MS 3.7 3.7
8-SPP MV-LS/LV-LS 3.3 3.3
9-WSCC MV-LS 3.3 3.3
HV - High Volatility
MV - Medium Volatility
LV - Low Volatility
MS - Medium Sulfur
LS - Low Sulfur
3.8 3.5
3.4
3.8 3.5
3.8' 3.5
3.4
3.8 3.5
3.8 3.5
3.4 --- ,
3.4
aAssumes 90% sulfur removal using a limestone scrubber I. Data
are grouped by size of the plant requiring the FGD unit'.
40
-------
4.6 Methodology for Calculation of Incremental Loss-of-Load
Probabilities
In this section an overview of the analysis used to
compute the incremental loss-of-load probabilities (ILOLP) is
presented. The methods are discussed in detail in Appendix A.
The presentation here is organized as follows in the subsections
below:
(1) calculation of the probability distribution
of capacity available at a given unit, taking
into account both boiler and scrubber down-time,
(2) calculation of loss-of-load probability for a
single unit, and
(3) calculation of incremental loss-of-load proba-
bility for a new coal only in a regional
system in which the demand is distributed
" among the different generating units.
:4~ 6.1 Calculation of Distribution of Available Power at
a Particular Unit
It was assumed that a revised NSPS would require a
generating unit to be derated upon FGD module failure if a
spare module or bypass were unavailable. Therefore, for a
given generating unit with FGD, the percent of total capacity
which can be operational at a given time is limited by two
factors. These factors are:
(1) the percent of total capacity at which the
boiler is capable of operating,
(2) the number of scrubber modules which are
available.
41
-------
Suppose, for example, that the boiler, because of a
partial outage, is capable of operating at 80 percent of
capacity, but that only three of five scrubber modules are up
This means that the scrubber is capable of generating
100%
or 60 percent of the output from a fully operational boiler.
Thus, the boiler-scrubber system is limited to operating at
' fin no-.~ooT.t- n-F t-rvf-oi ,. an a «-f * ir TT-M-io both factors mu's t be
***.*-* w ) v.*.*^ i_/\^ -L. -h, ^ .!_ hJW-1-UkLyk/^.U. t? JT C3 «_ ^JLU. J_ »3 .1. -UU.i_U t. *
only 60 percent of total capacity. Thus,
taken into account.
If r denotes the probability that a given scrubber
module is up, and if there are K modules, the probability
that exactly i modules are up is
P.
K:
(1.0-r)
K-i
(5)
This is the well-known binomial probability distribution
The distribution of boiler outage for different
sized plants was discussed in Section 4,3. The following
is the primary equation used to combine the probabilities
for the scrubber and for the boiler:
P(at least j70 of capacity can be used) -
P(at least j7a of boiler's capacity is available) x
P(enough scrubber modules are available to handle at least
j?a of the total capacity)
42
-------
In this manner, the distribution *of percent of
total capacity available was obtained. The values .r - .8
and .9 and K=5 and 6 were considered. When K=6, the sixth
module was considered to be a spare; that is, each module
was assumed to be able to handle up to 20 percent of the
emissions from a fully operational boiler.
The mean and standard deviation of the percent of
total capacity available were computed from the distribution.
The use of the distribution to compute loss-of-load probability
for a single unit is discussed in the following section and in
Appendix A. The use of the distribution to compute the incre-
mental loss-of-load probability for a regional system within
which the demand is distributed among different generating
units is discussed in Section 4.6.4.
4-6.2 Loss-of-Load Probability for a Single Generating Unit
The loss-of-load probability (LOLP) is simply the
.probability that not enough power is available to meet
demand. If we define as random variables
Y = amount of power available,
X = power demand, and
L = Y - X,
then loss of load occurs if L is less than zero.
X, Y and L are random variables_represented by~prob-
ability distributions: The distribution of Y was obtained
by using the analysis in the preceding section. The probability
distribution of demand, X, was discussed in Section 4.2. A
statistical method called convolution was used to obtain the
distribution of L. Convolution as applied to this problem
43
-------
is discussed in some detail in Appendix A. The loss-of-load
probability (LOLP) was obtained by summing the probabilities of
values of L less than zero (demand exceeds available capacity).
4.6.3 Expected Effect of FGD on the Reliability of a Single
Unit
Consequently, if LOLP with and without FGD can be
computed as above, the effect of various scrubber configurations
upon the reliability of an individual generating unit can be
estimated. Table 12 shows the effect of a scrubber on the
probability of meeting demand for each of the four unit sizes
and types studied. It should be stressed that this table con-
tains loss-of-load probabilities for individual new coal units
and not absolute probabilities for meeting demand by all units
in a system. Since individual units are typically part of an
interacting system, this analysis is not exact and is meant
only to provide qualitative information.
Table 12, then, shows the effect of the number of
scrubber modules and the module availability on individual
unit reliability. Module availabilities of 80 and 90 percent
were considered. In addition, five modules were used with and
without a spare module. The effect of these scrubber systems
upon unit reliability is listed for 300 to 599 MWe units and
600+ MWe units in base- and intermediate-load service. The
results expressed as (1.0 - LOLP) were computed by convolving*
the demand probability distribution with the availability
probability distribution for the boiler/scrubber -system. Case 3
which used 90 percent module availability and one spare module
was found to have the highest overall unit reliability while
The statistical method of convolution as applied to this
study is discussed in Appendix A.
44
-------
H
M
jti
>^
r^
a
M
H
w
a
w
o i
W :
13
0 :
L Mi i
< J
!
fa j
o i
^4
H
i)
r 1 i
&
j
-U
H
SJ r-!
3 J2
Q rH
2 -r+
^
Uj
O
03
M CJ
^2J W
2 2
§ 01
a
^^_
41
03
' CO
4J
03
01
S5
m o o>
CM O i 1
en CN in
c^ in co
\O VQ vf^
co *3" en
O CO O
r-- so r-»
»o en in
r*1* oo vo
vO vo so
i O O
O O . i
41
w
CO
CJ
4-1
03
O
3
* '
r^ CO
in cTt
vO SO
, .
m CN
m so
'
CSJ CVJ
-------
Case 4, which used 80 percent module availability and no spare
module, had the lowest unit reliability. It was found that a
base-load unit had lower'probability of meeting demand than an
intermediate-load unit, as one might expect.
The facts that (1) groups of individual units inter-
act and are committed, dispatched', and controlled simultaneously,
and (2) groups of generating units can suffer concurrent full
and/or partial outages, require that an entire power system be
considered as a whole. In this case, the analysis of a single
unit definitely cannot be applied to an entire system. Con-
sequently, each region was studied as groups of generating units
interacting simultaneously; the next subsections address this
approach and its resulting estimates of changes in system reli-
ability due to FGD.
4.6.4 Incremental Loss-of-Load Probability for a Power System
In this section, the calculation of the incremental
loss-of-load probability (ILOLP) for a system of units* in a
region is discussed. It was assumed that the system demand is
distributed among all units such that the load duration curves
of Section 4.2 were applicable to individual base-load and
intermediate-load units.
In a given set of calculations, the incremental loss-
of-load probability for new coal units was obtained for a .
particular region, assuming that all these units in the region
had the same number of scrubber modules and that the probability
that a particular module was up was the same for all modules.
Separate calculations were made for the following cases:
^"System of units" in this case representing only a portion of
total generating capacity.
46
-------
(1) no scrubbers;
(2) five modules per unit and r = .9, where r is
the probability a given module is up;
(3) six modules per unit and r = .9;
(4) five modules per unit and r = .8; and
(5) six modules per unit and r = .8.
These new coal generating units were classified as follows:
(1) large (600 megawatts or more) base units,
(2) small (390-599 megawatts) base units,
(3) large intermediate units, and
(4) small intermediate units.
The analysis presented in Section 4.6.1 was used to
calculate the mean and standard deviation of the proportion
of total power available for a given unit in each of the four
categories. Then using properties of the mean and standard
deviation along with other information, the mean and standard
deviation of ;total power available for a given region were com-
puted. The other information used was
(1) the mix of base- and intermediate-load new
coal units and
(2) the average sizes of the large and of the
small new coal units in the region.
', i
-------
Now define as a random variable:
Y - total power available at all units in the region.
The mean and standard deviation of Y are known, then,
through the analysis discussed above. Also'define.as a random
variable:
X = total power demand at all units in the region.
The mean and standard deviation of X can be obtained
by using properties of the'mean and standard deviation, the
demand characteristics, and the generation mix. Then define
as a random variable:
Y - X
(6)
As before, loss-of-load occurs if L is less than zero.
The probability of this event occurring was calculated; details ;
of the calculation are in Appendix A. Because of time considera-
tions, this probability as used in this study is an incremental
loss-of-load probability (ILOLP), since only that increment of
system capacity composed of new coal units was considered in
the calculation. ILOLP's were calculated for each region with
and without FGD on all new coal units and for each of the four
modular configuration/availability cases. THESE PROBABILITIES
MUST NOT, IN ANY WAY, BE CONSTRUED TO REPRESENT ACTUAL- EXPECTED
TOTAL LOSS-OF-LOAD PROBABILITIES FOR THE REGIONS; THEY ARE NOT.
However, if it is assumed that (1) the unit load duration curves
given in Section 4.2 represent operation with interactions among
all units (i.e., the total system is economically dispatched
and/or operates under automatic generation control or load
frequency control), and (2) expected mix of base- and intermediate-
load new coal units without FGD represents operation at a
48
-------
reasonable system total loss-of-load probability with adequate
reserves (i.e., the system is neither overbuilt nor underbuilt),
then these values of ILOLP can be used to compute a percent
additional generating requirement for the system with FGD. This
requirement represents the amount of additional generating
capacity necessary such that the computed ILOLP for the system
with FGD is equal to the computed ILOLP for the same system with-
out FGD. In other words, if the system studied is neither
overbuilt nor underbuilt without FGD, it has a particular accept-
able reliability level. If FGD is then required on all new coal
units, reliability is reduced such that a mean amount of additional
capacity equal to the mean additional generating requirement is
required on the average to restore mean system reliability as
measured by ILOLP to the previous level attained without FGD.
In mathematical terms, the above computation is as
follows :
P(L<0.0)
= Pi (7)
no FGD
and M MWe
o
and P(L<0.0)
= Pi
-------
Therefore, if M. is the additional generating requirement in MWe
for FGDi, then
P(L<0.0)
P(L<0.0)
FGD..^
and (MQ + M±) MWe '
(9)
no FGD
and MQ MWe
This value of M. was thus calculated for each region.
It should be noted that for a given region, L is the
sum of the power available at each unit and the (negative of the)
demand at each unit. Since each region contains at least 40
units, and in most cases, close to 100 units or more, it is evident
that L is the sum of a large number of terms. Therefore, the
calculation of M. could be a very complicated and time-consuming
process which might have been impossible given the time constraints
of this study. However, it is reasonable to assume that L is-
normally distributed: the Central Limit Theorem in statistics
states that under certain conditions, sums of large numbers of
randomly varying quantities are approximately normally distributed,
and the properties of normally distributed random variables are
well-known and easily computed. The normality property, along
with the means and standard deviations discussed above, then, were
used to compute easily the incremental loss-of-load probabilities
and, hence, the mean additional generating requirements M£. F^
is a mean value in MWe which was computed using the Central Limit
Theorem and, thus, other mean values. In simplified terms, M..^
was calculated from the relationship
Y - X
(10)
50
-------
where
L, Y, and X are mean values, and where
Y = Y A ' (11)
where
Y = total capacity available
A = mean new coal unit availability- for the region
without FGD.
Therefore, M. must be scaled to represent the total capacity
requirement M. due to FGD. This scaling is accomplished by
dividing M. by the mean new coal unit availability without FGD
for the region, A. These results neglect the effects of.FGD
units on any additional operating capacity required.
In the calculations discussed above, correlations be-
tween demands at different units in the same region at the same
time were not considered. Such correlations probably exist be-
cause of several factors; for example, all units would be
affected by the same type of diurnal cycle. The result of this
emission is to shift the incremental loss-of-load probability
toward 0.5 to some extent. The investigation of these corre-
lations was beyond the scope and limited timeframe of this
study; it is felt, however, that the effect of the correlations
is not drastic.
4.8.5 Expected Effect of FGD on System Reliability
This section considers the effect of FGD on the
effective generating capacity of a power system. Listed in
Table 13 are mean expected additional generation requirements
51
-------
Table 13. MEAN ADDITIONAL CAPACITY REQUIREMENTS IN PERCENT OF
NEW COAL CAPACITY TO OFFSET THE RELIABILITY EFFECTS OF FGD
Case
l(Base)
2
3
4
5
No.
Modules
0
5
5
5
5
No.
Spares
0
1
0
1
Availability
Per Module
90%
90%
80%
80%
Mean Additional
Generation_
Requirement M.
4
1
. 9
4
0
.5%
.2% (Best)
.9% (Worst)
.4%
aExpressed as a percentage of new coal capacity without FGD.
Does not include boiler/turbine/generator availability effects.
M. for the different scrubber module configurations and different
module availabilities studied. As was previously mentioned,
mean new coal unit availabilities for each region must be used
with these mean additional requirements to determine the amount
of actual total additional capacity required to compensate for
the generating capacity lost due to lower FGD availability.
It can be seen from this table that the scrubber system
of Case 3 had the least effect on the system performance, while
Case 4 caused the largest decrease in system/mean available
capacity. It must also be noted that Cases 2 and 5 had similar
mean additional generation requirements indicating that adding
a spare module compensated for the ten percent drop in module
availability. These results may be sensitive to future statis-
tical findings .
52
-------
These calculations were made for the expected
generation mixes in each reliability region and for other
additional mixes and capacities. It was found that the mean
percent additional requirements given in Table 13 were independent
of either the mix of base and intermediate units or the total
generating capacity in a r.egion. The expected probability of
excess demand* for new coal is sensitive to these considerations
and is tabulated by region in Appendix C. This result is
important because if the mean percent additional capacity re-
quirement per se were independent of regional considerations,
such as mix or load factor, then the results of this study could
be easily applied to many different cases. Regional considerations
would only be required to determine type and amount of total re-
placement capacity from the mean percent requirement expected.
More detailed .investigations should be made to verify and further
clarify this relationship.
^Probability of excess demand = ILOLP
53
-------
SECTION 5
EVALUATION AND INTERPRETATION OF RESULTS
The effect of FGD on reliability for units put on-line
before 1986 was first evaluated and found to be only marginally
important. Next, the effect on units put on-line between 1985
and 2000 was estimated for one individual unit and for an entire
system. This study of the net effect of FGD on power generation
led to several conclusions. First, significant amounts of addi-
tional capacity may be required to obtain the same ability to
meet load as can be maintained without FGD. Second, constraints
on power interchange among power pools may restrict necessary
power flows to deficient utilities during FGD-induced outages
such that interchange cannot significantly mitigate the impact
of FGD. Finally, an analysis of reserves indicates that systems
may be underbuilt by 1985 and thus may be unable to keep adequate
reserves or, in some cases, to meet demand. It was found that
further reductions in reliability caused by FGD would impact this
problem. Measures to mitigate these impacts of FGD were also
determined.
In this section the results leading to these conclu-
sions will be evaluated. This evaluation considers the following-
subjects:
Additional capacity requirements
Constraints on interchange
Reserve policies and requirements
Uncertainty and sensitivity of study
54
-------
5.1 Effect of FGD on Additional Capacity Requirements
The data and methodology presented in Section 4 were
combined to provide estimates of the expected effect of FGD on
individual unit and system reliability. This analysis assumed
that any revised NSPS would require scrubbers on all new coal
units not presently under construction. In this section,
results are presented which reflect the effects of this
implementation of FGD on the reliability of the nine regions
of the U. S. before 1986 and between 1985 and 2000.
5.1.1 Effect of FGD on Reliability - Units On-line Prior to
1986
. It was found that the proposed NSPS for coal-fired
power plants will not drastically affect units which are on-line
prior to 1986 because many of these units are already under
construction. However, to subjectively determine the effect of
any scrubbers on-line during this period, data furnished by
EPA11* for units and FGD systems presently either planned or
under construction were collected and analyzed. These data
are summarized in Table 14. As can be seen from the table,
only about 20 percent of the new coal units on-line in 1985
are expected to have FGD systems.
55
-------
Table 14. ESTIMATED SCRUBBER ADDITIONS BEFORE 1986
New Coal
Region New Coal MWe New
1
2
3
4
5
6
7
8
9
U.
a
b
- ECAR
- ERGOT
- MAAC
- MAIN
- MARCA
- NPCC
- SERC
- SPP
- WSCC
S. TOTAL
Teknekron data
EPA data cited
30,122
7,798
2,500
8,752
10,942
2,278
12,862
17,097
. 27,985
120,336
in above text
7= Total No. a
Generation Units
62%
54% '
127o
387,
1007o
137=
287,
567,
597=
467=
59
14
4
21
27
5
24
32
65
251
No. ' b
Scrubbers -
12
6
0
4
7
1
3
4
17
54
-..
From a more rigorous standpoint, NERC reports that 20
percent of the total additional generating capacity required in
1985 is presently not under construction.17 If it were there-
fore assumed that 20 percent of new coal capacity on-line before
1986 were also not under construction, then from Table 13, a re-
vised NSPS would only affect about 9 percent of the total new
national generating capacity. For the worst case studied (5
modules/no spares at 807= availability) this effect can' be
estimated as requiring about 5100 additional MWe nationwide.
This estimation was done using the methods described in Section
4.6.4 for new coal generation on-line between 1985 and 2000.
56
-------
Therefore, because of the small number of scrubbers
planned in each region prior to 1986 and the small percentage
of total additional generation which might require FGD, it is
expected that the implementation of a revised NSPS would have
very little effect on overall regional system reliability prior
to 1985. There are few possible exceptions to this statement:
one is WSCC, where overall reliability will probably not be
adversely affected, provided that another massive water shortage
does not occur. These conclusions are, of course, subject to
change if (1) more widespread use of FGD were required in this
period, (2) construction of new units continues to be delayed
(Section 5.3) such that systems are significantly underbuilt,
(3) scrubbers are installed without spare modules and with lower
modular availability (<80%) than is presently assumed, or
(4) another massive fuel shortage occurs at the wrong time of
year and puts a large number of units out of service.
5-1.2 Effect of FGD on Reliability - Units On-line Between
1985 and 2000
Assuming that the forecasts and data in Section 4
(without scrubbers) are correct insofar as reserves are concerned,
(i.e., future system reliability is within reasonable bounds
and required reserves are maintained), then the effect of FGD
would be to produce a deficit in available generation such
that reliability standards and reserve requirements would no
longer be met. The results in Section 4.6.5 were analyzed based
on this assumption to give ,the additional generating capacities
which must be planned for and added in order to make reliability
as measured by ILOLP equal to the base case (no scrubbers) ILOLP
values for each region. These mean additional capacity require-
ments M in percent (as in Section 4.6.5) were found to be
virtually identical by case regardless of system configuration.'
57
-------
Total additional capacity in MWe thus required in each region
with FGD was obtained by dividing the mean additional capacity
requirement in percent by the mean new coal unit availability
for each region and multiplying by the amount of new coal genera-
.tion MWe, as in Section 4.6.4. Thus, this study has recognized
that this additional capacity would .operate at the same mean
availability as new coal without FGD. Table 15 gives these
expected additional requirements in MWe to offset FGD reliability
only for the five cases studied by region and nationally. In
addition, energy penalties associated with scrubber operation were
determined; these total expected generation requirements to off-
set these penalties by region are shown in Table 16. As was the
case for reliability results, operation at the same availability
without FGD as other new coal was considered. Total additional
capacity required to offset FGD (reliability plus energy) is
seen in Table 17.
As will be discussed in Section 5.3, it was found that
this additional capacity will most probably be made up of new
units. If present trends continue, these new units will probably
also be new coal units. The additional generating requirements
due to FGD were therefore estimated in terms of equivalent
600 MWe coal units.* These estimates are summarized in Table 18.
It can be seen from Tables 15-18 that a large amount of
additional generation will be required to offset the-effects of
the widespread implementation of FGD. Since the 1970 NFS
estimates that demand during the period 1990-2010' will double
about every 10-12 years, these additional requirements after 2000,
will also increase in like manner. Therefore, it is important to
consider measures which would mitigate the impact of FGD. If it
is assumed that the United States will continue to develop its
600 MWe units were assumed in order to agree with assumptions
in the Teknekron data.
58
-------
Table 15. ADDITIONAL CAPACITY REQUIREMENTS IN MWe TO OFFSET
FGD RELIABILITY ONLY IN 2000
Region
1-ECAR
2 -ERGOT
3-MAAC
4 -MAIN
5-MARCA
6-NPCC
7-SERC
8-SPP
9-WSCC
MWe
Case 1
(Base)
0
0
0
0
0
0
0
0
.0
U.S. TOTAL 0
MWe
Case 2
(5/0@90%)
7,490
4,295
2,115
3,290
1,600
1,840
5,120
4,625
4,605
34,980
MWe
Case 3- a
(5/1(390%)
1,995
1,145
565
880
425
490
1,365
1,235
1,230
9,330
MWe
Case 4 b
(5 70(380%)
16,480
9.445
4,650
7,245.
3,515
4,055
11,260
10,170
10,130
76,950
MWe
Case 5
(5/1(980%)
7 , 325
4,11.5
2,065
3,220
1,560
1,800
5,005
4,520
4,500
34,190
a Best Case
b Worst Case
59 '
-------
Table 16. EXPECTED ADDITIONAL GENERATION REQUIREMENTS DUE
ONLY TO FGD ENERGY USE PENALTIES BY REGION
IN 2000
Region
1-ECAR
2 -ERGOT
3-MAAC
4-MAIN
5-MARCA
6-NPCC
7-SERC
8-SPP
9-WSCC
U.S. TOTAL
Maximum Additional Generation
Requirements Due to FGD Energy
Penalty - MWea
6,325
3,245
1,785
2,780
1,210
1,555
4,320
3,495
3,480
28,200
a Includes only effects of new coal unit boiler/turbine/generator
mean availability.
60
-------
Table 17. TOTAL FGD-RELATED ADDITIONAL CAPACITY
REQUIREMENT BY REGION IN 2000 - MWe
Region
1-ECAR
2-ERCOT
3-MAAC
4 -MAIN
5-MARCA
6-NPCC
7-SERC
8-SPP
9-WSCC
Case 1
0
0
0
0
0
0
0
0
0
Case 2
13,815
7 , 540
3,900
6,070
2,810
3,400
9 ,-440
8,120
8,085
Case 3a
8,320
4,390
2,350
3,660
1,635
2,050
5,685
4,730
4,710
Case 4b
22,805
12,690
6,435
10,025
4,725
5,615
15,580
13,665
13,610
Case 5
13,650
7,440
3,850
6,000
2,770
3,360
9,325
8,015
7,980
U.S. TOTAL 0(base) 63,180 37,530 105,150 62,390
a Best Case
Worst Case
61
-------
Table 18. TOTAL FGD-RELATED ADDITIONAL CAPACITY REQUIRE
MENTS BY REGION IN 2000 - 600 MWe COAL UNITS
Region
1-ECAR
2 -ERGOT
3-MAAC
4 -MAIN
5-MARCA
6-NPCC
7-SERC
8-SPP
9-WSCC
U.S. TOT^
Case 1
0
0
0
0
0
0
0
0
0
VL 0 (bas
Case 2
_ 23
13
7
10
5
6
16
14-
14
ie) 108
Case 3 *
14 _
7
4
6
3
4
10
8
8
64
Case 4
38
21
11
17
8
10
26
23
23
177
Case 5
23
13
7
10
5
6
16
14
14
108
Best Case
Worst Case
62
-------
coal resources as in the President's energy plan, then several
alternative measures to mitigate the impact of FGD are possible.
These alternative mitigation measures include:
1. Use of spare FGD modules
2. Allowance for bypass of scrubber modules on
outage status with reasonable restrictions, such
, , as when load cannot be met
3. Use of alternative technologies which meet a
revised NSPS without FGD or with reduced FGD.
Other possible mitigating measures include the purchase
of power from other utilities or the carrying of additional
reserves on older units. However, the next two sections indicate
that these latter two alternatives are too tightly constrained
to be highly significant mitigation measures.
5 . 2 Evaluation of Interchange Constraint's
In the event of an FGD-related outage on a single unit
or a number of units, it is possible in some cases for a utility
or power pool to purchase emergency power from another utility
or pool. Hence, this study has included an investigation of the
constraints on power interchange among utilities. Many of these
constraints must be evaluated using classical electrical AC net-
work theory, since the flow of power in a network is directly
related to the state of the network at a given time. Such an
evaluation was beyond the scope of this study. Furthermore,
the explanation of network constraints given in this section
is highly simplified. In addition, this section also evaluates
reported non-simultaneous power transfer capabilities for the
nine NERC reliability councils.
63
-------
Power interchange among utilities is restricted by
many different constraints. These constaints include excess
generation in a given area, maximum tie line power flow cap-
abilities, and contractual constraints. The capability to
transfer power requires accessible; available power generation
and, more importantly, accommodating .tie line and supporting
transmission networks. It should thus be rioted that regardless
of desires, contractual requirements, or net interchange capabili-
ties, power transfer is, in the end, directly controlled by the
instantaneous state of the generating system and the network.
If generation is unavailable, power cannot be transferred. More
importantly, if a network in a given instantaneous state
(regardless of its design or available generation) cannot
accommodate a desired power flow from point A to point B at a
given moment when it is needed, very little can be done to
alleviate the situation. Moreover, interconnection does not
eliminate reserve requirements and thus cannot eliminate the
additional generation requirements caused by FGD; it merely dis-
perses them. In the case of highly constrained interchange, as
exists in the U. S., it is doubtful that a reasonable dispersion ,
of the large amounts of additional generation required by FGD
could be achieved.
Several constraints resulting from system design and
instantaneous network state may be encountered during the trans-
fer of power.15 Network and generation configurations including
location of plants, generation type, transmission voltage, instanta-
neous plant power output, transformer tap settings, and -so on, can
strongly affect a utility's ability to interchange power. For ex-
ample, nuclear units require an extended outage (one month) for
refueling at predetermined intervals for optimum utilitization
of the nuclear fuel. These scheduled outages can cause problems
when forced outages occur coincidentally. Under certain normal
network conditions, large generating plants which are electrically
64
-------
close to tie lines can impair tie line flows to values far below
rated capacities. Furthermore, operating constraints can influence
tie line capabilities. Utilities which rely on only one major fuel
(usually at a uniform price) do not routinely interchange power
with other utilities because it is not economical. For example,
ERGOT, which traditionally has relied' almost solely on natural gas
at a uniform price has not engaged in significant amounts of
economy power exchange among its members. Utilities of this type
may be interconnected but generally do not have transmission net-
works which can accommodate massive power flows which might be
necessary in the event of a large outage. Conversely, when a
region does routinely interchange large amounts of power on an
economy basis, emergency transfers may be severely restricted be-
cause tie line and supporting network flows are already near the
capabilities of the transmission lines. Tie lines and network
design (or network state) may affect connections in another way.
One major impediment to efficient energy transactions in WSCC is
loop flow. The western part of this region is a relatively low
impedance network; the eastern part, high impedance. The result
of this situation is that the actual flow of power through these
interconnections may not match the scheduled transfer of power be-
tween control areas, and power may, in fact, loop completely around
the system. Loop flow can overload essential tie lines so that
in the event of an emergency, bulk power transfers become im-
possible. In fact, if loop flow should occur as a result of or
in conjunction with power loss at a marginally stable location on
a network whose state (or configuration) is also marginally
stable, the flow of power can even reverse itself away from the
areas where the need is greatest. This reversal generally has
a cascading effect which can result in a massive blackout.
In addition to generation and transmission constraints
contractual constraints on interchange also exist. In general,
these contractual constraints are controlled by regulatory
65
-------
agencies. While most utilities have emergency short-term power
exchange/payback agreements with their neighbors, utilities in
many of the power pools and/or utility companies state in their
interchange contracts that they are under no firm obligation to
supply power in the event of forced outages of equipment, missed
load forecasts, or fuel supply problems for any specific period
of time. This practice is, in many cases, required by regula-.
tory agencies in order to maintain adequate service to the
supplying utility's own customers. This practice may thus in-
convenience a receiving party should a power transfer be terminated;
such termination can and probably will result in a severe power
disruption. This type of disruption usually occurs when the
supplying utility experiences a coincident outage. Obviously,
the probability of this type of power cutoff increases with the
widespread use of FGD systems since it has been shown that FGD
reduces unit availability and system reliability. Also, in many
cases, interchange transactions must be scheduled. For example,
Middle South Utilities Company, located in SPP, has required in
the past that the receiving party furnish a schedule twenty-four
hours and, sometimes seventy-two hours, in advance of the inter-
change transaction. This scheduling delay, however, can be
avoided through mutual agreement if it is to the advantage of
both parties, as in an emergency. Middle South also burns
cheaper fuel for their own system load and reserves the higher
priced fuel for power transfers. This practice, generally regu-
lated to assure their customers of a cheaper rate, can cause
delays in emergency shipment to cover extended outages if the
price of power is too high. In addition, political constraints,
such as recent actions by some states to keep cheaper types of
power at home, can influence the power interchange transactions
from region to region.
66
-------
While it has been shown above that interchange of
power is strongly constrained by a number of factors, it is
still important to evaluate any maximum interchange capabilities
which might be required in the event of an outage. Present and
expected future power interchange capabilities vary considerably
from region to region. The following diagrams show the potential
non-simultaneous energy transfer capabilities of each of the
nine NERC reliability regions with neighboring regions and sub-
regions. l6'l7 Each of these capabilities represents the maximum
possible power transfer capability which might be expected under
ideal network conditions. Furthermore, each of these interchange
capabilities represents a non-simultaneous interchange; i.e., not
all transfer capabilities shown can be utilized at once. Actual
transfer capabilities thus can be expected to be somewhat less
than these values for reasons given previously.
Figure 9 shows that WSCC (Region 9), presently partici-
pates in no significant interregional power transfers; it is
anticipated that the minor interconnections with MARCA (Region 5)
which presently exist will not increase significantly in size due,
to stability considerations. WSCC does have strong subregional ,
power transfer capabilities among member utilities; this sub-
regional interchange capacity is expected to increase in the
future. Figure 10 shows that ERGOT (Region 2), the most indepen-
dent reliability region, still has no interconnections with any
other regions. It is anticipated that this situation will probably
not change in the future. However, ERGOT is divided into northern
and southern subregions which can,'and do, engage in limited emer-
gency power exchange among member utilities.
MARCA, Figure 11, can engage in power interchange with
MAIN (Region 4), SPP (Region 8), and WSCC. It appears that MARCA
will gradually increase energy import capability and decrease ex-
port capability to MAIN over the next six years, while power in-
terchange capability with SPP should remain relatively constant.
67
-------
1 O O O
§?S8
(0
I
ca
o
'01
Cl]
^.
ffl
C
co
O
t.
_
cd
O
'5>
£
2
'c
glf
SI CO
UJ
^
^
O
O
X
1
z
o
5
0
X
Q
ui
CC
1
CO
5
S
to
0
1
_l
o
o
N
cc
O
X
1
z
o
5
0
X
Q
UJ
CC
1
i
3
CO
X
H~
r^*
z
o
5
O
X
Q
UJ
cc
o
C/3
1
O
UJ
Z
_J
o
O4
o
CJ
o
S
Z
o
o
S
^
0
z
_J
o
X
D
P:
Z
o
5
X
Q
~
CC
0
03
o
O
g
5
Z
0
o
N
CC
<
CM
S
UJ
z
o
o
O
g
2
Z
0
o
N
cc
w
CO !C rH
%$&
a)
S
(30
68
-------
H
O
W CU
i 4J
U)
O
-n 5>
ffl 2
*"
co
c
CD CO
r*- w
W
rl H
4J ~ CU
r) x-. CU
r-l VO
H iH^S
Cd
Cd
O
CD
CO
§
4J
W
§
PdrH
H
O
to 13
C
-------
CO
c
£
o
c
O)
CD
a>
'c
O O I O O
o to o o
T <0 O O)
_
T- tn CO O4
o a o o
* to 10 in
co ^r co
Z
CO CO CO
<<<
o o o
O O O < < z z
^^!!I1
Q Q S22 5 5
bO
co 43
O vo
r-l
CU
cd cd
CD
h-
cn
o
GO
O3
70
-------
The seven subregions of MARCA can also transfer power among
themselves. SPP, Figure 12, can exchange power with SERC
(Region 7) and MAIN in addition to MARCA. SPP also can inter-
change heavily among its four subregions. Figure 13 shows that
MAIN can engage in sizable power exchange with MARCA, SPP, ECAR
(Region 1), and SERC. These exchange capabilities are expected
to increase by 1984. MAIN is divided into six subregions which
also can engage in heavy power transfer. SERC, Figure 14, is
subdivided into eight subregions which are also involved in
power interchange. SERC also can transfer power with SPP, MAIN,
ECAR, and MACC (Region 3). These transfer capabilities are ex-
pected to increase, with one exception; by 1984, energy export
capabilities to SPP should decrease considerably. This reduction
could indicate the installation of a new large generating unit
by a member of SERC in the area of that particular interconnection.
NPCC, (Region 6), Figure 15, power pools with ECAR and MAAC, and
is composed of six subregions which also can transfer large
amounts of power among themselves.
Figures 16 and 17 show the interchange capabilities of
ECAR and MAAC. These regions are not divided into subregions,
and are generally dispatched as single systems without regard to
individual member utilities. Therefore, their power interchange
capabilities per se are strictly interregional. ECAR has inter-
change capabilities with MAAC, MAIN, and SERC; and MAAC can inter-
change with SERC, ECAR, and NPCC.
From the above figures and text, it is apparent that
the eastern power pools are more tightly interconnected than the
western pools; however, no interconnections have been made
between the east and the west. It is doubtful that any major
east-west interties will be established before 2000, if at all.
71
-------
PM
PH
CO
1
^-^
CD
5
co
OJ
H
4J
H
r-f
H
O
rC 'f"
4-J !*^
1^- 4J
O
t3 (U
S «-l
Sw
vo i-l
r- <0
o^ cj
r-l O
H
>4J
/^N Ct)
C
o
o
C
to
o
'a
2
£
c
a:
O
Q.
CN
CD
I--
05
O
GO
O>
72
-------
CD
co
c
to
£
o
c
CO
g
§1 8 8 S 8 8 8
<7>
bQ cO
co
O
0)
i
H
CO
CO
i-H
0)
be
H
I r-4 4J
J-J
CO
CO
f-
O3
O
CO
O5
73
-------
o
f*
w
CO
H
Ft,
o
w
fl
o
.4J
. cd
cu
-u
>
74
-------
C
0)
£
o
c
m
c
o
"
O
CJ
OJ
a
H
o
0)
o| m in in
CM OJ r~ O O
| CO CM CM in CD CM
-------
03
g
'en
CO
r~-
O3
76
-------
JS i
c
£
o
c
to
o
Is
ffl
c
0)
CQ
CU
H
JJ
H
'cO
CO
0
CO
g
4-J
0)
I
CU
w
§
4J
iI
I
T-l
co
I
C
O
CU
n
3
CU
CU -
CO CO
CD
r^
cn
O
co
o>
77
-------
Power pools with more efficient tie line networks.
coupled with increased reserves could alleviate some disruptions
in power flow caused by FGD. However, because of fairly strong
constraints on the interchange of power among different regions,
it was determined that bulk power shipments most probably could
not significantly reduce the additional capacity requirements
caused by widespread implementation of FGD over long periods of
time.
5.3
Evaluation of Utility Reserves
Generation reserves for utilities are essential in
maintaining continuity of service. Shortages in power resulting
from forced outages, delays or maintenance on new generating
units, and fuel supply problems, require the use of reserve
generation to maintain a given level of reliability. Reserves,
therefore, must be maintained at levels sufficient to cover
anticipated possible power losses. Such power losses include:
an error in load forecast such that actual peak
load exceeds forecast by 15-20 percent,
the loss of the largest generating unit or plant
on a system,
the loss of the largest transmission line-on a
system, and
the loss of the largest interconnection with
imported power.
78
-------
The amount of reserve generation utilized in each individual
region ranges between 15 and 20 percent and is 'not uniform
across the nation. This range of percent reserves is generally
considered by the industry to be adequate, although 15 to 25
percent of capacity is preferred for- the mean unit sizes in
use today.
Since it has been shown in previous sections that use
of FGD on all new coal units will reduce system reliability, it
follows that FGD will increase system reserve requirements:
Hence, if a revised NSPS were adopted, additional generating
capacity would have to be added to the nation's utilities in order
to maintain levels of reliability achieved without FGD; required
amounts were seen in Section 5.1.2. Two factors regarding re-
serves, then, are relevant to a discussion of the impact of FGD
on power system adequacy. These factors are:
(1) how the additional generating requirements imposed
by widespread use of FGD will be met, and
(2) whether or not systems without FGD can be expected
to contain adequate reserve capacity in 2000.
These factors will now be addressed.
System reserve requirements are affected by many differ-
ent considerations. In order to increase generating capacity as
required by the use of FGD, several strategies could be employed;
these include the following:
79
-------
construction of larger or oversized generating
units
construction of additional generating units of
average size
utilization of interchange
modification of load, shape, i.e., load factor.
The effect of each of these strategies on system reserve require-
ments will not be discussed. First, as the average unit size
increases on a system, the percent reserve requirements must also
increase to maintain a constant loss-of-load probability. An
example of this requirement is shown in Figure 18. In general,
if all other variables are held constant, a unit larger than
average will affect reliability in a negative fashion in propor-
tion to the square of its size.2 For example, a 600 MWe unit
would have nine times the effect of a 200 MWe unit on reserve
requirements. Hence, oversizing of units reduces system re-
liability and increases reserve requirements. On the other
hand, reserve requirements are reduced as the number of units
increases because the magnitude of a failure is reduced. A
system with one unit with a capacity of X MWe would require at
least 100 percent reserves but a system with four units, each
with capacity X/4 MWe, would require only 25 percent reserves
to obtain essentially the same level of reliability. A similar
reduction in reserve requirements can be seen through -the use
of interconnections with other utilities. Load factor also
influences the reliability of a system; in general, the higher
the load factor, the greater the percent reserve requirement.
This is because a high load factor generally indicates a fairly
flat load shape with substantially equal loads throughout the
80
-------
z
Q
LU
o.
s
100,000
90,000
80,000
70,000
60,000
50,000
40,000
30,000
20,000
10,000
9,000
8,000
7,000
6,000
5,000
4,000
3,000
2,000
1,000
IM
1\
l\
il\
ll\
l\
1 \
1 \
I i
I
\
\
0 5 10
\
\
\\
A
\
V
\
( \
\
\
* X
\;
1
V
\
y
\
\
N
\
\
^
N
-
PERCEh
LOAD"
YEARS,
1.
2.
3.
4.
JT RESERVE REQUIRED FOR LOSS OF
rO BE' EXPECTED ONE DAY EVERY TEN
BASED ON:
GENERATOR FORCED OUTAGE RATE OF 2%
LOAD FORECAST ERROR - STANDARD
DEVIATION OF ±3%
SCHEDULED MAINTENANCE FILLS UP
SEASONAL LOAD VALLEY
SYSTEMS CONSIST ENTIRELY OF INDICATED
SIZE UNITS
"VyJ (JA
~*>+
\.%
\j
^
ff-
^
^C,.
1
\
\
L"/A-
X
I
"^N*.
^
\
's^x^^
--
15 20 25 30 35 40 45 50
REQUIRED PER CENT RESERVE
Figure 18.
Reserve requirements as a. function of unit size
Reprinted from 1970 NFS; no copyright.
81
-------
year. This type of load characteristic does not allow for
maintenance which is usually done during extended periods of low
load. Hence, the risk period is longer and additional reserves
must be maintained to cover units on maintenance status. Energy
conservation measures to increase base load and reduce peak load
as evidenced, for example, in time-of-day metering would, thus,
tend to increase percent reserve requirements from their present
levels. Reductions in load factor, on the other hand, reduce
percent reserve requirements at the expense of system economy.
Furthermore, it is presently quite difficult to control load
shape to the extent that the effect on reserve requirements is
highly significant. Therefore, it can be readily seen that the
feasible strategies for overcoming the effect of FGD are the
following:
(1) construction of additional generating units
of average size
(2) utilization of interchange
(3) modification of load shape
Since it is difficult to control load shape in order to reduce
reserve requirements and since it has been shown in Section 5.2
that interchange may be strongly constrained in the future, the
most feasible method of reducing reserve requirements or over-
coming a decrease in reliability caused by FGD would be to build
more generating units of average size. Since coal units are
presently and probably will continue to be the most economical
units to build and operate, it follows that decreases in reli-
ability (and, thus, increases in system reserve requirements)
caused by the widespread use of FGD, as called for in the revised
NSPS, will be met in almost all cases by the construction of new,
additional coal units of average size.
82
-------
It is important to consider the probable future
generation reserve situation in order to determine if any effects
of FGD might be mitigated or possibly amplified. Nationwide
reserve generation is expected to drop within the next few years
as a result of insufficient power generation expansion.17'18
According to NERC, this problem is a result of the long lead times
required for the approval and construction of new plants.
Figure 19 depicts large unit generating capacity to be
added nationally in the next ten years. Twenty percent of this
capacity is not yet under construction. It can be projected that
a certain amount of delay will be encountered with these gener-
ating units. NERC projects that coal-fired generating units which
are not presently under construction will probably be delayed
one year. Nuclear generating units which do not have construction
permits will probably be delayed two years, and those units which
are already under construction will be delayed one year because
of difficulties in obtaining operating licenses. These delays do
not reflect intentional delays caused by reduced growth in demand.
Figure 20 shows a potential deficit in national operating reserves
beginning in 1980 as a result of these delays in construction and
operation. If this trend continues, the expected peak demand,
as seen in this figure, could rise above generation capabilities
as early as 1990. Since generation reserves are not uniform
across the nation, each reliability council should be considered
separately. The following paragraphs discuss projections of the
probable ability of each reliability council to meet demand and
maintain sufficient reserves. These projections are based on a
future growth rate of 5.7 percent per year, as opposed to 7.0 -
7.5 percent per year used prior to 1977.
The individual reliability councils function according
to their own unique conditions. Consequently, the system reserve
policies and outlooks of these regions vary. Figure 21 shows
83
-------
FOSSIL AND NUCLEAR
GENERATING UMT ADDITIONS
(300 MW AND LARGER]
(CONTIGUOUS U.S.)
MEGAWATTS
X 1000
300i-
200 -
UNDER CONSTRUCTION
100 -
rt *L- ^- - - * 1^JL^.1^^^^^..^^^^
1977 1978 1979 1980 1981 1982 1983 1984 198S 1986
Figure 19.
NERC large unit construction forecast - U, S. through
1986. Reprinted from "7th Annual Review..." (Ref. 17)
by NERC, 1977, with permission.
84
02-2S71-1
-------
NERC (CONTIGUOUS U.S.)
CAPABILITY/PEAK LOAD/RESERVES
(SUMMER SEASON)
MEGAWATTS
X 1000
800J-
700 -
60O-
500
400
3OO
200
100
CAPABILITY
PEAK LOAD
RESERVES
1976 1978 1980 1982 1984 1986
PROJECTED
POTENTIAL
POTENTIAL DEFICIT
Figure 20. NERC reserve forecast for contiguous U. S. Reprinted
from "7th Annual Review..." {Ref. 17), by NERC,
1977, with permission.
35
02-2S72-1
-------
MEGAWATTS
X 1000
120
100
80
60
40
2O
PEAK LOAD
RESERVES
1976 1973 1980 1982 1984 1936
PROJECTED
- POTENTIAL
POTENTIAL DEFICIT
Fisure 21 NERC reserve forecast for Region 9 - WSCC. Reprinted
from "7th Annual Review..." (Ref. 17), by NERC,
1977, with permission.
86
02-2S74-1
-------
the reserve requirement of WSCC .(Region 9) . It should be noted
that a large portion of WSCC is supplied by hydroelectric genera-
tion where water availability is somewhat undependable. However,
'capability of supplying the peak load in this region does not
appear to present a problem in the near future, provided no
further water shortages are experienced.
MARCA (Region 5), Figure 22, could encounter a reserve
deficit as early as 1982. This region's generation capabilities
might not meet peak demand as early as 1986. Any further delays
in construction past about 1985 could result in severe power
shortages. Another similar situation is anticipated for SPP
(Ps.egion 8); Figure 23. The potential deficit for this region
could be as high as 10,000 MWe in 1986. SERC, (Region 7),
Figure 24, may also experience a deficit in 1978; however, a
downward trend in reserves might not actually develop until 1980.
SERC does not appear to have any difficulty in meeting peak
demand in the near future. Figure 25, MAIN (Region 4), follows a
trend similar to SERC. A potential deficit could begin as early
as 1979, but does not actually drop significantly until 1984.
SERC may, however, have trouble meeting peak load before 1990
if construction of additional generation capability is further
delayed. In the case of ERGOT, (Region 2), Figure 26, reserve
generation will probably remain high until 1984, when a potential
deficit could occur. Based on these NERC projections, it is ex-
pected that ERGOT will be capable of meeting peak loads through
2000. By 1990, ECAR (Region 1), could experience a power
shortage if peak load indeed becomes greater than generation
capabilities, Figure 27. ECAR's reserve generation may decline
in- 1980 with an increased deficit in 1984. MAAC (Region 3) and.
NPCC (Region 6), Figures 28 and 29, are the regions with the
most favorable outlooks. No trouble in meeting demand is pro-
jected for either region through 2000. The reserve generation
anticipated for both regions remains relatively sufficient;
however, after 1982, some deficiencies could occur.
87
-------
MEGAWATTS
X tOOO
sop
40-
30
20
10
PEAK LOAD
RESERVES
1978 1978 1980 1982 1984 1986
- PROJECTED
POTENTIAL
POTENTIAL DEFICIT
Figure 22. NERC reserve forecast for Region 5 - MARCA. Reprinted
from "7th Annual Review..." (Ref. 17), by NERC,
1977, with permission.
88
02-2578-1
-------
MEGAWATTS
X 1000
SO
70
SO
50
40
30
10
LOAD
RESERVES^
1978 1978 1980 1982 1984 1986
PROJECTED
- POTENTIAL
POTENTIAL DEFICIT
Figure 23.
NERC reserve forecast for Region 8 - SPP. Reprinted
from "7th Annual Review..." (Ref. 17), by NERC,
1977, with permission.
89
02-2573-'
-------
MEGAWATTS
X 1000
175
150
125
100
75
CAPABILITY
PEAK LOAD
RESERVES.
1376 1973 1980 1982 1984 1988
PROJECTED
- POTENTIAL
POTENTIAL DEFICIT
Figure 24.
NERC reserve forecast for Region 7 - SERC. Reprinted
from "7th Annual Review..." (Ref. 17), by NERC,
1977, with permission..
90
02-2575-1
-------
MEGAWATTS
X 1000
so-
40
30
20
10
CAPABILITY
PEAK LOAD
RESERVES
1976 1978 1980 1982 1984 1988
PROJECTED
POTENTIAL
POTENTIAL OEFICJT
Figure 25. NERC reserve forecast for Region 4 - MAIN. Reprinted
from "7th Annual Review.,." (Ref. I/), by NERC,
1977, with permission.
91
02-2579-1
-------
MEGAWATTS
X 1000
60
50
30
20
10
CAPABILITY
PEAK LOAD
RESERVES
1978 1978 1980 1982 1984 1986
PROJECTED
- POTENTIAL
POTENTIAL DEFICIT
Figure 26.
NERC reserve forecast for Region 2 - ERGOT. Reprinted
from "7th Annual Review..." (Ref. 17), by NERC,
1977, with permission.
92
02-2581-1
-------
MEGAWATTS
X 10OO
1201-
100-
SO-
60
40
2O
CAPABILITY,
PEAK LOAD
RESERVES
1978 1973 19SO 19S2 1984 1986
PROJECTED
-POTENTIAL
POTENTIAL DEFICIT
Figure 27. NERG reserve forecast for Region 1 - ECAR. Reprinted
from "7th Annual Review..." (Ref. 17), by NERC,
1977, with permission.
93
02-2576-1
-------
MEGAWATTS
X 1000
60 p
50-
30
10
LOAD
RESERVES^
1976 1973 1980 1982 1934 1988
PROJECTED
POTENTIAL
POTENTIAL OEFJCJT
Figure 28,
NERC reserve forecast for Region 3 - MAAC. Reprinted
from "7th Annual Review..." (Kef. 17), by NERC,
1977, with permission.
94
02-2S80-1
-------
MEGAWATTS
X 1000
60-
50
40
30
20
10
CAPABILITY/ ''
LOAD
1978 1973 1930 1932 1984 1938
PROJECTED
- POTENTIAL
POTENTIAL DEFICIT
Figure 29. NERC reserve forecast for Region 6 - NPCC. Reprinted
from "7th Annual Review..." (Ref. 17), by NERC,
1977, with permission.
95
02-2S77-1
-------
Should any further delays be encountered in new
generating unit or major transmission line installation and
operation, serious deficiencies in capacity could result. If
these delays result in systems being underbuilt in terms of
actual reserve capacity after 1985 as these projections
indicate, the impact of FGD would be amplified, since FGD has
been shown to reduce generating unit availability, and, hence,
system reliability.
96
-------
5.4 Sensitivity of Results to Projected Demand and
Generation Mix Statistics
The analysis in this study considered only new coal
units and their incremental contribution to system reliability.
Due to many unpredictable factors, the exact number of new coal
generating units which will be operational in the future and its
percentage of total generation cannot be known. The number of
units, moreover, affects the demand at any given plant and the
mix of base- and intermediate-load new coal units. Thus, it is
of interest to know how sensitive the results are to the demand
distribution and the mix of unit duty types.
Calculations were performed for a wide variety of cases
to investigate the sensitivity. These cases involved the follow-
ing load factors:
(1) 70% for base units and 55% for intermediate
units, and
(2) 60% for base units and 45% for intermediate
units.
The mix of base- and intermediate-load new coal units
was also varied. The fraction of new coal base units was varied
from 0 to 100 percent in steps of ten percent. This set of runs
was performed for the first set of load factors listed, above. It
was found that the following quantity Q was essentially invariant
as the load factors and mix were changed:
97
-------
where
X m available power minus power demand assuming no
scrubbers are used,
Xl =3 available power minus power demand assuming
scrubbers are used, and
* total estimated new coal capacity.
The quantity Q is the decrease in excess capacity
over demand when scrubbers are added as a proportion of total
capacity. Since demand is the same whether scrubbers are used
or not, however, Q can also be viewed as the decrease in total
power available due to the use of scrubbers as a proportion of
total capacity.
The invariance of Q is indicated by the fact that
for all load factors, mixes, and regions considered, the values
of Q were the same to -three significant figures . The conclusion
is that the proportional decrease in available power due
to the use of scrubbers is insensitive to quantities which had
to be projected into the future and which are therefore un-
certain. The variable Q, then, is a reliable output of the
study in this respect. The values of Q for various scrubber
configurations are given in Table 19- As expected, Q does
change as the number of scrubbers or the probability r changes
(r has been previously defined as the expected availability of
a scrubber module) .
The reason why Q is insensitive is that as the mix
o-r demand statistics are changed, the available capacity with
scrubbers and the available capacity without scrubbers are
affected in the same way, according to the assumptions in this
study. Actually, these parameters might vary slightly because
the FGD unit itself places a demand on the generating unit
98
-------
(energy penalty) which would not exist without FGD. From the
above definitions ,
XT = CNC - D(M) (13)
and
XT = Ck " D(M) (14)
where
CNC = new coal capacity available without FGD
C'
NC = new coal capacity available with FGD
D(M) = power demand (as formulated, D(M) is function .
of the mix of base- and intermediate-
load new coal units)
From Equation (12) , then
C - D(M) - (C' - D(M)
Q = _NC - - NC - (15)
UNC
and, thus
GNC ' CNC * D(M) + D(M)
c NC
From this equation, it can be seen that the power demand as a
function of mix cancels out, leaving Q invariant in terms of load
factor and mix of base-load and intermediate-load new coal units.
The incremental loss-of-load probability for a given scrubber
configuration, however, is very sensitive to such changes. The
incremental loss-of-load probabilities sometimes change by several
99
-------
tenths as the mix or demand is varied. The incremental loss-of-
load probabilities, then, are more sensitive to predictions of
future conditions and, thus, are probably less reliable than is
the quantity Q discussed above. However, these incremental loss-
of-load probabilities do not reflect system loss-of-load prob-
ability; they only represent the effect on. new coal units given
the assumption that the system without FGI) was designed in a
reasonable manner for a reasonable total system loss-of-load
probability.
The sensitivity of the incremental loss-of-load prob-
ability to load factor is demonstrated by Tables 20 and 21. It
is seen that as the load factors are decreased by ten percent,
the incremental loss-of-load probabilities sometimes decrease
dramatically. However, it should be noted that the 6070/45% load
factors vary considerably from the load factors associated with
the EEI outage data (Section 4.4); it may well be that unit outage
data used are invalid or nearly invalid for this case, since,
in reality, these load factors tend to reflect more intermediate
or peaking duty from an entirely different class of units whose
outage rates were not examined. The 70%/557o load factors used
as the basis for this study are more in line with the EEI data
and are thus, considerably more reliable.
100
-------
Table 19 . ADDITIONAL GENERATION REQUIRED FROM THE USE
OF SCRUBBERS AS A PROPORTION OF TOTAL NEW
COAL CAPACITY
Case
l(base)
2
3
4
5
Note: These
No. of Scrubber
Modules a
0/0
5/0
5/1
5/0
5/1
values apply for
Probability
r
.9
.9
.8
.8
all regions and
Proportional Def
icit in Genera-
tion
0.0%
4.5%
1.2% (Best)
9. 9% (Worst)
4.4%
different generation
mixes.
'(Active modules/spares)
Table 20 . INCREMENTAL LOSS-OF-LOAD PROBABILITIES (ILOLP)
FOR WORST CASE (5 ACTIVE MODULES/NO SPARES;
r - 0.8) LOAD FACTOR SENSITIVITY TEST
Generation Mix Used
Region (% Base/% Inter.)
1-ECAR
2 -ERGOT
3-MAAC
4-MAIN
5-MARCA
6-NPCC
7-SERC
8-SPP
9-WSCC
2- T _ J 1?-, _4-
25/75
50/50
0/100
60/40
70/30
10/90
10/90
50/50
90/10
ILOLP
Case 4a a
.30
75
.17
.82
.82
.27
.14
.74
.99
ILOLP
Case 4b b
.00
.04
.01
.12
.28
.02
.00
.04
".41 '
: c
-------
Table 21. INCREMENTAL LOSS-OF-LOAD PROBABILITIES (ILOLP)
FOR BEST CASE (5 ACTIVE MODULES/I SPARE; r = 0.9)
LOAD FACTOR SENSITIVITY TEST
Region
1-ECAR
2 -ERGOT
3-MAAC
4 -MAIN
5-MARCA
6-NPCC
7-SERC
8-SPP
9-WSCC
a Load
b Load
Generation Mix Used
(% Base/% Inter.)
25/75
50/50
0/100
60/40
70/30
10/90
10/90
50/50
90/10
Factors: Base = 70%,
Factors: Base = 60%,
ILOLP
Case 3aa
' .00
.08
.01
.17
.34
.00
.00
.07
.53
Intermediate =
Intermediate =
ILOLP
Case 3b b
.00
.00
.00
.00
.03
- .00
.00
.00
.00
55%
45%
102
-------
REFERENCES
Teknekron, Inc., The Integrated Investor and Municipal
Output (Derived from "An Integrated Technology
Assessment of Electric Utility Energy Systems,"
Study for EPA, Contract No. 68-01-194), EPA Reports
,PD 111 and PD 61M, Dec. 4, 1977.
Federal Power Commission. 1970 National Power Survey,
Washington: Government Printing Office,1970,
4 vols.
"1976 Annual Plant Design Survey," Power 1976 (November),
pp. S.4 - S.7.
"1977 Annual Plant Design Survey," Power 1977 (November),
pp. S.I - S.20.
"Design Criteria for a Standard Coal-Fired Plant for Ten
Western Utilities" - Fuel Outlook, Electrical World,
February 1, 1978, p. 52.
Edison Electric Institute, Prime Movers Committee, Equip-
ment Availability Task Force, EEI Equipment Availabil-
ity Summary Pv.eport on Trends of Large Mature Fossil
Units Categorized by Fuel and, in Commercial OperatTon
Prior to January 1, 1971. N.Y.. Oct. 1976.
Edison Electric Institute, Prime Movers Committee, Equipment
Availability Task Force, Report on Equipment Avail-
ability for the Ten-Year Period, 1965-1974. N.Y.,
November 1975.41 pp.
Federal Power Commission Bureau of Power Staff Report.
Electric Generating Plant Availability. May 1975.
60 pp.
Radian Corporation, The Energy Requirements for Control-
ing SO2 Emissions from Coal-Fired Steam/Electric
Generators.Prepared under Contract EPA-450/3-77-050a,
December 1977.
103
-------
10. Hitman Associates, Inc., Electrical Power Supply and
Demand Forecasts for the United States through 2050.
Prepared under Contract No. EHSD 71-43 USEPA, Office
of Air Programs. Columbia, Maryland, 1972. 57 pp.
(NTIS No. PB-209266).
11. Lyndon B. Johnson School of Public Affairs, the University
of Texas at Austin. Energy in Texas, Volume I;
Electric-Power Generation. Austin, Texas,1976.
116 pp. ~~
12. Resource Planning Associates. Energy Supply/Demand
Alternatives for the Appalachian Region, Executive
Summary!CEQ Report EQ 4AC-022, Cambridge, MASS.,
1975. 78 pp.
13. National Electric Reliability Council. Fossil and
Nuclear Fuel for Electric Utility Requirements and
Constraints.1977-1986.Princeton, New Jersey,
1977. 26 pp.
14. Letter from Ken Woodard, USEPA, to R. Dean Delleney,
Radian Corporation, Austin, Texas, June 15, 1977.
15. Conn, N. (ed.). Symposium on Scheduling and Billing of
Bulk Power Transfers. In: Proceedings of the American
Power Conference, Vol. 34, Chicago, Illinois, April,
1972, pp. 904-967.
16. National Electric Reliability Council. 6th Annual Review
of Overall Reliability and Adequacy of the North
American Bulk Power Systems!Princeton, New Jersey,
1976.28 pp.
17. National Electric Reliability Council. 7th Annual Review
of Overall Reliability and Adequacy of the North
American Bulk Power Systems.Princeton, New Jersev,
1977. 28 pp.
18. Javetski, J. W., "Central Engineers See Energy Conservation
as Central to Sound Engineering Economics," Power 1978
(January), pp. 60-62.
19. Fink, D. G. (ed.)- Standard Handbook for Electrical
Engineers, McGraw-Hill, New York, New York, 1969,
ppT 10-4.
104
-------
APPENDIX A
CALCULATION OF INCREMENTAL LOSS-OF-LOAD
PROBABILITY FOR NEW COAL GENERATION IN
A POWER SYSTEM
105
-------
APPENDIX A
CALCULATION OF INCREMENTAL LOSS OF LOAD PROBABILITY FOR
NEW COAL GENERATION IN" A POWER SYSTEM
In this Appendix, the analysis which was used to cal-
culate the incremental loss-of-load probability for new coal in
each geographical region is presented. An understanding of the
mathematics is not essential to understand the meaning of the re-
sults presented in the text of the report. The analysis is pre-
sented,, however, for completeness of documentation and for the
benefit of those who are interested.
The discussion in the sections below is organized
as follows:
(1) calculation of the mean and standard deviation
of capacity available at a given unit, taking
into account both boiler and scrubber down-time,
(2) calculation of loss-of-load probability for a
single unit, and
(3) calculation of loss-of-load probability for
a system of units among which demand within
a geographical region is distributed.
106
-------
CALCULATION OF MEAN AND STANDARD DEVIATION
OF AVAILABLE CAPACITY AT A PARTICULAR PLANT
In this section, the calculation of the mean and
standard deviation of available capacity at a particular
unit are discussed. The probabilities associated with scrubbers
will be discussed first, then (briefly) the probabilities
associated with the boiler, and finally, the probabilities of
various levels of power availability for the scrubber-boiler
system.'
Suppose there are exactly k scrubber modules and each
module has a probability r of being up at a given time . The
values r = . 8 or .9 and k =.5 or 6 were considered in this study.
Then, from the well-known binomial probability dis-
tribution,
P (exactly i modules are up) = m r1 (1...0 -r)k"i
LJ is ,
where L is , the binomial coefficient:
M _ k:
I il ~ IT(k
But it has been assumed in this study that only five
modules can be used at any given time. If there is a sixth,
it serves as a spare. Thus, if we denote the probability that
i modules are capable of being used by p., 0 <_ i<_5, then when
5,
and when k= 6 ,
107
-------
Pi " (i) r± (1-°-r)k"i
therefore, for five modules plus one spare module, it follows that
r6 (1.0_r)k-6
r5 (1.0-r)k-5+r6
Each scrubber module can handle at most 2070 of the
total capacity. Thus, when i scrubber modules are capable of
being used, the plant can operate at no more than [jj 100% of
total capacity.
The probability functions for percent of boiler
capacity available are discussed in Section 4.4. See Tables
9 and 10.
The probability functions for the boiler and the
scrubbers were then combined as follows :
P (0 capacity can be employed } = qQ + PQ - PQ qQ
where
q = probability 0% of the boiler is operational, and
P = probability no scrubber modules are available.
108
-------
Moreover, inverse cumulative probabilities can also be computed
P (at least j% of capacity can be used) =
P (at least j% of boiler's capacity is available)x
P (enough modules are available to handle at least
J7o of the total capacity)
where j = 20, 30, . .., 100.
These inverse cumulative probabilities were calculated
because of computational convenience; the probability
P (at least j% of capacity is used)
is obtained by performing a single multiplication, given that the
basic probabilities for the availability of the boiler and the
scrubber individually are known.
For the purposes of further calculations, however, in-
verse cumulative probabilities cannot be used. It is necessary,
therefore, to express the probabilities in the following form:
o^ = P (X..^ % of capacity is available)
where X. is a number between 0 and 100.
This is discussed in detail below.
The probabilities corresponding to intervals, such as
20 to 30% of the total capacity, can be obtained from the above:
109
-------
P (% of available capacity in between 20 and 30) =
P (% is at least 20) - P (% is at least 30)
The following special case was noted:
P (0 <% of capacity available <20) -
1.0 - P (0 capacity is available) -
P (at least 20% of capacity is available)
This case represents operation at levels usually requiring manual
control of the unit's boiler/turbine/generator system.
The result of the analysis above, then, is a table of
probabilities of the following form:
Probability
o-z
Percent of
Total
Capacity
0
.1-19.9
20 -29 . 9
Proportion
of Total
Capacity
0
.001-. 199
.200-. 29 9
Midpoint of
Interval of
Proportions
X! - 0
x2 - .15*
'x3 - '.25
90-99.9
100
.900-.999
1.0
x10 = .95
xn = 1.0
*Because of physical limits of the boiler/turbine/generator's
speed and voltage regulation control systems, it is felt that
running at very low percentages of capacity, such as 5%, is
unlikely. Thus, 0.15, rather than the interval midpoint, 0.10,
was used for this case.
110
-------
The mean and standard deviation of the proportion
of total capacity available were then computed as follows:
11
mean = E x.a.
i-l
11
std. dev. =« / I x?a. - (mean)2
where a and x are defined by the table above.
Ill
-------
CALCULATION OF LOSS-OF-LOAD
PROBABILITY FOR A SINGLE UNIT
In this section, the calculation of the distribution
of available capacity minus power demand for a single unit is dis-
cussed. From this distribution, the probability of not meeting
demand follows immediately. The basic probabilistic technique
used is called convolution; the mathematical solution requires
solving an integral known as the convolution integral.
First, we assume that the distributions for demand
and for available power are tabulated in terms of percent of
total power in steps of ten percent. Thus, the following data
will be used:
Percent of
Total Capacity (%)
0
10
20
30
40
50
60
70
80
90
100
Probability of
Demand
Probability of
Power Availability
PD2
PD3
PD*
PD5
PDS
PD7
PD8
PD9
PD10
PA 2
PA 3
PA 4
PA 5
PA 7
PA 8
PA 9
PA 10
112
-------
The probability of available capacity is obtained
by using the analysis discussed in the preceeding section.
Although, as calculated, this probability function is not tabu-
lated in steps of ten percent, such a table can easily be approxi-
mated. The probability that available capacity is 40 percent,
for example, can be approximated by
P (357, < available capacity < 45%) =
1/2 P (307, < available capacity < 4070) +
1/2 P (4070 < available capacity < 507»)
The Latter quantity is obtained directly from the
results of the last section. A similar type of approximation
was made with respect to 'the demand curves, which are discussed
in Section 4.2.
Now, we are interested in the characteristics of avail-
able capacity minus power demand, which we will call L. It is
desired to compute the probability of different values of L. We
will let X denote demand and Y denote available capacity. Then
L = Y-X
The random variable L, then, equals a particular value T
if Y equals a value YQ and X equals YQ-T, since
L - Y-X = YQ - (YQ-T) = T
To compute the probability that L equals T, then,
we will sum the probabilities of occurrence of the different
ways L can equal T. Consider the case L = 80, for example.
The following is a complete list of ways this can occur, given
that X and Y are tabulated in steps of 10:
113
-------
(1) X=0 and Y=80, which has probability (PDi)(PAg) ;
(2) X-lO and Y-90,'which has probability (PD2)(PAi0); and
(3) X-20 and Y=100, which has probability (PD3)(PAn).
The probability that L=80, then, is
(PDOCPAg) + (PD2)(PA10) + (PDaXPAn).
In general, the probability that L equals some value
T is calculated from the approximation of the convolution
integral as follows:
F(L-T) = z[P(Y=Yo) - P(X=YQ-T)]
where the sum is over the appropriate tabulated values of
Yo. The loss-of-load probability, P(L
-------
CALCULATION OF LOSS-OF-LOAD PROBABILITY
FOR A SYSTEM OF POWER PLANTS
In this section, the calculation of the loss-of-load
probability for a system of generating units within a geographical
region is discussed. It is assumed that demand is distributed
among all plants in the system. These calculations were used to
determine the incremental loss-of-load probability (ILOLP) for
only new coal units.
The demand probability functions for base and inter-
mediate units are discussed in Section 4.2. The mean p and
standard deviation cr of proportion of capacity demanded at a
particular site are computed as follows:
a = il Cj B± -(U)
where
C^ = i value of proportion of total capacity, and
B^ - probability that that amount of capacity will
be required to meet demand.
In the . manner indicated above, the following means
and standard deviations were computed:
115
-------
Standard
Mean Deviation
Base Units - u-g
Intermediate Units u-,.
To convert the mean and standard deviation from
proportion of capacity to capacity in megawatts (MWe) , the
following calculations are needed. The same equations apply
for demand as for available capacity.
Suppose a plant has capacity C. In the calculations,
one value of C was used to represent plants with capacity 390
to 599 MWe, and another value was used to represent plants, with
600 MWe or above. - The two specific values vary with region.
If the mean and standard deviation (of either demand or avail-
able power) in terms of proportion of. capacity are u and" a,
respectively, then the mean and standard deviation in mw are
GU and Ca.
Now, the mean of a sum is the sum of the means.
Thus, if there are N plants, each with mean available capacity
GU, the mean for the N plants combined is
N
E Cu - ITCu
i-1
The variance (standard deviation squared) of a sum
is the sum of the variances . The standard deviation of the
N plants is, therefore,
N
I (Ca)2 = Ca /N
-1 v
i-1
lie
-------
The above relationships, then, can be used to find
the mean and standard deviation of the total available capacity
or demand for a set of.power plants with the same total capacity
and demand curves. Suppose the following generation mix exists
for a region of the country:
no., of small (390 - 599 MWe) base units = NSB
no. of large. (600 MWe or over) base units = NLB
no. of small intermediate units = NSI
no. of large intermediate units = NLI
Then if Cg and C^ are the average capacities for the
small and large units, respectively, the following are the means
and standard deviations for the different classes of units.
Power Demand Statistics
standard
mean deviation
large base units £_ = c_ (NLB) M,, STW - C er
ijJj Li o LoJ L, B
small base units X-, - CL(NSB)un SeT}
ao a o SB
large intermediate units 5L T - C. (NLI)uT S,
JLI C^aT
small intermediate units JL C<,(NSI)nT SCT = CvaT
"* " i Oi. XI
The mean XQ and standard deviation SD for total demand
on the system, then, can be computed as follows:
AD ~ *LB ^B + ^1 + XSI
117
-------
Technically, if the random demands being added are
correlated, the variance of the sum is the sum of the variances
p lus a set of covariance terms. It is clear, moreover, that
the demands at two power plants in the same region are positively
correlated. Both demands would reflect the same general type
of diurnal cycle, and thus, for example, a cold front could
increase demand throughout the region.
An investigation of correlations among demands at
different plants is beyond the scope of this study. It is evi-
dent, however, that inclusion of positive convariarice terms
would increase the value of S^. The loss -of -load probabilities
which are reported in the text, and which are discussed below,
are influenced somewhat toward .5 due to the omission of the
covariance terms. While it is believed that the resulting
error is not excessive, further study of the effects of demand
correlations would be beneficial.
For a given case, all units in a region were assumed
to have the same number of scrubber modules (5 or 6) and the
same value of r ( . 8 or . 9) . The variable r is defined earlier
in this Appendix to be the probability that a given scrubber
module is up. For a given case, then, suppose the following
statistics regarding proportion of total capacity which is
available have been computed:
standard
mean deviation
small units
large units
Then the mean .and standard deviation of total power
available in MWe for the different classes of plants can -be
computed as follows:
118
-------
Power Availability Statistics
standard
mean deviation'
large base units
small base units UAS
large intermediate units XAT _-
ALI
small intermediate units X^-CgCNSI) u^ SASI=GSCTAS
The mean XA and standard deviation S. of power avail-
able for the system, then, can be computed as follows:
XA ~ XALB + XASB + XALI + XASI
SALB + SASB + SlLI
We are interested in the random properties of the
available power minus the power demanded. This difference L
has the mean value
' XA -
and the standard deviation
The probability that demand will not be met is the
probability that L is less than zero. The next question to
119
-------
address, then, is how to compute that probability.
Each region considered will have at leat 40, and, in
most cases close to 100 new coal units or more, and L is the
sum of
(1) the power available at each of the units, and
(2) the (negatives of the) power demanded at each of
the units.
Thus, L is the sum of approximately 200 terms or more.
Since sums of large numbers of random variables are generally
very nearly normally distributed, it is reasonable to assume that
L is normally distributed. (The Central Limit Theorem in
statistics is a rigorous statement of the normality property of
certain sums of random variables.)
The probability that L is less than zero, then, can
be expressed as follows:
P(L<0) = P /L - X. < -X, V. P Z <-X,
I ii L, | 1 L,
ST S,
where Z is a normally distributed random variable with mean zero
and variance one. The desired probabilit}?-, then can be found by
standard methods. For new coal units only, this probability,
expressed as ILOLP, is given by region and by scrubber con-
figuration in Appendix C.
120
-------
APPENDIX B
ESTIMATED LOAD FACTOR CALCULATIONS
BY REGION
121
-------
REGION 1 - ECAR
Unit
Type
Expected Total
Capacity MWe
Estimated Load
Factor
Old*
Pre-1986i
Coal
Oil
Nuclear
Post-1985-1-
Coal
0-390 MWe
390-599 MWe
600+ MWe
Nuclear
Other*
Regional Total
36,810
30,122
723
18,034
60.
1,990
111,200
70,254
18,357
287,550
67%
67%
67%
70%
50%
70%
70%
33%
67%
Conclusions - 100% base .
Used - 90% base/10% intermediate to reflect uncertainties
in load
^Estimated from 1970 NPS
tEstimated from Teknekran
122
-------
REGION 2 - ERGOT
Unit Expected Total _ Estimated Load
Type Capacity MWe ' Factor
Old* 12,038 57%
Pre-1986t
Coal -7,798 577,
Gas (/Oil) 1,718 577,
Nuclear 4,800 707,
Post-1985t
Coal
0-390 MWe 990 - 57%
390-599 MWe 3,679 "
600+ MWe 62,140
j
Nuclear 8,500 707,
Region Total 93,163 577,
Conclusions - 100% intermediate
Used - 10% base/90% intermediate to reflect uncertainties
in nuclear
^Estimated from 1970 NFS
tEstimated from Teknekron
123
-------
Unit
Type
Old*
Pre-1986t
Coal
Oil
Nucleart
Post-1985t
Coal
600+ MWe
Nuclear
Other*
REGION 3 - MAAC
Expected Total
Capacity MWe
22,590
2,SOOt-
3,020t
15,923
32,400
32,050
12,592
Region Total 121,075
Conclusion - 10% base/90% intermediate
Estimated Load
Factor
62%
55%*
55%*
70%
57.8%
7DT~
33%
62%
*Estimated from 1970 NFS
tEstimated from Teknekron
124
-------
REGION 4 - MAIN
Unit Expected Total Estimated Load
Type Capacity MWe Factor
Old* 27,536 61%
Pre-1986f
Coal 8,752 61%
Oil 2,755 61%
Nuclear 11,576 70%
Post-1985t
Coal
0-390 MWe 350
390-599 MWe 550
600+ MWE 49 , 200
Nuclear 9,350
Other* (mostly peakers) 10,886
Region Total 120,955
Conclusion - 70% base/30% intermediate
^Estimated from 1970 NFS
tEstimated from Teknekron
125
-------
REGION 5 - MARCA
Unit Expected Total Estimated Load
Type Capacity MWe Factor
Old* 14,529 61%
Pre-1986t
Coal
0-390 MWe 948 61.4
390-599 MWe 5,109 62%**
600+ MWe 4,885 65%**
Post-1985t
Coal
0-390 MWe 1,050 61%
390-599 MWe 500
600+ MWe 24,000
Nuclear 2,336 70%
Other* (peakers /pumped
storage) 5,929 30%
Region Total 59,286 61%
Conclusion - 70% base/ 30% intermediate
*Estimated from 1970 NPS
tEstimated from Teknekron
**Reflects large base-load Western coal units
126
-------
REGION 6 - NPCC
Unit Expected Total Estimated Load
Type ' Capacity MWe Factor
Old* 20,933 61%
Pre-1986t-
Coal
0-390 MWe 32 (probable peaker/
cycling) 50%
600+ MWe 2,246 61%
Oil 4,721 61%
Nuclear " 10,899 70%
Post-1985t
Coal
0-390 MWe 400
390-599 MWe 400
600+ MWe 27,850
Nuclear 59,800
Other* (mostly hydro/
pumped storage) 40,194 50%**
-'Estimated from 1970 NFS
tEstimated from Teknekron
**Reflects base-load hydro
Region Total . 167,475 61%
Conclusion - 100% intermediate
127
-------
Unit
Type
Old*
Pre-1986t
Coal
Oil
Nuclear
Post-1985t
Coal
0-390 MWe
390-599 MWe
600+ MWe
Nuclear
Other*
REGION 7 - SERC
Expected Total
Capacity MWe
30,640
12,862
4,659
27,946
2,789
2,273
76,209
146,361
12,656
Estimated Load
Factor
65%
657o
657=
7070
Region Total 316,-395
Conclusion - 307o base/7070 intermediate
*Estimated from 1970 NFS
tEstimated from Teknekron
12 3
-------
Unit
Type
Old*
Pre-1986t .
Coal
Oil
Nuclear
Post-1985t
Coal
0-390 MWe
390-599 MWe
600-1- MWe
Nuclear
Other* (hydro/pumped
storage)
Region Total
Conclusion - 100% intermediate
REGION 8 - SPP
Expected Total
Capacity MWe
17,712
17,097
225
11,774
1,975
2,068
68,820
19,150
Estimated Load
Factor
58%
58%
58%
70%
40%
58%
^Estimated from 1970 NPS
tEstimated from Teknekron
129
-------
Unit
Type
Old*
Pre-1986t
Coal
Oil
Nuclear
Post-1985t
Coal
0-390 MWe
390-599 MWe
6004- MWe
Nuclear
Other* (hydro)
Region Total
REGION 9 - WSCC
Expected Total
Capacity MWe
48,585
27,985
309
18,770
3,579
3,900
66,700
98,626
132,224
400,678
Estimated Load
Factor
63%
637.
63%
707.
637.
75.77.
707.
507.**
637.
Conclusion - 1007. base ...
Estimated - 907. base/107, intermediate due to uncertainties in
hydro
*Estimatect from 1970 NFS
tEstimated from Teknekron
**Reflects base-load hydro
130
-------
APPENDIX C
INCREMENTAL LOSS-OF-LOAD PROBABILITIES FOR NEW COAL
ONLY BY REGION IN 2000 - TEST CASE RESULTS
131
-------
APPENDIX C
INCREMENTAL LOSS-OF-LOAD PROBABILITIES FOR NEW COAL
ONLY BY REGION IN 2000 - TEST CASE RESULTS
This Appendix presents the incremental loss-of-load
probabilities (ILOLP) for new coal only by region for different
scrubber configurations and for different module availabilities.
Also, the base case (with no scrubbers) is given for each region
as a means of comparing the incremental ability of new coal units
in each region to meet demands placed upon them with and without
FGD.
132
-------
% of New Coal
Capacity Used
Scrubber
Region in Base Service
1-ECAR ' 90
2 -ERGOT 10
3-MAAC 10
4-MAIN 70
5-MAE.CA . 70
5-NPCC 0
Configuration
5/0*
5/0
5/1
5/1
0
5/0
5/0
5/1
5/1
0
5/0
5/0
5/1
5/1
0
5/0
5/0
5/1
5/1
0
5/0
5/0
5/1
5/1
0
5/0
5/0
5/1
5/1
0
Availability
of Each Module
.8
.9
.8
.9
.8
.9
,8
,9
.8
,9
,8
,9
,8
,9
,8
,9
.8
,9
,8
,9
,8
,9
,8
.'9
Incremental Loss-of-
Load Probability
(ILOLP)-New Coal Only
in 2000
.999
.885
.874
.535
.392
.165
.015
.014
.0022
.0011
.238
.0608
.0593
,0215
,0146
,900
.528
,516
,267
.198
,822
,528
,519
,338
.282
,186
,0493
,0483
,0185
,0129
*(No. active scrubber modulesAspares) - one scrubber of the given configuration
per each new coal generating unit.
123
-------
Region
% of New Coal
Capacity Used
in Base Service
7-SERC
30
8-SPP
9-WSCC
90
Scrubber
Configuration
5/0
5/0
5/1
5/1
0
5/0
5/0
5/1
'5/1
0
5/0
5/0
5/1
5/1
0
Availability
of Each Module
8
,9
.8
.9
.8
..9
.8
.9
.8
.9
.8
.9
Incremental Loss-of-
Load Probability
(ILOLP)-New Coal Only
in 2000
,407
.0562
.0534
.0091
.0045
.0795
.0045
.0043
.00048
.00021
.994
,833
.822
.534
.419
134
-------
BIBLIOGRAPHIC DATA 1. Report No. 2.
SHEET E-PA-450 73-78-002
4. Title and Subtitle
The Ability of Electric Utilities with FGD to Meet
Energy Demands
7. Authors) Dr. H. J. Williamson, Dr. J. B.
Dr. E. P. Hamilton III, Riggs ^ Miss T. j. Andersen
9. Performing Organization Name and Address
Radian Corporation
8500 Shoal Creek Boulevard
P. 0. Box 9948
Austin, Texas 78766
12. Sponsoring Organization Name and Address
Environmental Protection Agency
Office of Air and Waste Management
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
3. Recipient's Accession No.
5. Report Date
January, 1978
6.
8. Performing Organization Rept.
No-78-200-187-25-05
10. Project/Task/Work Unit No.
200-187-25
11. Contract/Grant No.
68-02-2608
13. Type of Report & Period
Covered
Final
14.
15. Supplementary Notes
16. Abstracts
Impacts of FGD on U.S. electric reliability and adequacy-through 2000 were
evaluated. C&al-fired-units on-line before 1986 and between 1985 and 2000 were con-
sidered for the nine National Electric Reliability Council (NERC) regions. Each
region's ability to meet power demand (with reasonable and typical reserves) as a
power pool with and without FGD was assessed. Different FGD module configurations
and assumed availabilities were considered. Power interchange capabilities which
might be used during FGD-induced outages were also evaluated, as were reserves.
It was concluded that a revised NSPS would have little effect on system
adequacy before 1985. By 2000, however, the NSPS would have a significant impact on
reliability and adequacy, requiring large amounts of additional generation to offset
the effects of FGD. Sensitivity of these results was analyzed and mitigating
measures were determined.
17. Key Words and Document Analysis. 17a. Descriptors
Electric Power Generation Reliability
Stack Gases-Desulfurization Processes
Outages-Electric Power Failures
Steam Electric Power Generation
Air Pollution Legislation-Regulations
Air Pollution Legislation-Cost Analysis
17b. Identifiers/Open-Ended Terms
EPA-450/3-78-002
Electric Utilities
Air Pollution Legislation
17e. COSATI Fieid Group 1QA
Release unlimited, additional copies from Radian
for $7.50.
19. iecuruy
Report'
UXC1
20. Security
Page
'-NC:
Class ' i his
-A5SIFIED
i-.ass ' i nis
-ASSlFtED
21.
22.
No. of Pa?es
147
Pries
1
THIS FORM MAY SE REPRODUCED
JSCOMM-OC
-------
------- |