ML20087P973
| ML20087P973 | |
| Person / Time | |
|---|---|
| Site: | Catawba  |
| Issue date: | 06/21/1983 |
| From: | Rosen R ENERGY SYSTEMS RESEARCH GROUP, INC., PALMETTO ALLIANCE |
| To: | |
| Shared Package | |
| ML20087P969 | List: |
| References | |
| 82-352-E, ESRG-82-59, NUDOCS 8404100164 | |
| Download: ML20087P973 (134) | |
Text
. .
w hcp c ESRG 82-59 r
(
TESTIMONY OF RICHARD A. ROSEN On behalf of Palmetto Alliance, Inc.
Docket 13o. 82-352-E June 21, 1983 ENERGY SYSTEMS RESEARCH GROUP, INC.
120 Milk Street Boston, Massachusetts 02109 (617) 426-5844 8404100164 840406 PDR ADOCK 05000413 N PDR
ESRG 82-59 TESTIMONY OF RICHARD A. ROSEN On behalf of Palmetto Alliance, Inc.
Docket No. 82-352-E June 21, 1983 ENERGY SYSTEMS RESEARCl! GROUP, INC.
I20 Milk Street Bosten, Massachusetts 02109 '
(617) 426-5844 -
T
__m ___m___ _ _ ____-_________m__ .
I QUALIFICATIONS Q. PLEASE STATE YOUR NAME AND BUSINESS ADDRESS.
3 A. My name is Richard A. Rosen. My business address is Energy 2
Systems Research Group, Inc., 120 Milk Street, Boston, 3
Massachusetts 02109.
4 5
O* ^
- A. ESRG is a non-profit organization specializing in research on 6
energy-related issues, particularly research related to 7
electric utilities. Among the electric utility issues which have been addressed by ESRG research are: demand forecasting, 9
conservation program analysis, electric utility dispatch and reliability modeling, generation planning, avoided cost analysis, financial analysis, demand curtailment modeling, rate design, cost of capital analysis, and district heating.
In addition, ESRG has done detailed analysis of r.uclear and coal plant capital costs, and nuclear plant capacity factors and operations and maintenance costs.
Q. PLEASE DESCRIBE YOUR BACKGF.JUND AND QUALIFICATIONS.
A. I am a senior research scientist at ESRG. In May, 1979, I completed directing a critique of the New England Power Pool Electric Demand Forecasting Model under contract to the New England Conference of Public Utility Commissioners. During 1980 I was project director of a study that culminated in 1
testimony by Dr. D. Shakow regarding " Generation Planning and Reliability" in Case #EO-80-57 before the Missouri Public Service Commission.
25
. _ _ ~ . . - =
1 I have presented expert testimony, in some cases on 2 numerous occasions, before the utility regulatory commissions 3 of Alabama, Indiana, Maine, Michigan, New Hampshire, North 4 Carolina, and Pennsylvania, as well as before the Federal 5 Energy Regulatory Commission and the Atomic Safety and 6 Licensing Board of the Nuclear Regulatory Commission.
7- My generation planning testimony before the state 8 commissions has included " Generation Planning and 9 Reliability" in Pennsylvania PUC v2 Philadelphia Electric 10 Company, Docket No. R-79060865 (the 1979 rate case), before 11 the Pennsylvania Public Utility Commission. I have also 13 submitted extensive direct and sur-rebuttal tastimony in 13 Cases No. I-79070315 and -317 ("CAPC Investigation) before 14' the Pennsylvania Public Utility Commission on generation 15 planning and reliability; in Case No. I-80200342 (the 16 " Limerick" Investigation); and on excess capacity of 17' Pennsylvania Power & Light in Docket No. R-822169.
18 I have also testified before the North Carolina 19 Utilities Commission on power plant performance standards and 20- fuel adjustment clauses. Further, I have recently presented 21 testimony before the Michigan Public Service Commission 22 analyzing the use that the Consumers Power Co. has made of 23 their own in-house dispatch model in preparing power supply 24 cost recovery factors.
25 Q. PLEASE DESCRIBE YOUR BACKGROUND BEFORE JOINING ESRG.
A. I received my Bachelor of Science degree from M.I.T. In 1966 2
and my Master's and Ph.D. degrees in physics from Columbia 3
University in 1970 and 1974, respectively. Before joining
.4 ESRG, I did research on industrial energy conservation at the National Center for the Analysis of Energy Systems at 6
Brookhaven National Laboratory, serving as principal investigator on projects involving industrial process energy 8
modeling for the U.S. Department of Energy. More information 9
on my background can be found in my vita, provided as Exhibit 10 (RAR-1).
i
l FINDINGS 1 Q. WHAT IS THE PURPOSE OF YOUR TESTIMONY?
2 A. The purpose of my testimony is to elaborate on the major 3 contentions made in the affidavit filed on behalf of the 4 Palmetto Alliance, Inc. in the recent court case '
5 83-CP-400044. A copy of this affidavit is attached as 6 Exhibit (RAR-2). The reopening of this docket permits me
'7 to present new evidence before the South Carolina Public 8 Service Commission which bears directly on its ultimate 9 finding-of mutual benefit of the proposed cale by Duke Power 10 Co. of a 25 percent undivided ownership interen, in Unit 2 of 11 the Catawba Nuclear Station to the Piedmont Municipal Power 12 Agency (PMPA).
13 Q. IN PREPARING YOUR TESTIMONY FOR THIS PROCEEDING, HAVE YOU 14 UNCOVERED INFORMATION THAT WOULD CAUSE YOU TO ALTER YOUR 15 CONCLUSIONS AS STATED IN EXHIBIT (RAR-2)?
16 A. No, I have not. My further review of the several analyses 17 that R.W. Beck has performed for PMPA over the last few 4
18 years has generally underscored the validity of the 19 conclusions advanced in the affidavit.
'20 Q. DO YOU STILL FIND A NEED FOR FURTHER STUDY OF THE ISSUES 31 BEFORE THE COMMISSION MAKES A FINAL JUDGEMENT AS TO MUTUAL 22 BENEFIT OF THb PROPOSED SALE?
23 A. Yes, I do. The record in this case is very incomplete as a 24 basis for any final determination as to mutual benefit.
- 25 Considerable additional study and review of the Beck 26' assumptions and mechodology is required. In addition, the
I intervenor (the Palmetto Alliance) should have the 2 opportunity to have the Beck analysis re-done with 3 alternative assumptions of its own choosing, to determine the 4 range of scenarios under which Beck's affirmative conclusion
~5 regording the proposed sale does or does not hold up. The 6 further analysis should include not only conventional supply 7- alternatives, but also the cost-effectiveness of an 8 accelerated customer conservation / load management program, a 9 viable resource planning option that (as I shall show below) 10 appears not to have been explicitly analyzed by R.W. Beck or
' 11 PMPA.
12 Q. WOULD YOU PLEASE SUMMARIZE THE ADDITIONAL CONCLUSIONS THAT 13 YOU HAVE REACHED IN YOUR RECENT REVIEW OF THE BECK RESEARCH?
14 A. Yes. I have reached six specific conclusions:
15 1. The load forecast upon which the R.W. Beck studies 16 are. based is flawed. It is founded.in significant 17 part upon utility personnel judgements that are not 18 made explicit, and therefore it cannot be
'19 independently reviewed or verified. The i 20- methodologies upon which it-is based are crude and do l _21 not represent state-of-the-art forecasting I
l 22. techniques. Finally, the forecast assumes a l
l 23 resumption of robust load growth after 1984, an l L24 eventuality against which a prima facie case can be i /
made, pending completion of any. systematic state-of-L '25 26 the-art load forecast that may be prepared for PMPA.
\;
y 2. Beck has not.made its assumptions regarding the operations and maintenance (O&M) costs for the 3
3 Catawba station clear. Beck may not have used 4
sufficiently high osM assumptiens in their most 5 recent study up date of November 15, 1982. If they 6 did not, this could prove to be a significant bias 7 in their study.
8 3. Beck has made no independent an& lysis of the capital 9
cost for Catawba, a figure that is crucial to its 10 economic conclusions. Furthermore, extensive analysis 11 that ESRG has performed on nuclear capital costs leads 12 me to believe that the cost of Catawba will likely 13 be about 20 percent higher than Duke and Beck have 14 assumed.
15 4. Beck has made no independent analysis of the assumed 16 capacity factor for Catawba. This parameter, too, has 17 an important bearing on the validity of Beck's 18 economic conclusions. Applying the results of ESRG 19 studies of nuclear capacity factors to Catawba 20 leads me to believe that the actual capacity factor 21 of Catawba (after the first few " immature" years) will 22 be about 10 percent below the level assumed by Beck.
23 5. Beck, again, has made no independent analysis of the 24 required capital additions for Catawba. Initial ESRG 25 analysis of this issue indicates that the Duke
' 26 estimates for the cost of capital additions, which
- l. Beck relies on, may ba underestimated by a factor of 2 four.
3 6. Not only are the benefits uncertain; there is a 4 serious' financial risk to PMPA ratepayers resulting 5 from such a large investment-in a single generating 6 source. The chances of incapacitating accident, 7 poorer than average plant performa'nce, or plant l 8 cancellation, are sufficiently high as to overwhelm 9 any possible economic benefit of the proposed 10 purchase. Thus the economic case for the proposed 11 PMPA purchase of 25 percent of Catawba #2 appears to 12 be fatally flawed.
13- Q. -WHAT ARE THE IMPLICATIONS OF THESE SPECIFIC FINDINGS?
14 A. Taken.together, they throw the Beck conclusion of a positive 15 benefit to PMPA ratepayers from the proposed project into 11 6 serious doubt. Just my alternative capacity factor estimate
_17 and the likely 20 percent higher Catawba capital cost 18 combined imply that the cost of power from Catawba could be 19 'about 40 percent more expensive than Beck has projected.
30 This is a substantial difference and could strongly reverse 21 -the benefit that Beck found in the sale of 25 percent of 12 2 Catawba to PMPA ratepayers.
23 Q. PLEASE DISCUSS THE SCOPE AND LIMITS OF YOUR REVIEW OF.THE 24 BECK RESEARCH.
p 25 A. I was able to~ analyze the basic Beck methodology, which in 26 general I found to be ' a L rease- ble one. However, the studies
-27' are quite complex, and of necessity they make many key
. . - . _ _ - - ~ ._- - - __..._._ _ .. _ --
i l' assumptions for the values of certain input parameters which a directly affect the results. There were many aspects of the 3 Beck studies that I could not review. For example, I have 4
not been able to review the details of how Beck modelled the 5 power interchanges between Duke and PMPA in the cases where 6 PMPA bought into Catawba versus the case when they did not.
7 Nor have I been able to review the financial modelling that 8 Beck did. Constraints precluded the type of review of these 9 areas that such a complex study demands. In addition, there 10 were insufficiencies in the information at my disposal.
11 I was able to review some of the methodology and results 12 that determined the values of certain key parameters used by 13 R.W. Beck in their original study and in subsequent updates.
14 The Beck parameters that I reviewed are:
15 1) The demand forecast for the PMPA service territory; 16 2) Operations and maintenance costs for Catawba; 17 3) The capital cost of Catawba 42 18 4) The capacity factor for Catawba #2 19 5) Capital additions for Catawba.
20 These are generally among the most important parameters in a
' 21. generation planning study. For each of these parameters with (22- the exception of the demand forecast, I have prepared alter-i
, 23 native estimates to the values Beck used, which were uniformly
'24 overly favorable towards the econoei of the Catawba purchase.
- 25 In the case of the demand forecast I will discuss reasons why
'26 I believe that the Beck forecast for PMPA is too high.
. = - .. .
1 Q. WHAT IS THE GENERAL IMPLICATION OF YOUR REVIEW OF THE BECK 2 ANALYSES?
3 A. Billions of dollars are at stake in this proposed Catawba 4 purchase. In the key areas listed above, I uncovered reasons 5 that throw Beck /PMPA conclusions into doubt. Thus, I believe 6 that the South Carolina Public Service Commission should 7 allow the intervenors in this case to perform a more thorough 8 review of the Beck study, and to request that PMPA have Beck 9- rerun its model with a set of input assumptions developed by 10 the intervenors so that it can be determined how sensitive 11 Beck's results are to the values they assumed for the key 12 input parameters. The hearings in this case should not be 13 concluded at this point.
14 Q. PLEASE DISCUSS THE FIVE AREAS THAT YOU HAVE INVESTIGATED IN 15 TURN.
16 A. I shall begin with a review of the demand forecast that Beck 17 has made for the PMPA service territory, discussirg its 18 adequacy in the light of historic load growth and changes in 19 load growth patterns during the past ten years.
l t
i l
l l
I l
I i
1
, ,w ,- ,-, ,, , , - - - -,----,.,r, m -n,---,- , ,-
LOAD FORECAST 1 Q. PLEASE DISCUSS HISTORIC LOAD GROWTH FOR THE TOTAL PIEDMONT 2 MUNICIPAL POWER AGENCY.
3 A. Table III-l in the 1980 preliminary report oy 9.W. Beck shows 4 a compound annual growth rate of 6.8 percent per year for 5 energy requirements from 1967 through 1979. Closer 6 examination of the numbers listed in that table reveals the 7 sharp discontinuity between the period through 1973 and the 8 years following 1973. The six-year growth rate in total
-9 energy requirements from 1967 to 1973 works out to over 10 10 percent per year, representing a very robust growth indeed.
11 On the other hand, the six-year growth from 1973 through 1979 12 works out to an annual growth rate of approximately 3.5 13 percent per year, thus representing a much reduced rate of 14 growth in energy requirements. There are numerous reasons, 15 absent any detailed and systematic projection of future load, 16 to expect a further decline in the rate of growth in the PMPA 17 region. Real price increases will continue, full industrial 18 recovery in the region is problematic, consumer. consciousness 19 of conservation benefits has increased, and energy management 20 firms actively solicit business from commercial and 21_ industrial anterprises concerned with their energy bills, l
l 22 among other factors. A declining rate of growth in energy 23 demand. translates, of course, into a declining rate of growth 24 in peak load, such as has also been experienced in the period 25 since 1973.
- 26 Q. PLEASE DISCUSS PMPA'S FORECASTS OF FUTURE LOAD GROWTH.
A. In the 1980 Beck report, as shown on Table III-3, the 1979 to 1
1990 rate of growth in energy requirements was forecasted at 2
3.2 percent per year. The rate of growth in the peak demand, 3
taking the 1990 forecasted peak of 414 megawatts compared to 4
the 1979 demand of 294 megawatts, was 3.17 percent per year.
5 Demand requirements were forecast to reach 580 megawatts in the 6
year 2000.
7 In the Beck supplemental report dated October 15, 1902,
! 8 there is no discussion of load forecast revisions, but 9
figures for future energy and demand requirements that differ 10 from those shown in the 1980 report appeared in the sections 11 entitled " Analysis A" and " Analysis B." The year 2000 dcmand 12.
is now apparently forecast to be 470 megawatts (Table B-1, 13 page 2), a full 110 megawatts less than in the 1980 forecast.
14 My review of the PMPA bond report " Preliminary Official 15 Statement Dated November,1982" shows that this reduction is 16 apparently due to a temporary halt in energy and demand 17 growth in the 1981-1983 period. Later, the rate of grow'5 in
-18 forecasted demand is almost the same in the 1982 report as it 19 was in the 1980 report. This may be seen by taking the years 20 1984 through 2000, where the annual growth rate'of peak 21 demand is 3.58 percent in-the 1982 report and 3.52 percent in 22 the 1980 report. Equally it may be seen by taking the years 23 1984 to 1990, where the rate of growth in peak demand is 3.73 24 percent per year in the 1982 report, and 3.65 percent in the 25 earlier report.
26
1 Q. DO YOU THINK IT IS REALISTIC FOR PMPA TO BE FORECASTING A 2 RESUMPTION OF THE 1980 FORECAST'S RATE OF INCREASE IN ITS I
'3 -PEAK DEMANDS?
4- A. No, in view of the sharp changes in load growth patterns we -
5 have seen in the recent past, both in the PMPA area and !
6 nationally, I do not. There is a prima facie case for 7 expecting that actual load growth will be lower. Of course, T
8 systematic, quantitative, and verifiable forecasting 9 procedures must be used to develop an accurate projection of 10 future energy and demand for use in system planning. i 11 Q. WHAT IS THE ROLE THAT A LOAD FORECAST SHOULD PLAY IN UTILITY 12 PLANNING?
13 A. A long-range load forecast Ls the basis of utility planning 114 to meet the service requirement of its ultimate customers.
15 Utility system planning should be geared to expected levels 16 of demand. Plans to purchase new capacity, as well as any 17 other system plans, should be predicated upon providing '
s 18 reliable service at the lowest cost to ratepayers. This in 19 turn entails' matching supply to demand.
L20 Q. FROM THE POINT OF VIEW OF SYSTEM PLANNING, WHAT NEGATIVE
'21 CONSEQUENCES MIGHT ENSUE AS A RESULT OF UNDER-FORECASTING?
22 A. The principal danger of under-forecasting is that the utility 23 will find itself obliged, on relatively short notice, to 24 purchase capacity or energy at a cost higher than that which 25 wculd have been incurred had it entered into such commitments
~26 to meet its. customers' needs at an earlier point in time.
c .
12 -
"- e' er *- +-w- y,qm-,g- tr w-y- *W--=meey-w=awW - s-v't g -EyyP e%y '-'y- g um-e m +vy*3y-r'y ye9
_ _ _ _ _ _ . . _ . _ . . - . . _. .. . ~ _
3- Q. WHAT IS 'THE PRINCIPAL DANGER OF OVER-FORECASTING?
3 A. From the. point of view of system planning, over-forecasting 3 can lead the utility tci make financial commitments to j capacity.that turn out not to be needed, or to be of a size 5 or type that is economically sub-optimal.
.6 Q. WE CANNOT FORECAST PRECISELY, CAN WE?
7 A. Of course not. Uncertainty is unavoidable; forecasting ;
8 simply endeavors to reduce it. State-of-the-art forecasting 9 techniques must be used to reduce the likelihood of over-10 forecasting or'under-forecasting. :
11- Q. HAVE YOU DETE.. MINED WHAT FORECASTING TECHNIQUE IS USED BY 12 PMPA TO REDUCE UNCERTAINTY CONCERNING FUTURE LOADS?
13- A. Yes, I have. I note'that in the R.W. Beck " preliminary 14 report" of 1980, it is stated (in Section III, pages 1-3) 15 that time trend analysis supplemented by econometric modeling 16 .and the judgement of utility personnel is used to forecast 17 future loads. More recently, in attachment V to Palmetto 18_ -Alliance interrogatory responses, the same approach is 19' restated: .the load forecast is produced by a time trend in 20 energy use, adjusted on tho basis of a single-equation 21 econometric model'and of "the judgement of eacli member's t
. 22_ . utility personnel."
4 :
23 Q. DOES THIS REPRESENT A STATE-OF-THE-ART FORECASTING
% ' METHODOLOGY?
'25 ;
- A. Certainly not. The only component of the PMPA approach.that i 26 comes close to so qualifying is the econometric component.
13 -
s Y
. - , . , , ,,,~,,u-,,,
. - ,.4 .e. y ,.,% ,...y, , , , , ,- ,.,.-e.,m._,,,,,,3om,,.m..,.,,-.~,..,..,., , . _, , , .,-.,+,3-. _
1 However, the use of one single econometric equation to a explain all energy use is not an acceptable forecasting 3 techniquer and 'in addition to using just one such equation, 4 PMPA does not tell us how it entered into the mix of trend, 5 equation and judgements that formed the resulting forecasts.
6 The PMPA forecasts are simply not based upon a reproducible 7 mathematical model. They are not accessible to independent 8 verification or analysis since they do not emerge from such a
- 9. systematic forecast model where the mathematical structure is 11 0 available for independent review. Even the particular
- 11. judgements and adjustments made with the relatively crude 12 technique employed, are not revealed or specified in any of 13 the available documentation. Consequently, there is little 14 reason.to regard the PMPA load forecast as reliable at this 15 time. In short, PMPA does not "make explicit the underlying 16 assumptions and judgements.that are inherent in any forecast,
-17 allowing them to be reviewed and evaluated in a systematic 18 .way, "-which is a criterion established by R.W. Beck itself 19 on page III-2 of the 1980 report.
20 Q. WHAT CHANGES ARE REQUIRED FOR PMPA TO PRODUCE A RELIABLE 21 FORECAST?
32 A. First, PMPA should move from forecasting energy and peak on
- 23 an aggregate basis and move toward disaggregating their load
, 24 - into its major components. Secondly, PMPA should move away 25- from trending and econometric tachniques and toward end-use 16 techniques in analyzing the components of load. Third,
. . ._ . _ - .. ~ ..-_. . ._-.
4 I
whatever PMPA does, all calculations leading to the final .
2 load figures should be fully documented, and the sources of ,
3 the underlying data upon which they are based explicitly 4 shown.
5 Q. IS A DISAGGREGATED APPROACH TO LOAD FORECASTING THEORETICALLY 6
SUPERIOR TO AGGREGATE ECONOMETRIC MODELLING? ,
7 A. Yes, it is. R.W. Beck is in error in identifying the j 8' '"statistical analysis" of econometrics with cause-and-effect 9' relationships, and equally in error in stating that the end-10 ' une modelling approach lacks "the causal sophistication of 11 the econometric approach." In actuality, the end-use 12 approach endeavors to identify the causes of electricity
' 13 consumption. Typically, an end-use model will obtain 4 14 electricity consumption in any given category by multiplying i
, 15 'the number of uses in the category times the average 16 consumption of each end-use. End-uses are the separate
. 17 components of final electricity demand. Examples are
. 18 refrigerators in homes and air conditioners in office 19 buildings. The disaggregated approach separately models as 20 many end-uses as ' feasible given the :1evel of data available.
21 To the extent that data constraints do not perm'it modelling
' l23 end-uses per se,'one.still endeavors to break down l
t - 23 electricity use into sectors and sub-sectors to get as close 7.
- 24 as poenible to t he physical processes through which l-l:25 electricity is consumed.
15 -
i r ,. . . . , , . , - ,,n... - - - _ _ .-,. n,-.- , . - - , , . , . - - . . , , , . , _ . , . , - - . , --c. ..m.. , n, - . + .. ,,,. n.,n .
i i
1 Q .' DOES THE UTILITY INDUSTRY USE THE END-USE APPROACH IN ITS
-2 FOk2 CASTS?
3 A. Yes, to a considerable and increasing extent. Despite recent !
4 and growing recognition that end-use modelling provides a 5 superior forecasting approach, PMPA does not rely upon this 6 approach even in part.
7 Q. IS AGGREGATE ECONOMETRIC MODELLING SCIENTIFICALLY SUPERIOR TO 8 END-UCE ANALYSIS?
9 A. No, the reverse is true. An aggregate econometric modelling i 10 strategy is based on developing quantitative relationships 11 between sales and a theoretically appropriate set of j 12 determining variables.- The quantification of these ;
13 relationships relies on estimating coefficients, using 14 techniques of-statistical inference. The problem of relying 4
l' 5 on a-limited 'ata d base to mathematically specify "best" 16- theoretical relationships between variables to be " explained" 17 and a set of determining (or independent) variables exists in P
1 16 all sciences addressing inherently stochastic systems, 19= whether they be physical, biological, or social. When 20 applied to forecasting, the procedure has come to be called "econometric" because of the presence of price and income l _ 21 -
1
- 22. variables among the set of independent variables, and the 23 reliance on a set of statistical tests and techniques often 24 used in macroecunomic analyses. But the validity of i 25_ econometric forecasting rests on the satisfaction of several L 261 basic premises. First, it must be assumed that the i-T' w- r- +W-' *-q, t ur- i--T+=* gew e$w 7+'-w r + vyag D- e t--v 'g w-------y se4 vpw'-yM-, y-WtwrwwwN+yi MM-Me' 'vw *w-*w-g-"y-Y--y yfw e'
- e L,I functional relationships between the variables gleaned from 2 historic data persists to good approximation into the future.
3 Second, it must be assumed that the primary causal structures 4 can be identified and adequately quantified by methods of 5 statistical analysis. Third, it must be assumed that 6 important causal factors that are not directly specified 7 within the econometric framework, or have no statistical 8- analog in the historic data base (for example, conservation 9 regulations, changes in the end-use mix in electricity usage, 10 the approach of many appliances towards full saturation, the 11 spread of energy-saving technologies, and fuel switching
~ 12' policies) either are adequately captured through the use of
' 13 . proxy variables like price, or are insignificant.
14 Q. ARE THE ASSUMPTIONS UNDERLYING THE APPLICATION OF AGGREGATE 15- ECONOMETRIC APPROACHES VIABLE IN THE PERIOD OF HISTORIC 16 ~ CHANGE BEGINNING WITH THE OIL EMBARGO OF 19737
. 17 A. No, they are not. The oil embargo in 1973, and subsequent 18 energy price increases and policy changes, have lead to
' 19 fundamental shifts in patterns of load growth, both 20- nationally and in the PMPA region. Stable relationships 21 between aggregate loads and aggregate indicators of 32 demographic or economic activity can no longer be assumed.
23 Thus, aggregate forecasting techniques of the type. employed 24 here are now generally not desirable. The problem with 25 aggregate econometric forecasting has been aptly summarized 1 26- by Dr. Charles Hitch, former head of-the Economics Division
[
1 of the Rand Corporation, and a member of the Electric Power 2 Research Institute (EPRI) advisory council. Writing in the 3 EPRI Journal of May, 1980, Dr. Hitch stated:
4 Certainly, projecting the relations that held before 1973 is not going to work in 5 the futurs because in the period up to 1973 we had very cheap energy plus, in 6 general, falling prices of energy and no period of rising prices. So we have 7 no historical period in which we have had rising energy prices. All the economic 8 projections based on relations before 1973 have led to gross over-estimates for energy 9 demand. Since we don't have any econometric projections and won't have any for some time, 10 except on a very short-term basis, we must use a different method: end-use analysis."
12 To Dr. Hitch's observation I would add that, because the 13 underlying technological factors are still in a state of 14 flux, it is not clear when, if ever, it might be reasonable 15 to place a basic reliance upon aggregate econometric 16 techniques to forecast a load pattern.
17 Q. DO YOU MEAN IT IS DESIRABLE OR POSSIBLE TO EXCLUDE 18 ECONOMETRIC TECHNIQUES FROM FORECASTING ALTOGETHER?
19 A. It is not, and may never be, possible to have a forecast 20 based on technological relationships alone. However, 21 forecasts should first be as disaggregated as possible, 22 consistent with the data available. Second, forecasts should 23 be based on direct technical relationships, rather than 24 aggregate econometric relationships, again.to the extent that 25 data permit. It may then be useful to use econometric 26 methods to predict particular components of demand,
incorporating such methods selectively within the framework of the disaggregated model. If any econometric techniques are used, one should strive to use the most specific data available and to structure equations so as to conform as closely as possible to the technical details of the process under consideration.
Q. YOU REFERRED PREVIOUSLY TO THE LACK OF APPLICABILITY OF PRE-7 1973 DATA IN THE CURRENT SITUATION. DOES THE PMPA FORECAST 8
RELY ON SUCH DATA?
9 A. We are not explicitly told. According to the attachment V referenced above, the 1967-1979 historical period is used as the basis of the time trending component. If the same years were used in performing the regressions that developed the one-equation econometric model, then the years 1967, 1968, 1969, 1970, 1971, 1972, and 1973 itself, or more than half of the years of data involved, would belong to the obsolete pre-1974 historical period. Moreover, it might well be that years before 1967 are used in the regression analyses; we simply are not informed.
19 Q. HAVE YOU EVALUATED THE FINAL ECONOMETRIC MODEL USED BY PMPA IN FORMING ITS LOAD FORECAST?
21 A. That model, with an evident typographical error therein, is given in attachment V referred to above. It is difficult to evaluate this model, not only because the years of data upon which it is based are not given, but also because the test statistics are not given and the way in which the equation's
1 results were applied in developing the actual load forecast 2 numbers is not given. However, it is possible, just on the 31 basis of the information given in attachment V, to establish 4 that the equation, while it might possibly be plausible as an
! 5 explanation of residential energy consumption alone, is i
- 6 not plausible as an explanation of commercial or industrial 7 energy censumption. The reason is that there are no terms in 8 the model to express any influence of either employment 9 growth or industrial activity upon energy consumption; and the 10 absence of such terms is not particularly plausible.
i 11 Q. MIGHT THERE BE SOME DIFFICULTY IN APPLYING A STRONGLY DATA-
, 12 DEPENDENT END-USE FORECASTING MODEL TO A SERVICE AREA AS 1 13 SMALL AND DISPERSED AS THAT OF PMPA?
- 14. A. Yes, there might. It is certainly far easier to develop
- 15 appropriate data bases for large, concentrated service areas.
j 16 However, there are two alternative ways in which the inherent 17 advantages of the disaggregated/end-use approach might be 18 brought to bear in forecasting PMPA's load. First, a
' 19 systematic forecast using an end-use model might be performed
. 20 for.the entire region within which the PMPA towns are 23 located. Then, the results of this larger load forecast l 22 ,
could be scaled to the PMPA region based on the historical
[ 23 ratios between energy sales in the PMPA towns and energy 24- sales in the rest of.the region.- As an alternative or
, - 25 supplement, secondly, it would be possible to develop end-use 26 ' forecasts-for_those sectors of consumption in the PMPA towns 1 for which data on appliance caturations and usage levels 2 could be developed. I would think this possible at least for 3 the residential sector. As a cross-check on either of these 4 approaches, econometric techniques might be used on a dis-5 aggregated basist they could be used to attempt to explain 6 residential, commercial, and indust y sales separately, and 7 the results of the distinct econometric equations considered 8 alongside the results of the end-use approach.
21 -
.p., yr* - --vv -- - + w -- ----- - - w --e -.c- -
OPERATIONS & MAINTENANCE COSTS
' F
.l' Q. ONE OF THE CRITICISMS THAT YOU MADE OF R.W. BECK IN YOUR
?-
2- AFFIDAVIT WAS THAT THEY DID NOT PERFORM A COMPREHENSIVE AND f
3 INDEPENDENT ASSESSMENT OF THE NUCLEAR O&M COSTS FOR THE -
'4 CATAWBA 92 UNIT. DOES THIS FINDING STILL STAND?
5- A. ' Yes, it does. In the course of reviewing the responses to 6 the interrogatories in this case, my initial view was confirmed.
~
7 The Duke response to Question 31 of Palmetto-
-8 Alliance (First Set) stated that "R.W. Beck requested that l
l 9- - Duke make an estimate of the Operating and Maintenance costs 10 for Catawba for the Piedmont Municipal Power Agency." Duke 11 then proceeds to explain briefly how this was done, and the J12 results are provided through 1992 in. Duke's answer to Question 30 of the same set of interrogatories.
13- Thus, R.W.
'14 Beck.did not..do their own assessment of Catawba OEM costs.
- 15 One noteworthy aspect of this Duke calculation is that 1
^
16 once OEM costs for the first full year (1988) were estimated at 17 $128.4.million for both Catawba units, this figure was then 18 escalated at 16.4 percent per year to 1992. However, the l
l 19 reason'for using 16.4 percent was not explained. In ;
.-20 addition, it is not clear whether or not R.W. Beck used this 21' LDuke estimate in calculating the results presented in their 22 - November 11, 1982 letter, nor is it clear whether or not they (23 continued to escalate the Catawba OEM costs at 16.4 percent l24 through the year 2000. This uncertainty arises from the 25 November 11 letter statement (p. 2), In preparation of the
- 26: Consulting Engineer's. Report, we increased the Catawba OEM
- 4
'-h-d *4 y py7 -m-we-eg h- g-pM g rm -w g wit- e ge-yq+gwr yo ig 1 org q
1 costs supplied on November 1, 1982, to system average nuclear 2 O&M costs for consistency and conservatism." The " Consulting 3 Engineer's Report" referred to is Appendix B of the PMPA 4 proposed Pre).iminary Official Statement draft of November 11, 5 1982.
6 Q. IS IT CLEAR TO YOU WHAT CHANGES R.W. BECK MADE IN DUKE'S 7 ESTIMATE FOR THE FUTURE O&M COSTS FOR CATAWBA?
8 A. No, because the basis for the O&M assumptions in the November 9 11, 1982 letter, though requested in Question #35, was never 10 received by me. The only new information on O&M cost 11 assumptions for Catawba that I have received from Beck is 12 Attachment III to PMPA's response to the first set of 13 interrogatories, Question 35, that seems to only have been 14 used in the Beck October 15, 1982 letter. In responding to 15 Question 31, Duke provided the methodology used in their most 16 recent update of November 1, 1982. From the PMPA 17 " Preliminary Official Statement" for their proposed November 18 bond sale (Appendix B, page B-29), it appears that Beck has 19 used O&M costs amounting to about 82 percent of those 20 estimated by Duke.
21 Q. HAVE YOU MADE AN INDEPENDENT ANALYSIS OF THE LIKELY O&M COSTS 22 FOR THE CATAWBA NUCLEAR STATION?
23 A. Yes. Using the results of the latest update to the ESRG 24 unalysis of nuclear plant operations and maintenance costs, I 25 have made a projection of the likely O&M costs for the 26 Catawba unit. The results in current dollars appear below, 27 compared to the Duke projections:
23 -
1 Year ESRG Estimate Duke Estimate 2 1985 36,290,000 28,493,000 1986 72,820,000 56,856,000 3 1987 116,190,000 91,917,000 1988 163,450,000 128,390,000 4 1989 176,930,000 149,446,000 1990 191,440,000 173,955,000 5 1991 207,040,000 202,484,000 1992 223,820,000 235,691,000 6
7 These ESRG results are derived from use of a multivariate 8 regression analysis that allows plant specific projections 9 for O&M costs to be made. It assumes a 6 percent underlying 10 inflation rate after 1983. The ESRG study is described in 11 Exhibit (RAR-3), appended below. The basis for the ESRG 12 results is a sophisticated statistical analysis of variables 13 that help explain the past O&M cost trends actually 14 experienced by almost all nuclear stations operating in the 15 U.S. through 1981. These variables include size, type, age, 16 cooling type, multiple unit status, and vintage. Thus, the 17 ESRG study prevides a more dependable approach to estimating 18 the likely Catawba O&M costs, given the high rates of 19 escalation in nuclear O&M costs during the 1970's, than does 20 either the Duke or Beck apprcach.
21 One example of this rapid escalation in O&M costs may be 22 found on the last page of Attachment III (provided in 23 interrogatory responses) where it is stated that Duke's O&M 24 costs for the Oconee Nuclear Station escalated at an average 25 rate of 30 percent per year from 1975-81. While the ESRG 26 regression equation projects a slowing down of this high growth rate, it still yields results that are 27 percent above the Duke estimates by 1988. By 1992, however, the ESRG result is somewhat below the Duke result. Therefore it is 4
important to know how Beck utilized the Duke results after g 1992 in performing any of their analyses.
Q. WHAT ARE THE IMPLICATIONS OF THESE DIFFERENCES FOR THE 7 CONCLUSIONS THAT R.W. BECK COMES TO IN ITS NOVEMBER 11, 1982 g
LETTER?
g A. Beck has found a narrowley positive economic benefit to PMPA ratepayers from the proposed project. O&M cost projections are important to this conclusion, but how they were actually arrived at.is shrouded in mystery. There is evidence that Duke's own OEM cost projections are not the best and if Beck is using O&M cost estimates for Catawba that are lower than the Duke estimates after 1992, it is probably projecting the cost effectiveness of the proposed PMPA purchase to be more favorable than it would be. Thus, the issue of how Beck treats future O&M cost escalation must be resolved, and g further analysis along this line conducted independently.
20 Al ng the way, certain specific puzzles with respect to Catawba OEM expenses must be cleared up. In response to the Palmetto discovery requests (Attachment B), PMPA has provided 23 a letter dated April 29, 1982 to Mr. Holmes of R.W. Beck from Mr. Hatley, Manager of the Catawba Special Group. In an attachment to that letter enclosed here as Exhibit
( R AR -4 ) , O&M expenses were listed for Catawba through the
1 year 1999. The interesting fact about these expense 2 estimates is that they are considerably higher than tha 3 November,1982 estimates made by Duke listed above, and they 4 are also higher than the ESRG estimates. Before any firm 5 conclusion can be drawn about the proposed sale, it is 6 important to know why Duke lowered its Catawba O&M estimates
-7 from' April, 1982 to November, 1982 (as provided in answer to 8 Question 30). Again, if the April, 1982 estimates generated 9 by Duke's Corporate Model Department were used in the Beck 10 analysis, the.results might change substantially. This is
- 11 especially likely if Beck did use the $30.28 per KW figure 13 for 1983 escalated at 9 percent as indicated in PMPA's answer 13 to Question 33.
- 14 Another puzzling fact is that Beck's November 11, 1982
'15 . letter;is the first place known to me where they mention that 16 Duke has changed its OEM escalation rate to 16.4 percent.
17 However, Exhibit (RAR-4) shows that this escalation was 18 used by Duke as early as April, 1982.
The question is why 19 Beck didn't use this higher escalation rate sooner.
I'-;
CATAWBA CAPITAL COSTS 1 Q. IN YOUR AFFIDAVIT YOU ALSO CRITICIZED R.W. BECK FOR NOT 2 HAVING MADE AN INDEPENDENT ASSESSMENT OF THE LIKELY CAPITAL 3 COST FOR THE CATAWBA 42 UNIT. DOES THIS FINDING STILL STAND?
4 .A. Yes, it does. As far as I know, R.W. Beck has made no 5 independent estimate of the capital cost of Catawba #2. Yet 6 the capital cost of the plant will probably have the single 7 largest effect on the cost effectiveness of the proposed 8 purchase for the PMPA ratepayers. Because the capital cost 9 projections have been a controversial and rapidly changing 10 aspect of economic analyses of nuclear power plants, the 11 absence of such an assessment is surprising. The capital 12 cost estimates for all nuclear units have been rapidly 13 increasing during their construction periods, often going up 14 by a factor of ten. Even on an inflation-adjusted basis, the 15 direct construction costs of nuclear units (the costs 16 excluding financing costs) have often increased by a factor 17 of four from initial planning to completion. This same 18 general pattern has affected Duke's estimates of the cost of 19 Catawba #2.
. 20 For example, as of May 10, 1983, Duke was estimating the 21 total direct construction costs for Catawba #2, without the 22 " profits on Contractual Entitlements" but with adpport 23 facilities, at $1,295,463,000 and for the entire Catawba 24 Station at S2,587,231,000. However, in the Preliminary 25 Engineering Report prepared by R.W. Beck for PMPA in August, 26 1980, the direct construction coste of the Catawba #2 unit
.. . . . - . ~ - . . - . - . - ... . .
1 were estimated at $1,063,856,000, with $2,114,055,000 3 estimated as the direct construction cost of the entire
.3 Catawba Station, presumably less some allotment for " Profits 4 on Contractual Entitlements." This represents an increase of 5- (at least) 22.4 percent in the cost of Catawba #2 in just 6 about 21 years. (These cost escimates cited here, and 7 discussed in this portion of my testimony, are exclusive of 8 any financing costs, costs for the initial fue7 '
,d, costs 9 for any of the special bond reserve funds, .c w, costs 10 that the PMPA bonds will have to go towards p , .. g . The 11 focus is on construction costs per s_e.) e 12 0 IS THERE ANY REASON TO BELIEVE THAT THE TREND OF INCREASING 13 COST ESTIMATES FOR THE CATAWBA #2 UNIT WILL COME TO AN END 14 WITH THIS LATEST MAY, 1983 DUKE ESTIMATE?
15 A. No, there is not. Generally when a nuclear unit still nas 16 'about four years to go until it reaches its commercial 17 perating date, as is the case for Catawba #2, projections of
.18 its total cost continue to escalate. One of the major causes 19 f this cost escalation are changing NRC construction 20 requirements for safety related purposes. Since there are 21 still many outstanding technical issues related to nuclear 22 power plant operations that NRC officially categorizes as 23 " Unresolved Safety Issues," the changes in NRC construction 24 requirements are likely to continue. In addition, further 25: delays in the construction schedule for Catawba #2 will also 26 cause the projections of direct construction costs for the 27 unit to increase.
1 ESRG has made a comprehensive statistical analysis of 2 the rate of change of direct construction costs for most 3 non-price-fixed nuclear units in the U.S. This analysis is 4 presented in Exhibit (RAR-5). The results show that even 5 before the Three Mile Island (TMI) Unit #2 accident, direct 6 construction costs for nuclear units in real terms (i.e.,
7 adjusted for inflation) vent up rapidly and steadily. This 8 general trend has continued through 1981, the most recent 9 year for which our analysis has been done.
10 Q. DOES THE ESRG NUCLEAR CAPITAL COST ANALYSIS YOU HAVE REFERRED 11 TO ALLOW YOU TO MAKE AN INDEPENDENT PROJECTION FOR THE COST 12 OF CATAWBA #2?
13 A. Yes, it does. The result of the analysis is the equation 14 listed on p. A-14 of Exhibit (RAR-5). The equation 15 results from a statistical fit to real dollar direct
'16 construction cost data for.52 nuclear units. Note that the 17 corrected R-squared for the equation is about 67 percent,
.18 which is quite high. Even more importantly, note that the 19 T-statistics of the key independent variables are significant 20 at greater than the 95 percent level. This equation allows 21 capital cost projections to be made for future nuclear units 22 as-a function of the date they were licensed (LICDATE), their 23 size, (MWDER), the experience of the plant architect-24 engineer-(AEEXP), the period of time over which they.are 25 constructed (PERIOD), whether they are a multiple station 26 (DUPLI), whether they are in the Northeast (NEAST), and 27 whether.they have a cooling tower (CTOWER). The equation 29 -
t new - . - , , , ,,, - e ,,--- - - , -- --
1 captures the key cost trends in the past that affect the 2 total unit costs, and projects them into the future. It 3 produces a probable value for the cost of any given unit 4 being forecast. Already there are many nuclear units being l 5 built in the U.S. that are currently being estimated by their 6 constructing utilities at costs above those predicted by the 7 ESRG equation. Similarly, there are many plants that are 8 currently assumed to cost less than the equation would 9 predict, 10 Q. HAS ANY PUBLIC UTILITY COMMISSION ADOPTED THIS ESRG NUCLEAR 11 CAPITAL COST EQUATION FOR USE IN A GENERATION PLANNING 12 HEARING?
13 A. Yes. In Docket No.81-114 concerning the economics of the 14 Seabrook Nuclear Staticn, the Maine Public Utility Commission 11 5 adopted this equation for use in estimating the cost of 16 Seabrook in its final order. The Maine Commission supported 17 my argument that by using this equation one was most likely 18 to obtain the most accurate possible prediction of Seabrook's 19 costs. They agreed that use of the other predominant cost 20 estimation procedure, viz. the architectengineering approach,
.21 .has proved to be unreliable precisely since cost estimates 22 generated via this approach are constantly changing over 23 time. (It is also noteworthy that in the same order the 24 Maine Commission supported use of our nuclear O&M methodology 125 as described above for calculating the future O&M costs for 2g Seabrook.)
y .Q. IF THE ESRG NUCLEAR CAPITAL COST EQUATION IS APPLIED TO 2 CATAWBA #2, WHAT IS THE PROBABLE COST OF THE UNIT?
A. At the end of Exhibit _ _(RAR-5), I present the results of 3
4 applying the ESRG capit.?.1 cost equation to Catawba #2. The 5 result is that for direct construction costs only, I project 6 that Catawba #2 has a probable ecst of $1,634,000. This 7 figure assumes 6 percent inflation after 1983. This is only 8 about 26.2 percent abose the most recent Duke estimate that I 9 have cited above. I believe that in the four years
~10 remaining f r the construction of Catawba #2, if it is yy completed by mid-1987, it is quite reasonable to assume that 12 Duke's estimate will rise by about 26 percent. In this 13 n ex , recall that Duke's estimate went up by about 22 g percent just over the last 23 years, so a 6 percent per year re-estimate, over each of the next fcur years, is r ot .
improbable. This would certainly be well below the average 16 g rate of change in Duke' estimate since the project began.
18 Two subsidiary conclusions follow from this ESRG cost 19 estimate. The first is that if interest during construction 20 is added to the direct construction costs at 10.3 percent, 21 the total cost for Catawba #2 will be S2.45 billion. The 22 second is that a higher cost estimate implies that Catawba #2 23- has reached a lower percentage of completion than the Company
- g. is presently stating. (When total cost estimates increase, 25; percent completion estimates drop.)
y --a ~ , , , - , - .- -y- - - - % -- w, w.y-w-n --,. - ,,,mw- . - , - - . - . . - w
1: Q. .IF THE ESRG COST ESTIMATE FOR CATAWBA #2 WERE CORRECT, WHAT g WOULD BE THE IMPACT OF THIS CHANGE ON THE ECONOMICS OF THE 3 ,
PROPOSED SALE OF 25 PERCENT OF THIS UNIT TO PMPA?
4 A. It is quite_likely that the 20 percent increase in the cost 5 of Catawba #2 indicated by the ESRG equation would reverse 6 the benefits claimed for the PMPA ratepayers by R.W. Beck in 7 their report and testimony. Consequently, PMPA should be 8 ordered to have R.W. Beck re-do its analysis with more 9 up-to-date key assumptions regarding O&M costs, capital 10 costs, and the capacity factor for the Catawba Station.
CATAWBA CAPACITY FACTOR 1 Q. WHAT DID R.W. BECK ASSUME IN MAKING THEIR ECONOMIC 2 CALCULATIONS FOR THE CAPACITY FACTOR OF THE CATAWBA UNITS?
- 3 A. As far as I am aware, Beck has used the same capacity factor 4 assumptions in all their studies, namely, that the Catawba 5 capacity factor would go from 56 percent to 69 percent over 6 the first nine years of operation. Assuming that Piedmont 7 fully enforces the new Duke buy-back provisions of the 8 contract in the period prior to the mid 1990s, the key period 9 is after the ninth year, when PMPA will be taking the output 10 of most of its 25 percent share.
11 0 IS THE INCREASE THAT BECK PROJECTS FROM 56 PERCENT TO 69 PERCENT 1
12 REASONABLE?
-13 A. No, according to an analysis of nuclear capacity factors that 14 I have done, it is not. A plant as big as Catawba is not 15 likely.to run as well as Duke or Beck is projecting, since 16 there is a general tendency for large nuclear units to have 17 poorer overall performance in terms of capacity factor than 18 smaller plants. This is confirmed by a third comprehensive 19 statistical analysis that ESRG has performed, dealing with 20 the capacity factors of almost all nuclear plants. This 21 study is presented here as Exhibit (RAR-6) and a complete
- 22 description of the methodology that ESRG developed can be 23 found there. Here I will briefly summarize the 24 considerations involved.
~ 25 A' great deal of data is now available concerning the 26 performance of base load generating units. With the aid of y this extensive data base, one can establie'i, via 2
multivariate regression analysis, connections between vari us haracteristics of a base load unit (age, type, size, 3
4 manufacturer, vintage, cooling system design, etc.) and the 5 performance of that unit as measured by its capacity or 6
availability factor over time. The regression equations 7 developed in the ESRG statistical analysis provide a guide to g the likely future average performance of units of a given 9
type, age, etc. These estimates indicate how well a 10 particular plant is likely to perform based on industry-wide 11 experience-for plants of a similar type. They provide an y objective standard cf plant performance that we might call 13 the " industry norm." This norm is calculated for nuclear g capacity factors-that have had the effects of refueling and 15 NRC-mandated outages removed, so we have labelled it an 16 " adjusted" capacity factor. The adjusted capacity factor is 17 the capacity factor over the hours of an operating year 18 exclusive of refueling and NRC-mandated outage hours.
1 19 From the ESRG study, and particularly from the resultant 20 regression equation for the adjusted capacity factor as 21 presented in Table C-4 on p. C-21 in Exhibit (RAR-6), the l
22 adjusted capacity factor can be calculated for each year of l- g operation of a nuclear unit such as Catawba #2. If the
-, 4 industry-wide average refueling rate of 13 percent is then 125 subtracted appropriately, one obtains the normal capacity
.26 factor for each year of operation. Given-the lack of data
1 for plants with ages of ten years or greater, the ESRG 2
equati n should only be directly applied up to that point.
en e ESRG capacity factor equation is applied to the 3
Catawba #2 unit, the result is a capacity factor of about 43 4
5 percent in its first year, increasing to about 59 percent in its ninth year. Again, this is a probable figure. The real 6
7 value has an equal chance of being higher or lower. However, I believe that it is very unlikely that the Catawba Station 8
w uld average 10 percent higher than the 59 percent capacity 9
factor that.I have projected after the ninth year, as Beck 10 g has assumed. I believe that no very large PWR unit like Catawba has actually run that well for any significant length 12 of time. Thus, I believe that Duke and Beck have certainly g been unduly optimistic in establishing their capacity factor assumptions for Catawba. Had Beck performed an independent assessment of this issue, a different assumption might have g been employed for this key parameter.
- ^
18
^
19 A. As is the case with changes in the capital cost, a change in 20 the capacity factor assumed for Catawba has a major impact on g Beck's calculations. The change from 69 percent to 59 percent alone increases the cost of power from Catawba #2 by
-23 about 17 percent. Coupled with the 20 percent increase in apital cost that the ESRG analysis indicates, the cost of 25 power from Catawba may actually prove to be about 40 percent 26 u
I higher than Beck has assumed due to only these two factors 2 alone. All of the Beck results and conclusions are thus 3 thrown into serious doubt. I believe there is little basis
-4 in the record of this proceeding for the Commission to
- 5 . conclude that the proposed PMPA purchase is in the economic 6 interest of PMPA ratepayers.
7 i
- n- =
c -m p-- N r--yy a er wa--' -r--, y +w- * -y-- + - g w -'+ye- g yw- =--1 e r- t W 4
CATAWBA CAPITAL ADDITIONS 1- Q. IS THERE ANY OTHER KEY PARAMETER THAT YOU BELIEVE HAS BEEN 2 SIGNIFICANTLY MIS-ESTIMATED IN THE BECK STUDIES?
3 A. Yes. I have reviewed the assumptions that Duke has made and 4
that Beck has relied on with regard to capital additions to 5 Catawba. The values projected are likely to prove far too 6 low. Again, Beck has done no independent analysis of this 7 issue. Moreover, Duke's responses to the Palmetto Alliance 8
interrogatories evidence little serious analysis as to how to 9- estimate such capital additions. Yet these capital additions 10 are very important in computing the cost of power from the 11 plant, and they are closely related to O&M costs. They 12 represent those aspects of equipment repairs and new l 13 equipment required to operate the plant that are put into the 14 rate base aach year. Thus, their impact on rates has both an i 15 immediate and long-term effect. From Attachment 3 to 16 Attachment B provided by PMPA in response to t,he Palmetto 17 interrogatories, we see that Duke has estimated that capital 18 additions for both Catawba units will rise from $5 million in 19' 1985 to $35 million in'1992. For a $4-$5 billion dollar 20 plant, this is much less than one percent per year spent on 21- capital additions by 1992.
i i -22 This is lower than the historic experience of the 23 utility industry would suggest. For comparison, note that
'2*- according to Mr. Eury, CP&L "had expended S167 million on
[:25 capital additions at Brunswick [from 1975 through 1982]. We 26 expect to spend an additional $233 million threugh 1986."
1 (p. 28 of Mr. Eury's direct testimony in NCUC Docket No. E-2, 2 Sub 461). Since these dollars are much less inflated than ,
3 the 1992 dollars of the Duke estimate will be, expenditures 4 by 1992 could well be almost double the $58.25 million dollar 5 a year average rate that CP&L projects for 1983-86 for 6 Brunswick, a much smaller plant than Catawba. This would 7 equal more than four times the level that Duke estimated on 8 per-kilowatt basis for Catawba.
9 Initial ESRG studies on the subject of capital additions 10 indicate that a realistic level of capital additions by 1990 11 for ' awba may be four to five times the Duke projection.
12 This factor will likely have a greater impact on the
.13 ' economics of the Catawba purchase than any inaccuracy in the 14 Beck estimate of expensed O&M for Catawba. This area clearly 15 warrants considerably more research.
4 . . - _ . .m. 4 .. ,
CONSERVATION OPTION
'l Q. WHAT EVIDENCE IS THERE THAT PMPA HAS ANALYZED THE FEASIBILITY 2
OF EMPLOYING A CONSERVATION AND LOAD MANAGEMENT PROGRAM AS AN 3
IMPORTANT ELEMENT ~OF ITS RESOURCE PLANNING?
4 A. I have found none, at least not in the documentation 5
available to me. Indeed, the R.W. Beck materials only refer 6-to conservation in the most general way, making little effort 1 -7 to estimate the levels of conservation that may have been
-8 attained in the several towns.
'9 Q. ARE THERE UTILITIES THAT HAVE INCORPORATED PROMOTION OF 10 cot:SERVATION AND LOAD MANAGEMENT AS RESOURCE PLANNING ELEMENTS?
11-A. Yes, there certainly are. Moreover, this is most commonly
- 12 the case in regions where load growth has been relatively
[ 13
! high, as in the PMPA area. In such regions of relatively 14 rapid load growth, the long-run marginal costs of production 15 tend to exceed the current average costs of production by a 16 significant amount. If the utility is going to be producing 17 new power at a higher cost than it is-charging on average, it 18
, . is losing money by expanding. An investment in conservation,
! 19' through a properly designed program, can stem the losses and 20' benefit the ratepayers. On account of such considerations, a
~ 21 number of utilities all around the country have embarked upon
' 22'
- l. deliberate programs to encourage and incentivize customer i ~23 l ; conservation and load management.- In fact, Duke Power 24 L : Company, itself, embarked upon such a strategy. For example,
. 25 houses that are weatherized to a high level qualify for a 26 special rate (the RC electric rate). In addition, customers 39 -
4
-a , - - - ,.,,-r.-- . ,e ,w. , , , - - - - - - ~.- .,- - . , - -
1 installing controlled water heating (where electricity is 2 drawn only during the off-peak periods when production costs 3 are lower) receive a credit from Duke Power. Many different 4 information programs, rate designs, and incentive programs 5 have been developed to accelerate the adottion by households 6 and businesses of energy management practices in these
. 7 various utility systems.
8 Q. BUT CAN SMALL MUNICIPAL SYSTEMS INCORPORATE CONSERVATION /
9 LOAD MANAGEMENT PROGRAMS?
10 A. They not only can, but a number have. For example, Vermont's 11 Burlington Electric Department floated bonds specifically for 12 the purpose of mounting a direct service conservation 15 program, in which utility personnel visit the house and 14 install certain no-cost / low-cost conservation measures 15 directly in the premises.
16 Q. IS THE BODY OF THE UTILITY EXPERIENCE WITH CONSERVATION 17 PROGRAMS DEVELOPED ENOUGH FOR A SMALL AGENCY LIKE PMPA TO'HAVS 18 EXAMINED FEASIBLE CUSTOMER CONSERVATION PROGRAMS DURING THE 19 TIMEFRAME OF THE R.W. BECK STUDIES?
20 A. Yes, it is, especially during the timeframe of the second 21 study. Currently, ample information is available to permit
. 22 PMPA to examine an. appropriate conservation / load management 23 strategy for near-term implementation. For the residential 124 sector, which is the largest single subcategory of PMPA load, 25 the body of experience is especially rich. I would recommend 26 that, in. addition to appropriate industrial / commercial 1
, .,.n. . ~ , - v.-. ~ ,
1- programs, PMPA immediately examine the cost-effectiveness of 2 offering such residential programs as:
3 1. Accelerated information to households, including 4 establishment of a1 centralized conservation " hot 5 line" manned by an energy conservation specialist.
6 2. Incentive programs to encourage insulation of 7 existing homes, in order to reduce demand for 8 air cooling and space heating. Programs that 9 have been successfully tried include Duke's RC 10 Rate, low-interest loan programs, and direct 11 service programs for installation of weatheriza-12 tion measures on the premises.
13 3. Simple rate designs to further encourage the use 14 of load management practices, such as time-15 controlled electric water heating.
16 4. Cash or credit rebate programs of temporary 17 duration to increase the percentage of customers 18 purchasing high-efficiency appliance models when 19 they are'in the replacement market, applying 20 certainly to air conditioners, and possibly to 21 refrigerators and efficient luminares as well.
22 I think some combination of such programs would 23 accelerate the rate of adoption of conservation practices by 24 customers, and reduce the pressure put on PMPA supply r
- 25 planning due to substantial load growth, which, though it is
-26 much reduced from past years, is still clearly likely to be 7 positive.
MAGNITUDE OF RISK TO RATEPAYERS 1
Q. IF FURTHER ANALYSIS DEMONSTRATED THAT A SLIGHT ECONOMIC 2
BENEFIT FROM THE PROPOSED PMPA PURCHASE WAS LIKELY OVER THE 3
LONG RUN, WOULD YOU THEN RECOMMEND THIS INVESTMENT TO PMPA 4
RATEPAYERS?
5 A. No, even in the unlikely event that a rigorous independent 6
analysis projected a slight economic benefit to PMPA 7
ratepayers from the proposed sale, I would not recommend that 8
these PMPA municipal systems would take the substantial risk 9
of making such an investment. After all we are speaking here 10 of an investment of about 015,000 or more per PMPA customer, 11 What if, for example, there is a problem with the North 12 Carolina Municipal Power Agency part of the purchase, and the 13 plant is cancelled after the PMPA bonds are issued? What if 14 after ten years the plant suffers a major incapacitating 15 accident such as TMI #2 has, or a prolonged NRC mandated
.16 outage? Or, more likely, what if the plant merely has below 17 - normal performance, i.e., below the levels that I have 16 forecast? After all, many plants'have had much poorer 19- lifetime performance than the average for any given plant 20 type. An example of very poor performance is quite close to 21 home.here, in the form of CP&L's Brunswick units. By 22 investing so heavily in a single project, PMPA is subject to 23 financial risks that are many times the size of the possible 2{ benefit-(if any) of this proposed purchase. Thus, from the 25 point of view of PMPA ratepayers, this proposed generation 26 planning decision makes little economic sense to me. PMPA
-. , , ,,we,, . y ..c - - - < ,,e+e , - - .
I can always get its power-from the complete mix of Duke's 2 generating units under their current type of rate agreements, 3 an arrangement far less frought with risk.
4 Q. DOES THIS CONCLUDE YOUR TESTIMONY?
5 A. Yes.
I l
i
, , , _ ~ . . , . _ - _,- _ _ _ , . . _ _ _ . , _ _ , - , . . , , , , , - _ , , . . . . . . . - - - _
EXHIBIT (RAR-1)
Richard A. Rosen Vita f
f l
V
Exhibit (RAR-1)
Sheet 1 of 5 RICHARD A. ROSEN i
)
4 Research Scientist Executive Vice-President Energy Systems Research Group Education Ph.D.: Physics, Columbia University, 1974 M.A.: Physics, Columbia University, 1969 B.S.: Physics and Philosophy, M.I.T., 1966 .
Experience:
1977 present: Energy Systems Research Group, Inc.
Responsibility for a broad range of research on industrial energy conserva-tion; electric generation planning issues; and modelling studies of long-range electric demand, utility system reliability, electric demand curtailment, and district heating systems.
1978 - 1980: Consultant to Brookhaven National Laboratory.
~1979: Consultant to the National Academy of-i Sciences, Puerto Rico Energy Study Committee.
1976 - 1978: Assistant Physicist, Economic Analysis Division, National Center for the Analysis of Energy Systems, Brookhaven National Laboratory.
1974 - 1976: National Research Council - National Academy of Sciences Resident Research Fellow, Goddard Institute for Space Studies, New York.
1973: Instructor - Putney - Antioch Graduate School.
I Testimony Case or AqEncy Docket No. Date Topic
~P&nnsylvania Pub- R-822169 Mar. 1983 Excess Capacity for lic Utilities Pennsylvania Power &
Commission Light Company-Ncrth Carolina E -10 0 , Feb. 1983 Power Plant Performance Utilities Com- Sub 47 Standards and Fuel
-micsion Adjustment Clauses Papago Tribal ER82-481 Dec. 1982 Overview of Conservation
- Utility Authority and' Generation Options
Exhibit (RAR-1)
Shont 2 of 5 Case or Agency Docket No Date Topic Kentucky Public- 83-14 Dec. 1982 Review of the Kentucky-Service Commission American Water Company Capacity Expansion Program Maine Public 81-276 Dec. 1982 As to the Avoided Costs Utilities Commission for Cogeneration and Small P w er Producers Public Utilities81-114 Nov. 1982 Maine Public Service Co.
Commission of the Investigation of Power Supply State of Maine Planning and Purchases Public Utilities81-174 Oct. 1982 Capital Costs of the Commission of the Seabrook Nuclear Units State of Maine Indiana Public Service 36818 Oct. 1982 An Economic Assessment of the Commission Marble Hill Nuclear Station New Hampshire Public DE81-312 Oct. 1982 Investigation Into Supply Utilities Commission and Demand of Electricity for Public Service Co. of N.H.
Michigan Public U-6923 May 1962 Consumers Power Co.
Service Commission Electricity Case Alabama Public Service 18337 Jan. 1982 Long-Range Capacity Commission Expansion Analysis State of New York SEMP II Nov. 1981 Conservation and Generation Energy Planning. Hearings Planning i Board l Pennsylvania Public 80100341 Sept. 1981 Operating and Captial Costs:
Utility Commission (Sur- Limerick Nuclear Station ebuttal) l I Maine Public MPUC 80-180 Apr. 1981 Electric Energy Costs:
-Utilities Commission (Sur- Seabrook Nuclear Power Plants rebuttal)
Pennsylvania Public I-80100341 Feb. 1981 Operating and Capital Costs:
Utility Commission Limerick Nuclear Generating S tation Ohio Public Utilities80-141 Dec. 1980 CAPCO Construction Program Commission EL-AIR (generation planning)
Michigan Public 'U-6360 Sept. 1980 Generation Expansion Plan-Service Commission ning: Consumers Power Co.
Pennsylvania Public I-79070315 Aug. 1980 CAPCO Construction Schedule Utility Commission (Sur-rebuttal)
... . .. , - - . ~ , . .
Exhibit (RAR-1)
Sheet 3 of 5 Case or Agency Docket No. Date Topic Ccnnecticut Power- F-80 June 1980 Renewable-Resource Fccility Evaluation Electric Generation in CT Council Pcnnsylvania Public I-79070317 March 1980 CAPCO: Generation Utility Commission Planning and Reliability t
Michigan Public U-5979 June 1979 Forecast Critique and SCrvice Commission Adjustments: Consumers Power Co.
Mtssachusetts De- -19494 Aug. 1978 Long-range Electric Demand partment of Public Forecast: Boston Edison Co.
Utilities Pcnnsylvania Public 438 March 1978 Long-range Forecast of
- Utility Commission Electric Energy and D emand (Philadelphia Electric Co.)
ESRG Research l October 1982: The Economics cg, Closing the Indian Point Nuclear l- Power Plants. ESRG Study No. 82-40. Principal investigator.
f Cctober 1982: Review of the Kentucky-American Water Company Capacity Expansion Program. Final Report of the Kentucky Public Service Commission. ESRG Study
, No. 82 -45. Co-author.
August 1982: Nuclear Capacity Factors: The Effects of[ Aging and Salt Water Cooling , A Report on Research in l Progress. ESRG Study No. 82-81. Co-author.
August 1982: The-Impacts of Early Retirement of Nuclear Power Plants: The Case of_ Maine Yankee. ESRG Study No. 82-91. Co-author.
April 1982: A Power Supply and Financial Analysis of the Seabrook Nuclear Station as a Generation Option L for the Maine Public Service-Company. ESRG Study No. 81-61. Principal investigator.
January 1982: Guidelines for Designing Rates for Sales to, g . Qualifying Facilities Under Section 210 of the l Public Utility Regulatory Policies Act. ESRG
- l. Study No. 81-32. Co-author.
Exhibit (RAR-1)
Sheet 4 of 5 June 1981: An Analysis of the NSSd for and Alternatives to-~~
the Proposed Coal Plant at Arthur Kill.
A Report to: . Robert M. Eerzog, Director, New York City Energy Office and Allen G.
Schwartz, Corporation Counsel for the City of New York. ESRG Study No. 81-21. Co-author.
. October 1980: The ESRG Electrical Systems Generation Model:
Incorporating Social Costs in Generation Plannino,
, ESRG Study No. 80-12. A Report to the U.S.
, Department of Energy. Co-author.
September 1980: Reducing New England's Oil Dependence Through Conservation and Alternative Energy, ESRG Study No. 79-29. A Report to the U.S. General Accounting Office. Co-author.
- July 1980
- Preliminary Economic and Need Analysis of the Proposed Brumley Gap Pumped Storage Facility for the AEP System, ESRG Study No. 80-08/P. Principal investigator.
July 1980: The Potential Impact of, Conservation and Alternative Supply Sources cp_ Connecticut's
- Electric Energy Balance, ESRG Study No. 80-09, A Report to the Connecticut Power Facility Evaluation Council. Co-author.
1 i November 1979: South Carolina Electric Demand Curtailment 4
Planning, ESRG Study No. 79-31, A Report to the
~
South Carolina Office of Energy Resources.
Principal investigator.
t
. - May 1979: Demand Curtailment Planning: Methodology, ESRG Study No. 78-18, Chapter submitted to Brookhaven National Laboratory and the Department of Energy for the Electric Demand Curtailment Planning Study. Principal investigator.
May'1979: Assessment of the New England Power Pool -
Battelle Long Range Electric Demand Forecasting
[ Model, ESRG Study No. 79-06, A Report to the New England Conference of Public Utility Commissioners. Co-principal investigator.
October 1978: - The Employment Creation Potential of Energy Conservation and Solar Technologies: The Implications of the Long Island Jobs Study for New England, 1978!1993, ESRG Study No. 78-16.
, Co-author.
4-
, , ,n, - ,,,-,-r .--n--wwn -c--, w ,,-wr, + w- vw m a, - n n w ,-c,,,,a ,,---rev,,--,ra, v-, +,,,a-,-
I Exhibit (RAR-1)
Sheet 5 of 5 Nov:mbar 1977: Profilo of[ Targets for the Energy Advisory Service to Industry, ESRG Study No. 77-09, A Report to the New York State Energy Office. Co-Author.
October 1977: The Effect on Air and Water Emissions of Energy Conservation in, Industry, ESRG Study No. 77-04.
Co-author.
July 1977: The Effects on Air and Water Emissions of Energy Conservation in, Industry, ESRG Study No. 77-04.
Co-author.
June 1977: Toward an Energy Plan for New York, ESRG Study No. 77-03, A Report to the Legislative Commission on Energy Systems. Co-author.
April 1977: Assessing Demand, Alternative Operating Strategies, and Utility Economics in the Service Territory cd[ Orange and Rockland Utilities , E5RG Report No. 77-01. Co-author.
Other Publications March 1978: The Use ci the Pulp and Paper Industry Process Model for R&D Decision Making, Brookhaven National Laboratory Report No. BNL 24134. Co-author.
1976: A Non-Linear Model for the Linewidth, Intensity, and Coherence of Astrophysical Masers ,"
Astrophysical Journal vol. 190.
Prpers July 24-28, " Energy Use Modelling of the Iron and Steel 1978 Industry," Summer Computer Simulation Conference.
November 12, " Energy Conservation in Industry," Northeastern 1977: Political Science Association meeting, Mt. Pocono, Pennsylvania.
Awards and Honors 1968 - 1974: Faculty Fellowship, Physics Department Columbia University.
1966 - 1970: New York State Regents Fellowship.
1967 - 1968: Adam Leroy Jones Fellow in Philosophy, Columbia University.
EXHIBIT (RAR-2)
Affidavit of Richard A. Rosen L
Exhibig (RAR-2)
COMMONWEALTH OF MASSACHUSETTS ) Affidavit of COUNTY OF SUFFOLK ) Richard A. Rosen I am Richard A. Rosen, Ph.D., Executive Vice-President of Energy Systems Research Group (ESRG), a non-profit multidiscip-linary research organization with offices in Boston, Massachusetts.
A description of ESRG experience and qualifications appears below at Part III, and my resume appears at Part IV.
This affidavit represents my preliminary analysis of the proposed sale by Duke Power Company of a 25 percent interest in its Catawba Unit 2 to the Piedmont Municipal Power Agency (PMPA).
In this analysis I identify the material issues that should be the subject of further detailed study in order to determine the " benefit"
! 'to the towns of the proposed purchase. I show, first, that recent studies by R.W. Beck are flawed and not adequate to reach a con-I clusion concerning benefit; second, what the components of a study adequate to reach such' conclusions would be; and, third, the will-ingness and ability of my organization, Energy Systems Research Group, to conduct and conclude such a study.
This affidavit shall make reference to the six elements that are the minimum basis for a determination of project benefit as listed in Section 6-23-60 of the Code of Laws of South Carolina,
- these being the economies and efficiencies from the project; the need for power and energy; project useful life; time required to implement project; alternatives to the project; and forecasted i
load. An 7.dequate study covering these six elements and other Olements relevant to the issue of benefit can be performed in a period'of three months. It would-recuire that a full and adequate discovery from Duke and PMPA be permitted.
I. REVIEW OF PMPA ANALYSIS.OF CATAWBA PURCHASE
Background
In August, 1980 R. W. Beck and Associate submitted its
" Preliminary Engineering Report" to the Piedmont Municipal Power Authority (PMPA) analyzing the economic feasibility of a purchase my PMPA of a 25% ownership interest in the Catawba #2 nuclear unit. - This initial analysis was then followed up with two up-dated revisions of the earlier calculations (October 15, 1982 and November 11, 1982) which involved changes in a large number of key assumptions. The central focus of the Beck analysis was the sconomies of the proposed purchase from the point of view of the
.PMPA ratepayers. The Beck studies contain no substantial analysis of the benefits of the proposed sale of 25% of the Catawba #2 unit from the perspective of the Duke Fower Co. ratepayers. To the extent, therefore, that South Carolina law requires such an analysis,-this requirement has not been fulfilled by the Beck studies.
In performing the type of complex generation planning study that i
R. W. Beck did on behalf of PMPA, it is very important that all the major numerical assumptions receive a full and independent review.
Major assumptions employed in the Beck studies were made by the l
l Duke' Power Co. with respect to the capital. cost of the Catawba #2 unit and alternatives to it, the cost of fuels, the cost of the purchased power that Duketwould sell to PMPA if they did not
-buy into Catawba, the cost of operating Catawba #2, and the cbility of this unit to operate efficiently within the Duke Power 2
l a - ,+y--=-g- 7-m>+- e.- aw -e-w q g y----
system. System load growth, as forecast by EMPA, is also a key assumption. Yet in most cases R.W. Beck did not meet the standard cited above: they did not independently derive values for the important assumptions utilized in their study. This critical omission will be described in greater detail below, in the context of a discussion as to what a more comprehensive study than the one performed by R. W. Beck would consist of for each key area of analysis.
Results of this Review In the most recent study update (November 11, 1982) Table 1 of the Beck letter to PMPA indicates that the changes in key assumptions that have arisen just in the month since mid-October have significantly reduced the earlier projected favorable econo-mics of the proposed purchase of the Catawba facility for PMPA ratepayers. In fact, without " rate stabilization" the proposed purchase is now projected to be a money los,er through the year 2000. Generally, the changes in assumptions just_between Beck's October 15 status report to PMPA and its November 11 report caused the 1983-2000 savings to drop by more than 20%. Yet, as we shall describe below, R. W. Beck had not made updates to independent estimates of the PMPA load forecast, the capital cost of Catawba 42, the likely capacity factor of Catawba #2, and the cost of conservation as an alternative to Catawba #2. Thus given these inadequacies in even the most recent Beck analysis, we believe that it is not possible to rely on their results to determine:
3
whether the sale to PMPA of 25% of Catawba #2 is to the benefit of the PMPA ratepayers. The remedial analytical measures that follow must be undertaken before such a study can be relied on by the South Carolina Public Service Commission in carrying out Section 6-23-60.
Base Case Load Forecast An independent " Base Case" or " business-as-usual" load forecast for peak demand and. energy sales must be the first stage in a generation planning study. Achieving an accurate estimate for future load growth is critical in terms of assessing how much new generating plant capacity is needed, and what type of capacity it should be. The R. W. Beck load forecast assumptions in the most recent study update are seriously flawed in two ways. First, the assumptions appear to rely on forecasts made in the " Preliminary Engineering Report" of August, 1980 thus are two years out-of-date.
As Beck states in their October 15 update "Our projections of electric power and energy requirements assume that the Piedmont region of the states of North and South Carolina will continue to experience moderate economic growth. . ." (p. 13). They do not mention here any revision to their 1980 demand forecast. Yet"by October of 1982 there was increased reason to qualify the long-term assumption of " moderate economic growth".
More importantly, in the 1980 report Beck states that their PMPA forecast was dependent on trend analysis and econometric techniques. However, we find that tnese are not state-of-the-art techniques. In the next sections of this affidavit, describing 4
the type of work that should be done, I will describe the need, for the sake of accuracy, to incorporate more detailed analysis into the load forecast than has apparently been done by PMPA.
The likely consequence of an inadequate forecast is that the need for Catawba #2 for both PMPA and Duke Power will be over-estimated. In fact, just last month the North Carolina Utilities Commission has indicated that they believe that Duke Power's demand growth rate will be about 1. 5% per year rather than the Company's estimate of 2.8% per year. These figures sharply contradict with the 3. 4 % that R. W. Beck has used in 1982 for the PMPA load growth rate; and calls this key Beck assumption into serious question.
The implication of this likely error on Beck's part is that the Catawba #2 unit may not be required or economical for either PMPA or Duke ratepayers (or other South Carolina ratepayers) in the 1983-2000 time frame.
Conservation Assessment It is generally accepted among professional energy analys ts l that there remains a substantial potential for conserving a kilowatt-hour of electricity through efficiency improvements at less than the cost of supplying new kilowatt-hour with a new generation station such as Catawba #2. Thus another major flaw in the Beck study is that the economics of utility promotion of customer conservation l or load management as an alternative (see 6-2 3-6 0 (n) ) to the
_ project was not addressed. Without such a prior assessment it is incomplete to proceed to analyze the cost / benefit of Catawba #2 for PMPA ratepayers as Beck did. In the next section of this 5
1 affidavit I describe the incorporation of conservation promotion as an " alternative" into the cost benefit assessment of the project. '
To the extent that the (properly forecasted) load growth can be moderated through conservation and load management, the rate of 3
load growth can be abated. Duke has a deliberate strategy to do so on its own system.
The Capital Cost of Catawba #2 R. W. Beck did not make any independent assessment of the likely capital cost of Catawba #2 in any of their studies. In each they used the current Duke Power Co. estimate. The Duke estimate for Catawba #2 has been increasing rapidly over the last few years, just as have the estimates made by all utilities building nuclear units. Since the completion date for Catawba #2 is now delayed until 1987, the cost estimate is very likely to continue to rise substantially over the next few years. Thus by relying on the present Duke capital cost estimate Beck is guaranteeing that
, this analysis will become rapidly out-dated. They must attempt I
to make a final cost estimate at this time based on cost trends
( that have affected nuclear units that have already been completed.
L
' Statistical techniques for making such incependent cost estimates l do exist and have often been applied to similar generation planning l
,ossessments by organizations such as ESRG. It is also imperative j for Beck to make an independent estimate of the likely completion dates for the Catawba units, which it has not done. Given redent n.
developments, these units are likely to be delayed by Duke again in the next couple of years.
6
.The Operations and Maintenance (O&M) Costs of Catawba #2 Just as for the capital costs of Catawba #2, R. W. Beck failed to make an independent estimate of the likely O&M costs of running Catawba #2. One way to develop such-an estimate that has been used by me is to extrapolate from actual historical data on the costs of operating other nuclear units. Again, techniques for doing this are available. In their most recent study update, R. W. Beck merely relied on Duke's new O&M cost estimates which they stated were "significantly higher" than those previously 4.
furnished. The changing nature of the Duke estimate further supports the need for independent estimates to be made. Further-more, Beck does not provide the basis for these new Duke estimates so their reasonableness can not be determined. Yet nuclear C&M costs can have a major impact on the attractiveness of the proposed
, PMPA purchase of Catawba #2 for PMPA ratepayers. Clearly, addi-tional discovery is required in this area, as in the other areas discussed here.
?
l l Projected Capacity Factor of Catabwa #2 The proposed sale of Catawba #2 to PMPA is a "take-or-pay" l contract, namely PMPA has to pay for Catawba whether it runs l
l well or not. This introduces an extremely high degree of risk I to PMPA into the proposed sale, for by the 1990's well over 50%
j ef PMPA's demand would be often served by this one station, in conjunction with the two McGuire units. If PMPA were to simply rely on purchased power from Duke, on the other hand, as they have in the past, their exposure to financial risk would be 1
7
substantially less. If the Catawba and McGuire stations generate more poorly than R. W. Beck has assumed, then PMPA ratepayers will have to pay the Catawba fixed costs in addition to replacement purchased power costs from Duke.
The operating efficiency of Catawba #2 is measured by it's expected " capacity factor", or the percentage of time it runs in a year. During 1982 Beck has assumed that the capacity factor of the Catawba Station will average over 60% and will rise towards 69% in the later years. Actual historical averages for large nuclear units, especially PWR's, have been lower thsn this, often considerably lower (as with Carolina Power & Light Company's three
- 2 nuclear units). Thus, again, it is critical to a study of the type that Beck has undertaken to do a thorough and independent review of the capacity factors for Catawba as projected by Duke.
In this area there is prima facie evidence that Duke (and there-fore Beck) is being overly optimistic as to these projections and that the proposed purchase would prove unfavorable to the
.PMPA on these grounds alone. Simple statistical techniques are available to expolate capacity factors for new plants from historical data on existing nuclear plants, but Beck did not use j these techniques, nor any other.
1 Fuel Prices Beck's report indicate reliance on Duke Power Company's Eos,t recent fuel price estimates. While these estimates may not be as problematic as those made for other key assumptions, it is incumbent on Beck to review these estimates carefully and alter 8
(-
than if-their judgment dictates.
l There is no discussion in the Beck documents of such a review.
Purchased Power Alternatives The entire focus of the R. W. Beck study is to compare the economies of the proposed purchase by PMPA of Catawba to PMPA buying power from Duke. However, Duke is r.ot the only alternative.
Lately, the cost of purchased power from Mid-Western utilities is falling and its availability is increasing, as demand in that region falls. This trend due to the economic downturn was well underway by November, 1982. Thus Beck should have also compared
( the economies of the proposed purchase to the economies of other long term purchased power arrangements that PMPA might be able to make with other utilities, notably the American Electric Power System. Such a long term purchased power contract would likely prove to be considerably less financially risky to PMPA ratepayers, even if its direct economics were only cor. pared to the proposed purchase, and therefore this alternative could well be preferable i
to PMPA. As a general principle it is financially wise for a small distribution utility not to get locked into a single power source for a large fraction of its requirements, so that it can make the most out of the changing market conditions for purchased power.
! The Economics of the Catawba Project for PMPA vs.
i- . Duke Power Ratepayers By selling 25% of Catawba #2 to PMPA Duke will have an easier time financing the remainder of its construction program. However, once Catawba #2 is complete, the generous buy-back arrangement
l-r l that Duke has offered PMPA will cause Duke ratepayers to have to pay most of the high fixed costs of PMPA's share of Catawba #2 in the fist 10 years of its lifetime, these being the years when the plant is not likely to be cost effective for any set of ratepayers. If Catawba #2 is cost effective at all to the citizens of South Carolina it is only likely to be so over the very long run (beyond the mid-1990's). Beck has not studied the issue of the impact of the proposed Catawba sale to PMPA on Duke ratepayers, a subject that must also be of concern to the South Carolina Public Service Commission.
l l
l 1
l l
i 10
II. ELEMENTS OF AN ADEQUATE GENERATION PLANNING STUDY An adequate study of the Catawba purchase project would not be greatly different from the Beck study in its methods of eco-nomic analysis, but the input assumptions (on load forecasts, plant costs, plant performance, etc.) would be independently developed, and the scope of the consideration of alternatives would be much greater, as it would encompass (a) purchase power from other sources than Duke, and (b) active PMPA promotion of customer conservation and load management as an alternative method of reliably meeting expected load growth or a substantial portion thereof.
Load Forecast While it is difficult to forecast load for dispersed communities, a good load forecast is the proper casis of all system planning. Ideally, a forecasting approach would establish I a disaggregated structure among and within major customer classes, preferably at the level of the major end uses of electricity, in
- an effort to analyze the effects of the diverse factors that chape consumption (employment changes, household growth, appliance saturation,
- onservation technology penetration, etc. ) ,
upon the various components of that consumption. Even if it is difficult to use a fully disaggregated end-use analysis for PMPA, it should be possible to incorporate elements of such an analysis, probably by performing a systenatic independent load forecast for the region within which PMPA is situated, then allocating 11
l appropriate portions of the forecasted demand to the PMPA communities, i -(_ trvation and Load Management In preparing a load forecast for planning purposes, an important step is to determine likely levels of penetration of demand-reducing practices and technologies during the forecast period, given current estimates of future market behavior and
- given estimates of the future effects of policies and programs now in place. An explicit identification of the likely levels of conservation and load management must be an important component
- j. of the load forecast analysis described immediately above.
I I Beyond that, it is,necessary to view the utility option of developing a deliberate strategy to induce higher-than-expected levels of conservation and load management as a resource planning alternative on a par with new capacity or purchased power. To the extent that this approach is a cheaper way to cope with increasing demand, it is the preferred alternative. Some l
utilities, ' including Duke Power, have embarked upon a policy of i
requiring and inducing customers to adapt demand-reducing practices I- ~
and technologies, a strategy that is usually embodied in a conservation and/or load management plan.
p A' study of the desirability of the project should therefore include an assessment of the potential for conservation, consisting i of an accounting of electric energy and electric peak conservation potential for each customer class, where conservation is defined in terms of measures that cost less to implement than would the 12
r supply of the energy otherwise required. This assessment of conservation potential would then form a basis for the development of long-range conservation achievement goals, including expansion of present programs and the development of new ones as found appropriate.
If'an end-use forecasting model is being employed as the
. load forecast model, which would be the preferred approach, a conservation scenario can be constructed by specifying changes that impact specific end-uses and groups of end-uses during the fore-cast period. The effects of such demandreducing measures as a reduction in the use of electric space heating in office buildings and an increase in the insulation levels of homes can be quantified
! explicitly in the conservation scenario. Using these scenario inputs, a " conservation case" forecast can interrupt base case
. computations to produce a second, slower growth year-by-year long-range forecast. When compared with the base case forecast, the Conservation Case forecast presents a quantitative estimate of the energy that can be saved and the winter and summer peak i
l reductions that can be attained if a deliberate policy of l
pursuing additional conservation is found acceptable and is
- successfully pursued.
The. conservation measures and levels incorporated in the scenario should. satisfy three criteria. They should be technically feasible; their incremental costs to electricity consumers will
'be less than the costs of additional electricity; and they can be
. effected through actions that it is feasible for PMPA to undertake.
13
k r
-Supply Analysis The supply analysis should be an integrated assessment of the economic and financial assessments of the proposed project and the viable alternatives thereto (including conservation promotion and power purchased'from other sources than Duke). It should include independently derived estimates, not only of the basic
- load to be met, but also of capacity costs, operations and main-tenance costs, likely capacity factor (for capacity options),
costs of fuels, and financial implications of alternative invest-
' ments, transmission constraints, etc.
Time Frame It is estimated that a full professional review would take three months to complete, from inception to final report and pre-
. filed. testimony. . Full discovery is crucial to incorporating the necessary accurate data regarding the Duke and PMPA systems.
ESRG is prepared and able to conduct such a study. The final section describes the qualifications of my organization to l
complete the required study.
l i
1-t 1
.14
III. ESRG EXPERIENCE AND QUALIFICATIONS Load Forecasting No organization outside the electric utility industry itself has had more extensive experience in the field of load forecasting than energy Systems Research Group. When ESRG was organized in 1976, it began a comprehensive review of forecasting experience and methods. Based on this review, ESRG commenced construction of its own forecasting model. From the beginning, the ESRG model was designed to improve the acebracy of the forecasting process by solving problems of mis-forecasting that were plaguing the l energy utility industry.
The ESKG forecasting model established a disaggregated end-use structure within which to examine the elements shaping consumption of energy and other resources including water. Beyond this basic commitment to the use of end-use disaggregation in order l
to examine the detailed components of demand, ESRG hold no brief for any one " Method" of forecasting.
Within its end-use framework, the ESRG model incorporates econometric, engineering, and trending methods in an eclectic endeavor to capture the diverse forces (economics, regulation, population growth, conservation awareness, technical change, etc.) that will determine demand in the future.
ESRG's forecasting expertise has been widely recognized.
On .the basis of this expertise, ESRG was awarded a bid from the New England Conference of Public Utility Commissioners to cerutinize and assess the forecasting model of the New England Power P? ol for the several state commissions of New England.
l 15
,_ In 1980, ESRG was retained by the General Accounting Office of the U. S. Congress to perform demand forecasts, conservation forecasts, and conservation option analyses for all New England states, a technical analysis that constituted the basis of GAO's two-volume Report EMD-81-58, issued in 1981.
ESRG has produced electric system forecasts for the utility commissions of Maine, Vermont, Connecticut, Wisconsin, Alabama, and Oklahoma, and has been technical consultant to the Public Utility. Commission of New Hampshire, assisting in forecasting model evaluation and development.
The ESRG forecasting model is designed to maximize reliance on data specific to the locality under study and minimize reliance l on generic national data. Numerous local data concerning weather, population, appliance saturation, employment, housing charac-teristics and other relevant factors are reviewed when ESRG performs a forecast critique and incorporated as input data when ESRG makes a forecast of its own. For example, a recent forecast l for the Sierra Pacific Power Co. service area, performed for i
-Nevada's Consumer Advocate, explicitly accounted for the growth L in employment and floorspace in the casino business, and its effect upon electricity consumption.
Economic and Technicel Analysis of Supply Alternatives ESRG has had wide experience assessing supply options facing utility systems on both a service area and a regional basis. These assessments have encompassed generation plant, transmission plant, purchases of capacity and energy, central station and decentralized cogeneration, and alternate sources of energy such as wind, biomass, 16
and solar energy connected to electricity grids. The assessments 1
have reviewed the technical, economic, environmental, regulatory, and financial aspects of supply planning, including the relations
, between supply planning, load forecasting, rate design, and revenue requirements.
In 1978/79 ESRG began developing the ESGEM Electricity Supply /
Generation Expansion Model, designed to develop optimal reserve margin levels for electric utility systems and to develop production cost estimates for alternative expansion plans. This model was enhanced under contract with the U.S. Department of Energy for reliability analysis purposes. Starting in 1979, ESRG conducted a series of investigations of the power supply alternatives of-electric utility systems. The ESGEM model was used in these
- Enalyses in 1979 and 1980.
By the end of 1980, ESRG acquired the SYSGEN Electric System Production Costing and Reliability Model developed at the Massachusetts Institute of Technology, and substituted it for the i
ESGEM model in power and supply evaluations. It has proved to be i a flexible, accurate and economical model to use to calculate l power production costs and reliability indices for annual, seasonal l
l and time-of-day periods for: purposes of power supply planning j enalyses and rate design / costing studies.
l A third methodological development was ESRG's introduction
(
of the ELFIN-electric utility financial model, developed by the i
Environmental Defense Fund, into power supply evaluation studies.
This model is a corporate financial simulation model which calcu-lates the revenue requirements and financial implications of power 17
! supply alternatives. In the proposed power supply study it will be used in tandem with the SYSGEN model.
Finally, there have been four major statistical studies undertaken by ESRG research staff over the past years. These have provided a statistical basis for predicting:
- 1. Nuclear power plant capital costs
- 2. Coal-fired power plant capital costs
- 3. Nuclear power ',lant O&M costs
- 4. Nuclear power plant capacity factors.
These analyses have strengthened our power supply data in specific areas.
From a methodological viewpoint, ESRG has retained flexibility by maintaining familiarity with other videly used power supply models such as the WASP and PROMOD models. We have critiqued the application of these models in specific cases.
From a practical viewpoint, ESRG staff have had experience with a wide range or utility systems including such major interconnected systems as AEP, CAPCO and the Southern Company, as well as a regionally dispersed range of other utility systems.
In addition to numerous analyses of conventional supply options, ESRG has performed numerous studies of alternative (i.e. non-conventional) supply, beginning with a survey of the potential for industrial cogeneration in New York State sponsored by that state's Energy Office in 1978. ESRG has studied the
_ potential for wind energy, wood energy, small-scale hydropower, cnd energy from solid waste, as in a report to the Connecticut
- Power Facility Evaluation Council in 1980. Currently, ESRG is etudying the feasibility of district heating from existing power 18 4
~
plants for the Boston Redevelopment Authority. ESRG has also developed a multi-resource supply planning system, the LEAP system, for the Royal Swedish Academy of Sciences, and applied this system to several African nations.
Numerous of ESRG's supply studies have involved integrated reviews of the technical and financial implication of various types and levels of capacity expansion in the light of anticipated loads, alternative supply sources, and the potential for conserva-tion. Supply studies in which the ELFIN model has been employed have included scenarious in which the implications for utility finances of utility investment in customer-side conservation have been traced in quantitative detail. An example is the comprehensive economic analysis of the proposed Arthur Kill Plant and the alternatives thereto recently completed for the New York City Energy Office.
Conservation Analysis ESRG has assessed conservation programs of utilities and state agencies, conducted quantitative studies of the potential for additional cost-effective conservations, and identified conservation program priorities in a series of governmental agencies cnd citizen groups. A study that is representative of many of ESRG's capabilities in this area is " Utility Promotion of Customer Energy Conservation: An Assessment of the Existing Programs and l
Potential Role of the Public Service Electric and Gas Company".
This study was prepared for the Public Advocate of New Jersey in August 1981, and has been widely circulated by the Department.
ESRG's PSE&G study contained: an original analysis of the regulatory 19
issues and practical problems involved in utility conservation program design, based on a nationwide review of utility conserva-tion program development experience; a detailed review of PSE&G's existing customer conservation programs; a quantitative analysis of the most promising options for further conservation program development, based on customer usage data specific to the PSE&G service area; and recommendations for specific regulatory actions to realize the potential for cost-effective conservation.
.It formed the basis for detailed conservation program recommendations that influenced, and were specifically referenced by, the New Jersey Board of Public Utilities in its November 1982 order establishing a new, comprehensive PSE&G conservation program.
ESRG has conducted several assessments of utility conservation programs based on its detailed research into conservation techniques, costs, and programs, and the particular circumstances of the given utility systems. Such assessments have also been conducted on a regional basis, as for example in the recent report " Priority Incentive Programs to Save Electricity in Massachusetts," funded l
by the. Massachusetts Executive Office of Energy Resecurces.
ESRG'has been extensively involved in the development of customer programs to promote cost-effective conservation in utility service areas. As consultants to the Advocate for Customers L of Public Utilities in Nevada, ESRG has provided guidelines for f fan optimal conservation effort.on the part of the Sierra Pacific Power Co. For the Papago Tribal Utility Authority, constructive proposals for enhancements to conservation program activity by the Arizona Public Service Company have been advanced in testimony l
20 l - _ _
i recently submitted to the Federal Energy Regulatory Commission.
Currently, as consultants to the Public Advocate in New Jersey, ,.
2SRG is reviewing the conservation programs of all investor-owned utilities to recommend expansion, contraction, continuation, or restructuring of programs to promote conservation where it is cost-effective for the utility to do so.
ESRG pioneered in the development of conservation case forecasts and analysis of the economic effects of conservation.
In each of the " conservation case" forecasts identified in the
" Forecast Experience" table in subsection A, below, ESRG iden-tified a set of technically feasible and economically attractive
- . conservation measures and levels which went beyond " Base Case"
. levels.. A conservation . scenario was constructed to e::plicitly j modify several of the input assumptions contained in the " business-as-usual" scenario underlying the Base Case energy forecasts.
A " Conservation Case" long-range forecast of electricity (and, in
^
most cases, heating fuel consumption) was developed to quantify the energy and capacity savings that can be realized through implementation of the additional conservation measures in the conservation scenario for'the region. In several of the studies, the chief direct cost tradeoffs involved in implementing the conservation scenario.also were quantified. The stream of costs cnd benefits were computed on an annual and cumulative basis, using our CONCOST model, and discounted to their present worth to the consumers.
21
III. A 1hB2 1 ISRG IDNG-RANT IfAD TORECAST EXPCMENCE*
Independent Crataque Inde; ender.t Base of Cbr.serva-Utility or Region State Chse Utility tion Case Forecast Ferocast Foreca:-
Not York Power Pools New York xx x x Central Hudson Gar. & Electric Co. New York xx x x Consolidated Edison Capany Ne# York xx x x Ieng Island Lighting CmWy New York xxx xxxx xx Not York State Gas & Electric Corp. New York xx x x Niagara Mohawn Pcuer Corp. Not York xx x x Orange & Rockland Utilities New York xxx x x Ptwar Authority of the State of N.Y. New York xx xx x lechester Gas & Electric Co. Nos York xx x x Statewide Forecast Ne# York x x Central Area Power Coordination Group (CAPCD):
Philadelphia Electr.c Ccupany Pennsylvania xx xx x Pennsylvania Ptwar C a pany Pennsylvania xx xx Duquesne Light Company Persnsylvania xx x x Cleveland Electric Illuminating Chio x 1blecb Edison Cmgany chio x Ohio Edison Ca pany Chio x American Electric Power Co. (AEP):
AEP Syste x Aggalachian Pcwer Capany Virginia and W. Va. x x x Celtanbus & Southern Onio Ohio x Chio Power Ccupony Chio x New England:
Not England Ptwer Pool x Boston Edison Canpany Massadiusetts x Maine Public Service Co. Maine x Central Maine Power Cefany Maine x x Mass. Municipal Wholesale Elec*ric Massachusetts x x Not Bedford Gas & Electric Massachusetts x Northeast Utilities System Connecticut xxx x xx United Illminating Ccagany Connecticut xx xx x statewide forecasts Eacn State x(6) x(6)
Vermont x x Not Hemishire x
! magionwide forecast Six-state region x others Alabama Pcwer Ca pany Alabama x x l
Arizona Public Service Co. Arizona xx Atlantic City Electric Co. Nor Jersey x x x Sierra Pacific Ptwar Co. Nevada x x x Cincinnati Gas & Electric Cmgany Ohio x Dayton Ptwar & Light Czapany Chio x Detroit Edison Ctapany Micesigan x x C ms m ar Ptwer Ctapany Michigan x x Indiana Public Service Indiana x x Public Service Ccmqmny of Oklahcna Oklahczna x x Dallan Ptwar & Light Ccmgany 'Notas xx Secas Electric service ccupany Secas x Utah Ptwar and Light Utah x x hastern Wisconsin utilities Wisconsin xx x (Dairyland Power Coop., Zake Superior l District Ptwar, Northern State Ptwer
- 02. )
Eastern Wisconsin Utilities Wisconsin x (Madison Gas & Electric, Wisconsin Ptwer & Light. Wisconsin Public Ser, rice Corp.)
- 1acn x regresents one stucty. 2/ t3 i
i
i l
i 1hB2 2 ESRG ELECTRICITY SUPPLY PRC0ECTS -
SYS7D4 RELIABILITY, GENERATION PIANNING, PRODUCTION CDSIS l
i Docket of Utility or Brief Regalatory Year Power Pool Project Title Agency 1979 Power Authority of the Greene County Power Plant N.Y. P.S.C.
State of N.Y. Generation Cast Analysis 80006 N.B.C.
50-549 1979 Detroit Edison Cmpany D.E.C.O. Systen Reliability / Michigan P.U.C.
Reserve Margin Requirements Case U-6006 1979 Long Island Lighting L.I.L.C.O. Janesport Plant, New York P.S.C.
. Conpany Generation Costs and 80003 Alternatives 1979 Philadelphia Electric P.E.C.O. Systen Reliab_lity, Pennsylvania Capany Generation Plan, Optimal P. S . C.
Costs (EEGEM Model) R-79060865 1980 Duquesne Light Co., CAPCD Systen Reliability, Pennsylvania Pennsylvania Power Co., Cast Analysis of Delayed P. S . C.
CAPCD Construction (ESGEM Model) R-79070315 through 79070331 1980 Union Electric Canpany U.E. Systen Reliability / Missouri Optimal Construction, P.S.C. Case Reserve Margin Analysis ED-80-57 (ISGEM Model) 1980 New York Pcwer Pool 7tstimony on Generation F.E.R.C.
Planning and Econanic Project 2729 Analysis 1980 Central Area Power Analysis of CAPCD/2I Ohio P.U.C.
Coordinating Group, Generation Cm struction 79-537-EL-AIR Cleveland Electric Plan 111taninating 1980 Constanar's Power 74chnical/F-ic Assess- Michigan P.U.C.
W ny ment of CPC's Generation Case U-6360 Construction Program (ESGEM Model) 23
_- m. m- - . , , . . ,_ .
_ _ - -. ...,._,-%- - - _,, , - , , ,.,,w- ,,_,,_ , ,, ,- - . - - .-y.m-- -. - - . ,-
l SUBIE 2 l (Ccntinued)
ESRG ELECIRICITY SUPPLY PRQIECIS -
TEM RELIABILITY, GENERATION PIANNING, PRODUCI' ION CDSIS Docket of Utility or Brief Regulatory Year Power Pool Project Title Agency 1980 -
D.O.E./B.N.L. Develognent -
and Refinement of ESGEM Model 1980 American Electric Power A.E.P. System Optimal F.E.R.C.
Systen Capacity Plan, PURPA Rate Project 2812 Structure Alternatives (Brumley Gap)
(SYSGEN, ESGEM Models) 1980 Iong Island Lighting Nuclear vs. Conservation N.Y. State Cm.pany Investment Cmparative P.S.C. 27774 Econanics Analysis; LILCD l Shoreham Plant 1980 CAPCO, Ohio Edison Analysis of CAPCD/OE Ohio P.U.C.
Cm pany Generation Construction 80-141-EL-AIR Plan 1981 Appalachian Power Analysis of APCO Systen West Virginia l Cmpany Reliability Generation P.S.C. 79-l Planning (SYSGEN Model) 140-E-42T 1
1981 Power Authority of Cmparative Econanics F.E.R.C.
State of New York Analysis; Conservation Project 2729 Investment vs. PASNY's Prattsville P mped Storage 1981 Idaho / Utah Idaho P.U.C. Innovative Rate -
! . Study - Interruptible Rates l (SYSGEN) 1981 Central Maine Power Econanics/ Financial Capari- Maine P.U.C.
Campany son of CMPC's Planning U-3238, Alternatives (ELFIN) U-3239 1981 Philadelphia Electric PECD lamerick Station / Pennsylvania Cmpany Alternate Capacity P.U.C.
Reliability, Econanic and I-80100341 l Financial Analysis (SYSGEN, l
ELFIN Models) 1981 Alabama Power Canpany/ Poer Supply and Financial Alabama P.S.C.
l Southern Canpany Systen Analysis of APCD System 18337
, Reliability / Generation j Expansion Planning (SYSGEN l and ELFIN Models)
'lhBLE 2 (Continued)
ESRG ELECTRICITY SUPPLY PRCL7EC75 -
SYS7D4 RELIABILITY, GENERATION PLANNING, PRODUCTION CDS75 Docket of Utility or Brief Regulatory Year Power Pool Project Title Agency 1981 Consolidated Edison Evaluation of the Travis/ U.S. Army Corps.
Co. and Power Authority Arthur Kill Coal Plant of Engineers of State of New York Proposal and Alternatives Environmental (SYSGEN and ELFIN models) Hearing 1981 Public Service Co. of Evaluation of Black Fox Oklahcma Oklahcma Nuclear Station - Corporation Cm pletion vs. Cancel- Ccmmission lation 27068 1981 Iong-run Marginal Cost Appalachian Power Co./ Virginia State of Power Supply American Electric Power Corporation (SYSGEN model) Systen Ccmnission PUE 80076 1981 Maine Public Service Power Supply / Financial Maine PUC Co. Evaluation of Seabrook 81-114 Nuclear Station (SYSGEN and ELFIN models) 1982 Public Service Co. - Power Supply Alternative New Hampshire of New Hampshire (ELFIN model) Public Utility Ccmnission DE 81-312 l
l l 1982 .Montaup Electric Evaluation of Power F.E.R.C.
Ompany Supply Planning ER 82-325-000 1982 Public Service Co. Power Supply Plan and Indiana P.S.C.
of Indiana Evaluation of Co. Plan 36818 1982 Public Service Co. Evaluation of Power Oklahczna Corpora-l of Oklahcma -Supply Planning tion Ccmnission 2/83 25
, . , , _ , . . - - _ _ ~ ,-,_w.y,. .- .-- - , - . - . _ . . . . _ , , ,
. . IV. RESUME RICHARD A. ROSEN Research Scientist Executive Vice-President Energy Systems Research Group Education Ph.D.: Physics, Columbia University, 1974 M.A.: Physics, Columbia University, 1969 B.S.: Physics and Philosophy, M.I.T., 1966 Experience 1977 - present: Energy Systems Research Group, Inc.
Responsibility for a broad range of research on industrial energy conserva-tion; electric generation planning issues; and modelling studies of long-range electric demand, utility system reliability, electric demand curtailment, and district heating systems.
1978 - 1980: Consultant to Brookhaven National Laboratory.
.1979: Consultant to the National Academy of Sciences, Puerto Rico Energy Study Committee.
1976 - 1978: Assistant Physicist, Economic Analysis Division, National Center for the Analysis of Energy Systems, Brookhaven National Laboratory.
1974 - 1976: National Research Council - National Academy of Sciences Resident Research Fellow, Goddard Institute for Space Studies, New York.
1973: Instructor - Putney - Antioch Graduate School.
Testimony Case or Agency Docket No. Date Topic Maine Public 81-114 Apr. 1982 Investigation of Power Utilities Supply Purchases and Commission Planning, Maine Public
. Service Co.
Maine Public 81-276 Apr. 1982 As to the Avoided Costs Utilities for Cogeneration and Commission Small Power Production Facilities on the Maine Public Service Company System
'26
Case or Agency Docket No. Date Topic Maine Public 81-114 Mar. 1982 Investigation of Power Utilities Supply Purchases and Commission Planning, Maine Public Service Co.
Alabama Public 18337 Jan. 1982 Long-Range Capacity Service Expansion Analysis Commission State of New York SEMP II Nov. 1982 Conservation and Energy Planning Hearings Generation Planning Board Pennsylvania Public 80100341 Sept. 1981 Operating and Capital Utility Commission (Sur- Costs: Limerick rebuttal) Nuclear Station Maine Public MPUC 80-180 Apr. 1981 Electric Energy Costs:
Utilities (Sur- Seabrook Nuclear Power Commission rebuttal) Plants l Pennsylvania Public I-80100341 Feb. 1981 Operating and Capital Utility Commission Costs: Limerick Nuclear Generating Station Ohio Public 80-141- Dec. 1980 CAPCO Construction Utilities EL-AIR Program (generation Commission planning)
Michigan Public U-6360 Sept. 1980 Generation Expansion Service Commission Planning: Consumers Power Co.
i Pannsylvania Public I-79070315 Aug. 1980 CAPCO Construction Utility _ Commission (Sur- Schedule rebuttal) l Connecticut Power . F-80 June 1980 Renewable-Resource Facility Evalua- Electric Generation tion Council in CT PGnnsylvania Public I-79070317 Mar. 1980 CAPCO: Generation Utility Commission Planning and Reliability
. Michigan'Public U-5979 June 1979 Forecast Critique and Service Commission . Adjustments:
Consumers Power Co.
s -
, .. . . .-_ . - . - - - . ~ . . . , - .
Case or Agency Docket No. Date Topic Massachusetts Depart- 19494 Aug. 1978 Long-range Electric ment of Public Demand Forecast:
Utilities Boston Edison Co.
Pennsylvania Public 438 Mar. 1978 Long-range Forecast Utility of Electric Energy Commission and Demand.
(Philadelphia Electric Co.)
ESRG Research April 1982: A Power Supply and Financial Analysis of the Seabrook Nuclear Station as a Generation, Option for the Maine Public Service Company. ESRG Study No. 81-61.
Principal investigator.
January 1982: Guidelines for Designing Rates for Sales to Qualifying Facilities Under Section 210 of the Public Utility Regulatory Policies Act. ESRG Study No. 81-32. Co-author.
June 1981: An Analysis of the Need for and Alternatives to the Proposed Coal Plant at Arthur Kill.
A Report to: Robert M. Herzog, Director, New York City Energy Office and Allen G.
Schwartz, Corporation Counsel for the City of New York. ESRG Study No. 81-21.
Co-author.
October 1980: The ESRG Electrical Systems Generation Model: Incorporating Social Costs in Generation Planning, ESRG Study No.
90-12. A report to the U.S. Department
.of Energy. Co-author.
September 1980: Reducing New England's Oil Dependence Through Conservation and Alternative Energy, ESRG Study No. 79-29. A Report to the U.S. General Accounting Office. Co-author.
O SO
July 1980: Preliminary Economic and Need Anal,y_ sis of the Proposed Brumley Gap Pumped Storage Facility Tor the AEP System, ESRG Study No. 80-08/P.
Principal investigator.
July 1980: The Potential Impact of Conservation and Alternative Supply Sources on Connecticut's Electric Energy Balance, ESRG Study No.
10-09, A Report to the Connecticut Power Facility Evaluation Council. Co-author.
November 197 9: South Carolina Electric Demand Curtailment Planning, ESRG Study No. 79-31, A Report to the South Carolina Office of Energy Resources.
Principal investigator.
May 1979: Demand Curtailment Planning: Methodology, ESRG Study No. 78-18, Chapter submitted to Brookhaven National Laboratory and the Department of Energy for the Electric Demand Curtailment Planning Study. Principal investigator.
May 1979: Assessment of the New England Power Pool -
Battelle Long Range Electric Demand Forecasting Model, ESRG Study No. 79-06, A Report to the I
New England Conference of Public Utility Commissioners. Co principal investigator.
October 1978: The Employment Creation Potential of Energy Conservation and Solar Technologies: The Implications of the Long Island Jobs Study Tor New England, 1978-1993, ESRG Study No.
78-16. Co-author.
L November 1977:-- Profile of Targets for the Energy Advisory i
Service to Industry, ESRG Study No. 77-09,
- A Report to the.New York State Energy Office.
l Co-Author.
October 1977.:
The Effect on' Air and Water Emissions of Energy Conservation in Industry, ESRG Study No. 77-04. Co-author.
. s.
e-
,. A a 6
! g
-. 9
- g. ,*
~
e -.y
.- s
~
July 1977: The Effect on Air and Water Emissions of Energy Conservation in Industry, ESRG Study No. 77-04. Co-author.
June 1977: Toward an Energy Plan for New York, ESRG Study No. 77-03, A Report to the Legislative Commission on Energy Systems. Co-author.
April 1977: Assessing Demand, Alternative Operating Strategies, and Utility Economics in the Service Territory of Orange and Rockland Utilities, ESRG Report No. 77-01. Co-author.
Other Publications March 1978: The Use of the Pulp and Paper Industry Process Model for R&D Decision Making, Brookhaven National Laboratory Report No. BNL 24134.
Co-author.
1976: "A Non-Linear Model for the Linewidth,
. Intensity, and Coherence of Astrophysical Masers," Astrophysical Journal vol. 190.
Pipers July 24-28, " Energy Use Modelling of the Iron and Steel 1978: Industry," Summer Computer Simulation Conference.
November 12, " Energy Conservation in Industry," Northeastern 1977: Political Science Association meeting, Mt.
. Pocono, Pennsylvania.
Awards and Honors 1968 - 1974: Faculty Fellowship, Physics Department Columbia University.
1966 - 1970: New York State Regents Fellowship.
1967 - 1968: Adam Leroy Jones Fellow in Philosophy, Columbia University.
g ,~ .. - - ,v.
W l
l CTATE OF MASSACHUSETTS)
) ss COUNTY OF SUFFOLK )
AFFIDAVIT OF RICHARD A. ROSEN Richard A. Rosen, being first duly sworn, on oath, deposes and says that the attached statement was prepared by him or under his supervision and the information contained in such is true and correct to the best of his knowledt e, information and belief.
lJ-~
Richard A. Rosen
~
Subscribed and sworn to me before this /D i day of March,1983.
( C.. M. 1 Notary Public
(
My Commission Expires Nov. 12, 1987.
e
EXHIBIT (RAR-3)
OPERATIONS AND MAINTENANCE L
Exhibit (RAR-3)
C. OPERATIONS AND ' MAINTEN ANCE C.1 Introduction The operations and maintenance (O&M) costs of nuclear generating stations together with nuclear fuel expenses comprise the total electricity production costs at these facilties. These costs are directly passed on to electricity consumers as direct expense items in required revenues. Other costs associated with nuclear power include the capital costs of the facilities sincluding capital additions) which enter required revenues through their inclusion in the rate base, upon which interest or a rate of return can be earned. In addition, the costs of decommissioning and spent fuel disposal can impact required revenues insofar as funds for their implementation are collected during the operating years of the nuclear station.
Nuclear power plant operations and maintenance costs fall into 13 broad subcategories, as reported by utilities to the Federal Energy Regulatory Commission (FERC) in annual Form 1 submissions and to the U.S. Department of Energy (Ref. Col). These are listed'below in Table C.1.
Data on these O&M costs for nuclear generating stations for the years 1970 through 1980 have been collected from utility FERC Form l-reports and the U.S. Department of Energy (Ref. C.1). A total of 49 nuclear stations, virtually all commercial units that have operated in the U.S., are included in this data base.
C -1 a
_ _ _ _ _ _ _ _ _ _ . _ _ _ . _ _ _ . _ _ . . _ _ _ _ _ _ _ ._.____._d
TABLE C.1 NUCLEAR O&M SUBCATEGORIES Operations Maintenance Supervision and Engineering Supervision and Engineering Coolants and Water Maintenance of Structures Steam Expenses . Maintenance of Reactor Plant Steam from Other Sources Maintenance of Electric Plant Steam Transferred Maintenance of Miscellaneous Electric Expenses Nuclear Plant Miscellaneous Nuclear Power Expenses Rents Some of the salient features of nuclear power plant O&M cost experience average directly from examination of industry-wide ave.: ages. In the industry as a whole nuclear O&M costs have increased from about one-half of nuclear fuel costs in 1970 and 1971 to about twenty parcent greater than nuclear fuel costs in 1979 and 1980. Thus, these costs have begun to dominate the production costs for nuclear facilities. Within the O&M costs themselves the split has remained rather stable at about 55 percent for operations and 45 percent for maintenance throughout the 1970-1980 period. Of the 13 subcategories, the two largest are miscellaneous nuclear power expenses (23.5 percent in 1980)
I and maintenance of miscellaneous reactor plant (20.5 percent in 1980). These two subcategories plus maintenance of miscellaneous nuclear plant have increased their share of total O&M costs from about 39 percent in 1970/71 to over 50 percent in 1979/80.
Total costs for operations and maintenance of nuclear stations have increased dramatically from about $20 million in C-2
1970 to about $1,400 in 1980, a seventy-fold increase. Table C.2 shows the industry-wide nuclear station, annual O&M costs from 1970 through 1980 on a per-kilowatt installed capacity basis in both nominal and constant (i.e. 1983) dollars. The second column, nuclear OEM costs in 1983 dollars per kilowatt, shows the growth trend in real per unit costs during the 1970-1980 period, thus correcting for both inflation and the increasing size of the industry. The increase was from about $12.5 per kilowatt in 1970 to about $35.9 per kilowatt in 1980. The average annual growth rate in real (i.e. above inflation) O&M costs per kilowatt for nuclear stations in the U.S. was 9.3 percent per year from 1970 through 1978 (the last full year before the TMI reactor accident)
- and 11.0 percent per year from 1970 through 1980.
TABLE C.2 OPERATIONS AND MAINTENANCE COSTS FOR NUCLEAR STATIONS IN THE U.S.
1970-1980 Average Industry Average Industry 1983 Dollars Dollar per Kilowatt Per Kilowatt 1970 5.25 12.53 1971 5.02 11.40 1972 6.91 15.08 1973 6.38 13.16 1974 8.73 16.58 1975 9.94 17.27 1976 11.98 19.78 1977 13.65 21.29 1978 16.78 24.39 1979 20.93 28.04 1980 29.21 35.93 Average Annual Growth Rate (Percent) 1970-1978 16.6 9.3 1970-1980 18.6 11.0 C-3
4 4 ~: _ y - , A--eL -
mJ.--s--..---1am4 C.2 Statistical Analysis Nuclear generating station operations and maintenance costs have varied widely by facility and year of operation. In the present analysis linear regression techniques'have been used to explain this variation in terms of independent variables expressing the characteristics of the nuclear stations. Various models or equations were selected for analysis. These equations expressed the dependent variable (o&M costs in 1980 dollars per kilowatt) as a linear combination of several independent or
. explanatory variables. Numerous independent variables were explored in various combinations with each other. These included plant size (in Megawatts) and age, chosen to test whether economies of scale and cost increases associated with aging have i ~
been occurring. Other variables which.were-explored for ;
statistical significance in explaining the variation in OEM costs were plant vintage (date of initial commercial operation),
~
l geographic location, demonstration unit status,' salt-water i
cooling, multiple unit siting, 1980 operation, reactor
! manufacturer, cooling towers, turbine manufacturer, utility size, i
and utility experience with nuclear plant operation. The last five variables were not found to have statistical significance.
Definitions of the variables for which statistical significance was found are provided in Table C.3.
l C-4
._. . , - - , . . , - . _ _ _ _ _ _ , . - _ _ . . _ - - . - ~. . .-_._ . --.. - . ____. .
C.2 Statistical Analysis Nuclear generating station operations and maintenance costs have varied widely by facility and year of operation. In the present analysis linear regression techniques have been used to explain this variation in terms of independent variables expressing the characteristics of the nuclear stations. Various models or equations were selected for analysis. These equations m
expressed the dependent variable (o&M costs in 1980 dollars per kilowatt) as a linear combination of several independent or explanatory variables. Numerous independent variables were explored in various combinations with each other. These included plant size (in Megawatts) and age, chosen to test whether economies of scale and cost increases associated with aging have been occurring. Other variables which were explored for statistical significance in explaining the variation in O&M costs were plant vintage (date of initial commercial operation),
geographic location, demonstration unit status,' salt-wcter cooling, multiple unit siting, 1980 operation, reactor manufacturer, cooling towers, turbine manufacturer, utility size, and utility experience with nuclear plant operation. The last i five variables were not found to have statistical significance.
l Definitions of the variables for which statistical significance was found are provided in Table C.3.
l C-4 f
l l
TABLE C.3 DLPINITIONS OF INDEPENDENT VARIABLE 3 USED IN NUCLEAR OEM COST REGRESSION MODEL l
Variable Definition AGESTEP Cumulative years of commercial operation to the end of the year for which the O&M cost observation is made. For multiple unit stations, AGESTEP equals the age of the first unit until the second unit comes on-line. With multiple units in operation the variable will equal the average age of all units.
NEMASK 1 'if station located in the Northeast 0 if otherwise
.DERSTEP The station's net design electrical rating (DER) in megawatts. For multiple unit stations, DERSTEP equals the first unit's capacity until the second unit commences operation. With multiple units operating j the variable will equal the average (DER).
SALT 1 if station is salt-water cooled.
O if otherwise DEMO 1 if station was built as a demonstration project 0 if otherwise MULTSTEP For multiple unit stations MULTSTEP is 0 until the year the second unit begins commer-cial operation. With multiple units l operating MULTSTEP will equal 1. O at all times for single unit stations.
BIRTHSTP Date of commercial operation. BIRTHSTP i includes the actual calendar on-line date through the use of fractional years. For multiple unit stations, BIRTHSTP equals the l birth date of the first unit prior to i commercial operation of the second unit, after which BIRTHSTP equals the average l birth date if both units are operating.
1 L TMI 1 if year of operation is 1980 0 if otherwise C-5
The model chosen, including the values found for the
. coefficients and the measures of statistical significance and
' goodness of fit ( t-statistics , R-Squared, etc. ) is given in Table C.4. All of the variables in the model show strong statistical significance.
Several of the variables are time related. The result for general age term (AGESTEP) indicates that real (1980 dollars) O&M costs have been increasing at over $3/KW per year for every additional year of operation of nuclear stations. The aging effect of operating salt-water cooled plants, (SALT x AGESTEP) is found to be an additional S.92/KW per year, probably the result of the corrosive impacts of salt-water in the cooling systems and steam generators of these units. Economies of scale were found to increase with age. The value for the coefficient for the size times age term (DERSTEP x AGESTEP) implies that for a 1200 MW plant O&M costs would be S.40/KW lower than for a 800 MW in the first year of operation. Another way of examining the effect of this term is to compare it with the general age term. For a 1000 MW plant the two terms together imply OEM cost increases of about
$1.20/KW per year, while for a 800 MW the increases would be about Sl.60/KW per year.
I l
c-6
TABLE C.4+
NUCLEAR STATION O&M COST REGRESSION MODEL Equation Independent value of Confidence Coefficient Variable Coe f ficien t
- T-S tatis tic
- Level A -139.21 -8.45 > 99.8%
+ B x AGESTEP 3.19 7.00 > 99.8%
+ C x NEMASK 5.15 5.75 > 99.8%
+ G x DEMOX AGESTEP 3.24 4.78 > 99.8%
+ H x DEMO -31.71 -3.87 > 99.8%
+ J x MULTSTEP -2.98 -3.21 > 99.8%
+ K x BIRTHSTP 2.02 9.28 > 99.8%
+ M x DERSTEP X AGESTEP .002 -3.28 > 99.8%
+ Q x SALT x AGESTEP .915 5.65 > 99.8%
! + N x TMI 8.59 6.25 > 99.8%
Number of Variables = 10 Standard Error of Regression - 7.16 R-Squared = .675 F(9/317) = 73.21 Corrected R2 = .666 COND(X) = 131.98
- Rounded
! + Dependent variable is nuclear station O&M costs in 1980 dollars l per kilowatt.
l l
C-7
t Two other time related variables proved significant. The variable TMI was found to have a coeffic'ent i of 8.59, implying that on average an additional $8.59/KW was experienced by nuclear stations in the year 1980 (the first full year of operation after the TMI accident). Whether this represents a one-time effect, a permanent shift, or an acceleration of O&M cost increase trends is difficult to determine at this time. Analysis of 1981 and 1982 data will be helpful in this regard. The variable BI.',THSTP is also time related.
It measures the calendar year (and fractions thereof) of initial commercial operation. The finding of an additional $2/KW for each year later of commercial operation indicates that there are higher costs for maintaining a kilowatt of capacity which is built later. This may be a result of greater complexity and more safety features embodied in later vintage plants.
i C-8
REFERENCES C.1 Steam- (Thermal-) Electric Plant Construction Cost and Annual Production Expenses (DOE /EIA-0323), U.S. Department of Energy.
C-9
EXHIBIT (RAR-4)
Excerpt from R.W. Beck Study i
2 .:. e.
,, , ne,e,.,,., ,,,m . 8. ,
~,.,,.,,e, ,e ,m a. n.
Exhibit - (RAR-4) 8+ *.4 t* M Q *E I
- S. . "P. @. em 8. *t. W. P=. nn. e. S. er. 3. en . e
==#, ce Pe ** m ee es em erg grg C em
- 3 4 8% m e. e a. g , F * *P.* te O e ee. esJ ** 3 omenHn*mg @@@m f*
>= *U e
e V eI c. .
r ss
- ne m g%e e a sp
[ @
- E. a. er* m. as.
- a m
- t. ii e eme =-
h !
e3
- t n ne en 8 o s' s e e o.e e a e v emoe#=
~-
m.e e ea se,e.t w. er
.e e~ e inm ee<en ow e .
- e. s . e. s. h. e =. e. e.
e em rg to .A y . ett e A tm W W e O em s em ma ' .e.
.W
==
48 8* h - - 4F @ e9 w Ne, f5 f* 8"'WmW@
ve en. . e c as
- w c: e p. O
@ ". @m-N
- Jn er g *f . 3, e. e. ne. F em se em f. en M e ** h se.r* te 884
. @ @ m. e.. . m. 8 .
e g == , # ts ee # t; em a en me p. ee e te ne go
- == a= e ce eg # e e mM y ge w pg yy
'd '
V' '
- er e-
-g an e ==gg,y e= *= y e =ame .m. am c. a.n e oet.e 3 sisi - gm
- ert C . E. **l. a. ce. ce. e. e- . m. F. r=. e. m. e. m.
- h e. . a en a- = e - se e o a w - e ei p
- *
- E 9m E C' ** eP m ee e . *. *e nO O O eO E se am F. es ce set m e e art te en es* ,
O < E,eE g ",**
- C. .
- 7. 2 s *Oc
- in vg.
O T ie. - c.
oc %
x % = ee em 4 6 to w ye e
ggm5 m n c* ,cE ry , a e m .- an e ae,oee. em m ea e~ ==
o e- es ee erwine ee.
- m; ,=_..
m e m e o. .%.m.e.w.m.e.e.m.
o gen. .en.,.e GQ
- m - . .e n N.
wa-
.wae .
Uc r
. m I b n%,Ee et's 6.
O Cj
>" Ye el in C !. ***
u-
- ""*l'
= O. to th ep
'O .::*
4 *96*
.&. c,
- 0 4 .
e'
%,l
- a. e.no o - e e o e o e o *m rw
.eooaooo- m e= o m esn =e b
WR m. W. D. e. e. **. m. 7. =S. P. t .D M. N. O. W.
C' =
- -*a*e=e* ce em ma mc. c eh,edce e em.cemee, i, o.
n h E es 6
e t
e ~c.515 m p=~ .mc. set
~. m - as "
g m es eea= es~ .n. m te an e
- m
=8 88%. 40. O. W. O. *. W. Wm M. O. W. %. O. 88. + g Qt .em vs 9 en m e o.m r ,ee eo en sna mse=Ors =a=m m esse r =een. a em p MMMmme@ @ he e
e.
N S
- m. m .p .. p W h. e. e. S m--
m.
peWh e,,ee ee
~
- =
- c. em f.'*. F sC'e*f= ** 5* F (*b
= = == == F==e.F F f,* e F.- e
@*@. On== f 1 se
EXHIBIT (RAR-5)
ESRG NUCLEAR COST ANALYSIS REVISED TOR 52 PLANT DATA BASE
Exhibit (RAR-5)
ESRG NUCLEAR COST ANALYSI,S REVISED FOR 52 PLANT DATA BASE November, 1982
ESRG NUCLEAR COST ANALYSIS ESRG Nuclear Plant Cost Data Base The ESRG nuclear data base was constructed in order to assess the impact of different parameters on the cost of commercial nuclear power plants in the United States.
The data base includes all of the commercial light water reactors built in the U.S. prior to 1982 with several exceptions.
Seven demonstration plants and fifteen turnkeys were excluded from the data base. Also excluded were Ft. St. Vrain, because it is a high-temperature gas-cooled reactor, and Sequoyah 1, because cost data was unavailable.
The 52 plants were arranged roughly in order of obtainment of a' construction permit from the NRC (or AEC). That date (LICDATE), expressed as a decimal -- April 5, 1967, for instance, becomes 67.26, as April 5 is the 95th day of the year; 95/365 = .26 --
was available in the Electrical World annual survey of commercial nuclear power plants.
The capacity of each plant is measured by its design electrical rating (DER). This figure was chosen because it is a good measure of the full size of a plant and because it remains fairly constant over time. For units whose DER's did change between the operation date and the present, the larger of the two ratings was used. DER's were obtained from the NRC " Grey Book."
The data base records the cost of each plant in 1980 dollars, excluding AFUDC. This data base cost was estimated on the basis of reported costs using a procedure that is described in detail below.
The input data for this estimation was available for most plants in Steam-Electric Construction Cost and Annual Production expenses.
Because these EIA/FPC documents report data by site and not by plant, many common-sited plant and duplicate costs had to be estimated.
Additionally, some plants' costs were obtained or calculated from-
-other sources.
A-1
The date of.first' commercial operation (COMODAT) is expressed as a decimal for each plant and was obtained from the NRC " Grey Book."~ Construction time-(PERIOD) is simply the period in years
.from the license date to the commercial operation date.
The data base also contains 4 dummy variables. The variable for duplicate (DUPLI) shows whether a plant is the second or third unit of a common-sited set to be constructed. The variable FIRSTl indicates whether or not a plant is the first unit of a common-sited set. A dummy variable is included which indicates whether the plant is equipped with a cooling tower (CTOWER).
Another dummy variable, NEAST, indicates whether the plant is located in the northeast. Northeast is defined as Connecticut, Maine, Massachusetts, New Hampshire, New Jersey, New York, Penn-sylvania, Rhode Island, and Vermont.
Finally, included in the data base is a measure of experience
' in terms of numbers of plants for architect-engineers (AEEXP). For A-E experience, each plant engineered by a given firm and licensed previous to the plant under question counts as one unit of experience.
(Thie ' includes turnkeys, demos, and Ft. St. Vrain, as well as un-completed plants licensed before Hatch 2). The number assigned to the plant in the data base is equal to the units of previous experience plus one. In the cases where more than one firm was involved in the
.5 architecture-engineering phase of plant construction, the AEEXP
-number refers to the firm with the greater experience. The firm with
.the lesser experience is nonetheless given credit for one unit of
' experience, which is included in the next plant which was A-E'd
' by that firm. A-E 's were reported in NUC Corporation's Commercial Nuclear Power Plants.
The data in the data base is listed in Table A.1, which-follows. Notes to Table A.1 are listed in Table A.2.
A-2
TABLE A 1 NUCLEAR COST DATA BASE
.*w. ..$... ..we*.. . .....
- . . . . .. . .re.
. ** h f.
e
.-ee . L T..
...... .. e
.....f..e a..
1, w.
. . . . f,w* * *n . -
. 4 e .*
v-M ==t-Tr?7?:rtr===r==l=rr=======?=rr=r!==T-r=Ter====?==*===rtTrsr==-trer
, **, t
- r 4*e I ,**FF ?
e
.i t ?. *. g. r
. e. . . 4,C.r? . .
.4t. .e . w a
y e
.e t. e. . hv.e. .
7rc,7*
.I t
- e. .1 e
e
. . r . ...
w .f .
.e
.....s.e 3 I wwwaw e e...*.
- .i.*. :. . -.
e .e .
C w. .
gr ..:r
.e.i 1.* '.15 r., e e.
.s w.... .. e e
. r..... .s
....e . 4 7. .
/.
... ? r... e .* .
e s
.t .
at e...... . ...
. w7.we 1 4 . ..
a.
e .. .
9 s ", a, , .s .) . I .e 9 . t... * * .? E A. ?. w. 4 e. r. .F..*
- e rile - i. .. .e ...e
. .r.
- .o.*e**,
r . .9 wve. ? 4 e
4 e1.wC C o r. . i e rwi. .s .e .s .s .e
- r. . ; . w: r. .
e f I *1*;C?
- e. / ee .*
- r., .
e .
. g. .
. ==..-..e s9.,1. ,..
.. c. <*f. -
. - ww .
. . ....* * . ., e ?...e e
- a. m. e. . ..
g ss /. s =.
% ! e .g
- *q = .;*...e
..... .....e =. .e,.* .* . e e ..
J. a
- e. 6 t
.e
.e
..o .e
.e,. .*.= e. r. .t. ./.. ". . /.
s 9
e.. 4. ..e
.: *1 w.
/.. I
.3 . .i V.?.".*.*- .. e e g .e l
-a.. .. m ,f ..,..,,
e 9 e e I w . . e ..w..e v..--*. . ..
er . * -rs:=.--. 2 2:= 'e=-re:--r:rses--:--m--= en?r- re:----,.e- =s--.-.
..e.
r...e.,.
> . ..- .e ... .... .
. e.
. 4 . . ...... .,
p.., . .
. . ee . . . . - . ..
.... -- ....,., ..,......n..r... ............,.,,ere.........,....
w.. ....,.
..i w.....
..e ... .
.w.e e ..
- e. **S2'
- Etr*
. . . ...e w..e
- r. i. .. .
F77 ......
.<e r. i. e. .
er.
ww
. .. .... . .g e
.r
- e ;*r q e
..we c. wr ?. e i $
. .* *e ? e . .. .. e
- . *"P.
g
.e .eg. r f. e l
n .O t.we r t.
- r o
.... 1. ...pwe l . . ...
. .we w.e p.... e
,.e ..e.e p , .
..e a wwwe
.g
...e e
,..e
.. y
. ...e..
e 6 .e
.e e .. .
e a w...e
. .=
ge..e . a 6 .C. .e g e *
. ...rwe s .. . .
r-.2 - , r a - -- -- e r r-- re-===rre=====r===*==s=r-=ssrere===*?r- ==-r==r====r
.. . .. ==..
P........
e.e.e...we..
.Fcs . m* f....
.***
- w.. .
.n.
7..*6, r..*.
.. e ...
- ?-- -
/
- 1 - T T ? 7 = = = = = r r 8 = = = 1 = = T r = = = r *. =
- r ' ": = ? - r = * * * * = = = = r ? Ter=TTr??e==rar:.
e ,
t e.
...w..
- w. e . . w e.
.4 i . . ,.
w- . .. ...
..oww w.
..ew.,
C ., .s,aw. .. . .
e.- . w.
e.. .e w w .. .w .
wwe .. woe
. .t ...
wwe4w. .. 1. . ..
.e Sw
. I .
.e %
9* I sc e w ww.*T l 4 a 4: r c.
- t. e w . .e .'l o.
. w .w l c. e l e 4 . .w . . .
..w.
. . . . . <we.w s w
.y . . . ..q= ..e... w e r. ,.
wwe .. ..
.. w
. o .w
. a g .- wwe
- t. w.e s g
. c.
e w.e
. og a w.
..ew. . . .
w 7* O
- * *..e t e w.e,w I
w
" ?.
e .. e.
- ", e*
.se 4.f t
- . *4 "r.S e . e w .-** e
- 4. .
,.. =. e, 9 e, e .e*w
- 4. e **
- e
. **. e. si e e **. ,e e . ee.
- i ~ ~ . **
. ee*
- ... e7 1. B
- e
. ww .
e E *. ..ew. e o . 4 e .e 4 e
wwes. *7..*. .
F**=*=&,=t=4-========r==;T===============l===?=*rr=*?R=t-=,T=?-T===TrT==T=*
A-3
TABLE A.1 (Continued)
.. t.e.r .r F r. a r g.S.
. . . .* . *. . . .r w.4 rc
. a t.i .t .? .r. f. . ..
,g
. . . r,iww., 7 .m.
$.i.J
.. .. ;r .w .r,h*. f. P=cr.U a , *^ e
.. . IU w lt<eta;szzzsztszzzzzzzzjazzarzzarzzzurzz[zzzszzzzzzzzz=zz'zzrazarazzz====rl t 11
, ,71.99 1 72.94 1 73 69 1 74.5E I
.,, I
. e.,
-,. 16 i, 3
, w . e.,. i ,.. 69 4.v4 ,
I.
. ! ,e . . , , . .w . .
. s, . c ,
.t . w,i I
. e
.4.4- ,
i
,4.,.
, . . .,4 .e.. v,,- v..3, ,
. , 1 .
i.
w 7.2. c. , -,. , .
,. ., . e. al ,, .. ,7 , , s.4 6
-. l
...w
,I
.,.. 4 .
t c.
w c.
s .s. !,
,..a -
3 .%
9 e4.
, Wr . w,I,C. - w.e
- e. w. G. M. ,.
- 4 . ",. .
a
. r.
- 9. w f .
- 4 &
w 4 ,"e ew.
~.
r w . .-w r
.I I -. ..w 3
,. **, w e ,w, .4 l , w . C. ." I *r..C
.w 4 3 79, . ... e
.* n.Jr . F. 7 . *. e. .wr . I *;...r.
I l.. w
,1 i
- l . . .. 0.. 9. e. . F 7 7 7. 5 .e ... .
3 o, ee - . ;. .M. * .+ "; ,.
. Ie.- . w t
t
- e. w)*.7 w g
...c 6 4 c. . w..ecc. ww ww #
. wo* .cw4 3 c
- i c
. .--7 j ? .21 ** r 6 =s tzza rt r = T * *. u s j t =Tr r r z rg rt z z z z z l 2 3 r z ar et t.r r e r. L n l :: . = ; " r i- - ?. r . s s = ,
. . # .*. *. .?.. ., .
- . *
- c=.F r s.r .?...* * .r. . *..**.. . 7*
.; .?..r i. r
fc.'..'. ;.*~;. FE0n 1 10 5 re-arr:=r i.=:r s e e r u r l : : r r m 3 3 - r --- ; = r u n r v, , r e r . n , = .r - r 2 : z r = = r r =
e.
w e- .,.e .w.
.* ..=.. a.g. e 7.
e
- - .. 4. 1 .4.
. .- *I.
.w .e . 2 s -
w 4 . .
-r .
. . a. . .
. . 4
. . c. l
. . 4 ... . .V.
. ;- . e
., 6 .
--.. .s . . .m . . e
.i .
. , .a s . , . . 9e . l
- . ,s .s
..,e
. l
!.r .
e
.c .,. . . l ..,.
w ...
-- e, w.
t.
. w.
m
- e. .
. w..
I %. .,. . 1,.. . .
- r-er-e-;er==r==raz===zz- r==rs=rezzr====rjaz=rzz=33r=razze = e r= =-es==nes
. ., ..-e. .
....- E F r..i .r,.c..e.... : e * ,,w.
c : a.
.. .r.. ...c.e .r.-EU
.. .. 1 . .e. w r ..
lr-r-wr-l-ete ?:r vr e .
r.
rzlmrezz=seretz=szzlrrrr=rs==m-ters=:rzr=-rr rurzzzzz I t . C. .
t ,
. w .. > .... e.. .
. e- .. . , ..
i
. .. t
.- , .s, . ,
.. A. . . ..
' e i .
l fii,
-r
.w e 6,:
w.
i.
5:
c .
1 5:
.s w .
i i:
,l v.
.7
. s
- v. . w. h.
w l s .S v .
- e. 77 i. V.
g
.. I
.. l I 7. c... 4' 1. I o. t
- t ... !. . .1.,' . e .a. .
w A.
3.. R
.v .
. . /.
v 9
J*T L. 1
.. { ".
1
. w. .w. .
l ar- - r= :- e r m e z z z a r s z =r ss i zn= s:s s === raz z z= : = = = =- z:z . r = = rr . = = r:3-:r:::::::: : .
A-4
TABLE A.1 (Continued) flu;LE AF;. FIT.5T1 -
PATE F.E'.'!EED: 11/07/82 AuuAL DATA FR07. 1 TO 52
- :=======, ==l==== s========= le=============== : =rsessassessant i 1 ! 0. I 1. I C. I 1. I
! 51 0. I 1. I O. ! 0. !
l 91 0. I 1. ! 0. I 1. :
! 13 1 0. 1 0. I 1. 1 0. :
1 17 1 1. ! 0. I C. I c. :
1 21 1 1. I 1. I O. ! 0. t I 25 t O. I 1. I 1. f C. !
1 29 1 1. I O. I 1. t O. !
! 23 1 0. I 1. ! O. ! 1. !
l' 37 1 0. t O. 1 0. ! 1.
I 41 1 1. I C. I C. : 1 l 1 15 1 0. I 1. I O. 1 0. !
! 49 t C. 1 0. I C. I 1.
- --rerre t s marra ssus-arass l a ze = r===rss=== l 3-: ==s3 =s ummas t eu erne sser==rs n l PU LE.^.F._00FL1 - DATE F;E'.'IEED: 11/03/S2
(.f!r."JAL Dr.T A FF.Or.1 TO 52
- -r-==selsenerser l--sm-servannstral= sezzensazzarustass=======
- : ==:
i !! 0. I O. ! 1. 0.
5l 1. ! /. I 1. t 1. '
.! ?I 0. ! C. I 1. ! (.
- 13 I 0. 0. : ;. ! 1.
I 17 1 0, i 1. I 1. 1 0. .
I 21 1 0. ! O. i 0. ! 0. !
l 25 1 0. ! 0. ! C. I 1. :
1 29 1 0. I 1. I O. I 1. !
! 23 1 1. 1 0. I 1. t 0.
37 1 1. 1 0. t O. ! 0. !
! 41 1 0. ! 0. 1 0. : 0.
I 4' I 0. 1 0. ! 1. ! 1. I i 40 ! 1. I 1. I 1. f 0. !
- s -sres::::: ==========st===srs 3 3s ::==l================! r====== sr== rst t UCLE %.NE AET -
DATE F E'.'*EED: 11'03/52 i.tn.U*L DATA FFO:1 1 TO 52 rer===:rsan==st
- s arav=ls===nesse -::armlessar===r=======::::=============11 0.
! I t 0. I O. ! 0. I
! 51 0. I O. ! 0. 1 0.
I ?I 1. I 1. I 1. ! 1. .
i 13 1 0. 1 0. I O. t 0. I i 17 1 0. I O. I O. I O. I
! 21 ! 1. I 1. 1 0. t O. .
I 25 t 1. I C. t C. t 0. .
I 29 1 0. I O. 1 0. 1 0. l 1 33 1 1. I O. I 1. I O.
! 37 1 0. 1 0. I 1. I 1. !
! 41 1 0. I 1. I C. ! 0. t 1 45 1 0. 1 0. 1 0. ! 0. :
I 49 1 0, 1 O. I 1. I 0. !
tr======l==== rs======amat==33 ===========[ ============== le===============l A-5
TABI.E A.1 (Cont.)
m.:t'.E A ..r tt T rm. -
1 F' AL I S Al'E 5 51 Salem 2
.' TU5hCf F0!NT 3 52 McGuire 1
- Tuhr.E f F O!ral 4 4 tteWra 5 FERRY t
- t. ihL.~d's FLkkf J 6 'J".E 4E E 1
? CCCf4EE O L' (d.o4H i 1 v clf 9]tjf T A N'. l E la
- Ara t.gr19n ;
li tt*LH 10 Tram .i
- . f e' E E M IL E ! 5'. A'.!: !
- J F C f.1 C A'. 4surd 1
- CQQ6 (6-
- p 0.' - e I t r. % .'i. . t a iAA:4 ? E l i L A*.:' I c ( 4 *. *. +..r ..,.q',
3 J '. s k t '.'ta i F E'l s 2
.... t :. J Av.:ur i.
.1 *;.u :n !
.. is.E9 !
.! (.;
- E f AL A l'/E k 2
- '. 4
- He SE(0 :
.". m A i r4 E r Arir.E L Ak N A :t Ai ;.MLE Ah Orat t t
. !! 0 r. 1
.. : 10r. .
/ r.v0s 1 3; C09f. .*
3: C 4. .'E R ' LLIFFS s s
.r..
- e. H...r.? . \ v. ... . t g c. s
- . /
- 8. .* ! ' 8 'J . P . ' ,
( .t se?;- :
.b 'r6CE 9!LE !!.Ar.:'
2a '#L W :(- 2 PL*;L: C' .
J tt A F :. y. :.
3' f ; ? ;8 .'.T A I C t.
4( ti. A*.'E F VALLEt 1 4: i'. LUCIE 1 C. a!'.tST0nc :
42 f f ' j ).*.rJ aa *: <'*t ..N.U.
1 4 's l' A'. : $ l'E 6 5.C 1 4 s. FAs . Et :
4 4 4 *. .*.*: t A t fiU C L E A i [ *. i 1.
4" stA'!s
-, .- 2, j2. * ..v (
e.
- . ,,., . . - ...g 9
4.
A-6
TABLE A.2 NOTES TO NUCLEAR COST DATA BASE
- 1. Turnkeys: Demonstration P.' ants:
Robinson 2 Dresden 1 Dresden 2 Indian Point 1 Dresden 3 Big Rock Point Quad Cities 1 La Crosse Quad Cities 2 Shippingport Connecticut Yankee Yankee-Rowe Oyster Creek Humboldt Bay Millstone 1 Monticello Ginna San Onofre 1 Point Beach 1 Point Beach 2 Indian Point 2 Nine Mile Point 1
- 2. Uncompleted plants covered for ASEXP Diablo Canyon 1 Sequoyah 2 Diablo Canyon 2 Fermi 2 Zimmer ,
Midland 1 Midland 2 A-7
TABLE A.2 (Continued)
NOTES TO NUCLEAR COST DATA BASE Source
- 3. Common sited plants for which individual costs were unavailable ~
Browns Ferry 1 -
Cost of each unit #1, year of comm. op., not reported Browns Ferry 2 -
Cost of each unit assumed to be half of 1975 total-site figure.
Oconee 2 -
Cost of units #2 and #3 assumed to be half of difference between 197'4 (total 3-unit site) and 1973 (onlysl unit) figures.
Peach Bottom 2 -
Cost of each assumed to be half Peach Bottom 3 of 1974 total-site figure.
Turkey Point 4 -
Cost assumed to be difference between 1973 total-site figure Surry 2 and 1972 (only unit #3) figure.
Cost assumed to be difference between total-site figure and 1972 (only unit #1) figure.
Prairie Island 2 -
Cost assumed to be difference between 1974 total-site figure-and 1973 (only unit #1) figure.
Browns Ferry 3 -
Cost assumed to be one-third of
' 1977 total-site figure.
Zion 2 -
Cost assumed to be difference between 1974 totalsite figure and 1973 (only unit #1) figure.
Calvert Cliffs 2 -
Cost assumed to be difference bet-ween 1977 total-site figure and 1976 (only unit #1) figure.
Brunswick 1 -
Cost assumed to be difference bet-ween 1977 total-site figure and
- 4. 1976 (only unit #1) figure.
Cost not obtained from Steam-Electric Plant Production Costs:
Hatch 1 -
Georgia Power Company Hatch 2 -
Georgia Power Company Arkansas Nuclear One #2 -
Arkansas Power & Light Company Indian Point 3 -
Report of Member Systems of N.Y.
Power Pool, 1977.
Fitzpatrick -
Farley 2 -
Komano f f Alabama Attorney General's Office North Anna 2 -
Virginia Puolic Utilities Commission Salem 2 -
Atlantic City Electric Company McGuire 1 -
Duke Power Company A-8
Calculation of Constant Dollar Direct Construction Costs As' mentioned above, published data regarding the cost of nuclear plants is in the form of " mixed current dollars" (i.e.,
not adjusted for inflation) and includes the cost of interest during construction in the form of AFUDC, in addition to actual construction expenditurco. For the purpose of predicting the future cost of power plants, it is necessary to separate out the effects of inflation, AFUDC, and trends in actual construction costs net of inflation. To do this, the direct construction cost in 1980 dollars of each plant in the data base was estimated on the basis of the available mixed current dollar costs. This estimation involved both the removal of AFUDC costs and correction for inflation in the cost of inputs to nuclear power plants construction.
The first step in this estimation procedure is to reproduce the flow of construction expenditures for the power plant over time.
This is necessary for two reasons. First, the conversion of the actual (" current") construct 3cn dollars to 1980 construction dollars depends on the year in which the expenditure was made. Second, the accumulation of interest during the construction or AFUDC depends in each construction year on total expenditures prior to that .
year. Therefore, the fraction of the observed total cost attributable to AFUDC (and hence excluded from the direct construction cost). depend on the pattern of expenditures over the construction period.
This pattern of construction expenditures was approximated by an equation developed by William Mooz for the Department of Energy (Refs. A.1 and A.2). Mooz took a typical " cash-flow curve" for nuclear power plants developed by ERDA and applied a standard
" curve-fitting" routine to yield the following equation:
Y t , 1 _ ,,, 3 t
.,3...],,:.,,,
(total)
- E(total) where . Yt is the total construction expenditure up to time is the ultimate total expenditure and t(total) is the length of the construction period.
A-9
To use this equation, it is necessary to define the construction period for a particular plant. The Mocz equation gives very small proportions for the1first and last years.* This is consistent with the interpretation that it includes the initial planning phase of the project-as well as final expenditures after construction is essentially complete. Therefore, for the plants in the data base, the total con-struction period is taken as comn.encing two years before the date the reactor was ordered and ending one year after the date of the commercial operation.
Using the Mooz equation, a yearly stream of expenditure propor-
.tions was estimated for each plant. Given this stream of proportions, the total cost of the plant is given by:
n n TC = I (DC) (FRAC g /D,) (1 + IDCg /2) u i= 1 j=1+1 (1 + IDC))
where: TC = total cost in mixed current dollars including interest during construction; DC = direct construction cost in 1980 dollars FRAC = fraction of DC spent in year i D1 = cumulative deflation factor for year i IDCg = interest rate in year i; and, n = total number of years in construction period.
What this equation says is that each year the amount spent on
. construction in constant dollars is the total constant dollar con-struction cost (DC) times the Mooz fraction for that year (FRAC g ).
.To convert this from 1980 dollars to current dollars, we divide by the cumulative inflation from that year to 1980 (D g )., This direct construction expenditure is increased by interest costs in the form of AFUDC. In the first year, this cost is one-half the interest costs-for that year (assuminc that expenditures are evenly spent over
- For example, for a ten-year construction period, .0061 would be spent in the first year, .48 in the second year, and .17% in the last-year.
A-10 l
the pe iod). Thus the actual amount spent in year i due to construction in year i is (1 + IDC /2)
, 1 times the actual construction expenditure in' year i (CDC) (FRAC g/D g) . There is however an additional cost due
_to this expenditure in all. subsequent years because of interest.
The original expenditure'in year i will be increased each subsequent year by an amount equal to that year's. interest rate. Therefore, the total cost resulting from year ll's' expenditure is the expenditure
.in year i. times the product of the series of terms of the form (1~+ IDC), with one such term forach year after year i, up to the
~
e last year (n)*.
Finally, the total cost (TC) is th'e sum over all years of each year's total cost impact.
This equation may be easily rearranged to give DC as a function of TC. Since DC appears in every term of the sum, it can be taken outside the summation sign. Dividing through by the sum gives:
~
' TC n
n DC = I (FRAC ggD ) (1 + IDCg /2)
- i=1 3=1 + 1 (1 + IDC))
This equation was used to calculated the 1980 dollar direct con-struction cdst (DC) on the basis of-the total mixed current dollar cost (TC).
The yearly inflation rates and interest rates used for this pur-
- pose are presented in Table A .3, inflation rates were calculated from Reference A.3. Interest rates are from Reference A.4.
l 2 .
- 'This is not strictly dorrect Secausb interest in Ehe last year will
~
generally not accumulate for"the full year. In the actual calculation, the interest rate for'the last year was reduced proportionately to j the fraction of the year before-the operation date.
.a ,_ .
'- A-ll i
4 Regression Results A series.of statistical analyses of the nuclear cost data base were used to.try to determine the factors affceting the cost of nuclear plants after inflation and interest have been removed. The variables that were considered as possible factors affecting the construction cost were:
LICDATE- -
construction permit issue date;
.MWDER -
Capacity of the plant in MW (design electrical rating);
AEEXP -
" experience" of the architect-engineer; PERIOD -
construction period (commercial. operation date minus construction permit issue date);
FIRST1 - whether or not the plant is the first of a common-sited set; DUPLI -~ whether or not the plant is a " duplicate"
< on its site;-
NEAST - whether or not the plant is located in the Northeast; CTOWER -
whether or not the plant has cooling towers.
The last four (4) variables are " dummies" having the value of one (1)-if the plant has the attribute, and otherwise having the value zero (0). For all the statistical analyses, the cost data were standardized to a cost (1980 dollar direct
, construction expenditures) per continuous KW basis.
i :-
f.
A-12 T
TABLE A.3 COST ESCAMTION AND INTEREST RATES Nuclear Cumulative Construction De f la tion Inout Cost Factor Interest Year Escalation to 1980 Rate (t) (%)
1965 2.7 2.832 3.8 1966 2.2 2.758 3.9 4.0
- 1967 3.6 2.697 1968 4.4 2.605 4.3 1969 5.8 2.495 4.6 1970 8.1 2.358 5.1 1971 10.1 2.182 5.5 1972 4.4 1.982 5.7 1973 6.4 1.898 5.9 1974 17.8 1.783 6.3 1975 10.8 1.514 6.8 1976 7.9 1.366 7.0 1977 5.9 1.267 7.1 1978 9.3 1.196 7.3 1979 9.4 1.094 7.6 1980 11.4* 1.0 8.6 0.898 9.6
=
estimate a
A-13 4
The eq ation ut:cd to is .esent *ho effect of these factors on Cost:
cost /kw = A + Al X LICDATE + A2 X log (MKDER) + A3 X log (AEEXP)
+ A4 X PERIOD + A5 X DUPLI + A6 X NEAST +
A7 X CTOWER + A8 X FIRSTl The results of simpic Icast square estimation of the coefficients in these equations are given below:
COEFFICIENT T-STATISTIC LICDATE 98.0901 6.84*
LOG (MFDER) -243.1260 -2.39**
LOG (AEEXP) -31.1383 -1.36 PERIOD 34.9633 2.18**
DUPLI -129.9980 -2.32**
NEAST 190.8200 4.38*
CTOWER 62.2815 1.61 FIRSTl 95.6245 1.92**
CONSTANT -4713.9300 -4.10*
2 0.720 R
2 CR 0.668 0- 132.52 P 13.80 DW l.51 significantly non-zero at 99.c. confidence; significantly.non-zero at 95.Si conficence:
A-14
This equation indicates that the cost per kw for nuclear plants increased $98 per year over the data base period. This effect ie' extremely significant statistically, and also by far the largest effect in magnitude in explaining the cost variation in the data base. All other things equal, a plant licensed at the start of the data base period (March, 1967) cost about $600 /kw less than one licensed at the end - (February 1973) .
The coefficient for the term 109 (MWDER) measures " economies of scale" in nuclear plant construction. It indicates.,that a doubling of plant size reduces the cost /kw by about S169/kw. This is a
+
fairly modest effect, since most of the plants are in the 800 to 1100 MW range. .Though small, the effect is quite significant statistically.
The-third term represents the reduction in costs that occur because the architect-engineer gains nuclear construction experience.
This effect is not very sitnificant statistically, and its magnitude is moderately small for most plants. AE experience in the sample ranges from one (1) to thirty-one (31); this corresponds to a variation of about $107/kw from the experience effect. Most engineers now'have experience with at least six (6) plants; the difference pre-dicted by the equation between a plant built by an AE with thirty-one -(31) plants compared to six (6) plants is about SS1/kw.
As.for the dummy variables FIRSTl and DUPLI, the effects of both cre quite significant statistically, and both are moderately large in magnitude. The variable FIRSTl indicates that plants which are the first of a common-sited set are S96/kw more expensive than others. The variable DUPLI indicates that second and third plants on a site are S130/kw. cheaper than other plants. It is not clear, however, to what extent this effect is an artifact of the'way the costs of common-sited plants are allocated. As noted above, reliable data was unavailable for some of these pairs, necessi-tating the use of scmewhat crude assumptions to derive separate cost
. estimates.
The dummy variable CTOWER is f air 1v- significant statistically.
It shows that cooling towers add about $62/kw to the cost of a plant.
~
1 A-15
[ The dummy variable for plants in the Northeast, NEAST, is very significant statistically. It is also large in magnitude, adding $191/kw to the cost of a plant.
Application of Nuclear Capital Cost Equation to Catawba #2 The equation described above was used to predict the direct construction cost of Catawba unit #2. For the currently forecast date of commercial operation, the resulting direct construction cost is $1525/KW (in 1983 dollar's) . The total cost of the unit in current dollars, including AFUDC, is
$2,455 million.
The direct construction cost was calculated according to the equation using the following data:
LICDATE. 1,975.60 MW 1,145 AEEXP 6 PERIOD 11.86 DUPLI 1 NEAST 0 CTOWER 1 FIRSTl 0 SCENT 0 Through 1982, a total of $722/KW had been spent on direct construction of Catawba #2. This figure is based upon information in the Preliminary Official Statement for the North
-Carolina Municipal Power Agency's Catawba Electric Revenue Bonds, Series 1983 A, dated May 10, 1983. After the expenditures through 1982 were taken out, the remaining construction costs were spread out over time based upon an inflation rate of 6%.
The Mooz equation, described above, was used to derive the shape of the: annual expenditure curve.
A-16 m
-- r- w
The AFUDC total cost through 1982 was estimated as 20 i
percent of the direct construction experienced.through that date. In future years, the AFUDC was determined based upon the annual direct construction cost projections and an AFUDC rate of 10.3 percent. This rate is the average rate on the. Series 1983 A Bonds.
The annual construction costs and AFUDC for Catawba
- 2 are listed in Table A.4.
A-17
TABLE A.4 CATAWBA UNIT #2 CONSTRUCTION COSTS (In Millions of $)
190 ) ifu. lon) J < > : !.i 190 193,'
T til fel' t I H
en t titl?. I . ' ? 'f .
. .*'o. .lH. I h e, . 6 .2. - 1.? .
(el lifel i 11. O. 1.'".'. 1. * * . 1H'. "'.
! s'il:
118 I (.I
,. 3 1. . _ /. . **l.
. I O '.' .
1 'l t t 's ' i 18 8.
t II I iel .'<,<,. 1 1 1 '? . 1 101:, 1,';l 1. 2 0 l '. . . *I1.
Ileatil it Il't t i t'ait M.Il'Itt. I t til! I i'
. 1 / l.14 1 c. ( ) . r'l1 1 Itif8 tiill I i s!' -- I 111 e il. I ' t . .'i t 8 I I t . l ie l sil t' tie 'l' ,il.
i.I's.ifte t ie .it 'ti l'..
REFERENCES A.1 William E. Mooz Cost Analysis of Light Water Reactor -
Power Plants, The Rand Corporation R-2304-DOE (June, 1978)
A.2 William E. Mooz A Second Cost Analysis of Light Water Reactor Power Plants R-2504-RC (De ember, 1979)
A.3 Handly-Whitman Index of Public Utility Construction Costs, Baltimore.
A.4 United States.E.I.A., Statistics of Privately Owned Electric Utilities in the United States; 1977. The
" Average Interest Rate" from Tables 12 and 13
" Interest on Long-Term Debt" was used.
2 i
G A-19 e ~ ~
V EXHIBIT (RAR-6)
Capacity Factors 1
i.
o b._ 4. . - _ _ , . ._. . . _ . . _ . . _ _ . , _ _ . . _ _ _ _ . . . . _ _ . _,._.,.__._._m... _ , _ _ _ _ _ . , , _ _ _ . _ . . . _ , _ _ _ _ _ _ _ _ _ _ . _ _ . . _ . . . . _ _ _ _ _ . . . ._
Exhibit (RAR-6)
B. CAPACITY FACTORS 3.1 Introduction The maximum output of a power plant over the course of a full year of operation is the product of the total number of hours in a year (8,760) and net full rated capacity of the unit. Thus, for a 1,000 MW plant it would be 8,760 GWH.* This output is never achieved for a number of reasons. First, power plants require outages for scheduled maintenance and equipment repair. Second, they often suffer unscheduled or forced out--
ages which require maintenance and equipment repair. Fi.nally, some power plants are operated to perform load following and are consequently brought on- or off-line as system loads experience upward or downward swings. Peaking units are an extreme example of this latter phenomenon, often being run only several percent of their available hours during the year.
4 Nuclear power plant operation differs from this in a
' number of respects. Nuclear units do suffer forced outages and require scheduled outages whose magnitude and character are specific to this technology and its requirements. But only rarely have some nuclear units been operated to load follow.
The costs and technical characteristics of nuclear units require that they be operated in the baseload mode, generating electricity at all hours when they are available to do so, and
- A GWH (gigawatt-hour) is one million kilowatt-hours or one thousand megawatt-hours.
B-1 y e., , - ---s--- ,*,m--
coming off-line only when necessary. For nuclear plants, l.
however,-the necessity of coming off-line extends beyond the
-imperatives of forced and scheduled maintenance and equipment outages. Nuclear units require rather long down times for
' refueling, on a twelve to eighteen month cycle. Moreover, Nuclear Reg'11atory Commission mandated outages for inspection,
- safety, training and licensing can also temporarily remove nuclear units from service.
Capacity factors are generally defined as the net
, electrical generation divided by the maximum possible generation over the course of a year (or any other time period). For a full year of service the capacity factor can be expressed:
i Annual Capacity Net Generation (MWH)
Factor (in Percent) Design Electrical' Rating x 100 (Net MW) x 8,760
- In effect, the capacity factor is the fraction of time (e.g., a '
year) a unit is generating electricty at full rated capacity.
Nuclear power plants have had capacity factors which on average have been far below the expectations of the industry.
'A simple compilation'of this industry-wide experience is provided by the U.S.-Department of Energy (Ref. B.1), and is reproduced here for the last ten years in Table B.l.below. A number of observations'can be made regarding this simple compilation. First.,'in only three out of ten years were the l ,
average industry-wide capacity-factors above sixty percent.
Second, both the ten year average experience of 56.7 percent and:the last four years experience of 56.1 percent fall far B-2 e
. - ~ - , - = , . . . - , , + _ . - - . - . . . . . - , - n . .. . - . - , , - - . . , - , , , , , ,
i i
below industry expectttions, which have ranged between 65 percent and 80 percent. There is no evidence here of an industry-wide learning process, that is, no general improvement over time.
Thus, on the basis of the experience of the industry as a whole, a capacity factor of about 57 percent would be a plausible assumption for a nuclear facility.
TABLE B.1 NUCLEAR POWER PLANT CAPACITY FACTORS IN THE UNITED STATES 1973-1982 CAPACITY FACTOR YEAR (Percent) 1973 63.2 1974 43.5 1975 55.2 1976 53.5 1977 62.9 1978 63.9 1979 57.6 1980 55.1 1981 56.6 1982 55.0 Average 56.7 B.2 Statistical Analysis While the above industry-wide average capacity factors are
' instructive in themselves, it is useful to explore further the rather wide variation in plant-by-plant and year-by-year experience. The first step in this process is to segregate the nuclear plants by reactor type and size to see whether major differences emerge. Another characteristic, whether salt-water is used for cooling, may be important insofar as corrosion B-3
i related problems could emerge. Table B.2, below, summarizes industry-wide average capacity factors from 1975 through 1981, segregating by reactor type (PWR or BWR*), size (greater or less than 800 MW), and cooling system (salt-water or not). For this table, maximum potential electrical generation is the product of hours in-service and the net design electrical rating (DER) of the unit.
TABLE B.2 NUCLEAR POWER PLANT AVERAGE CAPACITY FACTORS 1975-1981 (Percent)
PWR BWR ALL REACTORS SALT NO SALT ALL SALT NO SALT ALL SALT NO SALT ALL
< 800 MW 61.7 72.8 69.9 59.7 59.7 59.7 60.7 65.0 64.0 I 800 MW 59.2 55.3 56.8 46.4 57.7 55.6 57.5 56.1 56.5 ALL SIZES 59.7 60.2 60.0 54.5 58.8 58.0 58.4 59.6 59.3
B-4
Regression Analysis: Introduction More detailed examination and analysis of the variation in nuclear power plant capacity factors has been performed by ESRG.
This undertaking is applied to a data base consisting of 68 nuclear power plants (essentially all commercially operating units) in the U.S. and their electricity production and outage experience from 1975 through 1981. The source of this data is the Nuclear Regulatory Commission (NRC) " Grey Books" (Ref. B.2). The analytical technique applied is multivariate regression analysis, which explains the observations (experienced annual capacity factors) in terms of a linear combination of independent variables selected because they are believed to have a causal or associative relationship to the observations. The variables explored in regression analysis include unit size, reactor type, unit age, presence or absence of cooling towers, salt-water cooling or not, steam syste.m supplies, and commercial operation date.
The age (years of operation) variable is interesting in two respects. First, it can express a maturation effect, i.e.,
improvement after the first few years (or " shakedown" period) of operation. Second, it can express aging phenomena, i.e.,
deterioration of performance with age, after mature levels have been reached. In order to test for such phenomena it is necessary to use broken linear, rather than a single linear age variable. In addition, to examine whether aging effects differ with plant characteristics, i.e., reactor type, size, salt-water cooling, product terms of age times these variables have been employed.
B-5
A further discussion of the data base used for regression analysis deserves attention here. Since NRC mandated outages occur somewhat episodically or randomly over the data base period, these outage hours have been removed from the analysis. Similarly, refuelling outage hours have been removed, since these too add a degree of scatter or randomness because they sometimes overlap calendar years anc the cycle itself can vary rather widely. While regression could readily be performed on the raw or unadjusted capacity factors, here we have applied it to adjusted capacity l
i factors in order to obtain a better analysis of the factors which contribute to forced and scheduled maintenance and equipment outages. The adjusted capacity factor is:
Net Electrical Generation Adj. Cap. Fac.* = Design (Net) Electrical Rating x (8,760 - Refuelling and NRC Outages)
Average Refuelling and Regulatory Outages Before turning to the analysis of adjusted capacity factors, f
then, it is.important to examine the magnitude of these adjustments themselves. Once results are obtained for the adjusted capacity l factors, refuelling and NRC-related outages must be factored back l in'to obtain the net result, i.e., experienced capacity factor.
Table B.3, below, shows the magnitude of these adjustments on an industry-wide basis for PWR's, BWR's and all reactor types.
- For plants that commence commercial operation after January 1st, the start-up year adjusted capacity factor is calculated as:
Adj. Cap. Fac. = Net Electrical Generation Design (Net) Electrical Rating x [(8760 x fraction of year in operation) -
refueling and NRC outages]
B-6 m 9-- - -e w-p ,-4.- wyg,g+ we y w +- 9 g y r---wyy-y-- wyy--yyeg aww-wyw vs 9rvg--+gw-yv-'
TABLE B.3 L
NUCLEAR POWER PLANT REFUELLING AND NRC-RELATED OUTAGE RATES (1975-1981)
PWR BWR ALL REACTORS REFUELLING .13 .13 .13 NRC-RELATED .04 .01 .03 TOTAL ADJUSTMENT .17 .14 .16 The table shows-that, on average, nuclear units have experienced outages of 13 percent per year (6.8 weeks) for refuelling *, and about 3 percent per year for NRC-related
- shutdowns +, during the 1975-1981 period.
Regression Analysis: Discussion
- i. The independent variables selected in the multivariate
! regression specification for adjusted capacity factors are defined below in Table B.4.
The results of the multivariate regression analysis are a
provided below in Table B.5. The first column designates the 1
( variable,' defined in Table B.4, the second column provides the
[ regression coefficient, the third column gives the T-statistic, and L
the fourth column gives the confidence level. The regression coefficient is the measure of the magnitude of a variable's contribution-to the observed result (here, adjusted capacity-factor). Thus, for example, the values .066 for the TOWERSU i
- Since outage hours are-reported to.the NRC for each major outage l mode, refuelling outages can sometimes include time for scheduled
. maintenance.
+ Including NRC-mandated. outages, licensing and
- raining outages.
B-7
~ . - - , . - - - . -
variable and .157 for the TOWERSU x PWRU term show that a BWR plant with cooling towers will experience, on average, a 7 percent higher adjusted capacity factor than a once-through cooled BWR or PWR, whereas a PWR with towers will experience a 9 percent lower adjusted capacity factor-than a PWR or BWR with a once-through cooling system. The T-statistic is the measure of the significance of the variable in explaining the observed variation in adjusted capacity factors. The confidence level indicates the probability that the independent variable will have an absolute value significantly greater than zero.
Finally, summary statistics, including the corrected R-SQUARED and F-ratio which are measures of the goodness-of-fit of the entire equation, are also provided in Table B.5.
i B-8
TABLE B.4 l
' INDEPENDENT VARIABLES SELECTED FOR REGRESSION SPECIFICATION FOR NUCLEAR POWER PLANT ADJUSTED CAPACITY FACTORS Variable Name Definition DERU Unit size in megawatts PWRU 1 if unit is PWR 0 otherwise SALTU 1 if unit is salt-water cooled 0 otherwise AGE Years of commercial operation according to calendar years. The first calendar year of operation averages only one-half a year of plant operation TOWERSU 1 if unit has cooling tower 0 otherwise AGE 4 AGE-4 for Age 1 4 0 otherwise AGE 6 AGE-6 for Age 1 6 0 otherwise AGE 10 AGE-10 for Age i 10 0 otherwise BWSTM Babcock and Wilcox Steam System WESTM Westinghouse Steam System B-9 m -w- , -
w, --
TABLE B.5 REGRESSION RESULTS FOR NUCLEAR POWER PLANT ADJUSTED CAPACITY FACTORS 1975 Through 1981 /
Equation Independent Value of Confidence Coefficient Variable Coefficient
- T-Statistic
- Level A .697 8.24 99.8%
+ B x DERU -1.77 x 10-4 -1.32 80.0%
+ Z x DERU x PWRU -2.49 x 10-4 -3.15 99.0%
+ C x PWRU .418 4.79 99.8%
+ G x SALTU l.38 3.96 99.8%
+ E x AGE .007 .945 50.0%
+ X1 x SALTU x DERU -3.95 x 10-4 -3.02 99.0%
+ K x PWRU x TOWERSU .157 -4.23 99.8%
+ W x AGE x PWRU .016 -2.79 99.0%
+ D x AGE x DERU 3.63 x 10-5 2.48 -
98.0%
+ L x TOWERSU .066 2.35 95.0%
+ S x SALTU x AGE .112 -3.90 99.8%
+ F x SALTU x PWRU .085 1.17 50.0%
+ H x SALTU x PWRU x i AGE .020 -1.99 95.0%
+ L3 x . AGE 6 .026 1.06 50.0%
+ M2 x AGE 4 x DERU 9.28 x 10-5 3.44 99.8%
+ M3 x AGE 6 x DERU -6.54 x 10-5 -1.67 90.0%
+ N2 x AGE 4 x SALTU .064 -1.58 80.0%
l '
+ N3 x AGE 6 x SALTU .063 1.56 80.0%
+ N4 x AGE 10 x SALTU .093 2.44 98.0%
+ X2 x BWSTM .074 -2.34 95.0%
+ X3 x WESTM .019 .781 50.0%
Number of Variables = 22 Standard Error of Regression = .139 R-Squared = .333 F(21/398) = 9.48 Corrected R2 = .298 COND(X) = 114.43
- Rounded-B-10
I Regression Analysis: Results It is instructive to apply these regression results to l prototypical nuclear plants to examine the predictions of the entire equation. In Figures B.1 through B.4, the results are applied to four generic types (PWR and BWR, salt-water cooled and non-salt-water cooled). The impact of plant size (600 MW, 800 MW, 1,000 MW, and 1,200 MW) is also shown. The capacity factors used in these figures are the product of the regression result for each given prototype and a correction factor for refuelling and NRC-related outages (1 .17 = .83 for PWR's and 1 .14 = .86 for BWR's from Table B.3).* The cooling tower effects, about +7 percent for BWR's and -9 percent for PWR's, applicable to some non-salt-water cooled units are not incorporated in the results presented in Figures B.1 through B.4.
- The correction factor is not applied to the estimated capacity factor for the first year of operation.
B-ll'
~ . - . , . . . ,. . - . . . - - _ . -
t .
l Figure B.1 1
PROJECTED CAPACITY FACTORS FOR GENERIC PLANTS BWR: FRESH WATER COOLED 80 1000 MW 800 MW 60 50 800 MW 1000 MW 40 CAPACITY FACTOR (t) 30 20 1 2 3 4 5 6 7 8 9 10 11 12 YEAR B-12
Figure B.2 PROJECTED CAPACITY FACTORS FOR GENERIC PLANTS BWR: SALT WATER COOLED 80 70 60 800 .W 50 1000 MK 40 CAPACITY FACTOR
(%)
30 20 1 2 3 4 5 6 7 8 9 10 11 12 YEAR B-13 1
Figure B.3 PROJECTED CAPACITY FACTORS FOR GENERIC PLANTS PWR: FRESH WATER COOLED 80 800 MW 70 1000 MW 60 50 -
40 30 CAPACITY FACTOR
(%)
20 1 2 3 4 5 6 7 S 9 10 11 12 YEAR B-14
I Figure B.4 PROJECTED CAPACITY FACTORS FOR GENERIC PLANTS PWR: SALT WATER COOLED 80 70 60 50 40 80 0 .sn CAPACITY FACTOR (0) 30 1000 .T..
2C 1 2 3 4 5 6 7 S 9 10 11 12 YEAR B-15
REFERENCES B.l. Monthly Energy Review, DOE /EIA (83/04), April, 1983, p. 76.
B.2 Licensed Operating Reactors: . Status Summary Report (NUREG-0020), Nuclear Regulatory Commission, Office of Resource Management.
J.
g%.,
e l
L l'
l O %
- j. B-16 l
t-l'
I BEFORE THE PUBLIC SERVICE COMMISSION OF SOUTH CAROLINA DOCKET NO. 82-352-E In the matter of: )
)
. s Application of Piedmont Muni- )
cipal Power Agency for Author-) Certificate of Service ity to Acquire a Portion of )
the-Catawba Nuclear Station )
)
I, Michael F. Lowe, a hereby certify that I have this 21sk. day of June 1983 served copies of the attached " Testimony of-Nich~ard A. 51osen" on the belo'w named persons by hand-selivering same to'the addresses set forth below.
/ .
Robert T. Bockman -
~
James H. Still Eighteenth Floor Executive Director Bankers Trust Tower South Carolina Public Service Columbia, South Carolina ;.- , Commission
/ - 'N
~
lik Doctors Circle Columbia, South Carolina
/
'3 .
- ~. , _ , p ,; y e chaejp?'. Lowe Palmqcto Alliance, Inc.
x s' ,
m*(* y 4
N ,%
==
- i. . + %
_ft.., . .
O e 4
W e L
'se an ,
'% e
- r. 7, .
.w/ + 'y #
! O f a
- rr
[a r' -
l., , , , , . , - - - , - . ,