ML19329G216
| ML19329G216 | |
| Person / Time | |
|---|---|
| Site: | Indian Point  |
| Issue date: | 06/12/1980 |
| From: | Bickwit L, Hanrahan E NRC OFFICE OF POLICY EVALUATIONS (OPE), NRC OFFICE OF THE GENERAL COUNSEL (OGC) |
| To: | |
| Shared Package | |
| ML19329G217 | List: |
| References | |
| TASK-IR, TASK-SE SECY-80-283, NUDOCS 8007140182 | |
| Download: ML19329G216 (68) | |
Text
i UNITED STATES NUCLEAR REGULATORY COMMISSION WASHINGTON, D. C. 20555
'~* ~'*
3""' '*'
INFORMATION REPORT FOR:
The C issioners hw 4~
FROM:
E ward J. Ha an, Director Off e of Po cy Evaluatic Leonard Eickw t
General Counsel
SUBJECT:
REPORT OF THE TASK FORCE ON INTERIM OPERATION OF INDIAN POINT (DOCKET NOS. 50-24' AND 50-286)
CONTACT:
Robert Bernero, RES, 492-8528 (Section 1)
George Eysymontt, OPE, 634-3302 (Section 2)
George Sege, OPE, 634-3295, (Section 3 and general)
SECY NOTE:
Copies of this paper were advanced to each Comissioner on June 12, 1980.
One copy of NUREG-0340, which is referenced in the report has bebn provided to each Comissioner for information.
DISTRIBUTION Comissioners Comission Staff Offices Exec Dir for Operations ACRS Secretariat 8007140$M
l CONTENTS Page INTRODUCTION i
SECTION 1.
ACCIDENT RISK CONSIDERATIONS l
Population Distribution 1
Reactor Ac~ dent Risk Parameters 6
Site Aspects-7 The Effect of Design on Risk at Indian Point 23 The Sensitivity of Risk to Variations in Site, Public 35 Protection, and Design / Operating Characteristics The Risk of an Indian Point Reactor Compared to 38 Other Reactors Reduction of Operating Power Level 38 SECTION 2.
SOCIAL AND ECONOMIC IMPACT CONSIDERATIONS 41 Effects of an Indian Point Station Shutdown on Electrical 41 Power Reliability in the New York Power Pool Other Effects of Indian Point Shutdown 44 SECTION 3.
SUMMARY
OF PUBLIC COMMENTS 46 Safety Arguments 46 Impact Arguments 53 APPENDICES A.
Sample Generation of a Complementary Cumulative A-1 Distribution Function - CCDF B. 'Rebaselining of the RSS Results B-1 C.
Letter to Edward J. Hanrahan, Director, Office of
'l Policy Evaluation from Richard E. Weiner, Department of Energy-
)
e INTRODUCTION This report is submitted in response to Section D. The Task Force on Interim Operation, of the Commission's Order of May 30, 1980, in the Matter of Con-solidated Edison Company of New York, Inc. (Indian Point, Unit No. 2) and Power Authority of the State of New York (Indian Point, Unit No. 3)., (Docket Nos. 50-247 and 50-286.)
The May 30 Order established an approach, including adjudication, for resolving the issues raised by a petition by the Union of Concerned Scientists (UCS) that called, among other things, for shutdown of Indian Point Units 2 and 3.
The Director of the Office of Nuclear 'eactor Regulation had issued a decision regarding that petition on February 11, 1980.
Section D of the May 30 Order directed the General Counsel and the Director, Office of Policy Evaluation, to establish a task force to prepare a report to the Commission on information available at this time that bears on the question of whether to permit, p-4.ibit, or curtail operation of Indian Point Units 2 and 3 during pendency of the adjudication. The task force report was to include information on at least certain specified topics listed in the Order. The topics fall into two categories:
accident risk considerations (items 1 to 4 of Section D, at pages 6-7 of the Order) and social and economic impact considerations (item 5, at page 7 of the Order).
The accident risk considerations are addressed in Section 1 of this report.
Those considerations. include comparative site demography; accident risk comparisons; effects of emergency response; and effects of differences between Units 2 and 3, of changes ordered by the Director of-NRR, and of
o, i i power-level reduction.
Effects of uncertainties are discussed.
Some explanatory details are appended.
(Appendices A and B)
Social and economic impact considerations are addressed in Section 2.
The principal considerations addressed include effects of shutdown or power reduction on (a) reliability of the electric power supply for the region, including New York City, and (b) sources and cost of electrical energy.
Supporting information from the Department of Energy is appended.
-(Appendix C)
Public comments relevant to interim operation or shutdown, received in response to the Commission's February 15 solicitation of comments, are summarized in Section 3.
The principal contributors to this work were Robert M. Bernero, Roger M.
Blond,'W. Clark Pritchard, and Merrill A. Taylor, of the Office of Nuclear Regulatory Research; and George Eysymontt and George Sege, of the Office of Policy Evaluation.
9 9
s c
r w
'. ~
SECTION 1.
ACCIDENT RIS_K C_0NSIDERATIONS This section presents estimates of the accident risk posed by operation of the plants in their present condition; a comparison of the risk from other sites and designs; the sensitivity of th'at risk to emergency protective meastgres, and the sensitivity of risk to a reduction in power level during operation.
THE POPULATION DISTRIBUTION The Indian Point Power Station, with New York City less than 50 miles to the south, has the largest population in its immediate surroundings of any nuclear power station in the United States.
" Demographic Statistics Pertaining to Nuclear Power Reactor Sites," NUREG-0348, tabulates all U. S. nuclear power stations according to the total population within a circle of given radius from the reactor.
Tables 1, 2, and 3 show the j
populations at distances of 10, 30 and 50 miles based upon the 1970 census.
The region around the Indian Point station is the most densely populated as shown by these data.
When considering reactor accident risk, the population in a given direction, (i.e., in one 22h degree sector), is often more significant than population density averaged over all directions.
Reactors have been ranked by theb sector population in Table 4.
Here too, Indian Point ranks among the highest. However', 'a number of other U. S. reactor sites, for example, l
Zion and Limerick, also have relatively high populations in their vicinity.
e
TABLE 1 4
Population Statistics Between 0 and 10 Miles
~
POPULATION STATISTICS 19 79 REVISION 5/79
- BASED ON TIIE YEAR 19 70 POPULATION STATISTICS WITHIH 0-10 HILES TOTAL NUMBER of SITES = 111 IIINIlluH POPULATION =
0 HAXIMUM POPULATION = 218398 '
HEAN POPULATION =
36931 HEDIAN POPULATION =
24269 90% PERCENTILE POPULATION =
83557 STANDARD DEVIATION =
39164.6 COEF. OF VARIATION =
1.060 No.
_ SITE NAME POPULATION NO.
SITE NAME POPULATION No.
SITE NAME POPULATION 1
SUNDESERT O
38 DAVIS BESSE 153?O 75 OYSTER CREEK 36797 2
WPPSS 2 455 39 SKACIT 16038 76 FORKED RIVER 36797 3
PEBBLE SPRINGS 878 40
. CALVERT CLIFFS 16827 77 STERLING 37705 4
PALO VERDE 1892 41 WOOD 16889 78 OCONEE 37831 5
WPPSS 1&4 2648 42 FORT CALHOUN 17401 79 HCCUIRE 39374 6
S OUTil TEXAS 3254 43 PHIPPS BEND 17665 80 ERIE 40206 7
VOCTLE 3500 44 RIVER BEND 19147 81 NEW ENGLAND 41882 8
HATCil 4803 45 PRAIRIE ISLAND 19401 82 HUMBOLDT BAY 45403 9
WOLF CREEK 5260 46 BYRON 20377 83 CREENE COUNTY 45786 10 COMANCllE PEAK 5353 47 MARBLE HILL 20959 84 SAINT LUCIE 46066 11 SUMMER 5656 48 POINT BEACH 21073 85 CINNA 46325 12 RANCHO SECO 6061 49 YANKEE ROWE 11763 86 SUSQUEHANNA 50436 13 LACROSSE 6209 50 BRAIDWOOD 21942 87 PILCRIM 51203 14 DIABLO CANYON 6302 51 BELLEFONTE 22709 88 COOK 53006 15 COOPER 6363 52 ZIHMER 23023 89 SHOREHAM 54251 16 CRAND CULF 7245 53 VERHONT YANKEE 23030 90 HADDAM NECK 60374 17 BIC ROCK POINT 7551 54 ELK RIVER 23890 91 TROJAN 61E55 18 WATTS.BAR 7674 55 ARKANSAS 24141 92 HIDLAND 62000 19 NORTH ANNA 7713 56 NEW HAVEN 24397 93 CATAWBA 65901 20 H AL LAH 8365 57 SAN ONOFRE 25725 94 S U RRY 66630 21 FORT ST. VRAIN 8366 58 PEACH BOTTOM 25984 95 HAVEN 67981 22 TYRONE 8632 59 MAINE Y ANKEE 26000 96 PIQUA 72560 23 YELLOW CREEK 8828 60 ROBINSON 26016 97 PERRY 73600 24 CALLAWAY 8914 61 QUAD-CITIES 26739 98 DUANE ARNOLD 79310 25 FARLEY 9528 62, BROWMS FERRY 27215 99 S EAB ROQK 79478 26 WPPSS 3&S 9767 63 SALEM 28562 100 BAILLY 83608 27 BRUNSWICK 10000 64 HOPE CREEK 28562 101 PATHFINDER 84117 28 BLACK FOX 10404 65 PALISADES 29528 102 TURKEY POINT 88000 29 II ARTS VI LLE 11340 66 DRESDEN 31126 103 BONUS 89000 30 CRYSTAL RIVER 11699 67 CHEROKEE 31877 104 BEAVER VALLEY 105000 105619 31 CLINTON 11889 68 DOUGLAS POINT 32020 105 MILLSTONE
- 134206 32 CVTR 12029 69 SEQUOYAH 32145 106 FERMI 33 SHEARON llARRIS 12132 70 JAMESPORT 33200 107 THREE MILE ISLAND 136400 34 H0HTICELLO 12344 71 PERKINS 34369 108 SHIPPINGPORT 143371 35 KEWAUNEE 12759 72 WATERFORD 3567P 109 LIHERICK 152644 36 LASALLE 13343 73 NINE HILE POINT 36000 110 ZION 190314 37 CARROLL COUNTY 13999 74 FITZPATRICK 36000 111 INDIAN POINT 218398 9
e
TABLE 2 4
P0pulation Statistics Between 0 and 30 Miles
. POPULATION STATISTICS-19 79 REVISION 5/79 BASED OH Tile YEAR 1970 POPULATION STATISTICS WITIIIN 0-30 HILES TOTAL NUMBER OF SITES = 111 HINIMUM POPULATION =
87 HAXIMUM POPULATION = 3984844 IIEAN' POPULATION = 531127 HEDIAN POPULATION =
321647 90% PERCENTILE POPULATION = 998939 STANDARD DEVIATION = 645852.0 COEF. OF VARIATION =
1.216 NO.
SITE NAME POPULATION NO.
SITE NAME POPULATION NO.
SITE NAME POPULATION 1
SUNDESERT 87 38 HAINE YANKEE 197000 75 SHEARON HARRIS 495900 HILLSTONE 496143 2
PEBBLE SPRINGS 4752 39 TROJAN 197480 76 4
3 PALO VERDE 20039 40 HAVEN 208201 77 COO K 522000 4
CRYSTAL RIVER 32055 41 VERMONT Y ANKEE 211630 78 SURRY 524100 5
SOUTil TEXAS 40950 42 LASALLE 215680 79 NEW ENGLAND 563343 6
BIC ROCK POINT 46538 43 PALISADES 216535 80 DRESDEN 568123 7
COOPER 58916 44 NALLAH 218551 81 FORT CALHOUN 589809 8
WOLF CREEK 61905 45
-DUANE ARNOLD 232995 82 PERKINS 622997 9
COMANCHE PEAK 65049 46 KEWAUNEE 245806 83
'SUSQUEHANNA 631467 10 ARKANSAS 76582 47 PRAIRIE ISLAND 264432 84
~
JAMESPORT 655123 11 HATCH 81252 48 BROWNS FERRY 265532 85 DAVIS BESSE 672000 12 CRAND CULF 90049 49 HONTICELLO 271182 86 PERRY 703553-13 IIUMBOLDT B AY 90330 50 STERLINC 275717 87 CATAWBA 707512 14 WPPSS 2 92185 51 VOCTLE 280137 88 HCCUIRE 805535 15 WPPSS 164 98886 52 CREENE COUNTY 288026 89 PEACH BOTTOM 830276 36 YELLOW CREEK 104404 53 YANKEE ROWE 303271 90 CINNA 870591 17 BRUNSWICK 108479 54 PHIPPS BEND 308144 91 PILCRIM 883583 18 DIABLO CANYON 114014 55 NEW NAVEN 309178 92 SALEM 893626 19 BELLEFONTE 114998 56 CLINTON 334115 93 HOPE CREEK 893626 20 FARLEY 119394 57 CVTR 360589 94 PIQUA 895367 DOUCLAS POINT 900652 21 SAINT LUCIE 120843 58 BRAIDWOOD 360694 95 22, C A I.L AW /.Y 122389 59 OCONEE 363543 96 RANCHO SECO 907789 23 WPPSS 3&S 124551 60 REVER REND 371036 97 TURKEY POINT 909916 24 P ATil FI N DE R 135451 61 SUMMER 378538 98 WATERFORD 957223 25 II A RT S VI L LE 135984 62 SAN ONOFRE 408362 99 THREE MILE ISLAND 995200 26 WOOD 138451 63 QUAD-CITIES 415500 100 SEABROOK 1003843 27 LACROSSE 143321 64 SEQUOYAH
- 432375 101 ZIHMER 1052883 28 SKACIT 151774 65 FORT ST. VRAIN 434802 102 ELK RIVER 1202027 29 NORTH ANNA 152432 66 OYSTER CREEK 451606 103 ZION 1262593 30 TY RO N E 153801 67 FORKED RIVER 451606 104 SHIPPINCPORT 1677889 31 WATTS BAR 161537 68 BONUS 453000 105 BEAV8R' VALLEY 1700000 32 POINT BEACH 187086 69 BYRON 455409 106 SHOREHAM 1760382 33 CALVERT CLIFFS 188755 70 MARBLE HILL 457928 107 HADDAM NECK 1763975 34 ROBINSON 192140 71 BLACK FOX 459832 108 BAILLY 2200000 35 HINE MILE POINT 195143 72 HIDLAND 470000 109 FERHI 2371808 36 FIT Z P AT RI CK 195143 73 CHEROKEE 475129 110 LIHERICK 3836244 37 CARROLL COUNTY 196357 74 ERIE 483519 111 INDIAN POINT 3984844
-~s
.,m
TABLE 3 P0pulation Statistics Between 0 and 50 Miles POPULATION STATISTICS-1979 REVISION.
5/79 BASED Oil THE YEAR 1970 POPULATION STATISTICS WITilIN 0-50 IIILES TOTAL NUHBER O F SITES =,111 HINIHUll POPULATION =
-7784-MAXIHUM Pn*"I.ATION=17471479 HEAN POPULATION = 1705750 MEDIN.".oruLATION=
948747 90Z PERCENTILE POPULATION = 4085400 STANDARD DEVIATION =2196315.2 COEF. OF VARIATION =
1.287 NO.
SITE'NAME POPULATION f*0 SITE NAME POPULATION NO.
SITE NAME POPULATION 1
SUNDESERT 7784 38 SEQUOYAll 659015 75 SUSQUEHANNA 1537373 2
PEBBLE SPRINGS 74814 39 CVTR 661462 76 YANKEE ROWE 1538765 3
-IlUH80LDT BAY 100728 40 PHIPPS BEND 691304 77 SURRY 1550000
~4 BIC ROCK POINT 128631 41 FORT CALHOUN 711117 78 PIQUA 1654093.
5 ARKANSAS 150464 42 SUMMER 724009 79 TURKEY POINT 1660498 6
WOLF CREEK 165677 43 OCONEE 73029 1 80 ZIMMER 1786790 7
CRYSTAL RIVER 169908 44 CARROLL COUNTY 733928 81 NEW ENGLAND-1862933 8
COOPER 171895 45 CLINTON 768171 82 THREE MILE ISLAND 1868000 9
-BRUNSHICK 174066 46 C0HANCHE PEAK 783124 83 MONTICELLO 1956232 10 WPPSS 1&4 181928 47 NORTH ANNA 827109 84 DAVIS BESSE 2052000 PRAIRIE ISLAND 2057725 11 WPPSS 2 184296 48 NINE MILE POINT 843775 85 12 SOUTH TEXAS 196206 49 FITZPATRICK 843725 86 ELK RIVER 2101115 t
13 DIABLO CANYON 209444 50 BELLEFONTE 845838 87 CALVERT CLIFFS 2305635 14 PATHFIHDER 242751 51 II ARTS VILLE 869776 88 ERIE 2411857 15 HATCH 251612 52 BY RO N 881721 89 PERRY 2583218 16 CRAND CULF 269314 53 LASALLE 918803 90 MILLSTONE 2591658 17 CALLAWAY 299254 54 NEP HAVEN 921367 91 DOUCLAS POINT 3167529 18 II A LL AN 307945 55 HAVEN 927246 92 JAMESPORT 3173531 19 SAINT LUCIE 318784 56 W JO D 970248 93 IIADDAM NECK 3267732 20 FARLEY 320667 57 PALISADES 984252 94 OYSTER CREEK 3290000 21 LACROSSE 321073 58
. BONUS 999000 95 FORKED RIVER 3290000 22 PALO VERDE 328088 59 HIDLAND 1000000 96 SAN ONOFRE' 3572478 23 Y EI.L OW CREEK 344716 60 SilEARON H ARRIS 1062200 97 SEABROOK 3605493 24 WPPSS 3&5 345935 61 COOK 1120000 98 SilIP P INC PO RT 3735300 25 SKACIT 366247 62 TROJAN 1146188 99 BEAVER VALLEY 3900000 26 TYRONE 372980 63 VERil0NT Y ANKEE 1149200 100 BRAIDWOOD 4088663 27 VOCTLE 456631 64 STERLING 1154607 101 PEACll 80TTOM 4121297 486000 65 CINNA 1215870 102 FILCRIM 4234545 28 HAINE Y ANKEE 29 ROBINSON 530817 66 HARBLE IIILL 1245001 103 SALEM 4773288 30 DUANE ARNOLD 552745 67 CATAWBA 1245504 104 I;0PE CREEK 4773288 31 POINT 8EACH 564251 68 CHEROKEE 1308327 105 SHOREHAM 4940868 32 KEWAUNEE 574631 69 HCCUIRE 1380228-10f FERMI 5446957 l-33 QUAD-CITIES 601843 70 RANCHO SECO 1381581 107 DRESDEN 6305057 34 BROWNS FERRY 625608 71 GREENE COUNTY 1383978 108 BAILLY 6747815 35 RIVER BEND 627983 72 FORT ST. VRAIN 1396284 109 LIMERICK 7036199
?
36 BLACK FOX 641797 73 WATERFORD 1479345 110 ZION 7083759 1
17 WATTS BAR 657836 74 PERKINS 1506152 111 INDIAN POINT 17471479
~
. TABLE 4 SITES WITH HIGHEST SECTOR POPULATIONS 0
Population in Highest 221/2 Sector (s)
A.
Based on 1970 census data at 10 miles 1.
Zion 65,000; 43,000; 41,000 2.
Millstone 39,000 3.
Duane Arnold 38,000 4.
Three Mile Island 35,000 5.
Indian Point 32,000 6.
Trojan 32,000 7.
Beaver Valley 31,000; 31,000 8.
Indian Point 30,000; 30,000 B.
Based on 1970 census data at 30 miles 1.
Indian Point 1,500,000; 820,000 2.
Limerick 1,300,000; 950,000 3.
Bailly 900,000 4.
Fermi 800,000; 770,000 5.
Waterfo rd 700,000 C.
Based on 1970 census data at 50 miles 1.
Indian Point 8,000,000; 2,900,000; 2,300,000 2.
Dresden 3,300,000 3.
Bailly 3,200,000 4.
Zion 3,200,000 5.
Salem 2,700,000 6.-
Sho reham 2,100,000 7.
Fermi 2,100,000 A!l
REACTOR ACCIDENT RISK PARAMETERS The accident risk to the public posed by a reactor at a particular site can be analyzed by carefully considering the design and operating characteristics of the reactor plant, the local meteomlogy, the population distribution around the plant, and the various measures such as sheltering or evacuation which could be taken to reduce the effect of a reactor accident on the public.
Ideally, thi's analysis should be plant and site specific.
Experience has already shown that plant design and operating characteristics are not so standardized that it is sufficient to analyze any one reactor, or any one type of reactor, or even any one reactor plant designed by a single supplier.
The estimated probabilities and scenarios of reactor accidents are so sensitive to differences in details of component reli-ability design and procedures, including human errors, that apparently similar plants can be substantially different.
The same nee'd for plant specific analysis holds true for the siting aspects of plants, i.e., the ireteorology and especially the demography.
Since there exists no exhaustive risk analysis of the Indian Point plants, the following analyses will deal separately with the siting and then the design aspects of the Indian Point plants comparing what we do know of them to similar risk analyses of other U. S. plants.
Understanding the overall acciderit risk of a nuclear power plant or comparison of the risk posed by it to"'that posed by any other plant requires consideration of the siting as well as the design and operating characteristics of the plant.
t
, 1 SITE ASPECTS
'The Reactor Safety Study _(WASH-1400), subject, to be sure, to large uncertainties, provides a basic accident risk model which can be 'used to.
assess the potential accident risk of a plant, at least in comparison to other plants.
The model was developed in the detailed review of only two plants, the Surry pressurized water reactor (PWR) and the Peach Bottom boiling water reactor (BWR). The Indian Point Unit 2 and 3 reactors are PWRs, furnished by the same nuclear steam system supplier 1
(Westinghouse), but of a larger size and later vintage. To compare reactor sites to one another, the Surry.PWR is used as a benchmark and, through the faoility of calculation, is moved from site to site calculating the overall risk for four principal risk measures: early fatalities; I
early (radiation) illnesses; latent cancer fatalities; and public property damage costs.
If the power of the benchmark reactor is held constant, then this set of calculations provides a good comparative measure of one site to another.
The staff has perfonned a set of these benchmar,k calculations using the Surry benchmark reactor with its power increased to 3025 MWT, the rating i
of Indian Point 3.
In general, the risk a reactor poses is proportional
' to its power level.
Six sites were analyzed for this comparison.
Four, Indian Point, Zion, limerick and Fenni, represent sites of relatively high population.
One, Palisades, represents what the staff believes is a typical or average population distribution.
The last, Diablo Canyon, represents a remote site, thatis', one with relatively low population
. density. The results of the analyses of the enlarged Surry plant at these six _ sites are shown in Figures 1 through 4 for the four measures of risk..
,n
--.,nn.
4
. The results shown in these figures are the complementary cumulative distribution-functions (CCDF)* which are the variation of the conse-quences of a reactor accident per year with their associated probability of occurrence.
The estimated risk of accidents for a given reactor, thc product of probabili' ties and consequences, is the area under the curve.
On Figures 1, 2, and 3 are listed the key assumptions about public protective action, namely that people within a 10 mile radius of the j
plant suffer the entire cloud exposure and then four hours of ground exposure before they are evacuated; people outside the 10 mile radius receive the entire cloud exposure and a subsequent seven day ground exposure assuming nonnal indoor and outdoor activity.
Before studying the curves consider for a moment the range of consequences that can be caused by a nuclear plant accident. For severe consequences, i
substantial amounts of radioactive material must be spread out over the surrounding area. The forces ejecting the material and the local meteorology will control how much gets out and how far it will reach.
The areas closest to the reactor will stand to receive the highest doses and those farther away, less. The Reactor Safety Study analysis showed that for severe accident releases, only those people within about 10 miles are exposed to fatal doses, beginning at about 300 Rem. Thus, the population within 10 miles.of a site will be significant to the early fatality risk for that site; the population beyond 10 miles will not.
This was a principal
- The CCDF shows the probability that a consequence will be equalled or exceeded. Appendix A discusses how a CCDF is constructed.
For further discussion of the consequence model used in these calculations, please refer to Overview of the Reactor Safety Study Consequence Model (NUREG-0340) and Appendix VI of the Reactor Safety Study (WASH-1400).
FIGURE 1 - EARLY FATALITY RISK FOR DIFFERENT SITES 10'4 i
ii6 inn i
a i iim a
iiisin i
i i i tiu i
iitag 1.
I.P.
- 2. ZION
- 3. LIMERICK 4.
FERMI
- 5 5.
PALISADES 5
- 6. DIABLOCANYON3 g
x N
T 10-6 e
2 u
~
5 6
=
U 3
U 10- -
=
- - --- 8
_=
4 2
3 10'9 0
I 3
4 5
10 10 10 10 10 10 X, EARLY FATALITIES (SUPPORTIVE TREATMENT)
NOTE: THERE ARE LARGE UNCERTAINTIES WITH THE ABSOLUTE VALUES PRESENTED IN THIS FIGURE I
ASSUMPTIONS: 1 SURRY DESIGN.
2 I.P. UNIT 3 POWER LEVEL (3025 MWT).
WITHIN 10 MILES - ENTIRE CLOUD EXPOSURE + 4 HOURS GROUND EXPOSURE l
NO SHIELDING BEYOND 10 MILES - ENTIRE CLOUD EXPOSURE + 7 DAY GROUND EXPOSURE SHIELDING BASED ON NORMAL ACTIVITY.
- 4) WIND ROSE WEIGHTED 1970 CENSUS POPULATION DISTRIBUTION.
l
- 5) IDENTICAL 91 WEATHER SEQUENCES FOR ALL SITES.
i
,10 -
> FIGURE 2 - EARLY ILLNESS RISK FOR DIFFERENT SITES j
10-4 i
- 116u a
i aaliu i
i iileu i
e i iin n i
i i64g
- 1. I.P.
2.
ZION
- 3. LIMERICK 4.
FERMI l
10-5 5.
PALISADES E
6.
DIABLO CANYON -
J 2
4 x
N ce 5
10-6 m
f h
6 m
a g
a.
_,. 10-7
=
=
N
=
2
~
3 10' 5
N 2
3 6
5 10-'
'I I
2 3
4 5
6 10 10 10 10 10 10 X, EARLY ILLNESS NOTE: THERE ARE LARGE UNCERTAINTIES WITH THE ABSOLUTE VALUES PRESENTED IN THIS FIGURE ASSUMPTIONS: 1) SURRY DESIGN.
- 2) I.P. UNIT 3 POWER LEVEL (3025 MWT).
- 3) WITHIN 10 MILES - ENTIRE CLOUD EXPOSURE + 4 HOURS GROUND EXPOSURE NO SHIELDING BEYOND ltMILES - ENTIRE CLOUD EXPOSURE + 7 DAY GROUND EXPOSURE SHIELDING BASED ON NORMAL ACTIVITY.
- 4) WIND ROSE WEIGHTED 1970 CENSUS POPULATION DISTRIBUTION.
- 5) IDENTICAL 91 WEATHER SEQUENCES FOR ALL SITES.
FIGURE 3 - LATENT CANCER RISK (ANNUAL) FOR DIFFERENT SITES 10~4 i
i i iiii, i i iiiii i
i i, i iii
,,,iii, i
i i i ig
- 1. I.P.
- 2. ZION 3
LIMERICK 4
FERMI 5.
PALISADES 10-5 6.
DIABLO CANYON :
x Ai 6
5
= 10-6 g
a g
a.
U
- -i 10-7~
~
E g
g g
4 b
10-8
)
5 4
A 10-9 I~
I '
0 I
2 3
4 5
10 10 10 10 10 10 X LATENT' CANCERS / YEAR *
- TOTAL LATENT CANCERS WOULD BE 30 TIMES HIGHER NOTE: THERE ARE LARGE UNCERTAIN 5[ES WITH THE ABSOLUTE VALUES PRESENTED IN THIS ASSUMPTIONS: 1)SURRYDESIGN.
- 2) I.P. UNIT.3 POWER LEVEL (3025 MWT).
- 3) WITHIN 10 MILES - ENTIRE CLOUD EXPOSURE + 4 HOURS GROUND EXPOSURE NO SHIELDING BEYOND'10 MILES - ENTIRE CLOUD EXPOSURE + 7 DAY GROUND EXPOSURE l
SHIELDING BASED ON NORMAL ACTIVITY.
I
- 4) WIND ROSE WEIGHTED 1970 CENSUS POPULATION DISTRIBUTION.
l
- 5) IDENTICAL 91 WEATHER SEQUENCES FOR ALL SITES.
FIGURE 4 - PROPERTY DAMAGE RISK FOR DIFFERENT SITES 10'4
,,,,,n i
,,,i,,.
i
,,,,n
,,,,,n
,,,,e Z
- 1. I.P.
2.
ZION
- 3. LIMERICK 4
FERMI 10-5 6.
DIABLO CANYON -
5.
PALISADES M
Ai m
10-6
\\
,c
!3 3
\\
5 5
4 m
g 2
5 U1 10-7_
(
1 g
E I
2 10' b
(6 (3) 3 4
5 s
10'9 10 10'7 10 10 10 10 6
8 9
10 X, DOLLARS TOTAL PROPERTY DAMAGE *
- BASED ON 1974 DOLLARS NOTE: THERE ARE LARGE UNCERTAINTIES WITH THE ABSOLUTE VALUES PRESENTED IN THIS FIGURE ASSUMPT:0NS: 1 SURRY DESIGN 2
I.P. UNIT 3 POWER LEVEL (3025 MWT) 3 WIND ROSE WEIGHTED 1970 CENSUS POPULATION DISTRIBUTION
- 4) IDENTICAL 91 WEATHER SEQUENCES FOR A.LL SITES.
. reason for selecting 10 miles as the radius for emergency planning zones (see NUREG-0396, Planning Basis for the Development of State and Local Government Radiological Emergency Response Plans in Support of Light Water Nuclear Power Plants).
Radiation injuries, caused by doses of 50 Rem or more, can reach farther out in the event of a severe reactor accident, to the population as far as 50 miles away.
Therefore, the population up to that distance away is significant in estimating the early illness risk; the population beyond 50 miles is not.
The estimation of latent cancer fatalities includes even low exposures so populations as far away as 200 miles will signiff-l cantly influence the latent cancer risk estimate.
Thus, for the latent cancer risk, the differences between sites are relatively small since the populations of such large regions are frequently similar.
Figure 1 shows that the three sites with the highest local population density, Indian Point, Zion and Limerick, have essentially the same risk profile for early fatalities.
The other sites show progressively lower risks. As was discussed,'early fatality risk is dominated by the population within 10 miles of the plant, so the large population of New York City is not a factor here. The absolute values of these risk estimates are subject to large uncertainties but the range should be noted.
For low probability--high consequence events, thousands to tens of thousands of early deaths are estimated for most sites.
Early illnesses are defined as radiation exposures in excess of 50 Rem, whole body _for an individual.
These flinesses or injuries, shown in
. Figure 2, are dominated by the size of the population within a 50 mile radius. Thus, New York City is important to the risk of early illness for Indian Point.
Zion, Limerick and Fenni also have enough population in the 50 mile range to be comparable to Indian Point as shown by Figure 2.
Also for this aspect of risk, the typical Palisades site and the Diablo Canyon site are not very diffe'.ent from each other but are substantially lower than the others.
For the sites with higher population density, thousands to hundreds of thousands of early illnesses are projected for the lower probability events.
The latent cancer risk, as shown in Figure 3, is dominated by the population within about a 200 mile radius of the plant.
Because of this, the individual site risk curves for latent cancers reflect the character of the region.
Remember that Indian Point is outside New York City, Zion outside Chicago on the north shore, Limerick to the northwest of Philadelphia, and Fermi near Detroit.
Palisades is on the western side of the Michigan lower peninsula an. Giablo Canyon is on the California coast well above Santa Barbara. The latent cancer risk for these sites, and probably all other sites is approximately the same.
The number of latent cancer deaths projected is on the order of hundreds per year or thousands per. accident for the lower probability events (on the order of 10'9/yr).
Please note that the latent cancer risk is presented throughout this discussion as latent cancers per year, that is, the average number of cancer deaths tha,t would be expected to occur per year in the population
~.
- which was exposed to the accident.
The total number of latant cancer deaths associated with an accident would be 30 times higher, reflecting the calculated rate of cancer death continuing for a generation.
For further discussion of latent cancer risk see NUREG-0340 at page 30.
The curves for property damage are presented in Figure 4.
The model still calculates in 1974 dollars; the correction for inflation is probably about a factor of 1.5.
The flatness of the curve at the upper left indicates that any accident with substantial releases will cause damage
]
of many millions of dollars. The projected damage for low probability events reaches up into the range of tens of billions of dollars.
However',
the property damage here does not include damage to the plant.
The Three Mile Island accident, which did no offsite property damage, caused several hundred million dollars worth of damage to the plant and replacement power costs, analogous to interdiction costs, on the order of a billion dollars.
The property damage risk estir.: ate is directly proportional to population density. With the present property damage model (see NUREG-0340 at page 22) the population out to about 30 miles is significant.
However, the use of more strict interdiction and cleanup criteria, as may well be warranted, would make 'mpulations beyond that distance impo rtant.
The estimated overall probability of core melts. for the benchmark reactor (Surry) rebaselined* from WASH-1400 is about one chance out of twenty
- The Reactor Safety Study plants were "rebaselined" for all the analyses presented in this report in order to take into account peer group comments (e.g., the Lewis Committee) and to use better data and analytical tech-niques which are now available such as the MARCH and CORRAL codes.
F0rther discussion of this rebaselining is presented in Appendix B.
' thousand (5x10-5) per reactor year. The CCDF curves have been constructed to. display the probability vs. consequence relationship for those cases of core melt accidents where offsite hann is done.
Note that the majority of core melts are not estimated to do hann offsite.
For example, in Figure 1 the benchmark Surry reactor at the Indian Point site is predicted to cause one or more acute fatalities at a frequency of 3.2x10-6/yr.
This means that only 3.2x10-6 i-5x10-5 =.064 or less than 10 percent of the core melt accidents are predicted to give lethal doses offsite.
Conversely about 90 percent of the core meit accidents are not expected to produce lethal doses for that plant.
For other plants a larger or smaller fraction of core melt accidents may be expected to cause lethal doses offsite.
Our ability to predict how often core melt accidents occur is very limited.
However, we are cuite reasonably confident from the work so far that most core melt accidents will not give lethal doses offsite.
Only certain accident scenarios in the plant, those entailing core meltdown and gross containment failure, coincident with particularly adverse weather conditions, will result in lethal doses or severe offsite ground contamination (i.e., property damage).
However, those few core melt accidents that do give lethal doses are likely to do r-iver a signifi-cant area (out to a few miles downwind).
If even one person receives a lethal dose offsite, it is quite likely that one thousand will receive a lethal dose.
However, in no case are more than a few tens of thousands predicted to receive lethal doses. No combination of weather conditior.s, ineffectual emergency response and severe accident can be found at any 1 --
a
-a.
. probability that is realistically expected to give lethal doses to as many as one hundred thousand.
There are, of course, higher numbers of latent casualties predicted for such acciden%, as can be seen in Figure 3.
Consider the differences among the curves; the curves-have been constructed on logarithmic ~ scale,'whidh tends to: minimize'small. differences.
~
There are a few perspectives which the CCDFs should clearly provide.
For illustrative purposes consider Figure 1; Early Fatality Risk for Different Sites.
The probability axis shows the chance of equalling or exceeding a number of early fatalities per reactor year.
At 10 fatalities, the range of probabilities for the sites represents the variation between sites of the likelihood of having at least 10 people receive lethal doses. At this level, there is about a factor of 30 difference in probability between the Indian Point and Diablo Canyon sites. Thus, the CCDFs ~ show the variation in probability for given levels of consequences.
The CCDFs also give the range of consequences for a given probability level. At the one in one hundred million (10-8) probability level, one would expect the Diablo Canyon area population to suffer at least 400 fatalities whereas the number of fatalities estimated at Indian Point would be about 10,000 or more.
In addition to the probability and consequence perspective, the curves give a sense of the importance of the consequences and probabilities.
When the curves have a clear knee in them, that is they have an approxir,ately horizontal slope out to some level of consequences and then fall off
. sharply (see the Indian Point curve in Figure 1, the knee is at about the 4,000_ fatalities level) the most impor^. ant part of the curve is the horizontal portion where one would expect to have about an equal chance of suffering consequences up to about that " knee" level.
When the curve drops off; the uncertainties become very large and the importance of perceived differences should be minimized. When the curves do not have I
a clear knee, as in the case of Indian Point on Figure 2, the probabili-ties are dropping at about the same rate as the consequences are increasing.
This result leaves _ a question as to the limit _of how many consequences could be expected. That is, the low probability-high consequence range (bottom right of curve) is clearly contributing to the overall risk.
The risk curves in Figures 1-4 can be reduced to probability weighted l
values, or expected consequences and these can be termed the likelihood of the consequence occurring in a year. Table 5 presents these expected l
consequences.
The principal differences between the risks at these sites is seen to be ir, early fatalities and injuries.
The Indian Point
\\
site poses about 20 times more risk of early fatality than a typical j
site such as Palisades. With respect to early injuries, the Indian Point site is about 10 times more risky than Palisades.
The differences in other aspects of risk are not so great..
The risks of early fatalities and early illnesses for the Indian Point site alone where only public protective measures are changed are shown in Figures 5 and 6, respectively.
For the Indian Point site alone, the sensitivity of early fatalities and early illness to no evacuation at all until a day after the accident, to differences in evacuation radius,
19 TABLE 5 EXPECTED ANNUAL CONSEQUENCES.(RISK) FROM 6 SITES WITH THE SURRY REBASELINED PWR DESIGN Probability of Early Early Latent Property Consequence Oc-Fatalities Injuries Cancer /Yr*
Damage $**
Site currence oer vr Diablo Canyon 1.6x10-5 2.5x10-4' l.8x10-4 1290 i
Palisades 2.9x10-4 1.2x10-3 2.7x10-4 2670 Fenni 9.2x10-4 6.3x10-3 3.6x10-4 4780 Limerick 3.5x10-3 1.1 x10-2 4.7x10-4 6980 Zion 4.7x10-3 1.2x10-2 4.3x10-4 6030
)
Indian Point 6.1x10-3 1.5x10-2 5.4x10-4 9550
- Total Latent Cancers Would Be 30 Times Higher
- Based on 1974 Dollars NOTE: THERE ARE LARGE UNCERTAINTIES WITH THE ABSOLUTE VALUES PRESENTED IN THIS TABLE.
ASSUMPTIONS:
1.
SURRY DESIGN.
2.
I.P. UNIT 3 POWER LEVEL (3025 MWT).
3.
WITHIN 10 MILES - ENTIRE CLOUD EXPOSURE + 4 HOURS GROUND EXPOSURE NO SHIELDING BEYOND 10 MILES - ENTIRE CLOUD EXPOSURE + 7 DAY GROUND EXPOSURE SHIELDING BASED ON NORMAL ACTIVITY.
4.
WIND ROSE WEIGHTED 1970 CENSUS POPULATION DISTRIBUTION.
5.
IDENTICAL 91 WEATHER SEQUENCES FOR ALL SITES.
i FIGURE 5 - EARLY FATALITY RISK AT INDIAN POINT FOR VARIOUS PUBLIC PROTECTION MEASURES 10~4 i
,i,,o
,,,in
,,,,m
,,,,,q
,ii,,c 1.
10mileevacuation3
~
2, 25 mile evacuation -
3.
50 mile evacuation -
4.
no evacuation for
~
1 day j
15.
sheltering 10-5 x
/u N
(1,2,3,5
~
~
10-6
\\
=
se
=
o x
g a.
\\
. _. - - $ l a
g g
- 10~ -
Z E
--10 0
1 2
3 4
5 10 10 10 10 10 10 X. EARLY FATALITIES (SUPPORTIVE TREATMENT) 4 t
NOTE: THERE ARE LARGE UNCERTAINTIES WITH THE ABSOLUTE VALUES PRESENTED IN THIS FIGURE ASSUMPTIONS: 1)SURRYDESIGN.
- 2) I.P. UNIT 3 POWER LEVEL (3025 MWT).
- 3) WIND RQSE WEIGHTED 1970 CENSUS ~ POPULATION DISTRIBUTION
- 4) INDIAN POINT SITE (POPULATION AND METEOROLOGY)
EVACUATION SCENARIOS - ENTIRE CLOUD EXPOSURE + EITHER 4 HOURS GROUND EXPOSURE, NO SHIELDING WITHIN GIVEN DISTANCE; OR 7 DAYS GROUND EXPOSURE, NORMAL SHIELDING BEYOND GIVEN DISTANCE NO EVACUATION
- ENTIRE CLOUD EXPOSURE + 1 BAY GROUND EXPOSURE, NCRMAL SHIELDING l
SHELTERING
- ENTIRE CLOUD EXPOSURE + 1 DAY GROUND EXPOSURE, SHIELDING ASSUMES BRICK HOUSE WITH NO BASEMENT.
FIGURE 6 - EARLY ILLNESS RISK AT INDIAN POINT FOR VARIOUS PUBL l
10-4 i3,,,o
,,,in i
i..m 3,3,,q
,,,,,g l.
10 mile evacuation 2.
25 mile evacuation 3.
50 mile e'vacuation -
4.
no evacuation for 1 day 10-5 5.
shelterino
.J
.g 10-6 b
,E a
E
(
e
. -- - - $ -10' ~
r- -~---
E 3
5 8
)
E 5
10-8_
M 1
3 2
-10~9-
' ' ' ' ' - ---l 1
2 ~
3 4
6 6
10
'10 10 10 10 10 X,-EARLY ILLhESS NOTE: THERE ARE LARGE UNCERTAINTIES WITS THE ABSOLUTE VALUES PRESENTED IN THIS FIGURE
~
ASSUMPTIONS: 1 SURRY DESIGN 2 'I.P. UNIT 3 POWER LEVEL (3025 MWT) 3 WIND ROSE WEIGHTED 1970 CENSUS POPULATION DISTRIBUTION
- 4) INDIAN POINT SITE (POPULATION AND METEOROLOGY).
EVACUATION SCENARIOS - ENTIRE CLOUD EXPOSURE + EITHER 4 HOURS GROUND EXPOSURE, NO SHIELDING WITHIN GIVEN DISTANCE; OR 7 DAYS GROUND EXPOSURE, NORMAL SHIELDING BEYOND GIVEN DISTANCE NO EVACUATION
- ENTIRE CLOUD EXPOSURE + 1 DAY GROUND EXPOSURE, NORMAL SHIELDING SHELTERING
- ENTIRE CLOUD EXPOSURE + 1 DAY GROUND EXPOSURE, SHIELDING
-ASSUMES BRICK HOUSE WITH NO BASEMENT.
. namely,'10, 25 and 50 miles and sheltering were analyzed.
For Indian Point, this last would include New York City itself.
In Figure 5 for early fatalities, only two curves are s'hown, one for no evacuation for one day and a second curve representing a range of the public protection options since their differences are too small to distinguish. All evacuations are assumed to include direct exposure of the people to the cloud and then four hours of ground exposure while evacuating. Obviously, if one assumed that the evacuees could leave before suffering less or evcn any cloud and ground exposure, the risk profile would be drastically lowered.
Since early fatalities are dominated by the population within the first 10 miles, evacuating beyond that range produces little reduction in early fatalities.
The early illnesses that could be suffered around the Indian Point site with varying public protection strategies is shown in Figure 6.
The lowest risk is with a 50 mile evacuation.
The alternative of sheltering for a period of 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> and then evacuating selectively appears to provide nearly the same risk reduction for the Indian Point environs.
The other alternatives depicted do not appear to offer as much benefit for the low probability-high consequence events.
l
~
t.
THE EFFECT OF DESIGN'0M RISK AT!INDI'AN POINT The extensive use of quantitative risk assessment for U. S. power reactors began with the Reactor Safety Study (RSS), WASH-1400, which studied a 3-loop Westinghouse PWR, Surry, and a General Electric BWR, Peach Bottom.
Since the Reactor Safety Study, other reactor risk assessments of somewhat lesser depth have~ been made.
For example, the NRC staff has been pursuing the Reactor Safety Study Methodology Application Program.
This program is considering. 63ur reactors: Sequoyah, a Westinghouse 4-loop PWR with ice condenser containment; Oconee, a Babcock-Wilcox 2-loop PWR with dry containment; Calvert Cliffs, a Combustion Engineering 2-loop PWh with dry containment; and Grand Gulf, a General Electric BWR with Mark III containment.
These designs are being reviewed with application of the f
Reactor Safety Study event and fault tree techniques.
The reports on these studies will not be complete until later this year but some of the preliminary results are available to the staff.
The staff recently began a new program, the Interim Reliability Evaluation P rogram.
The first plant covered in this program is Crystal River 3, a Babcock and Wilcox 2-loop PWR with dry containment.
The initial report
. on this study is now in peer review, and its preliminary results are l
available to.the staff. Also available 63r comparison are the results i
j of the German reactor risk study of the Biblis B reactor.
l The staff ~used the infonnation gained from these studies to guide a
- short term risk evaluation of the Indian Point 2 and 3 plants.
This
e..
I evaluation relies heavily on the judgement of the reviewer with respect to the accident sequences being considered and to the parts of the plants involved. The approach was to consider the key accident sequences which involve core meltdown
- or containment failure modes that would be expected to dominate risk. The Indian Point plants were briefly reviewed against these scenarios and their designs were surveyed for single point vulnerabilities such as single manual valves or human errors which can trigger or control a significant accident sequence.
Particular attention was given to common interactions which could cut across more than one syste.a or be caused by a single initiating event.
Rough estimates were made of the likelihood and consequences of various sequences using the data and release characteristics of previous studies, particularly the Reactor Safety Study and its follow-on work, the Methodology Application Program.
Prior risk studies showed that a handful of accident scenarios would most likely define and dominate a reasonably complete spectrum of l
core melt accident scenarios for the PWR design.
Table 6 lists the j
accident scenarios which were so considered and which were among-those quantitatively estimated for the Indian Point 2 and 3 study. We found no risk significant differences between the Indian Point 2 and 3 designs.
j An estimate of the overall probability of severe core damage or core l
melt as made for Indian Point 2 and 3 as of December 1979.
Then the estimate was revised to reflect those changes that were made or committed to in early 1980.
This very preliminary estimate for Indian Point indicates an initial probability of severe core damage of about 3x10-5
- Here, as in WASH-1400, the tenns core meltdown and severe core damage are used interchangeably. The analysis presumes procession to core melting once severe damage is suffered.
. TABLE 6 DOMINANT ACCIDENT SEQUENCES Sequence Code Offsite Consequences Accident Scenario From WASH-140,0, Expected LOCA and failure of ECCS AD Low to modest in injection mode SD SD LOCA and failure of ECCS AH in recirculation mode SH SfH Transient and loss of feedwater TMLX or serious failure and no feed TMKX and bleed on primary side (X)
V LOCA and loss of containment AG Intemediate heat removal with subsequent SG interactions with ECCS SG LOCA and failure of ECCS and AHF High containment ESFs in recircu-S HF 1ation phase due to comon SfHF cause LOCA and coupled damage to Event V ECCS and potential bypass of containment Transient involving loss of TMLB' all AC power (or possibly V
DC) and failure of auxiliary feedwater
)
o
4 per year.
The improvements made or committed to this year are estimated 5
to reduce that probability by a factor of three to about 1x10 per year.
For comparison, Table 7 presents the estimated probability of severe core damage for the Indian Point reactors along with similar estimates from the Reactor Safety Study and other studies mentioned previously.
The overall effect of the Indian Point improvements is estimated to be a three-fold reduction in.the probability of severe core damage if these improvements are successfully implemented. As it turns out, it is not important to this overall analysis to detennine whether each of the committed changes has been made and when. The changes committed to are clearly beneficial in reducing risk but it is questionable whether the factor of improvement, three, is statistically significant.
The probabi-lities of severe core damage listed in Table 7 are subject to at least a factor of 5 uncertainty in either direction due to uncertainties in the data upon which all this analysis is based. Therefore', one should be very careful about attaching significance to differences in these estimates which are less than about one' order of magnitude.
The effect on risk at the Indian Point site u best seen by comparison of the CC0F's.
Figure 7 shows the early fatality risk curves for five different reactor designs, all at the Indian Point site, including the early fatality risk curves estimated for the Indian Point 2 reactor before the 1980 changes and after the 1980 changes.
Figures 8, 9 and 10 display the same comparisons for the other risk indicators, early injuries, latent fatalities ud property damage.
~
.- TABLE 7 - ESTIMATED PROBABILITY OF F9/ERE CORE DAMAGE REACTOR NAME TYPE PROBABILITY
- OF SEVERE CORE DAMAGE PER REACTOR-YEA.R SURRY 3-loop PWR 6x10-5 PEACH BOTTOM BWR(MarkI) 3x10-5 SEQUOYAH 4-loop PWR (Ice '1ndenser) 4x10-5 0CONEE 2-loop PWR 2x10-4 CALVERT CLIFFS 2-loop PWR 2x10-4 CRYSTAL RIVER-3 2-loop PWR 3x10-4 BIBLIS 4-loop PWR 4x10-5 INDIAN POINT 4-loop PWR 1x10-5
- Reflects median values i
i 4
9
FIGURE 7 - EARLY FATALITY RISK FOR DIFFERENT DESIGNS 10-4 i
i i iim i
i i i iiii i
i i i eim i
e i iiim a
a i i ig 2
1 PEACH BOTTOM BWR REBASELINED
=
? SURRY PWR REBASELINED 3 SEQUOYAH ICE CONDENSER 4 INDIAN POINT BEFORE FIX 5 INDIAN POINT AFTER FIX 10-
=
~--
m N
p
~
w
. $. 6
^
w b
b A4 (5
2'
=
b D
N
.. 10-7' w
~ ~ ~ ~ -
g g
3
. _ _j o-8, E
E 1
2
--10-3
=
0 I
2 3
4 5
10 10 10 10 10 10 l
X, EARLY FATALITIES (SUPPORTIVE TREATMENT)
NOTE: THERE ARE LARGE UNCERTAINTIES WITH THE ABSOLUTE VALUES PRESENTED IN THIS FIGURE ASSUMPTIONS: 1)INDIANPOINTSITE METEOROLOGY - 91 WEATHER SEQUENCES WIND ROSE WEIGHTED 1970 CENSUS POPULATION DISTRIBUTION UNIT 3 POWER LEVEL (3025 MWT)
- 2) WITHIN 10 MILES - ENTIRE CLOUD EXPOSURE + 4 HOURS GROUND EXPOSURE NO SHIELDING BEYOND 10 MILES - ENTIRE CLOUD EXPOSURE + 7 DAY GROUND EXPOSURE SHIELDING BASED ON NORMAL ACTIVITY
~ 29" FIGURE 8 - EARLY ILLNESS RISK FOR DIFFERENT DESIGNS 10~4 i
,,,iiii
,,, i i iii i,,,,,,,
i i i e iliti i
i i i t i tz 2
1 PEACH EDTTOM BWR REBASELINED 2 SURRY PWR REBASELINED 3 SEQUOYAH ICE CONDENSER 4 INDIAN POINT BEFORE FIX 5 INDIAN POINT AFTER FIX 4
. _.._. 10'$
g x g
___.. N.3n-6 s
5
.. _ _ _ _U _..10'7-3 E
~
g t
2 1
3 8
\\
1.
=
10 E
\\
E
(
10 2 3
4 5
10) 10 10 10 10 10 X, EARLY ILLNESS NOTE: THERE ARE LARGE UNCERTATNTIES WITH THE ABSOLUTE VALUES PRESENTED IN THIS FIGURE ASSUMPTIONS: 1) INDIAN POINT SITE METEOROLOGY - 91 WEATHER SEQUENCES WIND ROSE WEIGHTED 1970 CENSUS POPULATION DISTRIBUTION UNIT 3 POWER LEVEL (3025 MWT)
- 2) WITHIN 10 MILES - ENTIRE CLOUD EXPOSURE + 4 HOURS GROUND EXPOSURE NO SHIELDING BEYOND 10 MILES - ENTIRE CLOUD EXPOSURE- + 7 DAY GROUND EXPOSURE SHIELDING BASED ON NORMAL ACTIVITY
30-FIGURE 9 - LATENT CANCER RISK FOR DIFFERENT DESIGNS 10-4 i
i isiii i i i i iiii i
i i e i iii; i
i i i i n sil i
i i s i i:
2 1 PEACH BOTTOM BWR REBASELINED 2 SURRY PWR REBASELINED 3 SEQUOYAH ICE CONDENSER M
3 4 INDIAN POINT BEFORE FIX s
.5 INDIAN POINT AFTER FIX
. -.... - ~10 '
C n
g Al 4
g W
- -.a-30-6 N
E R
=_
y e
g o.
5 N
-g j o-L q.
g 8
a-10.8_=
\\
)
d
=-
10 0
I 2
3 4
5 10 10 10 10 10 10 X, LATENT CANCERS / YEAR *
- TOTAL LATENT CANCERS WOULD BE 30 TIMES HIGHER NOTE: THERE ARE LARGE UNCERTAINTIES WITH THE ABSOLUTE VA.UES PRESENTED IN THIS FIGURE.
ASSUMPTIONS: 1) INDIAN POINT SITE METEOROLOGY - 91 WEATHER SEQUENCES WIND ROSE WEIGHTED 1970 CENSUS POPULATION DISThfBUTION UNIT 3 POWER LEVEL (3025 MWT)
- 2) WITHIN 10 MILES - ENTIRE CLOUD EXPOSURE + 4 HOURS GROUND EXPOSURE NO SHIELDING i
BEYOND 10 MILES - ENTIRE CLOUD EXPOSURE + 7 DAY GROUND EXPOSURE SHIELDING BASED ON NORMAL ACTIVITY l
l
FIGURE 10 - PROPERTY DAMAGE RISK FOR DIFFERENT DESIGNS 39-4 1
i iisti i i i i iiii i
i iiniin i
i i i n sig i
i i,iig 1 PEACH BOTTOM BWR REBASELINED 2 SURRY PWR REBASELINED 1
3 3 SEQUOYAH ICE CONDENSER 4 INDIAN POINT BEFORE FIX 5 INDIAN POINT AFTER FIX N
t-10-5
~'
~
~
2 x
4 g
10-6
~~
c:
X N
y W
5 j
5 a-
\\
\\,
i
- - - } 3 n-7.
_ _.___ A g
+
8 1
g
-_..--10-8 E
E
''l 10 '
' ' 'h
'I 6
7 8
9 10 ll 10 10 10 10 10 10 X, DOLLARS TOTAL PROPERTY DAMAGE *
- BASED ON 1974 DOLLARS NOTE: THERE ARE LARGE UNCERTAINTIES WITH THE ABSOLUTE VALUES PRESENTED IN THIS FIGURE f
ASSUMPTIONS: 1) INDIAN POINT SITE METEOROLOGY - 91 WEATHER SEQUENCES WIND ROSE WEIGHTED 1970 CENSUS POPULATION DISTRIBUTION l' NIT 3 POWER LEVEL (3025 MWT) 4
-n-----
9 The reactor designs whose risk profiles are considered here include the two reactors considered in the Reactor Safety Study, Surry and Peach Bottom; the Sequoyah plant with its ice condenser and the two versions of the Indian Point design.
The risk profiles are presented only for these reactors and not the others listed in Table 7 because there was I
not time to do the others.
When considering the CC0Fs presented in Figures 7, 8, 9 and 10, it is important to keep the uncertainties in mind.
WASH-1400 assigned an uncertainty of plus or minus a factor of five to analysis such as this.
The Lewis Committee questioned that small an uncertainty. We believe it is prudent to consider that these curves have an uncertainty, plus or minus, of about a factor of 10 at the higher probabilities and perhaps as much as a factor of 100 at the lower probabilities.
Thus, one can attach significance to the range shown but not to modest differences between curves.
As indicated by the curves, the risk of the Indian Point reactors appears to be even lower compared to the other reactors than the ratio of their core damage probabilities would suggest.
Table 8 presents the expected annual consequences or the risk from these five different designs at the Indian Point site.
If one postulates that the Surry design is a typical reactor, then " Indian Point After Fix" poses about 30 times less risk of early fatalities, about 50 times less risk of early injuries, about 30 times less risk of latent cancers, and about 50 times less risk of
~
property damage. At this time, not enough is known about the overall l
l
i risk profile of all the individual plants in the U.S. to say what is typical or even what the range is.
The variation of the design and operation parameter done in this analysis was based on infonnation available, not on identifiable bounds.
l f
a 4
a e
9 e
=
y
>. - +
~+w
. TABLE 8 EXPECT ANNUAL CONSEQUENCES (RISK) FROM 5 LWR DESIGNS AT THE INDIAN POINT SITE Prob.'of Conse-Early Early latent Property e o r vr. Fatalities Injuries-Cancer /Yr*
Damage $**
Desia r
IP After Fix 2.2x10-4 2.7x10-4 1.6x10-5 jgg IP Before Fix 6.3x10-4 9.5x10-4 4.4x10-5 700 Surry Rebaselined 6.1x10-3 1.5x10-2 5.4x10-4 9550 Sequoyah Ice 2.7x10-3 2.2x10-2 1.2x10-3 14800 Condenser Peach Bottom BWR 1.7x10-2 3.1x10-2 1.1x10-3 13500 Rebaselined
- Total Latent Cancers Would Be 30 Times Higher
- Based on 1974 Dollars NOTE: THERE ARE LARGE UNCERTAINTIES WITH THE ABSOLUTE VALUES PRESENTED IN THIS TABLE.'
ASSUMPTIONS:
1.
INDIAN POINT SITE METEOROLOGY - 91 WEATHER SEQUENCES WIND ROSE WEIGHTED 1970 CENSUS POPULATION DISTRIBUTION UNIT 3 POWER LEVEL (3025 MWT) 2.
WIT:IN 10 MILES - ENTIRE CLOUD EXPOSURE + 4 HOURS GROUND EXPOSURE NO SHIELDING BEYOND 10 MILES - ENTIRE CLOUD EXPOSURE + 7 DAY GROUND EXPOSURE SHIELDING BASED ON NORMAL ACTIVITY l
THESENSITIVITY'0FRISKNO. VARIATIONS.'.I'! SITE,'PUBLIC: PROTECTION, AND DESIGN /0PERATING CHARACTERISTICS: :
4 In the preceeding sections the risk.'as considered for variation of three basic parameters, the reactor site, the public protection measures taken, and the different reactor plant design and operating character-istics.
For the first, a single reactor design, Surry, was placed at six different sites. The degree of uncertainty in this site comparison is act as great as for the design comparison because, although there are substantial uncertainties in the model, the sites differ only by two relatively well understood parameters, demography and meteorology.
The demography differences dominate the comparison.
The same degree of uncertainty exists for the public protection measure variation, since no evacuation logistics analysis is made here.
The model used for these analyses works just on the demography.
For the design variation there is much gi ater uncertainty. The compari-son of one plant to another involves different levels of study, different dominant accident scenarios, and the use of a great deal more judgment by the analyst.
Previous work by the staff in evaluating the reliability of auxiliary feedwater systems in many PWRs was done on a more consistent basis, where each plant received approximately the same depth and scope of analysis.
The results of that analysis showed reliability variations for that one important system from plant to plant ranging over two orders of magnitude, about as much as was shown here for site variation.
. Figure 11 was drawn to display the range of variation for the three parameters of this analysis.
On each of the four graphs shown in Figure ll, the solid lines show the bounds of variation when the same reactor was moved from site to site.
The long-short-long lines with shading in the first two graphs show the bounds for variation of public protective action options, all with the pessimistic (or realistic) exposure assumptions described previously.
The dashed lines on all four graphs show the range of variation of a few reactor. designs that were analyzed. We expect the full range of variation of risk due to design factors from the best to worst plant in the country to be broader than the small sample shown here.
Figure 11 suggests that the most significant parameter affecting risk is the design and operation of the plant.
The site is a significant variable more for early effects and the public protection options as shown here are the least significant.
m..
FIGURE 11'- RANGES OF~ RISK VARIATION 10
'4 i n disi 6 a
..i.asg
+64+]
=
i4,, in i i s i ii iisi i a a 4.iin
~
~
~
t o-I 10 5 N,,
g
{ - '-- ~ - "
W., %
' '
- 7t?*.**"
' g kd\\
~
\\
~
^
E 3
! ",,i 3e.
! b N)p'N
.\\
t M'
\\
E k
h.
\\
y T
V N
g
- 5..
g s
\\
\\
N L
\\
1 g
\\
~
I,
{h !y,
\\
k\\
3.s
-m 3
\\
I.
\\
5 j
sy;.
'7, } \\-
\\
~. n.
N
. g
\\
}.
.j l\\ _
\\
. \\
\\,
e i
s i,\\+,"
m^
4
,,.m
,,,i,
,,,i i,,te, f=F g-,
I t, '
2 3
5 0
10 a
10 10 10' 10 10 10 10 10' 10 10 8
l 3
3 8
- f. EARLY FATALIT!($ ($5ppWTIVE 14AMNfl I. EAALT ILLDE15 d
10**
10 i,,i,
,,,,,u
=,
,,,,,,,2 I'~,
-~~~-
10-5
_s
~
10-5 N
\\,
E b
i n
\\
n z
\\-
a s
E
\\\\
'I h, -s $ -
\\
.s h _
5 N
k N
\\
9 s
s s
s,
\\
g
\\
~
g, g
r N
g 3
E u-7
\\
\\
\\
\\
E u-'
s
\\
5 g
\\
3 l
E
\\
\\
i l
5-g
\\
\\
E 3
s-s
\\
6 g
g g
\\
s 3.s g
u.a i
\\
5
\\I
\\
5 5
\\
'E g
\\ :
g!
\\
s
' ~
3 g
r L,,,,
),,,,,,
3., j 3
,A,
,i,,n 10 10 10
. 10 10' 10 10 10 '
10 10' 10 10" 0
I 2
3 I
8 8
M X. LATENT CANCERS / ftAR* -
I. 00LUWt310TAL pit 0PDtTT QAftet *
'f0fAL LAftNT CAncet$ leULD DE 30 TMS N!GMER eg, *0 OR 1974 COLLAA$
NOTE 1. THE RANGES REPRESENT BEST ESTIMATES ON A COMPARATIVE BASIS. THERE ARE LARGE UNCERTAINTIES WITH THE ABSOLUTE VALUES PRESENTED IN THIS FIGURE.
- 2. PUBLIC PROTECTIVE MEASURES'HAD NO SIGNIFICANT IMPACT ON TOTAL LATENT CANCER OR ON TOTAL PROPERTY DAMAGE.
1 ESTIMATED RANGE OF CONSEQUENCES FOR VARIOUS DESIGNS CONSIDERED AT
____f INDIAN POINT SITE.
1 ESTIMATED: RANGE OF CONSEQUENCES FOR 6 SITES CONSIDERED WITH f SURRY DESIGN, Q"ECW l ESTIMATED RANGE OF CONSEQUENCES FOR VARIOUS PUBLIC PROTECTIVE weegysMG f MEASURES CONSIDERED AT INDIAN POINT SITE._
L.
. THE RISK OF'A 'IkDIb P0bT? REACTOR COMPARED TO'0THER REACTORS The preceeding sections examined the risk of the Indian Point site and the Indian Point reactor designs separately.
From those examinations it appears that the site is about an order of magnitude more risky than a typical site and the design about as much less risky than a typical design. There is much more certainty in our comparison of the relative site risks than there is in the comparison of the design risks.
It is reasonable to conclude that the two about cancel, that is, the overall risk of the Indian Point reactor is about the same as a typical reactor on a typical site. We recognize that such a comparison makes no explicit compensation for the Indian Point risk entailing notably higher consequences even if at lower probability than is typical.
It is not unusual in risk aversion to demand lower risk as the potential consequences increase -
as the stakes get higher.
Accordingly, one might argue that the probability should be more than a magnitude lower if the consequences can be a magnitude higher.
REDUCTION OF OPERATING POWER LEVEL Obviously, reactor accident risk can be essentially eliminated by shutting down the reactor. Reducing the operating power level car, reduce risk in two ways, by reducing the potential consequences of an accident and by reducing the probability of an accident occurring or running its course.
Reducing the operating power level of a reactor does not reduce the potential conse-quences proportionately until long after the power level reduction is enfo rced. A typical PWR core is divided into three sets of fuel assemblies.
One set is replaced at each refueling, so that each fuel assembly experiences
4.
three operating cycles in its period of use.
The accident risk posed by a reactor arises from the inventory of fission products which builds up in these fuel assemblies.
Based on the WASH-1400 analysis, about half that risk comes from iodine isotopes with half-lives of no more than eight days.
For these iodine isotopes, the equilibrium inventory level is proportional to power level, and is reached in about a month at that power. After about a month, then the iodine contribution to risk is going to be directly proportional to steady state power level.
The other half of the estimated accident risk is dominated by isotopes of elements such as tellurium, cesium and strontium, having fairly long half-lives, e.g., of years.
Some of these isotopes never reach an equilibrium level in the fuel as do the short-lived ones but continue to build up in proportion to both power level and the time spent at that level, in essence, in proportion to the number of fissions.
Therefo re, an operating power level reduction will not proportionately reduce the risk from these isotopes unless there is also a reduction in the fuel burnup allowed.
The reduction of operating power level can also have an effect on accident risk by reducing the fuel operating temperature levels and by reducing the amount of decay heat which must be removed after shutdown. At lower power levels the heat output of the fuel is lower.
Since the coolant tenperature remains essentially the same as at full power, the result is j
lower temperature of the fuel and much of the metal surrounding it.
The advantage of reduced fuel temperature in an accident is the fact that
)
\\
' the fuel has that much more capability of absorbing heat before it reaches severe damage temperature or melts. Thus, the core can tolerate longer periods without proper cooling before damage is done.
l Continued operation at reduced power level will also reduce the amount of decay heat generated after shutdown, in proportion to the degree of power reduction.
This, as well as lower fuel temperatures, increases the length of time the core can run without proper cooling before damage occurs.
With increased tolerance of poor core cooling, there is more time for corrective action by the operators in the event of an accident.
No quantitative analyses were perfomed to estimate the degree of risk reduction that can be achieved by reduction of the operating power level but, from the factors involved, it appears reasonable to say that risk would be reduced in proportion to the reduction in power level.
l
, SECTION 2.
SOCIAL AND ECONOMIC IMPACT CONSIDERATIONS l
EFFECTS OF AN INDIAN POINT STATION SHUTDOWN ON ELECTRICAL POWER RELIABILITY IN THE NEW YORK POWER POOL The New York Power Pool (NYPP) coordinates the generation and delivery of electric power for the State of New York.
Its members operate according to j
certain standards, including the requirements that NYPP members maintain an installed generating capacity reserve equal to 18 percent of maximum one hour net load. There are seven investor-owned and one state owned utility in the NYPP with a total capacity as of Summer 1979, of nearly 30,000 W.
Consolidated Edison represents about 31 percent (9400 W ) and the Power Authority of the State of New York (PASNY) about 22 percent (6700 W ) of the total capacity of the NYPP. The electric service area of CON ED consists of the five boroughs of New York City and a major part of Westchester County, an area of 600 square miles with over eight million customers.
PASNY does not have any geographically defined " service territory" but serves particular classes of customers in all parts of the State of New York.
Southeastern New York State is a summer peaking region.
CON ED's summer pea' load, in particular, is about 40 percent higher than its winter peak load mainly due to the widespread use of electric air conditioning.
The remainder of New York State is a winter peaking region.
The total NYPP System peaks in the summer.
. i According to a recent DOE analysis,M attached here as Appendix C, the fore-cast of the combined 1980 summer peak for CON ED and PASNY is 9403 % as shown in Table 1.
Total PASNY and CON ED capacity is approximately 16,000 N.
If Indian Point Units 2 and 3 are removed from the system and an 18 percent reserve margin is added to the forecast summer 1980 ~ CON ED-PASNY peak, there is still an apparent excess capacity of about 3000 N.
However, much of PASNY's capacity is not in the Southeastern New York area, but elsewhere in the State. Major PASNY facilities in Southeastern New York include Indian Point #3 and Astoria #6 with a combined megawatt rating of approximately 1740 N.
If u assumes that one-half of the projected summer peak demand for the PASNY system originates in the New York City ardak and if the location of PASNY's generating capacity is taken into account, then the reserve picture changes considerably as shown in Table 2.
It should be noted that of the total capacity of some 9300 N nearly 2000 Mw are combustion turbines which are generally not planned or designed for prolonged operation. Given the projected summer load for the Southeastern New York area, the shutdown of Indian Point #2 and #3 would result in insufficient capacity (by some 250 N) to maintain an 18 percent reserve. All of the reserve capacity disappears and energy would have to be imported from other parts of the NYPP if scheduled outages, sumer capacity reductions and historically experienced forced outages of some 1500 N are accounted for.
In addition, if the largest unit (Ravenswood #3 - 928 N) is lost, the_ DOE analysis concludes that the utilities would be forced to use all available capacity and interties to the maximum reasonable extent.
Ifletter to Edward J. Hanrahan from Richard Weiner, Director, Division of Power Supply and Reliability, Economic Regulatory Administration, DOE,
- May 15,1980.
2]LettertoHanrahan,op. cit.,
p.2, DOE states that PASNY's projected sumer 1980 peak load is 2503 N "of which less than half is in New York City and Westchester County areas".
s Table 1 Reserve Situation for the CON ED and PASNY Systems (Summer,1980)
(Mw)
CON.E0 PASNY TOTAL (1)
Sumer Peak,1980 6900 2503 9,403 (2)
Sumer Peak,1980
+ 18% reserve margin 8142 2953 11,095 (3)
Capacity with Indian Point 2, 3 9441 6740 16,181 (4)
Capacity without
,.775 14,367 5
Indian Point 2, 3 8592 (4) - (2) Apparent Excess
. Capacity 450 2S22 3,272 Table 2 Revised Reserve Situation for CON ED and PASNY Systems (Summer, 1980)
(MW)
Sumer Peak,1980 6900 1251 8,15i (2)
Sumer Peak,1980
+:18% reserve 8142 1475 9,617 (3)
Capacity without Indian Point 2,~
3 8592 775 9,367 (3).- (2) Excess Capacity
'450
-700
- 250 9
i L
I 4
l l
i 1
The bulk power transmission tie line capability above scheduled transfers is limited as shown in Table 3.
All but LILC0 is expected to have sufficient excess capacity in summer 1980 to transfer to the limit of the intertie.
LILC0 is expected to be able to supply an average of only 100 Mw. There also may be some contingency support through the submarine cable from Connecticut, but this would be limited to only 145 Mw.3/
i Table 3 Bulk Power Transmission Capability Above Scheduled Transfers (Mw)
SUMMER WINTER FROM 1980 80-81 Upper State New York 500 2200 Pennsylvania-New Jersey-Maryland Interconnection (PJM) 150 50 Long Island Lighting Co. (LILC0) 475 550 TOTAL 1125 2800 OTHER EFFECTS OF INDIAN POINT S!iUTDOWN Aside from reliability consideration, the coc.s to the service area of the CON ED and PASNY Systems of a shutdown of the Indian Point Station include expected increases in cost of service.
Indian Point provides electrical energy to the system at a cost in between hydroelectric and oil-fired generation.
These types of facilities along with the Fitzpatrick nuclear plant provide
_3./
Letter to Hanrahan, op. cit., p. 2.
. almost all of the power for the CON ED and PASNY system. The least expensive method of_ replacing power lost as a result of the shutdown of Indian Point station appears to be PASNY's hydro facilities as well as the purchase of hydro-generated power from the NYPP and Hydro Quebec if available. These facilities, however, are not in the Southeastern New York area, and the transmission facilities into that area are 1imited according to the DOE analysis.
Assuming that oil-generated power replaces the energy lost by shutting down Indian Point station, it is possible to calculate an upper bound to the economic costs of such an action.
If Indian Point station operated at its historic capacity factor of 60 percent, it would produce about 800 mil' ion kilowatt-hours per month. Approximately 1.4 million barrels of oil per month would be needed to produce the equivalent amount of oil-fired electricity.
At $31 per barrel this would amount to approximately $42 million per month in fuel costs without adjustment for ' differences in non-fuel operating costs and uranium fuel costs saved. The major impact would be the bill for oil, much of which would likely be imported. This, of course, assumes that none of the energy shortfall could be made available from non-oil generated power.
9 9
. SECTION 3.
SUMMARY
OF PUBLIC C04MENTS This section summarizes public comments that bear on interim operation.
Numbers in parentheses accompanying the comment summaries refer to the connent numbers assigned in SECY-80-168, which contains a full compilation of public comments on the Director of NRR's Indian Point decision received in response to the Commission solicitation of comments.
Considerations in the Director's decision that bear on interim operation are also summarized.
SAFETY ARGUMENTS Director's Decision The Director relies on two considerations in not ordering interim shutdown for the one to two-year period required to determine and install required additional design safety features:
First, several compensating features for the high population density already exist in the design of Indian Point 2 and 3.
These include:
1.
Containment we16.hannels and weld channel pressurization system.
2.
Containment penetration pressurization system.
3.
Isolation valve seal water system.
4.
Extra containment fan cooler capacity.
5.
Post-LOCA hydrogen control capability by both recombiner and purge.
6.
Third auxiliary feedwater pump, providing added assurance over a twice 100 percent capacity system.
7.
HEPA and charcoal filters for containment atmosphere cleanup.
8.
Confirmatory actuation signals +- swer operated valves which are not required to changa position.
9 9.
Extra margin in service water and component cooling water supply.
- 10. Redundant electrical heat tracing on borated systems.
Second, a number of extraordinary interim measures are to be implemented by the licensees -- some imediately and others within various deadlines (30, 60, 90, and 120 days, and 6 months).
These measures are specified in Appendix A of the Director's Order.
Included among them are matters dealing with modes of operation, shift manning levels, enhanced training of operators, l
and special containment tests. Some of the numerous specific requirements are:
A.
Effective immediately:
1.
Limit power _ level to keep peak fuel clad temperature at or below 2000 F under large LOCA conditions.
2.
Operate in base load mode only, without load following.
3.
Have at least two senior operators in the control room during operation or hot shutdown.
B.
Within 30 days:
1.
Have vendor representative on site for engineering consultation.
2.
Assure control room habitability under accident conditions.
3.
Enhanced training and retraining provisions.
4.
Special diesel generator tests.
Coments Favoring Interim Shutdown Commenters' safety arguments for interim shutdown relate to emergency plans, timing of long-term fixes, interim measures, short-term risks, dense popu-lation, and psychological impact.
~
. 1.
UCS (#85) argues that no plans exist today to evacuate the public within even 10 miles of the site.
(#85 at 8 and 13.) Both UCS and Mid-Hudson Nuclear Opponents cite testimony by the County Executive of Westchester County that existing plans are not workable.
(#85 at 13 and #86 at 2.)
UCS argues that there has never been an assessmunt of the consequences of a major accident at Indian Point, implying that a basis for emergency planning is lacking, despite NRC's post-TMI commitment to improve emergency planning.
(#85 at 8.) They refer to great difficulty of making effective emergency plans for the area due to physical and demographic characteristics.
(#85 at 8 and 13.) They further comment that the staff has not clearly found that the licensees' emergency plans comply with the applicable Regulatory Guide (1.101) and that, moreover, Regulatory Guide 1.101 does not require evacuation plans out to 10 miles -- a requirement that will not become operative till 1981.
(#85 at 20-21.) They conclude that today, in the absence of effectiv0 protection, the risk is too gredt to permit the plants to operate.
(#85 at 34.)
Mid-Hudson Nuclear Opponents (#86) urge interim shutdown in view of the large population density and absence of adequate evacuation plans for a reasonable distance (15 to 25 miles) (#86 at 4).
New York Public Interest Research Group asserts that it would take an estimated two weeks to evacuate The Bronx, whereas only 1-1/2 days would be available in case of a disaster at Indian Point (4-1/2 days with a " core catcher").
(#67 at 4.)
i
l
. 2.
Timing of long-term fixes:
UCS contends that there is no licensee comitment and no requirement established by the Director's order for licensee implementation of the protective-action time-buying provisions (filtered vented containment and core ladle): only a review of possible modifications is required.
(#85 at 10-11.) They see evidence of a dispute between the staff and the licensee concerning possible imposition of Class 9 accident related requirements.
(#85at11-12.)
UCS argues that the mere possibility of future protection offers no protection today.
(#85 at 11.)
l l
Mid-Hudson Nuclear Opponents refer to post-accident monitoring, aging, and asymmetric LOCA loads as serious unresolved safety issues.
They consider it insufficient for control of present risks to merely say that these issues are being examined -- with an unspecified schedule.
)
(#86 at 3.)
3.
Interim measures:
UCS coments to the effect that (a) the special safety measures already j
present at Indian Point 2 and 3 are of little real value and (b) that the special interim measures yet to be implemented (which, in any case, they regard as inadequate for the long term) should not be counted now, b6cause implementation is largely deferred.
(#85 at 15-21, 27-34, and passim.) With respect to the special safety features identified in the Director's Decision as already present, UCS coments specifically on each.
(#85 at 15-20.) They impugn each, usually on one or both of two grounds:
(a)_that they do little or no good -- or are even counterpro-ductive -- and (b) that they merely reflect implementation of present
~
i j requirements or correction.of inadequacies that could not be tolerated anywhere. Thus, for weld channel and penetration pressurization and the isolation-valve seal-water system, they argue that these measures merely compensate'for bad welds or leaky valves.
(#85 at 16.) For containment atmosphere cleanup, they contend that NRC regulations (Design Criterion 41) require such provision for all planfs.
(#85 at 18.)
Purging for hydrogen control is criticized as counterproductive.
("[T]he staff proposes to seal the containment normally but to vent it during an accident with no capability to filter
....")
(#85 at 17.) For further interim measures, they argue that they are neither extraordinary nor sufficient, and not yet in place.
(#85 at 33 and passim.) Yhe interim measures leave the safety issues raised by UCS unresolved.
(#85 at 33.) They stress fire protection, post-accident monitoring, equipment aging, and asymmetric LOCA loads.
(#85 at 26-31.)
4.
Short-term risks:
UCS asserts:
"Little by little, the short-term grows into the long-term."
(#85 at 32.)
Dean Corren, of Greater New York Council on Energy, expresses the view that distinction between short-term and long-term risks is "an improper and misleading use of the notion of statistical risk assessment."
(#80 at-1.) He contends that any safety improvements that are deemed necessary at all are necessary forthwith.
Brooklyn SHAD offers a similar argument.
(#63)
t Westchester People's Action Coalition views the risks pending completion of fixes as excessive even "for one more day."
(#19 at 3.)
5.
High population density:
UCS stresses the high population density as an obstacle to effective emergency action. They cite Robert Ryan (NRC's Director of State Programs) as characterizing Indian Point as an " insane" site, "a nightmare from the point of view of emergency preparedness," with difficulties exacerbated by severe traffic problems.
(#85 at 8-9.)
Westchester People's Action Coalition argues that dense population inevitably makes Indian Point 10 times more dangerous than the average plant, since plant safety improvements practical at Indian Point should be made nationwide.
(#19 at 5.)
Mid-Hudson Nuclear Opponents ask for suspension oi the licenses pending the Commission's decision, in view of the large population density and inadequate emergency plans.
(#86 at 4.)
6.
Psychological impact:
Westchester People's Coalition calls for consideration of human responses to minor mishaps, rumors of accidents, or threat of accident.
fhey write of human costs in anxiety and potential panic.
(#19 at 3.)
f l
Coments 0pposed to Interim Shutdown Arguments against interim shutdown relate to risk estimates, evacuation, and population density.
. 1.
Risk estimates:
Power Authority of the State of New York (PASNY) (#66) maintains that the staff's risk estimates for Indian Point overstate the risk.
(#66 at17.) They argue that special plant features already existing (identified in the Director's Decision) distinguish Indian Point from average PWR's and lower the Indian Point risks substantially below those derived from WASH-1400.
(#66 passim.) They present plots of Indian Point risks with and without adjustments for plant-specific features.
(#66 at Appendix 2.) The plant-specific adjustments include elimination of some WASH-1400 sequences that PASNY contends are not significant contri-butors to core melt probability. These include loss of auxiliary feed-water after shutdown and reactor transient followed by failure of reactor trip.
(#66 at 16.)
PASNY also asserts that in-vessel steam explosions now appear less likely than estimated in WASH-1400, so that containment failure due to missiles from such an explosion is also less likely.
(#66 at 17.)
2.
Evacuation:
Scientists and Engineers for Secure Energy (SE 2) (#62) describes the emergency evacuation of Mississauga, Canada, a city of 240,000, in November 1979, in connection with dera'ilment of a train that included l
11 propane tanks.
SE 2 cites that experience as showing that massive evacuations are feasible.
(#62 at 3.)
Corren (#80) encloses a statement of PASNY before the Committee on Environmentel Protection of the New York City Council, dated December
9
. 14, 1979, in which PASNY argues evacuability to 10 miles and also argues that a likelihood of evacuation being required for New York City residents under any circumstances.is not 'ealistically foreseeable.
(Page 6 of PASNY enclosure to #80.)
3.
Population densit':
y SE'2 argues that population density around Indian Point is not unusually high by world reactor siting standards. They cite Canadian, French, British, and Japanese practices of siting reactors in densely populated areas.
(#62 at 2-3.)
Differences Between Units 2 and 3 UCS contentions that Indian Point Unit 2 lacks some important safety features of Unit 3 suggest that their arguments for interim shutdown would apply to Unit 2 a fortiori.
(#85 at 21-23.)
IMPACT ARGUMENTS The Director's Decision does not reflect consideration of social or economic impacts of interim shutdown.
Comments on this general subject deal with need for power, cost of power, l
and effect on oil imports.
C 1.
Need for power:
Westchester People's Action Coalition (#19) contends that Indian Point's power is not needed. They assert that there is 50 percent excess capacity in New York; 30 percent without nuclear facilities. They further assert l;
. that there have been no capacity-related blackouts, even though Indian Point Unit 2 has been off-line for four months since last June, ano Unit 3 for five.
(#19 at 6.) They enclose a New York Times article from which they draw their assertions.
Dean Corren, of Greater New York Council on Energy (#80)' contends that there is no need for the Indian Point capacity.
(#80 at 2.)
He presents capacity figures that assertedly show that there is a 3,026-MW unutilized excess capacity (on top of an 18 percent reserve over peat demand), as compared with a total Indian Point capacity of 1,838 MW.
(Page 3 of first enclosure to #80.)
Corren states that Con Ed still claims a 1.8 percent annual peak demand growth, although that growth has slowed to 0.1 percent.
He also states that 69.3 percent of the system was idle in 1978, on the average.
(Page 4 of first enclosure to #80.)
He con-cludes that ability to meet demand would not be compromised by closing Indian Point 2 and 3.
(Page 5 of first enclosure to #80.)
Corren (#80) also encloses statements by UCS and PASNY. The UCS state-ment (at 1) argues that the Indian Point plants are often out of service, yet New York City does not go dark. The PASNY statement (at 7 and passim) argues need for power on economic (not absolute or reliability) grounds.
2.
Cost of power:
Stanley Fink, Speaker of the New York State Assem'iy, comments that o
shutdown would cause economic hardship in the Metropolitan New York i
area.
He considers it the responsibility of NRC to work with FERC and 4
. others to secure replacement non-oil power at comparable cost, if NRC orders Indian Point temporarily shut down.
(#1)
The New York State Building and Construction Trades Council sees a threat to " local economic livelihood" in any Indian Point shutdown.
(#7)
PASNY contends that shutdown would be an economic calamity for New York City, costing PASNY's and Con Ed's ratepayers about $700 million in 1980 alone.
Increases would escalate with imp'orted oil price increases.
(#66 at 20-21.) According to PASNY, 45 percent of the power cost increase would fall on public customers -- New York City and its Metropolitan Transportation Authority (MTA). These entities are already financially hard pressed. MTA's projected $200 million deficit for 1980 would increase by $100 million for increased cost of electricity for subway and commuter rail-lines.
(#66 at 21.)
Corren estimates that shutdown of Indian Point would cost the average residential ratepayer between $2 and $4 per month.
(Pages 11-12 and passim, first enclosure to #80; calculations at Appendix A to that enclosure.) Corren also encloses a concurring analysis by UCS.
In addition, he encloses a PASNY statement (with which he takes issue).
That PASNY statement is generally consistent with PASNY's comment on the Director's decision.
(#66) l i
4 3.
Oil imports:
Fink states that shutdown of Indian Point would exacerbate the region's dependency on impo'rted oil and calls on NRC to work with FERC and others to secure non-oil. replacement power in event of Indian Point shutdown.
(#1)
PASNY comments that the region depends on oil and nuclear sources for electric power generation.
(#66 at 19.)
Indian Point shutdown would require 20 million' barrels of imported oil per year for replacement power.
(#66 at 20.)
Corren presents a " worst-case" replacement-power-cost estimate of $5.21 per month for an average residential customer, based on oil at $30 per barrel.
However, he maintains that replacement fuel is likely to be a more economical mixture of oil, gas, and coal.
(Pages 7 and 8 of first enclosure to #E0 and Appendix A to that enclosure.)
Corren (#80) encloses a statement by UCS, which contains an estimate that replacement fuel would cause a 0.7 percent increase in total U.S.
imported oil consumption.
Corren's (#80) last enclosure includes a remark by Commissioner Bradford that nuclear electric generation frees up " residual oil, of which there is something of a surplus anyway."
l
A-1 APPENDIX A SAMPLE GENERATION OF A COMPLEMENTARY CUMULATIVE DISTRIBUTION FUNCTION - CCDF The CCDF is used to present the risk of reactor accidents in the fonn of a plot of probability vs consequences.
The average reader is unaccustomed to studying risk in this form of. presentation.
To facilitate understand-ing of the CCDF, consider generating a CCDF for the risk of death from air crash from high altitude using the attached figure.
If an airplane crashes from a high altitude, it is virtually certain that all on board will perish. Thus, Figure A-1 is a reasonable first approximation of a CCDF for such a crash; it shows a probability, P,
g that 300 deaths, the seating capacity of the aircraft, will occur.
Pg is the probability that the plane will crash; 300 is the limit of those on board who will die in a crash.
For this simple CCDF curve the expected risk is P, say 0.33 crashes per year, times 300 deaths per crash or 100 o
deaths per year.
The CCDF can be corrected first to show that the falling aircraft might strike and kill people on the ground.
Figure A-2 shows a tail on the CCDF curve reflecting that if the plane crashes, it will most likely not kill many people on the ground. At lower and lower probability, there is the chance of killing crowds in buildings or gatherings so the curve tails off toward some higher number of deaths.
Presumably there is a I
limit to the ground deaths that can be caused by the crash of a 300 passenger aircraft, perhaps 10,000 or 20,000 if it crashed into a 1
s A-2 crowded sports stadium. At that limit, the curve would no longer tail off to the right but become a vertical line showing a physical limit analogous to the seating capacity limit.
A second stage of refinement in this CCDF can be obtained if the airline gives us. figures on the actual passenger loads the aircraft usually carries.
If the data are limited, they might simply be reduced to the approximation that on 1/3 of the trips the plane is 1/3 full, on another 1/3 of the trips it is 2/3 full, and on another 1/3 of the trips it is completely full.
The CCDF can now be refined as shown in Figure A-3.
One hundred deaths occur at probability P, the probability of crash, o
because the plane is always at least 1/3 full.
At 0.67 P, the curve shows 200 deaths because the plane is at least 2/3 full 2/3 of the time.
And the curve shows 300 deaths at 0.33 P because on one third of its g
flights all seats are filled.
We can reflect the probability of ground deaths by putting soft tails on the sharp steps of the curve.
i As more accurate flight data are accumulated, the steps in Figure A-3 l
l can be refined into a more accurate curve as shown in Figure A-4. This last curve would represent the most accurate distribution of the likeli-hood of death from high altitude air crash.
l i
CCDF FOR AIR CRASH FROM HIGH ALTITUDE SEATING SEATING CAPACITY CAPACITY P,
P, N-N s
a E
E 2
2 E
E o.
o.
i i
i i
i 3
100 200 300 100 200 300 L
DEATHS DEATHS A-1.
A-2 P
SEATING 5, O SEATING o
CAPACITY CAPACITY
{ 0.67 P,
]
d 0.33 P 8
g a.
o l
l 1
1 1
1 100 200 300 100 200 300 DEATHS DEATHS A-3 A-4
m.
B-1 APPENDIX B REBASELINING OF THE RSS RESULTS The results of the Reactor Safety Study (RSS) were updated for nurposes of this comparative study.
The update was done largely to incorporate results of research and development conducted after the October 1975 publication of the RSS and to provide a baseline against which the risk associated with various LWRs could be consistently compared.
Primarily, the rebaselined RSS results reflect use of advanced modeling of the processes involved. in meltdown accidents, i.e., the MARCH computer code modeling for transient and LOCA initiated sequences and the CORRAL code used for calculating magnitudes of release accompanying various l
accident sequences.
These codes have led to a capability to predict the transient and small LOCA initiated sequences that is considerably advanced beyond what existed at the time the Reactor Safety Study was completed.
The advanced accident process models (MARCH and CORRAL) produced some changes in our estimates of the release magnitudes from various accident sequences in WASH-1400.
These changes primarily involved release magnitudes for the iodine, cesium and tellurium families of.
iso topes.
In general, a decrease in the iodines was predicted for many of the dominant accident sequences while some increases in the release magnitudes for the cesium and tellurium iso?. opes were predicted.
I It should be noted that the MARCH Code was used on a number of scenarios in connection with the 'TMI-2 recovery efforts and for Post-TMI-2 investi-gations, e.g., Rogovin) to explore possible alternative scenarios that TMI-2 could have experienced.
l l
B-2 Figures Bl and 82 show a compari. son of the original RSS and the rebaselined PWR and P.WR designs for the individual risk versus distance of early fatalities snd latent cancer fatalities, respectively.
These figures show the expected values conditioned upon a core melt accident of 'about 4
one chance in ten thousand reactor years (1x10 ).
This particular conditioned value reflects an average of the core melt probabilities estimated from a number of LWR designs.
Entailed in this rebaselining effort was the evaluation of individual dominant accident sequences as we understand them to evolve rather than the technique of grouping large numbers of accident sequences into encompassing, but synthetic, release categories as was done in WA$H-1400. The rebaselining of the RSS also eliminated the " smoothing technique" that was criticized in the report by the Risk Assessment Review Group (sometimes known as the Lewis Report; NUREG/CR-0400).
For rebaselining of the RSS BWR design, the sequence TCY' was explicitly included into the rebaselining results.
The accident processes associated with the TC sequence had been erroneously calculated in WASH-1400.
For rtoaselining of the RSS PWR design, the release magnitudes for the Event V and TMLB' sequences were explicitly calculated and used in the consequence modeling rather than being lumped together into Release Category #2 as was done in WASH-1400.
In both of the RSS designs (PWR and BWR) the likelihood of an accident sequence leading to the occurrence of a steam explosion (M) in the reactor vessel was decreased.
This was done to reflect both experimental
B-3
?
and calculational indicatforw that such explosions are unlikely to occur in those sequences involving small size LOCAs and transients because of the high pressures and temperatures expected. to exist within the reactor coolant system during these scenarios.
Furthermore, if such an explosion were to occur, there are indications that it would be unlikely to produce as much energy and the massive missile-caused breach of containment as was postulated in WASH-1400, i
As can be seen from Figures 81 and 82, the net (or overall) change in consequences predicted from the rebaselined RSS results are quite small.
In general, the rebaselined results led to slightly increased health impacts being predicted for the RSS BWR design. This is believed to be largely attributable to the inclusions of TC#'.
The rebaselined RSS-PWR led to a small decrease in an individual risk of early fatalities and latent cancer fatalities below the original RSS PWR.
This is believed to be largely attributable to the decreased likelihood of sequences involving vessel steam explosions (a).
In summary, the rebaselining of the RSS results led to small overall differences from the predictions in WASH-1400.
It should be recognized that these small differences due to the rebaselining efforts are likely to_ be far out-weighed by the uncertainties associated with such analyses, i
i I
~8-4 3e FIGURE B1 - RISK OF EARLY FATALITY TO AN INDIVIDUAL VERSUS DISTANCE GIVEN A CORE MELTO 10-2
=
Z Z
~
ff
.un
- - RSS - DESIGN REB EMDN 10-3 N:k.
<c..
I mg
[h, g
c g
\\ \\
" 10-4 5
k, b
5
~
PWR b
Y...
'#Ib b
h.N- ~
$h b
ik E
-5 3.,
10
' hlIiI
- h o
- 4 M,, k..
^yh)
E
!E.
/
f o
"% 'li;...
p g
~
Z 10-6
=
10-7 0.5 1.0 1.5 2.0 2.5 DISTANCE (MILES)
- CORE MELT PROBABILITY ASSUMED TO BE 10-4/ REACTOR ASSUMPTIONS:
RSS-DESIGN 1.
ALL RSS CORE MELT ACCIDENT RELEASE CATEGORIES 2.
ALL RSS ASSUMPTIONS (E.G., SM0OTHING)
REBASELINE DESIGN
~
1.
SM0OTHING ELIMINATED 2.
EXPLICIT ACCIDENT SEQUENCES 3.
NEGLIGIBLE PROBABILITY OF VESSEL STEAM EXPLOSION EXPECTED CONSEQUENCES FROM 91 WEATHER SEQUENCES WITH 3200 MWT POWER LEVEL ENTIRE CLOUD EXPOSURE + 24 HOUR GROUND EXPOSURE SHIELDING BASED ON NORMAL ACTIVITY
-B-5
,.,r FIGURE B2 - RISK OF LATENT CANCER FATALITY TO AN INDIVIDUAL VERSUS DISTANCE GIVEN A CORE MELT
- 10-3
=
2 a.
g..
a
?fs.
. v.
RSS - DESIGN "13
. 5!:g REBASELINE DESIGN
. 10,4 5 :l*
-\\
m5
- j!
E r
b 10 g....
m.
I g.
\\'
a E
X'9:
- d A*3.h,A BWR W
/
O 10-6 a-N....
g.
2 O
[
)%
o..
f..
,ssa 10-8
!O.
PO 30 40 50 0
DISTANCE (MILES)
- CORE MELT PROBABILITY ASSUMED TO BE 10-4/ REACTOR YEA ASSUMPTIONS:
RSS-DESIGN 1.
ALL RSS CORE MELT ACCIDENT RELEASE CATEGORIES 2.
ALL RSS ASSUMPTIONS (E.G., SMOOTHING)
REBASELINE DESIGN 1.
SM",0 THING ELIMINATED 2.
EXPLICIT ACCIDENT SEQUENCES 3.
NEGLIGIBLE PROBABILITY OF VESSEL STEAM EXPLOSION EXPECTED CONSEQUENCES FROM 91. WEATHER SEQUENCES WITH 3200 MWT POWER LEVEL ENTIRE CLOUD EXPOSURE + 24 HOUR GROUND EXPOSURE SHIELDING BASED ON NORMAL ACTIVITY
-..c A
APPENDIX C
~
O !$
Department of Energy vAY I 5 EO Washington, D.C. 20461 Mr. Edward J. Hanrahan, Director Office of Policy Evaluation Nuclear Regulatory Commission Washington, D.C.
20555
Dear Mr. Hanrahan:
This 1etter summarizes the views of the Economic Regulatory Administration's Division of' Power Supply and Reliability (DPSP) regarding the electric s'ystem reliability impact of various modes.
of operation of nuclear power units Indian Point 2 and 3 as described in your April 28, 1980, letter.
Indian Point 2 is a 849 MW (summer rating) PWR unit owned and operated by the Consolidated Edison Company of New York (CON ED).
Indian Point. 3 is a 965 MW PWR unit owned and operated by the Power Authority of the State of New York Inc. (PASNY).
The units are co-located in Westchester County, 25 miles north of New York City.
Both units are included in their respective entity's planned rescurces available to meet customer demands in 1980.
A sh2tdown of Indian Point 2 and 3 would impact the reserve capacity in th'e New York subregion (New York Power Pool) of the Northeast Power Coordinating Council (NPCC).
Without Indian Point 2 and 3 the available reserves in the New York Power Pool for the summer 1980
.and winter 1980-81 seasons would decline from 46.6 and 58.2 percent to 38.0 and 49.0 percent respectively.
This level of reserves is still considered adequate to. provide reliable electric service when viewed on a state-wide basis.
Eowever, due to limited transmission capability into the New York City and,Westchester County areas, the complete shutdown of these two Indian Point units during the 1980 summer peak period could adversely impact the-system reliability of CON ED and PASNY.
The electric. service area of CON ED consists of the five boroughs of New York City and a major part of Westchester County, an area of 600 square miles.
CON ED supplies electricity to over eight million customers.
CON ED's summer peak load is about 40 percent higher than its winter peak load due mainly to the widespread use Lof electric air conditioning.
CON ED's projected 1980 summer peak load is 6900 MW.
e 8007140\\it)
I.
+
C-2
+
2 PASNY does not have any geographically defined " service territory" but serves particular classes of customers in all parts of the State of
-New York.
PASNY's projected summer 1980 peak load is 2503 MW of which 1.ess than half is In New York City and Westchester County areas.
The DPSR collected data concerning the latest e'lectric system conditions, maintenance schedules, expected forced outages, and expected peak loads ~for CON ED.and PASNY.
The data was compared to historical data contained-in various DOE documents, and revised where it was felt necessary.
The conclusions drawn in this letter are based upon our analysis of this data.
The adverse impact on reliability due to the status of IP 2 and 3 results from the limited transmission system capability for importing power from other parts of the state, or from neighboring states, into the area in which these units are located.
CON ED has bulk power transmission tie lines with neighboring utilities having a megawatt transfer capability (above scheduled transfers) as shown below:
SUMMER WINTER FROM 1980 80-81 Upper State New York 500 2200 Pennsylvania-New Jersey-Maryland Interconnection (PJM) 150 50 Long Island Lighting Co. (LILCo.)
475 550 TOTAL 1125 2800 Energy transfers from areas outside of PJM or Upper State New York (USNY) would have to rely on the same transmission ties as transfers directly fgom Upper State New York and PJM.
Therefore, the only time capacity available from these outside areas would need to be considered would be if the PJM and USNY areas did not have sufficient capacity to supply the transfer limit.
Overall New York State generating capacity will be adequate in this time frame and all possible transfers from the north, up to the limit of the trans-m'ission system, could come from this area.
PJM also projects sufficient available excess capacity during the 1980 summer to be able to provide the 150 MW of capacity for transfer.
Long Island Lighting Company -(LILCo) does not have suf ficient available capacity this summ.er to provide the full 475 MW of the transfer capability and would only be able to supply an average of 100 MW.
There are no other systems to the east of the area that could supply power over these Long Island transmission connections.
There may be some co,ntingency support through the submarine cable from Connecticut but this would be limited to only 145 MW.
'I D
e-v
C-3 4
3
- The four possible modes of operation of IP 2 and 3 which DPSP was requested to analyze are detailed below indicating their impact on CON ED and PASNY:
o Operate IP 2 and 3 at 50 percent capacity for a 3 month period beginning June 1, 1980.
The loss of 907 MW f rom the CON ED and PASNY systems during the summer peak load period along with the expected amounts of forced and scheduled outages, will enable the companies to supply their expected peak loads plus withstand the loss of the largest unit (Ravenswood #3 - 928 MW).
This should be adequate for maintaining reliable electric service since normal tie transfers were considered, o Shutdown IP 2 and 3 for a three month period beginning June 1, 1980.
The loss of 1814 MW f rom the COM ED and PASNY systems during the summer peak load period will force the New York City and Westchester County areas to depend very heavily upon the transmission interties with neighboring areas.
Given the projected loads and expected forced and scheduled outages, the loss of the largest unit (Ravenswood #3 - 928 MW) would forcc the utilities to use all available capacity and interities to the maximum reasonable extent.
Further facility failure, or loads greater than forecast would force the utilities to institute voltage reductions, load curtailments, or other actions as required to prevent widespread loss of customer load.
Sustained high loads during the summer per.iod would force CON ED to operate its 1987 MW of combustion turbine generation capacity for longer periods than the units are planned and designed to operate.
This mode of operation places the system in a very vulnerable position and is not considered consistent with providing reliable electric service.
o Operate IP 2 and 3 at 50 percent capacity for a 12 month period beginning June 1, 1980.
The loss of 907 MW in the CON ED and PASNY systems for 12 months will have its greatest impact on system reliability during the summer months.
This situation is discussed above.
During the remainder of the year, CON ED and PASNY will have suf ficient capacity to provide reliable electric service to their customers.
a
- s
+
C-4 4
o Shutdown IP 2 and 3 for a 12 month period beginning June 1,-
1980.
The loss of 1814 MW f rom CON ED and PASNY will have its greatest impact on system reliability during the summer months as discussed above.
During the other months of th'e year, CCN ED and PASNY will have sufficient available capacity' on their own systems and from transfers from other areas to provide reliable electric service.
This analysis deals only with electric system reliability and energy supply; it does not consider the need to reduce operating costs and conserve oil or natural gas.
The out. ages of any large non-oil generating unit in Southeastern New York will result in increased costs to the consumers of electricity because of the resulting increased use of low sulfur oil-fired generation.
I would appreciate being notified of the decision regarding Indian Point 2 and 3.
S' rely,p Richard E. Weiner, Director Division of Power Supply and Reliability Economic Regulatory Administration 6
0 e
a e
8 e
e
.