ML19289C818
| ML19289C818 | |
| Person / Time | |
|---|---|
| Site: | Atlantic Nuclear Power Plant  |
| Issue date: | 01/12/1979 |
| From: | Jacqwan Walker CALIFORNIA, STATE OF |
| To: | |
| Shared Package | |
| ML19289C819 | List: |
| References | |
| NUDOCS 7901250231 | |
| Download: ML19289C818 (151) | |
Text
{{#Wiki_filter:X"C FIT ' i 't THIS DOCUMENT CONTAINS POOR QUAUTY PAGES UN1IZED STATES OF AERICA NUCLEAR P2(TJIATORY COMMISSICIP F0 :.U.2_ In the I-htter of: )
)
0FFSHORE P0h'ER SYSTEMS, )
) DOCKET No. STN 50-457 (Manufacturing License for )
Floatir4 Nuclear Power Plants). ) ,,g
) -I,h, .h
)
~ '.,g
[G , REQUEST OF THE CALIFORNIA ENERGY RESOURCES M~ CONSERVATION AND DEVEIDPENT COMMISSIOF d 9 @ '~ ., a-ON TE ISSUE OF CONSIDERATION OF CLASS 9 ACCIDEITfS Yp :.. , - .y b /"= T ; m-~ l - The California Energy Resources Conservation and Develope nt ' Co==ission by this petition requests leave to file co==ents in this case on the issue of consideration of Class 9 accidents. It is our understanding that the Nuclear Regulatory Co==ission has requested co==ents on what Class 9 accident policies should be considered in license applications. The co==ents submitted with this request address that issue. . Pursuant to California law, California Public Resources Code Section 25524.3, the Energy Co==ission has completed and held hearings on its extensive study of the necessity for, feasibility and desirability of requir.ing undergrounding of nuclear reactors licensed for construction in California. Although the Co==ission ultimately decided not to require undergrounding, we feel the infor. cation which was produced in the study is of significant value to state and federal agencies responsible for licensing nuclear reactors. We are therefore pleased to be able to share this information and expertise 19o GSo%3 (
with the Nuclear Regulatory Co=ission and hope that you will find it useful in your deliberations in this and other cases. As a part of the study, the Co=ission investigated and araly::ed accident sequences, their potential consequences, and the potential for citigation. We would therefore particularly direct your attention to the enclosed reports. We have included, in addition to the selected reports fro = the Co=ission's Undergrounding Study, a brief su=ary of the Co=nission's co Jents on the specific issue under consideration. Respectfully subnitted.
~
si (d 'n a.es A. Walker
- cutive Director valifornia Energy Resources Dated: 1/12/79 Conservation & Developnent Co=nission
COIEUTS OF TE CALIICEIEA ENERGY COIGESSION ON TE ISSUE OF CONSEERATIOU OF CLASS 9 ACCEENTS The following responds to the HR0 request for ce=ents concerning the issue of Class 9 accidents at nuclear power plants and what related policies should be adopted by the URO. The co=ents are intended to be general in nature and are not specifically directed to floanng nuclear power plants. The " defense-in-depth" concept has been the foundation of nuclear safety. perhaps a most significant ele =ent of defense-in-depth is the concept of design basis accidents, events considered unlikely, yet postulated to occur for the purpose of design of ergineered safety features. The safety of co=ercial nuclear facilities is integrally dependent on the choice of design basis accidents. The co=plexity of modern nuclear facilities, operating experience, and the findings of the Reactor Safety Study (WASE-1400, Im 75/014) confir: that accidents currently assumed for safety design purposes are only a limited sa=ple of potentially hazardous events arai that it is possible to overlook significant accident sequences. We agree with the HRC that it is time to rethink carefully the safety design basis of co=ercial nuclear power reactors and reco==end that Class 9 accidents be considered during licensirs. The reco=cendation that Class 9 accidents be more fully considered in the design and siting of nuclear power plants ste=s not from a belief that current designs may be unsafe, but, rather, from a position that the present level of safety of conte =porary plants, in co=parison to other societal risks, is uncertain. (The current safety record of operating nuclear power plants is not evidence that extre=e accidents will not occur. The recent co==ercial airline tragedy in California illustrates that an industry with an excellent safety record is not i=une to accidents.) The Risk Assessment Review Group Report to the U. S. Nuclear Regulatory Co==ission (I"JEEG/CR-0400) concluded similarly when they noted they were unable to determine whether the absolute probabilities of accident sequences in WASH-1400 (the Reactor Safety Study) were high or low, but believed the error bounds on those esticates to be " greatly understated". That the current level of reactor risk is known mathe=atically only to be less than the risk derived from the operating experience of current generation large power reactors argues for very conservative design and siting policies. Accident prevention and control through sound engineering and conservative operating policies are the principal guarantors of public safety. The California Energy Co==ission believe.3, however, that regulatory policy has i= properly limited the consideration of Class 9 accidents and that public safety could be served by proper treatment. A policy to include Class 9 accidents as a licensing consideration =ust be properly focused. The underlying objective cust be reduction of public risk from reactor accidents. Wa believe that objective can be cet through passive designs which need not be engineered and constructed to the rigorous standards of the present Engineered Eafety Features. Such an approach could make Class 9 accident consequence reduction an achievable goal.
Cc==ents of the California Energy Cc==ission er the Issue of Consideration of Class 9 Accidents The California Energy Co==ission has recently co=pleted a study of the feasibility of underground siting. A copy of two pertinent reports frc= the underground siting study are enclosed for consideration in these proceedings. The first , an Energy Co==ission Staff Report , su==arizes the entire study. The second report, prepared by a study contractor, addresses the phenomenology and potential consiquences of core-melt accidents. The work concluded that properly i=ple=ented underground siting, and other concepts, could be effective in reducing the public consequences from Class 9, including core-melt, accidents. The design proble=s posed by the hypothesis of a core-=elt accident are =any, but were found tractable. A second means to reduce the potential for accident consequences is through siting policy. Current licensing procedure limits the potential population which could be exposed to an extre=e accident by controlling the distance fro: plant site to population centers. The population-distance criteria, however, could be made core conservative to reflect increased attention to Class 9 accidents. We appreciate the co=plexity of specific site evaluation and that current co=puter codes =ay not be adequate to this task. In co=menting on the Energy Co==ission's Underground Siting Study, the NRC noted that a progra= was in progress to evaluate the relationship between wind patterns and population distribution and their affect on consequence modeling. This activity should be extended to determine desirable characteristics of sites which might minicize public consequences from Class 9 accidents. Such ntnlysis should be perfor ed and potential sites evaluated early in the site review process. In su-:ary, the possible consequences from Class 9 accidents should receive incceased attention during licensing. Studies sponsored by the Energy Cc==ission have suggested promising design approaches for mitigation of possible public consequences from extreme accidents. In addition, these studies have noted that it =ay be feasible to use design criteria for Class 9 =itigation syste=s less stringent than those used for Engineered Safety Features. Siting policies can also significantly affect the magnitude of potential public consequences. There are, considering Class 9 accidents, good sites and bad sites. The formal regulatory evaluation should be extended to separate the former from the latter.
DRAFT-UNDERGROUND SITING OF ~ NUCLEAR POWER REACTORS: - ~ AN. 0PTION FOR cal.lFORNI~A ~
% M/.
STAFF REPORT T go b h
. WQ:l h A
SUMMARY
OF THE
~ TECHNICAL AND ECONOMIC IMPLICATIONS WITH RECOMMENDATIONS 3
.I U N E : 9 7 8 Ob w.wh _
DRAFT UNDERGROUND SITING OF NUCLEAR POWER REACTORS: AN OPTION FOR CALIFORNIA A Sumary of the Technical and Economic Implications with Recomendations DRAFT Nuclear Assessments Office June, 1978 California Energy Commission e*
- M
_, 4mmqog e Mw W
DRAFT NOTICE This report was prepared by staff of the State of California Energy Commission, but does not necessarily represent the ', iews of the Commission except to the extent, if any, that it has formally been approved by the Commission at a public meeting. For information regarding any such action, coc:municate directly with the Commission at 1111 Howe Avenue , Sacramento, California 95825. The purpose of this study was to determine the general economic and technical feasibility, and environmental implications of underground -iting of nuclear power reactors. Although detailed cost figures, designs, and analyses were conducted, they were prepared only for this use. Neither the Commission, State of Cali fornia , nor any officer, employee thereof, or any of its contractors or subcontractors intends that the information herein be used for any other purpose, and makes no warranty, express or implied, or assumes any legal lia-bility whatsoever for the contents of this document. i
DRAFT FOREWORD A number of individuals and organizations have contributed to the project which this document describes. A substantial body of technical investigation and analysis underlie the major project conclusions. In general, study participants have prepared, and assume responsibility for, conclusions pertinent to their area of endeavor. This report summarizes that information and melds the tech-nical and social concerns related to underground siting to a policy framework. The interested reader is referred to the source documents listed as references at the end of each chapter to verify that work by others has been faithfully reported. This report summarizes a great be,dy of material. It is not intended to be a technical document and includes supporting explanations where considered necessary. It is not an exhaustive treatment of reactor safety or risk al-though, necessarily, these subjects are considered. Recommendations presented in this report represent soleiv the views of staff of the California Energy Commission. 11
DRAFT UNDERCROUND SITING OF St' CLEAR P0kTR REACTORS: AN OPTION FCR CALIFORNIA EXECUTIVE
SUMMARY
General Conclusions This study has developed solutions to the technical proble=s associated with so-called reactor core-=eltdown accidents. A principal conclusion of this report is that it is feasible to design passive systems to reduce potential consequences f ro= core-=elt accidents to very low levels. Underground siting is only one of several alternatives. The two underground siting concepts developed, each aug=ented to prevent buildup of excessive pressures, can effectively eli=inate public health consequences frc= hypothetical, highly unlikely, yet physically possible reactor accidents. In additica, alternatives to underground siting, such as re=ote siting and controlled release of excessive pressure through si=ple, engineered filter systees, appear feasible. k*h ile not as fully effective, these alternatives capture so=e of the benefits of underground siting at less cost. The =ajor staf f reco==endation ste==ing from this work is that underground siting not be candated due to (a) the uncertainty remaining over costs, construction ti=e and possible licensing concerns; (b) the existence of what appear to be coderately ef fective and less expensive technical alternatives; and (c) the opportunity to i=ple=ent re=ote siting within California. Introduction In the fall of 1975, the California State Asse=bly Co==ittee on Resources, Land Use, and Energy, conducted a series of hearings on issues related to the safety of nuclear power generation within the state. Subsequent to these hearings , in June 1976, California enacted three laws which directed the California Energy Co==ission (CEC) to exa=ine issues concerned with nuclear vaste disposal, reprocessing of spent fuels, and, the subject of this report, underground siting of nuclear power reactors.* Faced with contradictory evidence and testi=ony over the actual safety of large power reactors, the Legislature was =otivated to require this study. The intent was to evaluate the abtlity of underground siting to cope with the =ost extre=e and, a s ye t , hypothetical, reactor acci-dent: the so-called core-celtdown. The work, which this Executive Su==ary highlights, is a thorough review of the feasibility and effectiveness of underground-sited nuclear power reactors. It is, perhaps, the cost in-depth analysis yet done by a state in the general area of nuclear reactor safety. Although designed to address directly the requirecents of the candating legislation, e.g., that the Co==ission study "... the necessity for, and effectiveness and econo:ic feasibility of ..." (Public Resources Code 25524.3) underground nuclear power reactors, it has also exa=ined
*The ter: " underground" is used in a generic sense to include two concepts: a) construction of reactor and related syste=s in caverns cined in rock; and b) so-called ber=-contain=ent wherein a structure housing a reactor v uld be covered with soil after construction. Both concepts were studied and will be core fully described in the tollowing pages.
iv
DRAFT basic considerations of accident processes, safety, power plant siting, and the means by which safety is measured and accepted. In this light, underground siting represents one avenue to safety and embodies an approach whien can contribute meaningfully to enhancing nuclear reactor safety. Program Structure The overall study program was directed by the staff of the CEC. For the highly technical and complex analyses necessary, the Commission turned to experienced engineers and nuclear specialists. The two engineering teams selected for the preparation of underground nuclear power plant designs, cost estimates, and evaluation of potential licensing and construction problems , had unique quali-fications and prior experience with the underground concepts. The accident analyses were performed by a team of specialists with extensive experience in evaluating the safety and reliability of nuclear power plant designs and in estimating radiological health effects from hypothetical reactor accidents. Organizations previously involved in studies of underground nuclear power plants provided technical program management and coordination of other contractor efforts. The various contractor teams produced an extensive technical evaluation of underground nuclear power plants, the most complete yet performed. Concepts Studied Two general underground concepts were examined. The first is often referred to as rock or mined-cavern siting. In this concept, large caverns, of approxi-mately 100 feet in width, 200 or more feet high, and hundreds of feet in length, would be excavated in solid rock. The nuclear reactor and associated systems would be built in the large caverns af ter all rock movements , which occur in most underground excavations , had ceased. A rock-sited nuclear power plant is pictured in Figure A. The second concept, termed berm-containment in the law requiring the study, is sometimes called cut-and-cover, pit siting, or by other names. In this concept, shown in Figure B, a large open pit would be excavated, approximately 400 feet in diameter at the bottom, about 150 feet deep, and the nuclear power plant built. The plant would subsequently be covered with the soil initially removed from the excavation. Studv Objectives and Limitations The principal study objective was to evaluate the ability of underground nuclear power plants to cope with extreme reactor accidents, so-called core-melt or Class 9 accidents. Other arguments for underground sitiag, such as improved earthquake resistance, easier decommissioning, improved resistance to cabotage, urban siting, and others, were also evaluated. A formal risk analysis, where specific accident prcbabilities and related The directing consequences are combined, was not done for several reasons. legislation, taken in conjunction with its historical origins, would not have been satisfied by such an approach: considering safety, underground siting is necessary to combat only the most extreme reactor accidents. A directive to V
y (\
-r W
te. g m ' 'qe, ---
}, .-7 ~
'. y ==
~~
- g. l, ,
-l=
.i Lo o e
%g ffjf
':~ , ._/
,n L,r'
?
b fY _ re ' # j L,
<6,TN( Tf t
{ R[ACIOR CAvt1Nn, I s f i ICON 1ROL l h guxittARY CAvtRN 4 C[ Nit 1 e l I g i ' '# %Q l , k, N w *
.) N h y ^ t - ,4 / LIP ' 1 m ', >
f
^
S L M ' #
~s, EXPAN510N TRA
- f . /. , .
L- -
- g
} t 2)
CA - ltjittiNE' Cit /RA 4 . W151 ACCESS ,*f,' tAsi ACCESS CA N~ , 10NN[L, #,/' ' , TUNN[L - "
/,Af -/
M gure A ROCK-s1TED HUCLEAR POWER PLANT <
pp
/
e "Q- .
'gE[,ff_j_'!;jji L jji L J
- f ,, ,ggy'uQ .xQvy w ElliLlg g%%
.; . .v;T=M gq
/>M..If).
' (, 'u, -
x
.u[gr[!f 1
-- --! 2 3 0 k.r. 3. . t r
\
J'
- o
{" Y ,!l4 % y if; ! g
. . . j,
,q
?g k. '4
{ 7 j
'I.}
<.4i%y.
yjln I)k . l : .<
~
emLLcpk 1 4m W s . ggggg n k g 4 ll.c.,jpg ' w ": W a-
'lH c l' r
. l (
_ t.smrvae 1"
*"** iTTMK .'[i-L; l ;' .
1 p f .o. . , ga.u. ,u
.v
. q' .
l I f;l l 3 [ b , 5 -- :. M !d k i d gg{
.< ;.gj 7 p j j Ip., lIllill;llu i
i
.. ..,.. .ag q + - ...
mp g . s_~oi~g,ao
~ ~ -
-;.,. U. ] L ! I. STATE 0F CALIFORNIA g" ._
'E Energy Resources Conservation
' ' ~
and Development Commission Nuclear Undergrounding Study Figure B Berm-Contained Pressurized Water Reactor Plant DRAFT
DRAFT study underground siting, by implication, is one to study core-melt accidents. Secondly, the underground nuclear power plant designs were developed to the conceptual design level only and are not sufficiently detailed for an exhaustive risk analysis. A prototype plant would be necessary for this purpose. Alternatives were considered to permit a balanced answer to the question of the " necessity for" underground siting. Studv Approach The CEC study objectives were to develop, if possible, underground plant designs which utilized the earth (both rock and soil) as a direct means for controlling and reducing potential reactor core fission product releases. No attempt was made to " bottle up" accident-generated radioactive gases and vapors in the underground plant itself, a fundamental difference between this e f fort and previous studies of underground siting. Many accident sequences were examined to identify those which would place the most extreme physical demands on the underground facilities. An envelope of extreme conditions was used to design the pas sive (no parts require external power or activation signals) accident mitigation systems of the underground plants. The design concepts of this study utilize an " engineered failure pathway" for temperature / pressure management. Basically, by controlling the mode and pathway of containment failure, potential accident consequences can be greatly reduced. The engineered failure pathway concept was also applied to a surface facility and cost estimates prepared. A release of radioactive material similar to PWR-2 of the Reactor Safety :itudy was used to estimate public health consequences should similar accidents develop in conventional surface plants and the underground designs of this study. For the underground designs, the accidental release of radioactive gases and vapors was directed into the soil, in the berm-containment configuration, or, in the rock case, into special expansion / condensation chambers. The radioactive material was allowed to migrate through the soil or rock driven by the accident generated pressures. When the material reached the ground surface, it was assumed to travel through the atmosphere in a manner similar to the release from the surface plant. The dif ference in estimated accident consequences between the surface and underground plants is due solely to the enforced travel of radionuclides through the soil or rock for the underground concepts. All other modeling parameters, including methods of calculation of health effects, are the same for the surface and underground plants. The consequences which might result from accidental releases of radioactive material were estimated using the most current data for radiological health e f fects. Accident consequences calculated included early and latent health fatalities and other e f fects , e.g. , cancers , property damage , and evacuation and relocation costs. viii
DRAFT Cost estimates were prepared according to the assumptions and approaches normally used by each engineering firm. Estimates for reference surface plants, designed for California conditions of potentially large earthquakes and the need for closed-cycle cooling systems, were prepared. Thus, each underground concept could be compared to a related surface facility. The cost estimates prepared are, without doubt, the most detailed yet done for contemporary studies of underground siting. The underground plant designs reflect state-of-the-art engineering and are based on a 1300 MWe nuclear steam supply system. Both pressurized and boiling water reactors were considered. To as great an extent as possible, surface plant systems were incorporated, without modification, in the underground plants. All currently existing regulatory requirements were considered in the preparation of the underground designs. Engineered Safety Features (ESF) required for surface plants were incorporated in the underground designs. The Accident Mitigation Systems of the underground plants were intended to supplement, but not replace, contemporary ESF. Earlier studies of underground siting had concluded that little or no opportunity existed within California for rock siting and that berm-contained siting was nearly unconstrained. A screening effort was conducted to verify the existence of technically feasible siting areas for each underground concept. Specific Results and Conclusions The findings of this study are grouped generally along the major divisions of the mandating legislation. Feasibility o It is technically feasible to berm-contain 1300 MWe pressurized (PWR) and boiling water reactors (BWR). o Cavern span requirements for a rock-sited 1300 MWe PWR are approximately 100 feet and within the state-of-the-art. Cavern spans for a similar sized BWR would be approximately 150 feet. Although there is no apparent tech-nical reason to consider spans of 150 feet in feas ible , a demonstration program would be necessary before such a span could be declared attainable. In general, however, rock siting has also been found to be technically feasible. o The basic underground concepts developed in this study are generally compatible with most existing regulatory requirements and guidelines. No major alterations were made to Engineered Sa fe ty Features (ESF) or the currently used safety design philos'phy. The Accident Mitigation Systems added to the underground designs supplement, but do not interfere with, the operation cf currently required ESF. Although detailed questions would undoubtedly arise during normal licensing, the general designs contain no significant departures from present requirements. o Few operatiens or maintenance penalties are anticipated for the conceptual designs of this study. The underground plant layouts incorporated spatial requirements for normal operations , main tenance and safety, comparable to or greater than the reference surface plants. ix
DRAFT o Problems due to groundwater infiltration, notably reported at the Lucens, Switzerland, underground nuclear plant, can be avoided by engineered intercept systems. o Costs for an underground facility are somewhat higher than for a comparable surface plant. On the basis of engineering cost es timates , berm-contained project total costs are predicted to be, at minimum, 14* more expensive, and can increase depending on plant layout and design specifications. On the same basis, mined-cavern plants are estimated as 25 more costly. The uncertainties of underground construction suggest that prototype costs will ' be higher. Fur the more , first plants will be more expensive than later,
" mature industry" facilities, due to the need to establish details of design, licensing, and construction. For the earliest plants, the cost percentages cited above must be considered preliminary, lower bounds of ~
actual costs. Probabilistic treatment of time schedules suggest that, for prototype facilities, at the 85 confidence level, berm-contained plants would be approximately 37 more expensive. o Construction is estimated to take approximately 22 months longer for berm-contained facilities and 19 additional months for mined-cavern con-cepts. The time extension is due to primarily two factors: the need for large initial excavations and reduced accessibility during construction, o Technically acceptable areas exist within California for both bero-contained and mined-cavern siting. Effectiveness o Augmented underground siting can reduce to insignificant levels public consequences from nearly all extreme reactor accidents. No evacuation is necessary. As a safety system, underground siting can fail; however, because of the manner of failure, such occurrences would contribute little to overall risk. o Controlled, filtered venting for surface facilities is an attractive alternative to safety-motivated underground siting and is worthy of further investigation. As a concept, it is closely related to the augmented underground designs of this study. It is not, however, as broadly effec-tive as underground siting. o Remote siting is also a less effective alternative to safety-motivated underground siting. It is possible to reduce prompt adverse health effects due to extreme accidents through remote siting although long-term conse-quences would probably not be as significantly affected. o Accident consequences are extremely site dependent , showing great variation for the hypothetical California sites examined. o The anticipated level of earthquake-induced ground shaking for an underground plant is less than for a surface facility for a given earth-quake at comparable sites. However, the cost of underground siting, itself, was more than savings due to reduced design levels of ground shaking for underground plants. x
DRAFT o The consequences of a major release of radioactivity into groundwater are highly site dependent. For a wide range of realistic conditions, radio-nuclide travel times to an assumed release point one mile from the reactor site, ranged from 7 years to 4 million years. Groundwater contaminativn, while a locally severe problem, appears amenable to control measures. o Underground siting may lessen the attractiveness of a nuclear power plant as a target for sabotage since the plant can be effectively sealed from the outside. However, no definitive statements can be made pertaining to a reduced, or enhanced, resistance to sabotage, o No conclusions may be drawn concerning decommissioning of und2rground facilities, pending clarification of preferred decommissioning options. Recommendations The following are staff recommendations based on the findings of this study. They reflect the uncertainty of basic data in some areas and suggest means by which this uncertainty may be reduced. In addition, several concepts which have been identified as promising are suggested for further investigation. o Underground siting should not be mandated. This recommendation stems from (a) the uncertainty remaining over costs, construction time and possible licensing concerns; (b) the existence of what appear to be moderately effective and less expensive technical alternatives; and (c) the oppor-tunity to implement remote siting within California. o In view of the potential safety benefits of remote siting, the Commission should consider the relative safety benefits of alternative proposed sites under all accident conditions for nuclear plants proposed for California. o Further basic work should be sponsored, or actions taken to have such efforts sponsored by industry and the appropriate Federal agencies, in the following areas: o Evaluation of the concept of a passive " engineered failure p a thwa y" , possibly similar to the berm-containment Level-2 designs of this study, as applied to current generation surface facilities. o A detailed assessment of controlled, filtered venting for accident management. o Development of risk methodologies suitable for analysis and relative comparison of alternative reactor sites which reflect popu la tion , meteorology and topography. o Continued development and use of for=al techniques to identify major contributors to relative risk. o A program of continued safety improvements so that the estimated total level of risk for all currently operating reactors re=ains acceptable as new reactors become operational. This will require that the fundamental question of acceptable levels of technological risk be addressed; e.g., that the question of "how safe is safe enough" be answered. xi
DRAFT I&BLE OF CONTENTS PAGE II Foreword . . ... ... . ... . ...... . . . . . . . . * * . Executive Summary . . . . . . . . . . . . . . . . . . . . . . . . . . iv List of Figures . .. . . . ... .. . . . . . . . . . . . . . . . . xv xvii List of Tables ... .. .... ..... . . . . . . . . . . . .. 1.0 Introduction . ............. . . . . . . . . . . . . 1-1 1.1 Why Study Underground Siting? . . . . . . . . . . . .. . . 1-3 1.2 The " Necessity" for Underground Siting . . . . . . . . .. . 1-5 1.2.1 An Overview of Risk Analysis . . . . . .. . . .. . 1-6 1.2.2 Means to Reduce Risk .. . . . . . . . . . . .. .. 1-7 1.2.3 How Safe is Safe Enough? . . . . . . .. . . . . . . 1-8
- 1. 3 Ceneral Study Approach . . . . . . . . . . . . . . . . . . . 1-9 Study Reports . .. ... .. ... . . . . . . . . . . . . . ... 1-11 Study Participants ... . .. . . . . . . . . .. . ... . .. . 1-12 References . .. . . .. . .. ..... . . . . . . . . . . .. . 1-13 Concept Overview .. .. . .... . . . . . . . . . . . . . . . 2-1 2.0 2.1 Mined-Cavern Concepts ... . . .. . . . . . . . . .. . . 2-1 2.2 Berm-Contained Concepts ... . . . . . . . . . . . . . . . 2-4 2.3 A Review of the General Arguments for Underground Siting . . 2-4 2.3.1 Enhanced Seismic Performance . . . . . . . . . . . . 2-5 2.3.2 Improved Accident Mitigation Capability . . . . . . . 2-5 2.3.3 Possibility of Urban Siting . . . . . . . . . .. . 2-6 2.3.4 Improved Decommissioning . . . . . . . . . . . . . . 2-6 Study Guidelines . 2-7 2.4 . . . . . . . . . . . . . . . . . . . . .
2-10 a.5 Summary .. . .. . . .. . . . . . . . . . . . . . .. . .
.. . . .. . . ...... . . . .. . . . . . . 2-11 References .
3-1 3.0 Reactor Accidents . .. .. .. . . . . . . . .. . . . . . . .. . 3-1 3.1 Accident Classification . . . . . . . . . . . . . . . . .. 3-3 3.2 Probability and Accidents . . . . . . . .. . . . . ... .
- 3. 3 Design Considerations for Extreme Accidents . . . . . . .. 3-5 Containment Failure Modes .. . . . . . . . . . . . . .. . 3-7 3.4 3.5 Summary .. .... . ..... . . . . . . . . . . . . . . 3-11
. . . , .. . . .. . . . . . . . . . . . . . . . . . 3-12 References xii
DRAFT PAGE 4.0 Siting Implications . . . . .. . . . . ... . . . . .. . .. 4-1 4.1 Criteria . .. . . . ... . . .. . . ... . ... . .. . 4 4.2 Observations . . . . ... . . . . . . .. . .. .. ... . 4-2 4.3 Summary .. . .. . .... . .. . ... .. . . . . . .. 4-10 References .... . . . .. .. . . . . .... . ... ... . 4-11 5.0 Conceptual Plant Designs ... . . . . . . . .. .. ...... 5-1 5.1 General Considerations . . . . .. ..... . .. .. .. . 5-1 5.2 Berm-Contained Plant .... . .. . .... . ... .... 5-2 5.3 Accident Mitigation Systems in Berm-Contained Facilities . . 5-2 5.4 Mined-Cavern Plant . . ... . .. ... .. .... .... 5-4
- 5. 5 Accident Mitigation Systems for Mined-Cavern Facilities .. 5-8 5.6 Technical Feasibility of Controlled, Filtered Venting . .. 5-8 5.7 Summary .. . ... ...... .. .... ..... ... 5-10 References . .. . .. . ... . . . . ..... . ....... 5-11 6.0 Economic Implications of Underground Construction . . .. .... 6-1 6.1 Reference Surface Plant Costs .. . . .. .. ... .. .. 6-3 6.2 Undarground Plant Costs .. . ....... . . ... .. . 6-3 6.2.1 Baseline Berm-Contained Concept ... ... . .. . 6-3 6.2.2 Berm-Contained Concept Variations . . . ... . .. . 6-4 6.2.3 Mined-Cavern Concept . . . .... . . ... . ... 6-4 6.2.4 Mined-Cavern Concept Variations . . . . ... . .. . 6-8 6.3 Construction Schedules . . . . . . .. .. . . .... ... 6-8 6.4 Project Cash Flows . ... . . . ..... . ...... .. 6-10 6.5 Surface Plant Accident Mitigation System Costs . ... ... 6-10 6.6 Implications of Mature Industry Assumption . .... . ... 6-13 6.7 Variance in Estimated Costs . . .. .... . ... ... . 6-14 6.8 Impact of Escalation Rates on Relative Costs . .. ... . . 6-16 6.9 Summary .. . .. . .. .... . .. ... .... . ... 6-17 References ... . .. .... . .. . ..... . ....... 6-18 7.0 Consequences of a Major Radionuclide Release . .. ...... . 7-1 7.1 Consequence Model . . .. . .. . .. .. .. .... .. . ?-2 7.2 Criteria and Rationale . . . .. . ..... . ... ... . 7-2 7.2.1 Radionuclide Release .. . .. .. . ....... . 7-2 7.2.2 Exposure Pathways . . .. . .. .... .. . ... . 7-4 7.2.3 Dispersion, Transport and Deposition . . . . .. . . 7-4 7.2.4 Site Characteristics .. . ..... .. .. .. .. 7-5 7.2.5 Evacuation and Interdiction Criteria . .. ... . . 7-5 7.2.6 Health Ef fects Indices and Models . . . ... .. .. 7-5 7.2.7 Economic Costs .. . . . . ... . .. ... .. . . 7-7 7.2.8 Health Effects Nomenclature ... .. . .. .. . . . 7-8 viii
DRAFT PAGE
- 7. 3 Results of Consequence Calculations .. . . . . . . . . . . 7-9 7.3.1 Health Effects .. . . . . . . . . . . . . . . . . . 7-9 7.3.2 Economic Effects .. . . . . . . . . . . . . . . . . 7-11 7.4 Groundwater Contamination . . . . . . . . . . . . . . . . . 7-12 7.5 Summary .. . . . . . . .. . . . . . . . . . . . . . . . . 7-13 Reforences . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-14 8.0 The Ef fectiveness of Underground Siting . . . . . . . . . . . . . 8-1 8.1 Containment Failure Modes .. . . . . . . . . . . . . . . . 8-1 8.2 Division of Risk by Release Category . . . . . . . . . . . . 8-1 8.3 Mitigation of Specific Failure Modes . . . . . . . . . . . . 8-3 8.3.1 Steam Explosions .. . . . . . . . . . . . . . . . . 8-4 8.3.2 Overpressure . . . . . . . . . . . . . . . . . . . . 8-4 8.3.3 Foundation Melt-Through . . . . . . . . . . . . . . . 8-5 8.3., Loss-of-Isolation . . . . . . . . . . . . . . . . . . 8-5 8.3.5 Excessive Temperature . . . . . . . . . . . . . . . . 8-6 8.4 Failure of Pressure Relief Systems . . . . . . . . . . . . . 8-6 8.5 Summary of Accident Mitigation Capability . . . . . . . . . 8-6 8.6 Alternatives to Underground Siting . . . . . . . . . . . . . 8-7 8.6.1 Systems Hodifications . . . . . . . . . . . . . . . . 8-7 8.6.2 Siting Options . . . . . . . . . . . . . . . . . . . 8-8 8.7 Site Characteristics and Reaccor Risk .. . . . . . . . . . 8-9 8.7.1 Latent Fatalities and Site Characteristics . . . . . 8-9 8.7.2 Early Fatalities and Site Characteristics . . . . . . 8-12 8.7.3 Summary of Site Characteristics and Reactor Risk . . 8-15 8.8 Sabotage . . . . . . . .. . . . . . . . . . . . . . . . . . 8-15 8.9 Decommissioning. . . . . . . . . . . . . . . . . . . . . . . 8-19 8.10 Summary .. . . .. . . . . . . . . .. . . . . . . . . . . 8-21 References . . . . .. . . . . . . . . . . . . . . . . . . . . . 8-23 9.0 Options and Implications . . . . . . . . . . . . . . . . . . . . 9-1 9.1 Spectrum of Possible Decisions . . . . . . . . . . . . . . . 9-1 9.2 Safety Implications of Underground Siting . . . . . . . . . 9-2 9.3 Implications of Decision Options . . . . . . . . . . . . . . 9-2 9.4 Recommendations . . . .. . . . . . . . . . . . . . . . . . 9-4 Re fe rence s . . . . . . . . .. . . . . . . . . . . . . . . . . . 9-5 Appendix A - California Assembly Bill 2821 . . . . . . . . . . . . . . A-1 Appendix B - Underg ound Design Concept Specifications . . . . . . . . 3-1 xiv
DRAFT LIST OF FIGURES Page Fig. 2-1 Typical Mined-Cavern Concepts . . . . . . . . . . . . . . 2-2 Fig. 2-2 Typical Berm-Coutainment Concepts . . . . . . .. . . . . 2-3 Fig. 3-1 Frequency of Man-Caused Events Involving Fatalities . . . 3-6 Fig. 3-2 Possible LOCA Sequence with Massive LWR Core Melt . . . . 3-9 Fig. 3-3 Containment Pressure-Time Histories (ECCS Inoperable) . . 3-10 Fig. 4-1 Quaternary and Pliocene Unconsolidated Deposits . . . . 4-3 4 -4 Fig. 4-2 Mesozoic Granitic Rocks . .. . . . . . . . . . . . . .. Fig. 4-3 Quaternary Fault Zones ... . . . . . . . . . . . . . . 4-5 Population ....................... 4 -6 Fig. 4-4 4-7 Fig. 4-5 Presence of Water . .. . .. . . . . . . . . .. . . . . Fig. 4-6 Representative Areas - Berm Containment Siting . . . . . 4-8 Fig. 4-7 Representative Areas - Rock Siting . . . . ... . . . . 4-9 Fig. 5-1 Baseline Level-2 Berm-Contained Plant . . . . . . . . . . 5-3 Fig. 5-2 Cross Section of Level-3 Berm-Contained Plant Showing Filtered Vent Stack with Charcoal Iodine Filter . . . . . 5-5 Fig. 5-3 Mined-Cavern Underground Plant . . . . . ... . .. . . 5-6 Fig. 6-1 Plant Construction Schedule Summary . . . . . . . . . . . 6-9 Fig. 6-2 Normalized Berm-Contained and Reference Surf ace Plant Cash Flow Curves ..... . . . . . . . . . . . . . . . 6-11 Fig. 6-3 Normalized Mined-Cavern and Reference Surface Plant Cash Flow Curves ...... . . . . . . . . . . . . . . 6-12 Schematic Diagram of Consequence Calculations . . . . . . 7-3 Fig. 7-1 Fig. 8-1 Probability of Equaling or Exceeding Latent Fatalities of Number "X" for PWR Release Categories 1-8 . . . . . . 8-11 xv
DRAFT Page Fig. 8-2 Cumulative Population as a Function of Distance From Nuclear Plants . . . . . . . . . . . . . . . . . . . 8-13 Fig. 8-3 Probability of Equaling or Exceeding "X" Acute Fatalities for PWR Release Categories 1-3 . . . . . . . . 8-14 Fig. 8-4 Total Organ Doses "s. Distance from Reactor - Stability Category A ... . . . . . . . . . . . . . . . 8-16 Fig. 8-5 Total Organ Doses vs. Distance from Reactor - Stability Category F .................. 8-17 Fig. 8-6 Mortality Probability vs. Distance from Reactor - Stability Categories A and F . . . . . . . . . . . . . . . 8-18 xvi
DRAFT LIST OF TABLES Page Table 3-1 Reactor Facility Classification of Postulated Accidents and Occurrences . . .... . . . . . . . . . . . . . . . . 3-2 Table 3-2 Reactor Safety Study Accident Probabilities . . . . . . . . 3-4 Table 6-1 Previous Cost Estimates of Underground Nuclear Power Plants ....... . .. . . . . . . . . . . . . . 6-2 Table 6-2 Surface and Underground Nuclear Plant Cost Summary . . . . 6-5 Table 6-3 Berm-Containment Concept Cost Sensitivity Analyses . . .. 6-6 Table 6-4 Mined-Cavern Concept Cost Sensitivity Analyses . . . . .. 6-7 Table 6-5 Probable Plant Costs ................... 6-15 Table 6-6 Impact of Escalation Rate on Relative Cost . . . . . . . . 6-16 Fission Product Release Characteristics . . . . . . . . . . 7-4 Table 7-1 Table 7-2 Radioactive Dose Conversion Factors for Health Effects .. 7-6 Table 7-3 Accident Health Consequences .. . . . . . . . . . . .. . 7-9 Table 7-4 Economic Consequences . ................. 7-12 Table 8-1 PWR Release Categories - Dominant Sequence Description . 8-2 Table 8-2 BWR Release Categories - Dominant Sequence Description. . 8-2 Table 8-3 Risk by Individual Release Categories . . . . . . . . . . 8-3 Table 8-4 Relative Risk for Alternate Containment Designs Normalized to Current Surface Plants .. . . . . . . . . 8-8 xvii
DRAFT Underground Siting of Nuclear Power Reactors: An Option for California 1.0 Introduction in the fall of 1975, the California State Assembly Co=mittee on Resources, Land Use, and Energy conducted a series of hearings on issues related to the safety of nuclear power generation within the state. From these' hearings, three laws evolved which directed the California Energy Commission to exa=ine issues concerning nuclear vaste disposal, reprocessing of spent fuels, and, the subject of this report, underground
- siting of nuclear power reactors. Faced with contradictory evidence and testimony concerning the actual safety of large power reactors, the Legislature was motivated to require this study.
The work, which this document briefly suranarizes, is a thorough review of the feasibility and effectiveness of underground-sited nuclear power reactors. It is, perhaps, the most complex state-sponsored effort yet done in the general area of nuclear reactor ssfety. The study, performed in response to a specific enactment of the California Legislature, has moved beyond the bare technical questions of underground siting to a more fundamental level. In this light, underground siting represents an approach to safety, and embodies a philosophy which can achieve considerable reduction in risk. Although direct.ed primarily to th requirements of the mandating legislation, e.g., that the Cotsission s tudy ". . . the necessity for, and effectiveness and economic feasibility of.. ." (Public Resources Code 25524.3) underground nuclear power reactors, this work has considered underground siting as an alternative among many. In the area of accident analysis, this report deals primarily with the so-called core-melt or Class 9 accident. Recent years have seen considerable attention given to these unlikely, yet potentially large-consequence accidents. Under-ground siting has long been suggested as a means to make nuclear power reactors increasingly safe. Advocates argue that sub-surface siting has the ability to limit, or possibly eliminate, consequences to the public from core-celt accidents. As will be seen, there are several =eans by which consequences may be limited, among them underground siting. For - the purposes of this study, a core-melt accident was assumed. Such accidents are generally regarded as unlikely, yet physically possible. No independent evaluations were made to determine the probability of such events. The design concepts which evolved during the course of this study are intended to reduce the potential consequences of a core-melt accident under the as-su=ption that the accident has occurred. In other words, they attempt to mitigste consequences. The historical regulatory emphasis has been on crevention of severe accidents and towards keeping minor secidents from becc=ing major. TR two complimentary approaches are suggested for risk reduction:
*The term " underground" is used in a generic sense to include two concepts: a) construction of a reactor and related systems in caverns mined in rock; and b) so-called berm-containment wherein a structure housing a reactor would be covered with soil after construction. Both concepts were studied and will be more fully described in the following pages.
1-1
DRAFT prevention and mitigation. The latter philosophy, mitigation, can be imple-mented at several levels of effectiveness and at various costs. This study has found that underground siting, when coupled with a simple design modification, can reduce public consequences from the most extreme reactor accidents to negligible levels. Whether or not the investment in underground siting is warranted, however, is dependent on several complicating factors: the uncer-tainty over the actual level of risk posed by contemporary power reactors; the existence of less effective, but less expensive technical alternatives to underground siting; and, the possibility that, under some circumstances, remote siting may so alter the consequence characteristics of extreme accidents as to make the residual risk acceptable. The Commission will have the opportunity to balance these considerations and alternatives for, at the completion of the study, the Commission must decide whether or not to require underground siting of nuclear power reactors. The California Legislature has reserved a one year review period following comple-tion of the study in order to review the findings. This report summarizes the contract work sponsored by the Energy Commission, as well as staff efforts, and presents staff recommendations pertaining to underground siting of nuclear power reactors. As such, this document draws extensively from reports prepared by Commission contractors,* and is not intended to be overly technical. (A list of fourteen available s tudy-rela ted reports follows Section 1.) Project documentation is too voluminous to be bcund as a single report. The interested reader is referred, as appropriate, to the other documents. The reader must bear in mind that this project has been primarily a study of underground siting of nuclear power reactors. It is not to be construed as an indictment of previous reactor safety studies nor is it an exhaustive risk analysis of current reactor technology. It is not a treatis.e on the costs of surface-sited nuclear reactors. It is directed toward one end: to provide a base of information for consideration by the Commission. With that objective, certain considerations are treated quantitatively while others are & cussed salitatively. The s truc tu re of this report generally follows the legislative directives. The first portion discusses the concepts in a general sense and outlines the various motivations for sub-surface siting. The basic groundrules, study objectives, and areas wherein this effort has departed from similar work, are noted. The next part presents the technical requirements for underground concepts including siting in California; designs evolved in the course of this work; and assesses their general regulatory and environmental implications. This section also sets forth cost data for underground concepts developed specifically with respect to California siting requirements.
*The full complement of study related material is available from the Commis-sion Publications Unit by either written or telephone request: California Energy Commis s ion , Publications Unit, MS-23, 1111 Howe Avenue, Sacramento, California 95825; telephone 800-852-7516 within California, 916-322-3725 outside California. See p. 1-11 for a listing of this study material.
1-2
DRAFT The third division deals with the effectiveness of underground siting prin-cipally in terms of accident mitigation, sabotage, and decommissioning. The implications of the failure of sub-surface siting as a safety ' concept and the efficacy of alternatives are also considered. The final section presents and discusses various decision alternatives. The principal study conclusions and staff recommendations are contained in the Executive Summary bound with this report. 1.1 Why Study Underground Siting? Not only has the California Legislature had an interest in underground siting, but many others have as well. Organizations and individuals within the sci-entific community urging a serious examination of this siting concept have ranged from advocates for, to groups opposed to, the further development of nuclear power. Their reasons are varied and give some insight into concerns over nuclear power. The American Physical Society's study group on light-water reactor safety reviewed many issues surrounding nuclear power. Among their major suggestions was that a "... careful assessment should also be made of the benefits and costs of altery,tgvp siting policies, such as remote, underground, and nuclear-park This recommendation stems in part from the stated opinion that siting." the statistical tools used to establish levels of risk in recent reactor safety analyses, while valuable for assessing relative strengths and weaknesses of reactor systems, cannot presently establish absolute levels of risk. Thus, a general inability to determine levels of risk coupled with, in their opinion, remaining uncertainties regarding the efficacy of reactor safety systems and other factors, prompted a recommendation for a careful study of alternative siting policies, including underground concepts. A long standing proponent of nuclear power, Dr.~ Edward Teller, in his testimony before the California State Assembly Committee on Resources, Land Use, and Energy, argued that !.h e study of underground siting would be an element in continuing efforts to enhance reactor safety:
"The safety (of nuclear power] as it exists now is great. We must continue to increase it. . . . What the best way is I don't know.... I would like to see an engineering study...to see 1 ow much more expen-sive it is and whether it is at all more expensive to put a reactor underground and to supply it with enough safety so that nothing is at risk except the reactor itself." (1-2)
His testimony also suggested that underground siting, if effective, might erl.ance the insurability of power reactors to the extent that federal subsidy and limitations on liability would not be needed.
- Numbers in parentheses denote references listed at the end of each chapter.
1-3
DlAFT President Carter, a nuclear engineer by training (,7 gile a candidate, in a July 11, 1976 address to the Washington Press Club, suggested several alter-natives for increasing the safety of nuclear power plants:
"For instance, nuclear reactors should be located below ground level.
The power plants should be housed in sealed buildings within which permanent heavy vacuums are maintained. Plants should be located in sparsely populated areas and only after consultation with state and local officials." Although it is not clear if "below ground level" was meant to imply fully underground, it is commonly interpreted as such. Several points raised by then candidate Carter were the subject of a Petition for Rulemaking recently decided by the United States Nuclear Regulatory Commission. The Commission denied the petition thereby not requiring underground siting in all future cases. The Ford Found a recent analysis of issues pertaining to the nuclear debate. gig sponsored The authors of that report stated:
"On balance, we have concluded that the risks associated with nuclear accidents are acceptable since the predicted average rate-of-loss due to nuclear accidents compare favorably to those associated with competing fossil fuel technology...."
This somewhat optimistic conclusion was, however, significantly qualified:
"However, the number of reactors is expected to grow. Even though the benefits of nuclear power will increase correspondingly, the likeli-hood of occurrence of an extremely serious accident should not also be allowed to increase correspondingly. Therefore, there should be a continuing effort to imnrove actual reactor safety." [ Emphasis added]
The study group, as one method for reducing risk, suggested:
"A more restrictive siting policy would increase somewhat the costs of nuclear power in sote locations, but we balieve it is warranted by the uncertainties in the probabilities of accidents and by the large risk raductions that are possible. Special measures such as underground siting should be considered if nuclear power is to be sited at high risk iocations."
A similar suggestion was offered by Yellin(1-5) in his review of the Nuclear Regulatory Commis sion's Reactor Safety Study. Yellin was struck, as was the Ford / MITRE group, by the wide range of consequences possible for the same assumed accident at various reactor locations. Using prompt fatalities as a measure, he found consequences to va.y from approximately one to over 1,000, three orders of magnitude variation. He concluded: 1-4
DRAFT "The discussion above underlines the need for detailed nuclear Furthermore, accident consequence analysis on a site-specific basis. the evident large potential accident consequences and the striking risk differences between existing sites suggest that remote nuclear siting policies merit further NRC consideration, and that serious study of the advantages and disadvantages of underground siting is warranted." Underground siting uay represent one plateau of acceptability to some organizations now opposed to nuclear power:
"The Sierra Club opposes the licensing, construction, and operation of new nuclear reactors pending... resolution of the significant safety problems inherent in reactor operation, disposal of spent fuel, and possible diversion of nuclear material capable of use in weapons manu f a c tur e . . . ."( 1-6 )
A central theme to all of the suggestions for a serious study of underground siting has been reactor safety. Whila sub-surf ace siting has no impact upon the issue of waste disposal, or many aspects of possible diversion of nuclear material, it is an effective means to reduce public risk from extreme reactor accidents. Answering the questions of at what price, and with what necessity in light of alternatives, is the purpose of this report. 1.2 The " Necessity" for Underground Siting Later report sections will discuss the general concepts, costs and effectiveness of underground siting and discuss alternatives. This sub-section, however, provides background for considering the " necessity for underground siting". Necessity may be compelled by many factors. The preceding Section suggested however, that the most frequently cited reason for underground siting is for safety augmentation. The general argument is that the overlying soil or rock provides very great passive containment in the event of a severe reactor accident. In addition, if fission products were to escape into the soil or rock, the low migration rates through these media would retard their release to the surface biosphere. The net effect, it is argued, would be greatly reduced accident consequences to the public. To address safety one must first define the idea and consider a very related concept, risk. Risk, in the context of reactor safety, to borrow Lowrance's definition, is ". ..a measure of the probability and severity of harm to human "a thing is safe if its risks are judged to be accept-health able."g'Qurthermore ,From these few words, many observations may be drawn. First, certainly no technology is completely safe, absolutely free from risk. Second, to evaluate risk, one must establish the numerical measure of the likelihood of occurrence, the prob ability , of adverse events and combine them 1-5
DRAFT with measures of the severity of the event. Third, the determination of safety involves human judgement about acceptability. Thus, the answer to the question "how safe is safe enough?" lies in societal acceptance of the nature and cbtracteristics of the risk. A risk analysis cay provide useful information pertaining to safety but it cannot, by itself, determine "how safe is safe enough". 1.2.1 An Overview of Risk Analysis There are two broad approaches to establishing levels of risks. The first is to evaluate, statistically, historical data. For example, on the average about 55,000 traffic fatalities occur each year distributed over a U. S. population of approximately 220 million. A simple division g fatalities by total population yields an annual idividual risk of about 3x10 . Neglecting the consideration that traffic fatality risk is not evenly disgributed over the population, since not all persons use automobiles, the 3x10 figure (or, alternatively, three chances out of 10,000), represents the likelihood of an individual becoming a traffic fatality in a period of one year. Key to the matter is that the pre-ceding calculation has been based on historical data, facts, matters of record and public knowledge. The validity of the basic data is unquestioned. When historical data is lacking from which to derive measures of reliability, or a from which to explicitly calculate risk, an alternative (ggroach must be used. Such a method was used in the Reactor Safety Study In that effort, techniques in use for reliability analysis of comple: systems were applied to predict the likelihood and consequences of severe reactor accidents. This methodology is termed risk analysis. There are two principal components to a risk analysis: o identification of all significant adverse events (e.g., types of accidents) and calculation of a numerical probability for each; and, o determination of the consequences of each adverse event. The final risk is determined from che sum of the products of each accident probability with its associated consequences. Risk analysis as employed in the Reac tor Safety Study is an outgrowth of reliability analysis techniques in use for a considerable time. There is a fundamental difference, however. That difference lies in predic tions of absolute risk versus relative contributions to risk. Generally, reliability analysis attempts to refine complex systems by identifying principal contri-butors to failure and suggesting modifications to improve reliability. The Reactor Safety Study has attempted to predict risk in an absolute framework so that risks fras nuclear power facilities may be compared to other sources of risk, both to society and to individuals. That reliability analysis is a useful tool is not questioned here. That risk analysis, in its present fo rm , can accurately estimate risk from complex technologies has been debated in many forums including hearings before the Subcommittee on Energy and the Environment of the U4.gHouse of Re p r e s e n t a t iv e s ' Committee on Interior and Insular Affairs At this time, whether or not risk analysis can accurately predict future risk, must be considered an open question. 1-6
DRAFT Fundamental to the debate over the ability of risk analysis to accurately predict absolute levels of risk is the choice of methods to enhance reactor safety. A common recommendation of most reviewers of the Reactor Safety Study is that we avoid complacency; that we continue to improve the reliability of safety systems through efforts such as risk analysis and consider more than only impovements to mechanical systems. An emerging trend seems to be increased Recent consideration of extreme accidents such as a reactor core-meltdown. the American Physical summary gn47ses1 of reactor safety, such as t Society and the Ford / MITRE study group,g)ofhave, by the nature of their recommendations for underground and remote. siting, and consideration of severe reactor accidents on a site by site basis, conveyed a feeling of cautious regard for predictions of reactor safety. In this context, underground siting may prove the criteria of acceptability, for, as will be shown, it has tha potential for almost total mitigation of consequences to the public from severe reactor accidents. 1.2.2 Means To Reduce Risk There are two paths by which total risk may be reduced: either the probability of a particular event may be lessened through improved, more reliable or more effective safety systems, or the potential consequences may be lessened. It does not Underground siting is a technique for consequence reduct. ion. directly address the probability with which an extreme accident might occur; it focuses on controlling the outcome under the assumption that an accident might occur, irrespective of its probability. The two possible approaches to risk reduction, decreasing the probability of extreme accidents and limiting potential consequences, are not necessarily mutually exclusive. Attitudes, however, seem divided by these approaches. Neither is totally adequate. Accident probability calculations for nuclear Until power plants are necessarily imperfect, but they do provide guidance. verified by an extensive operating history (something substantially more than the approximately 300+ reactor years accumulated to date), the confidence which In this light, may be placed in probabilistic estimates of risk is limited. very low accident probabilities do not imply absolute protection against accidents in any given period. Even though the estimates of accident frequency may suggest a probability of say one in a billion years, this does not guarantee that it will be a billion years before an accident occurs, or even that it will be a billion years between two such accidents. No matter how reliable calcula-for deter-tions demonstrate safety systems to be, they are an incomplete base mining absolute safety. Conversely, a strategy of risk reduction through limitation of potential consequences would tend always toward ' olated siting for surface facilities. Studies, including the present investigacten, indicate that there will always be There does not appear to be a public at risk no matter how isolated the site. site so remote as to be free of public consequences in the event of an extreme accident. The re' .ve insensitivity of some forms of accident consequences to How then, does one establish the limits of location is discussed later. acceptability? The prudent path is neither total reliance upon numerical prognostications of there is a middle hazard nor Cassandra-like concentration on extreme accidents: 1-7
DRAFT ground. A combination of the philosophies of prevention and mitigation would appear to be a most effective approach to risk reduction. 1.2.3 How Safe is Safe Enough? Central to the issue of reactor safety is a recognition of what is meant by safety. As a society, we declare something to be adequately safe if we accept the risks. It is significant that social acceptability does not imply indivi-dual acceptance. Thus, socially, airplane travel is used extensively while some persons refuse to fly; individuals may accept the risks of smoking while the Federal government embarks on a social program to reduce cigarette usage. That which is safe is determined on both a social and personal level. Such decisions are complex in both regimes. Slovic has commente igp0yhe difficulty of assessing the public acceptability of He notes that "... intelligent people systematically technological risk violate the principles of rational decision making when judging probabilities, making predic tions , or otherwise attempting to cope with probabilist .c tasks." 4 To a large extent the departura fram 5:tional decision making" is due to the nature of the risks being svaluated. A convenient terminology has been coined to broadly characterize the distinction between those who are satisfied with probabilistic treatment of accident risks and those who are not. Zivi and T a- have termed the Actuarial View as one of sati with calculation of risks through probabilities and conse-quences.fagonAlthough the methodology may require wide error bands, funda-mentally the Actuarialist accepts the results of probabilistic fault tree analyses when they suggest that high consequence events have low probabilities. Counterpoint to the Actuarialist is the Catastrophic View. The latter perspec-tive is built on a belief that the worst will happen, that a severe reactor accident is inevitable, and that decisions regarding the acceptability of a technology must reflect such events. Society at large tends to adhere to the Catastrophic view. Thus, dam failures, earthquakes, floods, airplane accidents, and major fires receive much attention, while sources of greater mortality such as heart disease and automobile accidents go almost unnoticed. To quote Lowrance:
"...the social and political impact of a single catastopha affecting' many people at one time is usually greater than that of a chronic hazard affecting :he same number of people just as sericasly but over a long period."
The dichotomy of the Actuarial and Catastrophic view complicates the discussion on nuclear power. Each view has its adherents with a spectrum of acceptable consequence limits and each is supportable. At the extree , adherents of each, however, view the same event, from irreconcilable perspect.ves. Consider , for a moment, ef forts to reduce risk. As noted by Zivi and Epler, an Actuarialist would concentrate on efforts to reduce probability by improving the 1-8
DRAFT reliability of various plant systems. A Catastrophist, on the other hand, would concentrate on evacuation procedures, or perhaps siting policies, to achieve the same result: a reduction in the product of probability and consequence. This section has attempted to demonstrate that perception of risk is a funda-mental element in determining how safe is safe enough. The question of whether or not nuclear power plants should be placed underground, or even if the tech-nology should remain an option, may very well hinge on the issue of perception. The determination of risk is an appropriate scientific endeavor. The judgement as to the acceptability of risk transcends the technologist and must involve society at large. T Reactor Safety Study:ysgoint was well stated in the closing paragraph of the "The question of what level of risk from nuclear accidents should be accepted by society has not been addressed in this study [RSS]. It will take consideration by a broader segment of society than that involved in this study [RSS] to determine what level of nuclear power plant risks should be acceptable. This study (RSS] should be of some help in these considerations." [ bracketed material added.] 1.3 General Study Approach This study was not intended to be g ric evaluation of reactor risk in the manner of the Reactor Safety Study For the evaluative purposes of this study, the events and conditions necessary to result in a melt-down of the reactor core were assumed to have occurred. Such an accident is generally regarded as highly unlikely; however, there is disagreement over actual prob-abilities, and over what value of probability should be considered acceptable. No efforts were made to quantify, in an absolute sense, the likelihood of a reactor core meltdown; although, whether or not such an accident was rela-tively more or less likely in an underground plant than in a surface plant was considered. To evaluate the feasibility of underground concepts, designs were developed by experienced engineering organizations and cost estimates were prepared. The designs reflect the current regulatory requirements for surface nuclear power plants, the space requirements for normal operation and maintenance, and the realities of construction processes and techniques. Cost estimates were prepared using the Nuclear Regulatory Commission code of accounts, and, in general, are quite detailed. In addition, the cost estimates consider the recognized fact that "first-of-a-kind", or prototypes, are more expensive than
" mature industry" versions.
The design concepts of this study depart, in at least one major respect, from previous studies of underground siting. They utilize an " engineered failure pathway" for management of temperature and pressures generated by accident conditions. Fundamentally, they do not attempt to prevent an accident, but, instead, to control the manner of containment failure and in so doing, greatly reduce the potential for public consequences. The engineered failure pathway is alternatively called an Accident Mitigation System. 1-9
DlAFT ~ A spectrum of hypothetical reactor accidents was examined to find those se-c.uences which generated the extremes of physical parameters such as pressure, rate of pressure rise, tempera ture , and rate of energy delivery. The Accident Mitigation Systems were conceived and designed on the basis of an envelope of extreme conditions. To evaluate the effectiveness of the augmented underground designs of this study, radioactive material was assumed to be released from containment for both a surface and underground plant. For the underground plant, the release was to the Accident Mitigation System. This resulted in either forced travel through the overlying soil, in the case of the berm plant, or, for the rock plant, in release to special expansion / condensation chambers. When radioactive material reached the ground surface, it was assumed to be transported through the atmosphere. The consecuences of the radiological release were estimated using current data with conventional health effects models. Accident consequences calculated included early and latent health fatalities, and other effects such as cancers, property damage, and evacuation and relocation costs. They should be regarded only as estimates since the state-of-the-art for such calculations is inadequate to provide precise values. Evaluating and determining the necessity for underground siting requires consideration of the limitations of current methods to estimate risk, the process by which safety is determined, and the effectiveness and cost of alternatives. Balancing these factors is a most difficult process, one which ultimately requires subjective consideration of t't e nature and perception of risk, in general, and reactor risk, in particular. The following pages present design concepts and associated costs, measures of effectiveness, a presentation and limited assessment of alternatives , and a brief discussion of possible decisions pertaining to the " necessity for" underground siting. 1-10
DRAFT STUDY REPORTS Conceptual Design and Estimated Cost of Nuclear Power Plants in Mined Caverns, Underground Design Consultants, January, 1978. Conceptual Design and Es timated Cost of Buried " Berm Contained" Nuclear Power Plants, S&L Sugineers, January, 1978. Armistead, R. A., et al., Analysis of Public Consequences From Postulated Severe Accident Sequences in Underground Nuclear Power Plants, Advanced Research and Applications Corporation, December, 1977. Environmental Impacts of Underground Nuclear Power Plant, Environmental Science Associates, Inc., September, 1977, (Prepared for Advanced Research and Applications Corporation). Seismic Assess.ent of Underground and Buried Nuclear Power Plants, Applied Nucleonics Co.. September, 1977, (Prepared for The Aerospace Corporation). Alternatives to Underground Nuclear Power Plant Siting, The Aerospace Corporation, R. port No. ATR-77(7652)-1, April, 1977. Systems Management Support for ERCDC Study of Undergrounding and Berm Containment - Interim Report, The Aerospace Corporation, August, 1977. Anglin, Richard L., Jr., Syd J. Ireland, and Martin Goldsmith, Analysis of Federal Regulations Governing Siting and Construction of Nuclear Power Plants as Applied to Underground or Berm Containment, Jet Propulsion Laboratory, Document No. 5030-88, September, 1977. Goldsmith, Martin G., Ed., Literature Survey of Selected Topic? Pertinent to Underground Siting of Nuclear Power Plants, Jet Propulsion Laboracory, Document No. 5030-83, May, 1977. Miles, Ralph F., Jr. , and Martin Goldsmith, Decision Analysis Framework for the Undete,round Siting of Nuclear lawer Plants, Jet Propulsion Laboratory, Document No. 5030-92, May, 1977. Evaluation of the Feasibility, Economic Impact, and Effectiveness of Under-ground Nuclear Power Plants, The Aerospace Corporation, May, 1978. Smith, Je f f rey H., Ralph F. Miles, and Martin Goldsmith, An Application of Multi-Attribute Decision Theory to The Underground Siting of Nuclear Power Plants, Jet Propulsiot. Laboratory, Document No. 5030-224, March, 1978. Terasawa, K., R. O'Toole, and M. Goldsmith, Probabilistic Analysis of the Cost for Surf ace-Sited and Underground Nuclear Power Plants, Jet Propulsion Laboratory, Document No. 5030-223, March, 1978. Terasawa, K., R. O'Toole, and M. Goldsmith, Methodology for the Estimation of Cost of Underground Nuclear Power Plants, Jet Propulsion Laboratory, May 1977. 1-11
DRAFT STUDY PARTICIPANTS The Aerospace Corporation The Jet Propulsion Laboratory Advanced Research and Applications Corporation Intermountain Technology, Inc. Environmental Science Associates, Inc. Underground Design Consultants Gibbs and Hill, Inc. S&L Engineers (affiliate of Sargent and Lundy) And Individuals From: Stanford Research Institute Lawrence Berkeley Laboratory University of California, at Santa Cruz l-12
DRAFT REFERENCES SECTION 1 1-1 Reviews of Modern Physics , Report to the APS by the Study Group on Light-Water Reactor Safety, American Physical Society, Vol. 47, No. 1, Summer 1975. 1-2 Hearings on Proposition 15, Volume III: " Technological Concerns in Nuclear Reactor Safety, Part I," California State Assembly Committee on Resources, Land Use, and Energy, Sacramento, California, October 1975. 1-3 After Petition for Rulemaking, PRM 50-19, United States Nuclear Regulatory Commission, Federal Register Notice, May 1977. 1-4 Nuclear Power Issues and Choices, Keeny, S.M., et al., Ford Foundation / MITRE Corporation, Ballinger , Cambridge, Massachusetts ,1977. 1-5 Yellin, J., "The Nuclear Regulatory Commission's Reactor Safety Study," Review, The Bell Journal of Economics, Vol. 7, No. 1, 1976. 1-6 The Sierra Club and Nuclear Power, The Sierra Club, San Francisco, Cali-tornia, September 1976. Of Acceptable Risk, William Kaufman, Inc., Los Altos, 1-7 Lovrance, W.W., California, 1976. 1-8 Reactor Safety Study - An Assessment of Accident Risks in U.S. Commercial Nuclear Power Plant, WASH-1400, NUREG-75/014, October 1975. 1-9 Oversight Hearing Before the Subconsnittee on Energy and the Environment of the Committee on Interior and Insular Affairs, House of Representa-tives, July 11, 1976, Serial No. 94-61, U.S. Government Printing Office, Washington, D.C. 1976 1-10 " Psychological Determinants of Perceived and Acceptable Risk: Implications for Nuclear Waste Management ," Conference of Public Policy Issues in Nuclear Waste Management, Chicago, Ill., October 1976. 1-11 Zivi, S.M., and E.P. Epler, Enchancing Public Acceptance of Nuclear Energy by Improving Reactor Safety Systems, Institute for Energy Analysis, Oak Ridge Associated Universities, Oak Ridge, Tennessee, November 1977. 1-13
DRAFT 2.0 Concept Overview Before discussing the motivations, benefits and disadvantages of underground siting, the concepts must be clearly in mind. There is no universal concept for an underground plant. This causes difficulty because total project costs are related to what is built underground. Very of ten, arguments for and against underground siting depend on details. Different plant concepts have evolved from dif ferent motivations. A discussion of the merits and problems of under-ground siting must be keyed to a particular concept. This section will briefly describe the various concepts included within the category of underground-sited nuclear power plants. In general, there are two variations on the underground concept, each of which goes by a variety of names. The historical concept of underground siting involves the mining of caverns in rock and subsequent construction of the power plant within the excavated volume. The four nuclear power plants which have been built underground are representative of this approach. The concept has been variously termed rock or mined-cavern siting. Modern construction techniques, with the ability to excavate large pits and move tremendous volumes of soil, have given rise to the second general concept, herein termed berm-containment, but also called cut-and-cover, cut-and-fill, pit siting, or mounded. Each general concept has unique benefits and drawbacks and one must be careful in any discussion to define what is meant by underground siting. Siting requirements differ greatly between the rock and berm-contained concepts. The former requ.res rock of high quality with high compressive strength, few fractures, and of large areal extent, to ensure ease of construction and long-term cavern stability. Few areas of the United States possess rock masses of the requisite quality, although some were identified within California. The berm-containment concept requires earth material that can be easily excavated such as soil. Thus, a site suitable for a rock concept would be undesirable, in general, for a berm concept, and conversely. A generalization often made about underground siting is that it cannot be implemented in most areas of the United States. This is true only for the rock siting concepts and is not accurate for the' berm configurations. Within this report, the term " underground" will be used in a generic sense to include both the rock and soil concepts. Where a conclusion or point applies to only one concept, the more restrictive terminology " berm-containment," " mined-cavern," or " rock' sited" will be used. 2.1 Mined-Cavern Concepts Figures 2-1 and 2-2 show two general variations on rock siting and three alternative berm-containment concepts, respectively. Considering the rock concepts first, an important distinction based on topography must be noted. In Figure 2-1, the drawing labeled " Horizontal Entry" depicts the facility as constructed within a hill. Access to the plant is essentially horizontal, with main transport and cooling water tunnels nearly flat. Although the turbine generator is shown underground, it could be sited outside the rock. Several variations on the location of major plant systems were considered in this study and will be described later. 2-1
HORIZONTAL ENTRY VERTICAL ENTRY Relay / Switching Nuclear Reactor - and Control Auxiliaries s9-
,y; .y \
u L ' ,/, )& // ..
- a QN' 1,
/
, ,/ '
ECCS Nuclear Turbine ee Relays Switching i afr Figure 2-1 TYPICAL MINED-CAVERN CONCEPTS DRAFT
CUT AND COVER HILLSIDE COVER ONtY TURBINE CENERATOR S0lt REACTOR AUX 1tl ARIES k 8ACKFill TURBiff CENERATOR ACM SOIL TURBINE GENERATOR
\ \ AUXttlARIES / T:
MOUND REACTOR AUXitIARIES \ \ f 50tL COVER . cN# # i' , i.; O.,. ? Y.# r. ,U$ P 'l COVER a i e .
. < ie L jyg < . . ;
grac-. -Q Jt874?d: "S 7 w , ., .:,
@.B.a .: F..
fd Q ';> ., ,. , , %'? F . . .,,qR>)$tA
'. WT' i.s By.. 'dr$i i,4if;y'Es' 5??
.; lr4 ..z!F/.k * -iy ....' Qf -
. ,v:,;.; P W e, r V - .t
't;g*kpp7;2 ~ {pQ.c,.,, p,<vp.w i, +&g.:.dy' 3,g, .,.
6
- l' e,'pp:3.+.g:rr:,i W d u
"i.6+~ct~.g:.r;4y;.gs;.24n,.. *.i?$$hr
* . %~.
Figure 2-2 TYPICAL BERM-CONTAINMENT CONCEPTS DRAFT
DRAFT The second rock concept, also amenable to a surface turbine, requires vertical access to the major plant components. Such concepts have been studied in detail by Swedish investigators. For the sites studied, the littoral rocky plateau was well suited to a vertical access rock concept. The vertical access concept, however, has been criticized as being prone to flooding. Since many plant caverns would be below sea-level in such layouts , the possi-bility of flooding by ocean waters must be considered. If the turbine generator were also located below ground and below sea-level, turbine condensers could be subject to high hydrostatic heads, further aggravating the possibility of plant flooding. Of significance is the common association in the literature of underground facilities with an ' increased hazard of plant. flooding. This is strictly true for only one underground concept and not a general characteristic of sub-surface siting. For this reason, an analysis of the general merit of underground nuclear power plants is closely coupled to the specific concept under consideration. 2.2 Berm-Contained Concepts Figure 2-2 shows three variations on berm-contained nuclear power plants. Observations similar to those made for horizontal and vertical access mined cavern plants could be drawn for the " berm" and " cut-and-fill" concepts. The hillside concept attempts to minimize the volume of earth material excavated and subsequently replaced. The soil-covered concepts, in general, use the same construction procedure: an excavation would be made, the plant constructed, and the major structures then covered with soil. Procedures would be essentially similar to conventional surface plant construction and access during construc-tion would also be similar.
- 2. 3 A Review of the General Arguments for Underground Siting Generally, recommendations for sub-surface siting stem from a perceived ad-vantage relative to a surface plant. Examples of such debated advantages include:
o Enhanced seismic performance; o Improved accident mitigation capability; o Possibility of urban siting; o Inherent protection from sabotage; o Psychological benefits - the so-called "out-of-sight, out-of-mind" syndrome; o Improved decommissioning; o War protection; and o Isolation from external hazards. 2-4
DRAFT The list above is certainly incomplete. It is offered to clarify an important point; that is, recommendations for underground siting must be analyzed in light of the motivation (s) for the recommendations. Each perceived benefit must be weighed carefully and cannot be separated from a geographical context. Argu-ments quite valid in the densely populated northeastern United States, for example, may not apply to California, and conversely. Furthermore, the dis-advantages of underground siting must be balanced carefully against the dominant motivation. Each of the motivations listed above is discussed in the following pages and its applicability to California assessed. The objective will be to narrow the field of concern so that principal advantages may be later viewed in light of general disadvantages. 2.3.1 Enhanced Seismic Performance A number of investigators have suggested that underground siting may ' have seismic advantages over a comparable surface installation. (See for example, 2-1, 2-2) The impetus is generally one of reduced cost through reduced design requirements. The cost of providing seismically resistant design for a surface plant can be significant. An early project effort included a review of the adequacy of analytical methods, design criteria, and general advantagggd disadvantage 3 of underground nuclear plants with respect to earthquakes. Although this work concluded that ground motions for buried installations were reduced, it was clear that the general cost impact could not be evaluated until designs were completed. Many earlier reviews of underground siting have based their optimistic assess-ment of seismic bene fits on the response of plant equipment. However, since the principal contributors to the cost of seismically resistant design are the massive cone: ate structures, not the plant equipment, the foundation for the conclusion is poor. In particular, for the berm-contained plant, structures must be massive to resist not only the ever present weight of the overlying soil, but also dynamic soil loads wnich would be generated during an earth-quake. In the rock plant , since the rock itself functions as the major plant
" structure", economy can be achieved by using cavern walls to reduce earthquake-induced motions and loads. Thus, the net cost impact was determined during design evaluation.
2.3.2 Improved Accident Mitigation capability In general, two arguments are offered for enhanced accident mitigation from undtrground plants. First, that overlying soil or rock creatly increases the ability of a containment structure to withstand accident induced pres-sures. The question of ,thether or not the containment structure is adequate to ensure complete entrapment of fission products following a severe accident is rarely posed, but this adequacy was considered in detail during the course of this study. A second argument for increased accident management, often associated with soil-covered concepts, is that the inherent filtering capacity greatly retard and reduce post-accident fission product migration.g_gil would A crit _yconsideration, which is often ignored,appropriately, is discussed confinement, by Allensworth, of et al To achieve containment, or more 2-5
DRAFT fission products, whether it be within the plant itself or the sub-surface media, sudden failure of isolation systems must be prevented. If these systems, such as closure valves on surface routed ventilation shafts and piping, and seal doors on equipment and personnel tunnels, fail to function, a release of fission products to the biosphere could occur, possibly comparable in magnitude to some surface plant releases. The containment designs, which have evolved in the course of this study, attempt to use the underground environment to advantage. Engineered, passive systems were developed which capitalize on the filtering effectiveness of soil and, to a limited extent, on the enhanced pressure resistance offered by the overburden material. These designs will be discussed in detail, subsequently. 2.3.3 Possibility of Urban Siting Underground siting has been suggested as a method by which power reactors could be placed closer to population centers. Implicit in current surface plant siting policies is the use of distance from major population centers as an added safety dimension. Underground siting has been suggested as an alternative to remoteness, reflecting a trade of surface distance for added confinement. Of course, as Crowley, et al., note, the effectiveness of underground siting as a safety augmentation y must be unequivocally demonstrated before urban siting is realistic.g4 stem The urban siting motivation may also have great weight in those countries with relatively high population densities. In some countries , there may be no such thing as a " remote site." The possibility and opportunity for urban siting was investigated with specific reference to California conditions, e.g., urban regions, areas of seismic concern, and geologic considerations. 2.3.4 Improved Decommissioning The necessity to decommission and decontaminate a nuclear facility at the end of its operational lifetime must be considered in any siting decision. Underground siting has been suggested as a possible technique to facilitate decommissioning particularly if entombment were the selected option. The massive underground structures could serve as a ready-made vault. However, uncertainty exists regarding suitable methods and procedures. The general impact of underground siting on each of the four recognized means of decommissioning will be discussed in Section 8.9, and briefly presented here. Sub-surface containment may lend itself to decoc:missioning in a variety a f. ways Underground siting, by isolating the more radioactive portions of the pl- could facilitate conversion to an alternate heat source, especially with sur ace-sited turbine generators. In addition, the sub-surface containment might provide a ready made entombment for low-level waste and contaminated equipment. However, underground siting may complicate total removal of contam-inated material, due to limited accessibility, if such were the selected decoc=tissioning option. The ability of underground siting to facilitate decommissioning is very depen-dent on the selected option. Decommissioning, however, was not explicitly 2-6
DRAFT evaluated for the design concepts of this study but is discussed, in general, in a later section. 2.4 Study Guidelines This study has examined two underground siting concepts: (a) siting in mined-caverns using the horizontal entry or " hillside" concept; and (b) a berm-contained configuration in which plant foundation levels have been adjusted to balance the volume of excavated materials with that required to later cover the plant structures. The designs and subsequent cost estimates reflect conditions representative of California. For example, closed cycle cooling systems were assumed to reflect the trend in currently proposed nuclear plants within the state. The seismic design levels are very high in relation to other portions of the United States. Rock descriptions which form the basis for the mined-cavern designs and estimates were developed from actual rock corings made in certain areas of the state. Construction costs assumed labor rates currently prevailing within a certain region of California. While many study results may be valid for other areas, this work has reflected the realities of power plant siting within California, and strictly, applies only to this state. As noted previously, the benefits and disadvantages of underground siting are, in many respects, coupled to specific containment concepts. Two examples have been given, that of flooding hazard and site availability, which are extremely concept dependent and often advanced as arguments against underground siting. Similarly many conclusions regarding the effectiveness of underground siting for mitigating the consequences of extreme accidents are valid only with respect to the concepts studied. Two experiencen engineering firms were retained to prepare facility designs, estimate costs, and evaluate licensability of the designs according standards and regulations. Each effort is reported separately.gxigng A major study guideline was that a conscious effort be made in the engineering studies to use to advantage the natural benefits of the cverlying soil or rock. Soil is an effective filtering medium, yet this advantage will go unrealized if there is no pathway for fission products from within containment to the soil. Compared to soil, rock is relatively impermeable. Unfortunately, the pressure-resisting and nuclide migration-inhibiting characteristics of rock will not be realized if failure of tunnel seals, ventilation shaf t-closures, etc., leading to a loss of isolation, permit escape paths to the atmosphere. The designs which evolved in the course of this study utilize venting to a prescribed area to avoid uncontrolled failure of the containment building. Thus, an
" engineered" failure path was provided which proved to be exceptionally effective in limiting the release of radioactive materials following an extreme accident.
The following sections briefly describe study guidelines and considerations. A more complete description is contained in Reference 2-9, and summarized in Appendix B of this document. The guidelines common to both rock and berm concepts will be discussed first, followed by those specific to each siting mode. 2-7
DRAFT The engineering design teams used the following as general guidance:
- a. The underground facilities were to be designed to protect against the e ffects of a reactor core-meltdown accident--without first order regard for ef ! improbability of the accident.
- b. Public safety ( through prevention of catastrophic containment failure) was taken as the primary design goal of the engineered Accident Mitiga-tion Systems for fission product control in the conceptual underground facilities.
- c. After consideration of public safety, a supportive design goal was estab-lished for achieving a facility design of reasonably low cost--although trade-offs between operational convenience, maintainability, and in plant safety were required to be considered.
- d. Accident Mitigation System designs were required to be developed with the recognition that they would be compared against a standard of total, passive protection to the public from the results of a broad range of postulated severe accidents. Consequently, where possible, designs were sought for the Accident Mitigation Systems which would be demonstrably effective against perceived catastrophic nuclear power plant accidents.
- e. Development of thoroughly credible facility designs and analyses of public consequences was required, based upon documented assumptions and well-defiaed engineering methods.
Every effort was made to ensure objective treatment of designs, schedules, and cost estimates. In several instances, it was not possible to decide beforehand which was the appropriate direction. In such cases, the impact of variations of schedules, plant layouts, and certain design assumptions, such as groundwater volume and rock and soil quality were determined. A major cost estimating assumption was the existence of a " mature" underground industry. It was believed that this would minimize the number of judgeents and speculations over the time and cost of first plant licensing. The number of plants which must be constructed before the underground concepts could be declared " mature" is itself, highly speculative. The impact of the mature industry assumption will be discussed subsequently. A second major specification, and a departure from previous studies and cost estimates of underground siting, was that a common commercial operation date be used for both surface and underground concepts. This is more representative of actual planning processes where estimates of future demand are made a nd .' schedules developed to provide the needed baseload at some future time. The cost impact of this assumption can be significant and is covered in Section 6. Specifications for Accident Mitigation Systems were defined to ensure a syste-Five levels of s a fe ty matic development of augmented underground designs. augmentation were specified: Level Incorporation of only conventional Class 8 surface facility design features and methodology in an underground environment. 2-8 ,
Level Conventional surface facility Class 8 design with protective isolation devices on all tunnels, shafts and other penetrations to the surface. Level Elementary Class 9 accident mitigation system using a pressure relief mechanism (high permeability tunnels, expansion volumes, etc.) with a directed path to the earth medium surrounding the facility, in order to achieve pressure and fission product management through utilization of the natural external environ-ment of the unterground facility. Level Class 9 accident mitigation systens with an added pressure management system incorporating a controlled, filtered venting system. Level Class 9 accident control with complete fission product retention within underground containment using : ore catchers and long-term energy dissipating systems, as necessary. The Accident Mitigation Systems presented above have been defined in terms of accident " Classes" developed by the NRC. The accidents, which might occur in a nuclear power plant have been subjectively divided into 9 " Classes." Accidents in Classes 1 through 9 are categorized in ascending order of severity of conse-quences. Prior to licensing, a facility must be shown capable of controlling all categories of accidents up to, and including, a Class 8 accident, (repre-senting the most severe " Design Basis Accident") withcut substantial releases of radioactivity. A Class 8 accident can involve a partial meltdown of the core and substantial releases of radioactivity to the confines of the reactor containment building. However, before approval of a construction permit for a nuclear plant, the containmene building housing the reactor must be shown to be capable of retaining its integrity under the Clasa 8 accident conditions. Much of the debate over reactor safety centers on the mechanisms and the like-lihood of extreme, or Class 9, accidents. A massive failure of the Engineered Safety Features through whatever cause, could result in a Class 9 accident, which, by definition, is more severe than a Class 8. Study efforts concentrated on Level-2 and 3 design ccacepts. Earlier studies have shown that the extreme pressures possible and the diffggy of ensuring reliable closure systems, make Level-1 designs ineffective. The Level-2 designs attempt to utilize the natural material sutrounding the plant for pressure and radionuclide mitigation, while the Level-3 designs are a fully engineered attempt to provide similar benefits. An estimate of the cost of adapting a Level-3 concept to a conventional surface facility was also under-taken. Level-4 concepts were not evaluated. The designs were based on a nominal 1300 MWe nuclear see am supply system. Both pressurized water (PWR) and boilieg water reactors (BWR) were considered. To ensure comparability of estimating procedures, each de sign team prepared cost estimates for a similarly sized and located conventional surface plant. Both surface and underground designs were based on mechanical draft cooling towers, and used consistent labor rates valid for a specific locale in California. No transmission line or site acquisition costs were include.d in the overall cost estimates for either underground or reference surface designs. 2-9
L)KAr 1 2.5 Summary The preceding section has discussed the variety of underground design concepts; various arguments generally advanced in favor of underground siting; and, the basic guidelines by which the engineering teams approached their design tasks. The two general siting modes, b erm-contained and mined-cavern, are differaa-
- tiated by the geological constraints of the desired siting area. Mined-cavern plants require relatively massive formations of competent rock, whereas easily excavated material is preferred for the berm-contained configuration. Many of '
the criticisms of underground siting, such as increased potential for flooding, are concept specific and do not apply to underground siting in general. Similarly, the ef fectiveness of sub-surface siting in terms of accident miti-gation is also concept specific. Recommendations for sub-surface siting have generally steemed from some per-ceived advantage over a conventional facility. These- supposed advantages must be carefully analyzed on a site-specific basis. Valid motivations for underground siting, such as enhanced seismic per fo rmance , improved accident mitigation capability or improved decommissioning, are detailed in subsequent sections. Their impac t , validity, and applicability in a California context will be discussed. The designs and cost estimates evolved during the course of this study reflect conditions representative of California. Seismic design levels; rock and soil descriptions ; and, construction and labor costs are all California specific. Consequently, the conclusions of this report are strictly applicable only to California. A major study guideline was the assumption of an " extreme" reactor accident, i.e., a Class 9 or core-melt accident. The study goal, under this assumption, was to determine the capability of underground facilities to mitigate the resultant consequences should these unlikely yet severe, events occur. In marked contrast to earlier studies, the Accident Mitigation Systems of this effort use an " engineered failure path" to take advantage of the natural filtering and radionuclide capturing properties of the underground overlying soil and rock. The resultant ability of sub-surface facilities to achieve a remarkable level of reduction in accident consequences is discussed in subse-quent sections. 2-10
DRAFT REFERENCES SECTION 2 W.A. Kammer, N.P. Langley, L.A. Selzer, and R.L. Beck, 2-1 Watson, M.B., Underground Nuclear Power Plant Siting, The Aerospace Corporation and Environmental Quality Laboratory, California Institute of Technology, California, EQL No. 6, September 1972. 2-2 Blake, A., V.N. Karpenko, E.W. McCauley, and C.E. Walter, A Concept for Underground Siting of Nuclear Power Reactors, Lawrence Livermore Laboratory, California, UCRL-51408, May 1973. 2-3 Howard, G.E., and P. Ibanez, Seismic Assessment of Underground and , Buried Nuclear Power Plants, Applied Nucleonics, Inc., 1977. 2-4 Crowley, J.H., P.L. Doan, and D.R. McCreath, " Underground Nuclear Plant Siting: A Technical and Safety Assessment," Nuclear Safety, Vol.15, No. 5 , Se p t . -Oc t. , 1974. 2-5 Karpenko, V.N., and C.E. Walter, " Underground Siting of Nuclear Power Reactors," Lawrence Livermore Laboratories, Preprint, UCRL-75589, October 1974. 2-6 Allensworth, V.A., et al., Underground Siting of Nuclear Power Plants: Potential Benefits and Penalties, Sandia Laboratories, SAND 76-0412, NUREG-0255, 1977. 2-7 Conceptual Design and Estimated Cost of Nuclear Power Plants in Mined Caverns, Underground Design Consultants, January 1978. 2-8 Conceptual Design and Estimated Cost of Buried " Berm-Contained" Nuclear Power Plants, S&L Engineers, January 1978. 2-9 Evaluation of the Feasibility, Economic Impact, and Effectiveness of Underground Nuclear Powar Plants, The Aerospace Corporation, Final Technical Report, May 1978. 2-11
DRAFT 3.0 Reactor Accidents The plant designs, accident analyses, and discussions of risk which follow require an understanding of the nature of reactor accidents , certain physical attributes , and the manner by which the public may be affected. That the occurrence of an accident does not necessarily result in consequences to the public must also be understood. This section will briefly discuss these items. 3.1 Accident Classification There is no single "resctor accident." The complexity of a modern power reactor gives rise to large numbers of possible accidents of different likelihood and severity. To systemize licensing and evaluate the potential consequences of accidents, or need for improved safety features, a classification system was developed and is presented in Table 3-1. In general, lower-numbered accident classes display smaller potential for public consequences and the higher-numbered, much larger. In addition, as the class number increases, the relative frequency of occurrence decreases. A spectrum of assumed Class 8 accidents, termed Design Basis Accidents (DBA), is used to evaluate safety systems, such as the Emergency Core Cooling System (ECCS). If safety systems function as intended, an accident will be arrested and consequences to the public limited. However, should safety systems fail to function properly when called upon, for whatever reason, whether it be due to
' equipment failure, design error, or human failing, a Class 8 accident becomes, by definition, a Class 9 event. The descriptors core-melt and Class 9 accident are of ten used interchangeably. It should be noted, that an accident of the core-melt class has not occurred in a United States commercial power reactor.
Thus, they are at present hypothetical, yet physically possible, events. 3-1
DRAFT IABLE 3-1 REACTOR FACILITY CLASSIFICATION OF POSTULATED ACCIDENTS AND OCCLTRENCES No. Of Description Example (s) Class 1 Trivial incidents .Small spills Small leaks inside containment 2 Misc. small releases outaade Spills containment Leaks and pipe breaks 3 Radwaste system failures Equipment failure Serious malfunction or hu=an error 4 Events that release radioactivity Fuel defects during normal into the prLnary system operation Transients outside expected range of variables 5 Events that release radioactivity Class 4 and Heat Exchanger Leak into secondary system 6 Refueling accidents inside Drop fuel element containment Drop heavy object onto fuel Mechanical =alfunction or loss of cooling in transfer tuba 7 Accidents to spent fuel outside Drop fuel element containment Drop heavy object onto fuel Drop shielding cask - loss of cooling to cask, transportation incident on site Accident initiation considered Reactivity transient 8 in design-basis evaluation in Rupture of primary Pipirg the Safety Analysis Report Flow Decrease - Steamline Break Hypothetical sequences of Successive failures of multiple 9 failures more severe than barriers nor= ally provided and Class 8 maintained Source: Ref. 3-7. 3-2
DRAFT This study has evaluated the ability of underground nuclear plants to mitigate the potential consequences of extreme accidents which are assumed to result in melting of the reactor core. The accident analysis was conducted assuming that a series of particular, very unlikely, events had occurred. Since one of the major arguments advocating underground siting presumes that such a plant could control a meltdown accident, substantial effort went into testing this presumption. 3.2 Probability and Accidents In the past distinction was of ten made for safety evaluation purposes between events deemed " credible" and those considered " incredible." While useful in limiting the scope of safety evaluations to those accidents considered likely, the labeling was unfortunate for two reasons. First, " incredible" events seemed to occur too frequently and, secondly, use of the word " incredible" established a psychological coloration. This induced frame of mind prevents careful consid-eration of extreme accidents, and their potential consequences, which are possible, albeit improbable. An emerging trend, and one which should be encouraged , is the explicit cot.4ideration by regulatory agencies of extreme , yet unlikely events, in conjunction with those accidents normally considered for safety evaluation. The rate at which events are expected to occur is expressed as a frequency, the average number of events per time period. Of importance here is the predicted frequency of extreme accidents. Perhaps the most comprehensive, although not universally accepted, analysis of (tgeg )1ikelihood of extreme reactor accidents is the Reactor Safety Study (RSS), designated as WASH-1400. That effort used techniques from reliability engineering, fault and event tree analysis, to calculate the expected rate of occurrence of extseme accidents. Table 3-1, from the RSS summarizes the predicted rate of occurrence for categories of extreme accidents for both pressurized and boiling water reactors. Only accident sequences which resulted .in melting of the core are listed, and the table summarizes 58 different accident sequences for PWR's and 63 different sequences for BWR's. Several criticisms of the analytical technique used to develop the accident frequency estimates will be discussed subsequently; however, _fgr now it (PWR-7) should begoted that the individual table entries range from 4 x 10 to 5 x 10 ( PWR-4 ) . Expressed alternatively, the predicted frequency of PWR-7 is approximately once every 25,000 years and PWR 4, once every 2,000,000 years. Two questions arise: a) are the estimates accurate?; and, b) if accurate, do they indicate adequate safety or unacceptable risk? Be fore discussing the preceding questions, two points need stating. First, the frequencies cited above are for rate of occurrence of an accident in a given reactor which results in varying consequences to the public. Since the radio-active material is inside a massive containment building, whether or not, and how the containment fails influences the severity of an accident. Thus, a PWR-4 is not equivalent, necessarily, in consequences to a PWR-7. In fact, a PWR-7 accident would be expected to have relatively trivial consequences when compared _. 2: consequences of accidents in the first three PWR categories. 3-3
DRAFT Table 3-2 Reactor Safety Study Accident Probabilities (per reactor year) PWR Release Category 1 2 3 4 5 6 7
-6 ~0 4x10 -5
~ ~
~7 7x10 6x10 Probability 9x10 8x10 4x10 5x10 (50% value)
BWR( > Release Category 1 2 3 4
-0 -0 -5 -6 Probability lx10 6x10 2x10 2x10 (50% value)
(1) Re f. 3-1, Table 5-2. (2) Re f. 3-1, Table 5-3. Second, the frequencies are expressed in absolute terms; i.e., so many occur-rences, on average, per reactor-year. Another way to look at such data is relatively; that is, PWR 7 is relatively alout 100 times more likely than PWR-4. There have been many t.ho have argued that reliability analysis caanot produce absolute predictions of risk. Whether or not the methodology can fully account for human errors, design-deficiencies, and failures of multiple systems due to a single cause (common-mode failure) is questionable. The American Physical Society, in theu review of the RSS expressed a similar view:
....we recognze that the event-tree and fault-tree approach can have merit in highlighting relative strengths and weaknesses of reactor systems, particularly through comparison of different sequences of reactor behavior. However, based on our experience with problems of this nature evolving very low probabilities, we do not now have confidence in the presently calculated (RSS) absolute values of the probabilities of the various branches (or events) ." (Ref. 3-2)
(Parenthetical notations added by editor.) It is prudent, at this time, to regard analyses of the RSS type as optimistic, lower bounds to actual risk. An upper bound may be derived from the accumulated operating history of commercial nuclear plants, approximately 300 plant years. 3-4
- DKMT Thus th frequency of potentially severe accidents is probably less than 3x10 g3 (e1/300) and possibly greater than the values listed in Table 3-2.
Public policy makers must recognize this uncertainty in their decisions. The second question posed earlier addresses the acceptability of risk. An event cannot be dismissed because the frequency is seemingly low. Figure 3-1, after Re ferences 3-1 and 3-3, displays the calculated results from the RSS and the historical record of other major accident sources. Let us differentiate between accident consequences and accident rate .of occurresce. Risk combines both accident rate of occurrence and magnitude of consequences. Thus, Figure 3-1 curves show that more sevare accidents, those which result in larger numbers of fatalities, occur less frequently than less severe accidents.7 Given a generally
- constant rate for accident occurrence, the distribution of conse-quences in Figure 3-1, for the curve labeled RSS Results. F.arly Fatalities , is due to variation in the quantity of radioactive material which could be released in different accideat categories; the differences in meteorological conditione at various sites; the varying population densities at reactor locations; and other factors. Decisions about relative risk to the public cannot be made on the basis of accident probability alona but must reflect the potential for consequences to the public. The latter are, in many respects, site dependent.
Generally, debate over the validity of the RSS centers on the probability that extreme events and associated consequences will occur. (The latter category includes arguments over health models, etc.) Little disagreement is seen on the shape of the risk curve. For example, the curve in Figure 3-1, labeled "UCS Estimates Reference (3-3)," is similar in shape to the " Latent Fatalities" curve from the RSS, only displaced upwards. ne shape of the risk curve is extremely significant for it suggests particularly useful strategies for risk reduction, independent of the actual probability assigned to extreme accidents. This will be discussed more fully in Section 8.0. 3.3 Design Considerations for Extreme Accidents Accidents of the class considered in this study generate high temperatures and pressures. He precise valus of the calculated temperatures and pressures is dependent on the details of the specific accident sequence assumed to occur. As noted earlier, the RSS identified some 58 individual sequences of events which 1 This point deserves amplification. Not all reactors are the same. Differ-ent hardware configurations result in different accident probabilities. It is probably true that newer designs are less accident prone than early designs since the newer plants incorporate the experiences of the old. There is some question whether site specific influences, such as seismicity, flooding, tornadoes, etc., could affect the general frequency of accidents. However, at this time, withouc an exhaustive examination of the relationship between initiating events and accident rates, the statement stands on available documentation.
- 3-5
DRATE 10 Air Crashes Total Fires 1 1 N ? 10 ho e,j n r<, **s,*'%,, dp
- g 10 -
$ e# 'O, 4 . N E -3 -
% e, y 10 *o e, S o, h c
- 4 ^ 's UCS 10 -
o%"g 5- \ imates f, g (Ref. 3-3) Latent % 5 10 Fatalities Early Fatalities Id - RSS Results gg7 i i e i 10 100 1,000 10,000 100,000 1,000,000 Fatalities, X Figure 3-1. Frequency of Man-Caused Events Involving Fatalities. (Annualized Basis for 100 Nuclear Power Plants)*
- (Base figure reproduced from RSS--Ref. 3-1) 3-6
DKAFf could lead to core-melt in the PWR of that study and 63 sequences for the BWR examined. The cited numbers ingde only Since those sequences considered " dominant"a each sequence would generate (probability greater than 10 ). unique time-dependent progression of temperature and pressure, the physical parameters of interest, any concepts considered capable of mitigating such accidents, would need to be effective against each and every one. All 121 sequences were not examined. Instead, elements of certain accident sequences were selected to envelope physical extremes and are described in detail in Re ferences 3-4 and 3-5. Portions of three principal accident scenarios were selected to establish requiremtnts in two general time domains: the short term (periods less than one hour following onset of an accident); and the intermediate and long term (periods of days, weeks, and months). The large Loss-of-Coolant-Accident (LOCA) with effective emergency core cooling system (ECCS) was used to determine the short term response of the underground plant and accident mitigation systems to rapid increases of pressure. A transient-induced loss of of f-site and on-site power was assumed and appeared to bound energy release rates for system design. Compared with a LOCA, this sequence produces pressures and temperatures which rise slowly until about three hours af ter accident initiation. At that point, following melt-through of the pressure vessel and discharge of ECCS accumulator water onto the molten core, one approximately 100 million BTU's of thermal energy can be released in about minute. This extreme rate of energy discharge represented limiting conditions for pressure and mass flow-rates in accident mitigation system ductwork. The third assumed event was a Loss-of-Coolant-Accident withIninitial thisfailure and scenario, subsequent restoration of emergency core cooling systems. water would be introduced into containment and converted to steam at approx-imately the maximum possible rate. This process would develop the highest pressure at long times in a sealed containment or with the mitigation systems which evolved, size the heat-sink necessary to reduce system pressure by condensation. Several points must be emphasized. First, the selected scenarios are all unlikely events. They were chosen to place the maximum physical limits on the underground plants. Second, they do not form the basis for a risk analysis. Third, they were used to evaluate the e f fectiveness of augmented underground siting for mitigation of extreme reactor accidents. Fourth, they did not form the specific basis for calculation of differences in accident consequences between surface and underground plants. 3.4 Containment Failure Modes A significant finding of the RSS is that failure of safety systems As does not discussed automatically result in failure of the containment building. previously, an accident which releases radioactive material inside the contain-ment building does not, necessarily, result in consequences to the public. In 9 1 Consider curreng estimates Jr the age of the univege, 4.5x10 years. Thus, 1/(4.5x10 ) = 2.2x10 , and is bounded by 10 . The reader is cautioned against rash conclusions regarding acceptability of accidents which occur, on predicted average, once in the age of the universe. 3-7
DRAFT fact, current licensing procedures assume that a substantial portion of the reactor inventory is released to the containment building and that the contain-ment subsequently leaks at a prescribed rate, verified through testing, to the b ios phere. This assumed leakage forms the basis for potential doses to the public and setting of exclusionary boundaries around the plant. Typically, under these assumptions and current health models, potential doses to the public are low and would not result in significant health consequences. The key assumption is that the containment retains its integrity, although not 100" since it is assumed to leak slowly. There are several ways for a containment building to fail. Ihree of these are shown in Figure 3-2 and include breach by a missile, perhaps generated by a steam explosion (rapid conversion of water to steam); through buildup of pressures beyond the capacity of the containment structure or penetration seals, termed overpressurization; by melt-through of the plant concrete and steel foundation mat by a molten core; or by failure of the containment isolation system. The containment isolation system is designed in part to seal all piping which penetrates the walls of the containment building and which might provide an escape path for radioactive material under accident conditions. Failure of the containment isolation system could result in a breach of cantainment. Containment failure through overpressurization is particularly significant. Figure 3-3 displays several containment building pressure-time relationships for typical surface plants following a LOCA with inoperable ECCS. The curves share two features: a) pressures increase with time; and b) pressures appear to be still rising at the end of the evaluated time period. Similar calculations performed for this study have confirmed this behavior. The net consequence is that under some accident se quences , containment failure can be anticipated due to g This conclusion was reached early in the g internal pressures. and directed later concept design efforts towards passive systems program for pressure relief. A fifth containment failure mode, due to extreme temperatures which would cause the failure of con g ent penetration seals, resulted from several of the Temperatures well in excess of conventional seal accident sequences design limits (approximately 400*F) can develop. Containment penetration seal f ailure by high temperatures would not be as sudden as an overpressurization failure. Nevertheless , the augmented underground designs consider this failure mode through a secondary sealing system which provides an additional barrier to release of radionuclides. The manner by which containment fails overwhelmingly influences the magnitude of accident consequences. The designs developed for this study provide an
" engineered failure path." The objective is, under the assumption that a severe accident has occurred, to avoid those failure modes such as breach by missile or overpressurization which are likely to result in major public consequences. The engineered failure path directs the failure in a graceful fashion to minimize consequences.
3-8
DRAFT m m e w 3 ms
- =-
se & si
=8 g 8W *E" s e -23=
5 C Wag! sh wr w507- Es s, == 80 A Sc ss - u= 2 a .. . s * *
% :y u i
e a s 5 g 4 ed - as - S E~e s> r [~s \ E* z (e at 2
- a
=@
=
A % w N n
,4- - < -
=
ge-il wa s i C da gaa- ar = g . o
$5 7 c v5<'
8 O =>
-o- r
... m o.
O A w 3 e a m se* 8 " 5:
- s C e5 y
"'8 1 n1 m m au- um
= m z o 3
- s :
zas - C ggs C sg! i
==4 gs:
o8' eee mo-3-9
200 200 FRENCH STUDIES 1965 AEROSPACE CORP. 100 100 -
.g 1973 a FAILU w2 w - $ ' ' ' ' 0' ' ' '
O' W o_ 1 2 3 4 5 6 1 2 3 i' $ g g 200 m 200 E AEC, WASH - 1400 NORWEGI AN STUDIES 1975 1975 100 100 - ( 0 ' ' 0 1 2 3 4 5 6 1 2 3 4 5 '6 hr hr U NOT E: TYPICAL SURFACE CONTAINMENTS DESIGNED FOR 50-60 psig Figure 3-3 Containment Pressure 3 rime Histories (ECCS Inoperable) N r-]
DRAFT 3.5 Si== = rv Section 3 has briefly described the classification, one esti= ate of probabil-ities, and physical attributes of severe reactor accidents. In addition the accident generated extreme pressure and te=peratures, which influenced under-ground plant designs, was discussed. Reactor accidents can be ranked numerically from 1-9. As the rank order increases potential public consequences also increase. However, as the rank order increases, the estimated frequency of occurrence decreases. A spectru= of Class 8 accidents is currently used to evaluate the safety of conventional nuclear power plants. Should the Engineered Safety Features fail to function properly, a Class 9 accident loosely defined as an event more severe than Class 8, but of ten considered to include melting of the reactor core, could result. Since a Class 9 accident has not yet occurred in a commercial U.S. Power reactor, it re=ains a hypothetical, yet possible, event. De techniques of reliability engineering, fault, and event tree analysis have been used, in previous investigations , to predict the expected rate of occur-rence of extreme accidents. Yet, controversy exists regarding the accuracy of these estimates, and , if accurate, the acceptability of these levels of safety. Therefore, decisions about public acceptability of risk should not be =ade on the basis of accident probability and potential public consequences alone, but should reflect the uncertainty of these measures. Accidents of the class considered in this study generate high te=peratures and pressures, which are dependent on specific accident sequences. Since, the large nu=ber of accident sequences pouible prohibited individual treatment, ele =ents of certain dominant accident r- ; ences were utilized to envelope the physical extre=es of the accidents investigated. Selected accident scenarios were chosen to place maximu= physical limits on the underground designs. They were not the basis for a comparative risk analysis or the calculation of accident consequences. Contaic=ents =ay f ail in several ways, the manner of which will influence the
=agnitude of accident consequences. The engineered failure path of the present study seeks to avoid failure codes which could result in major public conse-quences such as overpressurization and missile penetration. The ability of underground siting to achieve this goal is discussed in a later section.
A novel containment failure mode , penetration seal failure due to extre=e temperatures, was identified during the course of the study. The high-te=perature failure mode has not yet been discussed in the reactor safety literature. _ 3-11
DRAFT REFERENCES SECTION 3 3-1 Reactor Safety Study - An Assessment of Accident Risks in U.S. Commer-cial Nuclear Power Plants, WASH-1400, October 1975. 3-2 Reviews of Modern Physics, Report to the APS by the Study Croup on Light-Water Reactor Safety, American Physical Society, Vo l . 47, No. 1, Summer 1975. . 3-3 The Risks of Nuclear Power Reactors, A Review of the NRC Beactor Safety Study WASH-1400, Unio n of Concerned Scientists, Cambridge, Mass., August 1977. 3-4 Evaluation of the Feasibility, Economic Impact, and Effectiveness of Underground Nuclear Power _ Plants, The Aerospace Corporation, Final Technical Report, May 1978. 3-5 Analysis of Public Consequences from Postulated Severe Accident Se-quences in Underground Nuclear Power Plants, Advanced Research and Applications Corporation, December, 1977. 3-6 Systems Management Support for ERCDC Study of Undergrounding and Berm Containment--Interim Report, The Aerospace Corporation, August 1977. 3-7 The Safety of Nuclear Power Reactors and Related Facilities, WASH-1250, AEC, July 1973. 3-12
DRAFT 4.0 Siting Implications There are two general questions concerning underground nuclear power reactors and siting: (a) Do reasonably appropriate sites exist within California; and (b) Will underground siting generate any major new opportunities? Previour studies have tended to either eliminate Cgi_gia completely as a potential locale for mined-cavern underground plants or to characterize the entire state as opportune for berm-contained concepts. Reality, as will be shown, lies between these opinions. As discussed earlier, underground siting is of ten viewed as having potential for urban siting of power reactors. If this were possible, and more impo r-tantly, acceptable, then the cost of construction and visual intrusion of the facility and major transmission lines could be avoided. Urban siting, however, , was found to be a non-existent opportunity for California due to a unique combination of population and earthquake hazards. This t io t, . g*tgog summarize s a more extensive treatment of the siting ques-Specifically, this aspect of the study dealt with the various parameters, both physical and social, which could influence or limit the siting of sub-surface nucl2ar plants within California. The screening criteria developed for this effort were judged, by the staff, to be reaconable and s t ra igh t fo rwa rd . It must be noted, however, that these factors were not conceived as absolute " exclusion" criteria, rather, they should be viewed as representing areas of varying desirability for nuclear plant siting. Further-more, the intent was not to identify specific potential sites or pass judgment on existing or proposed locations for nuclear power reactors. Ultimately, the conclusions reached depend on the criteria used and the criteria, in turn, were developed for the purposes of this study alone. 4.1 Criteria Fourteen factors, divided into the two broad categories of Engineering and Geological / Geophysical Influence, and Social Influence were examined. The e f fort was limited in scope, although adequate to its purpose. Factors con-sidered included: o Engineering and Geological / Geophysical o Quaternary Fault Zones o Quaternary Volcanic Activity o Flood Potential o Dam Failure Inundation Areas o Known Geothermal Resource Areas o Oil and Gas Fields o Quaternary and Pliocene Unconsolidated Deposits o Mesozoic Granitic Rock o Presence of Water o Groundwater o Earthquake Epicenters 4-1
DRAFT o Social Influence o National Parks, National Monuments, Indian Reservations, and State Parks o Wilderness and Wild River Areas o Population Reference 4-2 defines each of the preceding factors in detail and explains the process by which criteria were formulated and large-scale maps produced. Some elaboration of the Quaternary Fault Zone criteria and its significance is necessary, however. Five-mile half-wid th zones were drawn along identified Quaternary Faults. He use of Quaternary Faults as a screening parameter is somewhat more restrictive than current federal regulatory requirements (which are based in part on Holocene faulting) but the general delineated areas would be similar. More important, however, is the inter'pretation given the mapped zones. They represent areas with the potential for surface ground rupture, not just ground shaking, during an earthquake. Since, at the present time sites where the potential ex ground rupture are considered unsuitable for nuclear power stations,g_gfor the mapped zones are in effect avoidance areas for both surface and underground facilities. Berm-contained facilities require material which can be easily excavated. Quaternary alluvium, lake, playa and terrace deposits , nonmarine Pliocene and Pleistocene sandstone, and loosely consolidated shale and gravel were considered desirable material. The locations of such materials are shown in white on Figure 4-1. Mined-cavern concepts, on the other hand, demand large areas of high quality rock. For this study, only Mesozoic Granitic rock was considered to be, in general, suitable for large caverns. Other rock types may be adequate but were not considered. Mesozoic Granitic Rocks are shown as the light areas on Figure 4-2. . 4.2 Observations ne three dominant screening factors were Quaternary Fault Zones, Population and the Presence of Water. Figure 4-3 dramatically indicates the impact of faulting and potential earth-quakes on nuclear plant siting within California. Particularly significant is the effect on coastal siting. Population is presented in Figure 4-4. The three dominant urban areas of San Francisco, Los Angeles, and San Diego are clearly visible as are the urban centers in the Central Valley. With today's pbwe r. technology, water is necessary for cooling purposes. Figure 4-5 shows the location of inland water within California. If the map of Ouaternary Fault Zones is overlain with that of Population, a startling conclusion can be drawn: Californians prefer to live in areas that, coincidentally, have high potential for earthquake activity. More relevant, however, is the observation that seismic factors alone exclude most urban areas from siting consideration. The urban regions remaining af ter seismic exclusion 4-2
,,& N
/ tls 4 4 4 4
] 4 i
^*I'
'" ~
.t G' .
y v 4'
- x. 4 4 -
\ '
+
'., . +j =
\
.~~'
p y
~ ,
- x. ^ ;y 3y,W s;
)
QN g s ss
/ O #
N
*w, - -
g s , J' ~ gg
$A '
; g
-} -
~
)v, l ~ ' . , _
- prjy; fj. Y,
- .- r
,/
e J: s
;r;s -'.
, r
) :), , . * ] .,-
(j
.,w
, . , " . w + !: _'
- e. . _
"r N l e
r
~ ..
v L . , ' [ ,
- y. e
, * ,:: y ~i, ' "f.-' '
,/ .
^*
A' 4 (' ,,. c: ,. +Q
)
4j',, ~. * - w .i-
+
v. j
.'s
t l-y
; E Y';
._ - 4 l' v _
w,.. . n ._ x V ' . . l ,, _ ( n ;f. y ll ; a
>L}QR,g' .
,x . ,s % l.,. v
)
.s -w
,~~' '
' :,/ ^
;j . c
'E ', .r .. m
.s e
q Q
,W p -
gi' ~ ' 3 ( 1 n %,l. 'y v
^
- .- : j r
- ?-l': 4
# ',; ' ;j s
,, AP s
f ., ' k
\ a ,,a # \
J
/ s ' 1 e' y
\
,/ j _ N 4 8 . & h [ g N 4-3
++
/
s s , . . 4 4 4 / 'N
/s / %
b N 0 4
~
Y i e x s N' y . s s 9 , s.. s' s . si$pf'y \ a s/
" '/ +
'j y / A e
- ,l ,
* , 4 6 4 4 . . . ':
4-4
DRAFT Figure 4-3 i , ( j
.s '
t -
,t .
4~ t
! ~ '*-' -
QUATERNARY FAULT ZONES ,,
~ .
g
.:- 4 j
\, --- ~-*" . .:.?
a' i
. .-- , . . . .. . ~
s' as i r v- .% o' 3
's
..-~~~ N
\
N
. '. s
's
- - ~ ~ ~ - * -
W.N. - aft t
'N4
-. s
. m -
r pi,. a.m.p \
,,,se s mmsmo ter79 amasensie ammemme ---
asse1
\
I cammamester71 emem=4 em 8 % i-T c,m,.,--===== =a
, - - ===- ...: .
i g capangg Osuneum of h WW W ma . - % casmesmessff. issesmass res*8
- m gfM e - W 8e AstW* 08" esa .%
0 50 10 0 malts % g are 4-5
aeF1 - Figure 4-4 W W
- W W ~
f s a
- 6 d f
9
- t. "
t
- a .
8 *
% S .
6 s o' 'g* ? d g t *
, (
,~ .,. .
. .' C.
- h ..
,e* -***.-
*
- t, n ,
*%0
.$.* ?s,
,sk *;,.'* ..l '
**'.n.,
s gM' " , .. Y ***2
,5 , 'A f4*\. x' -
' Cfx &
^
. , & ,,.') e
- s'~ W
,f
, H,' , . .'. ' *
*y',,.
\.,,'
* . $ / ~.'-/,* % ' } ? ,,
- J , r t ..
; ,e -;'*"e ...*j, ~~., %
,s .
e
',Y !
i
' e s~
- z
%v.,
*/ s
' g' ) *
*.. i
'y,
-Q e
,e ;:q,,v'
- s. ,."-~j ,. '-.. gq.
, m j
f",. r
, X ,,,
s., N'
?~, ,.N,
.y '., ~
.';.m;
- s. -
; - 9$g.
~~
~
*sx .
e y-**--- j , WMM10"
% ~
6
**,g
%._ .h
't g
f.- ' --.. ,, %
\' Q.-{p,osooo***'?fj-9e ' . .
f*
)'
\
".o,== e.,
\' /,,
k we aQW
?
w.o.uJ'*
- 8,g.
to #:,,o , c# a= y* p o so ioo **'es w^ , -- _ _ ,,.-- ~*'*~.- .)\ g d
. m ,
\v '
- 8 s,, ". '
.-~ e d d d d 4 1 < #
i d e a 4-6
DRAFT Figure 4-5 7 C8L8'euna zwunas y asses ase 33,ogy 12/arr7 i s. 4 l PRESENCE OF WATER ; i e I s/' ' . - est l h -- j e j i l s . 6
.,r,5 O I s
i .. ,
; _. . .......4
.m a
_. A s., s t s .. . _ m , k % s., s s s as - N =1 O'- s i,
. T s e s
* \,
p N -
;, {T -
s l' M ,- ...- , LEGrie0 , i l ., - - - _ I ) s
.-.e-trMasertow h g.
. N s
W s .n
.w b ,
j I 6 . N
- s '
onovecearts ase Last stomaar *---- - - - - - - - - - - - - - - - - - - - ----------~~~'N O e.. - % g , ,
- a. '
og ,
% ?
l 0-t=*= .e
. s l o .. % <
,e aeastaaruman eas?t entse 6-
-) O cJ l s. _ .
. ' ~ ~ '. . . . . . '
I g . . ,,,,, _ ,, ,
..- . - ~
O SeudICES O
., n == w . c e a e eos .on o p.,-
y- % . g e.*se e s a
.s i e-
' r, t e se ,ee awas .s '
1 m _ ~ *J ? l 9 s e .* e e e o
=
-s........-- ,
4-7
A m&
/ J g . . . , < , 4 4 Y /~ i / O a
/ #Y %
N , s y
.,; 'y 4 /,
p x . a . 4 s s s 4 4 p , l' m. o Af .. l . c ' + '
- MP
. s' 4 k
A j , ;., , ,
.l ,
s ' . '
& 4 4 d . #
- 4-8
9Y 4
/ ?
4 , , 4 < , 4 J . l s/ e / s S y .
/ /*
s N s' g 4
\
s bM x
. ~$ $.'
/
s 0 A
'x
?
4 J', s
's s
4
\
s s ,. s e .. j
' .. ~
4 3
's s
4 4 4 . 4-9
DRAFT are, in general, water poor. The inland urban areas are not major constraints on siting and, for the most part, are not large enough to justify an urban-sited, underground nuclear power plant. Thus, a perhaps unique set of factors combine to eliminate urban siting as a significant motivation for underground nuclear power plants in California. Figure 4-6 combines all screening f3ctors, even though some are not strictly exclusionary. The areas shown in white may be considered, at this level of evaluation, to be technically feasible for berm-containment siting. Similarly, Figure 4-7 illustrates technically feasible mined-cavern concept siting areas. As a general observation, proposed nuclear plants for California fall within areas found reasonable for berm-contained siting. In addition, after elimi-nating high, alpine regions, the most realistic mined-cavern siting opportunity is probably in the foothills of the Central to Southern Sierra Nevada Mountains. 4.3 Summary Previous studies of the feasibility of underground siting opportunity in Cali-fornia have concluded th a t no appropriate sites existed for mined-cavern concepts and that an abundance of sites existed for berm-contained concepts. Fourteen criteria, which considered engineering and geological / geophysical factors, and measures of social influence were mapped. For California, three factors dominated siting. They were the presence of recent earthquake faulting; keeping prudent distance from urban concentrations; and, the presence of water. Recent earthquake faulting precluded most of California's coastline frem siting consideration. Given present cooling tech-nology, this made the presence of inland water of particular siting importance. Interestingly, California's urban regions coincided predominately with areas of geologically recent earthquake faulting. Because of the imprudence of siting, whether above or below ground, in areas of recent fault activity, urban siting of underground nuclear power plants was not considered an opportunity in California. Technically appropriate areas for each major underground concept were identified within the state. In general, these areas were proximate to recently proposed sites for surface nuclear power plants. Thus, underground siting should be considered a technical alternative to those surface proposals. The berm-containment concept was, in general, less constrained than rock-siting. 4 4-10
DRAFT REFERENCES SECTION 4 4-1 Review of Underground Siting of Nuclear Power Plants, UEC-AEC-740107, January 1974. 4-2 Underground Siting of Nuclear Power Reactors - Determination of Site Characteristics and General Site Availability in California, California Energy Resources Conservation and Development Commission, January 1978. 4-3 Geological Study of the Underground Siting of Nuclear Power Plants, California Division of Mines and Geology, January 1978. 4-4 General Site Suitability Criteria for Nuclear Power Stations, U.S. Nuclear Regulatory Commission, Regulatory Guide 4.7. 4 4-11
DRAFT 5.0 conceptual Plant Designs The general study guidelines were set forth in Section 2.4. There are several important guidelines that should be restated. First, and perhaps foremost, the designs were intended to be realistic. For this reason, the CEC turned to qualified engineering organizations for conceptual plant layouts, schedule analysis, licensing assessment and cost estimates. The design organizations were not asked, "Should nuclear power plants be placed underground?" but rather, "If a nuclear power plant were to be built underground, what form would it take and what is the best estimate of cost?" Second, the designers were to attempt to use to advantage the underground environment. In this regard, plant designs include an " engineered pathway for failure." Clearly, accident prevention is of foremost concern, yet a complementary concern is to consider the potential pathways leading to conse-quences. This study has shown that by controlling the manner of failure, and thus the release pathway, accident consequences can be dramatically reduced. It is significant that of the many previous studies of underground siting, few have considered extreme accidents, as h Those that have did not engineer a manner of failure.g_gisinvestigation. In general they have concluded that the accident mitigation capability of sub-surface siting is a technological Maginot Line, consisting of defenses which are easily bypassed. Given the designs evaluated in earlier works , simple extrapolation of surface designs to an underground setting, such a conclusion is reasonable; it is not however, a complete assessment of underground siting. 5.1 General considerations As discussed earlier, within the two general underground concepts, berm-containment and mined-cavern, a large number of plant arrangements and access methods are possible. For example, the facility may be entered (and con-structed) through nearly horizontal tunnels, or by vertical shafts; various elements of the plant, such as the turbine-generator and related systems, may be above.or beneath the ground surface; or the facilities may be built in hillside or flat terrain. Each possible plant arrangement has associated individual costs, merits and penalties. To avoid an unintentional bias in costs or tech-nical feasibility, several variations from a " baseline" concept eere evaluated. Thus, the economic impact of both surface and sub-surface location of turbine-generators, various seismic design levels, and other factors were examined. The complete baseline plant description for both berm and rock concepts and design variations considered are given in Appendix B. The safety design basis for the underground plants was equivalent to that used for current licensing of surface facilities. All requirements and safety features applicable to surface nuclear facilities were considered in the design of the underground plants. The passive systems added for the mitigation of extreme accidents not normally included in licensing were considered to be supplemental to the Engineered Safety Features. As such, only the interface between the mitigation systems and safety category facilities, e.g., the containment structure or cavern, was expected to be impacted by licensing requirements. 5-1
DRAFT 5.2 Berm-Contained Plant The final concept, portrayed in Figure 5-1, is fully described in Reference 5-2. Only highlights will be discussed here. The final concept evolved from preliminary study of the functional relationship between various plant systems and the cost of various alternative arrangements. Six variations on plant arrangement were considered which led, ultimately, to the final concept as the most preferred. The dominant plant feature is the large dome which encompasses the principal radioactivity-containing plant systems. The dome is an efficient structural shape designed to resist soil loads due both to overburden-weight and earth-quakes. The cylindrical, inner domed structure would be considered a reactor containment building for a surface facility. This underground plant containment building was designed to all applicable regulatory requirements regarding leaktightness and fission product retention capability. The nuclear steam supply system was essentially unchanged from that which would be used in a conventional surface plant. In general, the arrangement allows for incor-poration in the underground plant of equipment identical to that of the surface plant into relatively similar spatial relationships and it provides approximately comparable maintenance space. Piping, cabling, and ventilation ducts are routed from the underground facili-ties through a tunnel building to the turbine generator building. Fuel ship-ments gain access by a short tunnel through the berm. This tunnel would be normally sealed. These are the only interfaces between the underground plant and the surface. For reasons to be described subsequently, the tunnel building can be manually sealed. Thus, in this design concept, the domed facility can function as a secondary containment should the integrity of the primary containment building be lost. The elevation difference between the heat source and turbines coupled with slightly longer main steam lines result in a slight efficiency loss as compared to a comparable surface facility. This loss is estimated not to exceed one percent of net output. 5.3 Accident Mitigation Systems in Berm-Contained Facilities In the context of this study, underground siting is e f fec tive for accident mitigation because the plant designs make use of an engineered failure path. By controlling the mechanism of failure, accident consequences can be greatly reduced. It must be clearly understood that the systems incorporated within these plant designs are viewed as supplemental to the Engineered Safety Features (ESF) currently required for nuclear power plants and not as replacements. Fundamentally, the accident mitigation systems are designed to avoid over-pressurization of the primary containment. This is accomplished by venting of accident-generated radioactiva vapors and gases into the berm material at the foundation levels of the plant. The pathway from the primary containment to the surrounding berm consists of 24 large ducts placed radially in the massive foundation as depicted in Figure 5-1. Within the containment, the ducts are 5-2
E L + 99' E L 149'; - E L + 89' I EL+ 139' ,
' o L
^ N ' TURBINE BUILDING '
/ '
t.O -GRA *-- 150 l D.---+ a.o GRADE FLP ' O*2 [ l E L+ 2
- N .
/ [
RUPTURE DISK
\ l L.-32 '
N [ EL-4B , g \ r5, pEL.-85,
.a E L-135'y :"* *- 1:
m + u6
**u*:.
i _ E L-151'2 Nt 3 -' --~ "*',, 3 0 6* I.D. _ 338'QDL = ' 20' Figure 5-1 BASELINE LEVEL-2 BERM-CONTAINED PLANT DRAFT
DRAFT sealed by discs designed to fail at pressures above those anticipated in design basis accidents yet below those which sould jeopardize the primary containment. Re ferenc es 5-2, 5-3 and 5-4 discuss the rupture disc design and conclude in general that they are technically feasible at this time. The discs are also temperature sensitive, since it is possible in some specific accident sequggegg to generate excessively high temperatures with relatively low pressures. The designs are passive, that is, they require no external power or control signal for activation. They are redundant, amenable to random sampling and destructive testing to establish failure frequency. The simplicity and redun-dancy of the pathway permits engineering to any selected reliability criteria with great confidence that such design reliability would be achieved in actual us+. The Level-2 design concept calls for termination of vent piping in a rock-filled portion of the berm. In use , the rock would serve as a heat-sink and permit expansion, and' condensation of gases and vapors. Both functions would serve to reduce pressures within the primary containment. Subsequently, the gases would percolate through the berm material before reaching the surface. The effective-ness of the enforced radionuclide migration through the berm is quite dramatic and greatly reduces potential off-site consequences. This point will be more fully addressed in Section 7. The Level-2 design was the basis for evaluation of tha effectiveness of underground siting. The Level-3 concept provides, through an engineered system, what the bers provides naturally. The Level-3 design, as shown in Figure 5-2, terminates the vent pipe s in a graded-rock, sand-filled enclosure. The enclosure completely surrounds the plant at the foundation level. In addition, the Level-3 design provides a vent stack, filled with selected filtering material, to the at-mosphere. The virtue of the Level-3 system is that it can be optimally tailored to design requirements. In addition, the Level-2 concept is an underground application of a controlled, filtered venting system which has been suggested for surface facilities. (Cost estimates include application of a Level-3 design to a surface plant.) 5.4 Mined-Cavern Plant In contrast to the berm-contained plant where the large, domed rein forced concrete structure prevented intrusion of the overlying so il , in the mined-cavern concept the rock itself serves as the principal plant structure. Thus, the mined-cavern plant is designed as a series of parallel caverns which would house the major plant systems. The general arrangement is pictured in Figure 5-3. Construction of the plant in a three dimensional medium places certain restrictions on plant layout. First, the general orientation of caverns is dictated by the existing internal rock scress field. The cavern width is governed by equipment dimensions and the feasibility of construction. The spacing between caverns must be at least equal to the cavern spans to ensure stable openings. Lastly, the depth of sound rock overlying the caverns must be approximately equal to the cavern height. These considerations combine to produce the general arrangement shown in Figure 5-3. The relatively compact berm-contained f scility is in contrast to the large areal extent of the mined-cavern plant. 5-4
r ( o
,Im Y
M j J* ,, h .
*l- ' ' ' '- ls; , .rs ' ,'
n 1- >
'fi / / /,/ ,/ -' '
}c' ,,
/ -Q:-
-r'
., , : ' ,l l /, , '
, . . . . t.D y /
a
>~ M '
=
*. f z ee
-z ,
h M.9,,' 'f.* - * (. ,
,e
- 5# .
~
* *' '. ;,,* g,;,,,
, h k. ,, @$
E!
/
q p; '
- M - ~~*C* ;*
,. Hg I'
p
/ ...
}<. =,.
t ,
, - ,;. ... aa ,
o..D .
. w g . ./
'C- ,
6 g:: a . t$ ~~r , .
..~~ * .
Jz
'. 1.I ' 'CE
'% 'M j [, J -
. ,1 :". I Q'"'*
y . / , 4'. . ,
-. t 4g;
., x' < -
. j -;
, gQ
. * - Q," ..,. ,.... d' C . ,
, ; Og2 ,' ,
WM
,.', / . . ,
se
' r .
i , . Z
\,- '. ~. b .'- Q'.,h" h, * . ,,, hh
,/ .j t '- . . . ',
[- C j m M_ *" .
, , . zu , y . -
O cx: _., 1e. - . - -
.e , pg i
/
y .'.'; p
.; s
% . .;: . : T ~
"r*-- *
%,?
f6 w= l[ 1 , >. ,
- / . J.s : I .:
.:O.$.. --l% -' 'M. m '
. -b GD &
, 'yf' !
,..._ l*- .
E l }$
- -ih~; ,
- . m
- D.
-k2 i ,.'*'~,
~'
w
- .b,c
. .?3. , ..
I
,J _ -
3 Ps v,r=a_-= ; e wH ag g i
, ,T,I*, . D-iN L y'[_" _ _
.4 _. .
gg s
. ,7... , - .i x .
w , .. y. _
- s.
- 4 ozw
'y . 'g
\.
3 l l - * :p ' , ,, 7. D' .dr . .
, .. f n , E ::8" h )
. ' , hh f d.' " - ' ; '
/ C w
4
. g\;. s ,. . 7 O
l s .n. - 1ey;g 4
.-.g _% ,
u
- a. ' ,
- a Cx; ww mW
- a. .s . ~...a. . ,. , . , q, y
- g. . *. - - , '
m-
,e \ .
,y_.,,, .,
i-
. mw a
- s ,.- , . . e, * .
o U
= h J
\ .
k .' l l
' .,7p ' \ c, -
s'.Ud J, ', C ', '""
= ',{- -
> \ .
.= = ,. . ~
'y 4 ,- ; ,
= .* c cy
,. ,,.. ,,..f.-r.- - a
. .e.,
, , m
.Ti q ',* % - ; ,' ' ' >
c.
^, _
e
. 2l .
Q,1 s 3 . 1 >: . q g, y . . e.. :...,=........;..... w
,e
-Q
~ x :::..;.m...v'. .. .
y_
- . .9' . G O- : -
_~ . u;. *:
\ <
.,3*
- . x..
. $5
.. , y y - ,a
,+ [([
s
.\ d .i-4 P
- W' h g.. ,
4s 3 s
@; _ .c m :..-.h.
A A, ' 5-5
, 5.s f
e .:
~c
~
pf ' c f fp
;W:
g
\
A W + ' g
. . [ to" n= hf
\
e'. N [y s k
, pf' ?
y .
'_[f **
^
1+:f.;,, g a f',> W<<lf~Y - _a g nf, "' y ~ # y:),+<N~Q,%x # f 9 a i ' J,;, ' s {g N
\ ' y ip < '
j(6 ( ~PW. e Jhw t
/.m/.,
/
DRAFT Rock quality is important to a major facility such as an underground nuclear power plant. In nature, rock masses are not solid, but generally fractured, with numerous joints. The rock property specifications used as a basis for cavern design were developed from actual borings and tests peg g ed in connee-tion with earlier California non-power plant siting studies. In general, the rock quality is high and is believed to be representative of technically feasible underground nuclear power plant sites within California. The largest engineered openings to date for permanent facilities (as contrasted to temporary openings for mines) have generally been builtinclearspan.[5hf# f f
'I'**#i" plants and are of the order of 100 to 110 feet Although there do not appear to be technological or natural impediments to large open-ings, it was considered prudent to attempt to design the facility with caverns in this demonstrated cavern span range. Thus, a single large cavern housing all systse was not considered a realistic alternative, at this time, for the 1300 MWe facilities of this study.
Three major and several smaller caverns were used to house plant systems. The major caverns are the reactor cavern, the largest; the turbine cavern, for those alternatives with subsurface turbine-generators; and, an auxiliary cavern. The smaller caverns include transformer, diesel generator, and electrical caverns. A detailed description of plant layout is contained in Reference 5-3 and a summary presentation in Reference 5-4. For the Baseline case, the reactor cavern is the largest required, with dimen-sions of 212,101, and 190 feet , height, width and length, respectively. In the design of underground facilities, the principal controlling feature is cavern span. The cavern length nay be extended as far as quality rock exists. A 101' rock cavern span, subsequently reduced by a 3' thick concrete and steel liner to 95', is required to house a 1300 MWe PWR Nuclear Steam Supply System (NSSS) with no major modification to the primary loops. A cavern of this size would provide clearances for equipment, maintenance and inspection comparable to a surface facility. The cavern span required to house a BWR is substantially larger than that necessary for a PWR and is approximately 148 feet before lining. Since the BWR pressure suppression pool is a safecy related, licensed aspect of the NSSS, any modification to that system would require a major licensing activity. To avoid relicensing, the designs accommodate the pressure suppression pool in its current licensed configuration. The result is a larger cavern span. Although there appear to be no technological constraints to construction of such size caverns, conservatism would dictate a demonstration of feasibility and general stability. It may be possible to site smaller than 1300 MWe BWRs in rock, thereby avoiding the very large caverns. This possibility, however, was not evaluated in the study. The rock designs incorporate an extensive system of galleries i.nd intercept drains for groundwater control. Most, if not all, groundwater would be captured before it could enter major caverns. In addition, important caverns such as the reactor and auxiliary cavern, would include groundwater drains in the cavern concrete linings. Such a multiple intercept system would eliminate what has been offered as a major criticism of underground siting: problems caused by groundwater infiltration. 5-7
DRAFT 5.5 Accident Mitigation Systems for Mined-Cavern Facilities The objectives of the accident mitigation system in the mined-cavern plant are similar to those of the berm-contained facility, e.g., avoid uncontrolled failure of the primary containment through passive pressure and temperature reduction. In the rock case, two tunnels, partially refilled with crushed rock, function as a heat sink / expansion volume. Access to the expansion tunnels is controlled by pressure and temperature sensitive valves similar to the rupture discs of the berm-c7ntainment concept. The valves would opeu if the pressure / temperreure environment within containment exceeded the envelope of conditicas postu?,ated for design basis accidents. As with the berm-containment concept, the objective would be to control the pathway of f ailure as a final line of defense. The Level-2 Accident mitigation system is separated from the surface environment by a substantial depth of rock, in these designs approximately 300 or more feet. The surrounding rock would act as a heat sink and such heat conduction was considered in determining both the required tunnel volume and the proper size for tunnel rock fill. Certainly, migration through existing rock cracks and fissures would occur. (The possibility (5'-gt new fissures v uld be created was investigated and found not to occur.) There is no question that such migration is undesirable; however, it is a far better circumstance than direct release of fission products to the atmosphere under the postulated accident conditions. The Level-3 accident mitigation system is similar in concept to the Level-2 with the addition of a vertical, rock / sand filled shaft topped by a tall stack. In the Level-3 concept, noncondensable gases would eventually vent to the atmosphere. By permitting venting, long-term pressures are reduced. 5.6 Technical reasibility of Controlled, Filtered Venting The concept of a filtered vented pathway from, reactor containment is not new. Such filters are in use at the Savannah River Laboratories, in West Germany (SNR-300), and in the United Kingdom (SGHWR). A paper by Goesett, et al., detailed the requirements one could put upon such a system. A filtered pathway:
- 1. "...should, ideally, be as simple and robust as poseible. It should not depend on external services...."
Since this system is passive in nature and extremely simple, it would require little or no maintenance.
- 2. Should effectively remove 99% of the elemental iodine, volatile and particulate fission products.
It will be shown, subsequently, that filters of such efficiency do exist.
- 3. should be capable of being brought ". . . into operation...by local ac tion."
This criterion is interpreted to mean action within the immediate vicinity of the accident and is met by the functions of the rupture discs. 5-8
DRAFT 4.
...can be located conveniently underground, so that shielding for the activity which might be trapped in the bed does not present a problem."
Figure 5-2 illustrates the fact that the filter stack is indeed under-ground and sufficiently isolated from the rest of the plant.
- 5. will have a " . . .c apab ility to withstand the usual range of internal and external hazards..."
The responsibility for this parameter lies in the area of detailed plant design. The rupture discs already display redundancy in their design and emplacement, and could be further protected by construction of missile shields or labyrinths. The stack itself is seismically designed. Additional requirements for the filter would include an ability to deal with large volumes of hot moist air. The volume of crushed rock employed in the annular expansion chamber is expected to be "... capable of condensing the steam which would result from boiling of a in the reactor coolant system...with a large safety margin. gge water present In addition, ". . . work in support of the Savannah River Iaboratory sand filters suggests th filtration efficiency would not be diminished by water deposition."ghe If the moisture laden gases were expected to be a problem, desiccants could be added to the stack filter in conjunction with or in place of a portion of the sand. Dryers could also be an option. .The charcoal filter would require, approximately 10,000 pounds of charcoal. One drawback is the potential heat buildup in the charcoal filter from trapped activity. Whether or not this would be a serious problem, however, ' depends on the details of the design. " Silver plated copper-wool filters (could be) located just upstream from the H03 (hot of f-gas) charcoal filters. The primary purpose of these silver filters is to retain t the iodine, thus preventing an excessive heat load in the charcoal."g_gik of Materials other than charcoal are available. Decomposed granite may be one such substitute. The Gossett treatise concluded, "...it should be possible to find a material whch could be included in the sand bed itself to give sufficient iodine absorption without recourse to a separate charcoal bed (e.g., in the RSS, it is assumed that the backfill surrounding the foundation g provide a 1000-fold reduction in the I-131 release)."go,f a typical reactor will
" Published work has indicated that a particulate removal efficiency of 99.97 percent can be obtained in a sand filter...(and) that a filter area of 150 In square feet and a minimum filter depth of seven feet would be adequate. feasible..."gg application, a filter depth of more than 150 feet would be The physical height of the exhaust stack could also contribute to the efficiency of the filtering process. Using a stack of sufficient heigh i iP ate any residual activity "...by dispersion, diffusion, and decay."i5 gjld d 88 Solutions appear available to the potential problems of incorporation of a vented filtered pathway in a berm-contained, or surface plant.
5-9
D1 AFT 5.7 Summarv Qualified, experienced engineering firms developed realistic conceptual plant layouts, schedule analyses, licensiag assessments, and cost estimates for the sub-surface facilities of the present study. The primary design goal was to use to advantage tha underground environment through an engineered failure pathway for accident mitigation. This approach is unique, since most previous investi-gations sicply extrapolated surface designs to underground settings. The safety design basis for the underground plants was equivalent to that used for current licensing of surface facilities. The accident mitigation systems of the sub-surface plants are passive and supplemental to existing Engineered Safety Features. The berm-contained plant evolved during this study has an inner structure similar to existing containment buildings, and a massive concrete outer dome designed to . withstand overburden loads. This outer dome would function as a secondary containment should primary containment it.tegrity be lost. The engineered failure pathway from the primary containment of the berm plant is via 24 large ducts leading to the surrounding berm (Level-2) or to an enclosed filtered vent stack (Level-3). Access to this pathway is via pressure / temperature sensitive rupture discs designed to open only under C1sss 9 accident conditions. This pathway is simple, passive, redundant and inherently reliable. The principal plant structure in the mined-cavern concept is the native rock itself. The areal extent of the mined plant is much greater than that of the berm facility and detailed cavern orientation is dictated by the existing internal rock stress field. The cavern span required for a 1300 MWe PWR is 101' and, for a BWR, 148'. Accident mitigation for the mined plant is through two tunnels partially filled with crushed rock which function as a heat sink / expansion volume (Level-2). Access to the tunnels is via pressure / temperature sensitive valves similar to the berm plant rupture discs. As with the berm concept, the goal is accident management through a controlled failure pathway. Venting through a filter stack (Level-3) permits long term pressure reduction. Controlled, filtered venting is not strictly an underground plant concept. Such mechanisms are already in use at several facilities, and other studies have concluded that such a technique is a feasible method for control of accident generated pressures. Berm-contained nuclear power plant's appear technically feasible. The large cavern spans necessary for 1300 MWe BWR would require a demonstration of tech-nical feasibility, but, in general, rock siting may also be considered technically feasible. 5-10
DRAFT REFERENCES SECTION 5 5-1 A11ensworth, V.A., et al., Underground Siting of Nuclear Power Plants: Potential Benefits and Penalties, Sandia Lab oratorie_ s , SAND 76-0412, NUREG-0255, 1977. 5-2 Conceptual Design and Estimated Cost of Buried " Berm-Contained" Nuclear Power Plants, S&L Engineers, January,1978. 5-3 conceptual Design and Estimated Cost of Nuclear Power Plants in Mined Caverns, Underground Design Consultants, January, 1978. 5-4 Evaluation of the Feasibility, Economic Impact, and Effectiveness of Underground Nuclear Power Plants, Final Technical Report , The Aerospace Corporation, May 1978. 5-5 Analysis of Public Consequences from Postulated Severe Accident Sequences in Underground Nuclear Power Plants, Advanced Research and Applications Corporation, December,1977. 5 -6 " Siting Concepts and Investigations," Limited Distribution Reports for the Air Force Space and Missle Systems Organizations, K. O'Brien and Associates, San Bernardino, California,1965 to 1967. 5-7 Gossett, G., et al. , " Post-Accident Filtration as a Means of Improving Containment Ef fec tivene ss ," University of California - Los Angeles, Department of Chemical, Nuclear and Thermal Engineering. 5-8 Bin ford , F.T. and E.N. Cramer, "The High-Flux Isotope Reactor," ORNL - 34572, Vol.1, Oak Ridge National Laboratory, May 1964. 5-11
DRAFT 6.0 Economic Implications of Underground Construction Estimating the cost of a nuclear plant is an intricate, complex procedure. Several steps are involved, each of which requires a high degree of accuracy. Such steps include the determination of the basic plant construction costs, labor rates, and the indirect costs of management and licensing. In addition, projected cost estimates for conventional nuclear power plants have been, historically, much less than final costs which suggests an inherent uncertainty in the process. Thus, a determination of the economic impacte of underground nuclear plant construction, a project of great magnitude and complexity, is a difficult, involved, uncertain process. Historical estimates of subsurface nuclear construction projects have exhibited ' a wide range of values, from approximately 10% above surface plant costs to estimated increases of a factor of 2. Table 6-1 lists previously published estimated costs of both berm and rock concepts. However, most previous in-vestigations were based upon cursory plant designs, and did not include extreme accident mitigation features. Such historic uncertainty regarding construction cost estimates was one of the motivations for the detailed cost study undertaken for the present project. The experience and judgement of the engineering firms who developed the esti-mates were invaluable assets in the preparation of realistic cost figures. The cost estimates prepared for this study are probably the most detailed and accurate of any yet made for contemporary underground nuclear power plants. Nevertheless , they contain uncertainty. The values which follow should be interpreted as relative indicators, not as absolute predictors, of cost. The percentage change in cost due to underground construction can be compared to the arbitrary measure of surface plant con-struction costs. To enhance the accuracy of the percentage change, each engineering firm estimated the cost of a comparable surface facility designed according to their organization's procedures and philosophy. Cost estimates are given in dollars, so that the reader may appreciate the magnitudes of the changes i .olved; however, the relative percentage changes are more accurate indicators. Several key specifications directed the cost estimating effort. Costs are referenced to July, 1977 dollars. A common commercial operation date was specified for both surface and underground facilities. This was considered to be representative of actual planning practice where a demand at some future time is identified, and projects are init.iated to meet the demand schedule. As noted above, a " mature" industry with a well defined licensing base was assumed. Consequently, the engineering firms preparing the designs and depen-dent estimates, used best judgement tempered with experience in those areas where no precedent existed. The implications of this assumption will be discussed subsequently. Escalation rates for materials, labor and equipment, and indirect expenses were assumed to be 9% compounded annually. Allowances for funds used during construction (AFDC) were calculated using a simple interest rate of 10%. The impact of higher escalation rates on the differential cost of underground siting was also examined. 6-1
DRAFT Table 6-1 , Previous Cost Estimates of Underground Nuclear Power Planta Cost Year of (compared to Reference No. Estimate Plant Size Type
- surface plant) 6-1 1958 +3 to 7%
6-2 1000 MWe EM +8 to 21% 6-3~ 1971 1000 MWe RK +$6 to $10 million 6-4 1975 1000 MWe RK about + 10% 6-5 1974 1000 MWe BM 5% 6-6 1973 1000 MWe BM/RK +15% minimum 6-7 1970 1100 MWe Depress ed , +50% minimum no berm 6-8 1973 1000 MWe RK 94% minimum 6-9 1977 1100 MWe BM/RK +20% to + 40%
*BM - Berm Contained RK - Rock The elements which comprise the final project cost deserve explanation.
On the basis of project plans and specifications, a direct construction cost was estimated. The direct category included the cost of labor, and purchase costs of construction material and expendable equipment. To the construction cost was added the indirect cost. The indirects included the cost of engineering, management, licensing, and other items as well. There was no uniform allocation of items between the direct and indirect categories. This fact caused varia-tions in cost estimates prepared by the different engineering firms which are more apparent if detailed comparisons are made. The sum of the direct and indirect costs yielded the total construction cost, a sum expressed in dollars referenced to a specific date. Since the real cost of goods and services changes in time, and because indebted-ness during construction creates interest charges,' escalation and interest must be factored into the estimated total construction cost. Escalation and interest charges were determined from the project construction schedule, and reflected the purchase and payment dates for equipment, construction materials, payrolls, 6-2
DRAFT etc. The long time period required for planning and construction of a nuclear power plant, on the order of a decade, makes escalation and interest charges of paramount importance. As will be seen, for the 9% escalation and 10% interest rates assumed for this study, grand total projected costs for the surface plants increased by nearly threefold from the total construction cost level. 6.1 Reference Surface Plant Costs The estimated construction cost for the reference surface plant, as shown in Table 6-2, ranged from a direct cost of $637/kW to $708/kW of electrical output, depending upon estimating approach and design and construction assumptions. These values represent two independent estimates of surface plant cost. The reason for the range in values is more fully discussed in Reference 6-10. These estimated costs are consistent with other published investigations of n g g power plant costs when adjusted for plant size and reference Since the engineering firms developed their underground concept year. cost estimates using assumptions consistent with their reference surface plant estimates, the percentage change from surface to underground is a valid in-dicator of cost impact. 6.2 Underground Plant Costs A detailed discussion of the source for the increased underground plant costs will not be given here. Only major features will be discussed. The interested reader is referred to Reference 6-10 for more details or References 6-11 and 6-12 for explicit discussion of tha mined-cavern and berm-contained concepts, respectively. 6.2.1 Baseline Berm-Contained Concept Table 6-2 gives a total construction cost of $731/kW for the berm-contained facility as estimated by S&L Engineers, or $2.82 billion when escalation and interest costs are considered. The comparable surface plant estimates are
$637/kW and $2.48 billion. Thus, at the grand total cost level, the berm-contained concept is 13.6% more erpensive.
The large domed auxiliary building is a major contributor to increased cost. It accounts for approximately 38% of the total increase in direct costs. The structures category (account 21), overall, contributes about 61% of the increase in direct costs. Over 13% of the direct cost increase is due to the additional excavation and backfill requirements for the berm concept. The accident mitigation system cost listed in Table 6-2 is for a Level-2 design, as defined earlier, and represents about 5% of the direct cost of berm-contained siting. The cost of the accident mitigation system is incidental to the larger cost of this mode of underground siting; however, the Design Level-2 concept effectiveness is contingent upon berm-containment. Table 6-2 also displays a Design Level-3 modification to the reference surface plant. Basically a controlled, filtered venting system, it is functionally similar to the berm Design Level-2, and is estimated to increase surface plant direct costs by about 2%. Note, however, that the direct cost increase of 6-3
DRAFT the surface modification is, approximately $10 million, considerably less than~ the $97 million incurred through bera-containing. 6.2.2 Berm-Contained Concept Variations A large number of variations from the baseline plant configurations were examined. Considerations included foundation variations, meteorological conditions, seismi: design levela, depth of soil cover, changes in schedule, and others. Using a berm-contained , Design Level-2 facility as the baseline condition, the effect of variations considered on grand total project cost, as shown in Table 6-3, ranged from a 2.2% decrease for a configuration with 25 ft, of covering soil to a 17.6% increase for a postulated 19-month delay in schedule. A remote site increased direct costs 8.2%, over that for a Central Valley location, due to increased wage rates to attract workers, decreased productivity, and increased distances for material transportation. All other cost impacts were less than 2.7% of direct costs. This small variation suggests that the Baseline berm plant cost estimate is a reasonable central measure of the cost of berm-containment in California , assuming a " mature" industry. The sensitivity of grand total project cost to increases in schedule fur ther illustrates the major impact of escalation and interest charges. Variations in groundwater levels and associated control measures had little relative impact on total project costs. Either a BWR or PWR can be accommodated at approximately the same cost. 6.2.3 Mined-Cavern Concept Re fe re nce to Table 6-2 indicates a total construction cost (direct plus indirect) for the baseline mined-cavern cencept of $898/kW, or $1.168 billion, which is approximately 27% more than the reference surface facility. At the grand total project cost level, which includes interest and escalation, the cost increase is about 25%. For the mined-cavern concept, the rock itself serves as the principal structure. Thus the structures account of Table 6-2 shows about a 7% decrease for the mined-cavern case from the surface. The larger excavation cost for the mined-cavern plant, however, if reallocated to the structures account, would show a net increase of about 31% for the structures account above that for a surface facility. At the direct cost level, excavation increased approximately $73 million for the mined-cavern plant and this item accounts for about 47% of the direct cost increase. Unlike, the berm-contained estimates, the contingency (a sum included in all major construction to cover uncertainties of design, conditions, schedule and other factors) for the mined-cavern plants is shown as a total. If the contin-gency were distributed against individual costs accounts, about 15% would go to excavatica and backfill; about 37% against direct labor; and 48% would be allocated to cover uncertainty in direct cost of materials and equipment. As an item, estimated increased contingency requirements for the mined-cavern plant contributed about 35% to the direct cost increase. The accident mitigation system for the mined-cavern plant represented about 2% of the direct cost increase for this mode of underground siting. Thus, the cost 6-4
Table 6-2 Surface and Underground Nuclear Plant Cost Summary COST ITEM Pl. ANT COSTS (e) o Direct Construction Surface Subsurface Lvl. -3 Mined NRC Acc. No. Item UDC S&L Hod Berm Caverns 20 Land (a) (a) (a) (a) (a) Structures 196 193 196 252 182 21 22 Reactor 150 171 171 173 153 23 Turbine-Cenerator 133 199 200 207 150 Electrical 37 50 50 58 49 24 25 Hiscellaneous 14 20 19 20 19 27 Excavation & Backfill (c) 2 4 15 75 Accident Hitigation - - 4 5 3 28 Substation 3 4 4 4 4 35 construction Services (d) 8 8 10 (d) 7 u' 91
- Contingency Total 45 (b) (b) (b) 101 Total 578 647 657 744 736 o Indirect 343 181 181 206 432 Total Construction Cost 921 828 842 950 1168
$/kW 708 637 647 731 898 o Escalation, 9% (Compounded Annually) 1100 1033 1049 1105 1233 o AFDC 10% ( Simple Interest) 722 618 612 763 1029 Crand Total Project 2743 2479 2519 2818 3430 Percent Differential - -
1.6 13.6 25 - Notes: (a) - Land and transmission costs not included in estimates (b) -- Contingency cost included as part of individual direct cost item (c) - Not specifically known, estimated at about $2 million (d) - Included in indirect cost subtotcl h2d (c) - Costs are given in millions of dollars p.
Table 6-3. Berm-Containment Concept Cost Sensitivity Analyses Costs Cost Increment tariables Total Construction Grand Total Project (vs. Baseline) ($ million) ($ million) ($ million) (%) I. Baseline 950 2818 -- -- II. Reactor (PWR)*
- 1) BWR 934 2765 -53 -1.9 III. Accident Mitigation System Design (Level 2)
- 1) Level 1 949 2817 -1 negligible
- 2) Level 3 961 2841 +23 +0.8 IV. Site Char 9cteristics (B/L)
- 1) Soil Model 2 945 2805
-13 -0.5
- 2) Rock Model 1 954 2831 +13 +0.5
- 3) Dry Soil 949 2818 --
negligible
- 4) liigh Water Table 969 2875 +57 +2.0 2828 +10 +0.4 i' 5) Dry Climate 951 3050 +232 +8.2
- 6) Remote Desert 1027 V. Seismic Input (0.5g)
- 1) 0.67g 975 2895 +77 +2.7 VI. Minimum Overburden (50 ft)
- 1) 25 ft 929 2757 -61 -2.2
- 2) 100 ft 982 2912 +94 +3.3 VII. Construction Schedule (142 mo) a) Common Commercial Operation Date
- 1) Accelerated (136 mo) 990 2943 +125 +4.4
- 2) Extended (161 mo) 949 2782 -36 -1.3 b) Common Project Inititation
- 1) Accelerated (161 mo) 990 2819 +1 negligible
- 2) Extended (161 mo) 949 3314 +496 +17.6
- Data in parentheses represent baseline (B/L) conditions for variables cited. .
DRAFT
Tatie 6-4. Mined-Cavern Concept Cost Sensitivity Analyses Costs Percent Change Variables Total Construction Grand Total Project (vs. Baseline) ($ million) ($ million)
- 1. Baseline (PWR)** 1168 3430 --
II. Reactor / Turbine-Generator (B/L)
- 1) BWR 1177 --
+0.8
- 2) Surface T-C 1126 -
3311 -3.4* III. Accident Hitigation System Design (Level 2) ,
- 1) Level 1 1165 --
-0.3
- 2) Level 3 1166 -- -
-0.2 IV. Site Characteristics (B/L)
- 1) Better Rock 1167 --
-0.1
- 2) Poorer Rock 1172 --
+0.3 p 3) High Water Inflow 1174 --
+0.5
-4 4) Dry Climate 1170 --
+0.2
- 5) Flat Terrain 1191 --
+2.0
- 6) Remote Desert 1168 --
negligible V. Seismic Input (0.5g)
- 1) 0.67g 1181 --
+1.1
- 2) 0.3g 1159 --
-0.8 VI. Depth of Cover (200 ft)
- 1) 1000 ft 1260 --
+7.9 VII. Construction Schedule (123 mo)
- 1) Accelerated (104 mo) 1192 3560 +3.8*
.2) Extended (141 mo) --
3526 +2.8*
- Percent changes calculated with respect to Grand Total Project Cost. All other variations were calculated with respect to Total Construction Cost.
** Data in parentheses represent baseline (B/L) conditions for variables cited.
DRAFT
DRAF1 of the accident mitigation system is nominal once the major cost of mined-cavern siting is incurred. 6.2.A Mined-Cavern Concept Variations A number of factors were examined to establish their impact on plant costs. The factors and their net effect are given in Table 6-4. The baseline mined-cavern plant assumed a sub-surface turbine generator (T/G). However, as shown in Table 6-4, the surface placement of the T/G resulted in a decrease of total project cost by about 3%. The increased steam lengths required for surface placement were more than offset by the decreased excavation Costs. Increasing the depth of cover, which required that the plant be located further into the hillside, increased construction costs about 8%. Achieving pressure resisting capability through deep burial is costly. Three vaciations in rock quality were examined. Although all were considered adequate for caverns of the necessary dimensions, changes in general rock quality increased or decreased rock bolting and/or cavern lining requirements. The change in direct cost was from 0.1% decrease for the best rock to 0.3% increase for the poorer rock. The seismic design level was varied from 0.3g to 0.67g. At the higher level, direct costs increased 1.1% and at the lower level, decreased 0.8% from the baseline costs at 0.5g. The range of cost is small compared to that expected for a surface plant. The mined-cavern plant design used lateral bracing of main structures against cavern walls to decrease the amplificacion of accelerations which commonly occurs in surface facilities. Thus, plant costs increased nearly in proportion to seismic design levels rather than non-linearly as for surface plants. Similarly to the berm-containment concepts, the mined-cavern facility costs were sensitive to schedule changes. Money invested in a thorough, pre-construction site investigation would be a prudent investment to reduce sources of uncertainty and potential later delay. 6.3 Construction Schedules Figure 6-1, after Reference 6-10, summarizes four much mora detailed schedules from References 6-11 and 6 -12. Each underground concept construction schedule is compared to a reference surface plant schedule produced using similar assumptions. The engineering organizations which prepared the study schedules each approach nuclear plant design from slightly different perspectives. One assumes a standardized design whereas the other utilizes a modified existing plan. For these reasons, and others pertaining to licensing assumptions , the surrace plant schedules differ by 16 months. The underground concepts bo th have longer schedules than their respective surface counterparts; 22 months more for the berm concept and 19 months for the mined-cavern facility. In both cases, a major excavation program must precede actual in plant construction. 6-8 .
No, $ g O h of
- Surfac e (SkL) Mo. 12 11 10 9 8- 7 6 5 4 3 2 1 i i i i i i i i r De n
- Licensing to LWA 48 . CP RPV Construction to Te sting 66 Study A W 7 CO Fuel Loading to CO LWA h 6 g9 120 Berm-Contained Facility Start Load T-G Fuel Subsurface Erection Licensing to LWA 46 t_ 7 7 CP Construction to Testing 90 Be n g ; -,
^
{ Fuel Leading to CO d 142 Study p y Apa g Lo* I r ed i Surface (Gkli) " N Set Licensing to LWA 20 1 7 7 CP RPV Study w Constructior to Testing 78
- c '
CO Fuel Loading ,a CO 3 104 Mined Caveen Facility LWA Start [ g LOW l g T-G Foo L Subsurface Erection Licensing to LWA 20 Begin ." ; 7 CP f Construction to Testing 97 NSSS _i '
. m _,
Fuel Loading to CO __,_6_ Y L A Set
"~
SkL - Sh L Engineern GkH - Gibbs and Hill Engineers and Constructors Figure 6-1 Plant Construction Schedule Summary DRAFT
DRAFT From a construction perspective, the berm plant is more compact than a surface facility. The limited construction accessibility decreases the rate at which steel and concrete can be placed and increases the time for movement of men and materials. The berm plant requires additional quantities of reinforcing steel, concrete and formwork, principally for the foundation mat, thick walls of the dome, and the " tunnel" structure which connects the auxiliary building to the turbine generator building. For the mined-cavern plant, rock bolting and cavern lining can proceed almost concurrently with excavation. However, the limited access to the principal caverus and reduced use of construction cranes, as compared to a surface facility, also contribute to schedule lengthening. 6.4 Project Cash Flows Figures 6-2 and 6-3 present normalized cumulative cash flows for the berm-contained and mined-cavern concepts and also include the appropriate surface plant cash flows. If the money disbursed during planning and construction of a nuclear plant is added over time, the characteristic "S" shaped curve is obtained. While not unique to nuclear plants, the curve typically illustrates a gradual rate of increasing expenditure, reaching a peak rate about halfway through the project, and tapering off thereafter. The curves have been normalized far convenience. The actual total costs are about 14% more for the berm plant compared to the reference surface and 25% more for the mined-cavern facility. Commonly, it is assumed that if a project requires longer to build, then project costs will greatly increase due to escalation and interest. For the purposes of this study, the surface and underground plants were specified to have common commercial operation dates and the project initiation date was adjusted accordingly. Such scheduling tends to minimize accumulation of escalation and interest charges. For this reason, the underground concepts are not automatically more expensive by a minimum of 19 months escalation and interest applied to surface plant cost. 6.5 Surface Plant Accident Mitigation System Costs The construction cost increases for incorporating a modified Level-3 accident mitigation system within a surface nuclear facility have been given. In conparison to the reference surface plant, it appears that such a system could be added with relatively minor cost increases. Total construction costs were estimated to increase by approximately $14 million (2%). This nominal cost increase could make the modification of the surface plant an attractive concept for further investigation. It should be noted, howeier, that the berm-contained facility, as well as the mined-cavern concept , has an added barrier to radionuclide release, not inherent in the surface facility modification. The structure of the berm-contained plant would be available as a secondary containment upon failure of the primary reactor containment. Auxiliary, manual isolation systems have been added to 6-10
B 1.0 - M* (1
.9
.8 T
~
o _ LWA Issued
.7 g_ CP Issued A - R PV Se t
, .6 1;7 - T-G Delivered q y h O - Fuel Load 5 - A L - Surface Berm-4 - b
.3 %
g m o 2 *
*1 - $ $ '
o O
'/ ~ 3 -
C
,, 4 80 8z y,,,,
84 " " g
II-9 Fraction of Total Cost M - w w a m m w m e o 6;r w c
- I I I i 1 i i I 1 3 n
m i'" - D4 >e O
- W 0 F cM H,
- t 5 (>
[ n EO < ~ 3 eO m - 1 - o * % c a ir E lir EEa E g g Rock Cavern Plant Start g 5' L, p ' Surface Plant Start
# . O 2 N e y 4 y x
n
' h a
n * % F - e
; g o =>
0 O
,I m -
C k D W c O ' n @ W T ' O E m Jp E -4 o p "l" $* m o G - O c 4 . EE a o a
DRAFT assure a secondary containment function, if necessary. Fur thermore. , there are accident sequences which the underground concept would more effectively mitigate, such as breach of primary containment by an internally generated missile, than a comparable surface facility. Thus, direct comparisons between construction costs for modified surface and underground plants, under the assumptions made for the present study, may not be valid measurements of cost-benefit tradeoffs. 6.6 Implications of Mature Industry Assumption Cost estimates were prepared under the assumption that the conceptual study designs reflected a " mature" industry, e.g., one for which acceptable design procedures, solutions, and regulatory base were adequately established. Conven-tional wisdom holds that the time required to develop a complete licensing base would increase the estimated cost of underground nuclear power facilities by some great, but uncertain, amount. There is no doubt that nuclear capital costs have escalated greatly. Several factors have been advanced to explain the historical increase, including dramatic increases in labor costs; evolution of the environmental movement; increases in licensing costs; decrease in labor productivity; and unpredicted, unprecedented increases in interest and escalation. However, Reference 6-12 cuggests that the nuclear industry is ncw " mature" and notes that "with some 60 nuclear plants now in operation and over 200 in design; project cost estimates and actual installed costs differ now chiefly because of changes in the actual escalation rates experienced as compared with those assumed, and not because of significant design changes or construction technique modifications." In other words, if cost estimators could predict with certainty some factors over which they have no control, namely future interest and escalation rates, then estimates would be extremely accurate. This implies that licensing changes in mid-construction, so-called "ratcheting," are a less significant factor in increasing costs for surface plants now than previously. A recent study of schedules for twelve nuclear power units supports the preceding implication (6-13). The significant factor is when schedule delay occurs as differentiated from how it occurs. This point will be discussed further in the following section. It is tempting to draw parallels between the licensing period for the floating nuclear power plant (FNP) as proposed by Offshore Power Systems and the duration of licensing for an underground nuclear plant. Each is a novel siting concept with unique issues not yet encountered with conventional surface plants. The FNP licensing period has extended from April 1972 and is still underway. Much of the regulatory concern surrounding the FNP, however, has focused on the possibility and consequences of extreme reactor accidents -- precisely those events that underground siting can effectively mitigate. Thus, conclusions derived from the licensing experience of the FNP cannot be directly transferred to the underground concepts of this study. It is clear that some questions will need resolution, including design criteria for the berm plant auxiliary building dome, interfaces between the sa fe ty augmentation systems of this study and conventional containment, long term 6-13
DRAFT stability of rock caverns, and others , and that the first few facilities will bear these expenses. The concepts developed in this first study, however, use state-of-the-art technology, conventional arrangements of the nuclear steam supply systems, all currently required sa fe ty systems, and contemporary con-struction procedures. For these reasons, no major licensing issues are foreseen and the licensing period for the underground concepts would probably be less extensive than for other novel siting concepts. It is also possible that, once a complete licensing base has been established, underground, augmented facilities could be licensed more rapidly than surface plants. 6.7 Variance in Estimated Costs The preceding section discussed the impact of proceeding from an " immature" to " mature" industry. The maturation process includes refining criteria and designs, resolving licensing questions, and developing efficient construction procedures. This process can affect final plant costs. Another aspect of the accuracy of plant cost estimates is the variance of the estimate; that is, what is the likelihood that estimated costs will be final costs? This question was examined through use of a probabilistic model, completely described in Reference 6-14, which divided the licensing and construction period into five stages. Then, in conjunction with plant designers and constructors, cash flows for various cost estimate accounts were developed for each stage and the probability of achieving the cash flow estimated. Through this process, the schedule uncertainty was transformed into plant cost uncertainty. This was done for both surface and bem-contained plants. Although the uncertainty in the rock-sited plant cost estimates was not explicitly evaluated, the general conclusions obtained for the berm concept are believed applicable to that concept as well. Table 6-5 summarizes the results of the probabilistic analysis and compares surface and berm-contained plants. Costs are expressed as ratios as defined in the Table. The design costs were based on the conventional engineering estimates as reported in Reference 6-10. Slightly different assumptions pertaining to interest and escalation rates were used for this aspect of the study than for the engineering cost estimate. The relationships between various costs, however, remain approximately the same. The maximum plant cost reflects the most pessimistic scenario where, for example, delays in all development stages occur. The minimum plant cost would result when the assumptions used in preparation of the design costs are achieved during licensing and construction. Since a " mature" industry was assumed in the preparation of design costs, these probably represent lower bounds to estimated costs. The expected costs are those most likely to occur. The 85% confidence cost represents an 85% chance that the given cost would be achieved. Alternatively, there is a 15% chance that costs could be higher than the specified value. Column cae of Table 6-5 indicates an expected value for final cost approximately 12% higher than design cost for a surface nuclear plant. The comparable number for the berm facility, from column two, indicates a 19% increase. The greater variance in completion times for various stages of the bem plant construction result in higher expected values for final plant costs. Expressed in another manner , there is a probable 7% penalty (1.19 - 1.12) for sub-surface construction as compared to a surface facility due strictly to uncertainty in Construction. 6-14
DRAFT Column three of Table 6-5 is also instructive. This column reflects the higher design cost for the berm plant, approximately 18%, as compared to the surface plant. All other measures of cost increase commensurably. Table 6-5 Probable Plant Costs (Ratios to Selected Cost Measures) Berm (3) Berm (4) Surface (1) Berm (2) Surface Surface De sign 1.0 1.0 1.179 1.052 Expec ted 1.121 1.191 1.405 1.253 Maximum 1.433 1.537 1.813 1.617 Minimum 1.0 1.0 1.179 1.052 Standard Deviation .108 .144 .170 .152 Range .433 .537 .633 .565 85% Confidence Level 1.189 1.297 1.5 30 1.365 (1) surface costs normalized by surface design cost (2) berm costs normalized by berm design cost (3) berm costs normalized by surface design cost (4) berm costs normalized by surface expected cost In column four, berm plant costs have been divided by the expected cost of a surface plant. This column shows that the expected cost of the berm f acility is approximately 25% greater than the expected cost of the sur face plant. The range of probable berm costs are from +5% to +62% of surface plant expected costs. Furthermore, at any given level of confidence (risk) an underground plant is more costly than a surface plant. At the 85% confidence level, the berm plant is 36% more expensive than the surface, expected cost. The surface plant cost rises 6% from expected cost to achieve an 85% confidence level. In summary, conventional engineering cost estimates are only single values which do not convey any measure of the accuracy of the estimate. Based on expected costs, and given interest and escalation rates, the baseline berm plant is estimated to be 25% more expensive with a greater likelihood of exceeding its estimated value than a surface plant. 6-15
DRAFT 6.8 Impact of Escalation Rates on Relative Costs Appendix C of Reference 6-14 reports the sensitivity of total plant costs to various escalation rates, with constant interest rates. The long time periods required for planning, licensing and construction of major power facilities, both nuclear and non-nuclear, makes total estimated plant costs very sensitive to assumed escalation races. The objective of this project's cost estimating activity was not to predict, in an absolute sense, the cost of future nuclear facilities, but, rather, to determine the relative costs of various siting concepts. Consequently, the rate of escalation for labor equipment and material was varied from 5%, in steps of 5%, to 25%. Table 6-6 summarizes the con-E quence to total plant cost of selected escalation rates. Table 6-6 Impact of Escalation Rate on Relative Cost Plant Costs
- Escalation Surface Berm Differential Between Rate % ($ billions) ($ billions) Surface and Berm (%)
5 1.59 1.90 19 10 2.44 2.84 16 15 3.71 4.20 13 20 5.55 6.13 10 25 8.19 8.85 8
- Grand total costs are generally comparable to Design Costs of Section 6.7.
Small differences in assumptions of interest rates and the number of indi-vidual account cash flows considered cause slight variations. The numbers indicate that although escalation rates can have significant impact on absolute plant costs, the differential increase of berm-containment over conventional surface siting is less pronounced. The potential differential increases in plant costs for the reasons discussed in Section 6.7 overshadow the effect of escalation - considering relative costs, that is, percentage changes. The 9% general escalation rate selected for preparation of the principal plant cost estimates is considered reasonable and adequate for the purposes of this study. 6-10
DRAFT 6.9 Summary Previous estimates for sub-surface nuclear construction exhibit a wide range of values for increased construction costs. The cost estimates of the present study were developed, specifically, for California by experienced engineering firms. As a result they represent the most detailed and accurate of any yet made for contemporary underground nuclear plants. Costs estimated were referenced to July 1977 dollars. A common commercial operation date was specified for both surface and underground facilities. A " mature" underground industry base was assumed. A direct construction cost was estimated which included cost of labor, material, and equipment. Indirect costs included engineering, management, and licensing. Escalation and interest were factored into the grand total project cost. The berm-contained plant was estimated to be about 14: minimum more expensive than the reference surface facility for grand total costs. The massive concrete dome was the dominant contributor to increased cost. The level-2 accident mitigation system represented about 5% of the direct cost for aerm-contained siting. When the uncertainty in licensing and construction was incorporated in a proba-bilistic fashion into cost estimate, the expected cost of a berm-contained facility rase to 25% greater than reference surface costs. Mined-cavern siting was estimated to increase grand total project cost over surface cenetruction by about 25%. Excavation accounted for about 47% of direct cost increase. The cost of the mined-cavern accident mitigation system is nominal, representing about 2% of the direct cost increase. The engineering organizations approached their design tasks from slightly different perspectives. One assumed a standardized design, while the other used a modified existing plan. Thus, construction schedules for the two reference surface plants differ. Underground construction was estimated to take 22 months longer for berm siting and 19 months longer for mined-cavern siting. The cost of modifying a surface nuclear facility with a Level-3 accident mitigation system was estimated. The cost increase of such a concept appeared to be minor with direct construction costs increasing by approximately 22. This surface modification, however, does not display the secondary containment features incorporated in in the underground facilities. The designs developed use state-of-the-art technology and conventional arrange-ments, construction procedures, and safety systems. No major licensing issues are foreseen. The long time periods involved in nuclear plant construction make estimated cost sensitive to assumed escalation rates. Escalation rates significantly influence absolute plant costs. As escalation rates were increased in parametric studies, the relative cost differential of surface and underground siting narrowed. 6-17
DRAFT REFERENCES SECTION 6 6-1 Beck, C., Engineering Study on Underground Construction of Nuclear Power Reactors, USAEC Report,,AECU-3779, 1958. 6-2 Kroger, W., J. Altes, K. Kasper, " Assessment of Underground Siting of a Nuclear Power Plant With Pressurized Water Reactor", Institute for Nuclear Safety Research, Federal Republic of Germany, date unknown. 6-3 Rc'ers, F.C., " Underground Nuclear Power Plants: Environmental and Economic Aspects," Nuclear News, May 1971. 6-4 Kammer, W.A., and M.B. Watson, " Underground Nuclear Power Plants With Surface Turbine Generators," Nuclear Engineering and Design, 33, 1975. 6-5 Karpenko, V.N., and C.E. Walter, " Underground Siting of Nuclear Power Reactors," Proceedings, Symposium on Siting of Nuclear Facilities, International Atomic Energy Agency and OECD Nuclear Energy Agency, Vienna, Austria, 1974. 6-6 Holmes and Narver, Inc., California Power Plant Siting Study, Anaheim, California, May 1973. 6-7 Fogarty, D.J., B.R. Laverty, H.F. Brush, D.W. Hulligan, " Design Concept for an Underground Nuclear Generating Station at a Coastal Urban Site", Paper, 35th Annual Meeting of the American Power Conference , Chicago, Illinois, May 1973. 6-8 Ortega, 0.J., Manager of Generation Engineering and Construction (1972), Southern California Edison Company, prepared testimony be fore the U.S. Atomic Energy Commission Atomic Safety and Licensing Board Hearings, San Onofre Units 2 and 3, 1972. 6-9 Allensworth, J.A., et al., Underground Siting of Nuclear Power Plants: Potential Benefits and Penalties, Sandia Laboratories, SAND 76-0412, NUREG-0255, Albuquerque , New Mexico, August , 1977. 6-10 Evaluation of the Feasibility, Economic Impact, and Ef fectiveness of Under-ground Nuclear Power Plants, The Aerospace Corporation, ATR-78 (7652-14)-1, El Segundo, California, May, 1978. 6-11 Conceptual Design and Estimated Cost of Nuclear Power Plants in Mined Caverns, Underground Design Consultants, January, 1978. 6-12 Conceptual Design and Estimated Cost of Beried " Berm-Contained" Nuclear Power Plants, S & L Engineers , January,1978. 6-13 Ryan, R.G., "The State Side of the Siting Equation: Some Case Studies," Nuclear Safety, Vol-19, No.1, January-February, 1978. 6-14 Terasawa, K., R. O'Toole, and M. Goldsmith, Probabilistic Analysis of the Cost for Surface-Sited and Underground Nuclear Power Plants, Jet Propulsion Laboratory, Document No. 5030-223, Pasadena, California, March, 1978. 6-18
DRAFT
- 7. 0 Consequences of a Major Radionuclide Releast It was noted in Section 3 that reactor accidents can range in severity - that there is no unique reactor accident. It was also observed that as the severity of accidents tended to increase, the likelihood tended to decreasa. This study has examined the ability of safety-augmented underground nuclear power plants to reduce the public consequences from the most severe, yet least likely, reactor accidents.
In order to evaluate the effectiveness of underground nuclear plants a basis and set of measures for comparison is needed whether they are economic costs, health effects, cost to implement, or all taken together. This work has assumed that similar accidents would occur in conventional surface facilities and in the underground designs of this study. Consequences to the public were calculated and in subsequent sections, early- and long-term health effects, economic costs and environmental consequences will be compared. In addition, because certain actions such as evacuation can be taken to reduce consequences, and because of variations of results with health models employed, a number of parametric cases were considered. The numbers which follow, however, should not be construed as the total basis for evaluating the effectiveness of underground siting. The results of the consequence calculations to follow assume two important. conditions: (a) failure of the surface plant containment building; and (b) successful operation of the pressure relief systems of the underground plants. The implication of the first assumption is that only one event of a spectrum of possible accidents has been considered. The possibility that the pressure relief system might not function, the second assumption, will be addressed in Section 8. A numerical risk analysis, which would reflect all accidents and consequences, from benign to severe, was not performed for several reasons. First, a detailed analysis such as that of the RSS requires complete designs, and, probably an actual constructed facility for evaluation. Since the designs of this study were conceptual in nature, although far more detailed than previous studies, they are insufficient for a rigorous rir k analysis. Second, there are fund-amental questions over the ability of risk analysis to predict, in an absolute sense, future risk. Third, the question of interest here is the accident mitigation capability of a concept, not, necessarily, the frequency of major accidents. Fourth, since the conceptual designs were developed with all existing regulatory criteria and requirements foremost in the designers' minds, the ability of underground plants to control less severe accidents would be comparable to contemporary surface facilities. Thus, attention focused on extreme, Class 9, accidents. This section presents, in summarv fashion, the method by which accident conse-quences were calculated and the important considerations of such calculations and consequences. The following discussion is not an in-depth presentation . of radiological health models and criteria, consequence models, radionuclide transport calculations, or any of the subleues associated with the prediction of accident consequences. For the reader unfamiliar with health physics terminology, a brief explanation of some terms and their significance is provided. A more complete description of the subjects of this chapter may be found, in increasing depth of detail, in References 7-1 and 7-2. 7-1
DRAFT 7.1 Consequence Model Significant adverse consequences could occur following an extreme nuclear power plant accident and subsequent release of radioactivity. These consequences would be adverse health effects, socio-economic costs, and environmental impacts. Important factors, among many, which influence the severity of consequences are the radionuclides released; the radioactivity dis pe rsion, transport, and deposition modes; and, the radioactivity concentrations producing population exposure doses. An analysis of potential accident consequences was performed. The analysis was conducted for both the baseline surface and Level-2 underground plants. Results wre then compared to determine the relative effectiveness of under-ground siting in mitigating adverse consequences following a severe accident. The various parameters utilized in the consequence model represent realistic methodology and criteria. Radiological health effects were based upon current health physics standards. Economic data reflected the latest California specific information (1977) from the United States Department of Agriculture, the California Department of Food and Agriculture, and the California State Board of Equalization. Population and meteorology data were re flective of California conditions. In this light, study findings are pertinent and strictly applicable only to California. The major considerations in the consequence model are illustrated in Figure 7-1. A detailed discussion concerning the characteristics of these parameters is given in Reference 7-2. It should be noted that the only differences in the model, for assessing the consequences of a surface release as opposed to an underground release, are in the major radionuclide transport mechanisms. A surface release would be characterized by atmespheric transport, while sub-surface release products would be transported through soil or rock before recching the atmosphere. At that time, any remaining radioactivity would be available for atmospheric transport. Health effects calculated by this analysis are not absolute predictions of consequences following a severe reactor accident. As previously noted, however, the calculated health effects provide a useful scale for comparing the accident mitigation effectiveness of underground plants relative to conventional surface facilities. The following sub-sections briefly summarize principal features of the conse-quence model and criteria used to calculate accident consequences. 7.2 Criteria and Rationale 7.2.1 Radionuclide Release Should an extreme reactor accident occur, only portions of the initial reactor core radionuclide inventory would be expelled from containment. This stu somewhat similar to that of the PWR-2 category of the RSS.jygsumedThea release release was slightly modified to reflect certain features of the specific ac ident sequences used to construct the design envelope. Nevertheless, the radionuclide release used has characteristics suf ficiently 7-2
, GROUND WATER , ACTIVITY CONCENTRATIONS TRANSPORT MODEL ARRIVAL TIMES RADIONUCLIDE HEALTH EFFECTS RELEASES ECONOMIC COSTS M F ARRIVAL DPOSURE
- ENVIRONMENTAL METEOROLOGY
& CESSATION DOSE HISTORIES 7 IMPACTS AIRBORNE 6ROUND WATER =
CONCENTRATIONS DATA 6ROUND DEPOSITION ECONOMI'C DATA ATMOSPHERIC DENSITIES
, HEALTH EFFECT DISPERSION CONVERSIONS ENVIRONMENTAL DATA COUNTER MEASURES DEMOGRAPHY Figure 7-1 SCllEMATIC DIAGRAM 0F CONSEQUENCE CALCULATIONS DRAFT
DRAFT close to the PWR-2 release that significant health consequences are comparable. The radionuclide releases of this study are compared to the RSS PWR-2 release in Table 7-1, below. Table 7-1 Fission Product Release Characteristics Fraction of Total Inventory Released Radionuclides PWR-2 CEC Study
- 1) Xe, Kr 0.9 0.8
- 2) I 0.7 0.7 2a) Org. I 0.007 0.007
- 3) Cs, Rb 0.5 0.4
- 4) Te, Sb 0.3 0.3
- 5) Ba, Sr 0.06 0.05
- 6) Ru, Ho, Rh, Tc, Co 0.02 0.4
- 7) La , Nd , Y, Ce , Pr , Nb , 0.004 0.002 Am, Cm, Pu, Np, Zr 7.2.2 Exposure Pathways Exposure pathways were identical for the consequence calculetions for surface and underground plants. The predominant exposure pathways following a surface plant release would be as a consequence of airborne dispersion and subsequent deposition of radionuclides. Specific pathways include internal irradiation from inhaled airborne radionculides, external irradiation from ground deposited contamination and the passing radioactive cloud, and intake of contaminated food or water.
The radioactivity deposited on the ground following a surface release would be a source of long-term exposure. Two effects which tend to decrease ground source radiation intensity, decay and weathering, were included in the calculation of long-term exposure. Since consumption of contaminated food and water could be easily avoided by monitoring programs, this exposure pathway was not evaluated. 7.2.3 Dispersion, Transport and Deposition The X00D0Q computer code was used to calculate the average concentration of radioactive effluent in the air (X/Q) and the average deposition per unit area of radioactivity on the ground (D/0) with distance from the reactor site. Although the X00D0Q code is reasonably sophisticated in its treatment of a very complex problem, no techniques presently exist which can predict with precision the absolute values of reactor accident consequences. They can only be esti-mated. There are too many uncertainties in modeling and the required data to permit the consequence estimates to be used for any purpose other than showing the relative changes from site to site, or between concepts. The releases from the surface plant were considered to be at ground-level, " puff-like" in duration, and to occur at about 3.5 hours after accident 7-4
DRAFT initiation. For the underground plants, the releases were to the expansion / condensation zones with subsequent migration of the radionuclides through the overlying soil and rock. The latter problem, ground migration, was treated by two-dimensional numerical transport. In the underground cases, when radio-activity reached the surface it was assumed available for atmospheric trans-port. Soil properties from representative locations in California were used in the sub-surface transport calculations. A parametric investigation of groundwater transport was performed. It was found that the rate at which radionuclides moved through the soil was overwhelmingly dependent on site-specific soil properties. 7.2.4 , Site Characteristics ' To assure that the consequence calculations were reasonably applicable to California, four representative geographical regions were chosen and locations selected from within these regions. Three of the sites evaluated (A, B, and C) reflect different portions of the State's Central Valley.and the fourth, (D) the South-Eastern desert region. Meteorology and demography typical of these regions were used in the consequence calculations. For each of the postulated sites and each of four seasons, the weighted averages of six wind speed classes and seven stability categories for the most probable wind direction were used for calculation of accident consequences. In addition, a " severe" condition was treated which assumed winds blowing in the direction of the maximum population concentration within a distance of 50 miles. Beyond a distance of appros.imately 100 miles, the average California population density of 130 persons per square mile was used. This assumption tends to overestimate long-term consequences at some hypothetical sites and, possioly, underestimate the magnitude at others. However, since the study objective was to determine the relative performance between two siting concepts, this is not considered a serious shortcoming. 7.;.5 Evacuation and Interdiction Criteria Any persons who would accrue a dose of more than 25 rems within 30 years as a consequence of the accident were assumed to have been evacuated. A similar interdiction level was used. Interdiction refers to the level at which pre-ventative action is taken. The principal criterion was that, for modeling purposes, people were not permitted to return to an area if they would receive more than 25 rems within 30 years. Two variations of interdiction criteria, 50 rems and 100 rems within 30 years, were considered, since the cost of interdiction was a major factor in the overall accident-induced cost. Four different evacuation cases were considered. Times for evacuation com-pletion ranged from 1.5 hours to 24 hours and varied with distance from the reactor site. The base case, the results of which are reported subsequently, assumed four hours for evacuation completion from the time of containment failure. 7.2.6 Health Effects Indices and Models A generally conservative approach was taken in the consequence calculatiors. A linear dose-response model was assumed for the calculation of most health 7-5
DRAFT effects, although additional analyses were made using a 25 rem threshold model. In the general case, no dose rate effectiveness factors were used although sensitivity calculations were made assuming a 0.2 dose rate effectiveness factor. Table 7-2 shows risk factors, which convert average dose received to average incidence of a listed health effect, used in the calculation of general health consequences. Table 7-2 Radioactive Dose Conversion Factors for Health Effects
- 1. Early Fatalities Bone Marrow: LD 50
= 430 rem (Assuming Supportive Treatment) Lung Dose:
LD50""' #**
- 2. Early Illnesses Various 0
- 3. Latent Cancer Deaths 194/10 man rem 0
- 4. Latent Thyroid Cancers 134/10 man rem 6
- 5. Latent Thyroid Nodules 197/10 man rem 6
- 6. Genetic Disordere 132/10 man rem 6
- 7. Genetic Spontaneous Abortions 42/10 man rem
- 8. Prenatal Deaths LD = 80 rem (1st tr ime ster)
LD = 340 (2nd i 3rd trimester)
- 9. Temporary Sterility S 50
= 250 rem (ovaries)
S 50
= 80 rem (testes)
A " plateau" model, which assumes constant rates of cancer incidence due to earlier exposures, was used in the calculation of numbers of cancers at long times from initial exposure. Latent periods of from zero to 15 years, depending on cancer type were used. In contrast to the RSS, the plateau period, the time when cancer incidence occurs at nominally constant rate, was taken as the lifetime of the individual rather than a shorter (30 year) period. This has the effect of increasing the number of calculated cancers by approximately 1.6 ever the RSS values. - General uncertainty exists in the scientific community over many aspects of the calculation of health e f fec ts from radiation exposure. Among these uncertainties are:
- a. khether the absolute risk model or the relative risk model is appropriate.
- b. k*hether the presence or absence of a threshold exposure dose should be used for estimating late health effects; and, 7-6
DRAFT
- c. Whether late health effects depend on exposure rate.
As noted, sensitivity calculations were made to assess the impact of other basic assumptions on the number of calculated health effects. In all cases, however, the absolute risk model was used to calculate health effects. The radiation-health ef fects relationships selected for use provide a generally conservative, accepted basis for calculation of possible extreme accident consequences. The numbers to be presented should be regarded as only indicative of potential consequences since a lack of fundamental scientific information prevents precise calculation. 7.2.7 Economic Costs Any major accident has associated economic costs. In the case of a nuclear plant accident, the economic impacts include the costs of evacuation, reloca-tion, the interdiction of productive land, hospitalization of affected persons; and could include the cost of the plant itself and revenue lost from the loss of generating capacity. Reducing these economic impacts of an accident would be another measure of effectiveness. The latter factors, plant cost and generating capacity, were not included in this analysis, however. In the event of an extreme surface accident, certain segments of a population and certain locales within the release vicinity would be subject to exposure. Evacuation would be undertaken for those people exposed to potentially dangerous radiation levels. Consequently, accident related economic costs are sensitive to the number of people and amount of property permitted to receive a certain level of exposure; i.e., evacuation and interdiction criteria. As will be seen, one measure of effectiveness, economic costs, is inversely related to another, late health effects. The fulcrum of this relatiocahip is the interdiction criteria employed. For increased acceptable exposure levels economic costs are reduced, with a concurrent increase in health effects. This relationship is illustrated in Section 7.3, Results of Consequence Calculations. The hospitalization of affected people, following an accident, would also result in an economic impact. These costs would be due to short-term acute effects , such as early mortality and illness, as well as long-term effects , as in the case of surgical treatment of cancers. Similar to evacuation and relocation costs, hospitalization costs are sensitive to the interdiction level employed (that exposure level where public authorities would take action). Decreasing the allowable population exposure would decrease resultant hospitalization costs but increase evacuation and relocation costs. As will be seen, however, since evacuation and relocation dominate the economic impacts following a severe accident, any savings in medical treatment costs due to lower interdiction levels would be overwhelmed by increased evacuation and relocation costs. The loss of productivity for farmland which must be taken out of production would be another economic impact following an extreme accident. This cost is obviously directly tied to the interdiction criteria employed. Land and improved property value losses, crop and livestock losses, and medical treatment costs were included in the consequence calculation. In addition, the cost of evacuation and relocation, if necessary, was included. Any area for 7-7
DRAFT which the population would receive a dose of more than 25 rems in 30 years was evacuated and agricultural lands within the region were taken out of production (i.e., interdicted). "The value of the agricultural land was e s tima te d from actual 1977 average California figures and a nominal value of $1,000 per acre was assigned to f armland (average of irrigated and non-irrigated) and $300 to pasture. For farmland, the cost is assumed to reflect the discounted value of future crops. The cost of health services will vary according to the degree and type of illness. For this study, gross hospitalization costs were assumed to range from $3,000 to $15,000 depending on illness. Composite evacuation and relocation costs were calculated at the rate of $25 per person per day, decreasing with time. These values re flect $20 per person-day for evacuation, $3,500 for relocation, and $23,100 per person for developed property values. As an effective concept, underground siting seeks to limit accident costs by making it unnecessary to evacuate and relocate populations; eliminating immediate health ef fects with associated hospita'.ization; and, avoiding the need for interdiction. Its ability to accomplish these goals is detailed in Section 7.3. 7.2.8 Health Effects Nomenclature The health effects arising from a population exposed to greater than normal radiation are generally considered to be of two types, early and late, based upon the time of symptom appearance following exposure. For this study, early effects are considered as those which occur within one year following exposure, and late ef fects, those that occur later than one year after initial exposure. There are important differences in the consequent health effects of those exposure categories. In addition to a categorization of health effects by time of symptom appearance, distinction must be ma 'e as to the time of radiation exposure. Acute exposures, for this study considered to occur within one day follow an accident, give rise to somewhat different health consequences than chronic exposures, those which are received af ter one day. Early health effects generally develop from acute exposures. Late effects, principally cancer, may be generated by both acute exposures and chronic exposures. Chronic exposures can result from low levels of radiation over a considerable period of time. For example, evacuation may be effective in removing population from the most highly irradiated sectors, yet a large unevacuated area will receive radiation of lower intensity. This is significant since the prepoederence of adverse health effegsg in terms of simple magnitude, as noted by the American Physical Society arc contributed by latent cancer from chronic low doses. The magnitude of long-term effects is due, partly, to the calculational model used. This study has used a linear dose-response health effects relationship, although other models were examined and results calculated. With the linear-dose response model, low doses to large numbers of people will generate the same total consequences as will high doses to small numbers of people. If the multiplicative product of dose and people, expressed in man-rems , is large - even though due primarily to low doses at large distances from a reactor site - health consequerces will be significant. 7-8
DRAFT Thus to mitigate effectively the total health consequences of a severe release, an accident mitigation technique must substantially reduce long-term health effects. As will be seen neither remote siting nor evacuation is very effective in reducing the long-term effects of chronic exposures from severe accidents in surface nuclear power plants. 7.3 Results of Consequence Calculations 7.3.1 Health Effects The results of the calculations for accident consequences are summarized in Table 7-3 below. The results are presented as short-term and long-term f atali-ties and illnesses.- These categories summarize a much broader range of health effects which were actually calculated. The range of values is due not to calculational uncertainty, but to variations in surrounding population and general meteorological conditions at the four hypothetical sites studied. The principal entries in Table 7-3 are weighted consequences. They were obtained by averaging the health effects associated with the most probable seasonal wind directions together with " severe" consequence estimates. The " severe" values, shown in parentheses, represent the most adverse and unlikely set of conditions where the release was assumed to occur when the wind was blowing towards the most densely populated sectors of the several sites considered. Table 7-3 Accidest Health Consequences Weighted and (Severe)* Results Range of Effects Contributors Surface Sites Underground Sites ** Early Deaths (short-term) 17-450 0 (210-1900)* (0)* Latent Cancer Deaths (long-term) 3900-6300 0 (4300-7200)* (1)*
^
Early Illnesses (short-term) 160-7700 (2000-62,000)* Thyroid Cancers (long-term) 3300-17,000 <1-3 (6300-130,000)* (<l-27)*
- Exposure of the maximum population density sector within 50 miles of site.
** Berm-contained Level-2 Accident Mitigation System.
7-9
DRAFT The differences in accident consequences for surface and underground plants are quite startling. The effectiveness of sub-surface containment as a safety augmentation technique hinges upon the degree to which the underground environment is. utilized. This observation, as emphasized in Section 5, is important. Previous studies have maintained that merely placing a conventional surface facility underground (Level-0 and Level-1) withog system for pressure-temperature accident manage-ment, will be ineffective. Without a pressure relief system such as that of the Level-2 design for the present study, failure of the underground plant by over pressurization is likely. Venting of the accident-generated gases to the berm at the plant foundation level, or rock cavern in the case of the mined plant, greatly reduces the rate and quantity of fission products which finally reach the atmosphers. The actual quantities of radioactive gases to reach the atmosphere are given in Reference 7-2 and the underground plant health effects are based on those releases. A Level-3 design, by virture of its engineered failure pathway and incorporated iodine filter, would be expected to reduce adverse health effects even further. The enforced migration of fission products through the berm results in release of only noble gases and methyl iodide , the more significant of the two with respect to health ef fects. Rock is much less permeable than soil and the depth of overlying material much greater for the eined-cavern plants than the berm concepts. Consequently, diffusion to the atmosphere of radioisotopes other than noble gases was insignificant for the mined-cavern concepts. The health e f fects which do arise from the underground plant release are due principally to the long-term low-level releases of radioiodines. These would produce a small number of thyroid effects. For the surface plant, the main contributors to early fatalities (occurring within 60 days) are bone-marrow depletion and lung damage due to inhalation of radionuclides from the passing cloud. The late effects (occurring after latent periods of from zero to 15 years , and extending over the life expectancy of the irradiated individuals) arise from whole body doses due primarily to radio-activity deposited on the ground, and secondarily from the passing cloud. A more complete discussion of health impacts and their origin is given in Re ferences 7-1 and 7-2. There are several rignificant points to be noted about the surface plant health consequences. First, the adverse health ef fects from acute exposures in surf ace nuclear power plants were found to be highly site dependent. The range ,f values in Table 7-3, varying by a factor of about 25 for early deaths and about 50 for early illnesses, exhibits the extreme dependence of early consequences on site characteristics. Given the general unlikelines s of the accidents considered in this study, a site with probable consequences of 450 deaths might be considered unacceptable while one with the potential for 17 would be more in keeping with other hazards both man-made and natural. (It was noted in Section 4 that California's population is concentrated in the most seismically active areas of the state. Thus, there is an implied acceptability of the risk due ta major earthquakes. Certainly a repeat of the 1906 San Francisco Earthquake would result in significant numbers of fatalities.) 7-10
DRAFT Second, under unlikely circumstances - wind oriented toward high popriation sectors, ineffective evacuation, etc. - the number of early deaths increases. A possible means to reduce risk, in general, from surface facilities would be to site downwind, under prevailing conditions, from population. This would require, however, a broad meteorological analysis on a scale much larger than presently performed for surface facilities. Third, the long-:erm consequences are larger than the early consequences. Early consequences arise, for the most part, within about 25 miles of the reactor site. As will be discussed in the following section, the long-term effects arise from low-level exposures over long periods of time at distances of the order of 25 to 200 or more miles from the site. At this distance scale, population measured in most directions will tend to approach a gross average. Thus, long-term effects are generally independent of the site characteristics with the possible exception of coastal sites, where complex meteorology exists. Fourth, the largest category of consequence, thyroid cancers, is the most readily treatable and, probably, avoidable impact. Distribution of iodine tablets to the affected population shortly af ter an accident could block uptake of radioiodine and thereby reduce the numbers of thyroid effects. The validity of such an argument depends, however, on: (a) an adequate supply of iodine tablets being available; (b) timely distribution; and (c) ingestion. Neverthe-less, the calculated numbers are general upper limits on potential consequence:s. As a concluding observation, one must recognize that the long-term effects are presented as a total impact for the affected population. In actuality, the large numbers of latent effects occur at some rate for a long period of time, 30 to 50 years. Considering the range of latent cancers from Table 7-3, 4300 to 7200, an incidence period of 30 years would yield a rate of about 140 to 240. fatalities per year. This may be contrastedwiththeaverageagug incidence of cancer from other causes, 179 deaths / 100,000 population. 7.3.2 Economic Effects For the hypothetical accidents and sites of this study, the costs of evacuation and relocation, which includes improved property loss, farmland, and medical treatment are as summarized in Table 7-4, below. The surface plant consequences are expressed in billions of dollars while the underground plant consequences are measured in thousands of dollars. Values are presented as both weighted averages, determined analogously to health effects, and severe cases. As with tL health effects numbers from the preceding section, these values should be considered relative indicators and not as precise values. 7-11
DRAFT Table 7-4 Economic Consequences Weighted and (Severe)* Results Soure, Surface Plant Underground Plant (S Billions) ($ Thousands) Evacuation & Relocation 0.13 - 7.2 0 (0.23 - 34)* (0)* Farmland Interdiction 0.13 - 1.3 0 (0.31 - 1.6)* (0)* Medical Treatment 0.079 - 0.12 1-16 (0.1 - 0.6) (1-140)* TOTAL 0.34 - 8.6 1-16 (0.64 - 35) (1-140)*
- Exposure of the maximum population density sector within 50 miles of site.
The range of surface plant accident costs was from $0.34 billion to $8.6 billion under the most likely conditions, and could range to $35 billion under extreme circumstances. The comparable underground plant accident costs ranged from $1,000 to $16,000 with an extreme cf $140,000. These estimates are public impacts and do not include the value of the plant itself or the value of lost energy production. The major economic impact from surface facilities arose from the need to evacuate and relocate people for extended periods of time. Interdiction of farmland was the next most significant cost factor. Medical and health care costs were the smallest cost category. There is an implied relationship between interdiction criteria and evacuation and relocation costs. The reference calculations assumed that any area where a person might receive more than 25 rems in 30 years required interdiction. If this dose limit were raised, then evacuation and relocation costs would be less. Additional calculations were made at interdiction 1 _;s of 50 and 100 rems in 30 years. At the higher interdiction levels, average costs dropped 17%, however, predicted fatalities due to latent cancers increased about 70% over the reference (25 rems in 30 years) case. No evacuation, relocation, or interdiction of farmland was found necessary for the underground concepts. Consequently, no costs in those categories were incurred. The small number of health effects contributed the medical costs. 7.4 Groundwater Contamination Approximately 10 hours following onset of the accidents studied in this work, the molten radioactive core would melt through the plant's thick, concrete and steel, foundation mat. This would permit access of radioactive material to the groundwater. In an underground facility, as compared to a surface plant 7-12
DRAFT more of the core fission products could find their way into groundwater. A conventional surface plant is apt to experience containment failure prior to melt-through of the foundation and possibly release significant quantities of radionuclides into the atmosphere. Those radionuclides released would not be available for groundwater contamination. The time it takes radioactive material to travel from the release point to a withdrawal point depends on factors which can change markedly with site condi-tions. The question of groundwater transport was analyzed parametrically and, for a range of realistic conditions, radionuclide travel times were found to range from 7 years to 4 million years per mile. Since these times are long, compared to most radionuclide half-lives, considerable decay takes place before extraction occurs at an assumed distance of one mile. However, for some conditions, the longer lived nuclides such as Sr-90 and Cs-137 could exceed maximum permissible concentrations at a one-mile distance. In those cases, some control measures would be necessary. Possibilities for control of groundwater contamination include pumping ground-water from intercept wells for decontamination; altering physical properties of the soil to retard migration by injection of material into the aquifer; and others. It is concluded that control of contaminated groundwater is feasible and does not pose as significant a public health threat as release of the same quantity of radioactive material to the atmosphere. 7.5 Summary The consequence calculations performed for this study represent state-of-the-art analyses. Nevertheless, in many areas of the calculational process, a lack of fundamental knowledge introduces uncertainty. The applicability of present meteorological models to distances cf hundreds of miles; uncertainty in the health effects of low-level radiation, dose rates, and thresholds; variations in deposition rat 2s of radioactive material; the accuracy of evacuation and emergency response models; and other factors combine to make accident conse-quence predictions only estimates. No other connotation should be construed. Given the qualifications above, the analysis performed in connection with this study shows that an underground nuclear power plant, augmented with a passive containment pressure relief system, can greatly reduce the magnitude and severity of hypothetical, so-called core-melt accidetis. The reduction in health effects, referenced to a surface plant, and hypothetical site conditions, is nearly total for early fatalities and measured in thousands for latent fatalities. The reduction in economic impact is measured in billions of dollars. Under the assumption that a core-melt class of accident occurs, the underground facility does not require evacuation of a large populace. Of fsite health effects are negligible. Underground siting of nuclear power plants has been found to be an effective measure against extreme reactor accidents. Whether or not it can be considered cost-effective, or optimal, is the subject of the following section. 7-13
DRA?T REFERENCES SECTION 7 7-1 Evaluation of the Feasibility, Economic Impact, and Effectiveness of Underground Nuclear Power, Plants, The Aerospace Corporation, Final Technical Report, May 1978. 7-2 Analysis of Public Conseque .ces from Postulated Severe Accident Sequences in Underground Nuclear Power Plants, Advanced Research and Applications Corporation, December, 1977. 7-3 Reactor Safety Study - An Assessment of Accident Risks in U.S. Commercial Nuclear Power Plants , WASH-1400, NUREG-75/014, October 1975. 7-4 Reviews of Modern Physics, Report to the APS by the Study Group on Light-Water Reactor Safety, American Physical Society, Vol. 47, No. 1, Summer 1975. 7-5 Allensworth, J.A. et al., Underground Siting of Nuclear Power Piants: Potential Benefits and Penalties, Sandia Laboratories, SAND 76-0412, NUREG-0255, 1977. 7-6 American Cancer Society, "78 Cancer Facts and Figures" (1977). 7-14
DRAFT 8.0 The Ef fectiveness of Underground Siting The design concepts which have been developed in the cou rse of this project make substantial use of the underground environment. The underground concepts are able to vent excessive pressures and substantial amounts of radioactivity either to the berm or to rock caverns. Such action avoids uncontrolled failure of the containment building and prevents violent release of fission products to the atmosphere. Analysis and discussion to this point have considered those circum-stances where underground siting operates successfully; however, it must be recognized that underground siting can fail. This section will consider the overall effectiveness of underground siting for mitigation of extreme accidents including the manner and consequences of failure of the underground designs of this study. In addition, certain alternatives to underground siting will be considered. 8.1 Containment Failure Modes Section 3 reviewed the principal mechanisms for containment failure which included breach by accident-generated missile, overpressure, melt-through of the foundation mat, failure of the containment isolation system; and, as identified in this study, excessive temperatures. The manner of failure, to a great extent, determines the magnitude of accident consequences. The release categori s definto in the RSS combise accidents of a similar nature, includig de inventory of radioactive ma',erial released and the containment G11ure mode; it is convenient to discuss release categories as surrogate- .or specific accidents. The release categories are described in Table B ior PWRs and Table 8-2 for BWRs. Several studies have evaluated the contri- :on to total consequences of each major release category for both PWRs and BWas. 8.2 Division of Risk by Release Category Some reactor accident sequences are more likely to occur than others. A few accident sequences, although perhaps less likely, are far more serious in their potential for heal.h effects to the public. Since risk reflects both the probability that an event will occur and the consequences if it should occur, a low probability may be imbalanced by a high concequence. Such is the case with extreme reactor accidents. There is accumulating evidence which suggests that risk to the public from extreme accidents is concentrated in only a few of the many possible release categories. Table 8-3, af ter Reference 8-1, summarizes the distribution of risk according to three criteria, earl: tndlatefatalg_egandpropertydamage. Based on the release categories studied in the RSS, and as strikingly illustrated in Table 8-3, risk to the public is dominated by release categories PWR 1, 2, and 3, and BWR 1, 2, and 3. Other investigators have considered some improvements in reactor design since the RSS, and although the role played by PWR 1, 2, g in overall risk of early fatalities diminishe s , they remain dominant. The data of Table 8-3 suggest clearly those areas where the incremental safety dollar should be invested. In fact, one of the major benefits of risk assess-ments such as the RSS is to highlight areas where the greatest payoff in risk reduction is possible. 8-1
DRAFT Table 8-1 PWR RELEASE CATECORIES DOMINANT SEQUENCE DESCRIPTION
- 1. Loss of power with failure of heat removal systems; core melt followed by steam explosion which ruptures reeetor vessel; explosion generated missile ruptures containment.
- 2. Failure of core cooling systems, followed by failure of containment spray and heat removal systems and failure of the containment through over-pressure.
- 3. Overpressure failure of containment due to failure of containment heat removal systems; containment failure followed by core melting.
- 4. Failure of core-cooling and containment spray or heat removal system after loss-of-coolant accident (LOCA), and failure of containment isolation.
- 5. Failure of core-cooling after T.0CA (in either injection or recirculs. ein mode) and failure of containment isolation.
- 6. Failure of core-cooling and containment spray, but integral containment until melt-through.
- 7. Failure of core-cooling, but containment spray works and containment is integral until melt-through.
- 8. Large pipe break LOCA with failure of containment isolation, no core melt.
- 9. Large pipe break LOCA, no core melt.
1 Ref. 8-10 Table 8-2 BWR RELEASE CATEGORIES DOMINANT SEQUENCE DESCRIPTION
- 1. Steam explosion in reactor vessel with explorion induced failure of containment.
- 2. Transient event with subsequent failure of decay heat removal; failure of containment due to overpressure followed by core melt.
- 3. Failure of containment through overpressure with release of radioactivity through reactor building followed by core melt.
4 Failure of containment isolation--leakage of radioactivity.
- 5. Large pipe break loss-of-coolant accidents, no melt.
1 Ref. 8-10 n .,
DRAFT Table 8-3 1 RISK BY INDIVIDUAL RELEASE CATECORIES Percentage of Total Risk Release Type Early Fatalities Late Fatalities Damage PWR-1 35 9 5 2 41 46 53 3 24 39 21 4 0 2 1 5 0 1 0 6 0 1 3 7 0 0 17 8 0 2 0 9 0 0 0 BWR-1 93 9 7 2 7 32 46 3 0 58 46 4 0 1 0 5 0 0 0 1 After Reference 8-1. Those release categories which dominate risk have a principal feature in common: the containment building, the last barrier to atmospheric release of radio-activity in a surface nuclear powerplant, fails violently and suddenly. The containment fails either by an accident-generated missile penetrating the massive Containment structure or by overpressurization. Both result in atmos-pheric release of significant quantities of radioactivity. If, somehow, failure of the containment can be prevented, or the radioactivity directed to some less consequential path, the reduction in consequences can be dramatic. A commen-surate reduction in risk follows. The mechanism by which underground siting accomplishes this goal, namely, venting of fission products to either soil in the berm plant case, or to a rock chamber in the mined-cavern plant, has been described in Section 5. The reduc-tion in accident consequences was described in Section 7. The following sections consider the implications of mitigating certain release categories and the possibility that underground siting, as a safety concept, can fail. 8.3 Mitigation of Specific Failure Modes Under the assumption that the pressure and/or temperature relief devices permit venting to either the berm, or the expansion volume in the mined-cavern case , the underground plants will suppress the public health consequences from surface plant PWR release categories 2 and 3 and thereby reduce risk substantially. The ability of an underground plant to prevent breach of containment from an 8-3
DRAFT accident-induced missile was' not calculated. The commonly cited mechanism for generation of such missiles, steam (vapor) explosions is at present a testly understood phenomenon. Reference 8-3, a comprehensive review of core-meltdown considerations, suggests that it is possible to consider steam explosions only .in a qualitative fashion. Consequently, with no clear measure of the size, velocity or energy distribution of potential missiles, one cannot clair contain-ment. It is clear, however, that an underground plant is considerably less susceptible to breach of containment by accident-induced missiles than compar-
.ble surface facilities. In the berm plant, the massive outer dome and overly-ing soil, and in the mined-cavern plant, 200 feet of rock, provide additional missile restraint not present in surface facilities.
8.3.1 Steam Explosions Release Categories PWR 1 and BWR 1 include containment failure by an accident-generated missile. The energy source is considered to be a " steam explosion," rapid conversion of water to steam, caused by the molten core dropping into a pool of water. If this occurs within the reactor vessel, the high pressures generated could blow the vessel head off, and carry some additional components along. 325,000 lbs. for a The postulated PWR and about 650,000 missiles arelbs. for a BWR.ggmely heavy, about It must be recognized t ia' steam explosions are a complex, 19orly understood phenomena. The circumrv. aces under which such explosions can be generated are not fully known; the p .;ential energy release can only be approximarad; and the conditions necessa y to produce the largest events are considered highly unlikely. These ana other considerations led the authors of Reference 8-3 to conclude that available models could lead to only qualitative evaluation of containment failure. However, the RSS attempted to assess the impact of steam explosions since they are an avenue to high consequence releases of radioactivity. Under the most extreme and unlikely conditions the RSS estimates that the 325,0gg--PWR missile would reach a height, if unimpeded, or aearly 2,000 feet Whether or not a missile with this energy could be restrained by either a berm-contained or rock-sited underground plant is not known. With respect to the berm-contained plast. very preliminary calculations suggest that under these conditions the hypotheer. zed PWR missile has sufficient energy to lift a soil mound of diameter 150 feet and depth 50 feet nearly 60 feet. The calculation ignores the substantis.1 energy lost in penetrating the containment building and the auxiliary building dome; nevertheless, because of the high energy that could be released it carnot be unequivocally stated that underground siting can effectively mitigate all steam-explosion-induced releases of radioactivity. The degree to which underground siting would be more effective than surface facili-ties 'in this regard is not presently known, and cannot be determined without further basic research into the phenomena of steam explosions. The underground designs would, however, be more ef fective than conventional surface facilities due at least to the added restraint of the overlying soil or rock. 8.3.2 overpressure Release categories PWR 2, 3 and BWR 2, 3 are dominated by containment failure through overpressure. The underground cesigns of this study are e f fe c tive 8-4
DRAFT against this form of containment failure. The underground plants preclude containment failure through overpressurization by venting of accident-generated gases to a subterranean location, rather than permitting direct release to the atmosphere. The public health consequences which might arise from this fom of venting in an underground plant have been shown to be reduced to insignificant levels. Results were summarized in Section 7. 8.3.3 Foundation Melt-Through A molten radioactive core will eventually melt through the concreta and steel foundation mat whether the plant is surface-sited or underground. Of all containment failure modes, however, foundation me':-through results in the lowest public health impacts simply because the radioactive inventory is not dispersed in the atmosphere. It is confined to the plant environs below ground. Natural soil is an effir lent particulate filter and greatly retards the movement of fission products eimpared to transport in air. Eventually, however, a limited amount of radiative material of relatively low potential hazard to public health is released to the atmosphere although at extremely low rates. The degree of filtering effectiveness possible and public health consequences of material ultimately released to the atmosphere are discussed in Re ference 8-4. Whether containment failure by melt-through occurs in either a surface or underground plant, some form of post-accident groundwater interdiction and processing to remove still-present fission products could be necessary. It is difiicult to predict which siting concept, surface or underground, would be most problematic. It is likely that an underground plant would be sited in a drier location than a surface facility to ease construction, thereby reducing quantities of ground-water requiring processing. However, since a surface plant is more likely to suffer containment failure prior to foundation melt-through than is an under-ground plant, thereby releasing major amounts of radioactivity to the atmos-phere, less radioactivity is available for release to groundwater. Conse-quently, an underground plant is more likely to introduce greater quantities of fission products into the groundwater. The trade-off between quantities of radioactive materials released and volumes of groundwater to be processed cannot be determined without deta!1(d site specific information. However, radionuclide migration in groundwater is a very slow process. Monitoring capabilities could be provided af ter the fact to measure migration rates and determine what control measures, if any, were required to minimize the extent of land contamination, and public risk. 8.3.4 Los s-o f-Isolation If the containment isolativ: system were to fail, under accident conditions , the containment building would not be sealed, and might pe rmit escape of radioactivity. In a surface plant, the release would be either directly to the atmosphere or into another plant area and then to the atmosphere. In an under-ground plant, with the secondary contaimnent features of the designs of this study, a containment los s-o f-isolation release would result in venting into another controlled volume. Release categories FWR 4, 5 and BWR 4 can thus be mitigated by the underground plants. 8-5
DRAFT 8.3.5 Excessive Temperature The analyses of this study suggest a containment failure mode not discussed in the general literature. It is possible under certain extreme accident c ond i-tions, to generate high temperatures (> 400*F) which might occur in conjunction with moderate pressures of approximately 40 psig. The long time periods con-sidered in the accident analyses of this work, up to 600 hours following onset of the accident, led to conditions, after conversion of all available watg steam, where continued decay heat input would lead to high temperatures. Temperatures above 400*F may be assumed to fail containment building seals. However, pressure reductione achieved through the Level-2 or Level-3 Accident Mitigation systems assi i Pressure from containment to the other portions of the plant. glow d7 ving d Since the secondary containment structures of the underground facilities are sealed, releases to the outside environment would, if they occurred at all, be nominal and diffuse. A PWR-4, leakage-type release is shown in Reference 8-1 for a surface plant to result primarily in latent fatalities considerably less severe than PWR 1, 2, and 3 releases (c.f., Table 8-3). The consequences from an underground accident would be reduced further due to additional opportunity for delay of releases and removal of radioactivity associated with the secondary isolation features. 8.4 Failure of Pressure Relief Systems An element of the effectiveness of underground siting is proper performance of the pressure relief system. The rupture disc designs of this study are totally passive. They require no external power or activation signal and rely on t.ne physical environment during a potential extreme accident for mobilization. They are highly redundant and can be engineered to any design reliability through series installation, multiple paths, etc. Nevertheless, if accident sequences of extremely low probability are evaluated, possible pressure relief system failure must be considered. Complete failure of all redundant pressure relief devices (failure constitutes not opening) would result in generation of high pressures within containment with ultimate venting from the contain"nt building or reactor cavern into the other plant environs. Both the berm-ct. .ined and mined-cavern plants contain secondary isolation systems which effectively seal great volumes of the underground plant. The volume available for expansion, and surfaces for condensation, of high pressure gases would result in thermodynamic behavior comparable in many ways to the engineered failure paths of the accident mitigation sys tens . As shown in Reference 8-4, ultimately, high temperatures with low pressures result. In this event, the underground facility, with its limited number of penetrations and pathways to the surface, would f acilitate control of leakage. Thus, failure of the containment rupture discs, while unlikely and undesirable, results in circumstances with less possibility for consequences to the public than a surface plant failure. 8.5 Summary of Accident Mitigation Capability The preceding discussion of containment failure modes and release categories has noted the general effectiveness of underground siting in all cases except those 8-6
DRAFT which include steam explosions. The great uncertainty surrounding the latter phenomena precludes any de finitive statements. Conservatively estimate 6, a thoughtfully implemented underground plant could reduce risk to the public substantially by eliminating the possibility of containment failure through over-pressurization. An underground plant is not, however, absolutely f ail-safe. Because underground siting cannot make operation of a nuclear plant risk-free, undergrounding must be balanced against other alternatives for improving reactor safety. The following section will consider other options for risk reduction. 8.6 Alternatives to Underground Siting In the realm of extreme, or Class 9, accidents, most suggestions for risk reduction take the form of passive attributes which affect a physical character-istic of the accident. For example, controlled venting of the containment building attempts to preve .st overpressurization as does increasing containment design pressures or volumes. The key point is reliance on passive physical features rather than active engineered systems. An extreme accident cannot occur unless there is a failure of several existing, active, engineered safety systems. Consequently, and almost by definition of a Class 9 accident, more reliable, redundant, or conservatively designed pumps, motors, and valves, injection and spray systems, etc. , have not been considered as options. In a complete risk analysis, however, the effectiveness of such active means of risk reductim through accident prevention would be considered. They are philosophically in contrast to concepts such as underground siting which serves to control consequences. The alternatives to be discussed are all passive in nature and are related to consequence mitigation rather than accident prevention. They include means to prevent sudden, catastrophic failure of the containment building or methods to reduce the impac t should failure occur. In general, the alternatives may be loosely categorized as . systems modifications or siting options. Each category will be discussed separately. 8.6.1 Systems Modifications A recen examined several options intended to inidbit c onta inmen t failure.[8-grk The effort concentrated on means to reduce or eliminate the likelihood of containment failure through overpressurization and did not include failure due to steam explosions, los s-o f-isola tion, or excessive temperatures. Consequently, the rankings obtained reflect only one of several containment failure modes ag ainst which undergroand siting is e f fective. The underground plant designs of the Commission study combine underground siting with controlled venting, and would be the most effective of all concepts considered. For these reasons, the relative risk obtained by the investigators of Reference B-6 for both shallow and deep underground siting has been deleted from Table 8-4. Nevertheless , the results of Reference 8-6, summarized in Table 8-4, are highly instruc tive in illustrating those system modifications which show promise for mitigation of extreme accidents. It is noteworthy that no alternative reduces risk to zero. To rn.aferce the po int made earlier, risk reduction must be evaluated on an incremental basis since no present alternatives are absolutely risk-free. (Even shifting to another electrical generation technology is not risk-free. All bear some potential fe- harm.) 8-7
m m Table 8-4 RELATIVE RISK FOR ALTERNATE CONTAINMENT DESIGNS , NORMALIZED TO CURRENT SURFACE PLANTS Early Fatalities Latent Effects Current Surface Plants 1.00 1.00 Thinned Base Mat 1.00 1.00 Double Containment 0.91 0.91 Evacuated Containment 0.71 0.83 Stronger Containment 0.18 0.45 Increased Containment Volume 0.18 0.45 Filtered Atmospheric Venting 0.08 0.11 Compartment Venting 0.08 0.11 1 Summarized from Reference 8-6. The relative risk factor obtained for both shallow and deep underground siting has been omitted due to the dif ferences in design between the concepts of Reference 8-6 and this study. This study has estimated the incremental cost g_gitered atmospheric venting to be approximately 2% of direct plant costs. Such a system would vent containment to an engineered, passive, filtration system. The very inele-gant, yet efficient, soil / sand / charcoal system envisioned would capture most of the more hazardous releases of volatile fission products. The venting system would preclude containment overpressurization assuming reliable per-formance and avoid direct atmospheric release of fission products. A further caution is given in that the underground designs of this study, with inherent secondary containment and enhanced efrectiveness against steam-explosion-generated missiles , are more broadly effective than a filtered venting system alone. They should not be equated. Nevertheless, the low percentage cost associated with the modified surface facility is clearly an incentive for further study. In suumary, augmented underground siting, among current alternatives, appears to J i ve the greatest potential for risk reduction at the greatest cost. Other system modifications, however, suggest greater risk reduction per invested dollar. This conclusion is drawn from the risk distribution presented earlier in Table 8-3, the estimated costs of underground siting and controlled, filtered venting, and the general ranking of effectiveness as shown in Table 8-4 8.6.2 Sitine Options Two general alternative siting modes have been considered, offshore and remote land-based. Offshore siting of a nuclear plant on a large barge-like structure, termed a floating nuclear plant (FNP), is being considered for initial applica-tion along the Eastern U.S. Seaboard. Prior to this work, investigation of the sug_aglity of FNP for application to the California coastline was performed The study noted that the typical offshore conditions along 8-8
UKAF1 California's coast, deep undersea canyons , high ridges, with water depths much greater than the Eastern U. S. , generally were not suited for FNP. The work concluded that, at present, ocean-sited FNPs were generally not an option for California. The second siting option, remote land-based, holds more promise. Califor2ia's unusual population distribution, with about 90% of the people in the major metropolitan areas of San Francisco, Los Angeles, and San Diego, results in relatively low population densities in many areas of the State. It was noted in Section 7 that population totals around most existing and proposed California reactor sites are low in comparison to other U.S. reactor sites. California has the opportunity to implement remote siting on a consistent basis if deemed necessary. Results of the consequence calculations presented in Section 7.3, indicate that while ef fective in reducing early adverse health effects, remote siting does not appreciably affect the long-term health consequences of an extreme accident. However, "One might. . .choosp3_a9 ) site remote enough so that the hazard to individuals becomes small." In other words, since long-term health ef fects are calculated on the basis of an average population dose, individuals who would ultimately experience these effects cannot be identified. Individuals expected to experience short-term health ef fects could be more readily iden-tified, due to the characteristics of acute exposures, short-term effects , and smaller exposed population. It is suggested here that reduction, or elimination, of early fatalities will change the character and nature of reactor risk to the point where it is similar to other forms of commonly accepted risks. The reasons for this suggestion will be the subject of the following section. 8.7 Site Characteristics and Reactor Risk Preceding sections have noted the general dependence of consequence levels on the characteristics of reactor sites. This section will discuss the relation-ship between site and accident characteristics, and show how they combine to produce great variations in consequence levels. 8.7.1 Latent Fatalities and Site Characteristics Section 7.4 discussed the dependence of early fatalities on site character-istics, principally population distributions and wir.d conditions and the rela-tive independence of latent fatalities from these par neters. These findings are not unusual or unique. For excmple, - similar obsescations are noted in Reference 8-1 which reported on the RSS consequence model:
"he results of the calculations with the consequence model demon-strated that the early f atalities were generally limited to within 10 to 15 miles from the reactor even for the largest release magnitude."
"The calculation of latent cancer fatalities is performed on the basis of population doses. Since no threshold is assumed for cancer induc-tion from radiation exposure, it is reasonable to expect that the nenber of latent cancer fatalities would not be very sensitive to the various assumptions made in the consequence model. The single para-meter of primary importance is the total quantity of radioac tive material released in the accident."
8-9
DRAFT The reference above notes subsequently, that, with the e::ception of the PWR-1 and BWR-1 release categories, the number of latent fatalities is roughly propor-tional to the population between 25 and 100 miles from the reactor. Release categories FWR-1 and BWR-1 display a high Ru-106 release fraction, which in-creases the likelihood of inhalation induced lung cancer. Since evacuation e f forts are assumed to extend to 25 miles from the release site, and because reactor sites are generally chosen with low existing, and projected, populations within 25 miles, the lopulation principally af fected by long-term health impacts begins at that distance. It is very informative to separate the risk from individual release categories and present them independently. Table 8-3, presented earlier, shows aggregate risk from the nine PWR release categories of the RSS. Figure . 8-1, which pre-sents the probability of equaling or exceeding a specified level of latent fatalities, was developed from data which formed the basis of the RSS, Latent Fatality, 100 reactors, risk curves of Figure 3-1. This section will discuss only PWRs; however, the general observations are equally applicable to BWRs and possibly other reactor tee.hnologies as well. Since the data for Figure 8-1 were developed in connection with the RSS, probabilities are as determined in that work. (Absolute probabilities may be subject to change due to the uncertainty in their estimation and calculation. The relative probabilities, however, may be less sensitive to the uncertainties of the analysis. Thus , if necessary, the probability scale may be adjusted by the reader to reflect his perception of individual accident probability.) The curves may be visually grouped into four categories: PWR 1A, IB, 2 and 3; 4 and 5; 6 and 8; and PWR 7 by itself. The contribution made by each release category to total risk is visually indicated by the area under each curve. Curves lA, 1B, 2 and 3 dominate total risk although the use of logarithmic scales reduces the visual impact. The curves of Figure 8-1 reinforce the numerical distribution of risk given in Table 8-3. In addition, the distribu-tion of risk generally parallels the magnitude of the core inventory released. An interesting characteristic of the curves is the nearly flat initial portions for PWR's lA, IB, 2 and 3. For these release categories, the possibility of an accident producing latent fatalities from 1 to approximately 500 is essencially constant. Thus, for the magnitude of the release in these accident categories, a threshold of 500 fatalities, distributed over a time period of 30 to 50 years, must be considered a mi..imum. Note that this value is relatively independent of specific site characteristics but is dependent on aspects of health models, e.g., whether or not dose e f fectiveness fac tors are used.3 For an accident to result in greater latent consequences, adverse circumstances must exist. For extreme consequence lev el s , there must exist high surrounding populations, a prevailing wind blowing, at the time of the accident, in the direction of the densest population, and ineffective evacuation. 1 For example, the RSS used a dose-effectiveness factor which varied from 0.2 to 1.0 depending on dose rate and total dose received. Under a linear dose-response model, with no dose e f fec tiv . .s s factors, the 500 fatality threshold would increase to approximately 2,500, a factor of five. This level is generally consistent with the range of latent values calculated in this study, 3,900 to 6,340 (8-4). These are, equivalently, 130 and 210 f atalities per year for 30 years. 8-10
u.A u a.A- .s. Figure 8-1 10-5 2 s\ , 3 10-
\ 1A x
N E r 10' ~ 5 t
. a:
5 c. f 10- - 2
.8 E
7 8 6 5 4 10-' - k 10-10 1 10 100 1000 10,000 Latent Fatalities (X) Probability of Equaling or Exceeding Latent Fatalities of Number X for PWR Release Categories 1-8*
- Based on data from Reference 8-19 8-11
DRAFT It is possible, to a small extent, to control the magnitude of the po tential consequences by selecting sites with low population densities. However, since the major affected population is 25 to perhaps 200 or more miles from the - reactor site, it is probagly not possible to identify sites where there would be no latent health effects. The possibility of the most serious accident consequences, however, can be reduced substantially by avoiding those sites where the prevailing wind is towards major population centers. The meteorology over a wide area should be examined carefully prior to any site certification. It is significant that the population of California is concentrated in the principal urbanized areas surrounding San Francisco, Los Angeles, and San Diego. As a consequence, population density surrounding both existing and proposed reactor sites in California is generally below that of the average site nationwide. This is shown graphically in Figure 8-2, where the curve labeled WASH-1400 was constructed from the population data of that report, and is the average of 68 reactor locations. The sole exception to the generalization above is the Rancho Seco location which, in the distance scale considered, encompasses the Sacramento urban area. Several inferences may be drawn from the shape of the curves of Figure 8-1. It has already been noted that PWR 1, 2, and 3, and to a lesser extent PWR 4 and 5, exhibit a " plateau" where the probability of exceeding a given consequence level is nearly constant. There is evidently some threshold value of consequence which could be expected for these accident sequences. The other accident sequences do not exhibit this threshold to such a marked degree. A conclusion is that, for those accident sequences which exhibit a threshold level of latent consequences, the characteristics of the accident, e.g. , the quantity o f radio-active material released, dominate the latent fatalities and that conservative siting cannot reduce consequences below the thresMid. 8.7.2 Early Fatalities and Site Characteristics Figure 8-3 presents curves which relate early fatalities to the probability of exceeding a specific level of consequence for individual release categories. The curves of Figure 8-3 were also constructed from data developed in the RSS. (As with the latent fatality curves presented earlier, the probability scale may be adjusted if the reader deems it necessary.) The early fatality curves do not exhibit the " plateau" of the PWR 1, 2, and 3 latent fatality curves to the same extent. Site characteristics such as surrounding population, whether or not evacuation is ef fective, and meteorological conditions combine to influence the magnitude of the consequences. The accident characteristics, alone, evidently do not force a dominant threshold of consequences. Early fatalities.may be controlled, to s great extent, through siting policies. The principal reason is that the doses required to produce early fatalities are quite high and that concentrations of radioac tivity fall off rapidly with distance from the reactor site following an accident. This latter e f fec t is 2 An approximate distance scale for an accident may be calculated from uh/Vd where u is wind velocity; h is inversion height; and V i particulace deposition velocity. For u = 3m/sec, h = 1000 m, and Yd =s0.01 m/sec, uh/Vd = 300 km = 186 miles. 8-12
i e 5 10 _
~ Newhold Island 4
10 -
- ndian point _
;l j'j:!lllll,'!*.!!:!::F.'%k San Onotre
' F- Rancho Seco
\,
' WASII-34gg
'g ..'
** i.:
ii 8nJoaqula T y ** - u ..
.5 :ll5.' ;
g \\\\\\\\\\\\\\\ Ol*hlo Canyon
\ @ l0 i f/
f Sunde,,,, lo -
\
, ,\ , , , , , , , , a 0 5 to 15 20 25 30 35 40 45 50 55 60 Distance in Miles Figure 8-2 Cumulative Population As A Function of Distance From Nuclear Plants
DRAFT Figure 8-3 10-6 PWR 1B 10-I - PWR 2 x Al k
- PWR 3
$ 10' - PWR 1 A M e b o. 4a E E .2 10-9 2 o. 10:10 , 10'I ' ' 1 10 100 1000 Acute Fatalities (X) Probability of Equaling or Exceeding "X" Acute Fatalities for PWR Release Categories 1-3 Source: Reference 8-19 8-14
DRAFT shown in Figures 8-4 and 8-5 for two wind conditions.(8-20 Of interest is the change in dose with distance. TheseJagrs combine to produce mortality curves of the shape shown in Figure 8-6. Although Figure 8-
- particular circumstance, "a large, cold ground level release..."pg,g0f#the shape is characteristic for the reasons cited. Returning to Figure 8-2, it is noted tbat for all proposed and existing reactor sites in California (excluding Humboldc Bay) population within five miles of the sites was less than 1,000 persons at the time (1970-75) the data were obtained. For any given population distribution, early fatalities may be determined approximately by superimposing the population distribution over curves of the form of Figure 8-6.
Clearly, a simple strategy for minimization of potential early fatalities would be siting in sparsely populated areas where prevailing wind patterns avoid more densely populated areas. Such a policy would also reduce potential accident costs, as well. It was noted earlier that, for the accidents considered in this study, costs for evacuation and relocation were dominant. Since these costs are directly proportional to population density, remote siting would reduce cost by limiting population exposed. The difficulty with remote siting is deciding "how remote is remote enough." It is probably not possible to find locations within California feasible for nuclear powerplant siting and yet on the order of 20 to 30 miles from the nearest habitation. The merits, therefore, of remote siting lie not in to-tal elimination of early consequences, but in reduction of the scale of the accident. 8.7.3 Snne ry of Site Characteristics and Reactor Risk Certain conclusions may be drawn from the assumed characteristics of hypothe-tical reactor accidents. First, given the present understanding of radiation-health effects, it is not possible to eliminate latent effects from an extreme reactor accident due to their evolution from low-level radiation distributed over a large area. Second, the magnitude of early fatalities and early adverse health effects is very sensitive to site characteristics. Third, through conservative siting policies, the early fatality portion of extreme, high consequence accidents identified in the RSS which require large populations, ineffective evacuation, and unfavorable meteorology where prevailing wind is from the reactor site towards the population, may be substantially reduced in significance. 8.8 Sabotage Underground siting has been argued as being both beneficial and detrimental in the event of sabotage. For example, it is argued that the reduced number of plant access points, as compared to a surface plant, would make entrance more difficult and protection from within easier. That suggestion has been rebutted by arguing that sabotage is likely to be done by plant employees from within, who could then easily deny access to authorities on the outside. Determining the susceptibility of a facility g sabotage is difficult. It is also a highly subjective undertaking. However, evaluation of the consequences from sabotage is more tractable. In the latter respect, the underground designs of this study have a unique advantage. If saboteurs were to gain control of an 8-15
DRAFT Figure 8-4 - 6 10 , , ,, 5 ' 10
- Thyroid g
4 10 _ e Lung 8 s 3 - Bone . marrow
] 10 3 2
10 ,
\
10 I ' ' ' ' O.1 1 10 100 Miles from Reactor Total organ doses versus distance from reactor for hypothetical weather; stability A, wind speed of 0.5 m/sec. Thyroid dose = l-day ground + external cloud dose + 30-day inhalation dose; Lung dose = l-day ground + external cloud dose + l-year inhalation dose; GI tract dose = l-day ground + external cloud dose + 7-day inhalation dose (the GI tract dose is the dose to the regenerative cells of the lower large intestine); bone marrow dose = l-day ground + external cloud dose + 4 (7-day inhalation + 30-day inhalation dose). Source: Reference 8-20 8-16
DRAFT Figure 8-5 Thyroid . 5 \ 10 _
- GI :
trac - n 4 E 10 g : 3 8 m 3 Bone marrow 3 10 3
\ ( _
e : - 2 l 10 I -
~
i I ' ' ' ' ' 10 O.1 1 10 100 Miles from Reactor Total organ doses versus distance from reactor for hypothetical weather; stability F, wind speed = 2.0 m/sec. Thyroid dose = l-day ground + external cloud dose + 30-day inhalation dose; Lung dose = l-day ground + external cloud dose + l-year inhala-tion dose; GI tract dose = l-day ground + external cloud dose + 7-day inhalation dose (the GI tract dose is the dose to the regenerative cells of the lower large intestine); bone marrow dose = l-day ground + external cloud dose + 1s (7-day inhalation
+ 30-day inhalation dose).
Source: Reference 8-20 R-17
DRAFT Figure 8-6 1.0 i s i 3 Bon \ l ma"erow g b 0.8 .1 (A) I g i Bone marrow -
- - (F) 5 \
f0.6 I b I L g 0.4 . { h Lung (F) . E it\ \
\
0.2 . I p lung (A) { _ k\ \ GIJ\ (A)'s ,k
\
( GI (F) O , , , , , 0 1 2 3 4 5 6 7 8 9 10 Miles from Reactor Mortality probability for an affected population versus distance from reactor for two hypothetical weathers: stability category A, wind speed = 0.5 m/sec; stability category F, wind speed = 2.0 m/sec. . Source: Reference 8-20 8-18
DRAFT underground plant, the facility could be isolated from the external environment by actions from the outside. Attempts to produce a deliberate release of radioactivity could be thwarted. Thus, an underground planc designed to accommodate an ex treme accident through passive means, could not be used to jeopardize population at large by threats to destroy the plant. It is suggested that an underground nuclear power plant would be a less attractive target for saboteurs than a surface facility. 8.9 Decommissioning Reactors sited underground, just as their above-ground counterparts, must someday be retired and face decommissioning. Recent estimates of the costs of decommissioning today's large LWRs reveal that the expense involved in decon-taminating and dismantling a shutdovn reactor can be subgta,ngia1 g 8 k"8 Ef3S8#* to the original costs of plant construction and operation. If decommissioning costs for an underground facility were significantly different than for a surface plant, the overall economics of underground versus surface siting could be af fected. This study has not attempted to include a detailed analysis of the implications of undergrounding on the costs and requirements of decommissioning. Since the exact procedures that will comprise decommissioning are only presently evolving, no firm basis for evaluation of an underground facility exists. Until the more basic questions of the benefits and advantages of underground siting from an accident consequence mitigation standpoint are resolved, extensive consideration and speculation on the impact of sub-surface siting on the safe termination and disposal of a reactor would be somewhat premature. Nevertheless, there are some possible interactions of siting and decommissioning which may be qualitatively discussed at the present and whose implications warrant consideration in the overall assessment of the merits and penalties of underground siting. Essential to a discussion of the impact of underground siting on decommissioning is a definition of the various methods by which it may be accomplished. Decom-missioning as currently defined by NRC Reg. Guide 1.86 ge,gts several alterna-tive courses of action to qualify as decommissioning. Under the option of mothballing, an operator is required only to remove the spent fuel from the site, perform a certain amount of system decontamina tion, and provide active security to prevent entry of unauthorized persons and their subsequent irradia-tion by any remaining contamination. The entombment option is basically an extension of mothballing that includes the addition of physical barriers to restrict access to the containment for at least the immediate future. The third - option of complete dismantlement begins with decontamination but then continues with the complete removal and sa fe disposal of all equipment, components, facilities or structures still producing significant levels of radiation. If done properly, the site could be released to unrestricted use. As a fourth option, the NRC has defined conversion. Conversion involves decontamination and/or removal of the reactor itself with the conversion of remaining facili-ties, such as the turbines, for use by a new nuclear or non-nuclear steam source at the same site. . If methhballing is ra'.ected for decommissioning, underground siting would have relatively little impact upon labor and costs. Removal of spent fuel and facility decontamination should be insensitive to plant location. Active R 1Q
DRAFT facility security might be reduced because underground siting would provide fewer access points to high radiation areas. If entombment were desired, siting underground, especially in a mined cavern could be beneficial. Entombment of a conventional plant is generally envisioned as sealing containment by welding doors and adding concrete and steel barriers. For an underground plant, passive physical security could be provided by block-ing access tunnels. Sealing access tunnels in a mined site could be accom-plished inexpensively, perhaps through demolition. In addition, certain physical barriers may already have been placed in these tunnels to provide control for possible accident-related radiation releases. Since it may be an irreversible action, destruction of access tunnels to accomplish isolation deserves detailed study. If conversion were the selected decommissioning option, the ef fec t of undergrounding would depend upon the portions of the plant placed above and below ground. If the turbine building were placed on the surface and the containment and radwaste facility placed below ground, any advantage which undergrounding might provide for entombment would be transferred similarly to conversion. Isolation of the most radioac tive structures might be facilitated and linking of the existing turbines to a new above-ground, non-nuclear steam source might be more easily accomplished. There would be little immediate risk from the proximate, but physically isolated, underground portions of the former nuclear plant. The remaining option, dismantlemen t , is perhaps the most interesting. If dismantlement is defined as the process for safely, permanently, and completely dealing with the radiation hazard of a shutdown nuclear plant, underground siting may offer some substantial advantages over surface siting. If dismantle-ment is interpreted to mean total removal of all radiation to make the previous reactor site suitable for unrestricted use for a new facility, then undergrounding could present some substantial disadvantages. Considering the latter definition, total removal of all radiation, conventional concepts can entail destruction and transport of the containment structure to a disposal site. Removal of a containment structure, even those portions that have low or inconsequential levels of radiation, would be very costly and, as presently performed, could involve mechanical demolition or the use of explo-sives. If applied to berm-contained concepts, these procedures might first require removal of the earth overburden to permit easy physical access of demolition equipment to the outer dome structure and subsequently to the inner containment building. For mined-cavern concepts, the size of the cavern could make conventional demolition difficult and the use of explosives risky. The removal of the most highly contaminated portions of the plant, including the reactor pressure vessel, should be substantially una f fec ted by underground siting, since removal of highly contaminated hardware would represent a rela-tively small portion of the overall decommissioning effort. In addition, present dismantlement planning envisions cutting up large pieces of equipment, such as r pressure vessel, using remotely-controlled plasma torches (gg' g_gq);or Such activity could be readily performed below ground. 8-20
DRAFT If, on the other hand, the objective were to achieve maximum safe disposal of all decommissioning wastes, the costs might be substantially reduced by under-ground siting. Recent studies indicate that a large portion of the overall dismantlementcostswouldberelatedtotheremovalanddisposalofg4rf'iogsg the plant with relatively low radioactive contamination levels. Under present procedures and standards for weste disposal, nearly all radio-active decommissioning waste would be classified as low level and would, therefore, be disposed of by shallow burial. Several of the low level waste burial facilities, which have received similar mata. rial, illustrate a historic inability in preventing the migration of radionuclidesy_g the burial site or in providing physical sacurity for the disposed wastes Underground designs and siting considerations are concerned specifically with containment and/or prevention of release and migration of radionuclides. There-fore, detailed analysis could prove these sites to be superior to present low-level waste burial facilities, considering the level of public health protection they could provide. The tens or hundreds of feet of rock or earth overburden provided by the concepts considered in the present study could provide superior environmental isolation compared to the 4 to 8 feet of soil that present vaste facilities use to cover low-level wastes. It could be argued, therefore, that leaving the low-level radioactive portions of a sub-surface nuclear plant in place and simply sealing all access points with rock or earth, might provide better environmental isolation and protection of public health than would the demolition, removal, transport, and burial in presently licensed low-level waste burial facilities. In summary, this report has not attempted to undertake a thorough analysis of the interactions between siting concept and decommissioning. Preliminary assessment indicates, however, that the interaction could prove to be substan-tial and that the costs of decommissioning, viewed as a factor in the total costs of a nuclear power plant, could be favorably affected by underground siting. 8.10 Summary Section 8 has dealt with the overall effectiveness of sub-surface containment as a means to reduce consequences to the public from extreme reactor accidents. A convenient categorization of reactor cccidents is by release category. Pres-surized water reactor (PWR) and boiling water reactor (BWR) release categories 1, 2, and 3 appear to dominate risk. These release categories are characterized by sudden failure of the containment building. The underground nuclear power plant designs of this study, by changing the manner in which containment fails (ss contrasted to attempting to prevent containment failure) can eliminate the . most serious failure modes. Five containment failure modes were considered in the evaluation: breach by accident-generated missile, overpressure, melt-through of the foundation mat, failure of the containment isolation system, and, as uniquely identified in this study, excessive temperatures. Underground siting was found to be effective against all failure modes but the one due to accident generated (sceam explosion) missiles. However, should a steam explosion occur, an underground facility would provide greater protection than a surface plant. 8-21
DRAFT The effectiveness of the unde ground plant designs of this study were evaluated by comparisen against risk estimates developed by other investigators. The designs of this study were found to provide the greatest reduction in risk at the greatest cost. Alternatives provided less reduction at reduced cost. No alternative was found to be risk-free. The relationship between the characteristics of reactor accidents and sites was examined. It was concluded that, by remote siting, early fatalities could be greatly reduced but that latent fatalities were likely to be little changed. Thus, emote siting could be effective in changin2 the form of public health impact following an extreme reactor accident, but could not eliminate all consequences. No advantages were found which suggested that an underground facility would be less susceptible to sabotage. However, because an underground plant can be effectively sealed from the surface by civil authorities, if necessary, the consequences to the public from sabotage could be markedly reduced from those arising from a surface facility. The ability of underground siting to ease or aggravate decommissioning was found to vcry substantially depending on which decommissioning option was considered. Underground plants would have an advantage if entombment were the selected option, and a serious disadvantage if dismantlement were chosen. 8-22
DRAFT REFERENCES SECTION 8 8-1 Wall, I.B., et al., " Overview of the Reactor Safety Study Consequence Model," International Conference of Nuclear Systems Reliability and Risk Assessment, Gatlinburg, Tennessee, June 1977. 8-2 " Containment Venting Considerations for Light-Water Reactor Accidents ," Denning, R. S., et al., Technical Summary, American Nuclear Society Transactions, 1977 Winter Meeting, pp. 644-645. 8-3 Core Melt-Down Experimental Review, Sandia Laboratories, SAND 74-0382, Albuquerque, New Mexico, August 1975, page 6-46. 8-4 Armistead, R.A., et al, Analysis of Public Consequences from Postulated Severe Accident Sequences in Underground iuclear Power Plants, Advanced Research and Applications Corporation, December 1977. 8-5 Evaluation of the Feasibility, Economic Impact, and Effective.ess of Underground Power Plants, The Aerospace Corporation, Final TeIhnical Report, May 197 8. 8-6 "Value/ Impact Comparison of Alternate Containment Designs," Calson, D.D. , et al., Technical Summary, American Nuclear Society Transactions, 1977 Winter Meeting, p. 643. 8-7 Conceptual Design and Estima?.ed Cost of Buried " Berm-Contained" Nu: lear Power Plants , S&L Engineers, January 1978. 8-8 " Floating Nuclear Power Plants - An Overview," California Energy Resources Conservation and Development Commission, June 17, 1976. 8-9 Reviews of Modern Physics, Report to the APS by the Study Group on Light-Water Reactor Safety, American Physical Society, Vol. 47, No. 1, Summer 1975. 8-10 Reactor Safety Study - An Assessment of Accident Risks in U . S. Com-mercial Nuclear Power Plants, WASH-1400, Appendix V, NUREG-75/014, October 1975. 8-11 Testimony of Peter N. Skinner, New York State Law Department, to the New York State Public Service Commission, Case No. 26974, December 2, 1977, p. 21. 8-12 Testimony of G.R. Faust, Gilbert Associates, Inc., before the Connecticut Public Utilities Control Authority on the matter of providing for the costs of decocimissioning Millstone I and II. 8-13 R. Jon Stouky and E.J. Ricer, San Onofre Nuclear Generating Stations Decommissioning Alternatives, Report 1851, for Southern California Edison. (NUS Corp., February 1977). 8-23
DRAFT 8-14 " Testimony of Operating Licenses for Nuclear Reactors," Regulatory Guide 1.86, U.S. Nuclear Regulatory Commission. 8-15 R.M. Harmon, et al. , " Decommissioning Nuclear Facilities ," in Proceedings of the International Symposium on the Management of Wastes from the LWR ~ Fuel Cycle, CONF-76-070' sponsored by the Energy Re se arc h and Development Administration, Jenver, Colo. , July 1976. 8-16 William J. Manion and Thomas S. LaGuardia, An Engineering Evaluation of Nuclear Power Reactor Decommissioning Alternatives, AIF/ NESP-009SR, Atomic Industrial Forum, National Environmentai Studies Project, November 1976. 8-17 Improvements Needed in the Land Disposal of Redioactive Wastes--A Problem of Centvries, REP-76-54, U.S. General Accounting Office, January 1976. 8-18 Reactor Safety Study-An Assessment of Accident Risks in U.S. Commercial Nuclear Power Plaats e WASE-1400, Appendix VIII, NUREG-75/014, Oc tober 1975. 8-19 Reactor Safety Study - An t.ssessment of Accident Risks in U.S. Commercial Nuclear Power Plants, WASH-1400, NUREG-75/014, October 1975, supporting material. 8-20 Reactor Safety Study - An Assessment of Accident Risks in U.S. Commercial Nuclear Power Plants, WASH-1400, Appendix VI, NUREG-75/014, Oc tober 1975. 3-24
DRAFT 9.0 Options and Implications The law which required that the Energy Commission undertake this study also stipulated that the Commission determine "...whether to require by rules and regulations that nuclear reactors be either undergrounded or berm contained." (See Appendix A for complete text of AB 2821.) To assist the Commission in identifying and assessing the significance of the many motivations for under-ground siting, the discipline of applied. This Section draws substantially from that work.g_igon analysis was The decision analysis identified cost and safety as the primary attributes. All other considerations, such as the possibility of urban siting, decommissioning, war protection, improved- earthquake resistance, while perhaps imporcant in individual situations, were clearly secondary in the general decision of whether or not to require underground siting. Of major interest were two fundamental questions: what will underground siting buy in terms of enhanced safety and what will it cost? The following pages will present various decision options within the general context of cost and safety, underground siting, and alternatives. 9.1 Spectrum of Possible Decisions The Commission is not necessarily limited to either requiring that reactors be placed underground or not requiring such action. There is a spectrum of possi-bilities th.at range from, literally, all to nothing. Each decision option has corollary implications which, after a brief listing of the options , will be discussed. Possible decisions, and certainly not all, are as follows:
- 1. Require that all future nuclear power plants incorporate either berm or mined-cavern construction.
- 2. Zone the state such that underground construction is required in some areas but not in others.
- 3. Require that surface and underground construction be proposed as alterna-tives for each site in the Notice of Intention (NOI), with the decision to be made on a site by site basis.
- 4. Require that an NOI provide one or more alternative sites appropriate for underground construction.
- 5. Leave the siting-mode decision to the utility.
- 6. Prohibit berm and/or underground construction.
- 7. Make no decision requiring underground siting until a pilou plant has been completed.
- 8. Implement a less effective alternative to underground siting.
9-1
DRAFT 9.2 Safety Implications of Underground Siting In order to understand some of the implications of the listed options, a basic dilemma which relates to the safety criteria presently used in licensing nuclear power plants must first b3 understood. Applicants must demonstrate to regula-tory agencies that the proposed plant will operate safely without posing an undue hazard to the public. To ensure that accidents of the class considered in this study could not occur, engineered safeguards are provided. These engi-neered safeguards generally involve active systems composed of valves, complex plumbing, instrumentation and control systems and so forth. Moreover, the systems, if called to action, would be operating in an environment which is difficult to predict and under conditions where engineering behavior is not well understood. While every effort is made to provide reliable systems of the utmost quality and conservative design standards, uncertainty still exists as to whether these systems will operate when required and as, required. This study is motivated, in part, by these residual uncertainties. As described in Section 3, possible accidents have been classified in categories of from 1 to 9. Regulatory focus is on those accidents ranging up to Class 8, with some consideration of potential Class 9 events by virtue of the siting criteria used to isolate plants from urban concentrations. The regulatory intent is to prevent the most severe " credible," Class 8, accidents from becom-ing more serious, " incredible" Class 9 events. Massive failure of the Engi-neered Safeguards would be required before a Class 8 accident could degenerate to the most . severe Class 9. Underground siting is directed at mitigation of public consequences from, as yet hypothetical, Class 9 accidents. Thus, a determination concerning underground siting is, in effect, a finding regarding the need to consider extreme reactor accidents, e.g., Class 9 or core-melts. Alternatives to underground siting must be considered in the same light. 9.3 Implications of Decision Options For the first listed option, which requires that all future nuclear power plants incorporate some form of underground construction, it must be recognized that the law as presently written exempts any powerplant for which a Notice of Intention is filed before January 1, 1980. It is possible that many applica-tions for surface construction would be filed prior to the exemption termination date if this option were selected. After that time, due to the incremental cost, uncertainty of schedule, and licensing time for underground nuclear plants, based on the economic assumption of this study, it is probable that no applications for nuclear facilities would be filed. The cost increment for underground siting would greatly distort the relative economics of base load alternatives. Thus, a decision to require underg ound siting would likely force the state's utilities to consider and develop alternatives. If, under decision option one, an application for an underground-sited nuclear plant were pursued, assuming optimistically an additional three years of design and licensing time for the pilot plant with a 12-year " mature industry" licens-ing/ construction schedule, the first plant would not be on-line until 1995. Much of the motivation for underground construction stems from uncertainty in the actual safety of nuclear facilities. The intervening 18-year period would do much to reduce the uncertainty from surface-constructed plants in other states, if not other countries. It is unlikely that a utility would cake the incremental investment for underground siting with alternatives present. 9-2
DRAFT The second option, to zone the state and indicate those areas where underground siting would be required, is, in the converse, to indicate where surface siting would not be permitted. Presumably, in the remaining areas, surface siting could be considered. (Implicit in this option is the assumption that there exist acceptable surface siting areas.) In effect, this option is equivalent to implementing an alternative to underground siting, i.e., remote siting. . It was concluded in Section 4 that the most likely siting areas for underground plants in Californie are in the same general areas as current proposals for surface plants. These areas, by U.S. standards, tend to be remote. The third decision option, requires that for every proposed site in an NOI, an underground option be included. While, in general, a berm-contained facility may be constructed as an alternative to a surface plant at most sites, the same is not true for a mined-cavern plant. The latter requires a large expanse of high quality rock, and while such material is present in the state, it is not widely distributed. Thus, this option becomes, in general, a ques *. ion of whether or not a proposed facility should be berm-contained rather than surface sited. The fundamental question is the acceptability of the degree of risk posed by the surface plant at that site. If found to be unacceptable, it is likely that alternative sites would be sought rather than incur the incremental cost of underground siting. Decision option four, requiring that alternative sites appropriate for under-ground construction be included in the NOI, is fundamentally equivalent to option three. The same basic questions are posed. The fifth option, allowing the applicant utility company to determine whether or not to build underground, has actually two corollary alternatives. First, if no requirement were imposed to consider extreme accidents, the status quo would prevail. If such a requirement was mandated, then the burden wculd be on the applicant to demonstrate that his preferred option, whether it be underground siting, remote siting,_or an alternative such as controlled, filtered venting, is adequate. However, such an option requires a clear set of criteria against which proposals may be measured. Development of such criteria would be difficult since they answer the question of how safe is safe enough. This study has found no reason that underground construction should be pro-hibited, the sixth option. The concepts have been found to be technically feasible, effective in mitigating the consequences of extreme accidents, and not prohibitively more expensive. Option seven, delaying a decision until completion of a pilot plant, while interesting, is not responsive to the requirements of the law. The objective - would be to reduce uncertainty over costs, licensing and constructability, through construction of a pilot facility. However, it would take nearly two decades from the present to complete construction, and would leave current problems unresolved. The eighth option, requiring that a less effective alternative to underground siting be implemented, holds promise. This study, and others as well, have noted the general effectiveness of controlled, filtered venting (CFV) as a safety augmentation concept. CFV is attractive, since it - functions essentially analogously to the underground designs of this study, although it is not as 9-3
DRAFT effective in as wide a range of containment failure modes as is underground siting. It is, however, significantly less expensive than sub-surface siting and in terms of incremental risk reduction, more cost-ef fective. Remote siting is also an attractive alternative for general risk reduction if sites are carefully ci;osen to minimize the potential for acute fatalities. As suggested in Section 8, prudent siting carries the opportunity to change the character and magnitude of the most extreme reactor accidents. Assuming that the residual impact is acceptable, remote siting can be considered an alterna-tive to underground siting. Such a decision, however, can only be made on a site-by-site basis. This discussion has not been intended as exhaustive, but only indicative of the many decision options available. Certainly, there are others. 9.4 Recommendations The following are staff recommendations based on the findings of this study. They reflect the uncertainty of basic data in some areas and suggest means by which this uncertainty may be reduced. In addition, several concepts which have been identified as promising are suggest.ed for further investigation. o Underground siting should not be mandated. This recommendation stems from (a) the uncertainty remaining over costs, construction time and possible licensing concerns; (b) he existence of what appear to be moderately effective and less expensive technical alternatives; and (c) the oppor-tunity to implement remote siting within California. o In view of the potential safety benefits of remote siting, the Commission should consider the relative safety benefits of alternative proposed sites under all accident conditions for nuclear plants proposed for California. o Further basic work should be sponsored, or actions taken to have such efforts sponsored by industry and the appropriate Federal agencies, in the following areas: o Evaluation of the concept of a passive " engineered failure pathway", possibly similar to the berm-containment Level-2 designs of this study, as applied to current generation surface facilities. o A detailed assessment of controlled, filtered venting for accident management. o Development of risk methodologies suitable for analysis and relative comparison of alternative reactor sites which reflect population, meteorology and topography. o Continued development and use of formal techniques to identify major contributors to relative risk. o A program of continued safety improvements so that the estimated total level of risk for all currently operating reactors remains acceptable as new reactors become operational. This will require that the fundamental question of acceptable levels of technological risk be addressed; e.g., that the question of "how safe is safe enough" be answered. 9-4
DRAFT REFERENCES SECTION 9 9-1 Smith, Je ffrey H. , Ralph E. Miles, o.., and Martin Goldsmith, An Appli-cation of Multi-Attribute Decision Theory to the Underground Siting of Nuclear Power Plants, Jet Propulsion Laboratory, Pasadena, California, May 1978. 9-5
DRAFT Appendix A California Assembly Bill 2821 (adds Section 25524.3 to the Public Resources Code, ef fective January 1,1977) The people of the State of California do enact as follows: SECTION 1. Section 25524.3 is added to the Public Resources Code, to read: 25524.3. No nuclear fission thermal powerplant, including any co which the provisions of this chapter do not otherwise apply, but excepting any having vested rights as defined in this section, shall be permitted land use in the state, or, where applicable, be certified by the commission until the following conditions have been met: (a) The commission has undertaken and completed a study of the necessity for, and effectiveness and economic feasibility of, undergrounding and berm containment of nuclear reactors, and, upon completion of the study, the commission after public hearings, has determined whether to require by rules and regulations that nuclear rcr- ors be either undergrounded or berm aca.ained. The commission shall < .4 te such study and submit it to the Legislature, with conclusions and u mendations, within one year from the effective date of this section. (b) In the event that the commission determines that undergrounding and berm containment are necessary, effective, or economically feasible to meet the requirements of Sections 25511 and 25520, relating to enhancing the public health and safety at a site and a related facility, rules and regulations implementing such findings shall be suspended for a period of one year so as to provide the Legislature with the necessary time to evaluate and verify the results of the study. (c) In the event that the commission determines that undergrounding and berm containment are not necessary, ef fective, or economically feasible to ineet the requirements of Sections 25511 and 25520, relating to enhancing the public health and safety at a site, no nuclear fission themal powerplant shall be approved for one year thereaf ter, so as to provide the Legislature with the necessary time to evaluate the study for possible statutory implementation. The commission shall continue to receive and process notices of intention and applications for certification pursuant to this division, but shall not issue a decision pursuant to Section 25523 granting a certificate until the requirements of this section have been met. All other permits, licenses, approvals or authorizations for the entry or use of the land, including orders o f cour t , which may be required may be processed and granted by the governmental entity concerned but construction work to install permanent equipment or structures shall not commence until the requirements of this section have been met. A-1
DRAFT For purposes of this section, the vested right to construct a nuclear thermal powerplant shall exist if, prior to the date on which this section is chaptered, . an elec tric utility has performed substantial construction on such powerplant and has incurred substantial expense for construction and for necessary ma-terials for such powerplant, including, but not limited to, the following sites and facilities, with the associated estimated generating capacities. (1) As designated in the report of the Pacific Gas and Electric Company submitted to the Public Utilities Commission on December 23, 1966, pursuant to Section 1001 of the Public Utilities Code, one nuclear thermal power-plant, having a generating capacity of 1,060 megawatts, commonly known as Diablo Canyon Unit 1, to be located in San Luis Obispo County. (2) As designated in the report of the Pacific Gas and Electric Company submitted to the Public Utilities Commission on February 16, 1968, pursuant to Section 1001 of the Public Utilities Code, one nuclear thernal power-plant, having a generating capacity of 1,060 megawatts, commonly known as Diablo Canyon Unit 2, to be located in San Luis Obispo County. (3) As designated in the report of the Southern California Edison Company and the San Diego Gas and Electric Company to the Public Utilities Commission on July 16, 1970, pursuant to Section 1001 of the Public Utilities Code, two nuclear thermal powerplants, having a generating capacity of 1,100 megawatts per unit, commonly known as San Onofre Unit 2 and San Onofre Unit 3, to be located in San Diego County. Notwithstanding any provisions of this section to the contrary, the provisions of this section shall not apply to any nuclear 'ission thermal powerplant site and related facility for which a notice of intent has been filed pursuant to Section 25502 with, and accepted by, the commission within three years of the ef fective date of this act. SECTION 2. The provisions of this act shall not become operative if Proposition 15 of the June 1976 election is adopted by the people, whether or not this bill is chaptered before or after such adoption. Public Resources Code Sections Re ferenced in AB 2821 PRC Section 25502: Each person proposing to construct a thermal powerplant or electric transmission line on a site shall submit to the commission notice o' intention to file an application for the certification of such site and related facility or facilities. The notice shall be an attempt primarily to determine suitability of the proposed sites to accommodate the facilities and to determine the general conformity of the proposed sites and rel'ated facilities with standard of the commission and forecasts adopted pursuant to Sections 25216.3 and 25309. The notice shall be in the form prescribed by the commission and shall be supported by such information as the commissisn may require. Any site and related facility once found to be acceptable pursuant to Section 25516 is, and shall continue to be, eligible for consideration in an application or certification with fur ther pr;eeedings for a notice under this chapter. PRC Section 25511: The commission shall review the facto s related to sa fe ty and reliability of the facilities at each of the alternative sites designated in A-2
DRAFT the notice. In addition to other information requested of the applicant , the commission shall, in determining the appropriateness of sites and related facilities, require detailed information on proposed emergency systems and safety precautions, plans for transport, handling and storage of wastes and fuels, proposed methods to prevent illegal diversion of nuclear fuels, special design features to account for seismic and other potential hazards , proposed methods to control density of population in areas surrounding nuclear power-plants, and such othe r information as the commission may determine to be relevant to the reliability and safety of the facility at the proposed sites. The commission shall anslyze the information provided by the applicant, sup-plementing it, where nuessary, by onsite investigations and other studies. The commission shall determine the adequacy of measures proposed by the applicant to protect public health and sa fe ty , and shall include its findings in the pre-liminary report required by Section 25510. PRC Section 25520: The application shall contain the following and such other information as the commission by regulation may require: (a) A detailed description of the design, cons truc tion, and operation of the proposed facility. (b) Safety and reliability information, including, in addition to documentation previously provided pursuant to Section 25511, planned provisions for emergency operations and shutdowns. (c) Available site in formation, including maps and descriptions of present and proposed development and, as appropriate, geological, aesthetic, ecological, seismic, water supply, population and load center data, and justification for the particular site proposed. (d) Such other information relating to the design, operation, and siting of the facility as the commission may specify. (e) A statement of need providing information showing compatibility of the proposed facility with the most recent biennial report issued by the commission pursuant to Sect!'n 25309. ( f) A description of the facility, the cost of the facility, the fuel to be used, the source of fuel, fuel cost, plant service life and capacity factor, and generating cost per kilowatt hour. (g) A description of any electric traasmission lines including the estimated cost of the proposed electric transmission line; a map in suitable scale of the proposed routing showing details of the rights-of-way in the vicinity of settled areas, parks, recreational areas, and scenic areas, and existing transmission lines within one mile of the proposed route; and justification for the route and a preliminary description of the effect of the proposed electric transmission line on the environment, ecology and scenic, historic and recreational values. PRC Section 25523: The commission shall prepare a written decision after a public hearing or hearings on an application, which shall include all of the following: A-3
DRAFT (a) Specific provisions relating to the manner in which the proposed f a .ility is to be designed, sited, and operated in order to protect environmental quality and assure public health and safety. (b) Findings regarding the conformity of the proposed site and related facil-ities with standards adopted by the commission pursuant to Section 25216.3 and subdivision (d) of Section 25402, with public safety standards and the applicable air and water quality standards, and with other relevant local, regional, state and federal standards, ordinances, or laws. If the commission finds that there is noncompliance with any state, local, or regional ordinance or regulation in the application, it shall consult and meet with the state, local, or regional governmental agency concerned to attempt to correct or eliminate the noncompliance. If the noncompliance cannot be corrected or eliminated, the commission shall inform the state, local, or regional governmental agency if it makes the findings required by Section 25525. (c) Provision for restoring the site as necessary to protect the environment, if the commission denies approval of the application. (d) Findings regarding the conformity of the proposed facility with the 10 year forecast of statewide and service area electric power demands adopted pursuant to subdivision (b) of Section 25309. A-4.
DRAFT Appendix B Specifications for a Study of Buried, Berm-Containment Design Concepts for Underground Nuclear Power Plants Baseline Sensitivity I. Power Plant Characteristics Reactor and Thermal Capacity Westinghouse RESAR 414 (3800 MW) X General Electric GESSAR (3800 MW) (1) Babcock & Wilcox Standard 241 (3800 MW) (2) Combustion Engineering CESSAR (3800 MW) (2) Plant Electrical Capacity,1300 MW Net X NA Cooling Water Supply Concept - Forced Convection X NA Number of Units Sharing Common Facility Single NA Nuclear Steam Supply Configuration o Standard Vendor Configuration RESAR 414 (1) GESSAR o Alternative NSSS Configuration NA NA to Accommodate Dimensional Constraints of Buried Containment II. System / Component Locations Nuclear Steam Supply System Subsurface NA Waste Processing and Storage Subsurface NA Turbine Generator Sur face NA Main Power Transformer, Auxiliary Transformer, Substation Surface NA Spent and New Fuel Storage and aandling Subsurface NA Control Center Subsurface NA Emergency Diesel Generators Subsur face NA P-1
DRAFT Baseline Sensitivity III. Siting Data Soil Description
, Soil Soil Variation with Depth Model 1 Model 2 Soil Moisture Moist Dry Groundwater Specification Facility Location with Respect to Water Facility Water Table Totally Table Above at Water Ground Table Level Seismic Criteria - Surface Peak Acceleration 0.5 g a) 0.67 g Rock Characteristics NA a) Soil Model 1 Site Remoteness Central East Desert Valley Remote Rural Meteorological Conditions Humid Dry IV. Fission Product Containment Requirements Pressure-Time Relationships B/L B/L Energy-Time Relationships B/L B/L Pressure Management Design Concepts Level 2 a) Level I b) Level 3 Minimum Depth of Containment Cover 50 ft. a) 25 ft.
b) 100 ft. Surface / Underground Structure Interface Relationships Allowable Leakage Rates 0.1% per B/L day V. Schedule Variations Nominal a) Short b) Long VI. Reference Surface Power Plant Characteristics Power Plant Characteristics B/L (or NA equivalent)
. B-2
DRAFT Baseline Sensitivity System / Component Locations Standard or NA Replicated Surface Facility B/L Siting Data B/L NA Fission Product Containment Class 8 Level 3 (DBA) (1) BWR concept is a major design alternative for which plant layout must be prepared. (2) No plant layouts are required.
- B/L = Baseline. (Soil Model 1)
B-3
DRAFT Specifications for a Study of Mined Underground Design Concepts for Underground Nuclear Power Plants Baseline Sensitivity I. Power Plant Characteristics Reactor and Thermal Capacity Westinghouse RESAR 414 (3800 MW) X General Electric GESSAR (3800 MW) (1) Babcock & Wilcox Standard 241 (3800 MW) (2) Combustion Engineering CESSAR (3800 MW) (2) Plant Electrical Capacity,1300 MW Net X NA Cooling Water Supply Concept - Mechanical X NA Evaporative Number of Units Sharing Common Facility Single NA Nuclear Steam Supply Configuration o Standard Vendor Configuration RESAR 414 (1) GESSAR o Alternative PWR Configuration to NA YES Accommodate Dimensional Constraints Underground Containment II. System / Component Locations Nuclear Steam Supply System Subsurface NA Waete Processing and Storage Sub sur face NA Turbine Generator and Main Trans former Subsurface Surface Spent and New Fuel Storage and Handling Subsurface NA Control Center Subsurface NA Emergency Diesel Generators Sub sur face NA III. Siting Data Rock Description Rock Model 1 a) Rock Model 2 b) Rock Model 3 Groundwater Specifications Dry Dewa ter
- B-4
~
' DRAFT Baseline Sensitivity Seismic Criteria - Surface Peak 0.5 g a) 0.3 g Acceleration b) 0.67 g To; graphy Hillside Flat (40% Slope)
. Site Remoteness Central East Desert Valley Rural Remote Meteorological Conditions Humid Dry IV. Fission Product Containment Requirements Pressure-Time Relationships TBD NA Energy-Time Relationships TBD NA Pressure Management Design Concepts Level 2 a) Level I b) Level 3 c) BWR Level 3 Minimum Depth of Containment Cover 50 ft. 1000 ft. Surface / Underground Structure Interface Relationships Allowable Leakage Rates 0.1% NA per day V. Schedule Variations Nominal a) Short b) Long VI. Reference Surface Power Plant Characteristics (PWR) Power Plant Characteristics Baseline (or NA Equivalent) System /Comoonent Locations Standard or NA Replicated Surface Facility Siting Data (3) NA Fission Product Containment Class 8 (DBA) NA B-5
DRAFT Lifetime 40 years Condenser Pressure 3.5 in-See Hg. (1) BWR concept is a major design alternative for which a plant layout must be prepared. (2) No plant layouts are required. (3) Site Geology: Baseline-Soil (Soil Model 1) Groundwater: NA Seismic : 0.5 g Remoteness: Central Valley Rural Meteorological: Relative Humidity 22:; Wet bulb temp. 70*F. 3-6}}