ML20246F380

Design and Feasibility of Accident Mitigation Systems for Light Water Reactors.Final Report

ML20246F380
Person / Time
Site:	Limerick
Issue date:	08/31/1985
From:	Castle J, Catton I, Dooley J, Hammond R, Kastenberg W R&D ASSOCIATES
To:	NRC
References
NUREG-CR-4025, NUDOCS 8908300310
Download: ML20246F380 (214)

Text

.

-~m 5 y % : o;a ; ,

(l O'

RDA-TR-127303-003 NUREG/CR-4025f y ss* - $s.*/ sis l

DESIGN AND FEASIBILITY OF ACCIDENT MITIGATION f SYSTEMS FOR LIGHT WATER REACTORS

~

FINAL REPORT I

AUGUST .1985 l JAMES L. DOOLEY R. PHILIP HAMMOND WILLIAM E. KASTENBERG l: IVAN CATTON JAMES N. CASTLE WITHOUT PROPRIETARY INFORMATION R & D Associates P.O. Box 9695 i Marina del Rey, CA 90295 I

- Prepared For U.S. Nuclear Regulatory Commission 8

8908300310 850831 *2 hl i

DR ADOCK 0500 g

1. (

4 . .: ..

F L . RDA-TR-127303-003 NUREG/CR-4025 DESIGN AND FEASIBILITY OF ACCIDENT MITIGATION SYSTEMS FOR LIGHT WATER REACTORS FINAL REPORT AUGUST 1985 JAMES L. DOOLEY R. PHILIP HAMMOND WILLIAM E. KASTENBERG IVAN CATTON JAMES N. CASTLE

l l

R & D Associates

, P.O. Box 9695 Marina del Rey, CA 9C295 Prepared For U.S. Nuclear Regulatory Commission I

__ --_ - - - _ _- - -- a

A' ,'

i I

NOTICE This report was propered as an account of work sponsored by an agency of the United l

States Government. Neither the United States Government nor any agency thereof, .

or any of their employees, makes any .

warranty, expressed or implied, or assumes any legal liability or responsibility for any third party's use, or the results of such l use, of any information, apparatus, product I or process disclosed in this report, or represents that its use by such third party would not infringe privately owned rights.

I-The views expressed in tnis report are not necessarily those of the U.S. Nuclear Regulatory Commission.

l l

l l .

i' 1'

1

rg.,

ll -

L ABSTRACT

- Conceptual designs and. cost estimates are given for complete litigations' systems tailored to the BWR nuclear power plant design having the Mark II

- containment, the-BWR with the Mark III containment (GESSAR), and to-the PWR WSP-90 nuclear power

- plant. Each system is intended to intercept all the dominant ~ modes'of containment failure in a severe core-melt accident, and to remove essenti-l ally all the contingent risk to the public from such accidents. The incentive for such mitigation systems is assessed using value/ impact analysis based upon'the. human risk averted over the plant lifetime and the dollar cost of the system.

I l

d e

1 l

I

- - _ - _ _ _ -_- l

.-~ - _ - - . - - _ _ . ._- - - . _

i; EXECUTIVE

SUMMARY

PURPOSE AND SCOPE:

-The task of this report is,to assess the feasibility, h

costs, and-benefits of proposed systems for mitigating the

consequences to the public of severe core-melt accidents in o nuclear power plants. . Mitigation, as,used here, is to be L distinguishedifrom accident prevention-and refers primarily.

to methods of enhancing the ability of the containment structure to. resist failure or leakage of radioactive maturials.

Previous reports in this project have- been entitled Survev of Licht Water Reactor Containment Sys tems , Dominant Failure

. Modes , and Mi tica ti on Occortunit ies' (NUREG/CR-4 24 2 ) ,, and Survev of the State of ' the Art in Mi ti ca ti on Sv st ems .,

(NUREG/CR-3908). Thesa gave the results of exhaustive sur-veys of the ways-that various types of U.S.. containments tend to failfin severe-accidents, and of the technology available to improve.them.

l In this report the above results are applied to specific cases for more complete assessment. A detailed conceptual

' design is undertaken of a complete mitigation system for each of three types of plants--the Boiling Water Reactor l (BWR) with a Mark II containment, the BWR with a Mark III containment, or General Electric (GE) standard plant design (GESSAR), and the Westinghouse advanced Pressurized Water Reactor (PWR) known as WSP-90. Two additional reports, NUREG/CR-4243 and NUREG/CR-4244 give value/ impact assessment methodology developed for this work, and possible strategies for applying the results to implementing a mitigation policy.

For each type of plant the design and cost estimates are based on " complete" mitigation, meaning a system capable of

. intercepting all the expected modes of failure from the 1 dominant accident sequences. The requirement for complete

. mitigation arises from the observation that a severe acci-dent in progress has the capability of causing more than one mode of containment failure. If a mitigation system is installed that would prevent only one type of failure, the accident in progress would not be arrested, but would con-tinue until it caused a different mode of failure. Thus an investment in limited mitigation may have less cost / benefit justification than a " complete" system.

lit

A second design basis is applied where there are significant uncertainties as to the phenomena occurring during the course of a severe core-melt accident. Then the mitigation system proposed must be one that controls or forces the phenomena into a known path. This requirement is necessary since the results of this study may affect major regulatory policien, and the feasibility of mitigation should not be based:u;t.n questionable phenomenological assumptions.

OVERVIEW OF RESULTS:

The resalts of this work show conclusively that it is tech-nically feasible to install a system that will materially reduce the possibility of containment failure in a severe accident. This conclusion is based on the three plants studied, but ne reason was found that it would not be valid for any U.S. plant. The cost estimates for such systems were found to vary according to whether the plant was still on the drawing board, partially built, or in operation. the latter case being of course the most costly. The costs ranged frem very small amounts in certain special cases to S10-15 million for backfitting an operating plant. These costs include hardware and installation: they do not include replacement power costs for an operating plant.

The detallec assumptions and basis of the costing are fully described in the appendix to this report.

The assessment of benefits from a complete mitigation .

system was performed by methods developed in parallel tasks I of this project, described briefly in this report and characterized in full in accompanying reports. The work  ;

was based on updated Probabilistic Risk Assessment (FRA) l reports and reviews of such reports by Brookhaven National l Laboratory (BNL) and others. For each plant the dominant accident sequences and failure modes were classified and used to estimate the expected value of the public exposure to radiation from accidents over the lifetime of the plant.

both for a radius of 50 miles and for 500 miles. The ratio of dollars spent on mitigation systems to the man-rem of public dose averted by the system was calculated as a basis of comparison. The cases where this ratio fell below S1000 per man-rem averted were taken to be worthy of further study; such cases were found for each type of plant.

Although benefits estimated from calculated man-rem exco-sures are subject to considerable uncertainty, the indi-cations are clear that conventional mitigation systems are worthy of consideration as a part of severe accident colicy making.

l l

f iv l

I

Other types of benefits from mitigation besides the reduc-tion of public radiation dose will be studied in a later task. Such benefits would include reduction in required public service and emergency facilities, reduction in land use losses, lessened requirements for evacuation or emer-gency training, lower insurance costs, lower financing costs for utility investments and lower plant operating costs. Some of these benefits would also result in lower utility rates.

An example of such a direct monetary benefit from mitigation arose in this study in considering a new type of plant containment as a means of meeting mitigation requirements. In-this scheme the reactor system was surrounded not by a pressure vessel but by a low-pressure enclosure which was continuously vented to a large capacity, low. temperature filter systca. This chill-vent system was judged to be completely effective and reliable in trapping essentially all radioactive emissions, thus mitigating the consequences of both design basis and severe accidents. In addition it showed a direct monetary benefit.

That is, the overall system cost was almost certainly less than that of a standard high pressure containment with no mitigation features. The low pressure containment essen-tially removes the accident sequences ending in overpres-sure failure and bypass, with major reduction in the l remaining uncertainties 1. accident consequences. This finding was an important result of the work, and further

[ effort on the design, cost, and accident responses of I vented low-pressure containments is recommended.

The three specific plant types chosen for study represent a l variety of approaches to containment design, accident con-trol, and general arrangement, yet in overview the most striking result of this study is that all these systems ere vulnerable to the same types of containment failure, and l the mitigation systems required are similar in function.

Severe core-melt accidents in any type of plant require control of combustible hydrogen, means of limiting the I formation or accumulation of other non-condensible gases, and non-electrical means of removing denwy heat from the containment without release of radioactive materials.

UNCERTAINTIES Although the conclusion is positive that mitigation is feasible, the treatment of uncertainty will play an important role in any policy decisions. There is little uncertainty that mitigation features can be designed, built and installed with conventional engtysering components, 1

V l

l f

l L_____----_-_

9 lt that they will function as designed, and that their costs can be determined. .There is also little uncertainty that in the absence of a mitigation system.most containments would ultimately fail in one of-the several modes discussed in this report if a severe, core melt accident does occur.

What remains uncertain is whether the accident sequences will-have the frequencies predicted, whether the contain-ments will fail with the probability 6 distribution calcu-

' lated, and whether the ensuing consequences have been

~

correctly. estimated. Thus the removal of residual risk and the cost of doing so are relatively certain, but what that risk is and the benefits of its removal are uncertain.

However, as noted above, there are other potential benefits of mitigation, including those derived from the reduction of uncertainty itself. These considerations will be pursued in a later report..

vi h

L .

CONTENTS' Chapter' Page f: 1l INTRODUCTION AND BACKGROUND 1-1 1.1; Objectives'of the' Severe-Accident Mitigation Program 1-1 1.2 Scope ~and Definitions 2 R 1.3; Background- 1-4' l.4 Selection of Containments-.to be  :

Considered' l-5. l 1.

2 TECHNICAL ASSESSMENT METHODS' DEVELOPED' I FOR THIS-STUDY 2-1 2.1 Determining Dominant Failure Modes and Mitigation Requirements 2-1 2.2 LValue/ Impact Analysis of Mitigation y Systems 2-3 2.3 Design' Assumptions and Ground Rules 2-8 j 2.4 Costing Methods 2-9 2.5 Limitations to this Study 2-13 f 2.6 Uncertainties 2-14 2.7 A New Mitigation Concept 2-14 3 MITIGATION DESIGNS FOR MARK II PLANTS 3-1

(' 3.1 Description of.the Mark II Containment 3-1

-3.2 Containment Failure Modes 3-8 3.3 Mitigation Requirements to be Met 3-24 3.4 Component Designs, Descriptions, and Costs 3-26 3.5 Alternative Mitigation Strategy- 3-61 3.6 Discussion of Mitigation Options 3-70 3.7 Summary 3-76 vii f

f

l CONTENTS (CONCLUDED) l

' Chapter' Page 4 MITIGATION DESIGNS FOR BWR MARK III PLANTS 4-1 4.1 Description of the Mark III Containment' and its Engineered Safety Features 4-1 4.2 Containment Failure Modes 4 -7 4.3 Mitigation Requirements to be Met 4-8 ~

4.4 Strategies for' Conventional' Mark III 4-21 4.5 Alternative Strategy: The Unpres-surized Mark III- 4-34 4.6 Options'and Value/ Impact Comparisons 4-42 4.7 Effect of External Events 4-45 4.8 Other Mitigation Systems 4-45 4.9 Summary 4-47 5 MITIGATION SYSTEMS FOR WSP-90 PLANTS 5-1 5.1 Description of the Advanced Large Dry PWR Containtzent 5-2 5.2 Containment Failure Modes and Mitigation Requirements 5-8 5.3 Strategies for the Advanced PWR Con-tainments 5-11 5.4 Mitigation Strategy Using Conventional

! Components 5-11 l 5.5 Alternative Strategy--Unpressurized Containments 5-13 5.6 Summary 5-25 APPENDIX A. COST ESTIMATES OF MITIGATION COMPONENTS A-1 REFERENCES R-1 viii E___________.__________________________________.__________

s  :.  :.

i FIGURES Figure' Page 2

31 Typical Mark II Containment 3-2' 3'-2 Pertinent Dimensions 3-3 3-3 Decay Heat Rate and Heat Accumulation 1with Time 13-6 3-4 A Perspective of the Mitigation Energies (May.be Cumulative) 3-7 .,

3-5 Schematic Dual Heat Removal System 3-29 3-6 Plan of Dedicated Heat Removal Plant 3-30 3-7 Schematic - Double Isolation Valves for

-Intake 3-33 3 Sprays into Upper Drywell 3-36 3-9 Schematic of Externally Installed Drywell Spray System 3-37 3-10 Possible Debris Path on Diaphragm Floor 3-41 3-11 Core Debris Path after Melt Through - No l Protection- 3-42 3-12 Core Debris on Basemat below Downcomer Ducts 3 1 3-13 Schematic - Direct Water-Cooled Rubble Bed i Type of Core Retention 3-45 3-14 Access and Shielding Problems Illustrated 3-46

l. 3-15 Schematic - Detail of Dry Crucible Core Debris Retention and Underground Cooling Unit 3-51 3-16 Overall Schematic of Mark II Underground

.I Core Retention 3-52 3-17 Schematic - Pressure Relief Valves and

- Condenser / Filter 3-53 3-18 Steam / Gas Diverter Valve to Handle ATWS Flow 3-56

^

3-19 Schematic - Open Mark II Containment with Chilled Filter 3-63 3-20 Schematic of Chilled Filter System 3-65 l 4-1 Typical BWR/6 Mark III Containment 4-2 4-2 Mark III containment Dimensions 4-3 4-3 Free Volumes and Typical Pressure Capability 4-4 ix

_ _ _ _ . . _ _ _ . _ .i

H -

h.$

)

FIGURL'S (CONCLUDED)

Figure = Pace 4-4 Containment Scavenge Efficiency 4-15 4 Schematic Detail of. Dry Crucible Core Debris Retention and. Cooling' 4-32 4-6' Schematic Chilled Filter Installation ~ 4-38 E 5-1 Section Through WSP-90 Containment 5-3 5-2 'Section Through WSP-90 Containment 5-4 3 5-3 Plan of WSP-90' Containment 5 5-4 Schematic of Separate Mitigation Unit 5-18 5-5 Comparison .of Containment volumes and Design Pressures -(Typical 1200 MWe Plsnts) 5-20 j 5-6 Schematic - Separate Ice. Condenser and Filter System 5-24  ;

. 1

'l 4

i i

l x 1 a

I r  !

TABLES Table Pace 3-1 General Plant Information 3-4 3-2 Dominant Sequences by Containment Failure Class for Limerick (BNL-Review) 3-9 3-3 Key to Limerick Sequence Symbols 3-10 3-4 Summary of the Dominant Accident Sequence Frequencies which Lead to Core Vulnerable States (per Reactor Year) by Initiator and Class for Shoreham 3-12 3-5 Summary of Containment Failure Classes for Mark II Containment Systems 3-15 3-6 Definition of Containment Failure Modes (Limerick) 3-18 3-7 Conditional Probability of Containment Failure, Release Category and Class Frequency 3-19 3-8 Consequences for Each Release Category 3-20 3-9 Acute (2arly Fatalities)/ Year for Each l Containment Failure Mode - Internal Initiators (with ATWS-3A-Fix) 3-20

'l i

3-10 Latent Fatalities / Year for Each Containment Failure Mode - Internal Initiators (with ATWS-3A-Fix) Out to 500 Mi 3-21

, l 3-11 Man-Rem / Year (Gut to 50 Mi) for Each Con-l tainment Failure Mode - Internal Initiators (with ATWS-3A-Fix) 3-21 l

l 3-12 Man-Rem / Year (Out to 500 M1) for Each Jon-tainment Failure Mode - Internal Initiators (with ATWS-3A-Fix) 3-22 3-13 Man-Rem / Year (Out to 50 Mi) for Each Con-tainment Failure Mode - Internal Initiators (without ATWS-3A-Fix) 3-22 3-14 Man-Rem / Year (out to 500 Mi) for Each Con-tainment Failure Mode - Internal Initiators (without ATWS-3A-Fix) 3-23 3-15 Time Estimate to Install Wet Rubble Bed 3-48 3-16 Mitigation Options - Pressurized Containment 3-72 xi I l

~

l l

4 I

TABLES (CONTINUED) '

Table Page 3-17 Containment Mitigation - High pressure vs Low Pressure 3-75 4-1 General Information GESSAR II Containments 4-5 4-2 GESSAR II Accident Class Frequencies 4-9 4-3 GESSAR II Coremelt Sequence and Release -

Category Nomenclature 4-10 4-4 GESSAR II Consequences by Release Category 4-14 4-5 NRC Staff Risk Results for Selected Accident Sequences 4-14 4-6 Contribution to Risk by Containment Failure Mode from GESSAR II PRA 4-17 4-7 Characteristics of Liquid Nitrogen 4-26 4-8 Cost of Mitigation for Conventional GESSAR Containment 4-33 4-9 Cost Comparison Between Conventional and Non-Pressurized Containments 4-42 4-10 Mark III Containment Mitigation 4-44 5-1 General Information--Typical Power Reactor and Large Dry Containment 5-6 5-2 Dominant Failure Modes for Advanced PWR in a Pressurized Containment 5-14 5-3 Mitigation System Cost for PWR-Large Dry Pressurized Containment 5-15 5-4 Dominant Failure Modes for Advanced PWR in a Non-Pressurized Containment 5-19 5-5 Mitigation System Cost for PWR-Large, Dry Non-Pressurized Containment 5-22

~

A-1 Dedi7ated Surface Sited Heat Removal System A-5 A-2 Dedicated Underground Heat Removal System A-5 A-3 External Drywell Spray System A-7 A-4 Internal Drywell Spray System A,- 8 xii i

?

TABLES (CONCLUDED) l Table Page A-5 Core Distribution on Diaphragm Floor A-10 A-6 Central Basemat Core Retention System A-12 A-7 Dry Crucible Installation Costs A-14 A-8 Dry Crucible Installation Costs A-14 A-9 Clean Steam Venting to Stack A-15 l

A-10 Venting and Filtering System A-16 A-11 Combination Venting System A-17 l A-12 Large Hydrogen Recombiner A-18 A-13 Large Vacuum Breaker Valve A-19 I A-14 Large Chilled Filter System A-21 A-15 Large Chilled Filter System A-22 I

I I

I I

xiii i

. . l l

t l l 1 i

.i O

I I 4

i l

i I

^

i l

1 e

I 4

l i

l h l i

I . .

I CHAPTER 1. INTRODUCTION AND BACKGROUND The task of this report is to examine the costs and benefits of proposed systems for mitigating the consequences to the public of severe accidents in nuclear power plants. Such consequences derive almost entirely from escape of radio-active materials. In the " Defense-in-Depth" concept, which is the basis of the safety philosophy of all U.S. nuclear plants, the fuel cladding is the primary barrier or enclo-sure for the radioactive materials, while the reactor pres-sure vessel and the associated primary piping system form the second barrier. These elements are vital parts of the normal operation of the power plant, and accident pre-vention measures deal primarily with their preservation.

l The third barrier, the containment structure surrounding the entire reactor system, is almost independent of the operating portions of the plant, and forms a backup protection for the public in case of severe damage to the first two barriers.

The containment is thus a mitigation system. The containment of present reactors, however, has not been designed to resist extremely severe accidents in which the core melts and escapes from the reactor vessel: such accidents were deemed too rare to justify the cost of mitigating them. The present group of reports is part cf the Nuclear xegulatory Commis-l sicn's (NRC) reexamination of this point in the light of the Tnree-Mile Island (TMI) event.

1.1 OBJECTIVES OF THE SEVERE ACCIDENT MITIGATION PROGRAM l

The NRC published the " Advanced Notice of Rulemaking" in 3

the Federal Recister on 2 October 1980 to serve notice that a 8 long-term effort was to be initiated that would establish policy, goals, and requirements for core-melt accidents more severe than the present Design Basis Accident (DBA). It was l Very quickly discovered that the technological basis for decision making was weak. This led to the NRC augmenting its Severe Accident Research Program (SARP). The nuclear indus-try responded to the advance notice of a possible new design basis by organizing and implementing the Industry Degraded Core Rulemaking (IDCOR) program.

The first step in the decisionmaking process is to decide how safe nuclear power plants are under these new condi-tions, and whether this is safe enough. This judgment will be made by judicious application of risk assessment, i

1-1 i I

l l

I

l l analysis, and engineering judgment. Once this is done, the question of whether or not something should be done can be l addressed. If risk reduction is deemed necessary, then the I question of whether it is to be accomplished by prevention, mitigation, or administrative procedures will be undertaken.

The various research programs sponsored by the NRC (over 50 of these are relevant to the subject) and IDCOR are intended to show how best to achieve risk reduction if it is desired.

Separate efforts address the three methods of risk reduction.

In these reports, the focus is on mitigation.

A number of complex technical issues must be addressed before decision making on regulatory issues can be made without high uncertainty. At present, substantial uncertainties exist in both the phenomenology of a severe accident and in the public risk that it produces. Well-engineered mitigation devices can remove much of the phenomenological uncertainty, but the benefit of doing so depends on the risks inherent in the i unmodified system. The net result is certainty in the miti- l gation function but large uncertainty in the actual risk reduction. Mitigation of core-melt consequences by design changes or by the addition of various devices has been and is j the subject of a number of studies. A discussion of some of i the studies sponsored by the NRC Division of Reactor Safety Research is given by Castle, et al., (1984).

1.2 SCOPE AND DEFINITIONS 1

For the purposes of this study, the NRC has defined severe accident mitigation as those actions, devices, or systems intended to reduce, ameliorate, or remove the consequences to the public of a severe accident wherein the core is degraded or melted. Accident prevention, on the other hand, will be the term applied to those activities related to controlling the condition of the reactor and fuel before the core is damaged or melted. As with any such definition, ambiguous cases will appear. The TMI accident, for example, is a borderline case. Changes in the reactor equipment, instru-ments, and valves to prevent a repetition of such an accident would be prevention, but correcting the means by which small amounts of radioactive gas were vented to the atmosphere could be called mitigation. Failute to scram following an anticipated transient (Anticipated Transient without Scram (ATWS) event) represents another ambiguous case for some reactor containments, which in a BWR might cause overpressure failure of the containment long before the core is damaged.

In this particular case, the distinction has been resolved by allocating to prevention any measures to avert the initiation of the ATWS, and to mitication the steps taken to reduce the consequences of a core melt or overpressurization that might 1-2 l

i P

i follow later in.'the same accident. As an-examole, such mitigation might take the form of a self-closing pressure relief. valve in the containment of a BWR, capable of venting off enough noncontaminated steam during the ATWS event to.

prevent rupture. This valve would reclose before any core-melt occurred. Hence, any subsequent release of radioactive materials-would be contained.

As noted above, mitigation can include actions or proce-dures, devices, or combinations of components to form miti- l gation systems. In practice, a reactor meltdown accident would seldom pose only a single threat to the barriers I protecting the public or the environment. There is the risk of-overpressure failure of the containment due to steam, hydrogen formation, or concrete attack. There is f the need to prevent escape of the core materials by melting through the bottom of.the containment. ~ Finally, there is i the necessity of. discharging outside the containment the.

heat generated by the contained core without permitting the escape of radioactive materials. Unless the mitigation system covers all the threats in a particular ccenario, the o

amount of risk reduction will be less.than optimum, and may l not be cost effective. The actual risk averted will depend upon the dominant modes of containment failure, the plant arrangement, and the population density and meteorology of I the site; but in general, mitigation of a severe core-melt

' accident will require consideration of overpressure control, containment heat removal, and core debris retention.

For.the purposes of this report then, a containment mitiga-tion system will denote an interactive combination of I I

devices, subsystems, and components capable of dealing with the various modes of containment failure for a particular

) plant. The components of such a system will vary according to the threats deemed worthy of a mitigation by value/ impact analysis Operator action can be a part of such a system, l and it must be emphasized'that in this report description of

a. piece of hardware suitable to perform a required mitigation function does not preclude other means of performing the same function. Operator action or modification of existing equip-ment can possibly perform as well as dedicated hardware in some cases and at lower cost.

Planned intervention by plant personnel has recently been recognized to be an important possibility for accomplishing accident mitigation. Current activities include study of computerized operator adds for early diagnosis, improved instrumentation and display, alarm prioritization, extension of plant emergency procedures, and others. Operator training for such extreme conditions does not exist at the present 1-3 I

t t

}

y h ,

time, and only-rudimentary' operating procedures exist. Even without planned' intervention,.it.is difficult.to assess what the human error' factor might be. . Present-PRAs'are based.on.

judgment rather than on' solidly based actuarial data. There l is no doubt'that a well-trained-operator who. understands the physical processes guiding the. course'of the accident could intervene successfully in some cases. Only after a great deal more study will we know how he should be. trained and

- specifically what his procedures should be. It may be that the operator will be much more useful as part of a prevention scheme than a mitigation scheme. ,

L J

1.3 BACKGROUND

~

Earlier reports in this project'(Castle, 1984; Castle,L1984b) sum:garized the state .of the art of' containment construction; noted the' probable modes of failure in severe accidents: and outlined 1the incentives, possibilities and technology for upgrading present. containments so that they could better g resist. severe accidents. The conclusions were that the five l' major types of~U.S. containment systems were all vulnerable to. severe core-melt accidents, but that relatively minor modifications could improve resistance'substantially. The l type, design.and cost of such changes were shown to be dif-ferent for each containment type. Also, the benefits in terms of risk averted to public health varied with and were sensitive to the reactor site. Hence, a more detailed inves-tigation of particular cases was needed to ensure that a verifiable basis existed for a possible change in regulatory i policy.

This report presents the findings of such an investigation.

Three specific reactor / containment designs were studied to a depth considerably greater than was possible in the broad-gauged first report. Suitable mitigation systems were chosen and fitted by design and calculation to the needs of the particular containment type and its specific failure modes under severe accident conditions. Cost and benefit analyses were then performed te illuminate the type and degree of

. incentive available to implement a program for mitigating that type of reactor / containment.

l In the course of this program two unexpected developments occurred that affected the presentation of the results. The ,

first of these was the discovery that mitigation systems in l at least some' containments appeared to be made up of separ-  !

able elements, as far as cost was concerned. The initial assumption had been that since a complete mitigation system must be composed of highly interactive components carefully j designed to meet a specific accident end-state, then the l

l 1-4 f

r

cost of such a system should have to be determined separ-ately for each configuration.

This assumption meant that, because of limited resources, alternative designs must be narrowed down to a single f arrangement for a given plant on the basis of very pre-11minary considerations, and then a single cost determination made. Ecwever, the first preliminary study showed that it was possible to design and cost the system components separately, within a reasonable degree of accuracy, and then combine these components in various ways to meet different mitigation objectives.

This. approach permitted the creation of a " menu" of mitiga-tion components or subsystems, from which selections could be

! chosen to best fit the failure modes to the areas of dominant risk and hence maximum benefit. Although this method required more design and cost effort than for a single inte-grated system, the increased utility of the results more than justified it. The cost was a small fraction of that required to make design and cost determinations for an equivalent list of integrated designs.

}

The second discovery concerned the use of PRAs as the source of benefit determination in the first preliminary study in j

this series. The authors were familiar with a number of individual plant PRAs in some detail, and were aware that not all such documents were of the same quality. However, it had been assumed that these differences were minor and could be l allowed for in comparing different reactor plants. This turned out not to be the case, and comparative risk assess-ments between plan" types had to be foregone in the earlier

! r.tady (Castle, 1984).

For the Limerick study (Chapter 3), two specific PRAs and their review by BNL are used to form an assessment of bene-fits in terms of the risk averted to public health. Various sensitivity studies involving fission-product source term, site variability and accident phenomena were considered.

These costs and benefits were used in conjunction with the NRC's proposed S2OOO/ man-rem averted algorithm to assess the

" menu" of mitigation schemes mentioned above. The study in Chapter 5 involved a plant (WSP-90) for which the PRA was not yet available, so that the impact /value analysis for thic plant could not be completed in the same detail as at Limerick.

1.4 SELECTION OF CONTAINMENTS TO BE CONSIDERED In the following chapters three different plant types and their containments are considered:

1-5 t

I

1 at i

1.- A BWR with a Mark II containment, represented by I the Limerick Generating Station.

l

.I

2. A-BWR standard plant design with Mark III <

containment, known as GESSAR II. <

l

3. .n advanced large dry' containment.PWR, known as A

WSP-90.

These containments were chosen, in part, because of current and'pending regulatory actions concerning plants of these ,

types. .There are ten-Mark II BWRs' currently operating o" i under construction in the United States; and several of i them (e.g. Limerick, Shoreham) will be seeking operating. .

licenses in'the near future. The GESSAR-II Mark'III '

containment-is to be the.GE' standard plant of the future. ,

~

i In this case,'the NRC staff-has-prepared a Safety Evaluation 4 Report (SER), which- is required should a utility request- a - i construction permit.

l Last, the WSP-90 was chosen because it is intended to be the l Westinghouse standard plant of the future. In this case, ,

there is no PRA as yet, but ultimately the NRC staff will  !

prepare an SER. During the preparation of this report, certain design features of the WSP-90 containment were modified, and are still not.yet final, so that some of the details and illustrations presented here may not represent the ultimate arrangement. The study presented here can be considered preliminary in nature, but the design principles and. mitigation functions used are universal, and can be used as. guidance for further work.

1 l

1-6 l I f i

)

1 .

CHAPTER 2. TECHNICAL ASSESSMENT METHODS DEVELOPED FOR THIS STUDY 2.1 DETERMINING DOMINANT FAILURE. MODES AND MITIGATION REQUIREMENTS A degraded core or core-melt accident in a nuclear power plant carries a risk to the health and safety of the public because such accidents, although rare, initiate processes having a high probability of causing containment failure, with consequent escape of radioactive materials. To elimin-

.l-ate such consequences, an essential requirement is that the-final stage--containment. failure--does not occur. However, for any given type of containment, more than one mode of g failure'is possible, depending upon the particular accident i end-state that preceded it. The accident end-state condition is the name given to the situation within the reactor and the rest of the plant when accident _ prevention measures have all failed, damage to the core has occurred, and damage to the reactor vessel and ultimately to the containment is underway.

Each separate containment failure mode may represent the culmination of any of several accident end-states, and each end-state, in turn, represents the result of a much larger set of detailed accident sequences. The probability of any one particular accident sequence occurring may be very small, but the aggregate risk represented by a given failure mode may be substantial. Though only one accident sequence can occur in a real case, it is impossible to predict which one.

7 educe the risk, one must assume, in effect, that'all pc_sible sequences have occurred simu3taneously. Similarly, all the end-states leading to significant failure modes must be accounted for in the mitigaticn scheme. Thus, the governing requirements are established not by individual sets of failures, but by the envelope of them all.

. For example, if it is found that loss of all site electric power is a dominant contributor to a given containment

- failure mode, then none of the mitigation devices used should depend upon site power. On the other hand, if nene of the dominant sequences includes a failure to isolate the containment, then it is appropriate to assume that penetra-tions are properly closed in all mitigation conditions. The net result is that the requirements for mitigation are cumulative. Thus, for any given type of containment, the functions that any proposed mitigation design must fulfill 2-1 l

l

will be derived from a combination of the significant fail-ure modes.

Although this might seem an overly conservative approach, it becomes an essential one if mitigation efforts are to be credited with a real reduction in risk. In practice, it is found that many of the requirements are redundant (being the same for several accident sequences) and that certain miti-gation steps can fulfill several requirements at once. In addition, the incremental costs for overall mitigation tend to be lessened through use of common facilities and existing structures. The final result is that, after analysis of the .

dominant accident sequences and the end-state conditions thereby created for a particular type of containment, one can compile a set of " cumulative functional mitigation requirements." Such a list would comprise the necessary functions that a mitigation system must perform and the conditions under which it must operate.

It is important to distinguish between the " function" and the device that performs it. For example, to reduce the risk of containment overpressure failure from gases gener-ated by concrete attack, both a filtered vent that relieves pressure and a core catcher that prevents concrete attack would each perform the same function in a different way.

The choice between such options would depend upon other functions that each device could perform: which one fitted best into an overall, minimum-cost system; and which one averted a larger risk. In the following chapters, the opportunities for accident mitigation are given for each containment type in terms of these functional requirements.

From these requirements is derived a set of possible hard-ware choices that could fulfill them. This procedure is followed for each of the three types of containments.

Within the context of this report, the dominance of the various containment failure modes is established by their relative risk for each consequence of interest. In come cases, the dominance changes when a different measure of risk is used. For example, slow overpressurization failures

" of containment are the dominant contributor to latent fatal-ities or population dose, but steam explosions are the dontinant contributor to acute (or early) fatalities. In the following chapters the risks, as measured by various conse-quences, are determined for each containment failure mode using existing PRAs. For those failure modes that contri-bute more than 1 percent of the risk, a program of mitiga-tion is developed. A cost breakdown for each component is made to permit an overall cost / benefit estimate for each program.

I 2-2 l

I

The implementation of any program of mitigation for a par-ticular plant would then be determined by a number of regu-latory factors. One factor would be a value/ impact analysis.

2.2 VALUE/ IMPACT ANALYSIS OF MITIGATION SYSTEMS Cost / benefit analysis is a general approach used by decision makers as an aid in determining whether proposed actions should be undertaken. As the name implies, a key step in such an approach is the quantitative approximation of the benefits and costs of the proposed action. In the context of mitigating severe accidents by improvements to contain-ment, benefits are usually quantified in terms of reduced public risk and costs in terms of installation hardware and maintenance.

Within the framework of the NRC's published guidelines for performing the regulatory analyses required for a broad range of proposed actions (NRC, 1983), such cost / benefit evaluations are called value/ impact assessments. Values in this context represent a range of potential public benefits which may or may not be easily quantifiable. These values may range from attributable reductions in public health risks to increased public confidence in the regulatory pro-cess itself. Impacts represent the costs of such actions.

They may range from dollar costs of equipment and their installation to. occupational radiation exposures incurred in this installation as well as in its maintenance. In retro-l fitting an operating plant, the extra downtime required for modification can be an important part of the cost. Such cost is not included in the estimates given here.

In this section, the methods are described that will be used for the preliminary assessment of the mitigation systems developed in this report. This assessment cannot be final l for the following reasons:

1. The values employed are based on the review of l existing PRAs. These values, in terms of risk reduction, are therefore based on Reactor Safety Study (RSS) (NRC, 1975) source terms and method-ology, and are subject to change.

2. Tasks 4 and 5 of the project have as their focus the development of Value/ Impact measures (Task

4) for evaluating of mitigation systems, and the development of approaches for regulatory actions (Task 5).

2-3 1

4, . e 4

1 i

l l

2.2.1 Overview of Value/ Impact Methodoloav i

The Value/ Impact Assessment Handbook developed by Heaberlin (1983) has been used as a guide. .The objective of the analy-sis is'to identify and estimate the relevant values and l impacts likely to. result from a proposed action. The hand-l book defines the quantities that make up the values and l

[ impacts (i.e., the effects of a proposed action) as attributes, and gives the following list:

1. Public health. '

2. Occupational exposure (accidental). q i

-3. Occupational exposure (routine).

I

4. Off-site property costs.

5. On-site-property costs.

6. Regulatory. efficiency, i

7. Improvements in knetledge.

8. Industry implementation.

9. Industry operation.

10. NRC development.

11. NRC implementation.  ;

12. NRC operation.

The handbook cautions, however, that "in any particular app 11-cation, the analyst should carefully consider (1) whether these attributes are complete, i.e., whether they encompass all of the important consequences of the proposed actions; and (2) whether they are all necessary or appropriate for the particular action under consideration. The analyst should then supplement or modify the attributes as appropriate."

o rocedures are given for quantifying these attributes as l well as two methods for summarizing and displaying the I results of the attribute quantification: a ratio method and 1 a net-benefit method. These methods are briefly summarized as follows (Heaberlin, 1983):

i 2-4

)

I

h

1. Ratio Method. The total net public health value of the proposed action expressed in terms of the

. expected reduction in public exposure is divided by the total costs-(NRC, industry, etc.) of'the action. The units of the ratio are man-rem (averted)-per unit of cost. Other factors and special considerations are considered separately.

2. Net-Benefit Method. To the extent'possible, all-attributes are quantified in monetary terms, and the dollar values are'added together (with the appropriate algebraic signs). The result is the net benefit ~, in monetary terms, in units.of dol-l lars. Other. factors and special considerations are considered separately. When the net benefit 1

method is used, the factors used to convert non-monetary attributes to. dollars should be explic-itly stated.

I Neither of the numerical values obtained by the two methods is intended to be used alone as the sole basis for regula-tory decisions. Each provides an input to the decision maker on the " cost-effectiveness" of a proposed action. These, along with other considerations such as uncertainty, unquan-tifiable values and impacts, and other " engineering" and

" societal"-judgments, should be used in arriving at a regu-latory decision.

2.2.2 Discussion of Values The values that are usually estimated in a quantitative manner are reductions in risk (risk averted), which involve l

radiation exposure to humans, and on-site and off-site costs I due to accidents. The risk averted in terms of accidental public radiation exposure and off-site costs (e.g. decon-tamination) is usually determined probabilistically, i.e.,

l the potential consequence is multiplied by the estimated frequency of occurrence, and is found in various PRAs.

j Public exposure is usually measured in man-rem (population dose) and the hazards associated with the exposure in acute (early) fatalities, latent (cancer) fatalities, allness and thyroid cancer. In the net-benefit approach, these quanti-ties should be converted to monetary equivalents. This issue is somewhet controversial, and at present, there is no definitive resolution.

The alternative approach is to refrain from defining a -

monetary value for these quantities and utilize the value/

impact ratio as described above. It has been customary to 2-5

__ _ _ _ _ _ _ _ _ _ _ _ _ ______.__.____.____.___________J

use population dose and express the ratio as man-rem per million dollars, or if it is to be compared to the NRC benefit-cost guideline (NRC, 1983) of $1000 per man-rem averted, as dollars per man-rem (the inverse). Other alter-natives are to convert the population dose to acute and latent fatalities averted and express the ratio of dollars /

life saved to some appropriate criterion. However, none of these has been adopted for nuclear power.

The inclusion of on-site accident costs averted in a value/ impact assessment is somewhat controversial. Some argue that on-site accident costs do not influence the public health and safety, and therefore should not be part of regulatory decision making. Others argue that the on-site costs averted do affect the public, both directly (clean-up costs at TMI may be paid either by the federal government or by the consumer rather than the investor) and indirectly (replacement power costs are usually higher and affect prices to the consumar).

The values that are usually difficult to estimate in a quantitative way or are unquantifiable are such things as improvements in knowledge, increased public confidence, and a more efficient regulatory framework. Such unquantifiable values may be as important to the decision making process as the quantifiable ones, and as stated above, these shculd be considered as special considerations and displayed separately.

2.2.3 Discussion of Impacts The impacts that are usually estimated in a quantitative manner are the monetary cost (equipment, labor, mainten-ance) and any radiation exposures (occupational) associated with a proposed action. It is suggested in the handbook that monetary costs be discounted to the present value. For construction costs including labor, equipment, etc., this poses no problem because they represent present value.

Maintenance and/or other costs that may accrue at later dates pose a problem because there is no agreed upon dis-

, count rate. Values of 5 to 10 percent have, however, been suggested in the value/ impact handbook.

Occupational exposures should be given in terms of annual expected effect, and can be integrated over the plant life-time to obtain the total effect. The issue of the monetary value of occupational radioactive exposures is the same as is the case for accidental exposures, i

2-6 f

.s: .

i' It is important to reiterate that as part of the NRC Policy Statement on Safety Goals (NRC, 1983), a guideline of $1000 per man-rem averted was suggested on a trial basis for all types of exposures averted. However, its use with respect

! ;to reduction in occupational exposures has been in practice.

i for some-time.

2.2.4 Use of Value/ Impact for Scopina Studies In addition to a discussion and comparison of the net bene-fit and ratio. methods, the value/ impact handbook proposes that a scoping study be conducted before any detailed calcu-lations are performed. There are several reasons given:

1. The scoping study may prot'ide sufficient infer-mation to reject some of the alternatives.

2. The resources expended in the value/ impact study should be commensurate with the value of the

~l information to be obtained.

3. The study offers an opportunity to develop and communicate to all parties a clear understanding of the issues involved, as well as the nature of the proposed action and potential alternative actions.

The value/ impact assessment developed in this report shoulf be considered as a scoping study precisely for the reasons l presented above. The potential values will be based on public health risk averted, which in turn is based on RSS source terms and methodology. Hence, it is subject to le.*ge l uncertainty and change. Moreover, data obtained from cm.st-I ing PRAs will be used, and represents only public health effects. Risk in terms of on- and off-site costs are not easily obtainable from the PRAs although they have been l considered in a separate report by Strip (1982).

Given the nature of the information available, and the l difficulties in monetizing public health risks, the ratio j method employing dollars per man-rem averted will be employed in this study. As new source term information becomes available and as Tasks 4 and 6 develop, both the net-benefit and the ratio methods will be utilized,. .

Two other points should be noted. First, mitigation in a  !

strict sense assumes that core melt has already taken place.

In this regard, a mitigation system will usually not avert on-site costs unless it reduces core-melt frequency. An exception might be recovery or clean-up costs. For example,

)

2-7 l

f

l I

a system for directional control of the molten core, once it l 1 eaves the vessel such as proposed for Shoreham (Erdmann, 1983) J or a subbasemat core retention device such as proposed by i Hammond (1982) might significantly enhance the chances for prompt plant recovery, or lower the clean-up costs.

Second, the use of the 51000 per man-rem averted as an algorithm for monetization of population dose is thought to be conservative. In this sense, the criterion can be 4 thought of as a surrogate for other attributes, within the l context of the scoping study. Additional considerations ,

such as the nonquantifiable benefits discussed above are presented separately.

l 2.2.5 Value/ Impact Assessment l I

We are now in a position to state the ground rules for the value/ impact scoping study presented herein. As discussed above, the ratio method employing dollars per man-rem )

averted will be used for comparison with the S1000 per man- j rem averted criterion. Although the NRC proposes that only 4 man-rem averted out to a 50 mi radius be used, we will exhibit this ratio for both 50 mi and 500 mi because man-rem dose is being used as a surrogate for other public health j effects. j l

Since all the plants represented in the study are new, a 40- l year life will be used to determine total man-rem averted i over plant lifetime. Last, perfect mitigation will be I assumed in this scoping study. The implication here is that I if two approaches are proposed for elimination of a partic-ular threat to containment, quantitatively they are assumed l to have the same benefit. In order to improve on this crude j assumption, and to differentiate between options, both quan- '

titative (reliability) and qualitative (e.g., sabotage) considerations beyond the scope of this project would have ,

to be taken into account. An exception is the analysis of a j conventional filtered vented containment system (FVCS), I where the residual risk due to noble gas emission is )

accounted for by a risk reduction factor (Gazzillo and l Kastenberg, 1984). l l

2.3 DESIGN ASSUMPTIONS AND GROUND RULES l In undertaking the design of hardware components and systems l l intended to fulfill the mitigation requirements selected, '

j

certain assumptions and rules were adopted as the guiding l philosophy. These can be summarized as follows.

t l 2-8 1

I

I

1. Whenever the accident end-state conditions are g

uncertain, and the technical community is not in l

agreement as to the' outcome, the anomalous situation will be circumvented by a design that either svoids or minimizes these uncertainties to produce a known end-state. In some cases this means that the mitigation system must in effect intervene in the course of the accident to force a preferred result. In other cases, it would mean only that the system be capable of .

handling any of the possible outcomes.

2. To the maximum extent possible, the mitigation system should respond passively to the accident situation. Where a completely passive system is impossible or unreasonably costly, a quasi-passive system requiring no personal attention would be used. If motive power, as for driving pumps, is required, a dedicated source used for no other purpose would be used, preferably installed in duplicate and separated.

l 3. An uncontrolled overpressure failure of the con-tainment structure or its penetrations is deemed unacceptable. If the dominant accidents include the risk of such overpressure, the mitigation design would include a means of venting in a controlled, reversible manner.

4. It is considered essential to maintain control of the pathway, location and state of the debris from a melted core at all times. The escape of a melted core out of the containment or to random distribution within it is considered unacceptable, l 5. Mitigation equipment should be of the highest industrial quality and meet general design and seismic criteria, but need not have the documen-tation required of safety grade installations.

E. The mitigation equipment and its operation

, should present no interference to the normal functioning of the plant and its safety systems.

2.4 COSTING METHODS A major simplification in the presentation of results for this project was accomplished when it was found possible to consider separately the costs of the individual components 2-9 l . ..

V

~

used in the proposed mitigation. options.- These component

. costs can then.be ascembled in any combination to provide a cost estimate'for' desired system. options such.as those describedLin Section 3.6., This-assumed separability of:

costs cannot be1 complete for all cases,-but our study-shows that the uncertainty thus introduced is of lesser. order than that produced by local site effects.and estimating inac-curacy. .For policy-making purposes this procedure is far more effective than making separate designs and cost esti-mates for-each combination. -

For each component or system, the installed' costs'are esti- j mated for~three different plant. status situations: "A" costs are based on a new plant not yet completely designed and with no site' work started: "B" costs are based on a plant about midway in the construction process, i.e., miti-gation additions and modifications can be made with no radioactivity, minor rework, and no startup delay; and "C" where costs are based on going into a plant that has been operational for some *.ime but is currently down for refueling. However, it should be noted that no plant down-time charges are included. In case "C" the costs include radiation protection, draining of equipment, etc. In this j case access to some areas of interest becomes very difficult  ;

in a " hot" plant. While a number of Mark II BWR plants might I be retrofitted.where status "C" conditions would apply there are no existing Advanced BWR or PWR plants; thus only the "A" status cost need be considered. Costs of mitigation equipment applicable only to these new plants has a coat estimate only for this condition.

For consistency in all the alternatives, the cost estimates are based on the following conditions:

Labor Costs. Direct labor cost estimates include fringe benefits, portal time pay where necessary, insurance, incidental overtime and contractor mark-up based on forecast 1984 costs at $340/ man-shift for all shifts. One man-day is thus $1020.

~

Escalation. Inflation during construction is covered by a 4 percent addition to all labor and material costs. This is more than 15 percent annually in most of'our cases since these contracts are short term.

Desian Control and Field Administration. This item covers the control of drawings, changes, inspection, field engi-neering and administration of the projects in the field d ming construction. These charges are covered by a 12 per-ca .t: addition to all direct labor and material costs. In the 2-10 i

f f .

i cost breakdown shown this overhead component is included with inflation escalation and listed as " Field Ov6rhead--

16 Percent."

Purel.asino Costs. All direct purchases of equipment have a 15 percent handling charge added to cover alternatives, variations, shipping, tax, etc.

Supervision. For simplicity in costing, all on-the-job supervision is included at one average value regardless of the class of man involved. The S6000/ man-month figure used includes their own administrative overhead, employee bene-fits and vehicles where necessary. The more complicated installations--such as the underground work--will require one general superintendent, one project manager, one project engineer, an office engineer, a cost engineer, a field design engineer, a safety engineer, an office manager, a purchasing agent and an accounts manager. This team costs S54,000/ month and this amount is used for all supervision.

In many cases the supervision on the job will handle several different component installations under way at the same time .

so each mitigation component may have only a portion of the t total supervision charged to it. This is referred to as Supervision--chargeable in the tabulations.

Contingencies. On preliminary cost estimates such as these.

l where detailed drawings do not exist and the course of action is not specific, a contingency factor must be included. Twenty-five percent on the overall project is used on all components.

Profits. Profit is not included as a line item in this l estimate since it is not known how the work will be done or who will do it.

l Time Estimates. All work schedules are based on two shifts

' of 8 hr and one prendum shift of 7 hr six days / week.

Insurance. No liability insurance costs are included since I it is not known who will be doing the work. Many utilities have a general coverage for this.

Quality Assurance and Quality Control. At the request of the NRC project management our cost estimates have not included a factor for quality assurance or quality control but do provide the highest quality commercial grade equip-l ment and construction available. On components and construction materials where QA/QC formal paperwork is required the costs may be increased substantially.

2-11 1

' l L______________ . _ _ _ _

, .O Seismic-Ouality. The consensus of most knowledgeable engineers and construction men is that building a plant to a

. higher seismic rating does not markedly increase the'fabri-cation.and installation cost but it does add to the

engineering charges (e.g., vibrational resonance analysis of complex. structures).

Typical Specific Costs. Unless conditions indicate they should be. higher, the following specific costs are used in the calculations. The costs are "in-place" and include direct labor overhead only.

e Reinforced concrete ~ $600/yd 2

-e Gunite with steel mesh $6.00/ft -in e Structural steel--at plant $0.50/lb

--in place $1.50/lb e Machined components $4.00/lb e . Rotating machj nttry by component (mfgr. quote)

+ overhead e Electrical and control by component (mfgr. quote)

+ overhead Real Costs. The cost estimates of the mitigation components given in Appendix A cannot possibly represent the full cost at a specific plant with its own individual cost accounting methods for a number of reasons:

e All construction costs vary markedly with the time when work is done, the location where it is done and especially the construction business activity at the time the construction must be done. These imponderables cannot be forecast, so the costs shown are an effort to hit some mid-continent mean--circa 1984.

e Inflation is not included. These costs were generated in early 1984 and must be brought up to date if used.

o QA/QC costs have specifically been excluded since mitigation may not require " safety-grad 2" components.

e Most mitigation components do not require any extra land at the site. Although some cost accounting methods distribute land costs to each component, it has not been included here.

2-12 i

hy.. . ..;

L L

L '

o- Profit has' not been incidded but may be necessary depending upon'how the mitigation construction is handled.

e. The basis used is consistent and' valuable for-comparisons. 'However, some of the assumptions ~

may not apply in afspecific case. If all of .,

these factors are combined, an' upward adjustment of 1.5 to 2.5 times,our basic cost estimate may be required to arrive at a realistic figure for a particular plant.

- Uparadina Present Ecuipment. Almost.every plant incor-

. porates equipment. intended to perform the same function as the mitigation component called out-in this report. When-ever this standard equipment can be upgraded to meet the same mitigation' requirements it can and should be.used. It must have adequate capacity; have its own dedicated power source and be immune to the failures that caused the severe l accident in the first place. Where this can be done, the costs given here'would be reduced.

. 2.5~ ' LIMITATIONS TO THIS STUDY The' objective of the program reported here:is to assist the NRC in exploring and developing possible extensions to regu-latory policy with respect.to the mitigation of severe accidents. The' design and cost of work performed is intended to demonstrate that real containments can be pro-l - tected by specific hardware systems against the1most likely

- threats and accident conditions. The: designs are' based on

. carefully worked out structural and material assessments using standard engineering design procedures, and the costing is intended to 5e as realistic as possible, with actual equipment quotations used whenever obtainable.

I.

Although'we believe these objectives are fulfilled, the '

results are necessarily subject to certain limitations:

1- (1) The mitigation equipment described is untried for the use intended, since no 7e has ever " mitigated" a real accident. Hoktver, in tany cases the actual function of the individual components i, in itself quite standard and well proven. (2) The designs presented have been carried to a level termed " detailed conceptual design," but are not intended to represent actual construction designs, fitted to an existing building. (3) The limitations on the cost estimates were discussed above. It is expected that better designs could be achieved, and lower costs realized, in any real application, but such detail would not suit the present 2-13 M- .=.--_-_m_:__._._.____.____-_-__.-_._:.m- - - _ _ _ .

o .. e.

I purpose of the program. (4) The.most important limitation of the work, we believe, lies in the assignment of benefits to mitigation. The shortcomings of available PRAs have been mentioned, and the trend noted toward discovery of higher potential risks with continued assessments. The uncertain-ties from these sources, plus the as-yet unquantified bene-fits of mitigation that do not involve public risk, are together a significant weak spot in the overall assessment of benefit.

2.6 UNCERTAINTIES .

1 As will be discussed in later reports of this series, the .

handling of uncertainty will play a major role in decision 1 making. There is little uncertainty that mitigation  !

features can be designed, built and installed, that they will function as designed, and that their costs (impacts) can be determined. There is also a reasonable certainty that in the absence of mitigation systems most containments will fail in one of several modes in case of a severe core melt. What is not certain is that the accident sequences will have the frequencies predicted, that the containments will have the predicted distribution of failure probabili- ,

ties, and that the consequences of such failures will have j the human effects estimated. Hence, the function and cost of mitigation is relatively certain, while the quantitat.ue benefits remain uncertain. Other kinds of benefits, including the reduction of uncertainty itself, may provide I the necessary value to offset costs of mitigation in the regulatory process.

2.7 A NEW MITIGATION CONCEPT In the course of this study, an alternt.tive mitigation concept was developed that holds the pctential of lower costs, simpler construction, and reduced uncertainty. In j this alternative the containment is not required to hold high pressures during an accident, but is directly vented at all times to atmosphere through a very large capacity abso-

. Aute filter system. A reision of this concept is applied to j a BWR in Section 4.5, and to PWR's in Section 5.5. i i

l l

l l

2-14 i

1 ,

W.

)

' CHAPTER 3.' . MITIGATION DESIGNS FOR MARK II PLANTS-l

3.1 DESCRIPTION

OF THE MARK II-CONTAINMENT.

.: Although different Mark II plants vary slightly in detail, forLthis report the description is based on-the-Philadelphia Electric Company plant at Limerick,fPA. The typical con-tainment cross section is shown in. Figure 3-1, and the pertinent dimensions of this. containment in Figure 3-2.

Much of this data appears in the Limerick PRA (Philadelphia Electric.Co., 1981).

Relevant-information.about the plant and its. operating con-ditions are included'for ready reference in Table 3-1.

Figure 3-3 shows the approximate decay heat release rate in a core-melt accident and its accumulation with time when cooling has failed (Bowman. 1973). Figure 3-4 presents the

-l heat energy releases in the containment corresponding'to l various events or situations. Some of these events can be additive. More detailed description of the Mark II system can be found in the first report of this series by Castle, et al., ( 19 8 4 ) ..

l Safety Features Currentiv In-Place Certain safety features of the BWR. reactor have an impact upon mitigation requirements. These are listed below.

Emergency Core Cooling Systems (ECCS) e Low Pressure Core Spray System (LPCS) two loops, I]

6350 gpm per pump at 105 psi.

e High-Pressure Core Spray System (HPCS).

4 e High-Pressure Coolant Injection System (HPIC) two 5600-gpm pumps with turbine drives. l e Automatic Depressurization Valves (ADS) e Low-Pressure Coolart Injection System (LPCI) a four-loop, four-pump system at 10,000 gpm per pump at 20 psi. '

e The "ATWS 3A Fix."

3*

l l

>.* e e

?

l 4

I I

j I  !

i l

I ,

I I i i ' nM i e a j

~

l

! I i

, e a:m:i- -:: pas i

ass =s , ==

I .

e lt \ -

c nni pedestal g

/ ....,

l- l

.;g i

..,s

.g I'

-4 Y

g Diaohr.gm ficor I 8_ _

I _ _t i '.l E *. ;; sj: s .!.- ... u '. "

..E

{.'. .'

... . .c .

, '.; '" DownComers

= i r== ==== =.

[. ' O. _D *:

m i t.- . l . *.4 8

i It r s s n *

• 11 ,*ge#**  % g. g 4 gg p *d

...ns s *s. *s_

as,s. s e so.

. se

r. .. +... ss . ,, rs,..,, ,=,s..

Figure 3-1. Typical Mark II containment. ,

3-2 i

i

l

.,o< .

l i

L 9.

l 37 tt 10 in ID  :

2 El 325 ft 8 in less. same!

u  !*

4; - ,

L .y ,.

l El 312 ft 8 3/4 in '_* . . .

I

• l j -s a , .;.

l  ;:

.'.s .

3, .

muse .: Shield wall

.: s..-

- 5 El 266 ft 3 in

\.:

• 87 ft 9 m. c . '.

}..

El 272 /* - *' .

tt 9 in ..*.'.'

W \. *.; *;.-

. '- ' , " El 263 ft 6-5/8 in 'g.. ' . .

.,. =

El 252 ft J .';

,; 18 ft 5 in ....

10-1/2 in l ., 4 ft 5 in. .- e i J. ' i. .'

1 ...

g .

g 22 -

],

c'/ *?/86 *t 3 in OD ,

20 ft,3 in .- Diaphragm slab -

y of liner plate .

/  ;.

l

'*.I

& Liner R ,.' l '.' r 1  ; El 235 ft

.- 4, ......v.,....

. . ;' 10-7/16 in I El 234 ft 0 m ' . ' ,

r. ll .h. -g' s ..

. Holes 1 lI' l l 3 ft ' .., * "

6 in -

-- I / 204 ft 11 m

' .f .a . '-

/**

J .' l_ ] _ t ,' _ ._ ..' / ,

Suppresson

.*, _ 1 g-- . . .' . chamber l --

- 1,. . d '

columns

. 20 ft 1/2 m ,.

?' - . (3 ft 6 m 001 l .e . - - ..~

.-- . . ?. .

• , 5.' 4 ft *

>?.

.. N 9-1/2 in c.< -  ;* ',

.. . 1 .. 1

. El 181 ft 1.

in ..- T .

/- Holes I .-

u',

. . . ..; . .. . s. . .;. ';;. _....w,, .. ; . 2.o ' ' '?;.....;....;...' ...;

~

. * ; e..

j '

29 ft' 7 in

~

8 ft -e.- - =- 6 ft 2 in

= 88 ft 0 in  :-

1 Figure 3-2. Pertinent dimensions.

3-3

-;;it , ; , 1 , e. ,. e . .

d [

.y J l-J ABLE 3-1.' GFNERAL PLANT INFORMATION

.l.

l GENERAL-Type .

GE BWR/4 3'-

'& Electric output (Mw) .

1140 l, j NSS thornel output (Mw) 3440 l t

Approximate of fIclency (5). 32 ,

REACTOR PRESSURE VESSEL-Pressf e-test (psig) 1250 [

-operating (p sig) 1045 Temperature (CF ) 550

• Steam f low - max (IDAr) 15E+6 Size ID (f t) , 20.92 l-Well thickness (f t) 0.49 He i ght-I ns i de (f t) 70.60

. Bottom heed tnickness (f t) 0.71.

Weight (bottom head) (tons) -I volume (ft3) 103.8 f'

water in RPV  ; 11922 Steem in RPV- 10122 Water in recirculation loops 13200 l Steam in sein lines 1218 l water in f eed lines 1233 Total coolant volume - 26815 Saf ety/ relief actuation (psle) 1120 1143 psie end

. Rollet cepecity3 2.608 I b/t 1 (IbA) 838,900 .

Fuel -l UO2 chorge inventory (tons) 163.9

-volume (f13) 516 '

Zircellor cleeding (tons) 47.7

-volume (f t3) 235 Steinless steel and leonel in core (ID) 26980 Steintess steel below core (Ib) 66750 Steintess steel in control rocs (1b) 32750 Steel in top guides (Ib) 15200 ,

't 1 PRIMARY CONTAINMENT ,

Type Mark ll Construction Reinf orced concrete alth 1/4-i n, steel liner Size ID (f t)

Suppression chamber 88 Drymell bese 86.3 DryweiI top 36.3 -

Heightwetwell (f t) 52.5 j I

+

-d ry w e l l (f t) 87.8 Volume (f t3)

Drywell f ree volume 248,700 metsell f ree volume 149,400 wetwell pool volume 118,600 to 130,825 use 128,000 Yhickness (f t)

Besomet 8.0 Wells 6.1 Diephragm floor 3.5 Pecestal height (f t) 82.0 ID at besenet (f t) 20.5 3-4

..'g 5 y -

l TABLE 3-1 GENERAL PLANT INFORMATION (CONCLUDED)

I i Pressures (psig) l . Design (Internet) 55 i Design rollet 128 Max (below yloid) 120 Design (externet) 5 Leekage-tree volume delly (5) 0.5 Temperatures (OF)

Drywell Initial .135 Drywell mew design 340

.c Wetwell Inittel - 104

. [- Wetwell max design 220 i suppression pool death ~23 l Bottom area (ft )Z .

6082 Depth etter low rupture (f t) 3.4 Rollet possages (volves) 2-2 In.

2-4 in.

1-6 In, 1-18 in.

3-24 In.

1 4

$ECONDARY CONTAIf*4ENT i Type .

Rectangular bulld ..g 1 Construction .

Reinf orced concrete Design pressure (psig)

Internal and external 0.25 Leekage-tree volume delly (T) 5.0 l COOLING POND I: 2 418,000 8 i

Area (f (ft Volume t )3) 3.76 E6 Temperature (DF) 85 Elevation (ft) 239 HEAT REMOVAL SYSTD4 Duel electric motor-driven pumps  !

Capacity (gpm each) 10000 I. Head (psi) 20 Heat orchangers 2  !

Thermal cepecity (Mwt)- 12.3 each

• j l

1 i

i Tnis cepecity is given by the Brookhaven PRA. The General Electric -

Company states that under actual occident conditions, the high temper-

, ature dif ference would greatly increase the cepecity.

l I

1 l

3-5 l f l

jl e ,.

'? I;; ,

I 300 __ j l

14 1- . .,

(#

- 240 Based on: . . . j 3440 MWt BWR '

plant 9 80% power -l' j j

l 200 200 -.

, - Cumulative heet (MWht )

- 165 4

- 160 ,.

I i

118 100 --

-70 l

45 l i

- ,37 33.5 I

_r 29.5

' Decay heat rate (MWt) .24.5

~

l !

0' ' ' ' ' ' ' ' I ' '

0 2 4 6 8 10 l

Time (hours after snutdown) Source: Bowrr.an,1973 Figure 3- 3. Decay heat rate and heat accumulation with time.

3-6 i

i., 7 &

.p h7

-Es gm

1. yo C' *s F

i

, peuana c 7g peuout tou sownsey g -ve6oJpAy at

}. K#jNjk leets suol 09 E c

l a.hkb

!s-24

• i- a sig! sues

[ R S -spqop W "E 3 [ -<04 6{f k ..

i g

^ eJoo $

Teeeoe o f 8

'8 1 --

. g selgruang $

g :sg 5 a stesovoo .-

E gS

,e

'- a N 3- . m l g }*E gu!/ql p*1 + g, dwoosp j j Z oo gi-ggge aa ele;ovoo .,

. l ,

E

• l 0 h -

e e %00t uang .=-

fe popeut tou sewnssy 8

g ue6o;pAu JZ

>g w .a

'E

}

l 0a

. e l

l m

ci

%001 uang JZ

.- E

.  ; e i . as I a i ou l 0 l w %OZ h ~k '

• A @ SM1V W l

I- 0>

l c, e l E l

i i gi .s  !!no;io <-

1 -j v E A;ewyd I

I T

1. m Aeosp m J g cc e l

g ) N y ". N ko - eA! toto! Peg g u

j l =

. I e I I N Je6uryoxe 'g*

• t I IH

,a s

, R R a

=

(ygn) ASJoue lewJoyl i

j l

l 3-7 1

)

l

- _ _ - _ - _ - _ _ - _ _ _ _ __ _ _ . l

/

,. , .- 6 i

' Auxiliary. Systems e Residual Heat Removal-(RHR) systems--Dual pumps at 10000 gpm each at.20 psi, dual' heat exchangers-at 12.19-MWt capacity each.

e Reactor Core Isolation Cooling (RCIC); system--one 625-gpm pump at 1120 psi--turbine drive, o Fuel pool cooling and clean-up system--one heat exchanger of 3.3-MWt capacity.

Hydrogen control system l

e Nitrogen inerting of' containment' atmosphere, e Dual, low-capacity, hydrogen burners (~70 cfm).

l

-3.2 CONTAINMENT FAILURE MODES

~

The consideration'of.a mitigation system for'a particular containment begins-with an understanding of the existing design (as described in'Section.3.1) and its response to g severe accidents. For Mark II containments, the Limerick PRA (Philadelphia Electric Comeany, 1981) and the Shoreham l (Long Island Lighting Company, 1982) PRA, as well as the BNL review (Papazoglou, 1982).of.the Limerick PRA, yield valu-able insight inte the dominant accident sequences leading to 'l containment failure. This section summarizes the major

-accident sequences leading to core-melt, the dominant con-tainment failure modes, and their associated risks. l 3.2.1 Dominant Accident Secuences-The dominant accident sequences leading to core melt for BWR Mark II containments are: (a) transients with scram followed by loss of coolant inventory makeup and (b) ATWS. The dom-inant sequences as given in the Limerick PRA and BNL review i are summarized in Table 3-2 as a function of containment failure class. A key to the symbols is given in Table 3-3.

, Both the original PRA and.the review agree that the dominant initiators are transients due to loss of off-site power (TE) and transients due to loss of feedwater or closure of the main steam isolation valve (TF). The sequences involve loss of feed water (Q), loss of high-pressure injection (U),

failure of timely automatic depressurization (X), and failure of low-pressure injection (V), in various combinations.

I, -

3-8 j L

A>

't.

s i

}- m A

.- e s p ' e e e e e ee ee- e e.e e e. R 3' .

'.o. 'o- .o. 'o 'o 'o ' 'o 'e. .o.

- - - - . 'o.

'o b '.o.. 'o. 'o

'o . 'o.

.u a

,: M M x M M M M M M M M M M M M M M C-l- ,- J. '

, oen o-e ..n e.. e e ro e .- .n .

> J *

. A aJ J J ~d JA'A A A aaaJ B '5

7. . .- -8

. . N-

< w e e o = :F o. e 5 U h h h.

h.

p.-

z . ohxhhhh.hggJ. -

e.-

v ' h h h 0 $. .

Y > > . > - > > > e. - > > > t- 6 m

.J O Q. '

~

$ J aA 2-A-2 ,: i e' .&. .: a. . .A. . .J. A .i. . a. . .,.

6 0 2 ~

i n.

e ni n n -i n. n. - ni . n. a.

e e. e. e. e. S

-W e o o .o. .e o. . o. o. o e-

. .o. .o. o .o- u.

.l .I M M M M M M M M M M M M M M N 5 - n n e e . e e - e r N , . .

- = d A : aaaJ -

i i A aa S w -.

m M.

N N N y N q

.e @

u e. o - - c - y p N o a y .

y z

U hhhhhh'.hhh5hhhh

.ce., -

.- e .- e e 2 0

. z" .

J d A ,*. A 2 J &

o' &

- . J.'d A J -

I-f I<

- e.

e e . .

.. .. ~.e e.

.o. o o o e u

. p. - - --. . o. - .o. >

M M M M M M M M O

't 'u - r . . ., r n . -

6 f

5; -: -* J d~A  ; J ;

M

~

g - 4 -

~

0 5 5

- .5 5 2 p, 3 5 -

- p.-

f 5 5> - -

e m > e.- p. > p v.

0 D.

• T C N z

. 5 e . . . . i o o e e . .

e o i e e

.o. .o . .o. e.

. .o .e. - - - .o. - - .o. .o. u M M M M M M M M M M M M M M N e . ~. e. o. o. .e ,

-e - eo e e. e. .

~e o

. - m *n to e c N e @ e en m o e -

e M e ce

.E. m - .=

3 x - ,

p x x Q > U m = M Q U e

= >O W D M e D > D e <w

. O <m. -* -sQ Mb-3

@ m 3- w o w -

E

. "U m -

I N

$" 0 M. O N 9 M. h* D'b h. I*

eC" O *

- N O N O N @

- N.

. . .T.

3-9 f

,[ -.'. e I

. TABLE 3-3. KEY TO LIMERICK SEQUENCE SWBOLS

- Initiating Events A. . La rge, >4-i nch c laneter, LCCA -

.TT - Turbine trip with by-pass 4

T. E

- Loss of oH-s ite power Tg- - Inadvertent open relief velve Tu - Manual shutdown Tp - MSIV closure / loss of f eedwater System FeI tures AC - Power unavallebility DC - DC power unevelleellity

. D,Us - Inadvertent' operation of ADS or vessel overfill C. - Failure to bring reactor suberitical CE. - Failure of the reactor protection system (electrical)

C, ' - Failure of the reactor protection system (mechanical)

C2 - Fa ilure of both'SLC pumps M - Failure of saf ety/ relief valves to open P - Fallure of saf ety/ relief valves to reclose

.Q - Fa ilure of the f eeevetor U,UR - Fallure of the high-pressure injection system X - Untimely ADS actuation i V - Fe i lure of the iow-pressure ECCS 1

l W - Fa ilure of RHR system, RCtC

!" staan concensing, and pcmer conversion system (PCS) w(P) - Event W given that P has occurred WSW - Loss of normal and emergency service water 1 -

R - Fa ilure of recirculation pep trip system W2,wj2 - Loss of contelnment heet renovel l

C12 - Loss of poison injection C' - Untimely scram of reactor protection system p Source: Peperoglou et al., 1983 3-10 1

i

The frequencies given in Table 3-2 assume that the "ATWS-3A- 1 Fix" has'been made at-the plant. This fix is intended to I

/

lower the frequency of the ATWS event. From Table 3-2 it

}

appears that ATWS events (characterized by TC as the first two letters) represent about 3 percent of the core-melt 1 frequency. However, as will be discussed later, they may lead to early containment failure and high consequences, thus contributing significantly to ri3k. )

i t

The dominant sequences as given in the Shoreham PRA are

~

summarized in Table 3-4. Loss of off-site power is also a dominant initiator in this case, followed by ATWS events initinted by loss of Conde7wer/ Main Steam Isolation Valve 3 (MSIV) closure. In the PRAs for Mark II containmen'.s, 1* is convenient to bin these sequences into four containment accident classes. These are described in Section 3.2.2 From a frequeacy viewpoint, the Class I containment iailure i phenomena are dominant in the Limerick PRA (Table 3-4).

Their relative importance from a risk viewpoint is discussed in Section 3.2.3. In the Shoreham PRA, the same four con-tainment failure classes are employed, as well as a fifth class for loss of coolant accidents (LOCAs) occurring out-side the containment. This fifth class is not dominant (with frequency on the order of .0*9 per year), while the others are comparable to Lime ;R, except the Class IV l In the Shorenam PRA, the Class IV events (ATWS) events.

have a frequency of 1.4 x 10-5 per year because the ATWS-3A-Fix is not used. In this case, the Class I and IV contain-ment failure classes are dominant.

3.2.2 Containment Failure Classes In this section the four containment failure classes described in both the Limerick and the Shoreham are discus-red. The containment failure classes described below are mainly caused by pressurization of the containment. This pressurization can be the result of a steam explosion, a hydrogen burn or detonation when the containment is dein-erted, or the production of noncondensible gases due to core / concrete interactions. Moreover, the pressurization may lead to containment relief via enhanced leakage through

. seals, cracks, and penetrations. Because the containment has an inert atmesphere, it has been assumed that leakage during steady-state operations is negligible. Enhanced leakage occurs at high pressures, is large in volume, and leads to doses equivalent to structural containment failure.

In Section 3.2.3, the relative importance of these failure modes is discussed within the context of risk.

3-11

)

f

4 4 TAPtF. 3-4 SWWARf 0F THE DOMINANT ACCIDENT SEQUENCE FREQUENCIES wHICH LEAD TO CORE VULNERABLE STATES (PER REACTOR YEAR)

BY INITI ATOR AND CLASS FOR SHOREHAM CLAS S CLASS CLASS CLASS CLASS EVENT INITI ATOR CLASS I il til IV V Tr ens lents:

. Turbine Trip 2. 5E-6 1. 0E-6 -- -- --

Manual Shutdown 1. 4E- 6 1. 2E-6 -- -- --

MSiv Closure 7. 4E- 7 3. 5E- 7 -- -- --

Loss of Feedwater 2. 0E -7 4.2E-8 -- -- --

Loss of Condenser vacue 3. 2E- 6 2. l E-6 -- -- --

Less of Of f-Site Power 9. 9E- 6 1. l E-6 -- -- --

IDRV 6. 8E- 7 8. 9E- 8 -- -- --

1. 7E- 5 5. 9E-6 Loss of Coolant Accident (LOCA):

Large LOCA --

6.9E- 7 1. 8E-7 -- --

Medium LOCA --

2.7E-7 5.1E- 7 3.0E-8 --

Small LOCA 2.1 E- 7 2. 8E-8 1.5E-8 -- --

LOCA Outside Containment --

7. 2E- 9 -- --

3.6E-8 g Peactor Pressure vessel LOCA -- --

3.lE-7 -- --

2.1 E- 7 9.9E- 7 1.DE 3.7E-8 3. 6E - 8  ;

ATwS:

Tur Di no Tr i p 1. 2E- 6 --

8. 5E- 10 2. 3E-6 --

MSiv Closure / Loss of Condenser vacuum 8. 0E- 7 --

7. 5E- 10 7. 4E-6 --

Loss of Of f-Site Power 7.1 E- 8 -- --

6.9E-7 --

IDRV 1. 7E -7 -- --

1. 6E-7 --

Loss of Feed water 1. 8E - 6 --

2.1 E-9 3.0E-6 --

4.0E-6 1. 4E- 5 Other Transients:  ;

Cases involving the Re lease et Excessive water 3.1 E-6 7. 8E- 7 --

1. 7E- 10 --

Cases Initiated by the Loss of DC Power Bus 2. 7E- 6 7. 4 E- 8 --

4.4E-8 --

Cases involving an upset Condition with the Rx water Level Meassenent System 2.4E- 6 1. 2E- 7 --

1. 9E- 7 --

Manuel Shutdown due to Hi gn Drywell Temperature 1.4E-7 -- -- -- --

2. 5E-6 1. 2 E- 7 1. 9E- 7 Loss of Service water In itiated Events 3.1 E- 7 6.9E-7 --

4. 6E- 8 --

TOTAL 3. 2E- 5 8. 5E - 6 1. 0E- 6 1. 4E- 5 3.6E-8

\ote: Totals may not meten due to rouno-of f errors.

Source: Long is iand Lighting and Power. 19E2 3-12 I

f

i

'I e .

o Class I This containment failure class is characterized by sequences initiated by transients with scram and with loss of coolant makeup to the reactor vessel. The core is expected to melt relatively fast with the containment intact and at low pressure. The containment pressure just prior to melt-through of the reactor vessel is slightly higher than atmos-pheric, and the suppression pool is subcooled. The contain-ment is expected to remain intact through a large portion of the core vaporization phase. The containment is expected to fail either by overpressure in the wetwell or the drywell, l or by small or large leaks after vessel melt-through. If the containment is deinerted, it could fall via a hydrogen burn. The core power at dryout is less than 2 percent full power, and the containment pressure is at 17 psi. The  !"

suppression pool is still subcooled. The core melts about 1.3 h after scram, and the vessel fails at about 4.3 h. The diaphragm floor is penetrated about 6 to 6.5 h following l scram. Both the floor and the containment fail at this time due to overpressurization.

If significant quantities of core debris pass through the diaphragm floor shortly after vessel failure, the core would enter the suppression pool. The subsequent debris / water f interactions ceuld significantly change the time, place, and mode of containment failure. A strong interaction could result in early failure and potentially higher risk. If both cooling and containment failure occur on a very long l time scale, the risk would be significantly lower.

Class II The core is expected to melt relatively slowly, and the containment is expected to fail at the time of core melt.

The suppression pool is assumed to be saturated at core melt. The release of fission products from the gap between the fuel and the cladding and during core melt occur through the safety relief valves to the suppression pool. The vaporizat' ion release occurs in the drywell with an open containment. The containment failure is by overpressuriza-tion in the drywell or wetwell, or by small or large leaks.

This containment failure class is characterized by sequences initiated by transients with loss of containment heat removal (W) If decay heat cannot be removed from the containment building (via the suppression pool in the Residual Heat Removal [RER) phase), the pool becomes saturated and the containment is predicted to fail after 3-13

e - . ,e l

l-l^

l 1 about 30 h due to overpressurization. At containment f a i l --

ure, core injection is assumed to fail and the core melts. j The melt attacks the vessel and eventually penetrates the l diaphragm floor 43.3 h after scram. j d

C3 ass III J The core is expected to melt relatively fast and the con-tainment~to fail shortly thereafter. This case is very similar to the Class I sequence except that the suppression pool is saturated during the gap and melt radionuclides .

I releases. Therefore, the decontamination is lower for this I

class than for Class I. For Limerick, this containment failure class is characterized by transients with failure to i scram and with loss of coolant injection prior to contain-ment failure. The core is assumed to be at 30 percent full ,

power.. No operator action is assumed in these ATWS events, i The suppression pool is saturated prior to core melt due to  !

l steaming frem the core through the safety relief valves 1 (SRVs) The containment is at high pressure (65-psi) at  !

l this stage. The core melts about 0.9 h after initiation of i

the. event; the vessel fails approximately 4.3 h after ini-l tiation. The containment is calculated to fail by pressuri-ration between 6 and 6.5 h after initiation, depending upon the spreading cf the debris on the diaphragm floor.

For Shoreham, this containment failure class is character-l ized by large LOCAs with insufficient coolant makeup, and l small and medium LOCAs involving Safety Relief Valve (SRV) actuation with loss of inventory makeup, as well as the ATWS l events described above.

Class IV This containment failure class is characterized by tran-l sients with failure to scram, and with successful coolant l injection but without adequate containment heat removal.

The core is expected to melt relatively fast, with contain-ment failure prior to core melt. This case is similar to Class II,'except that the reactor is at a significantly l- higher power level at the time of core melt. The contain-L ment fails by overpressurization in the drywell or the l wetwell. Containment failure occurs about 40 min after {

initiation. The vessel fails about 4 h later so that the I debris enters a failed containment. l j

The four containment failure classes as described above are summarized in Table 3-5. In addition, containment leakage and by-pass are important pathways for radioactive release as described below. I 3-14 )

i

': l "d;

!  ! i,

. r-e llI.

i I 6 .

t's ,

m t>Q-3m* .= .

• .*=*-> 9 U +=

au e v >. . . .

" _g 8 e c =.=. 2C Ti '*

d E * * ::: ,

t,c -

.s > c -

+= r- .Eeb8U t

3 e

8**

-688 8 g t.

' 2 m" 6C e* *e s == _&

v e += += U D += 6==, ==*

UCe 5 C

• U*U* U@

>=@*== @h *@+- e == m av e ==>

6e *&e

• O>gE *e 1

O= 9C CDU .De- e * = = +-

i . ; W *>acp .Ce 9 .c = ==w he 606- 3 t2

-OO2 e.wg . W>-

-mgg

-6 Uv E

e O * *D ec-C 3 * == a 2 -e

>E C9 04/5 F Q => 0 .e-

• m 6 == @ W UeO6

.C U a'

.cceOE -

+- -

u= c .6 ye

- e > == 6

• 4= @c ' 4= .e.e m meoO C .

4 e- 6 .g w OcmO = 4@

_g g - - g==

.<. .m =m[Ocge O* .C;; kG * ,e6 0 *. , ;6 g Z .l*.8 O9 EU O>

>T

-3 U-me.W 6*

9 4

. . '8 O@ m* e t .. c e

-M w.

> == c .Ce& w +. 3 *- C

-4

-e*2 5 @ ==

N m-e @*c 8 ' b. '5 ,

~12 5 e

k~~

c.

c-U 82 Goo kB et X ===A* = < *. e 3 se. e O 6 2e *= c d e *a

> O . =U o - cr >ee cC rs -==c- 2

• e

.J -

p i= -- -ae> 0*c WX o m 0 -

• O =.

> c*@*

- .Os>T e@ e --* D T ce O

.@=0 Oc64

>-e y

h ==

e

.-@ c5Oe3 e e e3 g med

-.wM W

- e .O. e 6

0 6

39 eSemQ c .c*.9 c == a e teU

..=.g=

~

• => $94 9 e r <m & e.r60 >e= mOOO me*

<m 4- c *= @ t *+*

c e ==

Q T +- g ea +- p 6 *=

c

• p 13e w

M ce e - c == . 993 p#ce3 e .. .. @W @e.e cO -

E 3=>=>0 -0O* O +c= =c

>Z - -4 e e t - u* pk 3C..O e ee-E =

Z m m C == 6 c -- 0 0 n C * *-

@- ew@T ce cU eT- > eCc 6 e m* @ W >0 0 e-

~

O ' *-

g

=<- >b g *e C *

• E v* 3 m' * *E se ** e'--* sus

~5

$u -

c , e 6

. C c w .- E O eG & D 2-C @ 6 t e O *- - E @3 'O

- => * -ce == e9C C ee cg *=*-

E a-=

ee ge -.--C.h- @D

==

e .9

== .C 4 ge eS &

==

e*

4 3

et X * - = = @ h$. O m +- *- @ m

• 3ew +- ce C e- *C * .ee- 6 6e6 e 6 t *ea C. 6 e@ee

==

gg;v se aee ve em $

E 0 - e c.

a uCea e +. *a e v e.-e e c. v-ee W V1 w e 3

.. Uu=>

h -.===@ v E .. A 9 0 .. e G

• • 6 ,9 > *- *

$ 6 ,9 +- >e O @>

. 5 - c * - - e

.*.m6 .-

n.

m m-

-> .n e

C

-ee E*C @

- t .0 -

e--e tuc Dec

.. e .-

s o .e .c c

• g et ceO *geO w

t 12-a-

tv..

8-* Bet uu*e "t 'B "O - -

t* *E.8c b-6 e ,

eB*t, E c .,e u- - > u E,e v e .b. *e m ...= 0

,4C

= -6m u e z u .c

= e. . .e 8 c ., e

>. 5 ma wu

& e .. >

E.= t -

awa.-

cee a

=

a ev 3

Z zg 6 u6eBee

.Ee8p .

6 -6 .

- eb b e. . .

be e6 .e=*t c*e w  %. ?

• t. *-.

t oe>*g E. l 6 I.etl'

.g.5

= *= 6 6

. r.> ,*..O.** 8

== @C

>c-.u e>@

+- ======4 12

=w e-

>e.5. *E 6

C. ,6

==

S

• e

• - >L 3 eC Un - e e O.= *o e S*=+= c.

==ec&>0 *- 6

• = e=

m ==mger @ e e.

= = * * - U= - O =-C- D = = = = == G.

, e c 6 e> CC e@6 e == gecc0 e@ >eO eeO@ >

3--3a 32&&2 E E == - *- EEO3n E

• a- 6&

- .=.

4 k

- = - - >

M @ @ @ @

.=.e^<

. M M M @ m

=e < 4C et eC

=o eC 4Q w 6 e W W

.d Q

b U W 3-15

Bypass Leakace from Reactor Containment Another class of failures is potentially important, b'ut was not given weight in the PRAs that are the basis for this work. The reactor containment has a number of penetrations.

Personnel must be able to pass in and out through an air-lock, and ventilation air must be supplied into and evac-uated out of the containment. The closure cap must be removed and reinstalled during refueling. Feedwater enters the reactor and steam leaves it in pipes; cables for elec-trical power and signals must be brought out, and Control Rod Drive (CRD) hydraulic fluid must reach the rods. In the event of an accident with release of radioactive substances, passages through the containment wall are sealed by valves in pipelines. Each pipeline (or most) contains an inner and outer isolation valve for this purpose, and all valves leak somewhat. Furthermore, normal imperfections and wear in these valves lead to slowly increasing leak rates with time.

Through periodic inspection, testing, and maintenance, these leakage rates are kept to values below those stipulated in the plant technical specifications.

3.2.3 Contribution to Risk The objectives of containment mitigation systems are to reduce, eliminate or ameliorate the consequences of a severe accident that would otherwise lead to the release of radio-active materials to the environment. In this context, it is convenient to express risk as:

. J K s s j jk ki 3-1

]=1 k=1 where fj = the frequency of the j th accident sequence

, leading to core melt, Pjk = the conditional probability of the kth containment failure mode, Cki = the i th consequence of interest for the kth containment failure mode.

In most PRAs, several consequences are considered: acute (early) fatalities, population dose (man-rem), off-site property damage in dollars, and illness (usually thyroid cancer). These measures of risk are used in a cost / benefit or value/ impact assessment of the mitigation systems 3-16

. I e

n

-2.-

A i

.. described in-Section 3.6. However, it must be stressed that the risks presented L below contain varying degrees of uncer-tainty which-have not been assessed in detail. Such uncer-tainties fall into four-categories:

1. Uncertainties due toLthe statistical nature of failure data.

2. Uncertainties in understanding various physical phenomena or plant system behavior.

i 3. Uncertainties due to the lack of inclusion of such things as severe external events (floods, earthquakes, etc.,).

I

1. 4. Uncertainties due to unforeseen phenomena such.

as unusual common mode failures, design errors,

,., and bizarre human behavior (e.g., sabotage).

l The risks presented below should be viewed within this con-text. Their use.in value/ impact assessment will be tempered by their uncertainty and will be but one input to the deci-sion. making process. In this section the risks associated with each containment failure mode, as determined in the BNL p review of the Limerick PRA, are presented. These are used

! to determine the types of possible mitigation systems for Mark II BWR containments and to obtain preliminary cost /.

benefit or value/ impact insights. A more extensive value/

impact approach is being developed in Task 4 of this project.

i Table 3-6 describes the various containment failure modes used in the Limerick PRA and the BNL review, and Table 3-7 shows their conditional individual probability of occurrence as a function of containment failure class. In the Limerick PRA, the consequences are given by release category, as shown in Table 3-8, and their relationship to each contain-ment failure mode is given in Table 3-7. Hence, the risk for each , containment failure mode k, in each containment failure class j, is given by J

Rf= FP 3 jk ki 3-2 j=1 where Fj is the containment class frequency (the sum of the sequences making up that class), and Pjk and Cki are the i same as in Equation (3-1).

3-17 l

g. _ - . _ . ._ _ _ , _ _ _ _ .

.a . ' , 4 s

1

( TABLE 3-6. ' DEFINITION OF CONTAINMENT FAILURE MODES i ;. (L IMER ICK) ,

l' S ymbol Definition a In-vessel steem explosion

$ Ex* vessel steem explosion y Overpressure: cryweil f allure .

y* Overpressure: wetwo1i fallure

-Y" . Overpressure: metwel l f ai lu re, w ith loss of suppression pool 6 Small leek 6t See iI ieek,' SGTS f ai fure' 6g Large leek p Ove rpressure: hydrogen Durn p8 Overpressure: hydrogen detonet ton -

'SGTS.= Standby gas treatment system, i

Equation (3-2) has been evaluated for four consequences:

acute fatalities (Table 3-9), latent fatalities (Table 3-10),

man-rea out to 50 mi (Table'3-11), and man-rea out to 500 mi

l. (Table 3-12).- Examination of Table 3-9 shows-that the major p . contributions to risk as measured by acute f atalities are the steam explosions for Class I and Class ~II sequences and'the overpressurization due,to Class IV ATWS events.

Examination of. Table 3-10 shows that the major. contribution '

. to risk'as measured by latent fatalities is the slow over-pressurization failure due to class I sequences. The same is true when measuring risk in terms of man-rem as shown in Tables 3-11 and.3-12.

The'ATWS-3A-Fix is intended to reduce the frequency of con-s tainment Class'III and IV sequences. The BNL review gives the frequency for Class III and IV sequences with and without the ATWS-3A-Fix, respectively. These frequencies were used

- to construct Tables 3-13 and 3-14 which show the risk in man-ren for each containment failure mode by accident class, out

! to-50 and 500 mi, respectively. .As expected, the most-notable changes are the additional contribution to the slow-overpressurization modes for the Class III sequences and the i large leaks. '

3-18 L

L I l I

e p -; .:

>L

.l=

TABLE 3-7 CON 0lT10NAL PROBABILITY OF CONTAlt+aENT F AILURE, RELEASE CATEGORY AND CLA$$ FR[QU(PCY l

l' l

Mode'of Class 1 -Class Il Class til ' Class IV r- Containment (9.5x10-5 -(4,g,so-6 (3.4x10-6 (3.0x10*7 Fe l lur e - yr

• I) yr*I).

yr*I) yr*I) s 0.001 0.005 0.001 0.01 (0

• E) (0 *E) (0* E) (0* E) f I

d ,a ' O.002 0.05 0.002 0.09898 (0* E) (0* E) (0

• E ) (0

• E)

Y- 0.247 0.2245 0.247 0.445 -

t0PREL) (CPREL) (OpstEL) (Cey )

Y' O.1235 0.1105 C. 43 0.2225 (CPREL) (OPREL) (OP4L) (C4y 8 )

• Ya' O.1235 0.1103 0.123! C.2226 (OPR EL) (OMEL) (CA EL) (Cay ")

0 0.2223 0.500 0.2223 -

'j (n one) (n one) (n one) --

d 0.0247 -- 0.0247 --

(OPREL) - (CA EL) -

l l-Sc 3.247 -- 0.247 -

(CPREL) -- (CPR EL) --

u 0.009 -- 0.009 -

(OPREL) -- (0PREL) --

1 Total 1.000 1.000 1.000 1.000

. Oeeini+iens:

CSE is *me exicet ton release.

OMEL is t*e overcrossurization release.

C4Y is t ellure of tee arveell release f or ATws.

C4Y ' is f at ture of the etwe61 epove *me suppression pcci ret ense + ce ATWS.

C4Y " i s

• a l lur e e t *me etee l l belo the suceression pool ro t eese f or A ?ns.

3-19

( :.

H l

TA8LE 3-$. ' CONSEQUENCES FOR E ACH RELEASE CATEGORY (BNL REvlEa)

Refosse Acute' ' La t ent' Me nd em ' Wendem'

-Category Fete il tles Fatelities (500 miles) (50 miles)

CPREL 0 2.2 x 103 1.42 x 107- 0.7 8 x 107 0*E- 97- 1.9 x 104 4.90 x 107 2. 5 x 107 C4Y 75 1.4 x 104 7.88 x 107 4.7 x 107 C4Y 8 69 1.4 x 104 7.86 x 107 5.3 x 107 Car a- 138 .1.3 x 104 7.36 x 107 3.6 x 107 .

• Besee on w AS*1400 source-term and metnocology.

TABLE 3-9 ACUTE (E ARLY F ATALITIES)/YE AR FOR E ACH CONTAl*ENT F AILURE MODE - INTERNAL INITI ATORS (wlTH ATwS-y-Flx)

Failure Mode Class ! Ctess !! Csess til Class tv a 0.9 x 10-5 2. 0 x 10-6 3. 3 x 10-7 2.9 x 10-7

' S .u ' s.8 x 10-5 2. 0 x 10-5 6.6 x 10-7 2. 9 x 10-6 7 0 0 0 1. 0 x 10-6 Y' O O O 4.6 x 10-6 Y" 0 0 0 9.3 x 10-6 6 0 0 0 0 at 0 0 0 0 Og 0 0 0 0 u O O O 0 Total 2. 7 x 10-5 2. 2 x 10-5 0.1 x 10-5 2.7 x 10-5 Total r isk = 7.7 x 10-5 acute f atelities/ year '

l

• Besed on wAS+1400 source term enc methocologr.

3-20 t'

'g

, TABLE '3-10 LATENT FATALITIES / YEAR FOR EACH CONTAllesENT FAILURE MODE -

INTERML INITI ATOR$ (WITH AT%S-3A-FIX. ) OUT TO 300 MILES -

j 4

Fa i lur e .

Mode Class I Class il Class ill' Class IV a ' O.18 x 10-2 0.39 x 10-3 0.6 5 x 10*' O.57 x.10-8 S ,u

• 0.36 x 10*2 0.39 x 10-2 1.3 x110*' O.56 x 10*3 Y. 5.1 x 10*2 0.20 x 10-2 1.8 x 10-3 1.8 x 10*3

'Y* 2.6 x 10-2 0.99 x 10~3 0.92 x 10-3 0. 94 x ' 10-3 Ya ~2.6 x 10-2 g,99 x 30-3 0.92 x 10~3 0.8 8 ' x 10-3 6 0 0 0 0 q, ,

6, 0.5 x 10-2 o- 1. 8 5 x 10*' O 6

4 5.2 x 10*2 0 1.84 x 10*3 0 y 0.2 x 10-2 0 0.I x 10*3 .0

l .-

Tots 1 16.8 x 1C*2 0.83 z'10-2 0. 5 8 x 10-2 0.42 x 10-2 Total risk = 18.6 x 10-2 latent f etalities/ year l ..

• Besed on w A5+1400 source terms and methodology.

-TABLE 3-if. MAMEM/4AA (OUT 70 50 ulLE5) FOR EACH CONTAINMENT FAILURE.

.' MODE - I NTERML I NITI AT(Fl$ (WITH ATh$-3A K lX) *

-l' Failure Moce C ses I Class il Class til Ctess av

., s. 2. 6 0.56 0.09 0.08 6 ,a '  !.2 S. 6 0.18 0.08 Y 18L S 7.3 6.5 6.3 Y' 93.0 3. 6 3.3 3. 5 Y* 93.0 3.6 3.3 2. 4 3 -. -- -

l 6, 18.S - 0.67 --

6g 186.5 -- 6. 7 --

u 10.7 -- 0. 3 -

Total $910 20.7 21.1 12.4 Totsi etsa e 646 men-conheer (!O mileel

, *Beseo on s A$m la0C source

• erg one wtmecology.

l 3-21

= _ ___ _ __ - _ _ _ - _ - _ _ _ _ _ _ _ _ _ _ _ - - - _ - _ _ _ _ - _ _ _ _ - - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - -

,.'_.)

_I l

TABLE 3-82, MAMEwvtAR tout TO SCO wit.E5) FC4 EACH CCNTA t*ENT r48 LLRE -

200E - I NTERNAL INITIAT35, . (W ITH ATw$=3A 8 6 x)

• Failure moce Class l' Class 11 Class ill Class lv a 4.6 1.0 0.17 0.14 S ,u ' ' 9.3 10.0 0.35 0.14 T 326.0 13.0 11.6 10.5 l

1

!- y' 116.0 6.4 5. 9  !.2 Ya 166.0 6.4 5.9 4. 9 s

6, 33.0 -- 1.2 --

dg .333.0 - 11.9 -

u 19.0 D. 7 Totsi 1057.0 20.7 21.1 12.4 Total risk = 1152.4 wn-* emh eer (500 mi les)

• Based on m A$r*1400 source torno anc metmocology.

TABLE 3-13 MAMEM/YE AR (OUT TO 50 ulLES) FOR E ACH CONTAl*ENT F AILURE M00E - INTERNAL INITIATOR $ (wlTHOUT ATWS-3A-F 6X)

Faiture Moce Class I Class 11 Class ill Class lv l

2 2.6 0.56 0.94 0.20 S ,a ' 5.2 5.6 1. e 0.20 Y 182.5 7.3 68.8 16.8 l Y' 93.0 3.6 35.3 9.52 Y" 93.0 3. 6 35.3 6.5 1 6 -- -- -- --

6y 18.5 --

7.28 .-

6r 1 B6.5 --

72.2 --

p 10.7 --

4.0 --

Total 592.0 20.7 225.62 33.22 Total r Isk = B 72 ma n-r em/y r. (50 mi les).

• Based on w AS+1400 source terms anc motbocology.

3-22 1 4 i

l i

i lk;b[  ;

l}

j

. TABLE 3-14 'MA'N-REM / YEAR (OUT TO 500 MILES) FOR E ACH CONTAINMENT F AILURE MODE - lNTERNAL 1NITlATORS (WITHOUT ATh5-3A-FIX)

'Fallure.

Mode. Class 1 Class 11 Class 111. Class IV a 4.6' 1. 0 - 1.8- 0.38 S ,y ' ' - 9.5 10.0 3. 5 ' O.38 Y 326.0 13.0 123.0 28.0

,; Y'- 166.0 6.4 63.0. 14.0 ya- 166.0 6. 4 - ,63.0 13.0 6 ..

1,

'33.0 13.0 --

6t 333.0 -- 129.0 --

64 p . ts 19.0 -- 7.0 --

,- Total 1057.0 36.8 '403.32 56.0 Total r Isk = 1650 man-rem 4r (500 mi les).

• Based on WAS*1400 source terms and methodology.

The Severe Accident Risk Assessment'(SARA) for.the Limerick

.I Plant (Philadelphia Electric Company,-1983) shows that the major contributors to risk from external events are fires and earthquakes. Fires increase the frequency of Class I

sequences by 24 percent, while. seismic initiators increase

- [~ the frequency of Class I and III sequences by 3 percent'and 26 percent, respectively. Class-II and.IV sequences are 4

also affected; but as indicated above,.the major contribu-tien to risk'are from the Class I and III sequences.

~

Seismic events can also cause containment failure; however, these events are considered uncontrollable and are therefore treated separately.

3.2.4 Summarv l For the two BWR Mark II containment systems reviewed in this chapter. the dominant containment failure modes are slow overpressurization due to Class I and Class III sequences.

This was shown in Table 3-7. From a risk perspective, the Class IV ATWS sequences leading to steam explosion and slow overpressurization failure of containment, and Class I and II transient and LOCA sequences leading to steam explosion 3-23 i

L__ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ ___ _

a . .

2 l

E failures of containment, all contribute equally to acute fatalities. Latent-fatalities or man-rem are dominated by  ;

Class I and Class III transients leading to slow-overpres-surization failures of containment or large leaks _(Limerick)  :

as,well as Class IV ATWS events (Shoreham). 'It should also  ;

be noted that the dominant initiator of transients is the loss of off-site power. These considerLeions lead naturally  ;

to the mitigation concepts developed in Sections 3.4,-3.5 l' and 3.6 that follow.

I 3.3 MITIGATION REQUIREMENTS TO BE MET The preceding section showed that the dominant failure modes for the Mark II containment result from a limited set of l-severe accident sequences. These sequences, in turn, pro- )

duce a set of possible accident end-states,.or physical.  ;

i situations within the containment wherein accident preven- l tion measures have all failed and containment failure can be predicted. For any one accident there may be more than one possible end-state, but the total number of possible end-states is quite limited. A fundamental premise of the l- present project'is that effective mitigation requires the i

capability of intercepting the progress of any or all of

l. these end-states toward a containment failure, since it is >

I impossible to determine which one will actually occur, and. I since the act of foresta111ng one mode of containment j failure will not stop the accident in progress, but will instead increase the probability of some other mode occur-ring.

Thus, a " complete" mitigation system is favored over partial implementation, within the limits of cost-effectiveness, since the risk averted by a single type of failure preven-tion tends to be carried on into the subsequent failure to which the modified system may be subject. Only when there is a high probability that the accident is arrested and the containment stabilized without failure can full credit for mitigation be taken.

Listed bedow are the most probable accident end-states for the Mark II containment.

1. ATWS steam ceneration. The end-state results in an energy charged containment system and over-pressure with a continued high steam generation rate.

2. In-vessel hydrocen ceneration. The end-state results in additional noncondensibles in the containment. It also presents the possibility of l

3-24 1

N .,

{.b -

i a hydrogen burn or explosion when the. containment I j-' is deinerted (about 10 percent'of the time), or

• is vented.

4 3. Containment concrete decomposition. The condi- l tion results in steam and carbon dioxide to add  !

to the already high containment loading.  !

l'

-I 4. Ex-vessel steam pressure rise when the hot core

debris encounters water. A short term.but high rate of steam generation occurs when residual j sensible heat in the-core debris mass is released'

. to water, resulting in possible containment over-  ;

pressure..

] 5. Ex-vessel steam explosions. While this phenom-enon is still very much in controversy among the technical community, our philosophy requires that we consider it initially as an assault to be  ;

l dealt with in accident mitigation.

6. Ex-vessel hydrocen ceneration. This end-state

'l I- results when hot steel and any remaining Zircaloy in the debris mass contact hot steam and react,

' adding combustible noncondensibles to the gas l

loading of the containment.

i

7. Residual heat load. Thic condition occurs from  !

the radioactive fuel decay energy and can result in a containment overpressure in the long term.

To control these possible accident end-states., or some com-binatior. of them, a ecmplete mitigation system must be capable of the following functions. However, the actual choice depends on the results of value/ impact analysis for a '

'l specific plant.

2. Containment venting with overpressure relief valves to release relatively clean ATWS steam to

'the atmosphere and a diverting system to pass later and generally smaller flows of con-taminated steam and gas through a condenser /

filter system.

2. Adequate long-term heat removal from the con-tainment during the accident and as long as residual heat in the core is being generated.

3-25 f

l l

3. Core debris mass control during its course from the RPV.to a long-term retention area.

I

4. Adequate long-term cooling of the core debris once it leaves the RPV.

5. Vacuum breaker system to preclude containment underpressure as steam in the containment is condensed.

l 6. Hydrogen control from the onset of the core-melt i accident and as long as necessary afterward--

I generally until the combustible contents are below a flammable range, if the containment is l deinerted. Alternatively, reduce the amount of time that operation is permitted with a dein- l l erted containment. I

7. Missile shields to protect seals and penetra-tions from failure due to steam explosions. k As with most LWRs, long-term slow overpressurization repre-sents the main element of risk. However, for a Mark II system the most unusual aspect in mitigating the effects of severe accidents is the possibility that a core meltdown might emerge from the RPV into a containment already at >

l design pressure with a saturated suppression pool. One I j pathway to such a condition is through a sustained ATWS event: others include sequences beginning with a turbine trip or pipe rupture. Even though the proposed ATWS-3A-Fix j would reduce the probability of an ATWS event, the overall residual risk appears to be sufficient to justify provision for venting the clean steam from the ATWS, as well as provi-  !

sion against hydrogen formation (as a noncondensible gas) and concrete attack when the suppression pool is already saturated at design pressure.

l The following section provides detailed descriptions and i cost estimates of the individual mitigation components pro-posed f or use in a Mark II containment.

i 3.4 CCMPONENT DESIGNS, DESCR!pT!CNS, AND CCSTS l

The components described in this section fall into the l following classes:

j e Dedicated heat remova3--2 types.

I e Drywell cooling with sprays--2 types.

l e Ex-vessel core contro2--3 types.

e ATWS clean steam venting system.

l 3-26 f

% .c 1

p ~

J1;s

<

t. , , , ~

~

c.w a l

Qi '

m

s. s Overpressure venting through filter. ,

'#!' e Control of hydrogen. ,

I lp 4L 's: Control of containment' underpressure. -j

, - AnLaiternative non-pressurized' containment system venting 1 1- 4 through milarge chilled filter-is also' described and 'l evaluated.in Section 3.5.- {

>O ,

3.4.1 Containment Heat Removal' .j 1

~

Mitigation of'the consequences of a' cote-meltiaccident must- i sn . presume that; electric power at the p.lantais; unavailable,  ;

sinceEloss,of off-site power is a dominant initiator. In- i

' the case where:an:ATWS event.has occurred.1the reactorzcore- j s is no. longer critical, but the containment is thermally ls - charged and residual radioactivity in the coretis'still' generating. heat:that is not being. removed:from the contain- i ment vessel by the normal electrically powered cooling sys- l

' tem. 1This requires.a separate cooling system, independent 1

i

'O of the present(equipment and not. dependent upon the_presen*

- electric power.and-control in any way. This new system _

l should ue as' passive,.as: automatic,land as' simple asJpossible. .{

. b. i In some accident scenarios that end-up as core-meltisitua- 1 tions, a considerable amount of steam is delivered'to.the l drywell area together with core' debris and contamination in several. forms. A completely passive system of heat removal from.this area would be' desirable., but surface-type heat

'j ' exchanger equipment.to do this, such as heat. pipes or

-! thermosyphon systems, have the following disadvantages:

e The internal surface required to handle the heat i load is large due to the low temperature dif-ference and to gas " blanketing" by the non-condensibles.

'.l L e Many containment wall penetrations are needed.

o Air-cooled external heat exchanger surfaces are l extensive. l

. )

e Secondary containment buildings make the install- 1

- ation complex since not one but two walls must be penetrated.

a t; e Surface condensers do not wash contaminants and ]

E aerosols from the drywell areas down into the .,

suppression pool, e The resulting cost is high.

'1 3-27 I

l . .

l l-l Since passive heat removal does not appear practical, a-L dedicated ~ powered system is considered instead. The use of power, i.e.,' pumps, to' force suppression pool water through i L

the heat removal system not only permits spraying water throughout the drywell areas to. condense steam, cool core debris and-wash aerosols and other contaminants dor.'n into the suppression pool.but also results in a lower cost l installation because fewer pipes and smaller heat exchangers can be used when' forced circulation (i.e., high pressure -

drop) is available. Backfitting also appears to be more practical with forced circulation where fewer and smaller l pipelines suffice.

l The suppression pool water level could drop to about 4 ft  :

above'the basemat floor if the containment were ruptured at I

l that low level and the outer building were flooded with pool water. The intake location for the cooling system pumps I should take this into account to ensure adequate net'posi-tive suction head with this hot water.

The dual mitigation heat removal system should be completely redundant in every respect so either system can provide the necessary cooling and spraying. The plant should also be physically separated from the main reactor building for easy installation and must be designed for the proper seismic, fire, flood, and sabotage criteria. Since the pumping power will be supplied by direct-connected diesel engines, every precaution to ensure their starting and proper operation should be taken. One possible arrangement is shown in the schematic drawing of Figure 3-5 and the typical plan of Figure 3-6. i l

Shield barriers will protect operating personnel when.the i contaminated pool water fills the pipes, pumps and heat exchangers. All plumbing in this system should avoid valves  !

that can be mismanaged and cause a malfunction at a critical t ime -. Diversion valves used for checkout tests would be j manually held in the test position. When released, they  ;

would automatically assume the normal operating position. l The shell and tube-type heat exchangers may be formed of i multiple units in parallel in order to use lower cost bu* I high-quality commercial equipment, i.e., one standard unit j of 11.25-MWt capacity costs only about $16,000 bare at the i factory. Four such units would be required for cooling suppression pool water from 285'F to 190*F while warming pond water from 85'F to 175'F.

3-28 l

l l

l

i e

' 'I i

1 t

Containment suppression pool n hTo spreys,

.! pool or core catcher Isolation valves a

?

/ i i \

l I

/ l \s------

e-. .------- , -,

I e j l 1

i I

I i

e 8

8 I I i I i I l

I I i I

i l I i i i !s Suppression l

Heat i I Startup fg poog i

exchangers controls 1 circulating op -  ! W**

i a fiEX 1 a i

A![ p 1 HEX l c c ^w-l 1 1

- T 9 - l

\

}

L-- p / g p M. J g* 2 / \

• g" (v

s/ / \

/

Cooling water 7 9 circutsting E E g N

pumps Diesel engines f ~ M bhp o o iP if L 1 To and from cooling source Figure 3-5. Schematic dual heat removal system.

3-29 I

l.k 4

~* e i

4 l:

1 l

f l

l a

l 2

$ =

N 1

> l >

>I

.r s . - e> 43 *r TB l g- a-

. ,g R - it i .

I **5 '

1 g.

E '

S

• J g i

.. .m*:..

o ,

~.s

.65 .U

. , . ;i: t;: ,

l. 1

. f. . - <

, , - .,1. .

i

. . . . . . . - . . . . . . . . . . . ...... 1 I .

3 i

W a

,/,........... .. .

........ ......... .. .... 3 $

.a, j

)-

Ib o

. i

.. .i .,

g j t

... . m

, a ,

)ll i

. . . . . .. .. 2 r.*

Q.:

.. i 8 = l-

• t. 5x i' it -

A. O~ >

.i 1 j .. ..

O

.' yg.

• o . . . . * .

a t s

, .* t ' . . ? l  ;)_:: *. '. .;; ... g

' I y

., n  :::; !I

.e. .

l

'Lg 8E Y

2

..... l, '.:s .. . ..a I

• 0 g d f

, i6 i

.... 7. j......,e ,

.. '.......... 4..+................ ... / g .

• 1 Q

...e... /

,)

s C

3 'g

- .. O l

8 g .... .-- ................................. ....., .* j s . .

g 3 . . . . . ...i..........y..+................

3 , , ,

, 5 j~gg,,,----,+p----- g 'M i l .

. ?.,'; o g....... . . . .......

i y

% y ,

, 4 g

• - .,. i o n .: g o y 8!O 2 ll i =

.. . . e i .

2e * '

8 l

i

.' . ';: %.. . .'. ]

. ..: . :? : '.  : .

i 3 *

• i i

e T a.

I gC

• G -

- L&. c l

y m Sg .

'?: o i

i, ' . . . . . . . . . . . ......... ............ ... .. . ,)l6 9.

t s

% n

.. s...............:7, ....... ............. .....*S. .

D c

m

.w

• - .' 9 I

• • . . . ..- .. * =-

• ; " i ; , * .. .

1

.J_ . ';* ,, ' * : . .

o

> 1

'd

= 1 ec g .-'c l

.=

- a=

.2 G3lI Y

l R  !

J 3-30 l i

h' o ,

t The pumps would be directly driven from the diesel engine, and would be standard units of horizontal construction where their physical location is adequately low with respect to the suppression pool level. If this cannot be readily achieved, a deep well-type booster pump will be required (at extra cost). The alternate choice of an underground pump room at the end of the core-catcher tunnel would avoid the net positive suction head (NPSH) problem with the hot pool water, since this location would give ample suction head.

The suppression pool water side of the system would be dry and isolated until the engines start and cooling pond water is available at pressure (Figure 3-5 and 3-6).

Each of these two systems consists of:

I e A standard commercial diesel engine of about 350-bhp continuous output complete with automatic starting, fuel supply and exhaust silencing (Caterpillar, 1983).

1 e A direct shaft-driven circulating pump designed to take suction from the saturated suppression

{ pool and deliver water through the heat exchanger back to the spray nozzles in the drywell. This centrifugal pump should be positioned low to ensure adequate net positive suction head at the j

inlet under saturated pool water conditions.

Estimated operating conditions are 3200 gpm 150 ft total dynamic head, 160 bhp, and 75 percent l operating efficiency at 2000 rpm. Such pumps are commercially available (Ingersoll, 1983).

l e A direct shaft-driven circulating pump to take suction from the plant cooling water source (the ultimate heat sink) and deliver it through the opposite side of the heat exchanger and return it to the source. Flow capacity is about the same (depending upon the water temperature), but the head requirement generally will be less as no sprays are required. Low-level installation may help pump-suction conditions here too.

. e The two pumps will be baseplate-mounted together with the diesel engine driver. The pump shafts will be properly coupled to the driver so they cannot be 2ncoupled. There should be no valves in either circulating system, except for hydrau-lie isolation valves to be installed on both i sides of the containment penetrations. These l

normally closed valves will be pcwer-opened by 3-31

)

e l

i the pressure of the pond water circulating system as soon as the engine and pump start, as shown schematically in Figure 3-7. The suppression pool water circuit normally will be dry, but provisions are made for testing all equipment except the nozzles. Note that electric com-ponents are avoiced.

e All main flow piping would be about 12 in in diam-eter and of steel with adequate supports for the designated seismic action. Piping would be buried or protected against freezing or sabotage damage.

o Starting and control of this pumping / cooling plant should be completely separated from the ,

area grid and the in-plant electric power {

systems. There would be a battery and charger system for each engine cranking motor, good for four starting cycles, and with crossover cir-cultry between engines. Although rapid engine startup is not essential, either engine would be j expected to start and be operational in less than a minute. When started, the engine will come up to the governed speed setting and operate contin- j uously until manually shut off. Both engines I together will supply twice the essential heat  !

removal capacity, but only one is counted upon I for adequate cooling. l e The cranking / starting sequence will be initiated by either an overpressure or overtemperature signal from the containment vessel. This signal can be direct, operating through piping to avoid reliance on any electric power. It may be desirable to have a preselected short delay to avoid spurious starts. A manual override of the automatic startup is needed to avoid certain unwanted operation, but the voiding override must require manual reset every f ew minutes f or saf e':y.

e Heat exchangers of the horizontal shell and tube type capable of removing 45 MWt from the suppres-sion pool water pumped through the tubes while warming the cooling pond water being circulated in the shell. Although a variety of configura-tions may be considered, the costing used here is based upon four standard units in parallel (Young, 1983) with ~3996 ft2 of total heat exchanger surface.

3-32 l

l

1 -

1 y .; ,

j.

1 l

i l

i, i

i);;

~

5'bh: "" sit 8M;; N..

!:M)  %?}.Vh{hjff:+ } 27!.liiLhb r :4: : P V*k' *ctu*

.#T e NMn R.1:dith,. . S::;!r d sprine esased -

pressure opened h'en.

>:fl,f fy.j.d::k.,f.',[ie$'I;h,$$Ihy.!:[f

!j,,. :(,0 f

lg 'n  ::!}.?ih.':: ;r;?B(!::1f Poad t -

weter sll!" hj !67.g. =

preuure - -

.y "- M"h

ihili!W%Y

.ig. : li;g.pg/p:glf.;4,g@cg

. ,3

. g,q.:.t

=

=

& llq$.L.e "1 F F

W e ep;b fl. rur:s iv 3;.i:is.t.iyd:.::e

,n.::m:hy:aq ,

=

C 9 E Inlet j ( 3 screen Pool water \ =

to pump -

.gf,f =

1 suetbn 3;,g; 3;,4j -; :,;;y.;;.;:g:t: ggjggy;l{pg;.,;-

f;Mp?:'lg,ppjg 6 *tMiit(ti:i$,ff!::?,E g

l 0:iis

.' 'A . . .i :.! .\

P001

??n'::ti.'

U.W:.1p 2

. ?.u; M: j a'

i!..::

i : :w.

w 9: ij ':l: h.'G:f'n "Mi:::*i .~::'i?:n-l?$

!5$: 5'.l.5.i.i.Y.5.

.:w.::Qh:i?.'-;sf:

'lik'?, 6 v-  :;c  :.x - $.h:b.. llk:::::. b'?$.k,)l 5v 5[?!$y::'*:';.jiT0 x .!.h:v N.

5. ##"

-l

'5tW/ h,'p#"f'5'ik!C R:n. ~Q'<yis"'

ji'01 %if'.%. t$' i :? fiff.  !, 'A y 0S: 4':7p:4 ' n.:yi!:'i CO .: Y.

                                                                       $$!.f@iY5ib.$,{W))?.:e       i-                          TM:; -~!E,:!fll'Wif;f
                                                                            , ::. ;;ar            ~
  .i                                                                                 containment well 1

Figure 3-7. Schematic - double isolation valves for intake. 3-33 f

o Housing must be provided for the pumping equip-ment except in the alternatives where it is located underground near the dry crucible core debris retention unit. This dedicated pumping station should be designed to copo with seismic motion, sabotage, and flooding. At most sites i the surface unit will be sunken to provide good inlet conditions. ]

                                                                                       )

e Mitigation of the consequences of a core-melt event might include cleanup and recovery consid-erations after the crisis. In this vein, the pumping and spray system might include provisions for adding some form of water cleanup--such as ion exchange and filtering--to wash down the upper drywell areas with clean water. It need not be a full flow (3200 gpm) system. The 120,000 ft3 of suppression pool water will, theoretically, be j exchanged every 4.7 h with one diesel pump oper- l ating. Raf.ioactive cleanup provisions are not  ! included in the cost estimates. The estimated cost of the heat removal system located at the ground surface will vary with the time it is built, the s location of the plant, and the status of the plant where i is being installed. We have examined installations at plants in the design stage (A), installations at plants already under construction (B), and installations at operating plants with the critical operations done during normal refueling downtime (C). The totals are shown below, but the detailed breakdown of these figures is given in Appendix A.1 Plant Status A B C Total in S/1000 2085 2450 2770 In plants where the dry crucible core retention and cooling system (Section 3.4.3 Type 3) is installed as a retrofit and the underground tunnel approach is used as shown in Figure 3-16, it is possible to locate the dedicated heat removal system at the end of the tunnel near the caisson. The same equipment installed here is a little less expensive than a surface installation. Costs are given in Appendix A-2. It must be borne in mind that the costs given here represent the most costly version--a separate, dedicated installation in duplicate, taking no accour,t of possible use of existing plant equipment. In any real case, it is likely that the required functions can be fulfilled adequately by upgrading existing equipment or by merely providing a nonelectric 3-34 i t L_ __

t . . energy source. These opportunities will reduce the costs shown. 3.4.2 Drywell Sprays Proper drywell spraying with cooled suppression pool water is an important component in accident mitigation systems for all BWRs. While the suppression pool in these containments is a large heat sink to handle steam passing through it, the upper drywell spray system normally provided is of inade-quate thermal capacity to condense the steam generated under severe accident mitigation, even when the system has sur-vived the accident. Without electric power, the existing drywell spray system is useless, and we must assume that this plant power will not be available. The dedicated heat removal system just described in Section 3.4.1 is designed to deliver cooled suppression pool water under enough pres-sure to supply a ring of high-capacity spray nozzles high in the drywell. These sprays would not only condense steam j vapors in these sections but also cool the noncondensible gases such as nitrogen, hydrogen, carbon monoxide, and car-bon dioxide. This would encourage agglomeration of particu-l lates and aerosols, washing them down to the suppression pool for better radiation control. l Installation of this spray system, Figure 3-B, may take any i of several forms to provide good coverage and yet not inter-fere with existing equipment. Figure 3-9 shows one schema-tie alternative that has the necessary double isolation f valves, presents minimal hazard from internal missiles, and can be installed mostly from outside the containment. the cost of two ccepletely separate but essentially iden-tical spray installations each capable of handling about 3200 gpm of cooled suppression pool water, will vary with plant location and time of installation. The costs shown i below are for plants at different phases of construction. Plant Status A B C Total-in $/1000 565 728 860 A detailed breakdown of these costs 6s well as those for the i , internally supplied spray system is given in Appendix A-2. I 3.4.3 Decraded Core Control: Retention and Coolino l The consequences of a core-melt accident can be mitigated by requiring that the core material and the debris associated with it follow a carefully engineered course from their initial melt site in the RPV to permanent retention. When 3-35

                                                                             , f L____--_____

i 1 I y a I - Protect "" ' " penetrations Heeders

                                                                                                      .c with spray                       -                                                          .~
                                                         .                          :c
                                                                                     *s.3                            .

rY ?r n

                                                          *;                         y                                  -
                                                                        '             N
                                                       ..               j               ...

6 ' O' _ ._

                                                     .- * . <                                           ; *; .- : n.. P:. 3

m. , " . .n .

                                 ~,                                                           .
                                                            ;.a                              .                                                                 l swv                             .   .4 (3                ..

t.. from hast . . exchanges ,; ,'.. i. y

                         ~     .  -1          m.          . ' ' , __---
                                                                              -,         . ~ m-             J,',k
                                                               .                           .i
                                                                                                               .-                                              i
                             .              .                p                          .,. .!                   ,,

e ce r ,

                                                                                                               -                                                1 EE                                          ,              .

s s a *s... .,,, 4

                                                                                                  ,s,s.,s,             ,

t,,.<4 *

s. s ass *s. ..s ., s s , se ,+ o i

os,s. s-s s,.,s s. **  ; i _ ,,. ..s = , -- Sprays into upper drywell. l Figure 3-8.  ! 3-36 i

I

             .e                                                                                                                                                   i i

I l l' .ill.

 .1                                                                       ;                ?!&            ^:i;.gi;

• ~ 9:+."'- .

n.  :.m. ..

Flexible , .,..- connection  :::. - :: chs valve ,,.

                                                                                   ..            ~~ - ..
                                                                                  . $,i]E.*
                                                                                             'N?k?I ** . , ' .

ud ;ff f. Etel?MN 'Y x' y se.:.,i,:ii;; [..,. enecx valwe

                                               %                                                     ,.!l
                                                                                %                                                   Soray l                                                                                                                                " "'"

5  %

                                                              ..hf,ilh.it%                                                     e /

Y ::::S

?i's:{t...

Header- g; .' - e

                                                                                        . .;,g' 3.,

1. ,N,5'E d.j,f'.$5

                                                  -~

r ..  % iii@f.,**',;~..l

.f Y EY ,:5::.

ygg 'lL . . 3 5 hY5$.$

   '                           suppression       ~ ' $$N'                        '.
                               * ' ****'         WI,f:Ifi.?,jj%.'9E'::h.f,fE Containment wall i

Figure 3-9 Schematic of externally installed drywell spray system. 0 3-37 9 f

core cooling ability is lost in a reactor the residual decay heat generated raises the core mass temperature initially at about 3*F/sec. Although the heat generation rate falls rapidly at first, the 163 tons of fuel and about 55 tons of zirconium and associated steel will reach a mean 3300*F in about 15 min. The core center temperature will rise faster than-the periphery close to the cooler wa]). The core support structure and the RPV lower dome will not withstand these high temperatures very long, so that a failure of some type is certain. During the course of this heating, a large fraction of the zirconium fuel rod cladding may be oxidize" by any hot steam present to release hydrogen. The core may . be manually reflooded and incur even another dryout; thus, most of the zirconium must be considered exidized, resulting in about 35.5 lb of hydrogen generated per ton of fuel. Not only does this exothermic reaction (-18 kWh/ ton of fuel) add to the heat release inside the vessel but the resulting hydrogen gas poses severe overpressure and fire hazards in the containment. (Refer to Section 3.4.5.) Core Control Several factors in the design of the Mark II reactor ensure _ that the ex-vessel hot core debris flows downward into the

                                ~

center of the pedestal rather than horizontal,ly outward through the upoer pedestal wall into the drywell annulus. In their normal location in the RPV, several layers of steel, used for internal circulation control, encase the core and shield the 5.91-in thick RPV wall from the heating core mass. It is estimated that about 10 MWh of heat energy is required to bring this thick steel belt section to a weakened 1500'F from its normal operating 550'F. This steel heating takes at least 10 min after the core reaches molten temperature; but during the same time, the uncooled central internal structure will have failed by overtemperature and will have dropped the fuel to the bottom of the RPV where it will quickly melt the thin-sectioned control rod drives and seals. Once a bottom seal has given way, the core debris may be forcefully blown out if there is still pressure in the RPV,

 , or it may fall out by gravity to the diaphragm floor below.

It is noteworthy that even should a vessel melt-through occur in the cylindrical section, the core must still pene- , trate at least one foot of vertical concrete pedestal wall l before it can escape from the center pedestal area. , In the most typical Mark II containments, the hot, possibly semimolten, mass will arrive in some form at the diaphragm floor center section of the pedestal. Should a large sec- . tion of the RPV arrive with it, the violent impact could i 3-38 i

[4 . cause the center section of the diaphragm floor to break. I- If'it does not break.through, it will tend to heat up and 8 burn through the concrete floor, or spread out over the floor, through the openings, and out onto the diaphragm. t Core Catcher Tvoes Considered A previous report'in this series lists and describes several general types of core-catchers that have been devised over the years. Most of these units merely delay the radioactive core mass in the containment on its course into the bio-i sphere, not an acceptable outcome in this study. Retention for an indefinite time is deemed essential. This restric-tion, and the need for a practical retrofittable design, precludes aII but a few types of core retainers. The ex- [ vessel hydrogen generation problem cited above can be mini-mized two ways: by allowing little or no water contact.with the hot core, or by providing complete, rapid quenching to minimize the hydrogen generation period. The quenching l technique must preclude steam explosions and minimize pres-sure spikes. A key consideration with each type of core debris catcher is the time required to make the installation in a plant that has been in operation for some time and is down for refuel-l ing. Plant downtime costs about $720,000/ day per 1000 MW, a level that could become the dominant cost factor. For this reason, an estimate of the minimum installation time is included with our costs in some cases when it is critical. l The three most promising types of core debris retention units are judged to be: l

1. Retention in many small piles on a thoria-protected basemat floor and cooled by submersion l in suppression pool water after the core is dis-tributed by the downcomer pipes at the diaphragm level (see Section 3.4.3.1).

2'. Retention on a rubble bed within the central l pedestal area at the basemat level by controlled i introduction of direct water cooling (see Section l 3.4.3.2). l l

3. Retention in a dry water-jacketed crucible located beneath the basemat level. The cooling jackets use cooled pool water from the RHR plant ,

with forced circulation (see Section 3.4.3.3). l 3-39

r t Tvoe 1. ' Core Distribution on Diaphraum' Floor l This core. debris retention system briefly retains the hot material at the diaphragm 1 floor level while allowing it.to spread out from the center pedestalfarea through new ports constructed in the. pedestal wall. The bulk of thetcore debris could arrive on'the floor'within 2 to 3 min after

burnthrough (Fauske,"1981). . Once in the diaphragm annulus outside the central; pedestal, it is retained by a heavy steel,. ceramic, or-concrete dike ring. This circular dike j allows.the hot debris to melt the steel tops'of the enclosed l circle.of downcomer pipes to' clear the way and. allow debris H to flow into the. suppression pool below. Unt'ortunately, .

the.ce;-tainment wall in this section directly above the d31-phragm is penetrated by a large number of' vulnerable pipes, wires and instrument. lines. These penetrations.are not capable of' withstanding temperatures above about 300 F:for long. If the hot core. debris is to be allowed to' enter this outer annulus area, these penetrations must be protected. The dike ring'and shielding as shown on Figure 3-10 will be - used to preclude red hot material.from contacting the pene-trations, and a spray of cooling water will keep radiation heat loading at acceptable levels. _, Under conditions.where the RPV is penetrated and blows down at high pressure (-1045 psi), a portion of the core debris might be ejected vigorously (Figure 3-11). To get a per-spective.of the magnitude of this force, the 24500-ft3 RPV with its normal contents of fuel and steel might sprey the molten or semimolten debris at an estimated velocity of 380 ft/s for a maximum of -42 s (with an assumed 6.0-in-diam l hole in the RPV): then steam, hydrogen, particulate and aerosols would spew out at an estimated initial velocity of 1700 ft/s for an additional 200 s or so, not allowing for hole enlargement. These potentially violent conditions make 8 protection of the penetrations at the diaphragm floor level both mandatory and difficult. It is also necessary to provide some protection for the concrete floor of the central pedestal to minimize the con-crete attack as the hot debris flows.out the several ports to

             .                                                     be provided through the pedestal wall. To assist this flow, a cone-shaped pile of gravel faced with a high-temperature insulation material and clad with steel would be provided.

The many small piles of core debris material on the suppres-sion pool floor are expected to be successfully quenched (see Figure 3-12). For this approach the suppression pool may not have to be drained. l 3-40  ! l l 8 l t

 .;                                                                                                                                                                             g l                                                                                                                                                                        g                                                                                                                                                                                                          .1 B
                                                                                                                                                                                                                                                                                                                  ':isi:,
                                                                                                                                                                                                                                                                . e.                                                
                                                                                                                                                                                                                                                          $!h':fI.,f::*.

kS $l!'. ; *. ; '

                                                                                                                                                                                                                                      .

• l9:::.

                                                                                                                                                                                          ,, ,,;Y^          .. In . :::gy. .:

U e

                                                                                                                                                                 .:.q:9:Q:ig
                                                                                                                                                                                          .'i%: .. N'N: k*llll
                                                                                                                                                                            ,.i:.*:;: i:-s.                      . y.?:e!!::.;
                                                                                                                                                                                                                                                                                  .u :

1 ::

                                                                                                                                                    .9 l:                                                              /,(i
                                                                                                                                                                 ,                                                          ~

g.y,0 ,&'

• Jp.;,','., e

                                                                                                    . . .                myx; * .. v'"

uw::: w .

                                                                                                                                                                                                                                                             "                                               l k,i .                                                                                                                                  $ =g
                                                         ,. jl Mp,ig.l;i,I'                                                    -

6 + E d h :.:: g. IE 'k gg li y s_'V h . .p, -

                                                                                                                                                                                                                                                                                         ^
                                                                                                                                                                                                                                                                                                              \

1 _q .:. .

                                                                                                                                                                                                                            ..:p                                                                                                      t        g
                                                                                                                                                                                                                                                   *n                    5                                      -

r 7-gg h [O I $ h t9 e 4 KS I

                                                                                                                                                          \
                                                                                                                                                                                                                                      ~

L l , m e [Z iin W:  ?.,4li,  :;g:p;x,;q. ,7,y, - g g ;, . , ~ . . .. g., a a v

                                                                                                                                                                       ? : {$lf.fr.j$i$W.'O 5.h:.                                                                        ;                                in')                            ,
                                                                                                                                                                                                                                                                                                                             -l'[
                                                                                                                                                                                                                                                                                                                              ;                g
                                                                                                                                                                                                                                                                                                                                         '     o
                                                                                                                                                                                +                                                                                    j. r y.

y 93 l ~ * ; ,. '

                                                                                                                                                                                                                                                                                                          }

I j m

                                                                 ., ~ Ny                                               ,
                                                                                                                                     .: ,^
                                                                                                                                   - ~

a n - s

• J' 4,C m Le a En.u .

4., j /O j1 3 - 1;

                                                                                                                                                                                                                                                                                                                                             .M
                                                                                                                                                                                                                                            - -- S l

f *5 : l

                                                                                                                                   ?:

W.. . : . $

                                                                 .5 $ b e N".y.d
                                                                                                                                                                                                                                                                                                                                             .g jx.                                                                             .
                                                                                                                                                                                                                                                                                                                                             ,o ll
                                                                                                                                                          .% }[.f;,              !'-

W 1 m i O 1 ,

     <                                              ) e..                                                                                                                                                                                                                                ..30i'.                                     I       a.
                                                                                                                                                                                                                                                                                       .:x                                                       e Y}; .
                                                                                                                                                                                                                                                                        )                                                                     C

\ l C 0' i

 't
                                                                                                                                                                                                                                                        } .7
                                                                                                                                                                                                                                                                              , . ~. .
                                                                                                                                                                                                                                                                                                                                             "n
                                                                                                                                                                                                                                                        .                                                                                    M l                                                                                                                                                                                                                                                                        C.

l

                                                                                                                                                                                                                                                                                                                                     ,V        0 l

i ,. . l k [ _8 Ni:? h5

                                                                                                                                                                                                                                                                             ?

[$

                                                                                                                                                                                                                                                             .l ff" N.n.:M.g.               "-

g.y gg ,3

                                                                                                                                                                                                                                                                                                    }

c J j

                                                                                                                                                                                                                                          ....                3 fu            $                                   ]

f h BeW gl \ /; y 1

                                                                                                                                                                                                                                                                                                                              .        I "I .

3-41 > j f l l L _ _ ___ ____ _ __ t

r I t I p il L 1 h j a External sprays f c=:- <=n

                                                  ==                                        >    ==, as l

l I .

                                                                 '. v. ' ,                                                                                  =

l ' Y s p ,.4 f Internal

                                                               ,        j           J.2::3                                                                              sprays l                                                         *',                s       i .' '
                     ,                                                 31           a -

Penetrations . N ' L/ io. Concrete i its jeopardy , h- ' N00' assaulted __ S T W *.* S .td'e(.'-) A _ _ m .. n .y .

-. , . , :.c a .

                                    *'                                                   .                                                      ,c l
                          -s , s 1
                                      ; ms W             ?.

l

                                                                                                                                    .zi, . .

=

l ~ , J

                                                        .'       _ . _ _ ____          .<                                             J,i
                                                         's
                                                                                                                                                  ~

I *. \ .. .< j . ,

                                                                                    ,                                                             ,=        .      .

i 2 . e i . l .. p , 1 - W

             ;* + $ * * *

• v

                                          ,* ,8 ,,,,.                         f                     ,5 , s = ,, n ,*                                           , n. . . #.1
               . . n .+ .s .=,..,,              .                                                             * ... *..                                       .,* ,

ss - , + . , = , , + ,. , l Figure 3-11. Core debris path after melt { through - no protection. j l 1 l l l-l

                                                                                                                                                                                             )

3-42 l l I

j 11 1!l1al 1 1i I

         ..             -s N_                                                                                                                                                                                                                       1 y$"

9" 5 e2

                                                                          .9
                                                                                                  '         ~

w

                                                                                                                                           ,[j                         r_
                                                                         +w.

c :

j'f. s: W:l*.,.{$ , :q4 s gc.

                                                                                                                                                                ._i

2. i

                                                                                                                                                                    'lc y

f*qys 3 s ' e ht d p p e r o p e' e v - d'8 cy o

                                                                                                       . i's'
                                                     /                                                 .

w z. [: .; 3_ tw '8 e :f 5

                                                                                                                                                              ;[b;   j. ' g;ga    .

e d" I ihr o4, r  ; t s

                                      -                                                                          er8            T e                              ,'

c o - i i,.c u

  ~

w C~ n '.t'- g,#ga ? d

                                                                                                                                    /m-  ~-       g r

i r s tyc., e

                                                                                                                                                                                    ;p-                   m pI1                                          kf
                                                                                                                                      % i   -
                                                                                                                                                                   .^               7, g-           o c

il ', i,( i

                                                                                                                                'f,. r, i

g:}; n

     .                                                                                                                                      v7 f e5,gp.

j s r n

    ?                                 :                                                                                                                        :.?'.                          ,_5         w

A =

j:If-

. lJ. " yg.

i..{:i o

                                                                                                                                                                                             ;. v.       d
                                                                                                                                           ,m               S
                                                                                                                                                           ^

w

       -                                                                             ,. U
                                                                                                                                    .f                      Lll-                                          o v

l

     .                                                dr~ ga    i   :                  .

M

                                                                                                                                    . f_,!}}             !!

3;.y e b

                                       =                                                                                                                                                                 t
                                                                                                                                                                                           ,;g a

m

                                                                                                                                                                     .i

• l e s

l a at l ' a b

                                                                                                                                                         ~

r s ~ i,' t e dY - ^ _ nd ee Cp 5 n o 9: s

                                                                                                                                                                                     ~ ,                 i r

(w gj b n . d e g:, :. w e

                                                                                                                                 /J   ..
                                                                                                                                                  \ up}D, r

o s'".

                                                                                                                                                                                                .        C 3 ._

j* j'j

                                         -                                                                                              ~                              #                                     .

t.s 2

       -                                 -                                                                                            3 s3           .T                                                       1 j                                                                                          , <'

ih ,- -

                                                            -                  .j[
                                  ~                                                                                             /u              -

3 1 H e

                                         -                                                                                            'f      .

i

                                                                                                                                                                                         .5 e

r u tw~ e - n o m# _ i g W -- is s l F eo r o pp . c p . t e. v - u . e r.= S .mce-D. as ., -

                                                                                                                                                                         's..B <<

g

                                                                                         ~

W wi*W - i 1IIIlll,t!i ;I

3 1 The estimated cost of this type of core-container and cooling system is broken down'into operations and componei.ts in Appendix A-3.1. The totals are' summarized below: I Plant Status' A- B .C-

Total in $/1000 -1488 2305 3335 The estimated installation time at a Status C plant is 45 days.

Tvoe 2. Direct' Water-Cooled Rubble Bed In most Mark II BWR plants such as Limerick, t'he suppression a pool water extends into.the lower central pedestal area. In  ! these' cases, it could easily be'erranged to drop the hot core debris directly into the water by engineering a controlled core debris flow through the diaphragm floor.at the center. l I The inside diameter of the pedestal at the basemat is -20.5 ft. The volume of core material would fill this area to a depth of only 1.7 ft, but the debris may contain as much as 50 per-cent voids, with molten metal (steel and Zircaloy) and slag j on top of it, making a probable depth of 4 ft or more. The ' concrete walls should be water-cooled if hot debris is to be l L retained, in order to avoid excessive decomposition. This j L situation is schematically portrayed in Figure 3-13. Considerable simulation research and computer modeling have been done to demonstrate that hot core material poured into water breaks up into small particles and can be cooled if the water is-allowed to flow up through the resulting debris bed. , This. effort is not conclusive to date, and there is some j divergence of technical opinion on the heat removal mechanism ~ in this event. Therefore, this alternative core retention mechanism should be assigned a risk value higher than that of the dry crucible, for example, where cooling has been demonstrated. Also it is not completely apparent that this concept can be installed in an operating plant in an acceptable time (refueling downtime) nor that cutting access portals in the  ! diaphragm and pedestal wall is acceptable stresswise. Refer I to Figure 3-14 for an illustration of the installation problems. To preclude a steam explosion and minimize the steam pres-sure spike as well as ex-vessel hydrogen generation, this core debris retention system is kept essentially dry (1- or j 2-ft water depth maximum) when the hot mass pours into or on i it. Only after the material has penetrated into and been chilled a little by the thoria plates and gravel bed will i 3-44 ' I

o .

   ~l
   .I sd                                                               *W"
                                                  .W                                                          i Add funnel siis Present mess                                                                meterial           . , ,

m

t. ' '

3;, i

                            \          : fin >',,,jf}) MQ:QR).I ,

l{

v. '

Corium flow ' control Stoem a vents e i g) - Shelter ring r s Circumferential 4 waterwell tubes t

                                                              /

ifi8 _ --_-_w m- - Umited water docth f

                                                                          **                                                   Suppression pool Weter                                               *
                                                                                  ****CTrj inflow g                                                        -       rubble bed +

control l4 .y

                                                                        *3
                                                                        ,                   (          g.

Wria plates T I i 1 Basemst 1 (' f

          ,                                                                                                                            k
                       . . . . . . . . .... ~ .n . ~*..

• a : , , , , , r '_ a~..

                                                                                                 . . . ., ~ a , ~ .. , ,,
                  ..       .. .- . r ,~
                                                 , . . , ~ ..          . . . . . .           .n                     ..,*     ,

Figure 3-13. Schematic - direct water-cooled rubble l bed type of core retention. 1 3-45 f l

t l 1 l

                                                                                                                                       )

e Lead shot or sheets ,f* j

                       .   @E0@@E0@E@N
                                                                          ,q4 N
 .                                         Hoist
                                                /                             i     Temporary sup;mrt completed 1       ,

mv

               -                                                                                      Hoist at manwey t

1 1 I l l l _____ _____ Poolwater n' femoved

                                          #h" '#"9                                     Enlarge existing holes
                              '         '                                                                                              t
                                                                           -[ for access        - pave tsfore leaving
                                                                                                                                       ]

3 Y 9

4. . ,

i  !

         ,                    -;            i      -
                                                                                                 ,  q
                                                                                                                     .. [              ]

M_ _ _

           -                                     e       ~

9lTd@...Ns W M $ i!.p '

                                                                                          ,WM                 sjy  !s i

Figure 3-14. Access and shielding problems illustrated. { 3-46

                                                                                                                              .i I

w .. . ,

                                             +

more waterJbe allowed to enter-the central pedestal and'

inundate the debris.-

The estimated cost 1of this core catcher installation is v broken down inLAppendix A-3.2. The totals are summarized here: plant Status A B C Total'in $/1000 744 2058' 3445 In plant Status C.there.will'be an additional. downtime

                      -charge thatris not. included here, s;o Table-3-15Lis an.esti-mate of theLinstallation time. This slightly optimistic working schedule requires.about 50 days. so some' downtime must be. charged.

This' wet; rubble. bed' core retention and cooling system does leave'open the possibility of considerable hydrogen genera-tion if operations do'not proceed as planned. This must be considered and~ handled. Type 3'. ' Dry Crucible Since Limerick is an operating plant < the two types of core retention systems described.above are difficult to install. A third type described here is of interest for two reasons: first, its capability of successfully cooling the large hot-core-debris mass without the generation of additional hydro-gen gas or. steam pressure spikes is unquestioned by.most of the technical community as it is based on routine industrial practice; and second.,.it can-be installed almost entirely while the plant is operational, which minimizes or elimin-ates any plant downtime chargeable to the installation. The dry crucible is located well below the basemat of the containment to minimize interference with any equipment used j' in normal plant operation. It is installed where there is adequate room for ample cooling surface and without dis-turbing plant operation in any way except when basemat break-through is done during a refueling shutdown. In most Mark II containment it is not necessary to drain the suppression

      ~

pool. A previous study (Hammond, 1982) examined the feasi-bility of driving the necessary tunnel and caisson under several different nuclear power plants in different soil or rock materials. This work shows that it is feasible at u acceptable costs to do this at a site in either hard rock or wet sandy soil. In the vertical cylindrical crucible type of core-catcher, the water-cooled steel vessel can readily support a heat 3-47

                                                                                                                                                                                  ?

TABLE 3-15. TIME ESTIMATE TO INSTALL WET RtEBLE BED

                               -Step Description                                                                                                       Men     Day Started    ' Day Congleted 1    Engineering, design                                                                                                              --        --                --

2 Plant cooldown and preparation 30 0- 3 - 3 Draan pool and enter area 30 3 5' - 4 Install shields cranes, access ' 30 5 10-5.' Enter basemat level 30 10 12

6. . Drill access holes in drywell 20 12 .18 -

7 Center wetwell man access 10 12 18 ' 8.. L Install p iping d ist. system 30 18 24 9 Install piping to exterior 40 18 29 10 Install water well in pedestal 15 24 .30 J 11 Install shelter ring in pedestal 15 30 32 12 Add rubble bec 12 32 37 13 Remove debels, p lugs , etc. 4 18 40 14 Leave basemat and block holes 10 37 42 15 Remove shields 10 42 45 16 Install f uel control cone 5 45 47 17 Door barrier. 5 45 47 18 Leave drywell 19 Clean area 10 42 47 These tipos do not include Sunceys or holidays and include no time f or unf oreseen conp l i cat ions. 3-48 H l

}. .- I flux of about 250,000 to 300,000 Btu /h-ft 2 through the wall when the cooling water has a sufficient velocity that film boiling does not develop. The cooled wall in contact with the core debris must have sufficient area to remove the total 45 MW of thermal energy, or 45 x 3.4 x 106 /250,000 = 615 ft2,

!               which would be provided, for example, by a cylindrical cru-cible 4 ft in diam and 50 ft long. Actually the design and costing are for a truncated cone-shaped crucible 6 ft in diam 3

at the top, 3 ft in diam at the bottom, and 68 ft long to 2 allow for easy entrance of the molten mass. Experience shows that such a vessel would have a thin crust of frozen uranium oxide against the cooled steel wall, and heat would be trans-ferred to the crust by thermal convection of the molten l material circulating in the center. Calculation shows that the probable convection velocity near the crust on the wall is less than 1 ft/ min. For the vertical crucible type of core-catcher, there are no uncertainties of location, stability, or cooling adequacy as there might be with other mechanisms. The presence or absence of molten steel and/or water above the oxide layer does not affect its performance. As the rate of decay heat release falls, the solidified crust adjacent to the cooled wall becomes thicker, until eventually the entire mass be-comes frozen. l

+               At Limerick the central pedestal area at the basemat is filled with suppression pool water. This must be blocked off so the area is dry and core debris can drop down through the central diaphragm onto the first core-catcher steel l               barrier without contacting any water and without generating any hydrogen or steam. The hot debris will readily melt g               through a succession of thin steel barriers and drop into the lower cone. This cene is water-jacketed and supplied with forced circulation to remove all residual heat. The cooling water will be pumped and cooled by the dual dedi-cated heat removal system described in Section 3.4.1 using water frcm the suppression pool which is returned to the pool after it passes through the core catcher / heat exchanger and probably through the upper drywell sprays. Adequate water will be circulated to remove the residual core heat, and it is of little importance whether the cooled water picks up this thermal energy in the cone jackets, the upper drywell sprays, or cisewhere. If part of the debris is retained en the diaphragm floor, the sprays will pick up the heat.

This unit requires a hole about 6 to 8 ft in diam through the basemat, which accommodates the upper section of the core retainer and provides separation of this vital heat exchanger 3-49

9 from possible violent internal events that may have initiated the severe accident situation. This is shown in Figure 3-15. It is simplest to fill the caisson volume around the jackets with suppression pool water, and'this adds to the containment inventory. This can be done since the caisson is steel-lined and capable of!the same internal pressure as the containment itself--it is an extension of the containment. Installation of this core-catcher cone at Limerick is much

   . simpler than that shown in the referenced study,(Hammond, 1982) because the basemat area is not so radioactive as in a PWR and shielding can easily be installed. In a PWR'the         '

basemat plug is very radioactive, .so plug removal is diffi-cult and expensive. At Limerick the material can be broken up and moved out of the access tunnel--a much less expensive operation (Figure 3-16). The estimated cost of this installation is detailed in Appendix A-3.3. The totals are summarized as follows: Plant Status A B C Total in S/1000 2295 15,554" 15,670* 3.4.4 ' Containment Ventinc and F11terinc It appears that venting the containment is desirable under several possible mitigation scenarios to avoid containment rupture and the complete loss of radioactivity control. In some cases the vented steam is clean and can be released directly to the atmosphere; but in others, steam condensa-tion and filtering of the effluent noncondensibles are necessary. Venting can involve essentially c3ean steam when a large pipe break occurs under operation where the reactor may be operated.for some time at between 15 and 30 percent of full power with no steam usage, i.e., containment pressure will promptly increase to the point whe're venting is required j (see Figure 3-4). This ATWS steaming rate is an estimated l 720 lb/s. A cost estimate is made for this clean vent only system (Cybulskis, 1982). This can be vented to the stack as illustrated schematically in Figure 3-17.

    *Per reactor based on a dual reactor installation. For only a single unit, add approximately $5,000,000, l

j l i 3-50  ; I i t

                                                                                                                                                          =

-l _ y

l. ,

_1 - p Pedestal

                                                                                                                                               $     To Pool                                Diaphragms (2)                               I     sprays water                              -
                                                                   \             u                               . 6           .x
  . I '.                                                                                                                                                        us, present Basemat                      holes w                   -
                                                                                                                            * , g*. *w e ,,.+ . ., *+. * . . , * + .
                                                          .ga'gbu#,g***                                                  ,
                                                                                                                       ..4 p                                                                .e.. en                                                   .
                                                                      .+,
                                                                        *g                                             e..
                                                                                                                       .e*              Lined like the Core             *[                                             .*               containment catcher          *

N. . eo

                                                                      '.* , ]                                          ,' *y

• Suppression

                                                                       -*                                                              pool water
                                                                       ': t,
                                                                                                               )

W*

                                                                                                                   -+s..-.
                                                                       ++r,                                              n.+,
                                                                                                                         "+
                                                                      .'g,-  ,
                                                                                                                       ,,.1,...,,,,+.,,.,,,.-
                                                                                                                         , . . . . . , , * , ***+ '
                                                                                                                                 ,m e.u

• I a;

g4-en* Pump and hex room

                                                                           ,                                    x        1 -->
                                                                        , s 7lr1 l lsolation valves 1-g4
                                                                       +                                                                                    /
                                                                                                      .b 1,
                                                                        .,.   .                 ,~          '           .f s, a, ,V*
                                                                                                                                  ,, *.",e:",,, *a
                                                                                                                                                 . . ',..*,*i
                                                                      ,,.*                                             +,,,
                                                                       . * *, . 's . a e ' c s ' u + * *, io . ,* $ '
                                                                              .. . . . , * . . , *. . + n , . * : . -

Figure 3-15. Schematic - detail of dry crucible

• core debris retention and underground cooling unit.

3-51

5 o i i 1 i i i J i

                                                      ...                                                                                                                                                                      i
                                                        \...

s .... . N . .. E s - ..,

                                                                 \

3 C \........ .. ~

                                      >=    m                         's . . .:                                                                                                                       .C,                      }

s.*....

• e y l C

0

                                                     .,..,                              y   1; . . ,                                                                                                        J
                                                      =, ..                              \ .. .                                                                                                        y
                                                                                            \*               .,

W

. :. .'. ". . s

                                                             .                                    s,

.. . .. s E W

                                                      . . . ... ., ..., .                              s .. .                                                           e            -

W n..... .. . . . .. . - s . G gs .;.

                                                                                                                                                                   *1 7 .G.

F

                                                           ..e.....              . ..          *
                                                                                                                   \ *. .. . *,
                                                                                                                                                                   'E 4 *g A

U . i

                                                         .:P ... .P. . .l.      * * *.
                                                                                                                              . ....

• t

                                                                                                      . .                                                          *C 2P                                                ,
                                                                                                   , .. \                  .\
                                                                                                                              \.   ......*.  ..                                 I.

y. .. .:*: .

                                                                                                                                                                                                        "o u
                                                  .: v.v::::
                                                    . .     :.:              . .::::     .     :::::c :. y     . :.w. . \ s .                                                         *. .. .bp                                l
                                                                        . . .. . .... . .. ;. . . . :/ ..: , .. k. i :. . .
                                                                                  .. . . .. . . ,.. s- \                                           a
                                                                                                                                                                  *-                                  t
                                                                                                                                                                                                       ,J
                                                          ' ' * *                                                                               \,*:-         .

,'*:,.,*.;t.'"..-I..'"*...'..\ ..

C 1

                                                                                                                                   .  *                       ....t.
                                                                                                                                                                                                            -                  i
                                                 ,*. .=. .. .: . ...,*, c: .
                                                                                              ...n.'. .t....

m m I

                                                                                                 . ..              ..         ..........\
                                                                                                                                  .\
                                                                                                                            .                                                                                                  i
                                                                                                                                             .. g                  . .                                M                        ,
                                                          * . . : *- = . . . . . . * . >
                                                                                                                             . . ..\
                                                                                                                               .                 . . ......s                        .

m w v

                                                                                                                                                                                                                               \
                                                                                                                           ., ..::.    . .. .\... '. \

1:

                                                    .. .....                                            ~

• u

                              ~                 ::.:                                   :., .: . :: . .. ..                         . .... * * . .. * . \
                                                                                                                                                                            ,..                        g C                         *..' . . . .,.
                                                                                                   . .                           . . . , f,4 \
                                                                                                                                                                                             .                          l
                                                                                                 ..               .                ..                                                                  g
                     *t
                                                             . . . . ' .. ..?,:. .. . . * . . , * . "- *. li2 ; . *. .

. . " . /.y . . : . . .* * *. . l,. L. ..!' .z

                                                                                                                                              -                                                        C J

3: s ...a* . ,2 %. a .* * *. ~

                                                                                                                               ..:.                                            .**.                    :=
                                                                                                                          . ..   . . ,-.e, .. ..m .......

to 2 , g

                                                                                                                                                                                         ,                                    j
                                                                                                                                       -.-.g,,,             <   ig            ,'.*.                         ,                  ;

i

                                                                                                      ., . g                                                                                             0 o ..:. o.    *m                , <v .* .                                                     1

x. .

{

                                                                                                ' ,** *:, L % % ',.G. = .                                                       .- .
                                                                                                      .                                                                                                    e i                            -
                                                                                                 . ...  : e. y eu
                                                                                                         .        ee                           *.               ;              ..
                                                                                                                                                                                             .         Q
                                                                                                    . /. o*s; T .: ' * ..:         ;.                        ?*:                        6 o
                                                                                                                                            * 'l. :       j
                                                                                                                                                                              ,.l**..                  >

w

r. , .*

s l ; 7, 1 5 .

                                ~.

4 r -- __\ r . ; -

                                                                                                                                                                                             .         1 y_              -%                                                        .
                                                                                                                                                          . w:

m i . 1 - . y 2

       ,7  -._.

y

h. . .

s. . .. .

g .. .

                                                                                                                                                          *.                                           m
                                                                                      = , * .. ... , *..

I g-t. L I m em

                                                                                         .-.*. ...                                                                              =.

m * .

                                                                                                                                                                                      -                                        i
                                                         . ... . ..:                                                                                     5                                                                    3
                                                                                                                                     ,        .          .                      =g                                              i

. T

                                                                                                                                . n. '.e ra e., :
                                                                      ... .. *.* . .e d         .
                                                                                                        >se
                                                                                                                 -             i o*.x- - f.O             x .c, J> .:.

d Q . .~... .. -

                                                                                                        .D c-s                       S2 w-5 3-52 e

___ ________m__m____________w

    .        K

%. , q .;~ - . - l !. } , -f t o'9 e4 i .' g ene " - sb4 l ] 9 g , i 4

                                                                  ,                     e j                                     T             y
                                                                                ,h *

• w i

                                                               ,  {          .               a                     h         l N     ,. y                         a
                                                                                          .4 e
                                                                             .                                    .wt u

s +.<',. sW

                                                                              # m ++,*

4

                              %                               1  ,,               ,*g u# #

f < n C m g ,%, , .* g .< C e m. t

                                                                                 # yp+%+* , , -

u o u u C

                                                                                # ," =        ,u,,6-

• o O

                                                                    >                e *#       H  g H
                                                                    >            \    m   0 50 n               p h       #s q                  E
      <                                                                          g a ,*4 % $                      O
                                                                                 , " = 45!                          2-

• ra. s v e e*
• 4 ( g# M

                                                                                        ,..< . ' 5. ,8              e
                                                                     ,             +,e                              >
                                                                    .                  , + 's g   -

J *,*,,+ ,, s y

                                                                                                                  .e
                                                                    $                e              ,

e=4 9 F,* e*e*

• Q w , y + a q ** M

                                                                                          *#l y   n            .s,.l.*.*.*                    Q "4
                                                                                         *4 4**,   $a*              k
                                                                                     ' 0 ,g                         a m

r

                                                                      .              a.~:<
                                                                                      . . ,           .f m

m y

      ;'                                                            a d

a = *= *

                                                                                       # B e++r                   k a

s *

                                                                    'a                #

• e ,= e Q

                                                            .

• ep* t

                                                                    +,

f

                                                                                       #        ..                .m t

g *4,$ *s g m < #

                                                                                             #**s

• e A w

                                                                                          -       ,                 e 9

v

                                                                     +.                * + ,4                       o a                                **                 s*a   , # w e-          .ll" U

6 w

                                                                    * *                " e tt m %                 W
                                                                      +                *=* *s*
                                                                    >s                    o*,4 ,                e
                                                                    #,                  a    O,S g*

O e

• E

                                                      ~                  5              +g= &

• u " .,*E o

                                                                                             = .                  ~,
          ,                                                         # .                 *,+,s                5,7
                                                                   .-,.                                           r

_s

                                                                           >                        ,2 2 is
                                                                                                    .          =    W
           -                                                       7                                .
                                        ?        ,                         .              .                  y ~5   u
                                                                   >m                   ,,             m<          S x        >      r
                                                                                              ,; . N g

D

                                        \,            *             > * ' *44*a*                                  *"

T-h I , tv * // U 6* eme

                                          \        'W._                      n,
                                           %                                           [N O w        '          _

k *5*a - . . 3-53 l i l l l l l l

• C I

After the main steam flow from the LOCA or ATWS has ter-minated, a degraded core or core-melt accident may occur. While the core is in the RPV, contaminated steam, hydrogen, f other noncondensibles, and highly radioactive particulate and aerosols will be evolved into the fully charged contain-ment. This relatively low flow rate venting must be done , through a condenser / filter system capable of trapping most of the radioactive materials. The filter system would be sized l to condense the steam for a limited period to trap the particulate and aerosols and to pass the noncondensibles. This alternative is of special interest when the 3-A-Fix is installed, reducing the probability of an ATWS event. A cost l estimate is also presented for this low flow condenser / filter system. I i l When the hot core debris mass leaves the RPV. It conceivably could fall into deep water en mass. Knowledgeable technical l individuals are not in agreement about the result of such a l drop. Some predict drastic large steam explosions; others forecast merely a high rate of steam generation from the sensible heat. Another group centends that limiting the i I rate that water can flow into the hot core mass precludes the explosion and controls the steam spike. When this uncertainty exists. our policy is to design around it (Section 2.3). This uncertainty can be avoided in at least three ways:

1. Control the rate at which the debris encounters j the water.

2. Control the rate at which the water contacts the j debris mass.

l

3. Keep the debris mass away from water altogether and avoid both steam and hydrogen generation.

i l These three methods are used in the three different core retention units described in sect $on 3.4.3. l Under these conditions the relief valves and 3ine sizes are dictated by the LOCA and ATWS flow rates, while the filter system condensing capacity is sized by the total steam mass that can be generated by the sensible heat in the core debris mass, regardless of the duratien of the release. The large  ; clean vent plus the condenser and filter system would then ' handle all events except a violent steam explosion--which is avoided by limiting the water depth. The cost estimate of j this ccmbined equipment is also given. I 1 l 3-54 l l 1 i L____________________.. _

 .):

I1 a '. Apparently the Limerick plant currently has neither reclos-

   ,            able relief valves nor a system designed specifically for venting, but dcas have as many as nine different valves           }

(Limerick PRA, 1981) varying in size from 2 to 24 inches in  ! diameter that could be manually actueted to form a vent. This, of course, requires operator acti> st a critical time and therefore does not meet the requirer As set forth in Section 2.3. It is proposed anstead to have an automacic venting system of adequate capacity. The valve (s) (two for redundancy) would be set to open just above design pressure and handle full ATWS steam flow through a diverter valve that sends all the flow to the plant stack as shown in Figure 3-18. Direct venting lasts as long as there is ATWS steam generation. This steam will initially be from the feedwater normally circulated in the high-pressure system. Such water will have a small amount of 16N with a half life of about 7 s and a tiny amount of SOCo from " crud" in the flow circuit. When i venting these gases up the stack, radiation will be within a veptable limits at the plant boundary per 10 CFR 100. In u very short time the circulating feedwater inventory is t exhausted; then the makeup water going through the reacter for the first time will produce steam free from these irra-diated impurities. The relief valve design flow rate is based on 20 percent full power ATWS generation. Once ATWS flow eases, the valves would reclose to avoid unnecessary blo;&:wn and nitrogen loss from the containment. The subsequent lower rate venting of mostly hydrogen and carbon dioxide (IDCOR, 1983) will almost always have con-taminants, so it would be diverted by a valve from the stack to a long cylindrical gravel bed condenser / filter. This component would have sufficient espacity to condense the expected short-term steam flow and ffiter out the particulate matter. It would delay but not remove the rare gases. This filter is sized by the rate of hydrogen and/or carbon dioxide generation (Levy, 1981) and the needed thermal capacity to condense the maximum possible steam spike---75,000 lb steam (McCormack, 1982). After the vent recloses, containment steam will be condensed by the cooling sprays, causing the internal pressure to fall below atmospheric unless preventive steps such as vacuum release are taken. 7nese are analyzed in Section 3.4.6. Specifications for the vent and filter system for a typical Mark-II-type containment are as follows: 3-55 L_____.-_____

                                                                                                                                     -

• e

                                                                                                                                                                                        =    0 l'

a l Steam spike, hydrogen and contaminated gases to filter 4 j l 1 l m l Die - 36 in Value in 'up" position l K I i during ATWS. Radioactivity detection I i drops it to " low

• position I

1 i '

                                                                                                /

From ) ,- l containment 1 , \,, / relief . valve l l m b= J 0 n l l ATWS steam flow to stack - 500 lb/s Figure 3-18. Steam / gas diverter valve.to handle ATWS flow. l l 3-56 s I

    .W        }s I

Relief valve opening pressure (psia). , 70 Closing pressure (psia).  :>50 Number of valves (redundant) 2 Approximate throat size (in.) 24 Rated flow rate (Ib/s per: valve) 360 L Max flow rate.2x design pressure (Ib/s) 1236 Pipe diameter (ft)fte stack- 5 to filter 3 Filter material. gravel Filter length (ft) -1000 Filter area (ft2) 20 Steam condensing capacity (1b) 75,000 Filter _ pressure max (psia) 25-A breakdown of the ecst estimate for the three alternatives is given in Appendix A-4. The totals are.as follows: l p 1. Pressure relief valves only--vented to the stack Plant Status A B C Total in S/1000 1134 1387 1496

2. Low flow relief valves to condenser / filter l

Plant Status A B C L Total in-S/1000 1900 2336 2474

3. Complete system--large vent to stack later diverted to the condenser / filter Plant Status A B C Total in S/1000 2930 3599 3842 As a checkpoint, the specific cost of this complete smaller system with all piping and valves developed and installed is

                         $196/ft3 of filter bed compared to the much larger Swedish Barsebeck filter system at $56/ft3 (Filtra, 1982).

On a dual-unit installation, each containment must have its 3-57

                                                                                                     ?

L__ ___

                                                                                                                   *.   .o
         ~

own relief valves, but- the condenser / filter system can serve both units. 3.4.5 Hydrocen Control The generation of hydrogen gas in the course of a degraded core accident presents two types of problems, both resulting in containment overpressure The first is the generation of excessive noncondensible additional gas-(hydrogen) at a time when the containment volume is thermally loaded._ The socond is the possibility of ignition and combustion of the hydro-gen in sufficient quantity to raise the temperature (and

• pressure) of the containment gases beyond the vessel limits.

The generation of hydrogen results from the oxidation of either zirconium or iron by very het steam. These oxidizing conditions can occur inside the RPV at high or low pressure or as the core debris is on its course to the retention and cooling area (Keilholtz, 1976). Combustion of the hydrogen , may occur at any time after its release if the containment volume is air-filled (Berman, 1981; and Camp, 1983). Even when containment inerting is used (as at Limerick), oxygen may be drawn in when the vacuum breaker valve allows air to enter'in order to avoid collapse by underpressure after heavy steam venting has washed out a portion of the nitrogen (or other'inerting gas). In-Vessel Hydrocen Generation The Limerick Mark !! BWR reactor has 764 fuel elements with 64 Zircaloy clad fuel rods in each element. Each clad tube is 13.35 ft long with a zirconium alley area of 0.052 in2, , This is B.32 ind per tube or 1.95 lb of metal. The total core contains 95,500 lb of zirconium. The exothermic oxida-tion reaction with steam produces hydrogen weighing 4.4 per-cent of the zirconium consumed, or 4200 lb when all of the metal is consumed. The core may be dried out and reflooded , several times during the accident, thus tending to maximize i the amoun't of reaction. The maximum thermal energy produced  :

 . by the process would be about 3 MWh.                                                                                    .

i I During this same period it is also possible to " burn" some of the 70 tons of steel used as core support and shields. This fraction (Silberberg, 1983) weighs an estimated 30,000 lb. The high-temperature steel / steam reaction generates 5.4 per-cent of the iron weight as hydrogen, resulting in a possible additional 1620 lb of gas. 3-58 I i f

l' 5 8 t Tre technical community interested in hydrogen generation generally feels that about 30 percent of the metal involved with the core will be oxidized on any one dryout. If allow-ance is made for one reflooding and dryout, as much as 60 per-cent couJd be subjected to oxidation, resulting in the in-vessel release of (4200+1620) x 0.6 = -3500 lb of hydrogen. This gas release will increase the pressure of a cool con-tainment about 1.85 atm (27 psi). If the core melt occurs after an ex-vessel pipe break or after an ATWS event, the hydrogen generated would be added to a fully charged containment at design pressure (70 psia and 340 F). Gas must be vented to av,oid possible contain-ment rupture. The vented gas will be a mixture of hydrogen, nitrogen, carbon dioxide, steam and contaminants. The esti-mated maximum rate of generation is 350 lb/ min, which is 358 ft3/s requiring a relief valve with a minimum throat area of only 0.35 ft2--about 9 inches in diameter. This may not establish the valve size; other conditions may 1.eed greater capacity. This vent gas mixture will be contaminated, so the approach is to condense out the steam fraction and filter the remainder. The venting duration at this flow rate should be about 15 min maximum. Ex-Vessel Hydrogen Generation The 164 tons of hot, partially melted fuel together with the hot remnants of zirconium, and stainless and carbon steel from the core structure and pressure vessel could encounter hot water or steam during their descent to the core reten-tion and cooling unit (Keilholtz, 1976). The amount of hydrogen generated can only be controlled by limiting water access as there is ample steel available for reacting. This means that on-course from RPV to retention, the core debris mass either must be rapidly cooled as it enters water so that little steam is available for reacting or it must be kept reasonably dry to limit the steam available to form hydrogen. At Limerick it is quite practical to limit the pedestal central area water level. At some other Mark IIs it is more difficult. Hydrocen Control Ecuipment The removal of hydrogen gas from the Mark II inerted contain-ment is not required for immediate mitigation as the low volume condenser / filter vent system handles this condition. However, for the hydrogen gas to remain in the containment after the vent valve has closed produces a hazard to all future operations, and it should be removed. The rate 3-59 l l l l l l

                                                                         ?

a. .e 1

k l of removal should be low to minimize underpressure problems l as' cooling and steam condensation occur; therefore, the-I hydrogen equipment can be small in capacity. .Some plants already have such equipment installed (Henrie, 1972). An estimate is provided for plants that do not. Plant Status A B' C Total in $/1000 2996 34'29 3865 A detailed cost breakdown is given An' Appendix A-5. j 3.4.6 Vacuum Breaker System Once the containment has vented the large flow volume (567,000 cfm) of ATWS steam, the possibility of containment underpressure exists. This vented steam will carry with.it some of the initial containment atmosphere, inerted or not; and after the event has abated and heat is being removed from the containment, the steam will be cor.densed, leaving inadequate gas mass to maintain internal atmospheric pres ~ sure. Although of less magnitude, similar conditions can exist after any low-volume venting of steam, hydrogen and esrbon dioxide takes place. It, too, will carry some of the initial atmospheric gas with it and ac the steam pressure collapses and the hydrogen gas is removed, an underpressure could exist. Containment underpressure is not permissible if containment integrity is to be maintained; therefore, a vacuum breaker valve of adequate capacity must be installed if one does not exist. This inward flow relief valve must be large enough to permit the needed high flow rates with low-pressure drop. The valve must reclose and seal when the subatmospheric pressure no longer exists. The limiting condition is an ATWS venting of several minutes , duration (after the pool is saturated) that vents and washes an estimated 25 percent of the nitrogen gas out with it. Later, when the dedicated heat removal system extracts energy

                     .                     at a 45 MWt rate, the partial pressure of the steam in the containment could collapse in about 90 s. During this pres-sure drop, the in-flow valve must handle about 7500 lb of air        i with a very low pressure drop across the valve.       It should      i have the free flow area of a 26-in diam port.

Actually the Limerick plant already has vacuum breaker valves installed which may be quite adequate, but they should be checked for this use. A preliminary cost estimate of a large flow system is included here .in case they are not adequate. 3-60 i

- p- , x f.

  .j r The design, development, test and certification of this new valve together with a. dual-valve installation including all plumbing and wiring will be:                                           -

Plant Status A B C Total in $/1000 865 1081 1336 A breakdown of these costs is given in Appendix A-6. 3.4.7 other Possible Mitication Modifications There are a number of rather minor modifications that could have a sizable impact on the dominant risks if they are properly implemented. These are very site-specific but

  -             should be considered as they may have a high value/ impact ratio.                    Some of these are noted below:

e Continuous generation of normal load emergency electric power instead of the rapid-start standby system now in general use.

                                         .e  Improvement of certain critical component installations to resist seismic assault.

e Protection of certain penetrations from over- . temperature as a result of a core-melt accident or similar assaults.

 'I                                       e  Improvements in the severe accident emergency operation of equipment such as existing coolers and sprays, doors and seals, and latches.

e Improvement and training of some emergency pre-cedures such as use of alternate fire protection systems or controlled unfiltered venting. 4 e Improvement of operational procedures to minimize some transients. These modifications are not considered in this task but could be very worthwhile. 3.5 ALTERNATIVE MITIGATION STRATEGY A13 modifications discussed so far are considered applicable to the present Mark II containments, i.e., a structure capoble of pressures sufficient to retain all the vapors, gases, and other products of even a severe core melt acci-dent. In this section a completely different containment l 3-61 i

                                                               -        o
                                                                   .1 4

Ia concept is described, wherein the containment and its asso-ciated equipment form a tight but not pressurl=ed enclosure for the reactor. Any releases of steam, hydrogen, or other gases would os directed through a filter system open to the atmosphere and having very large capacity and high effec-tiveness, reven for noble gases. The maximum containment pressure would be only that required to drive the gases and steam through the f! Iter bed, and this overpressure lasts only a fr,w minutes until atmospheric conditions are achieved again. This concept is depicted in Figure 3-19. j This approach holds the possibility of making the contain- - ment t.ystem simpler, less costly, possibly safer and more acceptable to the public than containments that must with-

                                                                           )

stand pressures above their design rating during certain ' severe emergencies (as do most U.S. reactors) and must hold 'g there pressures for extended periods--months in.some cases-- 1 (as at TMI). Providing the containment with permanent com-munication to the atmosphere allows relaxation of the con- . tainment pressure specifications and would result in a Icwer l first cost and increased reliability. The gases and vapors ) leaving the wet well area, which would otherwise pressurize i the containment during an accident sequence, would pass freely into the filter tower to be condensed, adsorbed, or vented. No vacuum breakers would be required in this open system, and an inert atmosphere would not be needed. l Venting a containment through a filter is not in itself a i new concept--the new elements are the large thermal and flow capacity of the filter, the open venting with no relief valve, and the fact that it is maintained at a temperature low enough to delay the noble gas emissions significantly. When operated in conjunction with a suitable set of other mitigation components, the combined system should be capable of withstanding any of the dominant severe accident condi- I tions with minimal external effects. These components are as follows:

1. A large-capacity, dedicated, reliable system for removing heat energy from the containment. This

. may be an enlarged modification of the present equipment, but a completely independent dual system with dedicated diesel power is described for the Mark II in Section 3.4.1 and can be used here for costing estimates. The estimated cost of this dual 45-MW thermal capacity plant is

              $2,770,000 in a retrofit installation (Appendix A, Table A-1).

3-62 l

j ;I -l , iI i ll l)i l1iI i1 1 5.o r l r t i " e

                                                                                                            ,             ;b '                                                         t-f               -                                   l i

V: - f 4

                                               )
                                                                                                                                                    ~

d e I l

                                                                                                                                                                         .@             l i                                                      s~                                                 i
     .                                       , ;                                     r                                                              -                                   h o                                                                                                   c t

c a e

                                                                                                                                            * -                                         h
                                             ,;                                     R                                                                                                   t s                                   .f            i

_ l O w

p:- le w

_- t Y r n

                                                                                           .-           D                                             _

e s, . m

     ..                                                                                              f                                                _

_ n 7f::r - i a

                                                                                            .                                                                                             t
                            .                                                                                                                                                              n r g on                                                                                                                                                          o
     -                        t e cih c

ai & l eu I I b I I k r c e a t a M s 7n n en dr e y nu p ot Cs e o

                                                                                                                                     }         ~I          .

1): '

                                                                                                                                                      >           k                          -
                                                                   -                                               : : [

c {\ (: .+ ..

                                                                                                                                                            .                               i t

a g m w r P-a e

                                                   /j                                                                                                                                       h c

s r

                                                                     )

9 1 N n

                                                                                                                                      \     l g

S 9 1 o r - o - o 3 y li lte3 i g

                                                                             .                                        ita r

an e f i e wso d F e . ig r t n r u t f oa ll f e la wo p/ O i e l hH l F g Tis C( i il F . e d a r G wInw c , lfll ll' l'

                                                                                                               -l' b

h ~

              ~2. A spray' system in both'the.drywell and wetwell-containment sections using cooled suppression pool water. 1This equipment is similar to that described-in.Section 3.4.2. The retrofit
                    ' installation has an. estimated cost of-$860,000 l                     (Appendix A,' Table A-2). In some plants the                                             l existing spray system can be used at consider-                                             I able saving.

3. -Retention and lang-term cooling of the ex-vessel core debris mass after.a core-melt accident. A core-catcher. design to handle ^this hot mass and .

l

                    . avoid steam explosions as well as steam spikes'                                             l is describedifor the Mark II11n Section 3.4.3.

! The estimated cost of adding this alternative to an-existing plant.is S3,445,000 (Appendix.A, Table A-6).

4. Igniters throughout the drywell and wetwell sec-tions to ensure prompt burning-of any. hydrogen that evolves. .These igniters should be located l  !

to minimize the extent of any one burn and to attempt to burn the hydrogen as it becomes available (estimated rate -3.0'1b/s). These igniters should not depend upon plant electric power only. Some units can be powered from the  ; RHR diesel engine generator'or driven by the l

                    . spray water pressure. An estimated 100-igniters at an average installed cost of $3000 each would be recommended ~for a total cost of $300,000,                                              l i

5. A high-capacity filter system with a filter bed of chilled" rock and. activated charcoal to:

1

                     -Condense water vapors, i
                     -Entrap contaminated-particles and aerosol                                                   '

droplets, i

                     -Retain contaminated heavy gases such as iodine                                            !

and the noble gases by chilling and surface 8 j adsorption in the charcoal. As shown in Figure 3-20 and described in Sectivn 3.5.1, this filter bed is shielded by reinforced concrete walls and is uell insulated to maintain the low temperature (--80 F). , Dual, redundant refrigeration units maintain the low temper-ature. Installation of this complete filter system is esti-mated at S2,938,000 (Appendix A, Table A-14) plus about l 3-64 l t f

 .i
                                                              ,O.

weather prosesdon

                                                            /\                                       and solens 7 :-- 3 q I

7 / nsuletion

                                                                                                   - Activeted steronel - 40 T p           ..

3: Wl MC

                                                                                             -Concrete shielding l:n

.. *g ..

A temperature

                                        ~ 960 T                                      ;.               insutstion enti sover chilled basalt rock or                                ,

equivalent ') 0 t -80 F) g.j..j, D* 7 [.IO , Suilding Conteenment

                                                               .                                                            %      ,,gg Rupture

p. l

[ , diaphragm L

Refrigeration plant / (con be underground) '

                                                                                                       ,p        ;              =. Wetwell wee
    .                       /                          \          .
      .                                             [                                            !
                                                                                                                            ~

l i Figure 3-20. Schematic of chilled filter system. j 3-65 1 i

1

S300,000 for adapting an existing large penetration of the containment wall to connect with the large chill-filter, This brings the total cost of this retrofit open-type i

installation to S10,613,000, which compares to $13,712,000 fo'r the same mitigation capability using a pressurized con- j tainment. If other core retention and cooling methods are l I used, both system costs will be reduced (see Table 3-17). Any new BWR installation would undoubtedly be a Mark III l type which is treated in Chapter 4, so the Mark II plants I described here would all involve retrofitting if the unpres- . l surized containment concept is to be considered. Actually, q very little equipment inside these plants requires a change-- l ' I mostly valves and seals can be removed, locked open, or (in l the case of seals) just forgotten. Once the large chilled { filter building and the interconnecting duct are operational, I however, a number of definite advantages over the current heavy pressure design are manifest: j

1. The risk of containment rupture and contamination leaks or spills is almost eliminated.

2. Many of the presently critical valves are unnec-essary. Many of the penetrations now have double isolation valves whereas only one and sometimes none are required when the open concept is used.

Pressure relief and vacuum breaker valves are no longer needed. All of these items are costly, require continual checking and generally reduce overall plant reliability.

3. Since the existing containment and reactor building were designed for pressures 4 to 6 times those now expected, many testing and maintenance procedures can be reviewed and relaxed. Obviously, the existing seismic and internal missile capabilities are unchanged.

4. The several hundred penetrations through the containment shell can have reduced pressure and leakage requirements. All of this equipment will be subject to a light overpressure only during accident conditions with a duration of a few minutes.

5. In these retrofit installations the reduced maintenance will result in a cost saving.

Operational costs will also be reduced because the equipment in the containment will be more 1 3-66 i f L_-___.

       .                            - _ -_=       - _ . _. - - _ .    --       ._

h fe V i accessible without the requirement.of operating with an inert atmosphere.

6. Hydrogen generation during a core-melt type o accident is readily handled by'the installation- l of igniters throughout.the critical ~ areas. When j oxygen is available_the hydrogen burns to steam -'

                                                                   .which will be condensed.in the suppression pool, byLthe sprays, or.by the filter rockbed. No-
                                                                   . containment pressure is- generated.                                                 Ignition.

systems are. generally less costly to install and to maintain.than alternative hydrogen controls. 3.5.1 Mark II Retrofit Chilled Filter System  ; a When the open vented containment concept described in Sec- l tion 3.5 is considered.instead of the present high pressure' i closed systems,,it is imperative that.the filter be effective ,

         -                               under all conditions since it is the only component barring                                                                     j release of radioactive materials. This filter must have a high decontamination factor for condensibles, particulate and aerosols as.well as a good retention' efficiency for the
                                      .noncondensibles such as the iodines, kryptons,Jand xenons. .To                                                                    j achieve this effectiveness, a large vent / filter unit is pro-vided where-the' steam, gases and particulatas pass in series                                                                   3 over:    (1) cold rocks where the steam in the entering gases-                                                                 J condenses and drains back to'the: suppression pool. (2) very c                                cold rocks (--80 F) that condense.and freeze any-remaining                                                                      i moisture in the chilled noncondensible gases, and (3) very                                                                      1 cold'(-80.F) activated charcoal or equivalent that retains the contaminated noble gases.as well as the iodines by sur-face adsorption.               This charcoal section retains these gasep~                                                      ,
  'i' as long as low temperature is. maintained. Clean nitrogen,                                                                      '

oxygen, carben dioxide and perhaps a little hydrogen will ~ vent from the filter at the top through the two-way isolation doors. I'

                                       ' Figure 3-20 is a schematic of such a filter system. It is shown here vertically alongside the reactor building, but it                                                                   )

could be laid horizontally underground (for radiation , shielding) to extend to one of the cooling towers or take j

                  ,                       any other practical configuration. Preliminary design shows                                                                     '

that the -120-ft-high vertical filter stack may resemble a truncated cone with a ft-thick concrete wall for shield- )' ing and 2 to 3 ft of insulation to hold the low temperature economically. The diameter is -25 ft inside at the base and

                                          -10 ft at the top. This tapered silo holds two distinct types of filtering media. The lower section contains rock p e bble s ., in a layer about 80 ft deep, containing -1000 tons.

1 l-3-67 i

                                                                                                                                                                         )

I l j

1' The available-inlet flow area is about ~190 ft2 The rock can' absorb an. estimated 50 MWht for about an hour and still deliver exit gas below -40 F to the charcoal bed. t The charcoal bed is a key portion of the large unpressurized open containment concept. It is fully described below in Section 3.5.2. A weather shelter with swinging door type , anti-diffusion valves is located on top.' arranged to be L normally closed yet to allow gas flow in either direction at j low pressure drop when conditions demand. l Chilling to about -80 F will be done with two York or equiv- . alent cascade-type refrigeration plants using water cooled' condensing and electric motor drivers. Chilled inert gas I( will be circulated through the filter bed with fans. Either of the dual units of -3 tons (36,000 Btu /h) capacity should g handle the losses of the system, but both units will be used l for " pull down." Each compressor draws 22 bhp, for a plant total of 25 kW of electric power input including fans. It should be noted that the refrigeration plant operation is not required during a severe accident. A separate building l houses these plants. In normal reactor operations, this chilled filter bed is isolated by a thin, low-pressure rupture diaphragm at the large (-9 ft diam.) inlet duct and a light closure valve at the exit as shown in Figure 3-20. Any exigency in the plant causing a slight pressure change, either increase or decrease, will clear away these light seals and permit unim-peded full flow. After such a flow ceases, the light swing-ing doors at the top of the filter close by gravity, sealing the contamination into the filter tower for as long as it is kept cold. Even if both refrigeration plants are inopera-tive, the warming process would require several days. As the filter is allowed to warm up, the radioactive gases will leave the charcoal. If sufficient decay time has elapsed, these could be merely vented at a slow rate. Another possi-bility is to remove some gas from the containment building through the existing off gas treatment system, which would pull air down through the filter from outside and flush the

     . gases back into the containment for further decay.

Since the filter vent path is sealed off as described during - normal operation a normal ventilation and pressure equaliza- I tion system must be provided, and, indeed, already exists in most plants. During an accident this system could be iso-lated, as is presently done, or it could be provided with a small parallel chilled-vent filter. 3-68 I f I

i The estimated cost of retrofitting a complete chilled-vent system to an existing plant is detailed in Appendix A-7. The total cost is $2,938,000. 3.5.2 Performance of Chilled Charcoal Filters The performance of chilled activated charcoal filters is well understood and proven in practice. As is described below, this performance depends greatly upon the operating temperature and on the gas mixture present. The objective in the design given is to retain the heavy inert gases and iodine, and to pretreat the air from the containment so that no other substances reach the charcoal. This means that the very large bed of non-carbonaceous pebbles is designed to remove water and steam, carbon dioxide, and all particulate that could clog or choke the charcoal system. The condensation process accompanying the removal of water vapor is known to aid powerfully in removing such particulate. Another function of the pebble bed is to remove heat. The bed is kept at a temperature of -80 F and the mass is suf-ficient to absorb the heat of a hydrogen fire occurring after the suppression poo3 is already saturated. Thus in addition to drying the air flow, the bed ensures that radioactive gases are precooled before entering the charcoal. Under the worst-case hydrogen fire this entering temperature is below - 40 F, and would be even lower in most accident sequences. Although the volume of charcoal is sufficient that carbon dioxide would not compete significantly with the noble gases, since the rest of the mitigation system is designed to mini-mize concrete attack, a retreatment of the pebbles with a suitable alkali could reduce CO2 to a low level. References to the surface adsorption characteristics of krypton, xenon and radon on low temperature solid media are profuse and extend back more than half a century. A recent review of this work as applied to fission-gas absorption is given by Moeller and Underhill (Moeller, 1981). A process design patent by Thomas (1974) illustrates a practical operating system and Pence (1981) describes actual operating experiences with filters of this type. Several concerns can supply this carbon material on a large-scale commercial basis although our very large requirements may tax their facilities. The actual carbon material is not expensive--well under S1/lb. The behavior of the heavy noble gases at low temperature can be summarized simply: when a carrier gas containing some 3-69

__ __. - _ _ _ _ - - - _ - - _ _ - . - = _ - -- . _ , ,_ _ . _ _ - - _ _ _ R g-4 .,- heavy' noble gas flows over the chilled. carbon' bed' surfaces Lthese gases become." sticky," and move slowly across the

                                           . surface, being continually adsorbed ~and desorbed. When.the
                                           - solid medium:is activated carbon, with its enormous surface
                                            . area,fthe movement is extremely slow.. At even -5'F in'a.

large-bed the noble gases become so immobile that the bed can be evacuated of other gases.if desired. .Such a.means has been used to concentrate these elements; _W ater vapor reduces , but does not remove:the capabilities of the activated carbon, by blocking some of the macroscopic pores with liquid. How-ever, at our. temperatures where the water vapor pressure is - less than~0.1 mm Hg the liquid phase cannot exist. l The technology of iodine trapping has been. equally well explored. Recent work has clarified the nature of the com- t pounds that are likely to be volatilized, and the means for I trapping them. See Bellamy (1974). These volatile iodine compounds behave very similarly to the noble gases on chilled activated carbon. Special treatments for such adsorption beds have been found that will ecmbine chemically with the kl iodine.to reduce desorption. See Kovach (1982). The heavy inert gases that should be retained will be essen-tially immobile at the low temperature encountered in the

                                           ; dual-zone deep activated carbon filter bed. The economics of                        g the low pressure' containment with this chilled filter are so                     .

favorable that extra pebble bed and activated charcoal can be used to ensure low flow velocities and ample capacity. ,- j These two chilled filter beds.will be maintained cold by circulating a gas through them that has been precooled by a refrigeration plant. This gas should be one that has no tendency to " load up" the activated surfaces. It appears that dry nitrogen is suitable for this. Once the filter beds and structure have cooled down to temperature by circulating chilled nitrogen through them, the system qualifies as a

                                                " passive" mitigation, since the thermal capacity is so large that the refrigeration system need not be in operation during an accident.                                                                            )

l 3.6 DISCUSSION OF MITIGATION OPTIONS The mitigation components described in this report can be combined in various ways to create a modified containment system designed for specific threats. The basis for making I such selections and combinations depends primarily upon cost effectiveness--the amount of risk averted versus the cost of the system--but other factors may also have a strong influ-ence in the final selection. These factors may include 3-70 L  ;

4 i accident recovery cocts, local site effects, insurance costs, financing costs, and engineering judgment. The influence of these factors is discussed in Section 3.7. For the Mark II containment system, the mitigation compo-nents necessary for the dominant failure modes can be gracped into five major functions: a dedicated suppression pool cooling system, a drywell spray system using the cooled pool water, a core retention system, some type of overpres-sure limitation, and adequate underpressure protection. The components to accomplish these functions were individually described in Section 3.4, and their costs are given in . Appendix A. Various combinations of these components can be combined into " options" for accomplishing a given group of mitigation functions. Table 3-16 lists some possible miti-gation options, together with their estimated costs and expected benefits. The reader can easily make other combin-ations. The selections shown are all of the " standard" type of mitigation component associated with a pressurized containment. The benefits are shown as man-rem averted out to a radius of 50 and to 500 mi. The ratios of dollar costs to man-rem averted are also shown for both radii. The effectiveness of the several components for accomplish-ing the intended mitigation function has been considered, and each design is judged to have a high probability of suc-cessful performance, with the possible exception of the rubble bed core-catcher. Our evaluation of this device is that it will be dependable only if certain assumptions about the behavior of the core debris are fulfilled in an actual accident. There is both experimental and theoretical reason to support these assumptions, but there is a fair degree of uncertainty remaining according to the technical community. The dry crucible core-catcher, on the other hand, is expected to perform with high certainty under all conditions. For a new plant, where the cost of the two core catcher types is close, the choice would always favor the dry system, especially if accident recovery costs are considered. For backfitting to an already built plant, however, the costs are much higher for the dry crucible,

                       . because of the necessary tunnel under the foundation for access. We have no real basis for down-rating the rubble-bed core-catcher at this time, and we have shown the expected benefit from its use at full value. But the uncertainty remains, and only new information can resolve it.

With two exceptions used for examples, all the options shown in Table 3-16 include a dedicated heat removal system for the suppression pool, plus either a drywell spray system or a 3-71

nn._ _ _ _ _ - _ - - _ - - - . - _- ._. - --- _ _ 1 J

                                                                                                                                                                                                                                       .l 1

o'"'L*'$~~C: - a e c.

.... I s, k

o e R A.  ;

                                                       ' 3 Ilvon "O'806og                              ishts                                                                                      '
                                                                                         'hoj                          :       $          ?                     ?        $ 0          $  ** y           2 Q

• A a *
• O g i: 5=

z R. 4. R  ! ~n G z te nue, I z e e

                                                                                                                            ~

m l :R i s a c,

                                                                                                                                                                                               ~

2 : ;n. ] m 23 . o - ,, U h

                                                         'EUO;,

g,""+ 6:03 l e ~,.g . e g e > >aa g2-m 3

                                                                                                                            ~                   x                  ;

eja. ~< N , io,, Y

• OJn n ,* *A140 ajaa=

5 a

                                            !                                                                o         .    $

1 l  : =

                                                                                                                                                                                                        -9 m                                                                                                                           -                     A , ~R 4..

m A

                                                            "o<id,D'Do                           w..

g g , N$ i j$ 3 ,. K gf

                                                                                                                                                                   ; Os
                                                                                                                                                                                    !:                  m ~
                                                                                                                            ~               j                                            g t '. o.

8 i 4% m *" l x * *ka*f;' u

                                                                  ,,"~~                                                     e                                      ,        e a  a'o#n  t.

ns m - O ~ F n9 i l 3'%R  ? ( Uonay,10C.o144m 4 an E e k 4 $ $$ 23 z O loss U Eh lo m

                                                                                                                            ~
                                                                                                                                                                   ;1: a 1

si en se ):

                                                                a n.jdtz y

• g

                                                                                                                                                                                         $              hh e                                                                                                                                                    I C

N x }  ;

                                                                                                                                                                                 !                   . _  e r 7'1       111 -

5 e

                                                                                                                                                                       . -E     4 8    c        i% 1i *       =

i o E

                                                                                                                                                                                  "            3           -

m m w 2l ia - II a-- at vsli 2;53 a 3

                                                                                                                                                                                         =

e

                                                                                                                                                                                                        ==    11 E
                                                                                                                                                                                                              ~$

A 2 i .} e 3 5 g g s$ k  !

                                                                                                                       !                                        i                         -

r s s 5 2 jj 5 l le si i;i 2 >k 15 if *i

                                                                                                                                 !!                                           ss        *
                                                                                                                                                                                                             -I cs                              ;
                                                                                                                                                                                                             *l 3-72 1

i filtered vent system. These mitigation steps are essential for all cases because of the high probability of a site blackout and loss of heat removal. To these are added other Also, all cases except components for the various options. one are assumed.to.be equipped with the ATWS-3A-Fix already  ! planned for the Limerick plant. The advantage of the 3A-Fix can be noted by comparing options A and B. These cases include, in addition to the basic heat removal system, a large reclosing vent to releaso clean steam during an ATWS, a vacuum breaker to admit air after such a venting, and a large hydrogen recombiner to handle any hydro-gen formed later in the accident. The difference between the two cases is the presence of the 3A-Fix in A, and not in B. . Although the cost is the same, and the final level of safety is the same, the risk averted for B (having a higher initial level of risk) is substantially higher, showing the value of the 3A-Fix. However, for both these cases the final ratio of cost to man-rem averted is high so that neither of these options would be recommended. Options C and D are similar except that aOption filtered vent is D has the substituted for the hydrogen recombiner. ATWS clean steam vent, while C does not. As is evident, both these systems have a much higher mitigation value than in A and B, and at a lower cost, so that the final 500-mi ratios of 180 and 225 are well under the reference criterion ofIn

                                       $1000 per man-rem averted, a highly favorable result.

this case where the residual risk is very low, it can be seen that the ATWS clean steam vent is not cost effective--it does not add to the safety level. In Option E the overpressure control system consists of drywell sprays using cooled water from the pool cooling system. This low-cost arrangement has an even higher effec-tiveness than in C and D, resulting in costs per man-rem averted of only $153 at a 50-mi radius and S86 at a 500-mi radius. All the remaining cases shown include the drywell sprays since they are obviously an effective mitigation step. It remains to test whether other components can be added on to this base set at a justifiable cost to mitigate other accident conditions. In Option F the base system of Option E is augmented by a rubble bed core retention system. The costs and the benefits thus added are both substantial, and the final ratio of these choices (S298 at 50 mi and $167 at 500) is still well within the target range. 3-73

3

               . option G includes both a large hydrogen recombiner and an ATWS clean steam vent'with accompanying large vacuum breaker.

The cost per man-rem averted increases slightly, but is still below half of the $1000 criterion. Option G represents a close approach to " complete containment"--a containment highly likely to retain its radioactive contents under any combination of internal and external events.

               . Finally, options H and I are-included to show the effect of substituting a retrofitted dry crucible. core-catcher for the'

• rubble bed type. As mentioned above, a small phenomeno-logical uncertainty still exists for the rubble bed, which -

may cause it to be derated in later assessments. If this l should happen, the limit would be set by the dry crucible with its high certainty. Options H and I show that even with this expensive group of features the final ratio for "com-plete containment" is below the criterion at 500 mi, and very close at 50 mi. In developing a policy with respect to severe accident miti-gation, cost will certainly not be the only criterion. Many of the possible secondary benefits, such as liability insur-ance costs, costs of public resistance, and financial sta-bility, will depend not so much on actual risk averted but on the completeness of the mitigation system--the perception that all accidents are provided for. To realize such bene-fits, some version of a complete containment is required, but it is also required that the costs remain within some reason-able boundary with respect to the benefits achieved. Since we are unable as yet to assign monetary values to the second-ary benefits although we know they exist, we must use the values of the ratio given in Figure 3-16 in a generous way to j serve as their surrogate for the present. These unquantified secondary benefits should favor acceptance of as complete a mitigation package as possible, within the overall guideline, since they are at their maximum value under such conditions. This consideration leads to recommendation of incremental i benefits that would not normally be supportable by their incremental costs, but where the total cost / benefit ratio is I still within, or only slightly above, the $1000/ man-rem i criterion. ' The possibility of operating a Mark II containment in a non-  ; pressurized, noninerted open mode by venting it through a i large chilled filter is a new concept lacking sufficient - technical assessment at this time for recommendation. How-ever, the preliminary study that has been made indicates that large potential improvements in overall plant safety are certain. These gains are listed in Section 3.5. Table 3-17 shows a cost comparison of the conventional system and the 3-74

. j ; '. . .' TABLE 3-17. CONTAINMENT MITIGATION - HIGH PRESSURE vs LOW PRESSURE Options in S/1000 Low pressure Equipment open High prusure Function containment containment with per Option G chilled filter Dedicated cooling 03 + Separated 2770 2770

                                                                                                                                                             + Underground removal Drywell sprays SOY'                    + External feed                                      860                 860
                                                                                                                                                             + Internal feed
                                                                                                                                                     + Basemat rubble bed                                                           3445 Core control                                                                                                  3445
                                                                                                                                                     + Dy crucible ATWS "3Afix"                                            Yu                   Yes
                                                                                                                                                                                                                  -                  1728 g, ,

ATWS clean vent Filtered vent Large H 2 - 3573 pro on omb'n't C targe breaker - l 1336 Unoer Chilled filter 2938 - Botn- 300 - Open containment Impact - cost in S/1000 10,613 l 13,712 50 mun 25,2 0 25,2 0 Value or benefit - 45,064 45,064 man-REM averted 500 miles Impact /value S/ man-REM (50 mi) 404 543

• ratio S/ man-REM (500 mi) 230 304

                                                  ' Based on Figure 3-16 conditions 3-75

L, a a

 ,t o

l .' l A new open' containment,.where each has'a'" complete" mitigation system. The lower cost of $228'versus;S304 per man-rem quantifies 1the advantage of the proposed open system. 'How-ever,-the.non-quantifiable, advantages are probably more important. 3'7

SUMMARY

The BWR Mark II containment system is a well-engineered, compact system,= easily capable of withstanding the design - basis accident for which it was intended. Like most contain-

              .ments, it is. vulnerable to very-severe. accidents that' exceed    .

the original: design basis. However,'this study'shows that relatively minor improvements can substantially cure this vulnerability, and that the Mark II-lends itself readily to . their installation'at moderate cost.  ! Many.of these accidents, and the. subsequent risk of.contain-ment failure, stem from failure of heat removal systems or other safety devices that depend upon electric power being available at a particular place at a crucial time. Much of the mitigation equipment recommended in this work constitutes ,f provision of a non-electrical means of accomplishing-func-tions normally provided by electric motors. It should be noted that having power available.at the plant bus is notL enough--the fault can equally likely be in wiring, controls,- or switchgear in the accident zone. I The analysis shows clearly that the ATWS event should not be tolerated in a BWR--either the "3A Fix" or some improvement c of it is highly. cost effective and should be available at less cost than its mitigation. Other failure modes derive from slower processes related to heating and gas production by core residue after leaving the' vessel. The' equipment for intercepting these failures is also quite readily installed, feasible, and cost effective, according to the criterion used. Although the conclusion is definite that Mark II mitigation is feasible and justified, it is important in decision making to have a feel for the uncertainties that lie within the factors used. In the present case there is little uncer-tainty that the proposed mitigation equipment can be installed, that it wiJ1 function as designed, and that the cost can be determined. It is also relatively certain that in case of a severe accident the containment will tend to fail in the predicted modes, and that it will be protected if th'e mitigation equipment is there. What is less certain is that the aceit nts will have the frequency predicted, and that the consequen 's of failure are as calculated. Thus the situation l 3-76 l -

 ~j.

f L is that the costs and effectiveness of mitigation are cer-L tain, while the' projected benefit is uncertain. Let us examine this area of uncertainty. Both accident frequency and external consequences represent estimates that can never be verified by experience. Although experience can be used in determining individual component failure frequencies to some extent, there are many key steps that will remain judgmental, since the nuclear industry would not survive enough accidents to give the needed data. The only observation that we can offer about ' the numerical values available to date is that an observable trend exists toward higher frequene.ies and severities with centinued study. As to effects predicted, the revised source term will certainly reduce expected doses, but other calculational assumptions appear in need of review toward an opposite effect. However, there is another way of regarding the whole subject of mitigative benefits. This has to do with perceived risk. Although the NRC may be convinced that a given reactor will never harm the public, an opposite conviction by the public will lead to belief that NRC is not doing its job. And one could make a fair case that in this country an agency per-ceived generally as failing in its job is in fact not doing-its job. Thus the benefits of a mitigation system may rest entirely in public confidence, financial security, lower insurance and interest costs, and lack of harassment. 3-77

               - , . - - - - - - - - - - - - - - . , - - - - - , . - - - _ - - , , . n---   ,r,r_-   .-
          .T l

t-I i., I l.l 1-i 1-I

                                                                                                        ' k a

l l' I l l, l' i . I. 1 I l l l

                                                                                                                 .i I

i i i I l-

4; Tei " }? ,

n CHAPTER 4. MITIGATION DESIGNS FOR BWR MARK III. PLANTS-The Mark'III containment design is the' latest'BWR ce4cain-ment to date. Although only one plant.of this type--Grand Gulf I'in~ Port Gibson, Mississippi--has been' granted an operating license at this time, there'are 11 others in the United States under construction or planned. (A BWR/6 with Mark.III containment is also operating in Taiwan.) More- -

over, the future GE product will be a standard plant (GE, 1983) utilizing a BWR/6 Nuclear Steam Supply System (NSSS) l in a Mark III containment known as GESSAR II. The name

     'I            derives.from its description in the GESSAR Modification II.

In this chapter, design improvements aimed at mitigation of severe accidents are presented and assessed. Section 4.1 gives a brief description of a.BWR/6 NSSS in a Mark III-3 containment. Section 4.2 describes the potential contain-i ment failure modes and their risks, while Section 4.3 des- = cri'oes the requirements for mitigation. In Sections 4.4 and 4.5, possible mitigation strategies are described while j' Section 4.6 contains a cost / benefit assessment of various system options. A brief summary of mitigation opportunities is given in Section 4.7.

4.1 DESCRIPTION

OF THE Mark III CONTAINMENT AND ITS l ENGINEERED SAFETY FEATURES. The Mark III containment design can be characterized by the

      ]            Cleveland Electric Company plant at Perry, Ohio, and the l                   Grand Gulf plant at Port Gibson,, Mississippi. Other Mark i          III's as well as the new GE Standard Plant for the future L       l           (GESSAR II), will' differ from these only slightly. The general containment cross section and dimensions are shown in Figures 4-1 and 4-2 with the containment design pressures
       !           in Figu e 4-3. Much of this data is from the GE and the GESSAR II PRA, and may be proprietary.                                     General information on these plants is detailed in Chapter 4 of the Task 1                                                                             q report (Castle, 1984) but pertinent mitigation information about the plant and its operating conditions are included in Table 4-1 for ready reference,                                                                                                      j 4-1

   . -                                                                                                                                                                                                                                                     a' '. -e i      e '

9 i , w i e e

                                .                                                                                                                                                                                                              a I                                                                                                                                 .

m n 'I m B ___1 ___ . ____ r . - x- - n- .-

                                                                                                                                               %                                                                       ,                       f.
                                                           ~
                                                                                          .- ] I 4[i.             .~~M    .
                                                                                                                                                                .4 ms ,.*
                                                                                                                                                                                                                ?'
                       \         .
                                                                                                                                                                      .                                         [.                           .         /
                                                                                                                                                                        ;                                          J' 1

{..

                               , ..]      F'.                   3, . .
                                                                                                                                                                                                 .: /                                      J. ,
                                ,i i                                                   .
                                 ,                               .-                                                                                                                                  m
                                                                                                                                                                     *                                                         =           -
                                *]
                                 .w

[ . . +

                                                                                                                                                                      .i e
                                                                                                           .-Q+.                       .                                                                                                      .
                                                                                                                           ..=
                                ,                                ,                                                            e.                                 .
                                 , .L             f'T            .        ,

i *a M j *.

                                ..       w wm.

r __ Ih=== . .. ..

                               . .? .'
                                .t.

y l I 2.'..

                                                                                       .4*
                                                                                             '?-
                                                                                * * * * * . ;?

s,  :*'...,;..' a

                                                                                                                                                              -E.n

. .'?

                               ..,                               D                *..
                                                                                                                                       , ' . n. . .'.
                                                                                                                                       .:                     4  W         -
                                                                                                                                                                                                                                      .'.?,
                               ,. ;;.         . - .. . , ..'at. k.. :..gr.*:. 'e/ ..:.t. { '?*:y ,e7. ..=
                                                                                                                                            ' . ; . c.                                                                                               .
                                      *

• h. .;. -s .-.

                                       .c ," ;. .u.6.,?c::.* t .; ;: ..;; .,:g; c.s
                                                                        .                                                                         4. . ' ' r. . :t.: rf,.                 . . . . , 2: .y; l                                                                                                        130,000-160,000 ft 3 l                                                                                                        Pool water Figure 4-1.                                    (U) Typical BWR/6 Mark III Containment Ref:               GESSAR II PRA (May Contain Proprietary Infonnation) 4-2 i-I

 ).      .

I s i l l 2.5 ft '  ! ( J s

                  ?                                                                                                                                                                             1                              -

p h old building l

                          =                                   130 ft e
                   /              =

120 ft e r 6.i 3 ft M - . - DrYwell head

                     .                   . ... .                      ... .... .                      a. ...           .1.         .....

f,

                                                                                                                                                            ;       I ,,            ,,,,s.f           cesling
                      .            i;                                                                                                                                   ./                            slab Y.              *%
                                                                                                                                                                 /

J* f Steel

                                    .                 s                  _es                                          me--               .s                           ,                    ,

1.'

                                                                                           "'                      *                                                                       /          containment
                                                             . . . * * *                                                .                 .      .      t
                                                                                                                                                                      *I
                    *j
                                   ;\.' -~

_,f.*.

                                                            -,                                             p                                       j,.[

c

                                                                                                                                                                                             ;'       vessel 1.75 in
                    .                    ;                                                                                         4 ft            ,b               ;,
                                                                                                                                                                                           .          thick
                    ;               ..,.~.....,-                                           p,         q    ,        g                              ,t            .

i .., s k l Concrete j C (,y\ 4 i , i ((, 3 (-) 21 ft Y oh k 4

                                                           .                                                          p                                                                              S/RV
                  ,,                                     ' j. .                      p                                                        -   w ..,                      '

. . : i, discharge

                  '.-<'                                   .,c

• L .(" . line Elv E *

                                                                                                                                                   *:,k.

l-) 11 ft 5 in 1 '--

                                                         .$ , y                                   [e           I          - - - .
                              =_ _ 1;7_ _ :n , . -:            .
                                                                                                         )

9

                                                                                                                                               ,'         . . .gr. .                ..-
                         .: MfLb, y                                                             )                               ,< .     '
                                                    - w:                                      ,.               ,                             1Eli                          -
                            . -                           ;::l:S            ..
                                                                                                ,              ,        .., ::; .:. .n                                         - ,
                  , ,, .I%,                              :"3h .* 2..-. .* 5,                ' "--      '-T           **kf*\ *M L". .*,

r.- 5,,, . .. ir 1.9 ft ---* - -- --- 6 ft 5 ft --= - ~ Suppression --*- 19.5 ft --- pool = nhe l N: Elv (-) 28 ft 7 in Figure 4-2. Mark III Contain:nent Dimensions Reft GESSAR II PRA 4-3 l

  .,f:,

l l'

                 /
                                                                                                                                              \

i. Figure 4-3.

l (U) Free Volumes and Typical Pressure Capability Ref: GESSAR II PRA (May Contain Proprietary Information) { i 4-4 I {

1..', _ [_d : n- . ) .. i TABLE 4-1 GENERAL INFORMATION GESSAR 18 CONTAl *ENTS GENERAL Mene ano %cel GE Bdt/6

                                                   . Electric output - MW (not/gress )              1254/1306 NS$$ Thermal output - MW (rated / max)         3833/3995 Aporos. Efficiency - pet                      ~ 33 REACTOR PRESSURE VESSEL Pressure - test - psig                         1250                          -
                                                               - operating - psig                   1045 Temperature out - cog. F                       550 Design temp. - oog. F                          575 Steam fIou - Ib/h asa                          16.5 E+6 noter circulation rate

• lb/h 113.5 E+6 Size - ID - tt 19.83 well thickness - tt 0.49 Helgnt - Inslee - tt 72.2 Bottom heed thickness - ft 0.71 seignt (bottom head) - ton ~125 volumes - tt3 mater in IFV ~11922 Steam in RPV ~10122 mater in recirculation loops 13200 water in main lines 1218 motor in feed lines 1233 Total coolant volume 26815 Safety /rellet actuation - psia 1120 Relief capacigy- - 1143 psis and 2.606 lb/tt Ib/h B38,900 FUEL--002 Charge inventory - t m s 166

                                                                        - volume - tt 3             525 Zircoloy claccing - tons                        105
                                                                         - volume - tt 3            520
                                                     $tainless steel and inconel in Core - Ib                        ~30000 Stainless steel below core - Ib                ~67000 Stainless steel in control rods - 1b                                   ~34000 Steel in top guides - Ib                        16000 4

PRIMARY COWTAleceENT Type Me-g gli Construction Reinforceo concrete with

                                                                                                       ~1-1/2 in free standing stosI cyiIncer one come Size - Insico - tt Suppression enomber - diam.             120 Drywell - Amber - c l e s.              73 Melgnt - setwell - tt (to come)        ~200 Drywell - tt                           ~77 M AY QNT AIN PROPR I ET ARY INFORMAT10N 4-5                                         ,

i 1

E l TABLE 4-1 GENERAL INFORMATION GESSAR 11 CONTAINMENTS (CONCLUCEO) l PRIMARY CONTA:NMENT (continued) 3 bolumes - f t Orywell free volume 274.000 wetwell free volume - upper ~ 9 50,000 l Iowor ~ 190. 000 )* l wetwell pool volume 130,00C +o 160.000 j Thickness - ft l - casemat ~ 10

                          - walls - drywell                ~5                             4 Pedestai - height - ft (to RPV base)           25.8                               j 10 at casemat - ft               19.5                               -

Pressure - csig j l Design (internal drywell) 30 g j (Internal wetwell) 15  ; (external wetwell) 0.8 Rupture (Internal etwell) 58.3 calculated f Loanage - free volume daily pct 0.5 Temperatures - dog F , Drywell - Initial 135 j

                                - max cesign                330                          g     ]

wetwell - Initial 90 3

                                - max design                185                                j Suppression pool cepth - ft                   ~ 17. 5                             !

Bottom area - f t 2 7450 l Depth after low I level rupture - f t ~ 17 J SECONDARY CONTAINMENT i Type Round duIIO1ng Construction Reinforced concrete Design pressure-psig internal and external low fs LeeKage - free volume daily - pet 5.0 l COOL 1NG POND I j Area - square tt site specific Vol ume - tt3 site specific Temperature - Jeg F ~85 l l Elevation - ft site speeltic l HEAT REMOVAL SYSTEM l Duai electric motor orivon pumps Capacity gpm es:n 7000 Head - psi unknown Nggt gxchgnggrg 2 Tnermal capacity - MW, 14.6 max each l l MAY CONTAIN PROPRIETARY INFORMATl0N

                                                                                          .I 4-6                                     I l

f i' *

 't Safety Features Current 1v in Place Emergency Core Cooling System (ECCS) e Core Spray System (CS).two loops, 6350.gpm per pump at 105 psi, o High pressure coolant injection system (HPIC)

Two 5600 gpm pumps with turbine drives. o ADS--five. e LPCI a four loop, rour pump system at 10,000 gpm per pump at 20 psi. e The so-called "ATWS 3A Fix." Auxiliary Systems e Residual Heat Removal System (RHR) e RCIC--one 625 gpm pump at 1120 psi--turbine driven. e Fuel pool cooling and cleanup system--one heat exchanger of 3.3 MW thermal capacity. Hydrogen Control System o Dual, low capacity, hydrogen burners (-70 cfm). 4.2 CONTAINMENT FAILURE MODES The development of mitigation options for each containment design requires knowledge of the existing systems and how they will respond in severe accidents. For Mark III con-tainments, the PRA prepared for the Grand Gulf 1 plant under the Reactor Safety Study Methodology Applications Program (RSSMAP) gives valuable insight on accident sequences up to

    .                                 and including containment failure (Hatch, 1981). Off-site consequences for a specific site were not considered. For GESSAR II, however, GE has performed a PRA including off-site consequences for a composite site.                                                  In addition, the l                                       NRC staff, supported by BNL, has performed independent evaluations in preparing a SER for GESSAR. This section uses these information sources to summarize the dominant accident sequences leading to core melt, the dominant containment failure modes, and their associated risks.                                                                             This summary is based on the GESSAR PRA and the NRC/BNL review.

4-7

s.; , i  ? 3' - . p, l'

                       .4.2.1     Dominant Accident Secuences'and' Containment Failure Modes L                                                                                                         l The earliest examination of accident sequences'for a BWR/6'                  'l plant, was carried out for.the Grand Gulf PRA.within the RSSMAP effort (Hatch,E1981). The program identified twenty-two-accident sequences leading.to core melt with twelve ~

being LOCA related,:and the rest due to various transient L initiatcrs The most dominant: events.with respect to core l ' melt. frequency (about 65 percent) were transients with k. scram, and failure offthe RER..LCCALevents:with' loss of RHR accounted for 12 percent and-ATWS acc6unted for 15 percent- . of the core' melt frequency. Transients and LCCAs with loss

• of ECCS-accounted for 6 percent. These four sequences  !

accounted for 98 percent of the core melt frequency. i The. dominant total accident. class. frequencies for 3ESSAR- !, L based on'the GE PRA and the BNL/NRC review are given in l Table 4-2. The nomenclature,for'this and subsequent tables is l given in Table 4-3.. Class I transient initiating events, l-

l. denoted by I;, account for 58 percent of the sequences, i- according to the NRC calculations. Of these, 96 percent are L 'due to loss of AC power. Second in dominance are transient l-L s-events with loss of RHR, denoted by II;, which account for 34 percent of the cequences leading to core melt. l From a frequency viewpoint, loss of AC power is the dominant .f initiator of core melt, followed closely by all transients and LOCAs that are accompanied by loss of RHR. These types l of accident sequences account for 92 percent of the core melt L frequencies. l As noted in Table 4-3, containment failure may occur before l

core melt (42 percent) as well as after core melt (58 per- T l cent). There are two basic threats-to containment integrity i l that are considered in the GESSAR-II PRA: , e Hydrogan. l e Steam and/or non-condensible gas generation. l These threats can lead to the following containment failure modes: Y Slow containment static overpressurization (in . the order of hours) caused by either noncondensi-ble gas generation during care-concrete interac-tion, or steam generation following loss of con-tainment heat removal. 4-8 - I

2 TABLE 4-2. GE55AA 18 ACCIDENT CLASS FREQUENCIES M

                                                                                                                                                        /

t.. e 4 e N 4-9 f

TABLE 4-3. GESSAR ll CORD 8ELT SEQUECE 40 RELEASE CATEGORY O E C LATW E I Transients and LCCAS 84 Transients and LUCAS with Loss of heat remolded til ATWS without W V sekeup i IV ATir$ sithout $LC Injection v LOCA outsido drywell

                      $8   $aell break LOCA 48    Intermediate break LOCA LB   Large break LOCA                                                  .

T Transient E Early containment release , I intermediate containment release L Late containment release 1 Some suppression pool scrubbing of releases 2 Scrubbing of all releases until FY moitthrough 3 Continuous scrubbing of all reiseses B Loss of contelnment integrity followed by core damage C Coro demoge leading to overpressure f ailure of contelnment F Fast loss of containment Integrity followed by core damage Q Coro debels is quenched following drywell f ailure A ATWS with loss of residual heet removal Y' Loss of containment integrity caused by a con-tinuous burn. Y" Fast containment static overpressurization (within seconds to minutes) caused by global hydrogen combustion. u Containment dynamic overpressurization (within a fraction of a second) caused by local hydrogen detonation. u' containment dynamic overpressurization '( a [b w , w _;

          .w c                                                   Trom a risk perspective, basemat-penetration was not con-
           -                              '                sidered a failure mode, since it was assumed,that prior rupture of the containmentiby overpressure and atmospheric k>                                                          release would have occurred.'In the GESSAR PRA and the

'l BNL/NRC: staff review,.this.is considered a conservative l assumption in terms of' consequences. From a mitigation;

                                                           ~ viewpoint, however, the containment would fail by-basemat meltthroughLif'the containment is protected-from: earlier overpressure events..
                                                         . Lastly, sequences that bypass: containment'have aJso been         '

given a' low probability in the GESSAR II PRA. These potential-failure modes and their uncertainty are' discussed in the BNL report-(BNL, 1984) and are factored into their uncertainty analysis.'In'the next section, the risks associated with~the dominant accident sequences and containment failure modes dis-cussed above are~ considered. It should be noted that GE' con-siders'its' determination of risk and consequences to be "best . estimate" values.. The NRC. staff,-however, is'of the opinion l that the current state of knowledge permits only a range of-consequences to be calculated at this time. The NRC staff

                                                           ' lower bound corresponds to the GE "best estimate." The NRC upper bound effects are one to two orders of magnitude greater.

4.2.2 Release Catecories and Contributions to Risk During a core melt accident,. fission products may be released into the containment building where a number of systems are available-to mitigate potential releases to the environment. Prior to vessel failure, the dominant sequences, from a core melt frequency viewpoint, are transients that' release radio-active material to the containment through the SRVs. Hence

                                                          -they pass through the suppression pool and are essentially filtered or " scrubbed." For the Mark II containment dis-cussed in Chapter 3, vessel failure leads to molten' core material on the pedestal floor, with potential paths to the environment following containment failure. For the Mark III presented here, Figures 4-1 and 4-2 show that as long as the drywell wall and ceiling are structurally in place, the pool can still act as a mitigative system. Hence the general characteristics of the fission product release path for most Mark III accidents is similar. The BNL/NRC Staff review agrees with the GESSAR II PRA that for most of the core melt accident sequences, drywell integrity will be maintained, and all of the fission products released from the primary system will pass through the suppression pool.

For the highest frequency sequences (Class I transients), the pool will be subcooled and the fission products will be I 4-11 6 t

1 '.

                                                                                                        *~ ~*

I.l Cl . I, l subjected.to significant. pool scrubbing. When the pool is

                                ' saturated (e.g., Class IV transients without: scram), pool                    p scrubbing will'be reduced, but could still be substantial.                    p
                                .The ability of the M rk III containment to maintain drywell' L                                 integrity during a core-melt accident is a significant:fea-                   j.

L ture in terms of reducing fission' product 1 release.. This is.  ! because the drywell, which contains the primary system, is completely enclosed by the steel containment. .In the I l Class II and IV sequences (with less of containment. heat- f  ! removal) containment failure precedes core melt; but unlike

                                'the Mark I and:II containments,1where the core melts down-                *
                                .into'an1open containment, the pool scrubbing function is
                                -maintained in the Mark III.                                                   l' At issue then, is the potentia 1Lfor a hydrogen-detonation in the drywelli which could lead to a pool bypass.                 In the
                                                                                                            'j-GESSAR II PRA, pool bypass is assumed to occur late, and l                                 together with the assumption that water floods the cavity,                   g.
                                -limits fission product release.          Failure of the drywell is           l-considered at the drywell ceiling, and the water above the drywell-is assumed to quench the core / concrete interaction.

In-the NRC staff review, failure is also considered to occur L in.the drywell wall, which allows fission products to bypass the pool, and the core debris is not "ex-vessel flooded."' l Thus pool scrubbing is substantially reduced and' fission I product releases during. core / concrete interactions would continue, following vessel melt through and drywell failure. - Before discussing the release categories and their asso-ciated risks, it is important to note that the BNL/NRC staff review assumed that the maximum suppression pool decontam-ination factors (scrubbing effectiveness) were given in the GESSAR II PRA. The SPARC computer code was used to generate

                                 " minimum values" for these decontamination factors that are severel orders of magnitude lower.          The values used to ohtain the risks were considered mean values or best esti-mate, and are described in the SER (NRC, 1984).

The determination of risk for each containment failure mode, and the' sequences associated with them, follows from a set of containment event trees. The endpoints of the branches ' are then " binned" into release categories, i.e., endpoints whose fission product release characteristics (time, mode, location, etc.) are grouped together. Table 4-4 exhibits the consequences (in terms of latent fatalities and popula-tion dose) for the 15 release categories used in the GESSAR PRA (GE, 1983). 4-12

 -w m _mm__ __          _..m_--               _ _ _ . _ _ _ _ _

[1. t (early) fatalities Examination ofL Table 4-4 shows that acute This' are not. predicted to occur for any release category. i L result occurs because acute fatalities require a thresholdSuch threshold

radiation dose.

with early core melt-early containment failure sequences (e.g., a large-LOCA with a steam explosion inducedSince contain-ment where evacuation is not effective. eitherfailure) the suppression pool is sufficient'to scrub or the a poten- time tial release-(reducing it.below the threshold) (allowing effective evacua-for hydrogen formation is large tion), acute. effects'are essentially eliminated. As mentioned above, the NRC staff considered the possibility of suppression pool bypass, as well as a number of other. factors which could influence the consequence calculations, and hence the risk. These factors include:

a. Changes'in source term characteristics and behavior in the primary system.

b. Changes in the scrubbing efficiency of the suppression pool.

c. Changes in the core / concrete interaction models and assumptions.

d. A change in the site characteristics (GE used a composite site, the NRC staff used a specific, high population site).

The factors discussed above, as well as the potential for the uncertainty in deter-suppression pool and bypass highlight risk averted. (See next hence potential mining risk, section.) The NRC staff considers the GE consequence calculations as low range estimates, and their own calculations as high range Because of this, the NRC staff has only calcu-lated-consequences for selected sequences, as shown in Table estimates. The sequences were chosen to illustrate the effects of 4-5. plume characteristics, the timing of the release and the evacuation assumptions.. Two important insights are gained when examining Table 4-5. First, the 1-SB-El (an early containment exhibits early fatalities. This failure with minimal suppression pool scrubbing) , occurs because in 5 of the 91 meterological sequences sam- J pled, a small portion of the exposed population wasWhen calcu-1ated to receive bone marrow doses above 200 rem. 4-13

   ,f TABLE 4-4      GESSAA la CONSEQUENCES 6Y RELEASE CATEGORY Release         Containment          Freopency       Hean Latent       Mean Person Category        Failure Mode      -(per yeer) .       Fatalities         Ree (c) l-T-L3          Y(a)                                                2.0 x 10 l-T-E2          9'

• 7.0 m to I-T-O Y' 6.9 x to 1-7-12 V' 6.9 x to 1-Y Y. 5.4 x 10
• l-T-L2 Y,Y'(a) 4.9 x to ,

I-$8-L3 Y (a) 2.3 m to 1-LS-L3 Y(a) 3.6 x to I-58-El p '(a)  ; f.3 x 10 1-58-E3 Y" ' 8.1 x to I-58-LI Y, Y' $.8 x to 11-T-83 Y (b) 1.1 x to ll-L8-83 Y (b) 7.9 x 10' Il-A-83 Y (b) 8.6 x to IV-A-F3 Y (b)

• t.S 10 e

The nomenclature can be found (a) Following core melt. In Table 4-3. ee (b) Preceding core melt. May contain proprietary Inf ormation. (c) Dose calculated to 300 miles. i TABLE 4-5. *C STAFF RISK RESULTS FOR SELECTED ACCIDENT SEQUENCES 1 l CDNTAI M NT 8ELEASE FAILURE EARLY LATENT PER$0N FREQUENCY CATAGORY KCE FATALITIES FATALITIES REM (PER YEAR) a l-T-120 9'ta) 0 2.7 x 10 175 1.2 x 10 1-T-O Y"(a) 0 170 2.7 x 10 8.4 x 10' l-T-L3 ,Y (a) 0 40 0.7 x to -6 3.3 x 10 , ATWS Y (b 0 400 6.0 m to 3.1 x 10

                                                                                                               -6 l-T-12              y'(a)                0              480           7.6 x 10         2.8 x 10' ll-T-83             Y (h)               0               300           5.0 x 10         5.7 x 10 1-5B-E l            u'ta)            6x 10-3           600           9.0 x 10          7.5 x 10' Sources (NRC. 1964) and (NRC. 198 5) .                                      (e) rollewing core mit.

(t) c rece !m; core % t. I i

n___

         .)      <

k I i  ! l l statistically averaged, a small probability of death shared j

                     ,   by a small number of people, yields a small fraction of a death (0.006).

Second, the population dose, expressed in person rem, as calculated-by the NRC staff, is considerably higher than that l calculated by GE. The early release category (1-SB-E1) j differs by a factor of 7, while the release categories are ] two orders of magnitude apart. In aII cases, the NRC staff j calculates the higher values, which are considered " upper i bound" values. 1 4.2.3 Uncertainty in Suppression Pool Scrubbino - Most accident sequences result in the suppression pool acting as a scrubber of radioactive materials. Unfortunately there is a high degree of uncertainty about its effectiveness. , Further, estimates of its effectiveness vary from the low decontamination factor (DF) of'six used by the NRC (NRC, 1984) to the high of ten thousand used by GE (NRC, 1984). These results in very different estimates of risk. { 1 The basis for the large DF used by GE is open to question. Review of their work shows that many aspects of their modeling is empirical and that important physical phenomena have not been simulated. When the pool is subcooled and the gas flow from the core is mostly steam, the pool DFs will be large enough to scrub out essentially all the radionuclides. Near saturation or when there is very little steam being condensed the pool DF becomes uncertain and very particle size dependent. Some experimental evidence (Wassel, 1984) indicates a strong dependence on gas flow rate. Since the dependence on gas flow rate is a result of how the gas and aerosols enter the pool, quencher characteristics become important. Experiments with prototypic quenchers have not been conducted. Aerosol particle size is probably the most important param-eter and the most poorly defined. For particle sizes en the order of 0.1 micron, the bulk of the available data show a DF below 7. The question is how well one can predict the core effluent flow rate and the aerosol particle sizes reaching the poo2. The process by which radionuclides aerosols evolve from a degrading core is not understood nor is there much chance that it will be understood in the near future. Particle size is the most important variable. For a change in particle size from 0.1 to 0.2 micron the DF changes from 3.2 to 6.5. At present we have little knowledge of what size distribution to expect. 4-15 ) l l ___ _ - - - - - - i

   -j-                                                                                                  . .
       -)

Steam mass fraction seems to play an insignificant role unless it is larger than 0.15. This means that one needs to look carefully at the in-vessel thermal hydraulics to deter-mine the steam mass fraction and if it is between 0.15 and 0.5 the calculations must be accurate if meaningful DFs are to be obtained. It is not possible to do such calculations

                                        ,with codes available today.

Other uncertainties stem from the fact that radionuclides deposits in the pool will lead to heating the pool. A pool

• that is eliminating energy.by steaming will have a surface temperature that is at saturation. Strong mixing could
• result in an isothermal pool which means it will be slightly subcooled below the surface. If,,however,-the pool is being heated by radionuclides scrubbed from the core effluent, then it will be slightly superheated below the surface, which will reduce the DFs.

GE argues that the most'important path taken by the radio-nuclides is through the SRV because they evolve before vessel failure. During early phases when this is occurring the core may not be molten. Slow overheating and degradation may well p lead to small diameter aerosol generation and low DFs.

Another aspect of modeling a degrading core is the in-vessel DF. Values used in GE risk analysis range from 1 to 99.

This makes a very big difference in how well one must know the pool DF. With the present state of the art it is difficult to argue l that one can use a DF much larger than 6. Using a value of 6 I results in a much different risk picture than that presented in the GESSAR PSAR, as evidenced by the results reported here. 4.2.4 Summary Before addressing the mitigation requirements for BWR Mark III containments, it is important to differentiate ' between several of the containment failure modes. There are two types of sequences leading to slow overpressurization: those where the containment fails prior to core melt (Class II and IV ATWS) and those where the containment fails after core melt (Class I and Class III transients and small break-LOCAs). This differentiation is important because in the first case, the steam generated is " clean" (i.e., rela-tively free of radioactive material). In the second case (containment failure after core melt), the steam and/or non-condensible gas is " dirty," i.e., is contaminated with 1 4-16 l

r, 1 f , . .j '3,a- . *' j 1 4 L 1 - I radioactive fission products. 'Hence mitigatien for these two i cases might take very different forms. 1

                                                                                                                        -)

l Table ' 4-f.E shows . the dominant failure modes.for the Mark !!! I containment with this differentiation based on the GESSAR PRA. The: dominant contributors to risk using the GE values are global hydrogen combustion (68 percent), global hydrogen

                    . detonation (26 percent), slow overpressure after core melt (5 percent)'and slow overpressure before core melt-(1 per-1 cent).          The NRC staff values yield global hydrogen detenation (54 percent), global hydrogen combustion..(23 percent), slew
                    . overpressure before core melt (21 percent) and slow overpres-sure after core melt (2 percent).                            It should be emphasi=ed            ,

that containment failure before core melt is due primarily to the Class IV ATWS' events. Examination of Table 4-2'shows that the NRC staff considers the ATWS events to be a rela-tively larger contributor to core-melt frequency than GZ (10 percent-vs. 0.2-percent) in addition to its absolute. value.being a. factor of 300 greater. In terms of mitigation, the relative ranking of these con-tainment failure medes, as well as their absolute risks, will differ depending upon whether high or low range esti-mates are used. These differences, however, become less important'within the concept of complete mitigation. T 4LE 4-6. CONTRIBUTION TO RISK BY CONTAINMENT FAILURE WODE FROM GESSAR ll PvtA E WC Person Rom /Tr Percent Person Rom /Yr Percent Moce Y: Slow overpressure 3 21 21 before core mit 2.4 x 10 1 Th Slow overpressure -2 2 2 efter core melt 1.4 x 10 5

    ~

Y": Globel nycrogen .g 23 comoustion 1.8 x to 68 23 p-: Local nycrogen ---- - - - cotonetton --~ u's Global hycrogen -2 54 cotonetton 6.8 x 10 26 54 100 100.3 100 TOTAL 0.26 4 - 1.1

           ,                                                                                                        i 1

w =--- . . - _ _ _ _ _ _ . _ - . -

1 l' l l 4.3. MITIGATION REQUIREMENTS TO BE '

  -The preceding section showed that the dominant failure modes for the GESSAR II. containment result from.a limited set of severe accident sequences (primarily transients), which in turn produce a set of accident end-states or physical condi-                 a tions wherein accident preventive measures have failed and containment failure occurs. For any.one. sequence there is more than one possible end-state, .but'the total number of possible end-states is. limited to three: failure by.a                .

hydrogen burn or detonation,-steam generation or nonconden-

• sible gas generation and basemat penetration. A fundamental l premise of the present project is that. effective mitigation j requires the capability of intercepting the progress from j all of these end-states toward a containment failure, since 4 the act of foresta111ng one mode of containment failure will not stop the accident in progress, but will instead increase the. probability of some other mode occurring.

Thus, a mitigation system that covers-all dominant end-states is favored over partial. implementation, within the limits of cost-effectiveness. Only when there is a high confidence that the accident is arrested and the containment stabilized without failure, can full credit for any partial mitigation be taken. Listed below are the most probable accident end-states for the GESSAR II type containment.

1. In-vessel hydrocen ceneration. This end-state presents the possibility of hydrogen combustion or detonation since the containment is air filled.

2. Ex-vessel hydrocen ceneration. This end-state results when hot steel and any remaining Zircaloy in the core debris mass contact hot steam and react, adding noncondensibles to the gas loading in the containment, and/or leading to combustion or detonation.

3. ATWS steam ceneration_. This end-state results in an energy-charged containment system (i.e., the suppression pool) and overpressure with a con-tinued high steam generation.

4. Ex-vessel steam,cenerat.4on. This short term pressure rise occurs when the hot core debris encounters water. The long term pressurization 4-18

 ,4s.                                           o i
 >                                                                                                                                                                   J L

occurs when residual heat in the core debris mass I I is released to water.

5. Containment concrete decomposition _. This condi-tion results in steam and carbon dioxide (non-condensible gas generation) to add to an already

                                                                                        . pressurized containment.

6. Residual heat load. This condition occurs from the. radioactive decay energy of the fuel, and.

results in a containment pressure buildup with time. To control any or all of these accident end-states, a com-plete mitigation system must be capable of the following functions: 1

1. ' Hydrogen generation--both in the RPV and subse-quent ex-vessel--must be dealt with to preclude

                                                                                       , combustion (above -5 percent hydrogen volume) and possible detonation (above -15 percent). The Mark III is not inerted in normal operation.

2. Containment venting with overpressure relief valves to release relatively clean ATWS steam to the atmosphere and a diverting system to pass subsequent smaller flows of contaminated steam and gas through a condenser /fi'.ter system.

3. Vacuum breaker system to prec12de containment underpressure as steam is condensed in the containment.

4. Adequate long-term heat removal from the contain-ment during the accident and as long as residual heat in the core is being generated thereafter.

5. Core debris mass control during its course from the RPV to a long-term retention area.

6. Adequate long-term cooling of the core debris mass once it leaves the RPV and after its arrival at the holding area.

7. Missile shields to protect the seals and penetra-tiens from failure as a result of an explosion.

In the GESSAR II BWR system, as in most L*4Rs , the core melt-down debris can emerge from the RPV into an already i 4-19 l t

pressurized containment. A sustained ATWS or certain turbine trips or pipe ruptures can saturate the suppression pool and bring the pressure'to design values before core dryout and melt occurs. This means that extra steam'from residual ~ decay heat or noncondensible hydrogen generation or carbon dioxide i from concrete decomposition can possibly rupture the con-tainment vessel unless mitigative steps are taken. These mitigative actions are described in Section 4.4. When selecting and designing this equipment, the following condi-tions were assumed to exist: 1 l

1. All e'lectric power for operation of prevention and mitigation equipment has been lost. Controls' ,

and instruments may be inoperative. Note that over 90 percent of the transients leading to core melt are initiated by loss of AC power.

2. The normal and emergency core cooling systems are not functioning

3. The normal containment heat removal system is inoperative.

4. The core is essentially dry and the temperature has increased to initiate collapse of the core. ]

Melt through of the RpV lower dome is impending. I

5. Molten steel will accompany the core debris to the retention area. ]

6. A large pertion of the Zircaloy cladding on the fuel pins will have reacted with hot steam to form hydrogen gas.

7. plant operators are not necessarily available to initiate any corrective action.

Since the GESSAR II containment is still in the design and i approval stages it presents a unique opportunity to consider a different strategy to mitigate the consequences of a severe acqident. Two alternative strategies are considered l here. Section 4.4 describes severe accident mitigation equipment and costs applicable to the relatively high pres-sure, tight containments as they are currently designed and intended to be used today. Section 4.5 describes a com-plately different containment concept and 'ts costs wherein the containment is essentially not pressurized--only closed to ensure that all gases or vapors evolve <1 are directed through a very high capacity and very effective filter system. Although this new concept cannot be analyzed in the 4-20

\ If*. same detail as the mature pressurized designs, its cost and its risk reduction potential are presented for a first evaluation. 4.4 STRATEGIES FOR CONVENTIONAL MARK III For the conventional, pressurized Mark III containment to handle core melt accidents with minimal serious off-site consequences, the following new or modified existing com-ponents must be considered o Some form of hydrogen control.* (Mark III con-tainments are not inerted.) . e Increased capability for heat removal under acci-dent conditions to prevent containment overpres-sure. Hydrogen control alone will not prevent overpressurization, e Ex-vessel core debris retention and cooling to minimize concrete decomposition and hydrogen generation which can result in containment over-pressurization. Without proper retention, basemat penetration and possible radioactivity release will eventually occur especially if hydrogen overpressure control is included. e Venting excessive containment pressure through a relief valve and filter system, for both ATWS and non-ATWS events. e Underpressure control with vacuum-breaker valves, j 4.4.1 Hydrocen Control When severe accidents are considered, the Mark III contain-ment is at risk because it does not have an inert atmos-phere. As described in Section 4.2, hydrogen gas generation is almost inevitable during core dry-out and various hydro-gen phenomena (burns, combustion, detonation) constitute the I dominant risk. The Mark III containment currently uses only low capacity hydrogen recombiners designed for cleaning up the containment atmosphere during normal plant operation "The " hydrogen rule" already requires that hydrogen be controlled. 4-21

6 and/or design basis events. These recombiners.are quite inadequate to handle the 2 to 3 lb/s of hydrogen generated during a severe accident (IDCOR, 1983). A core melt acci-dent can be expected to generate enough hydrogen within the first hour to rupture the containment when there is global combustion or detonation. This problem has been the subject of numerous studies and these conclude that only limited defensive actions.can be.taken. Some are as follows: i

1. Ignite.the hydrogen to burn locally throughout the containment as it is formed or as it meets oxygen. Then remove the heat released (about 61,000 Btu /lb of hydrogen burned) in some way. -

The technical community is not yet fully in agreement on the efficacy of this operation l although it has been installed in at least one l containment (Sequoyah). There is insufficient oxygen to burn all of the hydrogen that might be generated. In Mark III containment, all enclosed oxygen will burn only about.2200 lb of hydrogen (only -500 lb in the drywell) whereas 3000 to 5000 lb might be generated.(Camp, 1983).

2. Inerting the containment at all times except during shut down and startup procedures. When service is necessary, plant personnel must enter with air-paes. This is a costly (estimated at $2 million/yr) and troublesome alternative during all normal plant operations and is not considered a desirable solution; however, it is currently done at Mark I and II plants--see Section 3.1.

3. Inerting the containment after the initiation of the accident can be accomplished by promptly injecting a material such as Halon, carbon dioxide, or nitrogen. In some accident sequen-ces, hydrogen generation can begin within about 5 min after core dryout, although this initial rate is low. This means that inerting, however it is done, must begin promptly--started upon the first indications of trouble.

Post-accident inerting is the least objectionable to plant operations and is a practical response to the hydrogen question It is the choice recommended in this study for the Mark III containment. Implementing it is not without its problems. The use of Halon is probably the most efficient and direct means since a smaller concentration is required than for any other inerting gas. The cost of the gas is so high, however (almost $5 million per charge), that reactor 4-22 , 1

i l L operators might be reluctant to begin the process until it l L is clear that it is needed--which may be too late. The use of Halon gas is not recommended for this reason.  ! Carbon dioxide inerting has-been thoroughly studied and widely used Industrially. Its main disadvantage.is that the gas dissolves.readily in water, especially at elevated pres-sure, forming a mildly corrosive acid. Another disadvantage is that under some accident conditions it can be reduced to ' carbon monoxide--a flammable, poisonous gas. The quantity needed is so large.that venting of containment air would be required during charging to prevent overpressure. The indicated choice at this time is the use of liquid nitrogen. This product is readily available to any plant and can be. stored efficiently. Its cost is not greatly different from that of carbon. dioxide, in that the handling , facilities represent much more cost than the material itself. To accomplish the purging of the Mark III contain- l ment volume to an oxygen content below 5 percent within 30 min requires high charging rates, and correspondingly high air venting rates to the filter. The inerting effectiveness can be improved by making use of the existing compartmen-talization of the Mark III. In a severe accident situation hydrogen may issue from-the RPV openings into the drywell (275,000 ft3 or 23 percent of the total containment) or it may exit the RPV via the PRVs into the suppression pool and rise to enter the lower wetwell area ( 330,000 ft3). Both areas must be inerted promptly after the initiation of an accident so each must receive liquid nitrogen spray. Actually all moist air displaced from the drywell by pure nitrogen passes through the pool. As drywell inerting proceeds this exiting air becomes almost pure nitrogen (less than 5 percent oxygen content by volume) which also helps purge the lower wetwell annulus. t The liquid nitrogen must be injected into the containment at  ; a high rate to inert the containment in 30 min. Roughly 20 percent of the nitrogen goes to the drywell, 13 percent to the lower wetwell and 67 percent to the large upper wetwell Each have ample dome. The wetwell areas present no problem. steel and concrete surface onto which the liquid nitrogen can 1 I be sprayed without complications, however, the drywell sprays must be more carefully considered because the areas are limited and much of the steel is hot and highly stressed. Cryogenic liquid must not be sprayed on these surfaces. An initial survey indicates there is ample surface and ample I J i

                                                                                                        ~

4-23 l i _ _ _ _ _ _ --_ _ __ 1

l l heat capacity in the drywell to handle the liquid nitrogen to q inert it but the sprays must be controlled. i A possible solution to this heat transfer problem in any of the areas is to mix about 0.6 lb water with each pound of liquid nitrogen as they exit from a spray nozzle so that snow is formed. This equipment is similar to that used at l ski resorts where snow machines use water and cold, com-pressed air. In the proposed process the nitrogen gasifies immediately and the snow melts a little later depending upon where it falls. In this way the cryogenic fluid contacts neither steel or concrete. The cold nitrogen gas then mixes . with the warm containment air, expands and displaces more ' air. Figure 4-4 is a scavenging efficiency chart showing dif-ferent theoretical processes for scavenging--perfect dis-placement and perfect mixing, and also a realistic path assumed in this analysis. The chart also indicates the scavenge ratio necessary to reduce the oxygen content below the volumetric 5 percent flammability level. Under these purging conditions the volume ficw to the con-tainment air filter bed is an estimated 2760 ft 3/ s at one atmosphere which is 18 to 20 ft/s velocity through the bed crevices at inlet but this gas has almost no radioactive { i contamination so the high velocity is not sericus. As soon as purging is nearing completion, the pressure relief valve can be returned to its initial pressure setting. The inerting process of nitrogen injection can be stopped at any 1 time if the crisis is remedied and atmospheric air can be pumped into the containment to restore safe working condi-tions. No harm will have been done, there will be no { deposits or corrosion, no seriously radioactive gases would  ! be released and the cost of the aborted operation would be I i low. l There would be adequate liquid nitrogen in storage to bring the oxygen content below 5 percent (by volume) in all these containment sections (drywell, 275,000 ft 3; lower wetwell, 330,000 f,t3; and the upper wetwell, 780,000 ft3). This value  ! is below the flammability level for hydrogen. A total of } 180,000 lb of nitrogen will be required. The liquid nitrogen  ; injection rate and duration into each area must be arranged j , to minimize the inerting time. This has not been optimized { l but critical areas can be inerted in 5 min and the whole 1 operation complete within 30 min after initiation.  ! Nitrogen is non-toxic but also non-life supportive. The pertinent characteristics of liquid nitrogen are given in 4-24 i b_______.__________ . . _ _ . _ _

rg, . - =3 .----- --- ----- t ,T f f .*[0 9-t.c  :. .

 ; i.

(ownloA Aq) uopmen ua6Axo Su!v!eweg -

                                   .8.

e. o k' o s o o l I i  ! 32 R Ra - 0

                                                                                                                                                                                                  *W
                                                                                                                                                                  -l                                0 02~                                w y8                                 u.

3 :a

                                                       -\
                                                         .\

j

                                                                                                                                                              .=}                                   4 g
                                                            \                                                                                                   c                                   >

to

                                                             \                                                                                                                            3
                                                                                                                                                                                         'E
                                                                                                                                                                                                  'O to (l                                                                                                          -

E!! #

                                                                  \j                                                                                                                        a
                                                                                                                                                                                              ~

i

                                                                     \s                                                                                                                    =      .
                                                                       \                                                                                                                   i      3
                                                                          \                                                                                                                         8
                                                                            \                                                                                                               y      u
                                                                              \
                                                                                 \                                                                                                         $

N- .

                                                                                                      \                                                                                            *
                                                                                                                  \                                                            .                   4 N                                             -

a o I \  % b

l. s -

.e

                                                                                                                                                      ',-                                         m I                                                                                                       .I *
         ,                                    6                                                                                                       E   .w*                                                      :,

1

                                   /,
                         =          s             a            s             =                                                            s                              =   =

a s s a (%) Anua!s!ys Su! 6ueAeos 4-25 4 f I

            =:_---_--_,______________- _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - - _ _ - - _ _

\ ,

                                                                                                                   :=

gf .

                                                                                                                                .I Table 4-7.      It will be stored-in a spherical, well insulated r               tank holding ample liquid for a complete flushing of the containment and will be located in a revetted building near L               the containment (s).                   Only one storage unit is needed for a multiple reactor installation. When inerting, this liquid                                                            4 l

will be delivered by an electric motor driven pump and/or a dedicated diesel engine driven pump system-into the proper - containment areas where it will be. sprayed against unstressed steel er concrete or' mixed with water to flash it to gas. In l the wel1~ insulated cryogenic storage sphere the thermal. - losses are low---1/4, percent / day--so even allowing 50 days of losses between fillings the storage required is only ~207,000.lb - f The 19.8 ft diam sphere holds'4090 ft3 of liquid and with :j the insulation is about 27 ft. diam overall'. It will weigh about 26,000 lb. Each pumping unit:will handle about SCO gpm at 150 ft head requiring -45 bhp. . All equipment is in an. isolated building l i and the connecting pipes (about 6 in diam) will be well insulated to deliver nitrogen to the containment where spray. :I

                                                                                                                           , I heads provide good distribution and rapid vaporization.

TA8LE 4-7 04ARACTERISTICS OF LIQUID MITR0 GEN p, , Solling temperature 77.3K

                                                                                    -95.2 cog C
                                                                                    -320.3 deg F Oensity-Lieuld                                     0.81 g/c 3                                     q 6.77 lb/ gal                                   1 3

50.6 lb/ft .l Heat of vaporization 199 J/g 85.6 Btu /lb i I Cost (approx. per trucklooe) SO cents / gal j 7.4 cents /lb 13.75/ft 3 4-26 , I I i _ _ _ _ _ . . _ - - -_ _ .-.__-_----______-__----_____---.__.a

e (( * . ( The cost of this liquid nitrogen.inerting systemicomplete is presented in Appendix A-8 and Table A-15 at $1,557,000. Based,onLanLevaporative~1oss of 1/4 percent / day and liquid i nitrogen at 7.4-cents /lb the maintenance cost is $37.50/ day Land the cost offeach full inerting operation is only about

                               $15,000.

4.4.2 Containment Heat Removal (Pool Coolinci The Mark III heat removal system, as currently. designed for

                        =all' design basis accidents,-is not adequate to handle severe accidents. However, it might be upgraded to meet the requirements-.              The thermal capa. city needed is.-45 Mwt while  .

the present design capacity is only 14.7 MWt for each of the' two units. The present pumps rely on the availability of-plant power but as stated.previously this is not acceptable for mitigation (see Section 2.3). GE advises that the present redundant system has increased capacity at the high temperature differences under severe accident conditions.

                         .Therefore, if one unit had a dedicated diesel drive, the present equipment.could be acceptable. This would be a low cost modification. .For value/ impact analysis however, a completely new and separate heat removal system with dual dedicated diesel power as described for the. Mark II.contain-ment in Section.3.4 is used, as are the costs from Appendix A-1.              The cost of these two 45 MWt heat exchange units com-plete with-both pool water and pond water circulating pumps, including pipelines and drivers is estimated at $2,085,000 for a new plant.

4.4.3 Containment Sorav system (Drvwell and Wetwell) In the new Mark III containments the sprays currently planned are either adequate for the core melt accident conditions or they can be easily upgraded to accommodate the larger flows required. The spray system arrangement depends upon how the hydrogen is handled. When hydrogen is deliberately ignited in the lower wetwell areas (as well as elsewhere), spray heads should be provided to pick up as much of the contain-ment head energy as possible to keep it from the filter system. Sprays should also cool critical areas such as the

     .                         penetrations that might otherwise overheat. For value/ impact analysis, however, the same extra spray system as recommended as in the Mark IIs is included.                The cost estimate is given in Appendix A-2 at $565,000 extra in a new plant.

l 4-27 i l 4

4.4.4 Containment Venting and Filterinc The' design of this' mitigation component is heavily influenced by:the ATWS clean steam flow conditions and the way hydrogen-generation is handled. 'The present Mark III containment.does not address the hydrogen issue since it does not becomefa serious problem until severe accident conditions are con-sidered. Section 4.4.1 considers alternative ways that accommodate this hydrogen gas-and post-accident inerting with nitrogen-is recommended. The ATWS clean steam flow conditions in'the Mark III system are very similar to the Mark II's as described in Sec-tion 3.4.4 and cost estimated in' Appendix A-4-2, Table A-9, except that the diverter valve mest be added (from Table A-11 at $200 K x 1.5 or S300 K) for a total cost of $1,529,000 in a new plant. This same venting equipment may be used during the nitrogen inerting operation.

   -When post-accident inerting with liquid nitrogen is employed the filter / vent system must accommodate the high exit flow of the containment air as nitrogen drives it out. During this operation the relief valve (s) sized for ATWS flow above should be opened so that the need for extra nitrogen to. bring the large containment volume up to relief pressure setting is avoided. It appears that the inerting operation must be completed in about 1/2 h. This means the average exit flow rate is 100 lb/s--well below ATWS flow. Even though the nitrogen enters the containment as cryogenic liquid (-320*F) the warm surroundings and contact with the suppression pool water will bring the air / nitrogen temperature to about 100'F.

(It is' proposed-to spray in water along with the N2 to reduce thermal shock to local areas.) This warm gas will be diverted up the stack until it shows contamination. Only then will it be diverted to the filter. If the filter is similar to that shown in Figure 3-19 the temperature of the

    -1000 tons of basalt rock will rise only 25 degrees maximum during the inerting operation. Even when warmed the filter bed would be dripping water from condensation of moisture laden air so it should have a good decontamination factor for particles' a'nd aerosols.

Once the inerting operation is complete the relief valve should be closed; it would open again only when the contain-ment pressure reaches its proper setting. There is never a l high flow rate through the filter when it must retain con-taminants. The cost of this vented filter system is given in Appendix A-4, Table A-10 at $1,995,000. 4-28

l' j 4.4.5 Underpressure Control Once the containment has vented off the large flow volume ) (-300,000 ft3/s) of ATWS steam, the possibility of con-tainment underpressure exists. This vented steam will have carried out with it some of the initial containment atmos-phere, inerted or not; and after the event has abated and heat is being removed from the containment, the contained steam will be condensed, leaving inadequate gas mass to maintain internal atmospheric pressure. Although of a lesser magnitude, similar conditions can exist after any low-volume venting: be it steam or hydrogen. It, too, will carry some of the original atmospheric gas with it and as ' the steam is condensed or the hydrogen burned their partial pressure collapses and an underpressure exists. Containment underpressure is not permissible if the contain-ment integrity-is to be maintained; therefore, a vacuum breaker valve of adequate capacity must be installed if one does not already exist. This inward flow relief valve must be large enough to permit the needed high flow rates with a low-pressure drop since almost no actual underpressure is permissible. The valve must reclose and seal when the subatmospheric condition no longer exists. The limiting condition is an ATWS venting of several minutes duration (after the pool is saturated) that vents and washes an estimated 25 percent of the nitrogen gas out with it. Later, when the dedicated heat removal system extracts energy at a 45 MW; rate, the partial pressure of the steam in the containment could collapse in as little as 90 s. During this pressure drop, the in-flow valve must handle about 7500 lb of air with a very low differential. It should have the free flow area of a 26 in diam port. Actually the Mark III containment has a vacuum relief valve but it may not have sufficient capacity. In case it is not adequate a preliminary cost estimate for a large system is included. The design, development and certification of this new valve,tpgether with a dual-valve installation including all plumbing and wiring in a new Mark III plant is estimated to be S865,000. See Appendix A-6, Table A-13 for details. 4.4.6 Core Retention and Cooling The Mark III containment design lends itself well to the inclusion of equipment to retain and cool the ex-vessel core debris mass. Because this equipment goes into a new plant, the design can be readily optimi=ed to accommodate the core 4-29

retentiot. system and it can be installed in the most expedi-tious construction sequence. Several factors allow this: e There are no radiation problems to interfere with the construction, o The design of the lower central pedestal area can be modified to fit a specific type of retainer. e The central pedestal area as designed is essen-tially dry (less than -2 ft of water accumula-tion). A dry sump precludes steam explosions and minimizes steam pressure spikes. e The steel heat exchange equipment such as the bottom pan, the water walls or the cone-shaped dry crucible can be shop fabricated and tested prior to site delivery where it only need be concreted in place. e The dry crucible concrete casement, if used, is made as an integral part of the basemat when cast, so that no effort is required to ensure a tight connection. The top of the steel crucible will be welded to the containment lining. o All cooling water piping and connections to the crucible water jackets can be installed when it is most convenient. Two types of core retentien and cooling are censidered--a pebble bed with direct water cooling by immersion in an enlarged central basemat area and a water jacketed cone-shaped steel crucible. Both of these have been described earlier for the Mark II type containment. Refer to Sec-tion 3.4.3, Type 2, page 3-44 for the pebble bed type j retainer and Type 3 page 3-47 for the dry crucible. While i there will be minor variations from these descriptions for the newer containment, the costs will be approximately the same. These costs are detailed in Appendix A-3 for the pebble bed system at S744,000 extra in a new plant. The design of a dry crucible type core debris retention and cooling assembly must include several features for assured performance: e The heat exchange process must be protected from damage by catastrophic, e: ternal events that may occur around the lower EpV area. The unit should 4-30 L___._.

I be well below the thick basemat and submerged in a hole. e The molten core debris must enter the crucible without encountering deep water to avoid explo-sive actions. The dry crucible design has multl-  ! pie diaphragms of thin steel that ensure a dry crucible but allow the hot debris to melt through each diaphragm in succession and pour or drop down into the crucible cone. e The inner conical sur* ace of the crucible must be ~ adequate to transfer the heat load from the mol-ten cere (>3400*F) through a crust layer at the surface, to the steel shell and thence into the coolant. This water jacketed dry crucible has been sized on the basis of 360,000 Btu /h-ft 2 heat flux but should be capable of considerably more. The specific thermal loading depends upon the porosity of the semi-molten to molten debris and the amount of steel involved with it. The needed heat flux falls as the residual heat decays. e To provide adequate surface for the heat transfer, the crucible is a truncated cone-shape with the top diameter -8 ft and extending down 60 ft to about 3 ft diam at the bottom. An B ft section, also with cooling jackets, is mounted at the top of the cone to accommodate molten steel and excess slag. This extra section also provides additional protection to the heat exchange surfaces by separ-ation from any violent actions around the RPV. The 1.0 in thick mild steel shell is water jacketed so that -3200 gpm of cooled suppression pool water will flow upward through the jackets at about 10 ft/s velocity. This forced cooling water is obtained from the dedicated RHR system and goes from the top header to the sprays in the wetwell.

                 .                                         A design that includes these features is shown in Figure 4-5.

The estimated cost of including this equipment in a new plant early in the design stages is given in Appendix A-3-3 Table A-B at $2,295,000. 4-31

                                                                                                                                                                         - - - - -               - - - ~ _ _ . , . . , , _ _          __

O To spreys g , Podestal gpy r d,

                                                                   \,y                                                1                                                                  .

e.s.

                        -.                      f                               e                                    .

br /.tpoe

                                                                                                                                             )
                                 ...         .         ;    e                                                          -                   -
                                                                                                                                                    .a s                                                                                     .
                                                                                                                                                                                              ; ,s s      .             ~              ..

• 89g:. . -

Qg .. n y .

                                                                                                                        .                    '3  :
                                                                                                                                                                                         .        b e, g =1:,                                                                                                                            ,[g                                    " , . ' . '; n,*
               's* . .J '.                                                                                                                         . ,i-

_ is .

                                                  $.c.2 , * ; . . . . ;. ;.                                                ;... ..                                                       . , ' <'"
                                                                                                                       . J.J :,;; ; ... .[ g l] ;

t.

               *" "' * ,                            *; g                                                                                                       . _ _
                                                                                                                                                                 . . . ; , ~,.; . , ' $ $ Buemet Cool water .,,, e, " J                         *
                                                                                                                           ,-*.},'.*...;
                                                                                                                            , _ _, ... .. . . . . . . ... .. .: .y      .e ,.      . ..;:* ..., , 18 c......... . .- .-. . . . -
                                                *                   ...s                                                              ,          .

trom ~~

                                                                                                                                                   ~...- #              #'~,",.                                     #'

g#s #, s ','g'

• s g 'ess ,. ,. .

MHR

   ,y m
                   -   g
                           ..              Is,                 #
                                                                                                                      ,, - =a-.,     ,-  .

3..T *g ,8,ssa'

                                                            ~                                                              .
                                                                                                                                                             * -= -
                                                                                                                              =
                                                                                                                              , .a

• s---s stw wagms #8 p ::i .. .

8

                                                                 ', C -:

l# -  ;. )

                                                                                                                   *~      *g Ic. ' .
                                                                                                                           **              Dry core detiris fi                   .,

teceotical

                                                                  ,o*
                                                                                   ~                            .. u *'

s.

                                                                  's,
                                                                                                ...\            
                                                                   *8                                     -
                                                                                                                 ' se                      Water W
                                                                                   ,. t d

se

                                                                                                                       =

r**g.;:

                                                                    ,u                              ' - ).; o*-
                                                                     *,*s. e,ow,-            o, g % s,!

Figure 4-5. Schematic Detail of Dry Crucible Core Debris Retention and Cooling 4-32 2

m.o

      ):.;

ci. [!f l- 4.4.7 : COST

SUMMARY

p Section 4.4 describes the mitigation components needed to 1 _ handle-a severe accident at a new GESSAR : plant. The. n' estimated cost of each component is given in.the description

             ,                 section-and is recapped in Table 4-8.

TABLE 4-6 COST OF MITIGAT6ON FOR CONVENTIONAL GESSAR CONTAl'etENT in $/1000 Adeouste boet removal (Sec. 4.4.2) 1 2,085 Contelament sprays (Sec. 4.4.3) 545

      <                                   Core retention end cooling--ory crucible (Sec. 4.4.6)         2,295 Nitrogen post-accicent inerting system ($ec. 4.4.1)            1,557 vent and filter system Ap p- A-4    Table A-10                 1,995 ATWS clear steem vent App-A-4     Toble A-9                    1.779
                                         . Large uncorpressure control velve App A 6 Toble A-13            865 Total                                                       $10,871 if the wet rubble bed cero retention is used               5 9,540 4-33 w-w                 ~ - - _ ,

l . f l [ 4.5 ALTERNATIVE STRATEGY: THE UNPRESSURIZED MARK III , A completely different containment mitigation strategy is proposed here. It is based on the concept of providing adequate venting. capacity so that the internal pressure in the containment can rise above ambient only enough to drive I the gases and vapors through a large absolute filter system to the atmosphere. This pressure lasts for a very short ,l

                    " blowdown" period. It requires an adequately large, very         I effective filter so that essentially no contamination passes        ,

this filter even when handling large flow volumes. This concept allows relaxation of many currently demanding and expensive specifications for the containment. It enables many current containments to cope with a severe accident without the risk of pressures above their original design value--in fact the maximum pressure incurred should always be just a fraction of the original design value. With this nonpressurized containment concept, a much lower first cost in a new unit, and reduced maintenance costs should result in a safer, simpler, more reliable, more easily serviced plant. The initial analysis of this new concept Indicates it can handle hydrogen generation, concrete l decomposition, ATWS steam flow and steam pressure spikes more easily and effectively than the standard pressurized shell. Some of the advantages are: e The probability of containment rupture by over- , pressure during an accident and the resulting I fission-products release to the environment is i drastically reduced. 1 e A number of currently critical valves are d unnecessary and can be eJiminated, for example: I l About half of all double isolation valves. All PRVs from wetwell to atmosphere. l Some of the rupture discs. All vacuum breaker valves except between the wet- and dry-wells. All these valves are costly, require continual l service checking and generally reduce plant reliability. 4-34 1

e The design pressure specifications on the containment structure itself can be relaxed to some value below 10 psig. Either the containment or the secondary building would still have the capability to handle internal and external missile ascaults and both must withstand seismic affronts, e The several hundred penetrations through the containment wall can have lower pressure and greater leakage specifications. They will be subjected to light overpressure only during acci-dent conditions which will have a limited dura-tion--minutes instead of hours or days. This is ' important for items such as the sleeves for the steam lines to the turbines. e Operational costs should be reduced because: The lower pressure equipment is lighter. Relaxed leakage specifications permit simpler equipment. There are fewer components to be serviced. Access for service is improved.

                                  --    Leak testing is almost eliminated.

It must be noted that the containment will still be com-pletely closed to ensure that all contaminated air, gas and steam (outflow and inflow) passes through either the new filtered vent system or the present low flow gas treatment units. Operation of the plant in this unpressuri:ed and filtered mode becomes especially attractive when certain other miti-gation components are added or the present equipment is modified to handle a core melt accident. 4.5.1. Open Vented Containment Svstem Details The open and unpressurized containment system will have:

1. Hydrogen control system using ignition and burning. Pre- or post-accident inerting is ur.necessary as ignition of the newly formed hydrogen gas is quite acceptable with the large open vent. In the more conventional closed and pressurized containment the possibility always 4-35

ye

     )                                   ,                                                        . .

L

 .)

l' exists that an. excessive amount of hydrogen has been generated before the igniters are started. The resulting global burn would be so extensive that'it could not be relieved without excessive pressure rise. This means that the operators l would have to determine that it is not too late I to ignite--the longer the wait to make the.deci-sion, the greater the chance of a global burn. In contrast, the vented system avoids this l dilemma.. The open vent duct to the large filter: - is. sized by the hydrogen burn situation. A global burn temperature rise of -20 F/s can be accommodated without excessive pressure rise. However, plenty of igniters would be placed throughout the drywell and in the wetwell sec-tions to ensure prompt burning of any hydrogen as it evolves--Iong.before any possible global fire. These igniters should be located to minimize the extent of any one burn and attempt to burn the hydrogen as soon as it becomes available and encounters oxygen. An estimated 100 igniter units at an average new reactor installed cost of L $3000 are planned for a total of $300,000. Tney , would be powered by a dedicated supply.

2. A reliable containment heat removal system of adequate capacity that is not dependent upon any plant electric system. This may be an enlarged modification of the present system or a com-pletely independent dual system with dedicated diesel power as described for'the Mark II con-tainment in Section 3.4.1 page 3-27. This system is included in the cost / benefit analysis. The estimated cost of this dual 45 MWt plant is

                                              $2,085,000 in a new plant (see Appendix A-1, l                                              Table A-1).

l 3. A spray system in the lower wetwell annulus sec-L tion using cooled suppression pool water. In a new plant the estimated installed cost of this equipment is $298,000 (see Appendix A-2, Table A-4).

4. A high capacity chilled filter as described in Section 4.5.2 designed with the following properties:

A chilled rock bed system to condense water vapor so that only dry air or gas enters the chilled charcoal filter. l l 4-36 l

[g +s C x

                                    --    Cendense the ATWS flow steam for'a very limited period.after the~ suppression pool becomes saturated.(-20 min of ATWS flow).
                                    --    Handle an ATWS overrun and pass' full ATWS clean steam flow without condensing at an acceptable pressure drop. At this juncture the rockbed is heated and the charcoal is warmed.
                                    --    Entrap all particles and aerosols' droplets-(contaminated or clean) and either hold them-                    ,

or wash them back.to the suppression pool.

                                              ~
                                    --    Chill all noncondensible gases to low temperature before they enter the activated charcoal filter system.
                                    --    Filter section to retain contaminated heavy gases by adsorption on the extended cold charcoal surface Both filcer beds are shielded by heavy reinforced concrete walls and insulated to hold the low tem-perature as illustrated in schematic Figure 4-6.
                                                 ~

Installation of this filter system at a new plant is estimated at $2,938,000 (see Appendix A-7, Table A-14).

5. Containment underpressure control is not required. Air as needed is drawn back through the large chilled filter. This will warm the activated charcoal somewhat, but any adsorbed gas that is released goes back into the containment proper along with the air. The refrigeration plant will draw the charcoal temperature back down to -40*F in -80 min after about 50 per-cent of the mass of air in the containment had been ingested. This is deemed soon enough to avoid any contamination release.

6. Proper retention and long-term cooling of the ex-vessel core debris mass regardless of where it finally stops after the melt down. Alternate core catcher designs to handle this hot mass and l.

avoid steam explosions as well as steam spikes

1. are described for the Mark II in Section 3.4.3 page 3-35. The estimated cost of adding this equipment to a new plant is $2,295,000 (see 4-37 f

C______________________________._________________________________________________________________ .

p:s -. ... In , u f 1 i,. fl

                                 }

fl k ,

                                                                   . } I j       Yd!!3,                          ,jj i'            ]  .
                                                                                                    };

a

                                                                                       \
                                                           . /c m
                                                                                                   ~\
                                                                                                                      \                     ,L l
                                              -w           ..                 .

m ,y .  ; g

t. . i. .

l A o< p '. . .

                                                                                                                                  . i.              3 u          -

e< .-

                                                                                                                             'm;.-                  2 e

i .m. ' o< j}.g

                                                                                                 .                   .e
                                                                                                                                                            .c
                                ,j                             j '
                                                                          "W'                                              y                e' i

r

                                                                                                                                                                         ~!'

ilt; ilI

                                                                         ~^                                                w.             -

ae " o-

                  ,o<

1Je Ir t

t. -

                                                                                                                          ...... t -

l' 0 , l ,_* N

                                                                                    ~

O< ' .. -.wy> 7

                                                                           'f e ,, g                          Cl

[ -

P!a!1 0 e .. n . 'J 5 ..,

3 w c . g

                                                                                                                                                           ..o
                                   ',.                                                                                                                     J m
                                   ' j ;;

1.. - ' E  ! o

                                           ,! If !                                             I              '

6 _ *h

                                   'i d. "                                                                                                                 m l

il l u i , lee [L

                                   -il                  d.; A.                   Ii5I                         O s                                      q*                  .i 1,                   a f                                  - 3 W
                                                                                               /                                                           3 i                                                                                                                                                           P 11                     __                                          7-l

rs

                                                                                       "                                                                ~

0 l' 4-38 l e L____-_-----____-----------

           .)*

• i l

I l Appendix A-3.3, Table A-8). If it can be used, the cost of the pebble bed catchment is much less, only $744,000 (see Appendix A-3.2, Table A-6). i However, the Mark III containments bave even a smaller central pedestal area than the Mark II designs which raises even more questions about the hot core coolability in the central basemat . as disc ussed in Section 3.4.3 Type 2 page 3-35.  ; 4.5.2 Mark III Larue Chilled Filter System j When the large open vented containment concept is considered as an alternative to the present high-pressure closed sys-tem, it is imperative that the filtering effectiveness be  ; high under every possible condition. It is now the only 4 component barring release of radioactive materials to the ) environment. This filter must have a high decontamination l factor for condensibles, particulate and aerosols as well as good retention efficiency for the noncondensibles such as the iodines, kryptons, and xenons. To achieve this effec-tiveness, a large filter unit is arranged so that the steam, gases and particulate slowly pass over very cold basalt type pebbles (-1000 tons) where the steam condenses and locally freezes, while the noncondensibles pass into a deep bed of cold activated charcoal. Any condensate in the pebble bed dradns back to the suppression pool in the con-tainment. Since the hot gas and steam enter the filter bed from the bottom, the rock warms from the bottom as the rising gas chills. The pebbles are prechilled to about -80 F which freezes out all moisture. The dry cold gas enters the

                                 -80 F activated charcoal bed that will detain the entering noble gases as well as the iodines. The chilled rock will accept about 5.0 MWH t energy in a reasonable time (to allow it to soak in) and still deliver -40 F gas to the activated charcoal bed. The capability of the bed to retain heavy gases is affected markedly by its temperature and -40 F is adequately low.

For an ATWS event, the steam always passes through the suppression pool in the Mark III arrangement; however this heat sink becomes saturated quickly (an estimated 20 min) at this high steaming rate. Actually, this is ample time to shut down the reactor with poisen and stop the ATWS flow. l If it is not stopped before pool saturation, this clean steam will flow to the chilled rock filter / condenser. The high flow rate warms the rock in about one minute and l l 4-39 l l l 1 I f

[y_ s, [ g te L-15 . proceeds to warm the clean 1 charcoal. Both filtere will pass b the ATWS steam but are warmed by it and.are less effective in trapping the noble; gases'until-rechilled--a period of-less than'two hours'. Preliminary investigation indicates o the chilled filter effectiveness for retaining the noble' _' gases should be excellent. (See Section-3.5.2.)

                           .In the-PRAs.for both Mark II and III containments it.is assumed:that, containment failure precedes and causes core                             i melt. This_ sequence is not possible:with this' proposed open               -

system. So long as there is suppression. pool water avail-L ables core melt will be delayed. -Clearly this interesting possibility.needs further study. I

                           ' Figure 4-6 is a schematic of such a filter system. -The
                             -120-ft-hign vertical filter stack is in the shape of'a                                ,

cylinder with a ft-thick concrete wall for radiation 1 shielding and'2 to'3 ft of insulation to hold the low temperature. The diameter is -30 ft inside at the base.

                           'This silo provides two distinct types of filtering pro-         .

cesses. 'The lower section is a layer over 80 ft deep, containingf-1000 tons of pebble rock. -The available inlet L flow area through the rock crevices is over 200 ft2 so that 1 l the gas _ velocity.As low. L l The-second filter section is a ft-deep layer of acti-vated. carbon in the upper region of the silo. This -40 tons i L - of active material delays transit of the noble. gases by I surface adsorption. A weather _ shelter with swinging door valves is located on top, arranged to allow gas flow either L in or out but remain closed with zero flow. The filter bed is maintained at the proper temperature:by circulating dry nitrogen at -80 F through it from a refrigeration plant, u Chilling the dry nitrogen would be done with two Freon l cascade-type refrigeration plants (York or equivalent) using water cooled condensing and electric motor drivers. (Streiff, 1984). The cold nitrogen is circulated with fans. Either of the dual units of 10 tons (120,000 Btu /h) capacity can handle the heat losses of the system, but both can be used together to cool down rapidly when required. Both units are expected I to be able to handle the heat generated in the filter by I radioactive decay in an accident aftermath. Each compressor draws 66 Bhp or 80 kW including the fans. The units are l housed in a small separate building. A dedicated diesel drive

. for these chilling units is not considered necessary since j operation of the chillers is not required during an accident l

4-40 s 4

                                                                                                                  'l

1 itself, and the thermal inertia of the filter is so large that several days could elapse without appreciable warm-up. As described earlier for the Mark II filter, the chilled beds are kept dry and isolated by a rupture membrane at each end. These are cleared away by gas flow if an appreciable under or overpressure develops in the containment. Normal ventilation of the containment would use the system normally provided for these plants. It could be isolated during an accident or provided with a small chill-filter. , In operation during a severe accident, the filter would accept heat as steam, and air from the containment, and , whatever particulate and noble gases were released during core melt. The only materials leaving the filter would be oxygen, nitrogen, and hydrogen. Once the rapid release of gas was completed, the swinging doors at the top of the filter would close. Later cooling of the containment would draw air backwards through the filter into the containment, with the doors swinging the other way. With no net flow,the doors would seal, and nitrogen circulation would resume, keeping the filters cold. After sufficient decay, these gases could be slowly vented by allowing the filters to warm up, or alternatively the off gas treatment system (GTS) could be used to draw them from the filter. 4.5.3 Cost Summary Table 4-9 is a cost comparison between a conventional pres-surized type of Mark III plant as in the GESSAR II PRA and essentially the same plant with the same degree of severe accident mitigation but using the non pressurized concept. In this comparison, however, no credit is taken for the obvious first cost savings and the reduced maintenance costs on the following: e Less steel in the containment dome structure-- estimated at 10,000,000 lb saved. e Reduced cost of lighter penetrations with simpler teals.

    -                   e  Fewer isolation valves.

e Easier maintenance without an inerted atmosphere-- estimated by GE at $24,000,000 saving over the plant lifetime. 4-41 9

lg I ', .- 1

                                   ' TA8LE 4-9. : COST CONPARISON SETWEEN CONVENTIONAL A@ NONJRESSURIZED CONTAltesENTS
 +

PRESStstlZED . NONJRES$URIZED IN 1/1000 'IN $/1000 r ADEQUATE TEAT REMOVAL 5 2,085 12,087 L CONTAlletENT $ PRAYS ' . 56 5 PRESENT.. CORE RETENTION AND COOLING-- DRY CRUCISLE 2,295 - 2,295 HYDROGEN IGNITERS 5100 UNITS NONE. 300 044LLED FILTER $1LO NOT REQu6 RED' 2,9385 WET RtS8LE BED FILTER 1,97$ NOME NATROGEN'INERTING SYSTEM 1,557. NONE ATWS CLEAR STEAM VENT 1,529 -NONE

l. LARGE VACULM BREAKER VALVES 865 NONE 1

TOTAL IN 5/1000 110.871 $7,618 , OlFFERENCE' 53.255-l

                             - 4.6    OPTIONS AND VALUE/ IMPACT COMPARISCUS-As discussed in Section 3.6, the mitigation components described above can be combined ~1n.various ways ~to create a modified containment system designed for specific threats.
                             .Although the basis for making such selections and combina-tions usually depends upon cost / benefit or valee/ impact con-
                             - siderations, other factors h' ave a strong influence. These p

include accident recovery costs, public percepti'on and con-fidence, reduction-of uncertainties, and engineering judgment. As discussed in Sections 4.2 and 4.3 there are two principal

j. threats to the containment:

l^ e Hydrogen burning (combustion or detonation). e Steam and/or noncondensible gas generation. l To mitigate these major threats, two alternatives were pre-sented: the conventional high pressure containment and the no pressure, open containment with a chilled filter. Either choice would be augmented by components to ccpe with: 4-42 _ _ _ _ _ _ _ _ _ _ _ . __________-___-_______-________--_a

3-x ' ,

a. Containment heat removal,

b. Core debris control,

c. Containment pressure protection.

Table'4-10 lists the cost of each component, the total "value" in terms of' man-rem averted over 40 yr based on the GESSAR PRA and the NRC staff review, the total " impact" in terms of system costs-and the-impact /value ratio which can be compared to the $1000/ man-rem averted algorithm sug-gested by the NRC Safety Goal Report (NRC, 1983). It must be stated again that each column must be viewed-in- , total. This point must be stressed because if the dominant failure mode (hydrogen combustion or detonation) were elim-inated, the containment would ultimately fail by overpres-surization. Although this type of failure would occur later in time, the risk averted does not change substantially because all of the releases are somewhat delayed. This delay, as was discussed in Section 4.2 isInattributed to the fact, examina-time it takes to generate the hydrogen. tion'of Table 4-9 shows only a factor of three difference in populationLdose and/or latent deaths between an early and a late failure of containment, except for the I-SB-EI sequence. Table 4-10 compares the conventional approach (three options) with the proposed new approach (the unpressurized, open system). In all cases, the impact /value ratio is slightly greater than $1000/ man-rem averted (between $1500 and $2700 per man-rem averted) when using the NRC staff upper bound values for risk averted. However, using the GE values for risk averted, the impact value approaches S10 5/man-rem averted. It should also be stressed that the results presented here do not include external events. As indicated in Chapter 3, the inclusion of external initiators (seismic, etc.) will increase the initiating frequencies and hence the benefits, provided that the mitigation feature itself is designed to survive the initiating event. This  : situation is discussed in Section 4.7 belcw.

                 -         The three options presented all have dedicated pool cooling and sprays. Recall that the Class II sequences had loss of residual heat removal leading to core melt, and were a sub-stantial contributor to core melt frequency (34 percent), but were a small contributor to risk (Table 4-4). The dedicated pool cooling and spray functions would reduce core-melt fre-quency and hence on-site costs.       This benefit is not explicitly accounted for in Table 4-10, but would be included 4-43

j- a se fI l: TABLE 4-10 MARK III CONTAI44ENT MITI3AT10N Options in 1/1000 Function Equipment Low pressure open High pressure containment contalwent .lth I chilied filter Option 1 option 2 Option 3 l Heat removal Pool Dedicated cooling 2,085 2,085 2,085 2,085

       ~Sprey                  Dry. ell sprays plus        Adequate in present             .

external food design 565 565 565 Core control Plus Desemat rubble bed --- --- 744 744 1 l Plus dry crucible 2,295 2,295 --- -- Pressure protection Over Ignitors 300 -- -- 300 ATwS clean vent -- 1,579 1,579 1,579 Filtered vent --- 1,950 1,950 - Nltrogen inortIng -- 1,557 1,557 ---

       .Under                  Larger breaker                       --                865          865          865 l

Both Chilled filter 2,938 -~ --- --- Open contelnment l Impact Costs ] (In 5/1000) 7,618 10,896 9,345 6,138 j

                           .                                                                                                 l Value or benefits             GE                             11                  11           11         11          )

non-REM averted NRC'* 5,240 5,240 5,240 5,240 i impact /va lue 5/ man-REM (GE) 6.9 x 10' 9.9 x 10' 8.5 x 10' 1.6 x 10' l Ratio $/ man-REM (NRC) 1,450 2,061 1,782 1,174 I e Assumes 500 mile redlus, 40 year plant Ilf e, no discounting and perf ect mitiget ton.

       .o NRC stoff: GESSAR, SER, Mod 1 1

1 4-44 i

j+ * -1 in an overall value/ impact approach. If this were done, the values presented would be greater. 4.7 EFFECT OF EXTERNAL EVENTS All the risk estimates in this report, including the results presented in Table 4-1G, are based solely on internally initiated failure sequences, in accordance with the available PRA's. In this section we attempt to estimate the extra benefits that could be credited to mitigation if external initiators had been included. A large earthquake represents the dominant portion of the external event risk. The NRC staff and its consultants (NRC, 1985) estimated that the frequency of core-melt due to seismic events is approx-imately 6.7 x 10-5 per year. The major contributor comes from the seismic event causing a loss of AC power in the plant (frcm relay chatter). Recall that in Section 4.2 it was determined by the NRC staff that the total core melt frequency of 3.8 x 10-5 per year was also dominated by tran-sient events with loss of electric power. Of all the seismic initiated core melts, only 5.9 x 10-6 per year result in containment failure due to the seismic event as well. Hence, 4 6.1 x 10-5 per year are mitigatable. That is, the contain-ment survives to protect against release following a seismic-induced core melt. To a first approximation, the NRC benefit given in Table 4-10 can be increased by the ratio of total core melt frequency (3.8 x 10-5 + 6.1 x 10-5 per year) to the internally induced core melt frequency (3.8 x 10-3 per year), or a factor of 2.6. This would reduce the impact /value ratio by the same factor, giving a range of $400 to $1000 per man-rem averted when seismic initiators are included. As before, this represents " upper bound" benefits. 4.8 OTHER MITIGATION SYSTEMS All work presented in this chapter presumes that the mitiga-tion components are separate from the plant accident preven-tion equipment, that these components do not depend upon inplant electric power and that they do not require an operator to function. In many plants some of these existing prevention systems can be upgraded to handle some of the mitigation needs, and would be simpler and less expensive. GE has proposed an alternative to provide added mitigation capability and to reduce the frequency of a core-melt acci-dent by a factor of 5 to 10 (GE estimates a factor of 10 while NRC predict a factor of 5) (GE, 1985). This additional equipment, called UPPS (Ultimate Plant Protective System), 4-45 1

l l ia o , l-f has been' presented-to the NRC staff'for1 review'and is intended to accomplish three functions: e Assure decompression of the.RPV. o Provide an assured source of reactor cooling water once the RPV. p:' essure is low enough. , e' Avoid containment overpressure from heating ~by venting steam and gases to the atmosphere.. . The UPPS proposal'is1 conceptual and no details arm available; j as yet but it proposes: e To' assure'RPV depressurization with air pressure l operated relief valves using a dedicated air ~ i supply.

                                                                                                                't

e To assure-a core cooling water supply by diesel powered pumps taking. suction from the plant fire; protection system. The engine is started by an l operator. Backup for this system is a fire truck  !

                          " pumper" using-the same' water source.

e To assure containment venting the relief valve will be actuated with a. dedicated air system when electric power is not available. e To supply critical pressure and water level readings under accident conditions. e To provide these facilities independent of plant electric power.

            . GE have submitted conceptual and preliminary designs of this system to NRC for GESSAR II with details deferred until a L             license applicant references GESSAR II. It should have the l             following advantages..

o It is simpler and less expensive than other miti-gation components. o Although it does not directly handle the hydrogen j; problem it reduces the core melt frequency enough I 1 that the CP/ML Rule can be met. o GE believes this UPPS will reduce core damage frequency from internal events by a factor of 10 l l l' 4-46

e :. 1 Y? ,

 -                                                                                                                                      l o

f- l

                                                                         -(GE; '1985)','while preliminary results presented to NRC show a reduction of 5.
                                                             ~ UPPS also has some drawbacks.

e' It requires the' action of. operators. o The depressurization. system can be inadvertently actuated. c e It appears'to heve quite a few shared components--shared with the system that may have failed initially. e During many accident scenarios the plant fire protection system is heavily taxed by other i exigencies, o Proposing to rely on the fire system and especially on fire trucks presents a poor image of reliability and forethought. 4.9

SUMMARY

The Mark III containment system is a well-engineered, compact design, easily capable of withstanding the set of design basis accidents for which it was intended. Like most con-tainments it is still vulnerable to very severe accidents that exceed the original design criteria. This study shows that relatively minor improvements can substantially cure these vulnerabilities. However, it is difficult to prove that the benefits are greater than the costs.because of the uncertainty in the risk calculations. Many of these accidents, and the eventual risk of containment failure,. stem from inadequate heat removal systems or other safety devices that depend upon electric power being avail-able at this crucial time. Much of.the mitigation equipment recommended in this work provides a non-electrical means of

                                                             - accomplishing a function normally done electrically. It should be.noted that having electric power available at the plant bus does not ensure that it is available at the a

critical equipment. The ATWS event frequency has been reduced by design change in the latest BWR/6 reactor, but its consequences are still

                                                             -large'enough to justify mitigation when using the NRC staff
                                                             . data. Other f ailure modes derive' from slower processes related te hydrogen burn phenomena (combustion and detona-tion), heating by steam, and gas production by core residue.

4-47

ip,N 1 3 LThe equipment'for intercepting these modes is' easily installed in new plants.and-appears to be_ technically feasible. Although the conclusion is definite that Mark:III mitigation' is feasible,-it is important to have a feel for the:uncer-taintiesfthat lie within.the factors used when-regulatory decision-making is involved. There'is-little uncertainty , that the proposed mitigation cumponent can be installed, that'it will function as, designed, and:that the cost can be . determined. .There is also little uncertainty that in case of r a severe. accident or core melt-, the containment.wil' tend to fail in one'of the modes discussed, and-that it would be

                        . protected ifnthe mitigation equipment is in-place. What is
                        .less certain is that the accidents will have_the frequency predicted, that they will have the conditional failure _ pro-babilities calculated, and that the consequences of' failure are of the magnitude estimated. Tables 4-2, 4-4, 4-5, and
                      . 4-10 only serve to highlight these uncertainties. For the GESSAR-II standard-plant, the difference between the GE-PRA and the NRC staff' review is three orders of magnitude (factor of 1,000)'in potential benefit. Thus the situation is that the financial impact'and the technical effectiveness of miti-gation are certain, while.the projected. benefit is quite uncertain.- Thi s uncertainty is examined a bit further.

Accident frequer.y and external consequences can never be verified by exp.riment: they must always remain best esti- i mates. Some - cr s tponents have enough operating data behind them to determiae their failure frequency, but many key inputs must~ remain judgmental because there can never be enough accidents to provide the needed data. There_is an observable trend toward higher calculated frequencies as in-depth study continues. As to radioactive material released, the source term studies underway will certainly tend to reduce the expected' doses, but other calculational assump-tions that appear in serious need of review would be corrected in an opposite trend. In fact, as individual PRAs are reviewed, projected risks tend to increase. Hence a  ; potential benefit of mitigation is the elimination of these l uncertainties, and hence of risk. Another major potential benefit of proper mitigation is its impact upon-the perceived risk of nuclear power plants on the l part of the public. Improved public confidence and espec- l I ially improved confidence of the financial community can be a very real benefit from mitigation. i

                                                                                                 )

l

                                                     . 4-48                                      !

I

w

 ; Y%                  '

1 CHAPTER 5. ' MITIGATION SYSTEMS FOR WSP-90 PLANTS The reactor and containment system described here is a version of the Westinghouse PWR reactor and the large dry

                         . type of containment. This system is under development by Westinghouse Electric Company with the designation ofLWSP-90, and is under review by NRC for a Preliminary Design Approval.

Some features of the design in its present status have been made available'to the NRC by Westinghouse, and these are used ' here for the. consideration of mitigation systems that would be.needed to mitigate the effects of severe accidents that exceed the design basis criteria. Since.the work on this project began, we have been advised that Westinghouse is considering changing the original cylindrical containment to a spherical shape. Because neither. version is necessarily final, we have chosen to keep the cylinder in our illustra-tion', since it is less expensive and more suited to mitiga-tion. It must be emphasized that the description of the plant and the illustrations provided may contain proprietary material, and that the design as shown may be changed before

                         'it is offered for full licensing review.

In'the previous two chapters, the Mark II and Mark III con-tainment systems for BWRs were studied, and'possible mitiga-tion systems considered. The mitigation features were assessed by considering the cost in dollars relative to the expected benefits in terms of population radiation dose averted over the lifetime of the plant. The basis for these estimates was a PRA assessment of dominant failure modes, expected accident frequencies, and possible dose consequen-ces. An equivalent PRA study has not yet been published for the WSP-90 plant design, because the characteristics. of the system have only recently been determined, and some features are still subject to change. For this reason only tentative and preliminary assessment of failure modes and mitigation opportunities is possible. Even with these limitations, however, the value of the assessment may be higher in the case of a design not yet committed for licensing than it is for a system having a substantial investment in design, development, and licensing activity. As the following sections will show, there is a possibility that containment failure in this advanced design  ! can be made nearly impossible even in very severe accidents, at a cost that is essentially negligible. 1 l 5-1 f

s 7 For the WSP-90 plant, the reactor, its fuel, and the contain-ment system are quite similar in most characteristics to the existing highly successful PWRs in large dry containments. In the absence of a PRA, it is a reasonable assumption that the same basic modes of failure will apply to the new design, and that the consequences will be similar, but that the accident frequencies may be lower. Thus we are able to determine with fair assurance what type . ; of mitigation system would be suitable in a severe accident, and the cost involved, but the bottom half of the cost-benefit ratio--the expected population dose that could be averted--will be missing. As the results have turned out, this lack may be of little consequence, for the actual  ! benefits seem to be of quite a different nature. In the other systems studied the cost of a suitable mitiga-tion system was compared to the reduction in population dose to obtain a ratio of dollars / man rem. averted. This was , deemed provisionally attractive if the ratio was below 1000. For the WSP-90 a new approach to containment seems to be the most feasible. If this modified approach, having inherent resistance to severe accidents, is actually cheaper to build l than a more standard containment, a direct monetary benefit would accrue, provided the modified system were equally licensable. In such a case, the mitigation features would have a direct justification. In fact, the system with greater accident tolerance should be easier to license, and should achieve greater public confidence. At this time, of course, all these benefits are prospective. Their further investigation should prove valuable and instructive.

5.1 DESCRIPTION

OF THE ADVANCED LARGE DRY PWR CONTAINMENT The advanced PWR reactor and large dry containment as typi-fled by the Westinghouse WSP-90 Nuclear powerplant design incorporates many preventive type improvements over earlier models. Although detailed information on these preliminary concepts is limited, it is known that there have been inter-nal changes in the reactor vessel and core to reduce the possibility of the core losing its cooling water cover. Other components have been moved into the containment to improve safety margins in an accident. Some of these features are shown in Figures 5-1, 5-2, and 5-3. General , information on this type of plant is detailed in Chapter 7 of the Task i report (Castle, 1984) but more plant specific pertinen:. mitigation data and its operating conditions are included here in Table 5-1 for ready reference. i 5-2 I

y:n

• 1 I E I E E I I I I I E I [ I W E 1 I I I I I I E I I E I Polar crane {

                  ,          ]                                                                                                              .

7 5

                                                            .O-             2 A El 14.8 El 18.9 N
                         \

l A El 11.3 g Water el 10.8 g Operating floor el 11.3 [

                                                                                        ~#=~

El 5.1

                    \  '                                                                 El 1.27                                     '

A El 4.9 El -1'48

                           -           "O"         %                                        [ El -3.1 El 7.9            .

i i 0.35 =;; El -2.3 O -- I-- N - - - f

                       '                                                                                               l A
                                              .                                                             ;g r--I                    Ei -6.3                         N s
                          #              C EWST                                      k                 t, , j I                                                             EWST EI-9A          \
                                                          \

El -8.3 El -15.7 El -16.0 El -19.0 Figure 5-1. (U) Section Through WSP-90 containment

         -                                         (May contain proprietary information.)

5-3 l-

.c I

- j 1

I i l pew' orene I l I T k l L l ^ ^ l \ l I g El 18.9 - g 1- l 1-l ' l \ I \ t l l El 11.3 El 11.3 l 1

                                                                         -                            1 I

n I i 4 1 7 nl E

                                                         ^                                J      F l

I i V _, V i El -2.3 % l [ l Y , El 1.27 I I El -3.9 ) l I r- , I g ' E! -6.3 I i '$ i EWST Lj I El -8.5 EWST

      }

i El -12.7 U El -15.7 El -164 i El -19.0 Figure 5-2. (U) Section Through WSP-90 Containment

              ~

(May contain proprietary information.) 5-4 i

+ ' 7 w l N f I  ! .i

                                                                $         l
                                  ,      v. e/W y    ,
                                                                                            \

f # ItN // \ K\ it --

                                                                                        +
                                  ~ "I".14J   l         //           '

D2 j

                                                           <~t
                                            ,            t %'

d,#

                                                             -     g      a F- .

Y l_ Y w - Figure 5-3. (U) Plan of WSP-90 Containment (May contain proprietary information.) 5-5

( 4 e i' TABLE 5-1, GENERAL INFORMAtl0N--TYPICAL PCwER REACTOR 'l; AND LARGE DRY CONTAINMENT { i j., i GEMERAL Make/Model' West./WSP-90 Electric output - WW 1,250 NS$$ thermal output - WW 3,816 i ' Efficiency - approximate a pct 33 REACTOR PRESSWtE YESSEL Pressure - psig l

                       - operating                                                           2,250                                        ;
                       - test                                                                2.485 Temperature
                       - mer out - dog. F                                                       625-                                      l      )
                       - design - dog. F                                                        600                                       p Primary circuit water flow rate - Ib/hr                                  150.6 x 106 Steam flow (from 4 generators) - Ib/h                                     17.2 x 106                                  .

Size

                       - Inside diameter    't                                                 16 .67
                       - wall thickness - tt                                                    0.83
                       - Inside height - ft                                                   65.53                                              )
                       - Outlet nozzle diameter - ft                                            2.56                                             (
                       - Bottom heed thickness - f t                                            0.56                                             1
                       - Bottom height weight - tons (est.)                                       65 volumes
                       - Water in primary circuit - tons                                      ~500
                       - Steam and water in secondary                                                                                     i. j circuit - tons                                                    ~200                                        j PRESSURIZER
                       - Volume ft)                                                          2,400
                       - Rollet valve prssure setting - psig                                 2,485                                                ,
                       - Capacity - 3 unl+s - Ib                                           A
                                                                                           ' 1,700                                              {

F1JEL - UD2

                       - Charge invento*y - tons                                                210                                              j
                       - Zircaloy cladding - tons                                                 60                                             l
                       - Steinless steel with core - tons                                       ~15 PRIMARY CDNTAIMWT SYSTEM Type                                                          Dome on vertical cylinder Construction                                             Steel reinforced concrete Size                                                                                                                           ;
                       - Main ch amber - Inside diameter - f t                                   150
                       - Height - pit to dome                                                   240 Free volume in containment - f t3                                          3x 106 Thickness - ft - basemat                                                       ~10 Pressures - psig
                       - Design Internal pressure                                               (60
                       - Rollef setting                                                           60                                               l l

SEC0tOART CDNTAIM Reinforced concrete building around containment ULTIMATE HEAT 51* Unknown - site speelf te , l HEAT REMOVAL SYSTEM Four units each with electric notor drive from two separated buses Capacity - 6.3 WWt each 5-6

B Safety Features in the proposed designs are' listed below-- these are all'part of the Integrated Safety System (ISS).

                          '. Emergency Core Cooling System (ECCS) e  High pressure-coolant injection system (HPCI)

Four pumps, 500 gpm min capacity each, 3285 ft head e Low pressure coolant injection system (LPCI) Four pumps, 1800 gpm capacity each, 500 ft head . e Pressure safety valves Three valves with a combined-capacity of 501,700 lb/h e Accumulators Four units --total water capacity of 7000 ft3 at'600 psig e Core reflood tanks

                                                   '/our units with total capacity of 1400 ft3 at 200 psig e  Containment spray cooling system.
                                                  -Electric power from two separate buses Two pumps take suction from each of two Reserve Water Storage Tanks (RWST)

Four pumps each 1800 gpm at 60 psig containment pressure, 2000 gpm at no pressure containment Any two pumps provide 100 percent coverage Auxiliary Systems e Residual heat removal system (RHR) Four electrically driven units--6.3 MWt each l e Reactor core isolation cooling system (RCIC)

               .             Hydrogen Control System o  Electronic hydrogen monitoring e  Dual, low capacity, hydrogen recombiners (70 cfm l-                                                each) e  Containment air mixing by electric circulation fans 5-7

n-i l i 5.2 CONTAINMENT FAILURE MODBS AND MITIGATION REQUIREMENTS , In earlier chapters, assessment of dominant failure modes and the corresponding mitigation requirements was made using PRA's prepared by utilities or other agencies. As discussed l Lefore, no PRA is available for the WSP-90 system. Instead, e the assumption has been made that, although accident fre-quencies are unknown, should a severe core-melt accident  ! occur in this type of plant, the phenomena resulting and the j response of the containment would be similar to those for

• i other large dry containments housing PWR reactors. This ,

assumption appears justified because although the WSP-90 l system has many improvements aimed at reducing the frequency l of plant failure, once a core-melt has occurred these pre- I ventive systems have no function. Thus the RPV and the l containment, being essentially the same as in other PWR ) plants, can be expected to behave in a similar way. In the absence of information to the contrary, we also assume that loss of all site electric power is a leading j factor in the initiation of transients or other accident I sequences, and that mitigation to be effective must function during site blackout conditions. If the final WSP-90 design incorporates non-electric protective devices, some of the . mitigation elements considered here may not be needed. l A basic premise throughout these studies should be restated I here, since it also serves as a ground rule for mitigation I requirements. This premise is that mitigation is not effec-tive unless all of the end-states resulting from dominant accident sequences can be controlled. Thus any proposed 1 mitigation system should prevent containment failure by any mode, and result in a stable post-accident situation. j Listed below are the most probable accident end-states for 1 the large dry type containment--their relative importance depends upon the specific accident scenario: 1

1. In-vessel hydrogen generation. This end-state results in the addition of noncondensible gases to the containment in amounts which the large volume WSP-90 can handle. However, it also introduces the possibility of a hydrogen con-flagration (5 percent & by volume) or even an explosion (15 percent +) when the containment is not inerted and this does become a threat to the containment structure.

S 5-8 I.

       . -4 it lw 2; . Containment concrete decomposition. Concrete-attack?by hot core _ debris.results'in the genera-tion of both; steam and: carbon'dioxidefga'sesJto-add ~to the'aiready heavy containment pressure loading.        (For each ft3 of. concrete destroyed, 8'lb of steam and 36 lb ofl carbon dioxide are                                                ,
                                   -formed.)

q-

3. Ex-vessel steam pressure increases. These events occur when the hot core debris encounters

                                   .'w a t e r . A short duration.butLvery high rate of-steam generation occurs when residua 11 sensible heat in;the' core debris mass ~is released to the-water,-resulting.in containment overpressure if it-is'already well charged by' previous events.

4. Ex-vessel steam explosions. This phenomenon is considered very unlikely and its magnitude is still very much in controversy among the tech-nical community. It must.be given consideration ast an assault'on the containment to be dealt with in accident mitigation, especially in a.

containment already pressurized by previous. events.

5. Ex-vessel hydrogen generation. This end-state ,

I results when hot steel and any remaining Zircaloy in the core debris contact hot steam. ., The reaction adds more hydrogen to the contain- i ment at a very crucial time.

6. Residual heat load. This is produced by the radioactive fuel decay energy and results in a slow containment pressure buildup to failure unless heat is removed.

7. ATWS steam generation. This end-state is almost impossible in the pressurized water type of reactor and is not considered further.

To control these accident end-states, a. complete mitigation  ! system must be capable of the following functions. However, value/ impact analysis may show that some components cannot be justified at certain sites.

1. The hydrogen generation problem--both in the Reactor Pressure Vessel (RPV) and subsequently in the core retention area--must be dealt with to preclude burns and explosions. WSP-90 5-9 l

t i systems, for example, are not inerted in normal I operation.

2. Adequate long-term heat removal from the contain-ment during the accident and as long as residual energy in the core is being generated thereafter.

The current WSP-90 design has 25 MWt removal l capacity with electrically driven pumps.

3. Adequate long-term cooling of the core debris mass once it arrives at the holding area. None ,

is currently provided in the design.

4. Control of core debris attack on concrete during its course from the RPV to the retention area.

5. Missile and thermal radiation shields to protect the seals and penetrations from assault are i necessary. This is generally done in the basic y containment design. f

6. Vacuum breaker valve system to preclude containment underpressure whenever steam is condensed in quantity in the containment.

The next stage is the design or selection of appropriate l mitigation components to meet the above requirements. As in i the preceding chapters, these designs were based upon the assumption that the following conditions would prevail:

1. All electric power for cperation of prevention and mitigation equipment has been lost. Con-trols and instruments may be inoperative. Note that 90 percent of the transients leading to core melt are initiated by loss of alternating current (AC) power.

2. The normal and emergency core cooling systems are not functioning.

3. The normal containment heat removal system is inoperative.

4. The core is essentially dry and the temperature has increased to the point of core collapse.

Meltthrough of the RPV is impending. l S. Molten steel will accompany the core debris to l 1 the retention area. 5-10 l

       .   .                                                                     I

6. A large portion of the Zircaloy cladding on the fuel pins will have reacted with the hot steam to form hydrogen gas.

7. Plant operators are not necessarily available to initiate corrective action.

5.3 STRATEGIES FOR THE ADVANCED PWR CONTAINMENTS Very little information is available on the containments and mitigation equipment details for the Advanced PWR's cur-rently under study and development by the reactor manufac-turers and other groups. We have information that the new designs include improvements for the prevention of severe accidents but little is known about what is being done to mitigate such an accident once it is in progress. Because of this lack of information some of the comments madeSec-here may be unnecessary once correct details are known. tions 5.4 and 5.5 describe two types of equipment necessary to mitigate a core melt type accident. Section 5.4 covers the additions or modifications necessary to provide complete mitigation for a conventional high pressure large contain-ment, while Section 5.5 describes an alternative--a smaller, basically unpressurA=ed containment that is always open tc the atmosphere through a condenser and chilled filter system designed to rett.in all radionuclides released, even during a severe core-melt accident. This alternative was originally designed to provide improved severe accident mitigation at a low cost but it turns out to have several advantages even during normal operation as will be shown later. 5.4 MITIGATION STRATEGY USING CONVENTIONAL COMPONENTS The preliminary design information available on the advanced pressurized water reactor has shown that much consideration has been given to improvements that reduce the probable frequencies of a serious core-melt type of accident, i.e., good prevention. In fact, the risk of serious accidents may well have been reduced to the point where the cost of many mitigative actions cannot be justified; however, this must be established. When these large dry containment systen.s must handle core melt accidents without serious consequences, equipment must be considered to handle the following conditions: 5-11 i

y , e

   ;t.

L 1 e In almost.all core-melt accident scenarios enough hydrogen is generated in the RPV to result.in sectionally burnable gas mixtures. This becomes i. a serious containment overpressure' threat.that represents the principal residual risk for this type of system. The risk can be avoided by I operating.the plant with an inerted containment atmosphere. Although normal operation with an inert containment is costly (Estimated

                 $2,000,000/yr), troublesome and pessibly hazard-ous to employees, it appears to be the most-             ,

practical solution. Post-accident inerting does not appear. practical for a PWR as the' steam from a primary circuit rupture is too contaminated to be. vented overboard while the flushing gas is being injected. .' Post-accident inerting of this l'I Very large containment (3,000,000 ft3).also may take longer than is permissible. Consideration of other alternatives confirms that pre-inertion i is least objectionable and least costly. The cost .I summary shows liquid nitrogen for this operation but other gases may be used. e The present heat removal systems (four 6.3 MWt units in some designs) are inadequate to prevent serious long term overpressure conditions after a core melt. Figure 3-3 shows'that the cumulative l, decay energy to be far in excess of the capacity with the four current units fully operational. In many dry containments this heat removal is done l' by surface type air coolers using electric fans for circulation. Without these electric fans in operation, the heat exchanger surfaces tend to j blanket with stagnant air or gas resulting in a serious reduction in capacity even though the cooling water circulation continues. Direct cooled water sprays are more reliable in this respect but this water will be contaminated and therefore must be kept isolated. For the value/ impact studies a completely separate redundant direct diesel engine driven pump and heat exchanger system to cool the water, similar to that described in Section 3.4.1 (page 3-27) is included. At some plants the present design can be upgraded, making this added system unneces-L sary. In containments having a free-standing steel shell, it has been proposed that spray l cooling the exterior of the shell could be a i useful heat removal system. We have not assessed 5-12 i

this possibility An external low pressure water supply may be necessary if the 3-250,000 gal tanks already included are considered inadequate. After a severe accident this water may be con-taminated so the tanks must be shielded.

              .       e  Ex-vessel core debris retention and cooling to minimize concrete decomposition as well as hydro-gen generation is needed. Without these pro-visions, this gas generation can result in long-term containment overpressurization. Also, with-out proper core debris retention and cooling,

• basemat penetration and a radioactivity release to the biosphere will eventually occur.

e The large volume dry type containment has no requirement to accommodate high capacity venting from internal steam generation. This is espe-cially true when adequate heat removal and con-tainment sprays are in operation. This equipment minimizes the need for air inflow as the internal steam is condensed, so that the present vacuum breaker valve system is probably adequate during a severe accident. All these mitigation requirements are recapped in Table 5-2 showing how each type of severe accident failure is handled. If all of these mitigation components are incorporated into a new plant the seriousness of a core melt type accident will have been alleviated. At this juncture it cannot be established whether this action is cost effective. The estimated cost of these proposed mitigation component systems is based on similar equipment previously estimated for other containment types and as shown in Table 5-3. 5.5 ALTERNATIVE STRATEGY--UNPRESSURIZED CONTAINMENTS A completely different containment strategy to achieve miti-gation is proposed here. It is based on providing ample and continuous containment venting to the atmosphere so that the internal pressure in the dry vessel can rise above ambient only very slightly--enough to drive any evolving gases ard/or vapors out through a large filter system, capable of trapping krypton, zenon, and lodine, as well as all particulate. This filter system is designed to handle all steam and noncondensible gas flow, so that the duration of the small pressure rise above ambient would be only for a few seconds. Thus any leakage through seals and penetrations would be 1 5-13 f

l

           .L l

4 l l l ' l TABLE 'e-2. DOMINANT FAILtRE CDES FOR MYACED NR IN A PRESSURIZED CONTAl*ENT Symbol Accident Event How Handled j 3 In-vessel stese explosion Rugged surrounding borriers Ex-vessel steen orplosion - Caref ul, expensive design 6 f ailure of opening and and test with good penetrations maintenance  ;' Mydrogen - displacement Operate with I orted Y - burn atmosphere  :

                                                                      - explosion                                              I Early occident overpressure        Adecuate, dedicated heat 61                                             removal f rom steam or hydrogen l

Late occident overpressure Adecuate, dedicated heet 62 from steam spike, carbon removal dioxir's or hydrogen C b5 emet penetration by hter cooled hot core debris dry crucible y laterfeclag systems %n-eltigable LCCA (RSS seys high risk) Use bettr maintenance I ATWS High steam flow rate of Containment rollef limited duration pressure in minutes , l Under Geses washed out by steam Venting not considered Pressure which then condenses hence small breaker valve (E l l l l 1 1 5-14 1

r. .

i-l TABLE 5-3. MITIGATION SYSTDI COST FOR PwR-LARGE DRY PRESSURIZED CONTAlWENT Function Mitigetton Component 1/1000 Dedicated Cooling Plant 2085 Heat Removal Shleided Wety Tank 1675' l Containment Spreys 565 ' Sub-Gosamet Dry 2295 Core Dobris Debris Watercooled Crucible inerted Atmosphere ** 2000 Over Pressure Large Cepecity' 3573 Protection Hydrogen Recombiners Under Large Brooker Velve Not Reculred Impact - Cost in 1/1000 12193 Value or Benefit 50 mi radius --- men-Rom Averted 500 mi redlus theect/Volue Antio 50 si radius - In 1/ men-Aom 500 mi redlus

                                                                  'Some plants alreacy have edeouste cepecity
                                                                "'First ccat shown - extra maintenance cost estimated at 5.2E6 annually 5-15 i

(

L

• L inconsequential in its effects. When this non-pressurized containment is coupled with a dry crucible core retainer and adeqvste heat removal the result is a plant _ with the capa-bility of coping with any severe accident without risk to the public.

Because current large dry containments have little internal thermal capacity other than a pressure change in their.large volume, i.e., no suppression pool, the, proposed external miti-gation alternative must also provide steam condensing capa- i. bility roughly equivalent to the suppression pool in BWRs or # the ice condenser PWR types. After.the steam content is' . condensed, the remaining effluent gas flows to a large and , very cold filter to detain the contaminated noncendensibles. (See description and schematic in Section 5.5.1 following.) Although it is possible to use a water suppression pool with the chilled filter, an ice condenser is recommended. The ice l system is preferred because it almost eliminates any water vapor entering the charcoal bed, and because Westinghouse has both design and operating experience with it. The impact of l 10 CFR 50, Appendix K, would have to be evaluated because there is now only atmospheric backpressure on the RPV which may affect the reflood phase of a Loss of Coolant Accident (LOCA) recovery.. Changes'are expected that would simplify this legal rule. All components of this vent and filter system have been installed and operated at one reactor or another. Ice condensers are in place at over 10 U S. FWRs. A large l-duct activated by pressure connects multiple CANDU reactors to the common deluge building. Charcoal filters are commonplace in reactor cleanup systems, and their operation at low temper-ature is thoroughly proven. A discussion of the!" application here is given in Section 3.5.2. While the exact specifications are not well established for the ice condenser system and the large fiJter system (being somewhat site specific), a preliminary cost estimate is given here for comparison with the rest of the mitigation system. The size of the ice condenser is based on that used at the Sequoyah plant (but enlarged to provide a reserve) and the activated chilled charcoal is based on extrapolation from prevdous chilled filter data. It should be noted that L only one condenser and filter system is necessary for a dual or even a multiple reactor installation. l i This combination of open containment, condenser and filter , l permits relaxation of many demanding and expensive contain ~ ment specifications, and, more importantly, provides a , I 5-16 i i

                                                                                     )

   .j*

practical way of disposing of the hydrogen gas generated (i.e., by ignition). This separate mitigation unit is shown schematically in Figure 5-4 and, together with a dry cru-cible core catcher and an adequate dedicated heat removal spray system (not shown in schematic) as described in Section 5.4, provides an adequate response to all of the accident events listed in Table 5-4. It appears that the cost of this separate condenser / filter system can be more than offset by other savings inas the initial cost and the operating cost of the plant, follows: e The main containment can be smaller, as shown in the comparison chart, Figure 5-5. This contain-ment need be no larger than necessary to house the machinery and provide needed working space. The minimum would be about the size of the existing ice condenser containments, without the ice space, or possibly as low as 900,000 ft3 vs the current 3,000,000 ft3 It appears that the containment diameter might drop from 150 ft to about 110 ft. There would be a substantial saving in construction steel and concrete. e This containment can be designed for some low pressure--possibly below 5 psig. Probably the shield building u'.ed to protect against aircraft and seismic assaults would handle this low prec-sure as designed. e Many complicated penetration seals would be sim-plified. These could be designed for low pres-sure duty of a few seconds duration. All leakage specifications could be relaxed. Frequent testing for leaks would become unnecessary. A great many of the lines penetrating the centain-ment wall could be protected with single instead of dual isolation valves, o Inerting the containment to handle hydrogen would

• not be necessary. Igniters would burn the hydro-gen as it is generated or wherever it contacts sufficient oxygen. The resulting hot steam would flow out through the large duct (s) to the exter-nal condenser if it is not already condensed by the internal cool water sprays.

5-17 1 a

         -.m_ _ _

i

 ~

r o t s r e . n o p ) e- 1

                                                                                                                  ?f/

m H r)]l'

                      .'l
                                     "          l

. .. d[

                                                                                                        ,31       f t

i i

                                      -                                                                                             n U

4 h-t ny

                                                                                                                                        )

i 1 qL J od w I b . i' h if. i u w*l .\- b\r l l tt

   ,.          N                                                               '

N 4, aS g

                                                                               -                                                '  i0 I

fl - ~ t9

                                                                                            /-

NI f $ i - e bYC $ I

k. MP m

                                                                                                       %. LpJ                       eW S

j'T M I T s Y= [ r o t t ay rr aa pn

                                                                                                                                               +

3"- ]i c

                   /
                                                ,

N a e ei Smi l o o c r t w 0 .% .. !. N. I

                                                                                                                    ?

N R fl oe r w e  : h cP h , C e i e  : t t e y , a: r c n o

                                      '                     ~

f.

                                                                       .,]/_

m. ef he C

                              , / i' tk                                                                                                                            cR nc eo                                                                    ,                                                                      -

S( vt s i

                    ,     i                                                                                                        )

s U f N f,

                                                                            ,U u

eg r a ur t h pp m ( 4 ue e - . L Rd 5 e n e _ n w r _ o ,u u it ,', l e g u i s n \ F _ I n _ i o t e t e' r de e "* r ir g 'P f ec e R I

                                                              /

wh" e

J TABLE S-4 DOMINANT FAILLRE MODES FOR ADVANCED PWR IN A MESSURlIED CONTAl*ENT ( Accident Event How Handled I Symbol l-in-vessel steam ocplosion Rugged surrounding barriers cs Failure of openings and impossible - no pressure g penetrations Wdrogen - displacanent Continuous Ignitors with Y

                                                      - burn                    open vent to condenser
                                                      - explosion                                                                           .

M pressure problem but Early accident overpressure need dedicated heet gg due to steen or hydrogen removal - vented to condenser Late accident overpressure from steen spike, carbon same as above 62 dioxide or hydrogen Release consequences reduced by no pressure Basemat penetretton by et w nthrough. Water _ C hot core debris cooled dry crucible  ! best Low pressure short V Interf acing systems LOCA duration pulse mlnleiros High steem flow rate of Overwhelms condenser if ATk3 limited duration longer then ~ 20 mins. No valves required draws Under High steen condensation air through filter and Pressure rat e condenser 5-19

l. j (S/guygt) aansoul uS sep  ! nuswueetuco

                                                                   .R          S       S           S                 R              R-

• a =

i i i i i i 8 4 l 5 t -- ,, e l-5 , e \ e 1 > b,4 , 4 A I c owntoA , , g > I b 1 \ m 6/g' 'u'dq'i o'g' . 's'; is';'d ' !5 i'd o e,, ,m ,,,,,,,,,,,, , .. , D ke owntoA c 'l

                                        -                                                                                                                    4 m
                                             .2                                                                                                              9
                                                  '                                                                                                          g                 \

1 [- $$$l....................f

                                                                                                      'M""a' 5" ""'" "'?'o                               -

O i g W ownfoA > , I, R M-La Cm oa gC t E' S/guyqt gg . einsesM usesec .

                                                       $$ '                                                                               Mnl0A              cc O2 0 3:

wo l

                                                                     ,g                5/guyqt g1. sanssead u6es*O g                                                                                       O owntoA            eH
                                                                     .p   6                                                                                .e4 C uo-4 e

41

                                                                                                                                                   '         EN
                                                                                    =

ea,

                                                                                                    'S'/gUi/higf.'da'nEa'5i[6 s'ao'      ]t
                                                                                  !d      ,

oW"PCA 3 m

                                                                                                        ......................                               t zae2N.                              ..                 .               . ... .

owntoA 4 )

                                                                        ,                                                                                    2                 4 m                 \

l I I l I l i I i 1 *(  !

o. * * *. N o e e v. N o  !

N - - - - - o o o o j CII9 0g ut mnton sea; tau tuewu!r2uoo 1 1 e 5-20 , i l l

                                                                                                                                                                         . i

[-g- - .. e Servicing-the external condenser and filter equip-ment could be done readily as it.would be.neither _ pressurized, radioactively norethermally hot. Little need for maintenance.is expected, and inspection could be done with. view ports. How-ever, if needed, personnel could enter the filter while the plant.i.s fully operational. This new arrangement would leave the main containment machinery bays much less. cluttered, which should facilitate maintenance there. The total cost of this complete mitigation system at a new , plant is presented in Table 5-5, using,an ice type condenser and including dual refrigeration systems. A detailed break-down of each cost component is given in Appendix A. While this set'of open containment mitigation components has about i the same' estimated total cost as does the standard pressurized  ! containment mitigation system given in Table 5-3, there are other savings that cannot be included in the estimate, notably the much cheaper containment structure. More importantly, survival of the nonpressurized system is inherently more cred-ible and it will be perceived as safer by the public. It is completely passive during an accident. 5.5.1 Description of the Condenser / Filter The basic components of this large filter equipment can be quite standardized but the building itself must be designed for a specific reactor site so Figure 5-4 can be only illustrative of the concept. The basic functions to be accomplished by this filter system are: , o Condensation of any water vapor (steam in the j entering gas mass). o Capture and retention of any particulate or aerosols entering with the gases. (A high decon-tamination factor is necessary and possible.) e Chill the noncondensible gases that pass through

  -                                                              the condenser section so they enter the chilled          )

filter very cold. I e Capture the heavier gas molecules (the contamina-tion) on the extended surface of the activated charcoal by adsorptica and allow the lighter clean gases to pass. Refer to Section 3.5.2 for details. l 5-21  ! l 1

t ., p- -. .M w; , . .i.

     ,   t i :-

3

                       ' TABLE 5-5        MITIG^ TION SYSTEM COST FOR PwR-LkRGE DRY 'NON-@ESSURIZED CONTAl*ENT -                       l Function-                   Mitigation Component           S/1000                           .

Dedicated Cooling Plant- 2085 . Heat Removal Shleided water Tank 1675* j* l i Containment Sprays 565 '! Core Doorts Control Sub-Basemat ky 2295 Watercooled "*ucible Hydrogen Ignitors 300 Throughout Pressure Over Protection and Large Ice Condenser 5746** - Under and Chilled Filter with Refrig. Plant impact - Cost in S/1000 12399 Value or Benefit 50 mi radius --- man-Rom Averted 500 mi radius -- impact /Value Ratio 50 mi redlus -- In S/ man-Rom 500 mi radius -- l

                         ' Sane plants stready have adequate capacity
                       **Detall cost breakdown not included                                                                t f

i i s-22

o Allow reverse flow through the system to avoid negative pressure in the containment. The first section will have about 2,000,000 lb of ice held in racks for good exposure to the water vapor laden gas passing through. The ice will be held at about -80 F by the circula-tion of cold air from a refrigeration plant. The low tempera-ture is chosen to improve the thermal capacity of the section, to minimize sublimation and melt / bonding of the ice itself and to ensure that cases enter the charcoal filter at low temper-ature. The layers of ice will have plenums between them to minimize short circuiting. Condensate and contamination will drain bacx to the containment sump. Preliminary designs indicate the ice annulus (Figure 5-6) will be about 20 ft wide around the central charcoal filter and 60 ft in depth. It is anticipated this ice section will be chilled by air circulation so it can be entered for inspection. The second section has several layers of activated charcoal with enough flow area to allow the noncondensible gas to pass at It' locity. The total charcoal content is about 50 tons. Eact. layer will be treated to efficiently capture specific radioactive gases, e.g., one section treated to capture iodines, another specifically for xenon and krypton, etc. All radioactive gas fractions are heavy and are adsorbed more readily. This filter is about 20 ft in diameter and surrounds the vent stack. The lighter gases (nitrogen, oxygen, carbon dioxide and hydrogen, for example) pass the filter and are vented up the stack to atmosphere. The two main sections must be isolated from each other and insulated from the atmosphere and the containment proper during normal plant operation. This is only necessary to minimize the refrigeration load and any small leaks will not preclude proper operation. This isola-tion may be effected by large light door-like valves or light rupture diaphragms. They must open to allow inflow of air as well as out flow of gases. Refrigeration of this absolute filter system will be done by a dual unit plant located close to the filter building. Each of the two units will be a cascade type freon system with elec-tric motor driven compressors and water-cooled condensers complete w?th evaporators designed to deliver -85 F dry air and nitrogen f.ft. (Air will be circulated through the ice racks and nitro-gun through the activated charcoal.) Each compressor provides about 10 tons of refrigeration (120,000 Btu /h) at the low l temperature and draws ~66 bhp. When the blower and water pump loads are added the plant requires about 85 kw maximum. 5-23

ro n s e a t s ee evesus _ e es pk s e D m Vs ue d ,s e t d /

                                                                                                                                                                           ,o           s.-   g-1 s

S y r i r o e m ' s t

                                                                                                                                                                                     -                                se n
                                 **                                ts e

rc mg o o lu c l i n o W s ,.' N' *'s, s ac F C hs Wtn e

                                                                                                                                                                  .V <

w s d i ,e, n vk

                                                                                                                                                                          >l            ..            '

d ee t h C a a 4 g l- r

                                                                              ,                     -                                                            .                                    '                   e s
                                                                                                                                                                                      ,3a j              ,                                :n $
                                                                                                                                                                                                             ,.}lg 4             d                                                                    ,'.
                        '1 m

T' t n t k i r c ~' _ Mw f M . T T e 3 E q t d e n eo 2% o a n Vt s a i s n a g ak i i sl-C W n e i 1 y u 1 n s 1 l i i c I I I l  : I l s 2 I I i

                                                                                                                                                                                                                 '        e g                ,
                                                                 '     ;               :            '                       :                                   jf 5'    ;?' *                     $:                   t f                   1     I' 1         1,,

5

                                                                                                                     +1)               I f
                                                                                                                                                         ~;
                                                                                                                                                         ]                       -

s a I '! I 8, r a m g I I EiI I p j I I I I I O, E Ex'S I e 6

                                                                                                                                                                                                               /

i i 1  : - s , l, g l a i kk..o ii . l o .. Ma 1 a a ' o . o 1 o i n c ra 5 ke . i c. 2i l 1 4a g c

                                                                                                                                                     .$     s                                               ,

o c h3 Mi . I

         ,                                                                                                             1                a                   s                                                          o

1. D h h t i

      ,.                                                 i /                     t c      1 C

1 C s e C i /' 'l 1 ' t

                         ~                                                              i                                               i is                               lu a

7 { . '

                                                                                                                                                                       ,9           n                                  c 1N r

e fp e i n I I i v c m e n I a n T n e e i I a g h u z I I i a i.o e o r c 1 t S [jC I I ' a . e

                                                 ,                                                                                                                                   e                                 n s                                              gc**                     I m

I I , x i eg d e ' s v e

                                                                                                                                                                                                                         )

v I e te l e v i s s I a n a l h e U e Ng I I fk f (

                                                                         ]

g c I  ; g a I I I I Y w n I -

                                                                                                                                                                . i,                                                         .

c ' 5 ) A; s 6 M[;v 5

                                                                                                                                                                                                   $*                     e t

e nrs e a iove so te o pt t c m u k', P- i F r u g w o msul " t tene em

                                                                                          ,, M                                                                                                       9 m,yA

i 1 i I l During normal operation only one unit will be operated at light load. Once this filter and condenser system has been brought down to temperature it is not necessary to maintain refrigera-tion plant operation during a severe accident event. It is a passive system in this sense. After such an accident, however, when the filter has been charged with radioactive materials that are releasing decay heat energy, the refrigeration plan' is needed. Several hours are permitted to bring it into operation. 5.6

SUMMARY

1 The Westinghouse advanced PWR and its large dry containment incorporate many improvements intended to reduce the fre-quency of sequences leading to severe accidents. The degree of success of these efforts is still to be evaluated. How- 1 ever, these studies make clear that severe accidents can still occur and that they will cause containment failure of 1 the same types as before--hydrogen fires and explosions, slow overpressure failure, and basemat penetration. In considering possible mitigation systems for these plants, an important factor is that none have yet been built, so that backfit installation is not a concern. The most out-standing characteristic of the large dry containment system for WSP-90 plants is that it has both the largest internal volume and the highest design pressure yet proposed. Thus it is by far the most expensive containment system. The design studies show that a conventional mitigation sys-tem can be adapted to the WSP-90 plant at a cost of about

                    $12 million. At this stage there has not btMn sufficient PFA analysis of the plant to determine how much, if any, of such a system would be cost-justified. For about the stme sum, however, the complete chilled-filter vent mitigativn system can be added. This system would reduce uncertainties to a lower level and produce operational savings in normal plant use. Its principal attractiveness, however, lies in that it need not be added to the very expensive standard containment, but a much smaller, low-pressure enclosure.

The savings in constructing this simpler enclosure should certainly far outweigh the chill-filter vent system cost. Although further study and analysis is needed, the conclu-sion seems clear that complete hydrogen control and mitiga-tion, with vastly improved enrtainty, can be had for the WSP-90 at a 3swer votal cost than for the unmitigated stan-dard desipr.. The cost-effectiveness of this kind of miti-gation thus does not depend uptn accident-frequency analysis. l 5-25

l i e. i I, i e t* 1 1 i 1 1

                                    )

1 1

q y s .n b I APPENDIX A. . COST ESTIMATES OF MITIGATION COMPONENTS. o This Appendir presents a cost and installation breakdown for each of the mitigation components and their alternates shown in_our Tables of: Mitigation Options. These initial cost estimates:are intended to permit an evaluation of the_value/ impact ratio as presented in Sections 3.6, 4.6, and 5.6. .In this way a selection can-be made of mitigation systems that might justify additional study. For each component or system, the installed costs are esti-- mated'for three different plant' status situations: "A" costs . are based on a new plant not yet completely designed and with no site work started: "B" costs are based on a plant about midway.in the construction process, so that. mitigation addi-tions can be made with no radioactivity present, no startup delay, and with minor rework: "C" costs _are based on going

             .into a plant that has been operational but is currently under-going a refueling outage. In this case it should be noted that downtime and replacement power charges are not included, since we believe that with adequate planning most mitigation installations can be made within normal outage periods. The "C" estimates do include the cost of radiation shielding, remote _ tooling in some cases, and the cost of draining the system to be worked on. A number of Mark II BWR plants are operating to which retrofitting under status "C" conditions would apply, but Advanced BWR or Advanced PWR plants have not been built, so that only the "A" status need be con-sidered. For components applicable only to these latter cases only the "A" costs are given.

Generalized cost estimating such as is presented in this , Appendix is difficult to compare with other experience, because the subject is general and costing methods vary widely. The most practical way to present these costs is to state clearly just how they were obtained and to permit the reader to adjust them as he feels is necessary. We believe the estimates given are conservative for high quality indus-trial practice, but we have deliberately excluded st:h fac- j tors as nuclear QA/QC, profit, insurance, and land charges.

   +          Our estimates show that inclusion of these if required could increase the overall costs shown by twofold or threefold.

For consistency in all the altern'atives, the cost estimates are based on the following conditions: Labor Costs. Direct labor cost estimates include fringe  ! benefits, portal time pay where necessary, insur?.nce , inci- j dental overtime and contractor mark-up based on forecast i A-1 l

n; I

i l l t 1984 costslat $350/ man-shift for all shifts. One; man-day is then $1020: (Sperry-1982). Escalation. ' Inflation during construction is covered'by a 4 percent addition to all labor and material costs. -This is more than 15' percent annually in most of our cases since the contracts are'short term.

                                                                                           )

J Desion Control and Field Administration. This item covers the , control of. drawings, changes, inspection, field engineering-and administration of the projects in the field during con- ' struction. These charges are covered by.a 12_ percent addi-  ; tion to all direct labor and material costs. In the cost breakdown shown this overhead component is included with l inflation escalation.and listed as " Field overhead--16 percent." Purchasing Costs. All direct purchases of equipment have a. 15 percent handling charge added to cover alternatives, varia-tions, shipping, tax, etc. Supervision. For simplicity, all on-the-job supervision is included at one average value regardless of the class of man involved. The $6000/ man-month figure used includes their own administrative overhead, employee benefits and vehicles where necessary. The more complicated installations--such as the underground work--will require one general superintendent, one project manager, one project engineer, an office engineer, a cost-engineer,-a field design engineer, a safety engineer, an office manager, a purchasing agent and an accounts manager. l This team costs $54,000/ month and this amount is used for all supervision. In many cases the supervision on the job will L handle several different component installations under way at the same time so each mitigation component may have only a portion of the total supervision charged to it. This is referred to as Supervision--chargeable in the tabulations. Contingencies. On preliminary cost estimates such as these, where detailed drawings do not exist and the course of action cannot be specific, a contingency factor must be included. Twenty-five percent on all components and their installation is used. Profits. Profit is not included as a line item in this estimate since it is not known how the work will be done or who will do it. [ A-2 i

i . 4 t i Time Estimates. All work schedules are based on two' shifts L - of eight hourteach.and one premium shift of seven hours six L days per week. Insurance.- No liability insurance is included.since it is not known who will be doing the work. Many utilities have a general coverage for this. Quality Assurance /ouality Control. At the request of NRC project management our cost estimates have not included a factor for Quality Assurance and Quality Control but do provide the highest quality commercial grade equipment and construction available. On components and construction . materials where QA/AC formal control is required the cost can be doubled. Seismic Quality. The consensus of most knowledgeable engineers and construction men is that building a new plant to a higher seismic requirement does not markedly increase the fabrication and installation costs but it does add to the. analytical engineering cost (e.g., vibrational resonance analysis of complex structures). Typical Specific costs. Unless conditions indicate they should be higher, the following specific costs are used in the calculations. The costs are "in-place" and include direct labor overhead only. j e Reinforced concrete $600/yd3 e Gunite with steel mesh $6.00/ft2.in e Structural steel--at plant $0.50/lb

                                                                --in place        $1.50/lb e    Machined components               $4.00/lb e    Rotating machinery                by component (mfgr. quote) e    E17ctrical and control            by component (mfgr. quote)

All construction costs vary markedly with the time when work is done, the location where it is done and especially the

     -                                construction business activity at the time the construction must be executed. These imponderables cannot be predicted, therefore the costs shown are an effort to reach some mid-continent mean for about mid-1984.

Cost estimates are provided for the following mitigation components: A-3

3 l l

1. Dedicated, separate containment heat removal.

2. Drywell spray system.

3. Core retention and cooling (three types).

4. Venting and filtering system.

5. Hydrogen control. .

6. Underpressure control system. *

7. Large chilled filter system.

A.1 DEDICATED SEPARATE CONTAINMENT HEAT REMOVAL I Two heat removal systems are considered. They each have the same equipment and the same operation and performance; how-ever, their location is different, which changes the costs. Regardless of the location, each of these plants is a dedi-cated, completely redundant, diesel-engine-powered pumping plant with equipment to handle both the suppression pool water being cooled and the cooling pond water to receive the heat. Each pump delivers its water through one side of a large heat exchanger system (-45 MW thermal capacity) before it is returned. Cooling water goes back to the pond but the suppression pool water (which may be contaminated under accident conditions) goes back to the containment vessel either directly or via the spray system (see Section A.2). As detailed in Section 3.4 and as shown in Figures 3-5 and 3-6, all the equipment is redundant. Cost estimates are made for a dedicated plant above ground but completely separate from the main plant, and also for the same equip-ment below the basemat at the junction of the access tunnel and the caisson containing the dry crucible core-catcher. Table A-1 lists these costs for the surface located plant. For the underground location, the costs are shown in Table A-2. Note that it is not practical to consider the underground siting cf the heat removal system for a new plant, since the tunnel would be superfluous. Hence there would be no access to the engine-driven pumping units. Accordingly, only Status "B" and "C" are shown. i [ A-4

                                                                                                                                    ;" {-
  ; j is ;                 I Id. l qj           ,

TABLE A-1 DED4CATED SURFACE SITED HEAT REMOVAL SYSTEM'

                                                                                 - (In 5/ 3 000)
                                                              ' Plant Status                              A           B-     C
                       ^

Field . survey, engineering and ' design (6+ men-months) 60 60 60

                                     ' Separete revetted protective shelter,1800' f t2 at $167+

(excavation included) 250 300 300 JTwo diesel-powered pumping units complete with:

                                           . Enginas (2) . 400 Bhp (Caterpillar,1983) 525,000 es. .        75.         75     75 Pumps (4) (Ingersoll, 1983) 57,000 es.                       45 -        45     45 Deepwell pumps (2)                                           -           50     50 -

l .Besepistes (2)- 15 15 15 Coup!Ings (4) 2 2 2 E.xheest silencer system (2) 25 25 25 ' .

     .c                                                                                                                                 -

Electric ster %.c betterles and chargers (2) 8 8 8 Fuel supply (2); 18,000 gal tanks, pumps, lines 40 40 40 Penetrations (2) 50 100 200 Pool inlet system with screens (2) 50 50 50 Piping to pool- 125/ft- 30 30 30' isolation velves et penetrations (4) 70 70 70 Control system-nonelectrical (2) 75 400 125 Heat exchangers--45 M"t 9000 ft2 (2) (Young, 1983) 5328,000 es. 250 250 250 Assembly and Installation (12 men conths) 120 120 120 Pond water piping with valves--installed 180 200 250 SUBTOTAL 345 1550 1725 Fleid overhead (165) 215 248 276 Supervi s ion-chargeable 108 162 215 Contingencies (255) 418 490 - 554 TOTAL g g g

   ,                                                                                                                                   j 1

l 4 A-5 1 I

  ' _ [-- .

l I j TABLE A-2. DE0lCATED UNDCRJROUND H?.AT REMOVAL SYSTEM (in 5/10001 I 1 Plant Status B C

                                                                                                   -l l-l Design and engineering                                      50     -50'                  -l Concrete-ilned, pressure-tight cavity added to tunneI:-                                                                         j Excavation                                             185     185 Lining                                                  75      75 Foundations                                             25      25 Plant equipment as above                                   210     210 Inlet piping to caisson                                     20      20 Isolation valves (2)                                        70      70 Controls as above                                          100     125 Heat exchangers (2 sets)                                   250     250 Assembly and install (12 man-months)                       120     120                    l Piping to spray system                                      80     150 Pond water piping Installed                                200     200
                    .SLETOTAL                                         1415    1560 Field overhead (165)                         226     249               .    ]

Supervision 162 216 Contingencies (255) 450 506

- TOTAL 12254 12532 l

4 i i 1 1 Y i A-6

                                                                                                        )

I 1

 . ,; v       . iI v

A.2- DRYWELL SPRAY SYSTEM

                                       'As described in Section 3.4.2, there are two types of drywell spray systems under consideration for mitigating Mark I containments. The first alternative consists of two circular-
                                       . pipe headers outside the containment structure, each fed frem one of.the diesel-driven pump units after the water nas been
                                                                                      ~

through a heat exchanger for proper: cooling. Each circular header pipe around the upper containment delivers water through about six small.(-6-in diam) penetrations ~(with double check valves) to internal spray nozzles. 'The cost breakdown in Table A-3 for the external system presumes that the cooling plant is.-300 ft away. TABLE A-3, EXTERt4AL DRYWELL SPRAY SYSTEM

                                                                                                                                             .(In 1/1000)

Plant Status A B C 30 50 70 Engineering and design

                                                                                                                                                                                                       -     72   72
                                                                   . Penetrations (12 small at $6000 esen) 60    60   60 Sleeves (12 et $5000 esen)

Pipes, checks, and nozzles 20 20 20 12 12 12 Flexible connects to heecer Two 12= Inch headers with supports 50 50 50 Two 12-inen pipes to heat removal unit 25 25 25 20 20 20 Breckets 80 100 125 Instelletlon (8 non-months ) 297 409 454

                                                                         $UBTOTAL a

48 65 72 Fleld overhood . (165) Supervision chargeetle 108 108 162 113 146 172 Contingency (255) TOTAL $565 5728 $860 l A-7

o 1 l l The second type of drywell spray system is only applicable to the cooled, dry crucible type of core-catcher where the heat removal system is below the basemat in the access tunnel. In this design four spray supply lines come off the dry crucible cooling jacket top header, extend into tne suppression pool and up 'hrough four downcomer pipes to the-drywell levels above the diaphragm floor. Each of these four lines has multiple spray heads that are directed into the levels as necessary. The cost of this completely inter-nal system presumes that the dry crucible is being installed . concurrently (see Table A-4). TABLE A-4 INTERNAL DRf* ELL SPRAY SYSTEM (ln 1/1000) Plant Status A B 0 30 45 45 Engineering and cesign 2 ;0 10 Basemat holes (4) Four sets of plaing (S In. at 150/ft) 25 35 50 Spray heacs (12) (5 mn-months instailed) 50 50 70 Radiation protection - - 50 SUBTOTAL 112 145 215 Fleid overhead (165) 18 24 35 Supervision energesele 108 108 162 Contingencies (255) 50 69 102 TOTAL $298 $346 1514 1 l I I I A-8 (

                                                                                            \

i

    < o A.3    CORE RETENTION AND COOLING A.3.1    Core Distribution Via Diaphragm Floor This core debris retention system requires minimal modifica-tion for installation in the power planc, but its successful operation depends upon complete overtemperature protection for the numerous penetrations at the diaphragm floor level.

The core debris is allowed to flow out of the contral pedes-tal area at the diaphragm level into an inner annulus sec-tion of the protected concrete diaphragm floor. The radial-ly outward flow of the core material is restricted by a circular dike to a diameter just larger than the innermost ring of the 24-in diameter downcomer pipes. The thin steel wall of these pipes will promptly melt when the hot debris contacts them allowing the material to drop through the pipe down onto the flooded suppression pool floor at the basemat level. This ensures the core is well distributed in small can readi-piles under the numerous downcomer pipes where it ly be cooled. A bed of thoria rubble or its equivalent can be located under each pipe to keep the hot material off the basemat concrete. The cost breakdown is given in Table A-5. A.3.2 Central Basemat Core Retention and Cooling This core debris control, retention and cooling concept guides the core material from RPV through a fusible delay in the center diaphragm to the suppression pool in the center pedestal area as described in Section 3.4.3.2. As the hot core contacts the limited water in the central area, it fragments into pieces small enough to ensure adequate heat transfer for cooling. A porous bed of thoria pebbles, gam-biens, or their equivalent keeps the core debris separated y from the concrete basemat floor while water-filled steel ' ducts lining the inner lower pedestal wall help prevent the hot core from contacting the concrete sides. Equipment cost for this installation is low; but the diffi-culties of getting crews and equipment into the required i areas, of providing shielding against radiation, and of assembling the pieces in place are both costly and time con-suming where a retrofit will be made. One possible arrange-ment for doing this work was illustrated in Figure 3-14 where lead-shot or foil shielding is installed on a scaffold under the RPV during modifications. A sumtaary of the time for doing this work as a retrofit was given in Table 3-15. The minimum time to make the installation is estimated at 47 days with no time added for Sundays or holidays: probably at least 50 days from plant shutdown is more realistic. BWR refueling normally requires about 40 to 42 days, so 8 to 10 days of downtime is A-9

 -:;   e
   " ;l {                                                                                               ,
                                                                                                                          -o   .; ,   .

'  ::ib l[e L 1

                                                                                                                                  .f-l 1

l '.

TABLE A-5.. TRE DISTRIBUTION ON DI APHRAGM FLOOR i (in-5/1000)- l i

4 t Plant Status A B C Engineering and cosign 60 100 150

                                                                                                                                        ,4 I
                                                                                                    --        --     100

- Plant cool-down ,

Suppression pool dratn/ refill! -- -- 50 Enter containment. - '--- 80 Bring.in oculpment/shieles, - 100 150 t Diamond drill vent holes (clachragm level) - 60 100 1 1 Remove concrete plugs and oculpment -- . 20 2 j Dike ring (5-ft high. 45-f t diameter, thoria f aced) 100 100 100  : Install dike 100 200 250 Protect and cool penetrations with shieles and sprays 200 300 400

                           '7horia plates and gravel plus steel cover in central p ecesta l --i nsta l l ed                                            400       450      500                    :

Shield workers -- -- 200 Hign-temperature Deccle bed under downcomers-24 areas 120 120 180 SUBTOTAL 980 1450 2300

                                    .Fleid overhead (165)                                          156       232     368 Supervision chargeable                                         54      162     162 Contingency (255)                                            298       461     667 TOTAL                                                          $1488     52305    13335 Notet This estinete does not include the cost of - the spray system for penetration cooling.             j i

1 I A-10 i

  ..$4             -k
 ,E chargeable to this. mitigation device. This cost is not included but is estimated to be S720,000/ day per 1000 MW,.

The cost' breakdown for this central basemat core retention system-is-given in Table A-6. A 3.3 Cost Estimate for Dry-Type Crucible An alternative to the two wet-type core retention devices is the subbasemat dry-type crucible that holds the hot core debris mass'in a water-cooled, steel, cone-shaped vessel (see Section 3.4.3.3-and Figures 3-15 and 3-16). This unit can ' be installed as a retrofit to an operating plant with almost no interruption of service. While its cost in the "C" retro-fit installation is markedly more than that of a new plant

                         .("A" status).the saving.in plant downtime may justify the extra expenditure. Plant downtime costs are not included here.

Since the case study plant considered is a dual instal-lation and since some very costly portions of the tunnel installation will serve both reactors, this cost estimate is made for two units. .The design allows for drywell spray systems in each reactor containment, but only one dedicated heat removal system for both reactors. The costs for these components are not included in this section (cee Tables A-2 and A-4 for these figures). The cost of fabricating and installing the long steel cone-shaped dry crucible is calculated separately; then two of these components are included in the following overall cost estimate. The estimated cost of shop fabricating the core catcher subassemblies and completing the assembly, inspection and testing in-place is given below. This covers one unit 69 ft long, 6 ft.in diameter at the top, 3 ft in diameter at the bottom, with a 1.0-in-thick carbon steel inner shell and with water jackets and headers about 3/4 in thick. The lining and diaphragms at the basemat are 8 ft in diameter and 1-1/2 in thick. All welds are inspected. The cost of this shop' fabricated assembly is estimated on a weight basis at $4.00/lb, which is equivalent to $5.85/lb when overhead is added. This is a conservative figure even for nuclear plants. l A-11 l l l

[ I, i. TABLE A-6 CENTRAL BASEMAT CORE RETENTION SYSTEM (ln 5/1000) Plant Status A B C Engineering ana design, 100 175 250 Plant cool-ocen -- - 100

 'Crain pool, Install shieles                                                                             -      -        50 Enter containment                                                                                       -      -     100                    -

Bore access holes, instell crane for basemat aree 20 100 150 . Enter besemat area - 25 100 Olomond crill to holes in pecestal wall 20 100 150 Manway to center pecestal - 50 250 Lower in steel walls 50 150 250 water walls Installed 50 150 200 Plpe to exterior cold water or equivalent 100 200 250 Shelter ring Installation 10 50 80 install pebbles and thorla 10 40 70 Clean up basemat level 10 80 90 Leave bottom-block and seal holes - 50 50 Remove shields - - 30 Install cones and barriers on diaphragm 50 80 18 0 Clean up, leave drywell - 30 50 l Refill suppression pool - - 10 SUBTOTAL 420 1280 2190 Field overhood (165) 68 205 350 Supervision energeable 108 162 216 Contingencies (255) 149 412 690 j TOTAL 5744 52058 53445 9 l l 5 A-12 i

                                                                                                                                                       )

I  !

p. ,

               .g       ..

j;' Weight estimate

                                                                               '1b e   Main cone                               _40,000 l

o Water. jackets .

                                                                            ~33,000 e  Top cone through basemat                 20,000-e   Diaphragms (3)      .   .

7,500 e Bottom section and pipes 3;000 e Supports and shock absorbers 15,000 Total 118,500~ The cost of each. dry cone-shaped crucible is then $474-000: , the cost of installing it in a plant is'given in Table A-7. , Two of these dry cone-shaped crucible core-debris retention units are then included inithe plant cost estimate shown in-Table A-8. If a single running; reactor per Table A-8 is retrofitted with , a tunnel-accessed dry-crucible-type core debris retainer, the . increased cost over Status C is approximately_$5,000,000. The basic costs of'the underground installation are taken from 1 Section 7.7 and Appendix B-11 of Hammond's report (1982). The ') items subtracted from these original costs either are not applicable in this installation or are included elsewhere. Plant Status A does not need the tunnel because the installa-tion is made prior to pouring the basemat. One of the advan-tages of this underground entry is that very little plant downtime is required to make the installation. Probably nor-mal refueling time'is adequate to plug the holes'in the center wet well since bottom entry is used. A.4 VENTING AND FILTERING SYSTEM Section 3.4.4 described the three overpressure control sys-tems in detail. Their costs are estimated in the following breakdowns. A.4.1 Containment Pressure Relief Va3ve Svstem This system has a vent line to the plant stack and includes dual redundant large valves (-24 in, throat diameter) that are reclosable when the containment pressure is reduced by the venting action. They are set to actuate and open slightly above the containment design pressure. Each valve should be capable of handling the calculated ATWS clean steam flow, although both may open partially. In many plants an existing penetration may be used for this flow, but the cost estimate includes a new one. The cost estimate  ; is given in Table A-9. A-13

o i t, . ;- e l ' p l-4-

                                                   ' TABLE'A-7. DRY CRUCIBLE twSTALLATION COSTS (in $/1000)

(

                                                ' Plant Status                         : A-          B        C
                                      , installation engineering                       100 ^     included in tunnel Core-catcher crucible                           474         474      474                     g-
                                   ' /Components in-place -                              50           60      80 -                  I Pressure test                                     20           40      50
                                           'TDTAL'                                    $644        $574     $604 Note: Overhead factors are adood later, s

TABLE A-8. . DRY CRUCIBLE INSTALLATION COSTS (In $/1000) Plant Status A B C Engineering and design -200 200 250 Crucible' installation 1010 -- -- Access tunnel (1) and two celssons (f rom NUREG/CR 2941) -- 19000 19000

                        -     less abeyance time                                                   --       (B00)     (800)-

less redletion shleid -- (500) (500) (nflation adjustment (1982 to 1984 at 65/yeer) -- 2124 2124 l Two complete dry crucibles from Table A-7 1288 1148 1208 I Piping to Internal sprey system 200 50 100 S UBTOTAL 2698 21222 21382  ! Field overhead (165) 432 3395 3421 -) Additional supervision 540 270 270 Contingencies (255) 920 6221 6268 TOTAL (two plants) $ 4590 $31108 $31341 l Cost per reactor 5 2295 $15554 $15670 A-14 a .

       . = , _ _ - _ .

q-y -. -

 <-yQ&[

y,

                                                                             ' TABLE A-9. . CLEAN STEAst VENTING TO STACK (In 5/*000) m

'*[m_ Y Plant Status, .A B C- .

                                                    ' System engineering:                                        . 25    l1'00   125'

Velve design, develop, test and certif y . 150 150 1150 200' 200 200 L' . Rollet valves' (2 et 17000 -lb et $5.8?/lb)

Large pipe' to stock (300 ft et $200/ f t) 60 60 60 - 200 200 200 I Penetration and sleeve ~ 100 300 400 , installation :1

                                                          $nsTOTAL-
                                                                                                                -735,    1010'  1115          .

118 162 178 F Field overhead (165) Supervisica (chergeable) 54 ' 54 .90 Contingencies (255) 226 307 3d5 TOTAL $t279 51533 51729 1 1 Estimetoa says et plant to retrofit ~50 i A.4.2. Low Flow Relief Valves to Condenser / Filter i

                                                'When.the ATHS-3-A modification'is made, the risk of an ATWS                                    !

event falls, so.the need for a high-flow' clean steam vent.is  ! reduced. Hence.the cost incurred may not be justified with -l this. reduced risk. This alternative venting system' handles l only the lower flow-rate containment overpressure relief l conditions and passes these gases through both a gravel. bed condenser.and a filter before they are released to the stack  ! because they.probably will be contaminated. Sufficient l gravel' bed is included to ecndense the one-time steam spike  ! from the sensible heat in the core debris mass when it i contacts' water, but there is insufficient condensing capacity to handle the residual heat release. The cost estimate is given in Table A-10. To save cost rupture discs may be used in lieu of these l relief valves, since they need not be reclosable--although it j 1-is preferred. j I A-15 L L. x-

n j < f , TABLE A-10 VENTING AND FILTERI% SYSTEM (In S/1000) Plant Status A B C Engineering and design 100- 100 125 valvo design, covelop, test certify 150 150 150 Relief valves (2 at 6000 lb at 15.85/10) 100 100 100 Piping 75 75 75 Penetrations and sleeves 100 100 100 . Filter duct (1000 ft long, 6-ft diameter) 200 .200 200

                                                                             ~

3 ravel (750 yd3 at 125/yd3) 20 20 20 Otten for filter duct 100 100 100 Condensate tank and punt, 125 125 125 Installation 300 500 570 SLsTOTAL 1210 1560 1705 Fleid overneed (165) .194 250 275 Supervision (chargeselo) 175 200 250 Contingencies (255) 395 500 557

                ' TOTAL                                          11975     12510   12785          I I

Estleeted days at plant to retrofit ~40 1 i A.4.3 Complete Larce Clean Vent and Low Flow Condenser / Filter System ') I Figure 3-17 is a schematic of the system designed to handle either the heavy steam flow during an ATWS event, the hydre-gen overpressure when metal around the hot core is being oxidized, or the short-term heavy steam generation when the large, hot, core-debris mass contacts water (the steam spike). The valves and pipe sizes are based on the most . l conservative flow circumstances. The case study plant cur-rently has a number of valves that may be used for this venting (nine are listed), but none appear to be automatic , reclosable relief valves. It is quite possible that an existing penetration can be used for venting without en-creaching upon its present function, but only a detailed  ; study of the plant drawings wi)1 determine if this is feasi- j ble. This cost breakdown is gisen in Table A-11. l

                                                                                                  . 1 A-16

 ).      . t ..

• L C:

l .. - TABLE A-11 COMBINATION VENTING SYSTEM (in $/1000) A B C Pfent Status Systems engineering and desipi 12' 150. 200 300 300 300 Velve design, develop, and tast (two types ) - . 250 250 250 Velve cost (2) 100 100 100 P1 ping 100 200 250 Penetrations and sleeves 100 100 100 Diverter volve (1) 40 50 50 Ouct (volve to diverter) 60 60 60 Larger pipe to stock 40 40 40 Sneller pipe to filter Filter duct (1000 f t, 6-f t diameter steel) 200 200 200 Gravel (750 yd3 et $25/yd3) 20 20 20 Oltch for filter pipe 100 100 100 Condensate tank and pump 125 125 125 500 800 900 Installation 2010 2475 2595 SUBTOTAL 320 396 432 Field overhoed (165) 216 216 266 Supervision tenergeable ) 636 772 84 0 Contingency (255)

                                                                                                            $3123   53859  $4242 TOTAL Estimated days et plant to estrofit         - 42 A-17

                           - _ 7 .- .
                                                                                                                              . . 4

j. ,h -

yp

    .W e
  . 4, i

I:

                                                                                                                                     l
                                 -A.5   HYDROGEN CONTROL While the 410,000 ft3 of free volume in the Mark                              II contain-            '

ment at Limerick is always inerted with nitrogen during .! normal operation, it is possible to allow' air to enter when underpressure conditions exist after an ATWS.and a core-melt accident at the very time hydrogen' gas may..h' ave been gener-ated. For this reason as well as for hydrogen. removal under ]l o any other service-type circumstance, a recombiner is recom- . mended as one of the mitigation. devices, Actually, recom- l, biners are now insta11ed'at Limerick, but they may not be at other Mark II plants. Hence th? Table A-12 cost estimate is '

                                 ' included.

1

                                                                                                                                           .1 TABLE A-12 LARGE HYOROGEN RECOMBINER                                  !

(In 5/1000) l Plant Status A B C

                       .e                                                                                                                  j Engineering and cosign of system                             50      150     150 Two complete hydrogen recomeiners (70 etm ea)

IMexi, 19831- 14C0 1400 1400 Access notes, valves, piping and electrical 75 125 175 Dedicated power supply" (110 kW di ese l') 50 50 50 l Installation (10 nen w nths for Status C) 250 300 350 SUBTOTAL Fleid overhead (165) 292 324 372 Chargeable supervision 108 162 in i Contingency (255) 556 627 714 i TOTAL $2780 $3138 53573 l i t,

• Limerick does not have a dedicated power supply for its hydrocen recombiners. , i j i

f i, 1 A-18 _ _ _ _ _ _ _ . _ 0

    ;  .          NP '

A.6 UNDER PRESSURE CONTROL SYSTEM Like hydrogen control,.most' plants of the Mark II type now incorporate some form of vacuum-breaking valve to preclude the possibilities of a damaging internal underpressure. Such a unit must be distinguished from the existing breaker valve-that is located between the wet and dry wells. Unfor-tunately, it is extremely difficult to establish,the needed capacity of this valve to' determine its suitability for handling in-flow air after a long-duration ATWS when size-able but indefinite amounts of inerting nitrogen are washed ' out of the containment along with the steam. While this is a-very necessary component for maintaining containment integ-rity after a-coru melt accident, it is relatively inexpen-sive and uncomplicated. Costs are estimated in Table'A-13. TABLE'A-13 LARGE VACUUM BREAKER VALVE (In 1/1000) ' Plant $tstus' A B C Lagineering end,. design 50 75 100 Valve design, develop, test, certify 200 200 200 Two volves (6000 lb each at 15.85/lb) 75 75 75 Penetrations and sleeves 75 150 200 Piping and Wackets 50 100 100 installation of sculpment 100 200 200 SUBTOTAL 550 700 875 Field overhead (165) 88 112 149 Chargeable supervision 54 54 54 Contingency (255) 173 21E 267 l TOTAL 5865 $1091 1133A A-19

4 4 i A.7 LARGE CHILLED FILTER SYSTEM 4 The large chilled rock and cold activated-charcoal filter system was described in Section 3.5.1. While the exact specifications of this large special-purpost filter system to ensure effective filtering of an unpressurized contain-ment are not yet well established, a rough cost estimete is made to determine if further analysis is justified. The cost estimate covers a round silo-like reinforced con-crete tower adjacent to the containment buildings. Only one filter is needed in a dual reactor installation. This silo will incorporate two distinct types of filtering media. The lower section will use basalt rock pebbles. It will be about 80-ft deep, hold about 1000 tons of rock and have about 180 ft of upward flow area through the rubble. There may be better filter materials than basalt rock, but they have not been examined yet. The lower section is topped with 20 ft depth of charcoal in the upper, smaller diameter section. This 40 tons of charcoal retaine the noble gases by adsorption. A weather shelter and swinging door valves are located at the top to allow flow in or out. The concrete wall averages about 2 ft thick and the cold sections have about 2.5-ft of low-temperature insulation and a protective metal outer case. Chilling to about -30 F will be done with two York or equiva lent cascade-type Freon refrigeration plants using water-cooled condensing and electric motor drivers. Inert gas (probably nitrogen) will be circulated through the filter bed with fans. Either of the dual units of 10 tons (120,000 Btu /h) capacity should handle the thermal losses of the system, but both units will be needed for " pull down." Each compressor draws about 66 bhp so each plant will need 85 kW electric power when auxiliaries are included. A separate building houses these two plants. Costs are estimated in Table A-14. A.7.1 Cost Estimat3nc Variations There has been considerable comment that the cost estimates presented here in Appendix A are unrealistically low. Even after rechecks we believe they are correct within the esti-mating framework proposed by NRC. In order to make a better comparison with other cost estimates the Large Chilled Fil- i ter System estimate of Table A-14 was taken as baseline with property costs, quality assurance and quality control costs l , i A-20

                        -_       _ - _ , - - _ - _ -      . - , . _ _             -- .-- . .-      -_          ._         . - - _            -=
   ,.y         :           t> ..

p TABLE A-14. LARGE CHILLED FILTER SYSTD4 (ln.$/1000)

                                                                      ~ Plant Statas                                              A, B, or C Design and engineet'ing'                                                         200 Development.and testing filter system                                            150-Concrete (900 yd3 at $600/yd3 ) '-                                               540 Insulation and cover (10,000 ft2 at 110/ft 23                                    joo Filter bed rock-sized and cleaned (500 yd3 at 550/yd3)                             40 100 Oiercoal-Installed 40 tons at $1.75/1b.

Piping-10 f t inlet duct and-refrigeration lines 90 Door type rollet valves with shelter (6) 160 Refrigeration plant--dual units , 3-ton capacity each with blowers and Hexes-remove 40 kW oech '(Streif,1964) 170 Building for refrigeration plant- (1000 f t2 at 5200/ft2) 200 Lhohrega seat at inlet and exit (2) 80 SunTUiA6 1850 Field overhead (165) 300 Super <l s Ion 200 Contingencies (255) 588 T07nt $2938 e A-21 i

1_ i. l l added. This is shewn in Table A-15. In this instance the costs increase by a factor of 1.75 when there extras are included. Even using the mest conservative estimates the i increase is about 2.43 times baseline. This cost accounting results in values close to that of other estimates. TABLE A-15. LARGE CHILLED FILTER SYSTEM (IN DOLLARS /THOUSAND) , 1 1 J l PROPERTY 43ST I PER + QA/0C CONSERVATIVE REPORT + CHARCOAL 8412 FROPERTY 9000 FT2 0 $50/FT2 - 450 450 , 1 4 DESIGN AND ENG4NEERING 200 200 400 DEVELOPMENT Arc TESTING FILTER SYSTEM 150 150 150 CONCRETE (900 YD3 AT 1600/YD3 540 540 600 l INSULATION AND COVER (10,000 FT2 AT 110/FT2) 100 100 200 l

                                                                                                               )

FILTER BED ROCK - SIZED AND CLEANED ) (500 YD3 AT 550/YD3 ) 40 40 50 l CHARCOAL - PREPARED (100 tD3 AT $250/YD3 ) 100 100 200 P APING - 10FT INLET "JJCT AND REFRIGERATION LjNES* 90 90 180 000R TYPE RELIEF WALVES wiTH SHELTER

• 180 180 200 REFRIGERATION PLANT DUAL UN4TS, 3 TON l CAPAC&TY EACH wnTH BLOWERS AND HEXES -  !

25 KW E/CH (STREIF,1984)* 170 170 350 1 BulLDING FOR REFRIGERATION PLANT - 1000 l (1000 FT2 AT $200/FT2) 200 200 200 l 04APHRAGM SEAL AT INLET AND ExtT* EO 80 100 l l j QA + QC AT 100$ ON STARRED ITEMS 1060 1430 l SUBTOTAL 1850 3360 4660 F4 ELD OVERHEAD ('65) 300 538 746

                  $UPERVISION (4 MEN, 8 MO)                   200            200            300 CONTINGENCY [S (255)                        588           1025           1426 TOTAL                                                            $5122         17133 12938        ( x 1.75 )    ( x 2.43)

A-22

   '. $ . 6-l REFERENCES Bellamy, R.R., 1974, " Elemental Iodine and Methyl Iodide Adsorption en Activated Charcoal at Low Concentrations,"

l Nuclear Safety, Vol. 15, No. 6, p. 711. Berman, M., 1981, Workshop on the Imoact of Hydrogen on Water Reactor Safety, NUREG/CR-2017 SAND 81-0661. - BNL, 1984, A Review of Bvm 6 Standard _ Plant Pr obabal is ti c Risk Assessment: . Containment Failure Modes __and_ Source Term 2 Resulting from Internal and__ Ex te rnal Events t _Vol. 3 LDraftl2 NUREG/CR-4135, Brookhaven National Laboratory. Bowman, F.L., 1973, Reactor Core Meltdown Containment for Offshore Application, Doctor &1 thesis, MIT. Camp, A.L., et al., 1983, Light Water Reactor Hydrogen Manual, NUREG/CR-2726 Sand 82-1137, Sandia National Laboratories. Castle, J.N., Catton, I., Dooley, J.L., Hammond, R.P., Kastenberg, W.E., and Swanson, D., 1984, Survey of Licht Water Reactor Containment Systems, Dominant Failure Modes, and Mitication Opportunities, NUREG/CR-4242, R&D Associates. Castle, J.N., Catton, I., Dooley, J. L., Hammond, R. P., Kastenberg, W. E., Swanson, D., 1984b, Survey of the State of the Art in Mitication Systems, NUREG/CR-3908, R&D Associates. Caterpillar Tractor Company Quotation to Dooley--R&D Associates, 1983. - Cybulskis, P., Wooton, R.O., Denning, R.S., 1982, Effect of , Containment Ventina on the Risk from LWR Meltdown Accident,, NUREG/CR-0138 BML-2002, Battelle Columbus Laboratories. Erdmann, R., et al., 1983, " Mitigating Severe Accident Consequences at the Shoreham BWR," Proceedings of the Inter-national Meetino on Licht Water Reactor Severe Accident Evaluation, Vol. 1, Sesbion 8. Fauske and Associates, 1981, "Phenomenological Assessment of Hypothetical Severe Accident Conditions of the Limerick Generating Staticn," R.E. Henry. (From Limerick PRA.) , R-1

                                                                                  *   . .iw
                                                                                            't 7

f f

        FILTRA" Report 1982, Filtered-Atmospheric Ventino of LWR Containments, AB ASCA-ATOM & Studsvik Energiteknik.AB, Swedish Nuclear Power Inspectorate.

Follis, R.M. 1984, Westate Carbon Co., Los Angeles, C A ,. private correspondence with J. Dooley and S.-Ridgway. Gazzillo, F., snd Kastenberg, W.E., 1984, " Risk Reduction by

                                                  ~

l' Filtered Venting in PWR Large Dry Containments," Nuclear. l Encineerino and Desien, g 77, pp 69-85. , GE, 1983, General Electr_ic Company S tandard Sa f e ty A na lysi s Reoort ( GESS AR II) , Docket No. STN 50-447. GE, 1985, " Description of the Ultimate Plant Protection System (UPPS)," General Electric Company. (Proprietary document.) Hammond, R.P., Dooley, J.L., 1982, Retrofittina Core Catchers to Nuclear Plants, NUREG/CR-2941 RDA-TR-119000-OO1, R&D Associates.

      ' Harper, F.T.,,and Benjamin, A.B.,           1982, Risk Assessment of Filtered Vented Containment Ootions for a-BWR Mark III Con tainment. Sandia National Laboratories, SAND-82-0403C.

Eatch, Steven W., 1981, Reactor Safety Study Methodology _

      . AJp_11 ca ti on = P ro J r_am : Grand Gul f #1, BWR Po_wer PI_an1 NUREG/CR-1659, Sandia National Laboratory, October.
       'Heaberlin, S., et al.,         1983, A Handbook for Value Impact Assessment, NUREG/CR 3568, Pacific Northwest Laboratory.

Henrie, J.O., 1972, Thermal Recombiner Demo Test, Atomics International Report AI-72-61. j IDCOR, 1983, Core-Concrete Interactions, Atomic Industrial Forum, Technical Report 15.3. IDCOR, 1983, Hydrocen Generation During Severe Core Damace Secuences, Fauske Associates and Argonne National Laboratories, Technical Report 12.1. J l Ingersoll-Rand Company, 1983, Phone conversations with - Dooley, RDA, re centrifugal pump. l Keilholtz, G.W., 1976, Hydrocen Considerations in_ Licht Water Reactors, ORNL-NSIC-120. 1 R-2 , I;

ev; t

   )t:

i.. Kovach, J.L.,-1982,'"The Evolution and Current State of'

                  'Radiciodine Control," Nu'elear Safety, Vol. 23, No. 1, p. 44.

Levy, S., 1981, Review of P;" posed Improvements Including Filter / Vent of BWR, etc., hPr!I-NP-1747, TPS-80-721.

                  'Long Island Lighting Company, 1982, Sho re ham Nu cl ea r Powe r S ta ti on , Unit I Probabalistic Risk Assessment . (Prepared by

• SAI, Inc.)

Maki, J., 1983, Atomics International Company. Personal , discussion with Dooley-RDA on hydrogen recombiners. Moeller, D.W. and Underhill, D.W., 1981, " Review and Evaluation of Factors Affecting Noble Gas Adsorption on Activated Carbon," Nuclear Safety, Vol. 22, No. 5. McCormack, J.D., et al, 1982, Testing of a Passive Submerced Grave 1' Scrubber for Containment Ventinc Applications, HEDL-SA-2637-FP, DOE Contract DE-AC06-767F02170. NRC, 1975, Reactor Safetv Study--An Assessment of Accident Risks in U.S. Commercial Nuclear Power Plants, WASH-1400, NUREG/75-014. NRC,-1983, Regulatory Analysis Guidelines of the U.S. Nuclear Reculatory Commission, NUREG/BR-0058, Office of the Executive Director for Operations, Washington, D.C. NRC, 1983, ." Policy Statement on Safety Goals for the Opera-tion of Nuclear Power Plants," Federal Recister 10772-10781, March. NRC, 1984, Safety Evaluation Report , Related to Final Desion Aggroval of GESS AR BWR 6 Nuclear Is_1and Design 2 NUREG-0979, Supplement 2, November; Supplement 4, July 1985. NRC, 1985, Sa f e ty Evaluation Report, Related to Final Design Approval of GESS AR BWR 6 Nuclear Island Design t NUREG-0979,

       .           Supplement 3.

Papzoglou, I., et al., 1982, A Review of the Limerick Generatinc Station PRA, Brookhaven National Laboratory, NUREG/CR-3028. Pence, D.T., 1981, " Critical Review of Noble Gas Recovery and Treatment Systems," Nuclear Safety, Vol 22, No. 6. R-3

w s.n_ t ,>c,  ;  ; ... TR - a

            ?

t

       ' %i f

Philadelphia Electric Company-, 1981, Limerick ' Genera ting-S tation PRA2. Volumes 1.and 2, 50-532/3.- , Silberberg,=M., 1983, "Phenomenological Standard ~ Problem for BWR's," NRC letter to Accident Source Term. Study Group,-RES Program Office. Sperry, R.E., 1982, Detailed Cost Estimates for Tunnel Access to Reactors, Subcontract report to~RDA. , Streiff, 1984, Private' discussion between Tom Streiff-york Manufacturing Co. and J. Dooley, re frigeration. Strip..D., 1982, Estimates of the Financial Risks of Nuclear Power Reactor Accidents, NUREG/CR 2723, Sandia-National Laboratories, Albuquerque, NM. l-Thomas, J.W.,.1974,:" Noble Gas Adsorption Process," U.S. Patent Disclosure S.N. 4 4 3 , 2 6 2 ~, . filed February 19, 1974. Westinghouse, 1983, Preliminary PRA and' technical data on WSP90 advanced reactor. Wassel, A. T.,-Hoseyni, M.S., Farr, J. L., Jr., 1984, Scrub-

bing of. Radionuclides Aerosols in Water Pools' ANS Topical ,

Meeting on Fission Product Behavior and Source Term Research,

                                               . Snowbird, Utah.

young Radiator Company, 1983, Letter quotation.from Coulthurst, N.D., Racine, WI. l ( i R-4 _ _ _ _ _ _ _ _ _ _ _ . _ . . _}}