ML20058B503

Safety Evaluation Report Related to Hydrogen Control Owners Group Assessment of Mark III Containment

ML20058B503
Person / Time
Issue date:	10/31/1990
From:	Kudrick J, Chang Li Office of Nuclear Reactor Regulation
To:
References
NUREG-1417, NUDOCS 9010300241
Download: ML20058B503 (44)

Text

__.

l'

~

NUREG-1417

~

Safety Evaluation Report related to Hydrogen Control Owners Group Assessment of Mark III Containments e

L U.S. Nuclear Regulatory Commission f

OfTice of Nuclear Reactor Regulation October 1990 gY 4>

l 9010300241

$$$7N""""901031 poa

l l

c-i AVAILABILITY NOTICE Availability of Reference Materials Cited in NRC Publications

_i Most documents cited in NRC publications will be available from one of the following sources:

1.

The NRC Public Document Room, 2120 L Street, NW, Lower Level, Washington, DC

.20555 2.

The Superintendent of Documents, U.S. Government Printing Office, P.O. Box 37082, Washington, DC 20013 7082 3.

The National Technical information Service, Springfield, VA 22161 i

Although the listing that follows represents the majority of documents cited in NRC pub;ica-tions, it is not intended to be exhaustive.

Referenced documents available for inspection and copying for a fee from the NRC Public Document Room include NRC correspondence and internal NRC memorands; NRC Office of

- Inspection and Enforcement bulletins, circulars, information notices, inspection and investi-gation notices; Licensee Event Reports; vendor reports and correspondence; Commission y

papers; and applicant and licensee documents and correspondence.

1 The following documents in the NUREG series are available for purchasc from the GPO Sales Program: formal NRC staff and contractor reports, NRC-sponsored conference proceed-Ings, and NRC booklets and brochures. Also available are Regulatory Guides, NRC regula-tions in the Code of Federal Regulations, and Nuclear Regulatory Commission issuances.

Documents available from the National Technical information Service include NUREG series reports and technical reports prepared by other federal agencies and reports prepared by the Atomic Energy Commission, forerunner agency to the Nuclear Regulatory Commission.

j Documents evallable from public and special technical libraries include all open literature

' items, such as books, Journal and periodical articles, and transactions _ Federal Register notices, federal and state legislation, and congressional reports can usually be obtained from these libraries.

Documents such as theses, dissertations, foreign reports and translations, and non-NRC conference proceedings are available for purchase from the organization sponsoring the publication cited.

Single copies of NRC draft reports are available free, to the extent of supply, upon written request to the Office of Information Resources Management, Distribution Section, U.S.

T Nuclear Regulatory Commission, Washington, DC 20555.

Copies of industry codes and standards used in a substantive manner in the NRC regulatory process are maintained at the NRC Library, 7920 Norfolk Avenue, Bethesda, Maryland, and.

are available there for reference use by the public. Codes and standards are usually copy-righted and may be purchased from the originating organization or, if they are American National Standards, from the American National Standards institute,1430 Broadway.

- New York. NY 10018.

1 L

a

NUREG-1417 Safety Evaluation Report related to Hydrogen Control Owners Group Assessment of Mark III Containments U.S. Nuclear Regulatory Commission Office of Nuclear Reactor Regulation October 1990

(" "*%,,

g

/

ABSTRACT Title 10 of the CodeofFederalRegulations (10 CFR),

provides its evaluation of the generic methodology pro-Section 50,44," Standards for Combustible Gas Control posed by the Hydrogen Control Owners Group. *lhis ge-System in IJght Water-Cooled Power Reactors," re-neric methodology is documented in Topical Report quires t' at systems be provided to control hydrogen con.

IIGN 112 NP," Generic Hydrogen ControlInformation centration in the containment atmosphere following an for BWR/6 Mark Ill Containments " In addition, the staff accident to ensure that containment integrity is main-has recommended that the vulnerability to interruption tained.The purpose of this report is to provide regulatory of power to the hydrogen igniters be evaluated further on guidance to licensees with Mark III containments with a plant specific basis as part of the individual plant exami-regard to demonstrating compliance with 10 CFR 50.44, nation of the plants with Mark Ill containments.

Sections (c)(3)(vi) and (c)(3)(vii). In this report, the staff i

i I

i r

1 i

t iii NUREG -1417

I I

l l

4 i

CONTENTS i

Page

)

ABSTRACT................................................................................

m 1 INTR O D UCrlO N AND B ACKG RO UND..................................................

1 2 GENERAL DESCRIFFION OF THE HYDROGEN IGNITION SYSTEM.......................

2 3 COu n uSTION tiG NITER TESTINo.......................................................

4 3.1 S mall. Scal e Tests...................................................................

4 3.2 Q uart er. Scale Test Facility...........................................................

4 l

3.2.1 Scalin g M e t hodology........................................................

6 3.2.2 Quarter-Scale Testing Approach..............................................

6 3.2.3 Quart er. Scale Test R esults...................................................

7 4 CONTAINMENT STRUCTURAL CAPACfrY..............................................

9 5 DEGRADED CORE EVENTS AND HYDROGEN GENER ATION............................

10 5.1 i n t rod u c t io n.......................................................................

10 5.2 Evaluation.........................................................................

11 5.2.1 Acceptable Hydrogen. Generation. Event Sequence...............................

11 5.2.1.1 The HCOG's Base. Case Scenario...............,....................

12 5.2.1.2 Station Blackout and NUREG-1150..................................

12 5.2.1.3 THU the " Acceptable Sequence".....................................

I'!

5.2.1.4 Hydrogen G eneration Profiles.......................................

13 5.2.1.5 Nonmechanistic Hydrogen Release Profile...... :......................

13 5.2.2 B WR Co re H ea t u p Cod e.....................................................

15 5.2.2.1 I n t rod uct ion......................................................

15 5.2.2.2 Phenort enological Assumptions......................................

15 5.2.2.3 S t eam G e n era t io n.................................................

15 5.2.2.4 Hydrogen Generation..............................................

16 5.3 - S u m ma ry and Concl usions............................................................

17 6 CONTAINMEKr RESPONSE-AN ALYTIC AL MODELING.................................

17 6.1 Localiz ed Com bu stion...............................................................

18 6.2 Containment Pressure and Temperature Calculations....................................

19 6.3 D rywe ll Analysis....................................................................

20:

6.4 Existence of Drywell Diffusion Flames.................................................

20 v

NUREG-1417 m

i l

CONTENTS (continued) f Page 7 SURVIVABILITY OF ESSENTI AL EQ UIPMENT..........................................,

22 7.1 Identification of Essential Equipment..................................................

22 7.2 Generic Equipment Survivability Analysis (localized Combustion)..........................

23 7.3 Diffusion Flame Thermal Environment Methodology.....................................

24 7.4 S pray Avail abili ty...................................................................

24 7.5 Press u re Effect s....................................................................

25 7.6 D e t o n a t io ns........................................................................

25 8 C O N C LU S I O N.........................................................................

25 APPENDIX A GENERIC HYDROGEN IGNITION SYSTEM TECHNICAL SPECIFICATIONS.....

27 APPENDIX B MARK 111 COMBUSTIBLE GAS CONTROL EMERGENCY PROCEDURE G U I D E LI N E.................................................................

29 APP EN D IX C B I B LI O G R AP H Y.............................................................

34-AP PEN D IX D ACR O N Y M LI ST.............................................................

37 FIGURES Figure 2.1 G en eral hyd rogen igniter assembly......................................................

3 Figure 3.1 Elevation views of the quarter. scale test facility............................................

5 Figure 3.2 Facility configuration tests S.01-S.11.....................................................

5 Figure 4.1 Typical Mark 111 Containment Configuration.............................................. 10 Figure 5.1 Hydrogen generation rate (150 gpm reflood)............................................... 14

. Figure 5.2 Hydrogen generation rate (5000 gpm reflood).............................................. 14 Figure 6.i T.Ty C LAS I X-3 Mod el................................................................ 20 Figure 6.2 SORV with no spray-wetwell temperature.............................................. 21 Figure 6.3 SORV with no spray-wetwell pressure................................................... 21' Figure B.1 Operator actions for primary containment hydrogen control.................................. 31 TABLE Table 4.1 Comparison of BWR Mark 111 containment characteristics...................................

9 NUREG-1417 vi

1 INTRODUCTION AND is greater. Typically, this would translate to a 1 to 5 per-j BACKGROUND cent metal wa er reacti n f the active cladding.

However, the accident at Three Mile Island Unit 2 on Followingaloss-of coolantaccidentinalight waterreac-March 28,1979, resulted in a metal water reaction that tor plant, combustible gases, principally hydrogen, may involved approximately 45 percent of the fuel cladding accumulate inside the primary reactor containment as a (i.e., about 990 lbs), which resulted in hydrogen genera-result of (1) metal-water reaction involving the fuel cle-tion well in excess of the amounts specified in 10 CFR ment cladding;(2) radiolytic decomposition of the water 50.44. The combustion of this hydrogen produced a sig.

in both the reactor core and the containment sump; nificant pressure spike inside containment. As a result,it (3) corrosion of certain materials within the containment became apparent that additional design measures were by sprays; and (4) any synergistic chemical, thermal, and needed to handle larger hydrogen releases, particularly i

radiolytic effects of post accident environmental condi.

for smaller volume containments and those with lower tions on protective coatings and electric cable insulation.

design pressures. The Nuclear Regulatory Commission (NRC) determined that a rulemaking proceeding should be undertaken to define the manner and extent to which To provide protection against this possib!c hydrogen ac.

cumulation resulting from an accident, Title 10 of the hydrogen evolution and other effects of a degraded core l

Code of Federal Rgrdations (10 CFR), Section 50.44, need to be considered m plant design. An advance notice g

" Standards for Combustible Gas Control System in Light-of the rulemaking proceeding on degraded core issues was l

Water Cooled Power Reactors," and General Design published in the Fcdcral Rgister en October 2,1980 (45 1

Criterion (GDC) 41,

" Containment Atmosphere I~R 65466). In addition, a new unresolved safety issue was Cleanup," Appendix A to 10 CFR Part 50, require that mstituted (A -48, ' Hydrogen Control Meas ires and Ef-systems be provided to control hydrogen concentrations fectsof Hydroner. rns on Safety Equipment ) to evalu-ate this new

.rn.

in the containment atmosphere following an accident to ensure that containment integrity is maintained. Conven-tional hydrogen control systems (e.g., hydrogen recomb-To formalu.- its requirements for additional hydrogen mers) historically have been mstalled to provide the capa-control measures to deal with degraded core accidents af-bility to control the relatively low rate of hydrogen fccting pressurized water reactors with ice condenser accumulation (or oxygen accumulation m merted contam-containments and boiling water reactors (BWRs) with ments) resulting from radiolytic decomposition of water, hiark III containments, the NRC published an amend-corrosion of metals mside containment, and environ-ment to the hydrogen rule (10 CFR 50.44) in the Fed-cralRegister on January 25,1985 (50 FR 3498). The mental effects on coatings and insulation. H,0 wever, the net free volume inside containment (or the inertness of amended rule required that a hydrogen control system be the condainment volume) controls the rapid hydrogen provided and that the system be capable of accommodat-production resulting from a metal water reaction of the ing, without loss of containment structural integrity, the fuel cladding. That is, the containment solume is large amount of hydrogen generated from a metal water reac-enough so that the hydrogen generated early would not tion of up to 75 percent of the active fuel cladding. In ad-reach the lower limit of flammability (or would not result dition, systems and components necessay to establish

+

m the formation of combustible mixtures).The rationale and maintain safe shutdown must be canable of perform-l for this approach is that the rate of hydrogen release re-ng their function regardless of hydrogen burning,.

1!

sulting from cladding reaction was assumed to be t Pursuant to the provisionsof the rule cach licensee with a l

rapid (on the order of minutes) following a postulated ac-cadent to allow for an acttve c Mark III containment has installed a hydrogen ignition gen control sys' ems (recomb, ontrol system. Thus, hydro-system and submitted a preliminary analysis and a sched-mers)would only be actuated ule for meeting the full requirements of the rule.The af.

l later m the event to control the slow hydrogen /or oxygen fccted plants with Mark III containments are Grand Gulf i

release associated with radiolysis and other reaction of materials mside contamment.

Nuclear Station, River Bend Station, Perry Nuclear l

Power Plant, and Clinton Power Station.The staff's in-terim evaluations of the licensecs' initial responses for To quantify the metal-water source for design-basis acci-their particular plant are documented in supplements to dents,10 CFR 50.44, codified in October 1978, requires the plants' safety evaluation report (NUREG-0989, i

that the combustible gas control system be capable of NUREG-0887, and NUREG-0853, respectively),

handling the hydrogen generated from five times the amount calculated in demonstrating compliance with In preparing their responses, the licensees were aided by 10 CFR 50.46, " Acceptance Criteria for Emergency Core the efforts of the Mark III Containment Hydrogen Con-

]

Cooling Systems for Light Water Nuclear Power Reac-trol Owners G. oup (HCOG). The utilities with Mark III tors," or the amount corresponding to the total reaction containments formed this group in May 1981 to collec-of the cladding to a depth of 0.23 mils, wNchever amount tively perform testing and analyses to demonstrate the i

NUREG-1417

l effectiveness and reliability of the hydrogen ignition that generated from a metal water reaction in-

[

systems.

volving 75% of the fuelcladding surrounding the l

active fuel region (excluding the cladding sur-in addition, each licensee with a hiark Ill containment rounding the plenum volumej.

is required to provide a final analysis [10 CFR 50.44(c)(3)(vii)(B)] to confirm the conclusions of the pre-The concept empkiyed by the licensees with a hiark 111 liminary analysis and/or,4 necessary, to institute modifi-containment, and similarly the ice condenser licensees, is cations to ensure compliance with the rule, ne scope of to intentionally ignite hydrogen generated inside contain-this analysis is specified in 10 CFR 50.44(c)(3)(vi)(B).The ment. This method precludes buildup of relatively high generic findings from the HCOG's program will be used concentrations of hydrogen during degraded-core acci-for (Tis final analysis, supplemented by plant specific de-dent scenarios.

sign considerations as addressed in each licensee's inde-pendent plant examination program.

,To accomplish this early igmtion, a hydrogen igmtion sys-tem (HIS)is installed in each of the four plants with hiark The following staff evaluation focuses on the assessment Ill containments. The HIS system consists of approxi-of the completed generic testing and analyses performed mately 100 agmter assemblics distributed throughout the by the HCOG in support of the plant-specific analysis.

The HCOG activitics were summarized in a topical report drywell and contamment regions.

transmitted by letter dated February 23,1987, correspon-dence identification iIGN-112-NP, " Generic Hydrogen The main element of the igniterassemblyis the hiodel7G Control Information for BWR-6 hiark III Contain-thermal igniter glow plug (commonly used in diesel en-ments." The topical report is a summary document of all gines) manufactured by the General hiotors AC Division.

of the individual generic submittals that have been sent to The same design of igniter assembly is used in each the staff by the HCOG. It should be noted that HCOG hiark III containment. Each igniter is powered directly correspondence identification designators with a "P" suf-from a 120/12-V step-down transformer and designed to fix (HON-XXX-P) are proprietary to HCOG. Whereas, provide a minimum surface temperature of 1700 *F.The those without a suffix or with an "NP" suffix are non-igniter assembly (see Figure 2.1) consists of a 1/8-inch-proprietary, thick stainless-steel box that contains the transformer and all electrical connections and is manufactured by the As part of the review of the HCOG program, the staff ob-Power Systems Division of hiorrison Knudsen. Igniter as-tained technical assistance from Sandia National Labora.

semblics are Class 1E, seismic Category I, and meet the tory (SNL), the principal contractor for the NRC research requirements of the Institute of Electrical and Electron-

- program on hydrogen control and combustion phenom.

ics Engineers Standard 323-1974 and NUREG-05SS for ena. SNL provided the NRC with an independent assess.

emironmental qualification.

ment of technical issues contained in selected HCOG submittals pertaining to hydrogen behavior.

The igniter assemblics are divided into two groups; each group being powered from a separate Class IE division The staff's evaluation of the generic considerations of the power supply The intent is to have at least two igniters i

hydrogen control system for the hiark III containment is located in each enclosed volume or area within the con-discussed in the remainder of this report.

tainment that could be subject to possible hydrogen pock-eting and to have each igniter powered from a separate power division. In open areas within the containment ano 2

GENERAL DESCRIPTION OF drywell regions, igniter assemblics at the same elevation are designed to have alternating power division sources.

THE IWDROGEN IGNITION Igniters are separated by about 30 feet when both engi-SYSTEM neered safety feature power divisions are operable or by about 60 feet when one power division is inoperable. Ig-The regulation,10 CFR 50.44(c)(3)(iv)(A), states:

niter placement is designed to be more widely spaced in the large open regions, such as above the refueling floor, and in the lower regions of the drywell that are subject to Each licensee with a boiling light-water nuclear power reactor with a hiark Ill type of contain-fl oding. Requirements of plaecment as well as other pa-rameters of the system are contamed in the techmcal ment..., shall provide its nuclear power reactor with a hydrogen control system justified by a suit-spectfications for the Hid, as proposed by the HCOG, and able program of experiment and analysis.The hy-addressed in Appendix A of this report.

drogen control system must be capable of han-dling without loss of containment structum!

Each igniter power division has a corresponding onsite integrity an amount of hydrogen equivalent to emergency diesel generator. Incorporation of the NUREG-1417 2

- ; _.4

[

i

=

,s

=-

o 6

a I..... 4.....r,,.......

.t.,-g,

=h.

%d p )p.g.. ww n

. s.

+;

L nn nn

' g 4

-=

a e

r v

i

! 4, i 4 i

......R..... !'.

l

• m

~;M O

O Le m

13,

3

i ij

l 4

x"

..... q l

st ;;

)

r4 s

j

,.Q

- =

~

unstwie [

l; w-mme c

L... ! t "d..

f y

Figure 2.1 General hydrogen igniter assembly emergency dicscis into the design addresses the question The HCOG provided its justification to support manual of igniter power for many sequences. but not for station actuation by letter dated March 5,1986 (HON-073). It blackout conditions. Under blackout conditions, the ll!S stated that actuation is linked to indication of the RPV would not be operable. On the basis of a separate evalu-water level, which is a key safety parameter and is closely ation of this possibilityin the context of the NRC corrtain-monitored by the operators. In addition, HIS actuation-ment performance improvement program, the staff has only requires the positioning of two handswitches. The recommended that the vulnerability of the hydrogen ig-HCOG further stated that operators should not hesitate niters to a pow:r interruption be evaluated further on a to energize the system during accident scenarios in which plant specific basis as part of the individual plant exami-the hydrogen threat is uncertain or marginal because nations of the Mark III containments (see. Generic 1.ctter there would be no adverse effect on the plant as a result of -

88-20, Supplement 3).

unnecessary igniter actuation. Time available to actuate 1

the HIS is the other significant parameter. On the basis of The HIS is designed for ma it al actuation from the main the hydrogen events considered, the HCOG estimated control room. Actuation by the operator is required by this time to be approximately no tess than 25 minutes; that plant emergency procedures when the reactor pressure is, after the water level reaches TAF to the lower hydro-vessel (RPV) water level reaches the top of active fuel gen flammability limit reached in the wetwell volume.

(l'AF).The proposed combustible gas control emergency The HCOG also noted that hydrogen would migrate to-procedures for plants with Mark 111 containments are fur-the upper portions of the containment before the wetwell ther discussed in Appendix 13 to this report, reaches hydrogen combustion conditions.This effect was 3

NUREG-1417 l

h seen in the quarter-scate tests.Therefore, the time inter-tainment walls. Ilydrogen was admitted through simu.

valis expected to be somewhat greraer than 25 minutes, lated quenchers and/or vents into the suppression pool j

and ignited by prototypical ignition sources.

The staff finds manual operator actuation acceptable be-cause the HCOG has shown that (1) manual actuation is a lhe most important result obtained from the 1/20 scale simpb task,(2) the operator has sufficient time to per-test was the confirmation of continuous hydrogen burning form the task,(3) and there are no negative effects if the in the form of steady diffusion flames above the suppres-system is inadvertently or unnecessarily actuated, sion pool.1hc significance of this mode of hydrogen burn-ing is the observed severe thermal loads that occur near "the staff finds hydrogen igniter systems currently in-the diffusion flames that could threaten the integrity of j

stalled in the plants with a Mark 111 containment to be ac-the containment and equipment. Diffusion flames were ceptable with the caveat that the vulnerability of the :.y-observed when hydrogen flow rates of 0.4 lb/sec (full-drogen igniters to power interruption be evaluated scale equivalent) or greater were used. Combustion was further on a plant specific basis as part of independent initiated by the igniters and rapidly propagated to the pool plant examinations of the Mark 111 containments (see surface and formed steady diffusion flames that were an-Supplement 3 to Generic Letter 88-20).

chored at the surface of the pool and located alx)ve the submerged spargers that released the hydrogen. Ilydro-gen burning was observed to intensify as the hydrogen in-3 COMBUSTION / IGNITER jection flow rate was increased. as evidenced by ta'lcr TESTING flames and higher temperatures.

As part of the 1/20-scale program to determine the sensi-Numerous research programs have been conducted since tivity of scaling, a 1/5 scale single-sparger mockup was 1980 to better understand bydrogen combustion behavior constructed. Asignificant retalt of these tests was that ap-and the performance of ignition devices. Sandia National proximately oac-half the Game height that would have j

laboratory summarized the findings of recent hydrogen been predicted from the 1/20-scale tests was observed.

combustion test programs in NUREG/CR-5079Jlhis re-On the basis of the results of the 1/5 scale tests, the port also provides additional background information and HCOG decided to conduct a larger scale test program to insights related to hydrogen behavior-obtain thermal environmental data more representative j

of a Mark 111 containment. Subsequently, the HCOG un-

'fhe specific tes; programs considered necessary by dettook an extensive program to better define the condi-HCOG to support the unique plant characteristtes of a tions that could exist duririg a degraded-core accident. A i

Mark Ill containment are discussed below, major element of tbis effort was the quarter-scale l

4 Mark Ill containment combustion test program.

. 3.1 Small-Scale Tests Small scale hydrogen combustion tests were perfor ned at Whiteshell laboratories and documented by the The quarter-scale test program became the major ele-HCOG in a letter dated June 7,1984 (HGN-017-NP).

ment of the HCOG's hydrogen research program.The The program was intended to investigate ignition and primary objective of this program was the investigation combustion behavior of mixtutes predominantly com-and characterization of the environment that could result t

posed of hydrogen and steam, that is, with limited avail-from diffusive burning on the suppression pool in a able OAygen. This condition may exist for a postulated Mark Ill containment. Ultimately, the information gath-4 drywell break event in which air is initially swept from the cred from the quarter scale test facility (QSTF)would be drywell and then later rcintroduced into the steam-used in determining the survivubility of select equipment.

hydrogen environment. These tests confirmed that such Test facility description and the combustion test results hydrogen air-steam mixtures can be successfully ignited may be found in the HCOG's letters HON-098-P, as long as the oxygen concentration exceeds approxi-July 18,1986; HGN-II5-NP, February 10,1987; and mately 5 to 6 volume percent.

HGN-121 P, July 22,1987.

Acurex Corporation conducted a 1/20 scale Mark 111 The test facility is a quarter-linear-scale model of a hydrogen combustion program, which the HCOG Mark 111 containment designed and constructed by documented in a letter dated February 9, 1984 Factory Mutual Research Corporation and located in (HON-014-NP). The objective of this program was to West Gloucester, Rhode Island. The test enclosure is provide a visual record of hydrogen combustion behavior designed to operate at pressures up to 40 psig and consists in a 360-degree model of a Mark 111 containment. Model-of an outside tank,31.5 feet in diameter and 49.4 feet ing included the suppression pool and major blockages in high, containing a smaller tank, about 21 feet in diameter the annular region between the drywell and outer con-and 23 feet high. The space between the two tanks is the NUREG-1417 4

i l

test volume and contains floors and ot her large b'ockager is used to modify the vesselinterior when needed. At the simulating the obstructions that exist in the actual con-bottom of the two tanks, the suppression pool is simu-tainments. Ilecause of the unique features of tne four lated. Several views of the facility are shown in Fig-plants studied, modular construction of the an.iutar floors ures 3.1 and 3.2.

31.C' LD.

464'

-rq\\

ur It'

}' M(

Y 7d m N s W n,,. _

nt L - sr to.fr-lo 6' w

+

s.

A

.a.

y y

3 hY

[

so.

Y o<

seo.

tro.

b.--ra.rr W497//7X41/7NN (W MAf7//17/H 7/7/77 F ll l

Figure 3.1 Elevation views of the quarter-scale test facility Source: 11GN-098-P STL58 0

,/"

CONtilleifNT RJICI4f.E.1 M.00*

PM' 135'

,/

r y l

I L

x N

M,,

s Q'&ll

,N. )n'

,d

/

m

. s ih'fil s

j m :.

\\

x r,

as,,n.

" ':{jjjp(

\\

b?[i' I' ' s

' y. ::

.f I-I

\\(/

9 N

i f JF

.c

0

_l h/

toch N

')

4 Pl9

./

1. s TESTS 8.07. 07.1 only \\\\

h GAAllC N

s' (for Clarity Only Grattivj at 6.$' til is shown)

Figure 3.2 Facility configuration tests S.01-S.Il Source: ilGN-098 P 5

NUllEG-1417

l i

l l

Facility design features include To assess the effect of these discrepncies, the HCOG i

provided a comprehensive analysis as documented by o

containment sprays llCOG lester dated May 5,1986 (llGN-085). The HCOG determined that compensatmg effects existed m o

simulated loss-of-coolant accident vents (top row the treatment of heat sinks.'Ihus, the data obtained from only, numbering 48) the QS'IT does provide a reasonably accurate description f the full-scale thermal environment when extrapolated e

simulated spargers (uniformly spaced every 15 de, by I roude scaling.

grecs azimuthally and totalling 24) e unit coolers (for the River llend configuration)

To assist the staff in the review of this complex matter, Sandia National Laboratory (SNL) studied the subject The facility has several instruments to measure gas and analysis and submitied various comments. On October 7 j

surface temperatures, gas velocities, gas concentrations, 1987, a meeting between the llCOG, SN!4 and the staff heat fluxes, and pressure, and five video cameras are used took place to resolve SN1/s comments. Subsequently, the for a visual record.

HCOG documented its responsen S a letter dated No-vember 6,1987 (llGN-128t un the basis of the addi-tional information, SNI. concluded and the staff concurs 3.2.1 Scaling Methodology that the application of quarter scale experimental data di-rectly to full-scale equipment survivability can be done The theoretical basis for modeling hydrogen flames in the conservatively in spite of the above discrepancies. SNL test facility is based on Froude scaling. The modeling as' documented its assessment in two letters to the NRC sumes that fully developed, buoyancy dominated, turbu-dated December 23,1987, lent flows are achieved to preserve the equivalent value for the Froude number in the model and in the full scale plant. The technique of Froude scaling is supported by 3.2.2 Quarter Scale Testing Approach numerous experimental demontrations in the field of fire research.

Tests were performed in the QSIT for the four different plants with a Mark 111 containment; i.e., the QS'IF was

" I'* * "

"" E Usin8 this type of modelinS, the full scale was scaled to characterist. E "f each Mark Ill configuration. Durmg the ics o quarter scale, a 4-to-1 linear scaling, resulting in tests, hydrogen was released through spargers used to simulate the automatic depressurization syste.n (AFZ,;

i e

32-10-1 reduction in mass and volume flow rates and a stuck open relief valve (SORV) or thinuch the simulated vents used during a loss of coolant accident 64-to-1 reduction in total hydrogen released (LOCA).Two hydrogen release profiles were used in the e

e 2-10-1 reduction in the time scale facility, a low reflood case (l$0 gpm) and a high reflood case (5000 ppm). A discussion of the development of e

2-10-1 reduction in gas velocities these profiles is contained in Section 5 of this report.

1-to-1 relationship for gas temperatures and gas concentrations However, before the plant unique or production tests were conducted, a series of scoping tests were performed Flame heights and global flow patterns also were deter-to ast ess data repeatability and the significance of va[ious i

P"fameters. Results from these tests formed the basis for mined by Froude scaling. Generally, Froude scaling was des. clop, g the final test matrix that was used in the pro-m used to reasonably and practically design the QSTF (e.g.,

duction test program. Also, m the early development of i

spray flow and droplet sizes, heat sink thermal character-p stulated degraded core events, spray availability was istics, blockages). However, the following discrepancies uncertam; therefore, tests were performed with and with-wem sted' out sprays activated. In the production tests, each plant had its own specific array of tests, focusing on the SORV The quarter-scale tests revealed that the insulation locations, the combination of ADS spargers and LOCA o

used in the facility became wet and its thermal prop-vent releases, and the effect of sprays / coolers.The data crties departed from dry insulation, obtained from these production tests formed a basis for e

The OS1T had only 30 percent of the mass as pre-determining the full scale thermal environment and be-scribed by Froude scaling.

came a central element of each licensee's final analysis.

l This mformation was used as input to analytical evalu-1 The scaling method did not rigorously simulate con-ations of equipment thermal response for assessing sur-e vective and radiative heat losses to structural heat vivability of critical equipment. Further discussions on t he sinks.

use of this data are contained in Section 7 of this report.

J NUREG-1417 6

q l

1

1 3.13 Quarter Scale Test Results Another effect that accompanied hydrogen burning was the formation of bulk air currents. llorizontal air flow was

'Ihc scoping test and partial production test results are created above

• he pool surface allowing diffusion burning summarized and presented in the iICOG's cortespon-to continue b; providing a source of oxygen. Another pat.

dence (llGN-0EP and ilGN-121 P). 'Ihese results tern of circulation was created by chimneys, which pro-demonstrated that the distributed glow plug igniter sys-y de for flow to and from the region of burning and ex-tem can providc an cifective means for limiting accumula~

change flow with the upper containment, that is, hot tion of hydrogen m plants with Mark 111 containments.

(upward Dow) and cold (downward flow) chimneys, liydrogen concentrations throughout the facility were maintained r car or below $ volume percent (dry basis) for I ocallred Combustion all tests, and steam concentrations were determmed to be about 10-15 volume percent for selected tests. Although Delow the flame extinguishment limit, flames on the pool low hydrogen concentrations were maintained, dJferent were not observed. The prevalent burning mode at very types of combustion behavior were observed during the low hydrogen release rates has been termed

• localized tests, depending on the synergistic conditions. 'Itc vari-combustion." This phenomenon is characteriicd as weak ous observed combustion modes are described * 'ow, flames or volume burning through a marginally combusti-ble hydrogen-air stcam mixture.This type of combustion UlfTusion Flatnes was detected nly in regi ns at r ab ve the llCU floor and concentrated mostly in chimney areas. 'fhis was evi-When hydrogen was released into selected spargers, hy-dent by temperature measurements; localized combus-drogen combustion initiated as a mild deflagration or tion was not observed by video tecordings. lecallied com-lightoff (pressure rise about I psi) in the wetwell region bustion appeared to be relatively benign (i.e., less tnan between the hydraulic control unit (!!CU) floor and the 250 'l' at instrumented locations). Ilurning was more suppression pool surface and persisted in the form of widespread and somewhat more intense at low oxygen stan& diffusion flames anchored to the pool surfacc.

conditions and was accompanied by slightly higher back.

'lhis was the dominant mode of combustion and occurred ground hydrogen concentrations (i.e., near $ percent),

for bulk oxygen concentrations of 8 volume percent (dry) and hydrogen injection rates greater than 0.15 lb/sec. It The QSTP was oriented to investigate the burning phe-should be noted that the hydrogen flow rates are full, nomena in the area immediately above the suppression scale equivalent values (i.e., a 32:1 increase). In this re, pool. As a result, the instrumentation layout above the gime of steady flames, combustion was essenthily com-IlCU floot was not sufficient ta provide a detailed map-plcte. l'or an injection rate of I lb/sec, flame height ping of the conditions in that region. Consequently, a reached about 8 feet (full scale).

rigorous investigation of kicallied combustion was not possible; however, the instrumentation that was present.

As the hydrogen injection rate was decreased to about along with IICOO's analytical effort (see Section 6), pro-0.15 lb/sec, combustion became less complete and the vided a reasonable charactcrization of the phenomena, flames less stable. As the rate was further decreased, dif, localized combustion is discussed further in Section 6 as I

fusion flames on the pool surface could not be main, it relates to the analytical methods used by the 11C00, tained. *lhis point is known as the flame extinguishment limit. Moreover, it was observed that this limit was Secondary Hurning strongly influenced by background gas concentrations.

During the quarter scale testing, an additional combus-tion phenomenon was observed late in one of the tests.

The following extraction from the QqlF scoping test te-When the bulk oxygen concentration dropped below 8 port illustrates the various relationships:

volume percent (dry), flames extinguished on the surface of the pool but formed at the I-lCU floor elevation.This The flame extingu. hment limit ranges from ~0.07 type of burning has been termed " secondary burning."

e is to ~0.15 lb/sec for ambient hydrogen concentra-

'lons below ~4.1 volume percent dry (high oxygen During a June 1986 meeting with the staff,11C00 re-(onditions).

vealed the presence of secondary burning in one of the o

Tae flame cr$i::pishment limit decreases with in-

% production tests. 'Ihis phenomenon was not ob-cresig i ydrogen concentration to a minimum be-served in previous production tests or m the scoping test tween ~0.025 and ~0.03 lb/sec at a hydrogen con-pham, Until this particular test, only in a scopmg test did centration of ~4.5 volume percent (high oxygen tM contamment oxygen concentration drop below 8 per-conditions.)

cent. Oxypc9 concentrations were generally mamtamed above 8 perant as a result of a unique need associated For comparable ambient hydrtgen concentrations, with the video coverage. Each of the five video cameras o

the flame extinguNhment liuit is slightly hi r at usc<. n me OSTF required a continuous air purge for the i

low oxygen conditio,'s.

ch ra lenses to prevent a condensation on the lens.This 1

i 7

NURiiG-1417

f resulted in a continuous inflow of oxygen in the facility, value, could reduce or possibly climinate secondary thus precluding the atmospheric oxygen concentration to burning.

fall below 8 percent. Ilowever, the camera air purges were not run continuously in the subject testi subsc.

(3) Also, not all conditions expected to exist in an actual quentlyglate in this test the oxygen concentmtion fell be-plant existed in the test u hen secondary burning was low S percent. Additional information is provided by obsened f or example, sprays were not activated IICOG rm Tittal llGN-106 P dated September 29, and the burning on the llCU floor was in a sector 1986, ano uso discussed in detail in the quarter scale where the hydrogen release was most concentrated.

combustion test report.

'Ihis sector was in the 45 degree chimney in which the SORY was located and the steam tunnel struc-To present its overall assessment of the significance of ture would reduce the upward cross sectional flow secondary burning, llCOG began by addressing the limi.

area.Therefore, it is expected that these factors con-tations of the QS'IF. The QS'IF has various physical and tributed to the locally high concentration of hydro-practical limitations associated with the investigation of fen that is required for secondary burning, it also secondary burning. The instrumentation in the facility should be noted that the overall shape of the flames was geared to defmc the thermal environment produced occupied a relatively small area, forming a flame by diffusion flames anchored to the surface of the sup.

zone near the corner of the steam tunnel and drywell pression pool, which is the dominant combustion rnode, wall. On the basis of these differences, the following

%erefore, more instrumentation would be needed to in, significant mitigating factors can be infcrred to re-vestigate burning above the 11C0 fkior. Since all plants duce the consequence of secondary burning:

with Mark Ill containments have different containment e

increased turbulenc inside containment volutnes, simulation of the expected oxygen depletion ugh spray opualm.en or unit cochg couM profile for each plant would be difficult. Therefore, the 11000 cvaluated the need to further consider the secon-potentially delay or preclude secondary burn-ing.,Ihts was evident to some degree m one of dary burning phenomenon by considering the following factors and their relationship to the four plants with Mark the scoping tests during which conditions werc ill containments:

similar to those during the Perry test where sec-ondary burning was present This scoping test (1) Secondary bu.aing is expected to occur over a nat.

had sprays functioning and the oxygen concen-row range of oxygen concentrations, approximately tmdon fell to approximat% 7.8 percent. Sec-from 6 to 8 percent. On the basis of the hydrogen ondary burning did not occur. Also, sprays / unit generated by a 75 percent metal water reaction, a coolers would provide cooling to mitigate the l

Mark 111 containment would experience this oxygen consequences of secondary burning if it were to "CC"'

contentradon interval late in the transient or not at Secon6ty burning appears to be extremely 10-all. Assuming the drywell air is not added to the con-tainment inventory, a metal water reaction of 55 to calizeJ. It eccurred in the region above the lo-1 67 percent would be reached before the oxygen con-catbn where three adjacent safety relief valves centration is expected to fall below 8 percent. His

(.cRVs) spargers released hydrogen. Further, range applies to three of the four plants. Ilecause of 8ccondary burning occupied only a small zone.

thelargercontainment volume to-power ratio, Clin-llecat:sc of equipment redundancy and separa-ton is not expected to have an oxygen concentration tion, secoadary burning is expected to affect i

fall below 10 percent; thus, secondary burning is not only one train of equipment.

anticipated. When the drywell air inventory is in' On the basis of its findings, the HCOG determined fur-cluded in the containment region, a metal water re-ther experimental investigation of secondary burning was action of 67 percent would be reached for Perry be-not necessary, fore the oxygen concentration is expected to fall below 8 percent: River llend and Grand Gulf would The staff's review of the evidence indicated that secon.

have already consumed the equivalent hydrogen re.

dary burning is not expected to present a significant addi-quired by the rule (l.c.,75 percent of the fuel clad-tional threat and,if this combustion mode were to occur, ding surroub %g the active fuel region).

it is expected that the thermal zone of influence would be limited. Therefore, the staff agrees that further detailed

?

(2) Considerable uncertainty is inherent in predicting study of secondary burning is unwarranted, llowever, the long term hydrogen profile, especially in the lat-since the redundancy of equipment (i.e., spatial separa-ter phase of the profile (refer to Section 5). For ex-tion of equipment performing the same function)is the ample, an alternate accident sequence such us a most important element, the staff requests that the Perry, drywell break sequence, different hydrogen release River llend, and Grand Gulf licensees (excluding Clin-rates, the use of the drywell mixing system, or not ton) confirm that sufficient separation (l.c., at least a even reaching a 75 percent metal water-reaction 90 degree azimuthal displacement) exists between the j

NUREG-1417 8

j

redundant equipment expected to be affected by secon-located between the containment wall and drywell wall, dary burning.

lletow the pool surface, horizontal vents are constructed in the drywell wall.*lhe principal difference between the four plants is in the characteristics of the containment 4

CONTAINMENT STRUCTURAL, shell asillustrated inTable 4.1.1:or arand oulf and Clin-L,Al % CI,I,i, ton, the primary containment is a stect lined, reinforced concrete structure consisting of a vertical cylinder and a hemispherical dome top. l'or River Hend and Perry, the

'Ihe burning of hydrogen inside containment has the po-primary containment is a free-standing steel vessel con-tential to induce pressure excursions in excess of the con-sisting of a vertical cylinder and a torus spherical ('r e tainment/drywell design values. To determine the pres-surrounded by a concrete shield building. 'lhe inte.ani sure capability of the containment structures, required by containment design pressure of 15 psig is the s;une for 10 CI'R 50.44(c)(3)(iv)(ll), each licensee provided its each plant. The ultimate pressure capacity was deter-plant specific analysis for staff review.The details of the mined to be alvut three times design (i.e., approximately staff's evaluations regarding the containment and drywell 50-60 psig) for each plant. Since the drywell structure is ultimate capacities are documented in each of the plant's designed to greater pressure values than the containment respective SIIR supplements. Rather than repeating vessel, t he drywell ultimat e capacities also are greater and these evaluations,a brief description of the hlark 111 con-are not limiting in the forward or reverse direction. *lhe taeme'.t will be provided.

containment pressure capacity, taking into consideration limiting containment penetrations, is used as the limiting in the hiark !!! containment design, the containment parameter when evaluating the consequences of hydro-completely surrounds the drywell. At the bottom of the pen deflagrations inside containment.1 igure 4.1 is an 11-containment, a 360 degree annular suppression pool is lustration of a h1 ark Ill containment configuration.

Table 4.1 Comparison of IlWit hlark 111 containment characteristics Characteristic Grand Gulf l'erry River llend Clinton llated thermal output, h1Wt 3.833 3,579 2,894 2,894 Number of fuct bundles 800 748 624 624 Drywell structure:

Design pressure, psig 30 30 25 30 lixternal design pressure, psid 21 21 20 17 Air volume, ft3 270,000 277,685 236,106 246,500 Suppression pool volume (includes vents), it3 1.3114 1.12114 1.3114 1.1114 Suppression pool surface area, ft2 553 482 522 455 3

50,000 40,564 20,353 33,804 l

Itoldup volume, ft Ilo! dup surface area, it2 3,145 2,617 2,564 2,490 Containment vessel:

Design pressure, psig 15 15 15 15 Ultimate pressure capacity, psig 56 50 53 63 Ilxternal design pressure, psid 3

0.8 0.6 3

Total air volume, ft3 1.4116 1.141116 1.192116 1.551116 Air volume below hydraulic control unit floor, ft3 151,644 181,626 153,792 173,000 Suppression pool volume, fL3 1.24115 1.06115 1.2 8I15 1.35115 Suppression pool surface area, fta 6,667 5,900 6,408 7,175 3

36,380 32,S30 0

14,655 i

Upper pool makeup volume, It l

Containment spray llow rate (1 train), ppm 5,650 5,250 0

3,800 Number of loss-of coolant. accident vents 135 120 129 102 9

NURiio-1417

/

y!%

i g

,p tDh*HWim lT mm

._.- m.- _ /

n Mm..au

' suDt*

yp M.

j 75%

a r

/

,ssnssa as i

Q pw sus

=

. s

.,f,

- :.= ';!)

V iSh~qf0*4 51MhN!:..j.Mb 4 ;?"'"

N n. m v. n.s. m : v e :. m.3 M Figure 4.1 'lypical Mark Ill Containment Configuration 5

DEGRADED CORE EVENTS AND The 1-1C00 analy7ed two degraded-core accident se-Ili,DROGEN GENERN1, ION quences (11C00 transmittals llON-003, -006, -018 P.

-031. -052, -055, -072, -104 P. -112.Np, -129 P and

-132).The base case scenario begins with a loss of offsite l

5.1 Introduction power, followed by reactor scram, isolation of both the containment and main steamline isolation valves, and the i

unavailability of the power conversion system. One dicsci To determine the consequences of hydrogen burning, the generator fails to start and the relief valves cyc!c on high hydrogen generation release rnust be addressed to estab-reactor pressure as a result ofisolation of the main steam.

lish a sepresentative hydrogen generation event and de-line isolation valves. Itclici valve cycling results in one fine representative hydrogen release profiles, stuck open relief valve (SOltV). The second scenario modcls a small break in the drywell by using the same to-l

'the regulation,10 CFil 50.44(c)(3)(vi)(ll), specifically re.

tal hydrogen and steam release histories as the previous quires that the following be considered in the analysis; case but the predicted hydrogen and steam release is mechanistically split between the drywell and the contain-l large amounts of hydrogen generated after the start ment.

e of an accident (hydrogen resulting from the metal-

.Ihc transients resulting in an SOltV were selected (1)to f

water reaction of up to and including 75 percent of the fuct cladding surrounding the active fuel region, ensure a rapid loss of inventory and (2) to account for and create a limiting local thermal containment environment excluding the cladding surrounding the plenum vol.

" * ')

for analysis and testing. Small break loss-of coolant acci-dents (SilLOCAs) were selected as an alternate sequence o

the period of recovery from the degraded condition to address the potential and consequences for hydrogen combustion in the drywell. Otherwise the SilLOCA se-accident scenarios that are accepted by the NitC quence is ider;tical to the SOltV sequence, e

staff and that are accompanied by sufficient support-I ing justification to show that they describe the be.

For the base-case scenario, all ac-powered reactor havior of the reactor system during and following an makcup systems are assumed to initially fail. According to accident resulting in a degraded core emergency procedures, the operator will depressurize the NUllI!O-1417 10

reactor w hen water level decreases to the top of the active (llGU) sequence and the acceptability of the fuel or when conditions requiring steam cooling are met, llWRCilUC to estimate hydrogen production histories.

Following vessel depressurization, low pressure system injection is assumed to fail. 'lhe scenario continues with 5.2.1 Acceptable Ilydrogen Generation-the core becoming uncovered and core heatup beginning I,. Vent Sequence about 35 minutes into the transient. Limited hydrogen is produced during core heatup. About 65 minutes int a the The analysis required by 10 CI'It 50.44 is to be based on event, the core is reflooded before it become, non-an accident sequence that is " acceptable to the staff" and recoverable (exceeding a 50-percent zirconium tr. cit frac-is at the same time limited to " recoverable" events. The tion). During reflooding of the core, a significa.it amount rule, however, does not provide criteria for the determi-of hydrogen is generated. *Ihis hydrogen is treasported to nation of " acceptability" or " recoverability."

the suppression pool through the safety tell,f valve spar.

gets and into containtnent where it is ignite J and burned.

The staff position with regard to recoverability is that there should be a reasonable expectation that the original

'lhe selection of the SOltV sequence was based on the re-core geometry is generally maintained, llowever, a quan-actor safety study methodology applications program titative definition of a degraded-core state that is recover.

(llSSMAP) study. In 1986, the staff questioned 'he ab-able is not required. The degraded core condition is a sence of the station blackout (S110) sequence l'etter condition in which the reactor core has experienced or is dated February 21,1986). IlCOG held the view e iSilO at the onset of experiencing damage from excessive tem-is not a likely 11011 contributor because of its remtively perature (including permanent deformation or localiicd low core melt frequeng (he itSSMAP study (NUltEG/

llGN-0$$). 'Ihis conclusion is melting). Inherently in the degraded core condition is an based on tue results 01 t extended loss of coolant injection without a chance of im-Cit-1659), which assumed Grand Gulf to be representa.

mediate recovery. The purpose is not to associate core tive of the four plants with Mark 111 containments. The recoverability with detailed phenomena of cladding or results of the GliSSAllll probabilistic risk assessment fuel melting and relocation, but rather to provide a rea.

(Pita)(NUitEG-0979) also found that Silo is a domi-sonab!c cutoff as far as the deterministic calculation of nant contributor to the probability of core damage, al.

hydrogen production is concerned. *Ihe total amount of though the core damage probability is quite low.The 111 A hydrogen production that must be considered is specified results were reinforced by the staff findings reported in an in the rule itself. It is in this limited sense that the term NitC report (NUllEG-ll50, second draft). In view of

" recoverable" is used in this evaluation, these studies, the llCOG revised its submittal to account for Silo. 'lhese revised results are contained in two re.

The 11C00 proposed a definition of secoverability in ports transmitted by letters dated January 8 (1IGN-114) terms of the fraction of Zircakiy cladding that has reached and September 9,1987 (llGN-123). 'Ihc staff's review or exceeded the Zircaloy melting temperature of 2170 K.

focused on those revised results and also drew on infer.

The staff accepted a Zircaloy cladding melt fraction of 50 mation from previous submittals.

percent as the cutoff point for " recoverability" on the ba-sis of IICOG's report that analyses indicate significant Additionally, the review effort focused on the hydrogen fuct melting is in progress at this point. It is the staff's release rate profiles that were derived using the boiling-judgment that the maintenance of the original core ge-water reactor core heatup code (llWitCIIUC). The com-ometry is unlikely after damage to this extent.*lherefore, puter code was described in Science Application Inc. and for the purposes of hydrogen rule considerations and hy.

International Technical Sersices reports to the staff; drogen generation rate estimates, the 50 percent Zir-IIGN-020, -031, -032, -034, -089, -096, and -132; and caloy melt fraction criterion is acceptable.

during an llCOG/NitC meeting on August 28,1984.'lhe objective of this review was to ascertain the capabilities With regard to the " acceptability" of sequences, the staff and the acceptability of the llWitCilUC for use in gener.

considered two criteria: the likelihood of a given ating the hydrogen generation profiles. Particular sequence and the contribution to risk from a given llWitCilUC concerns were (1) the Zircaloy oxidation sequence, On the basis of the second draft of model, (2) the transient simulation capabilities, (3) the NUllEG-ll50, the staff concluded that (1)the most ability to predict the maximum expected hydrogen pro.

likely llGEs would occur with the reactor vessel depres.

z duction rate, and (4) the ability to predict the total surized, (2) the potential for greater consequences is as-amotint of hydrogen produced in each transient.

sociated with IIGEs at high pressure, and (3)the risk from all llGEs is estimated to be extremely low.

5,2 Evaltiatioli in assessing which sequences should be considered by t he 11C00, the staff also considered the uncertainties associ-The evaluation was divided into two parts: the establish-ated with low-and high pressure events. For low-ment of an acceptahle hydrogen generation event pressure events, the requirements of the rule force 11 NUllEG-1417

conditions that are physically unrealistic (e.g., that the the staff by letters dated January 8 and September 9, core be recoverable when 75 percent of the Zircaloy is 1987.

oxidi/cd), which results in sequences that are somewhat artificial and, therefore. considerably uncertain. l'or high.

The results of analyses of Grand Gulf (NUllEG-ll50) pressure events, these uncertainties are further compli.

indicate that the most probable llGlis result from Silo.

cated by a further lack of experimental data.

'the most likely of these sequences (designated as TilU scquences) consists of loss of offsite power followed by The staff decided that it is sufficient to consider only low.

the failure of onsite ac power in divisions 1 and 2 the fail-pressure sequences because (1)the overall risk from ute of high pressure core spray (llPCS) and itCIC, and ilGIls is believed to be low (2)from a risk perspective depressurization of the reactor vessel. TilU represents the reduced likelihood of a high pressure llGliis likely to more than 90 percent of all 11Gli sequences and more offset the potentially higher consequences of such an than 93 percent of alllow pressure 11011 sequences.The event, and (3) the additional uncertainties associated with development of the loss-of offsite power (depressurized high-pressure event progression.

vessel) sequence begins by boiling off the entire reactor vessel coolant inventory. With the core dry but at pres-sure, the operator depressuri7es the reactor to increase 5.2.1.1 The llCOG's Dase. Case Scenario the length of time available to support core recovery be-The base case scenario proposed by the llCOG results fore the initiation of core damage. l'ollowing vessel dep-from a transient caused by loss of offsite power, subse-ressurization the core begins to heatup causing oxidation

• " " 7 "."O *

• d" * "E ' " EIO * *

• E quent :eactor scram, isolation of the main steamline isola-tion valves and the containment.and one SORV. All ac-pgrem to a pmnt where a nonrecoverable core ge-powered reactor makeup systems are assumed to fail ometry muW & clop, a papr vessel nDmd system is initially. llowever, the reactor core isolation cooling assed Io be remed. Dus n%od system then cmen (RCIC) system, which is de powered, and/or the fire truck the fuel region with water termmating the event with a diesel supply are available at Grand Gulf, limergency op-degraded, but recoverable, core geornetry, crating procedures require depressurization when the water level reaches the top of the active fuel region and ?

5.2.1.3 TilU the " Acceptable Sequence" tow pressure injection system is available. It is assumed lhe llCOG had considered the applicability of the vari-that following depressurization, the low-pressure system ous significant sequences identified in the draf t version of fails to inject. 'lhe core becomes uncovered and core tem-NUltliG-1150 to the llCOG program. The scenarios perature begins to rise about 35 minutes into the tran*

were divided into thnc categories: the short ter m (about sient. As the core temperature continues to rise some hy-I hr) damage statesTilU,TilVX,TCUX the intermedi-drogen is produced. About 65 minutes into the transient, ate (4-6 hrs); and the long term (8-10 hrs) sequences Til, the core is assumed to be reflooded at a high flow rate.

Tilu t, and TOUX (ilGN-123).' Differences, however During the refloodmg of the core, large amounts of hy-became apparent when the results of the revised drogen are produced and transported to the containment NURl!G were reviewed. The revised version of through the safety relief valves. At this point in time, the NURiiG-il50 estimates TilU to be the dominant 11011 core has reached thr recoverability crituon (i.e., the Zir-sequence, which accounts for 93 percent of core damace cahiy melt fraction is at about 50 percent),

frequency. The phenomenology of the TilU sequence'is similar to that of the llCOG basepase with regard to the A variation to the SORV sequence is the SHLOCA sce-SORV. Ilowever, the llCOG cxperimental testing and nario resulting from a hypothetical drywell break. The analyses, which encompasses the TilU sequence, as-drywell break is essentially the same as the SORV se-sumed the igniters were continuously power ed, including quence except that the hydrogen is discharged into the during the portion of the transient when ac powerwas not drywell as well as through the safety relief valves (SRVs),

available.

The staff considered this conservative sequence to evalu-ate the effect of hydrogen burning in the drywell where in adJition, the rule requires that the containment strue-essential control equipment cabling is found.

tural integrity and a safe shutdown be established and 5.2.1.2 Station tilackout and NURI:G-ll50

• I"pn**nunn aucunient n tation.the ie.rms Tit.111U,1110x, TliUI,1CUX.and IOUX' denote the folkwnp TDisstationblack-out. Ti1U n the has of offsite smer (tDSP)wnh failure of all high-The results of the GIISSAR li pR A (NURiiG-0979)also prenure f unctions. ne sarety rehef valves arc operational and the ves-found that Silo is a dominant contributor to the I,robabil-s :l is depressurized. TUUX is the IDSP with lau of all ac divisions and f ailure to de.:enurize.111U1 is the 1DSP with failure of ac divisions 1 ity of core damage, although the core damage probability nna 2 and d the llPCS. The FCIC operates for 6 to 8 hours9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br /> before is quite low.These results were reinforced by Grand G ulf faihnp a resuli of h(h preuure in the turbine exhaust.KUX is an findings documented in NURiiG-1150 Subsequently, f3$,Tg'fheY[s'a"N'IE"$'tN[he$Y[*f a$"' el'*n" Ic the 11C00 submitted information to account for SHO to core conhng spray functions execpt puer.

NURiiG-1417 12

_.~

maintained. 'the ability to satisfy these requirements de-reflood rate scenario is far less than the required 75 per-pend on both the total amount and the rate of hydrogen cent. For Grand Gulf, the active core region claddmg is production.To estimate the maxirnum hydrogen produc-79,100 lbs. For the oxidation to proceed as: Zr +

21I O--.ZrO + 211, the amounts of rirconium that cor-tion rate and the total amount of hydrogen produced, the t

2 2

rate of water supply in the recovery phase of the llGE is respond to the hydrogen released in Figures 5.1 and 5.2 critical. For purposes of the hydrogen control rule, the are 20,450 lbs and 13,700 lbs, tespectively, which repre-Ti1U sequence as described in NURiiG-1150 (which en-sent about 26.0 percent h1WR and 17.0 percent h1WR, compasses the SORV as described by the 11C00) is an respectively.* " A mechanism was needed to increase acceptable sequence leading up to core recovery. 'Ihe the release profiles to 75 percent hiWR of the active core staff finds the TilU acceptable for the time sequence of region cladJing, as required by the rule. The 75 percent events and for the hydrogen production rate and total h1WR of the zirconium at Grand Gulfis $9,300 lbs, which

amount, when oxidized will create about 2,600 lbs of hydrogen. It must be pointed out that mechanistic models that account 5.2.1.4 Ilydrogen Generation Profiles for 75 percent htWR of cladding oxidation result in a se-verely damaged core exceeding the recoverability crite-R 'covery of cooling water flow is cIfectively bounded be-rion. There are many possible scenarios that can be hy-twc'n 150 ppm from a single control rod drive cooling pothesized to yield 75 percent htWR claddmg oxidation; pumi. to 5000 ppm from the emergency core cooling sys-however, no attempt is made to estimate the phenomena tem lov pressure high flow. rate core recovery system, associated with such an oxidation level because it would The hydrogen generation profiles for these exttcmes are require an unreasonable recovery criterion, qualitatively and quantitatively different.'Ihc probability of a high flow rate recovery is expected to be higher than As discussed previously, after core quenching in the low-rate reflood case, the calculated maximum amount of that of a low-flow rate system, because there are more high-flow systems (or combinations of systems) to inject metal water reaction is limited to about 26 percent. To water into a depressurized vessel; hence, it is reasonable rnect the rule requirement of 75 percent h1 Wit, the to assume that the operator will attempt mor e of ten to re.

IICOG submitted a nonmechanistic model used to pre-cover one of the high. flow rate systems. A high-flow diet hydrogen production based on an energy balance in a reflood rate is associated with a high, narrow spike of hy.

severely damaged core. It assumed that such a core has drogen release, while the low flow reflood rate will yield energy losses at least adequate to remove decay energy la lower hydrogen production rates but for longer times (see the core, the energy produced by continued oxidation of Figures 5.1 and 5.2) (llGN-132). These profiles have Zircatoy, and excess stored energy in the core. It also as-a been estimated by the llCOG using the llWRCllUC sumed that termination of oxidation at 75-percent h1WR computer code, which is discussed in Section 5.2.2. The takes place by quenching the core and removing all excess total mechanistic estimated amount of hydrogen released energy (1IG N-034). Considering the above, the oxidation in the low rate reflood case is higher than that released in rate will suppor' a constant hydrogen release of about the high rate teflood case.The hydroger, peak release of 0.10 lbs/sec.The staff finds that this release rate is accept-the high of the hightefloodcaseisabout35secondswide able for hydrogen release for 75-percent hiWR, as re-at half maximum, while for the low reflood case signifi.

quired by the rule.

cant hydrogen releaselasts about 8.5 minutes. A summary Therefore, for the scenarios shown in Figures 5.1 and 5.2, of the main features of both cases is given below.

the " tails" respectively correspond to 1700 lbs and 2000 lbs of hydrogen,i.e., an extension of about 17,000 seconds (4.7 hours8.101852e-5 days <br />0.00194 hours <br />1.157407e-5 weeks <br />2.6635e-6 months <br />) and 20,000 seconds (5.6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br />), respectively.

Reflood Length of Total Peak Width at Rate Transient llydrogen Rate llalf hias The staff concludes that (1) mechanistic models cannot (gpm)

(min)

(lbs)

(lbs/sec) (sec) predict the required 75-percent h1WR of cladding oxida-tion in the active fuel region without core damage beyond 150 80.0 903.4 0.95 510 the recoverability criterion and (2) the use of a non-5000 25.8 604.8 8.00 35 mechanistic release model based on heat balance is rea-sonable and acceptable. 'Ihis leads to an oxidation rate producing 0.10 lbs of hydrogen per second, requiring an extension of about 4.7 and 5.6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br /> for the scenarios of 5.2.1,5 Nonmechanistle ll drogen Release Profile Figures 5.1 and 5.2, respectively.

3 The hydrogen rule requires consideration of metal-water reaction (htWR) for 75 percent of the Zirealoy cladding

' Qca p {""j"gj'g ""g'y y D g t g g e g e{n,ixw.

i surrounding the active fuel region. llowever, the esti-OM-0.07 and Nr 0.05-015.

mated amount of metal to-water reaction in either "ne htwit incoqwates channel tu and stainless sicci oenon.

13 NURl!G-1417

IREll.ASE IBS10ftt' il. SfMW (i!%I HI.II OOth p

8-7-

~

l N

I I

6-4-

3-2-

l-0=

i i

i i

i i

i 2

2.4 2.5 3.2 3s 4

4.4 1&lE i1110t5 ANDS OF $tCOND$)

Figurc 5.1 Ilydrogen generation rate (150 gpm reflood) j WDROGEN GENERATION RATE (RELEASE HISTO.RY.A - 150 Gf'M REFLOO.D) l 0.9 -

nn -

i

g 0.7

• h!g Os -

h es.

n4 -

Im3-f a2 -

as.

\\.. %

---[/

r- -,

o 2

2.4 2A 3.2 34 4

44 48 Int.

- t 1HotsAND5 OF SECONOS I Figurc 5.2 11ydrogen generation rate (5000 gpm reflood)

NUREG-1417 14

5.2.2 IlWR Core lietitup Code dicted m mh f uel assembly represented A separate les el f or the core bypass les el varies accordmg to the bun-

-=

$.2.2.1 loiroduc tion die power because the water m an assembly is assumed to 1

be sa:urateJ at the system pressur e Water les els aiso may i

'lhe bothng-water-reactor core heatup coJe tiiWit-un because the uJ fracnon of the water m an assernbly UlllT)was the computer code used by the iIC(Ki to cal-is a'iuncuan of awembly power 'the bypass lesel calcula-culate hydrogen release rate profdes for the hydropen non f urther assumes that water m the bypass is subcooled j

generation sequences described previously. lhe staf f"s and thus corresponds to the collapsed water les ci in the I

review of the llWl(CilUC is discussed below-core. 'lhete is a hydrauhc connecuon betw t en the assem-bhes anJ the bypass so that the water levtl m the by pass is The itWl(CilUC was not sahdated or benthmarked to rcJ,ncJ as the core water imenton n boileJ away by the m

phibal core esperimental data, rather it rehes on collte dec.n heat ptneratcJ within the assen.bhes. ihis repre-tive enginecrmy juJgment and undt rstarwimp of the phe-sent5 tion clost 13 corresponds to a partially cos ert J nomena takmp place m a core disruptne accident. ~lhe BWii b core at low pressun before any structures in the l

lack of betahmarkmg or vahdation is due to the absence core reach temperatures synd wantly picater than satura-of suitable experimental data lhis lack of benchmarkmf him

=

prevents the results of the code from being used directly wnhout appropriate consideration of selected mput pa-

$m Phenomenological ksumptions r

rameters 'ihe results of the llWitCilUC should be seen as an engmeer mg estimate of the anticipated phenornena A moJel for thannel bhsLage was mcluded in the code.

Accordmply, the cale review was aimed at the reason-but has not been emphsyed m the cak ulations because ex-ableness of the modehng, the physical sipmficance of the peronental results did not support total flow bhNkape.

4 assumphons, and posuble conservatisms m the estmuite-The bhs kape moJet assumed that the f uel rod cladding

!(casonableness was assessed m terms of moJels and h)-

melts whde the channel tu remams intact. Molten claJ-potheses that have been advanced by other researchers m dmp is then assumed to slump and refree/c withm the this field and any other evidence that could be pleaned channel f ornung a complete bhskape, which prevents f rom whatever hmited and parual exper mental mforma-steam f rom reachmp the Zucaloy surlates within the as-tion was avadable. I or (ode modelmg, the Till' sequence sembly. In aJJition, steam pencration below the blot Lape for an ll(il v as considercJ equivalent to the M)lW pressurves that portion of the assembly f orcing the two-sequence tSectrm 5.2.1.4i with respect to the depres-phase level m that assembly below the core plate Smcc sur vation and core uncmen time. thus similar as f ar as no stcam enters the channel. all oudauon would stop.1 &

hydrogen rener ition is concer ned. This sequence is the penmental results f rom the Pill tests (11('O(i presenta-simplest an'l most strayhtfonsard, thus havmg the high-tion to NI(C January 14,1%M mtheated that a reduction est probab hty of bemy modeled conectly-m the flow area as a result of /ircatoy slumpmr did occur, but that complete bhNkape did not f orm. Without the The llWitCilUC is a wc!I-wraten computer code m that ch. nel bloskape moJel hydropen proJuchon is man t1) it f anhfully represents the llWit peometne coie de-nu/cd, all other condaions bemp the same. The lack of sign and (2) the models meluded in the code are adequate cladJmp monon or channel blockage is a ven conserva-to cover the specific illil.s selected for analysis by the tive assumphon with regard to hydrogen pioduction.

11C00 and represented by the THU sequence. Malular arclutecture has been used extensn ely, where each mod-It is assumed that the control rods wdl remain mtact smcc ule (subroutme) m the code treats a diflerent phenome-u is consistent wah the recoven cnterion. SanJia Na-non or aspect of the problem. The code is budt by con-tional 1 alvratory performed an expenment that suggests nectmg the vanous modules w th executive routmes. 'lhe that, under certam conditions, llWil centrol rod blades numencal solution technique applied m the llWitCl100 could melt early m the core heatup phase of a transient.

is apparently as pood as any employed in sescre-accident This would lead to the possibihty of local loss of control; cales. Numencal stability. as reported by iICOG, is evi-thus, w hen the core is reflooded, local enucahty could re-denced by the graphs el code output and the fact that sult m intense heat produchon and core damage beyond refloodmg calculations can be run.

the hmits of recoserabilay. 'lheref ore, control rod melt would be beyond the scope of this program.

The llWit core peometry is very complex. Some subliches of the peometry have the potennal toaffect the predicuon 5,2,2,3 Sleam (;eneration of hydropen peneration. Therefore, it is appropnate that a best estnnate code contam a representanon of the pe-

'lhe modehnp of steam peneranon wahm the reactor i

ometry that is as complete as is reasonably achiesable.

pressure sev el titl"Jican has e a sigmficant effect on the This has het done in the itWiiClll'C. Considerabic at-quantity of bydropen gener.ded. While some aspects of tenuon has been drawn to the fact that the BWl(('lll1C steam penerauon are accurately moJeled m the allows for a ddferent two phase water level to be pre-ItWli('lIlC other sourtes of steam ate not moJeled at 15 N Ulti (i-1417

l all. "Ihe steam generation modeling generally is ronment (llCOG presentation to NitC January l4,1985).

incomplete; however, for the most likely llGE consid-

'Ihe 11C00 estimated Zircaloy oxidation versus Zirealoy cred, the steam sources not represented do not signifi-temperature and concluded that 2400K is a conservative cantly affect the generation of hydrogen.

representation to account for this effect (IIGN-032, item 4). On the basis of the evaluation performed by Interna-Within the llWitCilVC, the following five sources of tional Technical Senices, the staff has accepted the steam generation are modeled:

2400K as the irreversible oxidation cutoff temperature (NitC letter to IICOG June 4,1985).

deposition of the decay power from that portion of e

the fuel assembly below the two phase level into the in the reflood stage, quenching of Zircaloy that is at tem.

saturated water within an assembly peratures higher than the saturation temperature is non-r, heat transfer (by nucleate boiling) from portions of mechanistically estimated. 'Ihis can lead to overpredic-fuel rods, channels, control bladcs, and the core tion of the steam generation rate during the reflood shroud that are at temperatures greater than satura-phase.1,or nodes that are more than 100K above satura-tion when they are covered by the two phase level tion, quenching is assumed to take place in a single tirne step thus accelerating the heat transfer process and steam radtatise heat transfer from portions of control generation.

o blades that have temperatures higher than satura-tion to surrounding channel walls when the two.

Steam flow in the bypass region is underpredicted. Ilow-phase level within the channel is at or above the por-ever, the effect of this underprediction of the bypass tion (at elevated temperatures)of the controlblade steam flow rate on the overall prediction of hydrogen re-case is srna Oman swam pnnadon mks in an og e

fashing of water in the downcomer and lower m e cou e unduptch for tmnsients in pienur as a result of reductions in the llpV pres-which the two phase level is above the core support plate.

sure*

Again, the extent of this underprediction is small com-evapocation of core spray droplets entering the top pared with the uncertainties associated with predictions o

of a feel assembly during reflooding of the core of this nature.

'%sm ger.crotion resulting from flashing of the water in.

The staff concludes that steam generation before reflood ventory within the fuct assemblics and in the bypass re-is reasonably well predicted provided the IlpV pressure gion is not modeled. If system pressure would decrease, has been constant for approximately 10 minutes. In the flashing would occur, llowever, the selected ilGE se-sequence considered, the llpV is depressurized and quence does not involve changes in pressure vessel pres-steam and hydrogen production take place undes condi-sure after hydrogen generation has begun. It is assumed tions of constant pressure, that the ItpV pressure is constant for at least 10 minutes, which is the time required to remove the bulk of the heat 5.2.2.4 Ilydrogen Generation in the lower plenum sti tures.'therefore, the lack of a flashing model is not a factor.

As with the modeling of steam generation, the approach to modeling hydrogen generatior. is reasonable consider.

Downward rehication of molten Zircaloy can have a large ing the difficulty of representing the phenomenn in mod-effect on steam teneration. !I t he two-phase level is above cling techniques. *lhe lack of models for a few relevant the core support plate, molten Zircaloy can run into phenomena combined with sutne of the assumptions water. Quenching or relocating Zircaloy in water would made for phenomena that are modeled, leads to some un-enhance steam and hydrogen generation. 'lhis phenome-certainty with regard to the predictions of the hydrogen non is not modeled in the llWitCilUC,Ilowever, there is generation rate during the dominant ilGE.1-lowever, this no water above the cure support plate when Zircaloy uncertainty is expected to be negligible in the present melting occurs. Modeling of the melt relocation into the

context, water would not increase the quantity of hydrogen pro-duced compared to that which will be produced in the

,Re considerations related to hydrogen generation that core reflood because of the more favorable surface to-are not modeled or are underpredicted are listed below.

Volume mtio.

Oxidation below the kication at which melting oc.

An oxidation cutoff temperature is used in the curs is not modeled.

IlWitCilUC as a surrogate for the effect of cladding and Ilecause of the underprediction of steam in the by-e channel box tekication and subsequent quenching pass channel, oxidation of stainless stccl and the out-thereby removing the Zirealoy from the oxidizing envi-side of the channelis probably underpredicted.

llallooning of the cladding and localized failure re-e

• Premne.ume istory is paidnt t>y mer input.

sulting in simultaneous interior and exterior NUllEG-1417 16

oxiJation is not inodeled. thus hnntea h Jrocen un

% n n rum m n m a :

!>.9 t " o na mn 3

derprediction rna) result

( Mi ht underpa Jn wJ l b m cit r the staH ru hn cs that nt t et sc esan m tb

q, m Pu' tM it r no e

i um bodmp m a quendung nmJe n not rooJW,

' i d II L

, J., t

' t l. 'r l,

. L '.

1. t s

s I,his leads t() hi[her rate of h)d't *[t n p!t dik inif. I t '!

sIhirter time pcIll>Js It is nin clear tr.at ari in ct pm: all i hrmta,

under prediction wdl result ig j! u p,

g;,

,. s s

In the rentuamt stare. vapor mten 'd Jr+cu ~t mn im o m 1 r a

enter the top of f ac! awcrWes by raJ5

'1 ht t it i b l s'

+

transfer does not rcrimve heat trorn lucl rob It-rl sults m a consthilliic hs dro[Un pn d')

'h+ d l' ma\\1tnum tem perat Ur e is IM h m t ht. cuti '!ItrPih E,}

ture and phk not o mt natnt 0oe Mt !!

' ((jd ( (il}(l([s, h pl} %

s ((())[]}3 p g

cutof f t em perat ur e It r rn st ( h.ar i! t he i s c r '! t '

,t r.

1 ( ;e is rmnconsenatn c.

p

n.r

n, i

a m.m Itcachon rates of /ucahiy mJ stamW t't t I w c h ut e R'

-!U ate calculatcJ uurf the Ar r teruus iclato mstop t he i t m p,,

,g,

.p

,q tion rate constants useJ m these c\\premons w cr c Jt r o eJ

,,. s t-

..It

.s s

ID allbers Ittinl C\\pcIIInental l csults Ibls rlh dclin f 4 I! I s 4,

i a

, 6

.m actmn iateumJ the asseteJ heat rer eranon n appm s,,.

pnate and wrasteni wnn what n nscJ m oacr sn t u

,n

.,i s

. m.. t,. i accident trioJehny wJes A bJniren blantet my las.ti, a mtithicd m the formuhdion of the Arrhetuus rt ict1 m in ig in i

, n:. t 11,,, sg

, i:

atan, rate exprewon. llydropen htmketmp ref ers to tM pos+

17 i s y ot h a n, I h,1 1! mho.auo.ha

,t ble limitation ol the tl\\ldalliin rate lrton thP dtNusb W ralt w h,t y g 3( t,t wt t i, r m J j ; ;. l l( ( 4, (

A n,;tifs aigj of steam throuph the hydrorco enuttedirom ttm ouJi/mr y

. te n t us m J.f a d i s d N L R1 (

!If' 1,

s surlace. Whde the pro < css represented N the b Jn < t

[1 1 nt y u,

n!,,:t; :t te pe u,

,a blankrimp lactor n real, a redochon of the ouJauon rm i ! n,1 ds s,

sJ il' apmtr i

s n alu.mt t ertunty not realucJ under the unJn m c.

L'a,JN L'mt (J<

up,t' m6is>

'aJ' pected dormp tore damare m llWRs I uf f usion of steam

,m law o1 at } l p < s,. a la h lu int a g through the oude layet n the rate hnatmy pm c' wa lomJ ! i h m'm a t i W: M > R \\. w h. ! w as na:. h

'lherefore, the hyJtoren hhmketmp eficct we not t m pop, ca m j p ( n, 11 o t'i 2 imJ t'

i the 11:1 sidercJ m the I Ml M i udculation-whhhicpn';m

,,s

< a ju e<ip,,

,a ;u ( } g y,,,

sllf t Gunervalnm h

l ht ws 4 J it a' t n t nt e la l lo t r < ao s ; t J ats.d t

g itecause thc ouJahon rate varies e\\p;mentully with tt m tH 1m kt ill (

b t att !a J !

i < 1 t l at o in pn >

perature, the repf esentation of miact / M ca h ')

In d, 1;

lM tsu: t w s ', um t t; !n sl u ;qe!i art the s

reachmp temperatures siprnhcantly hirhci than the mt it-pcA ratt. its Jutal' s anJ d c h G n aof M ooren my temperature leaJs to highet ouJauon than woul I he pn,JostJ 'Ihc 4 m t nJs that 0 e p:oh;cs c o wit J N predicted if mehing were exphcitly treated I her clot t.

Ik (Mi u v Sc BWRt ill'l at e m s t pt..Ne lot u s in this n a conservatn c auumphon. Ilown er, because the Jt m,msn atny wmphant c w ah 10 ( l R 50 44 o\\iJe layer is rencraH) thick at these times, the actual quantity of aJJitional ouJahon is conuJercJ 10 he sinall 6

CO VI,'AI N M lW".I, Rl',Sl'OM', I, -

nus mereased oudahon may he unvcJ as a non 2

mechamsoc approach to icpresenung oie n ital mer ca-AN AIXTlCAl, MODlil 1NG m oudanon that prohabl awomp.unes slumpmp monen 3

Zacaloy in sicw of N quart < t de tt 4 pn nam 16 cmphasu on analy t nal rnt :hoJ U pr cJu tmr t atamnn ni r c<pansc ileatmg of the (ladJmy rcJuces the tenule suength anJ to byJmm n burnmy has syndh "uy Ja i shcJ lb.

mcreases ductihty. Simultaneous heating of the luel and tenduuon n tt scJ on the h:oaJ iaryt of h> J' ope n i t -

pasCs Within the cladding Ic;hls tii presstirt/ation isl the It a t.ato tit whhh J, flu:st stud,1*th" n t s ', n s i t J 1,4 rod from within llalloonmg of the claddmp anJ hicahecJ os s u ' lh si21 hcN s t s t h e t s a?m no o: A e s a ahl ladure may occur before inclunp. I adure of the (ladJmy in el t ' ' t u! (. 7 c o' e i.J N h eJ, s tr J o would allow the mtenor surf ace to be c\\poscJ to steam s.- 7 J'o i

s.

is im ny iO 11 i s

lhewfate,it o cntuely posuble that hoth the mtermt and I k itl

.1 L c Js m. yinai r R r

e\\terior su!Iacts of the sladdinf u dl tmJL f pi (WiJ.ili<1 sds

'b( l D!\\

' l!

m. [

l a-s i

h!!he this pilwthdll) is lhil nudeled in thc IM N( ll[ l n a$ a ti Ha adJi t %s ll! t 'l i -

L. ; i s 4 l

't' t cuds s las NL kl(i-141'

Cl.ASIX 3 code analysis penene.dly. As such, this ef f or t tion is sharacten/eJ as weak flames or weak wlume is only relevant at low hydrogen flow rates that are near burnmg throup h a marpmally wmbustible hydropen au-the flan.e otinpunhment hmit.

steam mature By it tien dated June 10.

1%h illliN-l@P L and I)eu mber li IW' ill(iN-lli-P L As documented by sarious stafi esaluanons performcJ the I M ( M i prm iJcJ various analy ses to demonstrate that bef ore the completion of the quarter-scale test program, the Cl.ASIN -3 moJel prenJed a bounJan calculation the ('l.ASIN coJe has been the principal analytical tool m f or the wmbustmn ou urrmp below the ddlusion flame predictmp the contamment response as a wnsequente of thrtshold I his mid1 empMy cJ a wmbustion mes ha-i butnmp hydropen for plant $. with an n.c conJenser or rusm that proJaces a more ses cre p!obal thc rou enuron-51 ark 111 containment The ( l.ASIN coJe or the merit than has been mesureJ laany m the ()SII for lo-

=

( l ASIL 3 code iwhich is the latest modif cJ s ersion that cah/cd combustion.

meluJes N1 ark Ill contamment features) deals with del-lagration (dscrete-ty pel byJropen bur nmg The coJe is a lhe stafi requested tne 5anJu Nanonal l aboraton multivolume contamrnent coJL that is used to udculate i W ) to m ww Ons approd in n'lh W K P subnub the contamment pressure and temperatmc response m IE Ik IN ("' comp.mJ ( l Ah 3 prWicuons for a separate compartments. hiorcoser; the code has the Ca-patnhts to moJel charactenstics that are umque to Mark spondmp m W (1 M N prNianms of the wetwell 111 con'tamments whJe tracking the distnbuuon of the at-

"U""'

" d O'n t kmperature profde eweedcJ rnosphenc constn uents (i c., my pen. mtropen, bydr open, be wlunewupW amare id npt rinwntal dant (in g,,

the basis of this analys:s, the 11( ( M.i wnsluded that

('l ASl% 3 yn lds coastnatn e predicuons ol thermal en-t he staff s desire to demonurate venfication vahdation yonnwnq yn drogn e con nmm r my w

of the ('I.ASIN coJe has been an extensive ellort.

w Ma l he MaH quNionN We appkMnW of uung Cl.ASIN r esults have compared well with results of other openmental data pert:uning to a local phenomena to

• ""*"*I

'" P"

"} "I ". "

• P

• I"*"

NRC-accepted analytical codes and hydrogen burnmp ev penments. I urthermore, the llc ()G has perf ormed ad-I *"" "

@IIwananat tk beal wduuion pknonwnon okmJ men diuonal code vahdauon by companny the more recent m Ow larre-scale hydrogen expenment5.at the Nevada test sue u

re uahon.

to Cl.ASlX 3 code predicuons. I his is documented m In its iIGN-111-P submn tal. the i1( ( X i provided a com-II('(M Ps letter dated January N 1%7 tilGN-113) The prehenute assessment of locah/ed combustion seen m f ocus on code vahdauon, with regard to hydrogen burn-ses crai quarter-scale tests. These tests showed that com-mg, has been or pressure predictions because tempera-bustion activity, as evidenced by thermomuple responses, ture comparisons are more ddficult to predict as a result is widespreaJ durmp penods of low hsdropen flow. ~l est of Iheir time anJ spanaHy dependent fluctuations data do not indicate that concentrateJ flame energy depo-sinon occurs at faed locations.1 nerp3 deposition appears As discussed cather, the mqor element of the llCOG's to be rapiJ and diffuse anJ is dommatcJ by comecove program is the quarter-scale test program. The data ob-maing, combusuon-mJuced turbulence, plume mflu-unned hom tests were useu to perform eqmpment surw ence, and backgrounJ gas flows. 'lypicany, peak tempera-abihty analysis (see Section 7 of this report t These tests tures recorded dunnp localved combusuon are iclain cly revealed that diffusion flames on the suppression pool low and persist for short durations. The temperature re-surf ace can exist at a hydrogen injection rate as low as 0.02 sponses are cyche and return to relatively low background Ib/sec under certain background conditions. As such,it is levels.

expected f or a significant poruon of postulated degraded core hydrogen profiles that diffusion flames will be the The 11C00 analy/cd five quarter-scale tests conducted dommant combustion mode. Smce Cl.ASIN-3 does not dunng the scopmg test phase of the program m an effort model diffusion flames, these results have a significant to better understand locah/ed combustion. At the low hy-bearing on the extent to which the Cl.ASIN-3 code can drogen flow rates, these tests demonstrated certam re-be rehed upon to predict contamment temperature envr peatable trends such as recurrent and pencrally predict-ronments, which further emphasves the importance of able thermocouple activay. Some of the findmps resulting the test program.

from the evaluation of hicah/ed combustion are bnefly desenhed below.

6.1 LOCali70(I COnihuStiOH

('ombusuon was generally widespread m tests with-out sprar When sprays were activated, combustion As noted m Secuon 31, testing performeJ in the QST) appeareJ to be suppressed m open chtmneys n.e.

revealed that combusuon occurnnp below the flame ed annular q..

antsi as a result of coohng effects and tmguishment limit is not global deflagration-type com-shifts m ploh flow patterns. Also, mereased mnmg bustion but "hscalued combustionJ f ocalved combus-as a resuh of usmp sprays. mJuced shphtly higher NURIE 1417 18

teriiperatures in some areas, but not appreClably the ll(T K, adequately aMrewed the itL(ly huitions hit higher than those recordeJ when sprap were off.

hvah/ed combustion and identihed reawruble tvunJs for the most threatemny thermal ensnonm(nt for (quip A comparistin o[ slopmp test results indicatcJ that 8

not hM u rme td hsMod oun% tam in m the locatiori of the S()lWs diJ niit has c as significant ik rMmt Wt Nums %ucin t r, the tWmW loal an effect as other parameters, such as variatnin in wmpmin m wenatun eh the IR (X >\\ discussen hydrogen fhm rates of halWJ mbum p d dye pf atmn e

Probably the most important finJmp was that tene that the El ASIN 3 therrud loaJ woulJ he mine sescre peratures close to igniters were pencrally no more than that experienscJ in the U51i f or low hyJropen m-severe than rectirJ:nts several Icet awas. The closest Jeanm ratet In tim!uncton wth Aumqu amLuwd m thermo;.,)uple located to a nearby irrater was about the QSil test repto't (IR sN 1:1-Pt ni awerts tkit 15 inthes laterath or about 6 mehes lateralk anJ l$

there n reasonable awurance that the thermal response mt hes ahos e the'ipmter Also, w hen compdrmp the at hW scale wf be no niort threaterung than that everi

( f f ects of bhs L7es abm c igniters to open ihm re.

erneJ m the OS il lhe staff alw has evahiatcJ the asue yo d A. a Mgr J h.G,1 triCrease in temperature Was not aIld cinh urs M b D I a%c"W Cn I-obv n cJ W musten encrpy dispersion was preva.

lent (In the basis (){ t le rmidelmy rileth.O d,)p) used in the ref-crenccJ submittals t e.y, hm hs Jropen lion rates and the lo f urthe! 'upp ;I t its [indmps til turbulent rThNlDr in-fiscm on t he w etw c!) phiftlel the stall finds that the dut ed by con,bw.an, the i R'( X i mcluJed a discuwton of

( i Ayx. 3 predhtuins would be am nMe m dett rmm-a test m w hhh p+ ul flames were observed A thermocou-mp the comunment (ns nonmental wnJhms m a wnse-ple Was placed a!biut } lLiot direct!) abosT the po d surIMc cNe o{ loQll/cd hs dhy en burnm:o AniuJin@, these user an active sp aper. }{cadmps mdicated that at hm h)~

probles could be uded u (vMuate the surs mhihty of dropen flows the !bmes at thn hicahon appeared to be m eqmpment.

terrnlitent and Unstable. Ilowever, the tempCrature Ic-sponse diJ not en eed 4M 'I as a result of these unsteaJs pool flames The i1( '( X i contends that, because of the ell 6.2 COlllailllllCllt EI'O%til'e alid ficient rnmnp. orn should expect local tadrogen concen-Teliipel at tii e Caletilatiolis trations elsew here m the f acihty to be h w than at the sup.

presuon pool surlaa Moreover, th:, ha t (oupled with I he IU ()(i's calculanons 01 the contalnment prewule temperature readmps tJncowcJ ahm e)anJ the absence and itmperature response were bascJ on the postulated ol visualindications regarJinp flanic f ormations abus e the degraJcJ-core scenarios thai are dm ussed in Sectam 5 11('l! floor,is not stronf y supportn e of a hypothesis that using the ('l.ASIN 3 code (letten illiN DCP and l

sustamed high-icmperature locah/ed combustion zones IMiN 104PL lo Jetcimme the adequaes of the h>Jro-will be estabhshed al ver) Iow h)dropen flow tatet in aJ-pen irmtion cystem tills L the 11( (i(i consiJered two dition,1IC( M i mJicates that the ten peratures penerated typ s of accidents m as pencra analysn: a stuck-open re-flom pool burning at low hydropen flow rates (i.e. about vahe (S()RV > tranMent and a small breal low.of-0.15 thwet i hom resultant hot plumes represent a more coolant acciJent i Sill ( M' A > in the drvw c!1. The compo-severe thef rnal environmeni than IoCali/cJ combustiim-nent of the hydropen release histors that n of init rest in thn analysis n ieferr cJ to as the " tail" portion and repre-As part of this aNsessment, llC( M's provlJed adJitional in-sents a nonmechanntically defined constant hyJropen formation with regard to the role of the ( l.ASIL3 c ac pencratton rate m its analyses and the conservatisms used for ontain-Inent inodeling. While plobal or larpe vollNe deflap-As dissussed above, the Cl.ASIN -3 results bound the ration, as modeled by Cl.ASIN 3, did ne occur m the thermal envuonment that may be produccJ for low hy-OS IT, the llc (Mi contenJs that the C'.ASIN -3 model-drogen release rates that are below the diffusion flame ing would conservatnely bound the observed locali/ed extmguishment hmit. The iIC( M i providcJ a peneric sen-combustion environment 'l o assess the severity of the en-sitany study usmp th( Pern Nuclear Power Plant con-vironment f rom an equipment survivabihty perspective, tamment characteristics for the El.A51N 3 model. In this the llc (Mi compared the thermal loads created by the senMtmty stuJy, parameters were varieJ to awess the ef-mor severe hicah/ed combustion measurements at the fects on the calculatcJ r(sults. ~lhe staff focused on the QS l I to the correspondmp Cl.ASIN-3 temperature pro-most important parameter considered, which was the as-file. 'the results of this comparison show the C1.ASIN -3 sumed availabihty of the contamment sprays The !IC( Mi profile penerates a sipmficantly more severe environment chose to use the (~l ASIV 3 code predictions without than that produced b) locali/ed comblistion.

spray s (i.e. the $()RY case) iri its pencili survivabilit) st ud) (diNcuwed in Secthin 7) Ior thC eQulprnent the staff requested that SNI review this twoc along with survn abihty analy us to be genene, it became necessary to the consiJeration of scalinp aspects. SNI determined that cornider the ntospray case because fan coolets rather 14 Nl'RI G 1417

than sprays are part of the 1(iver llend containment of the hydrogen to the suppression pool, the hydrogen

design, threat to the drywell appears to be relatively small. 'l he llCOG's thermal response analysis of selected drywell equipment is discussed in Section 7 of this report.The re-For the SOi(V case, all mass and energy releases were di-suits demonstrated that the equipment would survive the rected into the suppression pool.The CLASIX-3 model drywell burn.

used in the generic analysis simulated four compartments of the Perry containment: the drywell volume, the wet-of the suppression pool), the m,floorand the surface 6.4 Existerice of Dr}well DifTtistori well volume (bounded by the HCU termediate volume (bounded by the llCU floor and the refueling floor), and Flatiles the containment volume (above the refueling floor). Fig-ure 6.1 presents a schematic representation of the model, in the DWil case, air would be reintroduced in the liydrogen combustion was assumed to occur at a drywell through vacuum breaker actuation or operation 6-percent hydrogen concentration with combustion com-of the drywell mixing system. The drywell environment is pleteness of 65 percent.Thc CLASIX-3 sol (V base-case predicted to be a hydrogen.richloxygen lean mixture, model produces a transient in which the hydrogen is ig-When oxygen is reintroduced in the presence of an igni-nited in a series of burns in the wetwell volume. Figures tion source, a diffusion flame may result in the vicinity of 6.2 and 6.3 show the computed wetwell temperature and the oxygen source. This possible combustion phenome-pressure profiles.The early portion of the transient re-non is referred to as an " inverted diffusion flame."This is suited in the highest wetwell temperature, which is symp-a concern since the potential to establish a continuous in-tomatic of the hydrogen spiked release in the early phase verted diffusion flame at the oxygen source may result in of the release profile. Diffusion flames would be preva-locally severe thermalloads.

lent in this interval and would be beyond the usable range

[

for the CLASIX-3 methodology. For the major portion of the temperature profile, the wetwell burns produce a I

peak wetwell temperature above 800 'F.

ecany cusn-3 potn e o,....,

At the end of the hydrogen release period, the calculated hydrogen concentration in the containment volume did l

not reach the ignition criterion of 6 percent, in the

- 13u..i CLASIX-3 calculation, the HCOG assumed a hydrogen I

burn to occur in the containment at this lower concentra-l tion, which resulted in the most severe pressure excursion i

j to approximately 23 psig.

"q """

I en, cu

.,ca.co..n 1

( vok. nl 8 " 888 4868888 t v04.3 6 6.3 Dr3well Analysis

u=.j j

i g....:

j I

l For the base-case analysis of a small pipe break in the l

i l

drywell(DWII case), the CLASIX-3 containment model is similar to the SORV case except the hydrogen / steam j

~'"~,

I i

source terms are directed to both the drywell volume and

[,r.,,,,,,67 hoj the suppression pool. The DWil scenario was chosen be-p"' d Q~

cause of the potential for and consequences of hydrogen i

combustion in the drywell. Of the cases studied, only the 2-inch DWII case had conditions where a hydrogen burn was predicted to occur. Hydrogen ignition in the drywell is limited by lack of oxygen (i.e., below 5 percent) because

• ====

air is forced from the drywell by vessel blowdown. The m,,,,,,.

m,m,,,

only burn predicted by CLASIX-3 for the 2-inch DWil case resulted in a peak drywell temperature of about 1050 'F and a peak drywell pressure of about 13 psig. For the events consider <,d, where high steam flows are directed into the dry well along with the diversion of most Figure 6.1 Perry CLASIX-3 Model NURPG-1417 20 l

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

1.4 1.3 -

1.2 -

1.1 -

1-C 0.9 -

yj 0.8 -

'l 1

I l

I

=

l g

0.7 -

i C

l O'O ~

l N

0.5 -

l t l i

l l

l l

l

0. 4 -

3 I

j N " ' g fe ii Mh '. ' M a l

I l

'I 0.2 -

(

Jyjinti(ohaWW O.1 -

0 0

4 B

12 16 20 24 (ftoummis TNE (5Ec0NDS))

Irigure 6.2 SOltV with no spray-wetwell temperature Source: IlG N-109 1' 40 35 -

30 -

{

25 -

S l

w g l3m m m m e ta w s t M M W " W ' 1V

[

15 -__

10 -

O' 0

i i

i 0

4 8

12 16 20 24 TiuE ((SECONDS))

Thaumnos 1 igure 63 SOllV with no spray-wetwell pressure Source: ilG N-lD9-P 21 NUltliG-1417

In its submittal of June 25,1986 (llGN-091), the llCOG would be produced by diffusive combustion on the sup-discussed the criteria for establishing the existence of in-pression pool surface in addition, the CIASIX-3 code verted diffusion flames in plants with a hiark 111 contain-analysis is used to define a bounding environment for k)-

ment.'ihe 11C00 indicated that flames will not occur in calized combustion below the diffusion flame extinguish-the drywell when conditions are outside the flammability ment limit.

curve. In its submittal of June 10,19S7 (llGN-119), the 11C00 further discussed the low likelihood of achieving the necessary combustible conditions in the drywell based 7.1 Identification of Essential on the CLASIX-3 predictions, llowever, SNL reviewed Equipluent the initial submittal and determined that the 11C00 did not provide sufficient justification to preclude drywell "Ihe !! COG selected the equipment that has to sunive burning. Specifically, SNL commented that the flamma-hydrogen burning on the basis of function during and af-bility limit merely establishes the limits that will allow ter a postulated degraded core accident. Generally, all flame propagation; burns that do not propagate into the the equipment located in the containment that was con-mixture are not precluded by being outside the flamma-sidered to be in one of the five categories listed below was bility limits.17urthermore, it was not obvious that the considered to be essential for the safe shutdown of the burning mixture should be expected to follow the path

plant, predicted by the 11C00.

(1) systems and components that mitigate the conse-Generally, there is a lack of experimental data to support quences of the accident the llCOG's position. llowever, r ecent risk studics do not support the i WII case as a dominant core melt /de.

(2) systems and components needed for maintaining the graded-core event for plants with a hiark lit containment; integrity of the containment boundary therefore, further phenomenological investigation may not be warranted. Drywell break events are further dis, (3) systems and components needed for maintaining the cussed in Section 5 of this report. In addition, the ex.

core in a coolable geometry pected redundancy (i.e., spatial separation of equipment performing the same function) of the critical equipment (4) systems and components needed for monitoring the should compensate for possible kically severe thermal coursc oi thc accident and providing guidance to the

loads, opcrator for initiating action in accordance with emergency procedure guidelines

'Ihe staff concludes there is a reasonable level of assur-ance that the consequences of a drywell break event (5) components whose failure could preclude the ability would not pose a significant threat to containment integ-of the above systems to fulfill their intended func.

tm.n rity and would not preclude safe shutdown of the plant.

llowever, the staff believes, as part of the independent plant examination process, each licensee of a plant with a

.Ihe llCOG identified the following systems and compo-hiark lit contamment should confirm the kication of criti-nents that would be evaluated for survivability in its letter cal equipment with regard to potential oxygen sources dated hiay 16,1986 (llGN-084):

through the drywell vacuum breakers or a drywell mixing system to support the above conclusion for each plant, containment penetrations: air locks, hatch seals, e

electrical penetrations, vent valves, and vacuum breakers 7

SURVIVABIIIIT OF ESSEN,FIAI, drywell components: air kicks, hatch seals, and post-e accident vacuum breakers EQUIPMENT hydrogen igniter system e

In accordance with 10 Cirit 50.44(c)(3)(vi)(II)(5)(ii), each e mbustible gas control system: hydrogen recomb-licensee with a hiark Ill containment is required to dem-iners, drywell mixing system, and post accident at-onstrate that the essential equipment kicated inside the mosphere sampling valves containment will be capable of performing its functions e

containment cooling: spray isolation valves, low-during and after exposure to the environmental condi-pressure coolant injection valves, and unit coolers tions created by the burning of hydrogen.To support this e

automatic depressurization system objective, the llCOG conducted two programs to define the environment that would result from hydrogen com-e containment and reactor monitoring: containment bustion. As discussed earlier in this evaluation, the quar.

and drywell temperature instruments, reactor pres-ter scale test data is used to define the environment that sure vessel wide-range pressure instruments, and NUlt!!G-1417 22

tertor pre.sure vessel wide-range and f ucl-/one contamment volumes ~1hc stafl f mds that based on les ci instruments hnutations of the Cl.ASIN-3 methodology used in the penene analysis, there is no choice but to use the o

associated instruments, contr ols, cables, inteilos ks' e w 11 so ume lloweset, the staf f does recognve and terminal and junction blocks the selected proide is hmitmg.

The staff ImJs that llCOG's peneric equipment surm-e in the selected Cl.ASIN-3 case, there are no active abthly list conta ns the equipment essential for the miti-contamment coohng mechanisms O.e. the lack of pation of postulated degraded core accident conJitions.

aviolabihty of sprays or unit coolers). llecause of the liowever, as part of the fmal analystA cach bcensee with a typc of event considered, a recoverable degraded Mark Ill containment shoulJ provide plant specific infor-core, the HCOG cxpects that sometime durmg mation correspondmp to the peneric list in conjunct on these relauvely long transient events, spray ' unit with unique design f eatures that are relevant to the selec-coolers would become avadable.

tion entena' A set of equipment common to each plant with a Mark lit contamment was compiled from the hst of peneric equip-7.2 Generic Equipment Survivability ment. Subsequentiy, the most thermany sensiuve equip.

ment, such as cables, pressure transmitter, hydrogen ip-Atitilys.is (Loctil.ized L,orillnist.ioni) niter awembly, and solenoid valves of the automane deptessuri/ation rystem, was included in the peneric As discuwed m Section 6, pressure and temperature pre-nt mwabihty analysis. The results of the re-dictions were obuoned by usmg the ( l.ASIX-3 cale to nene analysis were used for the eqmpment response pronde the contamment envuonment or boundan condi-analysis for the drywell break case and the results were tions necessary to perform equipment response analyses M5 110 f r the SOltV case, the thermal re-for hydropen release ra'es below the ddlusion flame ex-sponse analysis indicated that the pressure transmitter ex-tmguishment hmit where burnmg is hmited to locah/ed eded its quahncation temperature bs 27 'l1 The siprufi-combusnon the 1il ()(i beheves that equipinent surw cance of this iesult is assened below.'

abihty can be estabhshed penerically and provided sup-portmp analyws m its letter dated August 7, 19X7 The wetwell Cl.ASIX-3 temperature ptofdes (discuned tilGN 118-PL The i1C00 stated that a penene ap-m Section b> were used m the equ:pment res;3onse anal)-

proach is sufficient because of the conservative nature of sis foi the $0ltV case with some modtheations. The tem-the combustion phenomena modeled by the ('t.ASIX-3 perature profdes were modified to exclude the few mittal code and the boundary conditions used in the penenc hydropen burns m which diffusion flames would exist and eqmpment survivahthly analysis.

the last hydrogen induced global burn. As a result, the moJtfied wetwell proble contams about 90 senal hydro-

.I he more important conservausms identified m the en burns. The calculated entical component of the pres-11( ()(i's genene analysis are paen below, sure transmitter exceeded its quahfication after 71 hydro-pen burns. Ilowever, the 11C00 indicated that the puure nansmuta is mM to sume W Wopen i cti is 11 io iit t id' rt n I m nt becauw of mnous conmaum in & anaW lease proble, is unhkely to occur af ter core recovery Also, on the basis of USTl tests, ddfusion flames on

.lhe staff acknowledges that conservatisms, as discussed the surface of the pool may cust aslow as 0 02 lb sec' earber, contained m these analyses could compensate for indicatmg hicah/ed combustion would not occur It the temperature exceedance over the quahfication of the is expected that the presence of ddfusion flames pressure transmitter. Nonetheless, the staff requested would probably be the dommant combustion mode' We llCOG tmovide additional data on the quahhcation possibly m combmauon with hicah/cd combusuon of the pressure transmitter. Ih letter dated April 5,1988 phenomena when the hydrogen flow rate is below (llGN-131-PL the 11C00 inihcated that durmg quahfi-the flame extmpuishment limit. Ihe significance of cation testmp the transmitter had operated without fail-these two dtflerent combusuon modes is the spatial ute at surface temperatures appioaching 350 *l for seu shif tmg of thermal loads, as such, a smgle piece el eral mmutes las compared to the quahfication of 320 i K equipment would not continually be exposed to hy-As part of its res;)onse, the llCOG re-evaluated the dropen burmnp resultmp in a lower temperature eqtupment response analysis of the pressure transmitter P

assummg contamment sprays were available. This re-The ('t.ASIN-3 wetwell temperature profile was sponse analysis indicateJ ihat the cquipment surface tem-used as the boundary conJnion for the equipment perature was about 'O ci less than the quahfication tem-response analysis although the most sensitive equip-perature of 320 'l.These results demonstrated the effect ment is hicated outside of the wetwell volume. The of sprays to cool the contamment environment, thus wetwell has the severest environment of the three mamtammg the function of ewential equipment. Itiver 23 N Ulll W l417

llend Station is the only plan t without cont unment

'I o vahdate and assess the heat uan& r methoJole a spravs but unit coolers are part til its design Whilc 5.pe-t omples ithic c-dinwnuonal peornctn I cahirimettf a

cific analyses bas e not been per f ormcJ to support quanto scmh6 was uscJ m sestral quarter-sude tests to suty,i fwation of the cotding efleit provided b) urlit i,tkilcf s s er-the caltirinletcI to ddlerchi hicathms anJ JJf trent tht r-sus sprays the 11( OG beheses that a reduction m nial cnvuonmer:ts A lli A llN(i b moJel o' the mmph s backgrounJ temperature would be aJequate to reJute calonmett r was constructcJ..nJ the calcuhited response the thermal loads on the pressure transnutter.

was compareJ to the measurcJ response to vahJate the a

methoJology. ~lhis chort is presented in the 11( ( H i's let-in summan. the 11((K mntenJs that f urther anJ us n ter datcJ August 20, lW ill(iN luf P.

3 unwarranted because, with the potenhal f or actne con-

"""""""nJ de te r nud Ad rnoM of tamrnent coolmr anJ the conservatisms mherent in the analyses, the pr' essure transmitter will function as de-mmp m

wm wngmu m tw mnJ-signed Jurmy recoserable derraded-core es er:ts that pno

1. P*'

I' * "

I " ' " '

press to metal water reaction for U percent of the claJ-kmmnwn J tNu w ht n Om gnm mWaidoMpy n dmp surroundmp the attne full repion Wah reparJ to thcsc anahses at hm hydropen flowrates, the stafI aprecs owd W Hanspenh ypnwm eMuauonN ayuew with the 11( ( Mi's posiuon that further efhirt m this area mnJuded ta ensure conwnauve speubcation of Om mn ary mnJitiont n not warranteJ. Moreover, the sta!! fmds it is more ap-propnate to deternune equipment surviv miht) on the ha-sis of data obt;uned f rom the OST) for diff uuon flames lhe MaH hnJs that the heat transtei rnethoJolors acahnp than on the haus of locahf ed combusuon condiuons.

un ddim e aimmten m pcmN m H(N ID3. pro-

@ an xcepthk hwJmm to F om @mphc d

eqmpment response anal) set Acwrdmply, each hcensee wit h a Mar k Ill wntamment mtends to use this methoQil-7.3 Diffusion Flame Thermal or3 as part of as Imai anavsn as reqmred n3 ihe h diopen 3

Eriviroririierit Metlioclology ruq iht sta!! aprees wnh SNi x rewmmendauon, that sulhuent detail of mput data shoulJ be provided by each in its Ittter dated Juls 30.1% ill(iN-103 L the ll( Oli nm to ensun as anahus o wnJuM in an appropn ate manner.

outlified the methoJology to be used by each hcensee us deter..ane the full scale plant speaht contamment ther-mal ennronments from the QS l1 data The full-scale en-7.4 Spray Availability vironmental conditions would be useJ as bounJan conde nons m the iU ATING 6 computer code to anaiy/c the in the prehmmarY esaluauons of hydropen irruter sys-response of containmt nt eqmpment during postulated tems (e p., see Supplement 3 to the Grand Gulf Si l(,

ddfusne combusuon events. As a result of these analvses.

NUld G-0531, July 1%2 L the staff allowed eredit for the survivabihty of essential equipment would be deter.

operation of mntamment sprays m the analyses of the nuned.

consequences of hydropen combusuon durmg degraded-core accidents on th( basis of several consiJeranons that I rom the production test senes conJucted f or cash Mark

" * " "" " " *

• W "} " P" "E

ill contamment, the test that produces the niost hminnp Gw p omnq cuduanons of ignun gMenn tht' MaH an focu on Om a ent gwn environment at the correspondmp (to full scale) equip-SOR\\ transient anJ the drywell pipe break. y for Ow ment location is used. Thermal profiles are constructed lhese acco by spatial mappmp of the test f acihty data. Specihc plant nt wquenm @ not nmmnh anph low of the con-p'rofiles are developed from average temperatures for nunment pay funcuon oNwy pumps M pumps time intervals of maumum hydrogen flow anJ low con _

may 'opna e, but Om m I ingcuon path may be m-stant hydrogen flow from the pnxJuction tests.This allows tmuW m MM i urther, at the time of the preliminary determmation of the plume loca..ans and the effects of uahonkt muaH tone d Om W R emyng pnw bhsckapes and sparpers.

mepm hnes O M was to fom on mntamment m-tegnty rather than adequacy of core coohnp at an earl er l'ull-scale vehicities arc tomputeJ from the quarter scale measured vehvines us.ng I roude scahng; test tempera-Smce the prehmmary evaluations w cre conducted, adde tures are used directly (scahng is 1:I). The com cetive heat tional mformation has been developed that raises ques-transfer and radiative heat fluxes are computed usmp the tions regardmp the vahday of assumpnons concerning scaled vehicines and temperatures. Smcc this approach avadabihty of the RlH( pumps m the contamment spray estabhshes an environmental map, the hea' transfer mode. In contrast to the earner focus on the SORV tran-modes that should be considereJ ar e dependent on the lo-sient and dryw ch pipe break, recent nsk analyws indicates canon of the affected equipment.

that stanon Nackout (SH())is a s:gnificant contributor 'o N U RI:G - 1417 3

hydrogen generation events. For the SHO, the loss of re-staff concludes that the potential for localized accumula-actor makeup is tied to the loss of pumps, including 111111 tion of significant concentrations of hydrogen is unlikely.

pumps, in either the 11C1 or containment spray mode.

Thus for the Silo sequence, the Rillt spray function car.-

not be reasonably assumed to be available until ac power 8

CONCLUSION is restored, in addition, the earlier emphasis tn the EPGs on containment integrity versus core cooling for contain.

On the basis of its evaluation, the staff finds the llCOG ment spray operation has been reversed. In llevision 4 to the llWit IIPGs (htarch 1987), the sequence of steps has topical report, *-Generic Hydrogen Control Information for 11WR-6 hiark 111 Containments' (IIGN-ll2-NP),

been modified. Use of RllR pumps in the containment spray mode, irrespective of adequate core cooling, is now dated February 23,19S7, provides an acceptable basis for techru, cal resolution of the hiark lli contamment de-directed as the last step, to control pressure rather than before the decision to vent.

nded core hydrogen controlissue,linch licensee should

. vide a plant 4pecific final analysis, as required by a CFR 50.44(c)(3)(vii)(!!), which will address the cle-l.or the above reason, the staff concludes the 11%,R ments specified in 10 CFR 50.44(c)(3)(vi)(ll). The hiark Ill owners should evaluate the containment and es-11C00 topical report, or portions thereof, may be refer-sential equipment response to hydrogen generation enced, where appropriate, taking into conside' ration the events assuming contamment sprays are unavailable, con-staff's recommendations as stated in this report. The sistent with SBO assumptions and the liPGs. Spray oper-plant specific analysis will use test data described in the ability can be modeled, but margins should be established topical tcport to confirm that the equipment necessary to for a variety of possible plant conditions. Similarly, as-establish and maintain safe shutdown and to maintain sumptions regarding availability of containment coolers containment integrity will be capable of performing its should be consistent with the basic premise of the Silo functions during and after exposure to the environmental accident sequence, conditions resulting from hydrogen generation in all cred-ible severe accident scenarios.

7.5 Pressure Effects One consideration in the staff's assessment of the ade-quacy of the hydrogen ignition system (lIIS) was whether in its letter of August 7,1987 (llGN-llS P) the iICOG an alternate power supply was appropriate. An important mdicated equipment located m, side containment is quali*

factor in this decision process is the level of risk associated fied to a pressure loading of at least 30 psig applied exter-with an SBO event leading to core damage. Recent risk nally.The CLASIX-3 predictions produced the most sc-studies reported in NUR110-ll50 have shown that the vere pressure rise of about 23 psig in the htark 111 overall core melt frequency for one plant with a hlark 111 containment. The staff concludes that pressure is not a containment (Grand Gulf Nuclear Station)is very low, concern, pending confirmation by cach licensee of the 30 i.e., ll!-6/ year. Ilowever, a potential vulnerability for psig capability. When the hydrogen ignition system is hiark Ill containments involves station blackout, during iunctioning, various containment subvolumes will be ran' which the igniters would be inoperable.This condition ap-domly affected by hydrogen burning: however, a large pears to dominate the residual risk from severc accident pressure spike is not expected to occur.

in the plants with hiark 111 containments. Under Silo conditions, a detonable mixture of hydrogen could de-velop, which could be ignited when power was restored 7.6 Detonat,onS and result in loss of containment integrity. On the basis of i

a separate evaluation of this possibility under the NRC The !! COG believes that a detonation is not a credib!c staffs containment performance improvement (CPI) phenomenon in the hiark Ill containment because (1) no program, the staff has recommended that the vulnerabil-rich hydrogen concentrations will accumulate inside con-ity of the hydrogen igniters to interruption of power be tainment since the distributed igniters will initiate com-evaluated further on a plant specific basis as part of the bustion as the mixture reaches the lower flammability individual plant examination (IPl!)of the plants with the limit and effective mixing will occur and (2) there are no h1 ark Ul containments.The staff has requested that the regions of the containment with sufficient geometrical licensees consider this issue as part of the IPli stipulated confinement to allow for the flame acceleration necessary in Generic Letter 88-20, Supplement 3.

to yield a transition to detonation.

With the above caveat, the staff finds that there is reason-The staff agrees with the llCOG position. As confirmed able assurance that the 111S installed in the plants with by the quarter scale test results, the atmospheric condi-hlark 111 containments will act to control the burning of tions inside the test facility were well mixed and burning hydrogen so that there is adequate protection against con-at low hydrogen concentrations was prevalent. Thus, the tainment failure.

25 NURl!G-1417

The staff concludes that the following key elements

($) confirm survivability of essential equipment should be addressed in each licensee's plant specific final analysis to resolve the degraded-cere hydrogen control is-identify plant specific essential equipment e

sue:

define thermal environment from quarter.

e (1) further ecaluate the vulnerability of the hydrogen ip.

scale testing niters to interruption of power perform equipment response analysis e

(2) confirrn npplicability to the generic effort confirm that redundancy exists for that equip-e ment affected by secondary burning and (3) provide quarter scale plant specific praluction test-drywell inverted diffusion flames ing results confirm pressure capability of equipment e

(4) confirm primary containtnent structural survivability through (6) state licensce's position regarding the proposed IICOG cmcrgerwy procedures for combustible gas e

quarter scale testm8 control pressure capacity analyses for drywell and con-e tainment, for example, confirm previous plant-(7) provide overall conclusions relating to conformance specific analyses of the hydrogen rule NURiiG-1417 26

APPENDIX A GENERIC llYDROGEN IGNITION SYSTEM TECllNICAL SPECIFICATIONS Generally, technical specifications (TS) of a particular moral of the surveillance requirement to determine the system consist of two distinct sections: surveillance re-location of inoperable igniters after the requisite number quirements to ensure system operability and limiting con-of failed igniters has been attained. In addition to these dition(s) for operation (LCO) to define the allovrable op-two changes, the 11C00 proposed to change the current crability range in conjunction with various plant actions acticn statement that allows 30 days to restore the igniter when needed. Each of the four plants with Mark 111 con-subsystem to operable status to 60 days because the tainments have similar TS for the hydrogen ignition sys-events during which the lilS is required to be operable tem (Ills)/Ihe following discussion is focused on the pro-are less probable than design basis accidents. Ilowever, posed generic 111S TS and their deviation from the the Nuclear llegulatory Commission (NRC) staff finds current TS.

this proposal lacks sound engineering judgment since it does not provide the rationale to support a 60-day inter-Currently, TS en igniter systems in plants with Mark Ill val.Thus, the 30-day interval is appropriate and should be containments prescribe two types of surveillance practicc.

maintained.

'the first type of surveillance is conducted at 184 day in.

tervals. All the igni;ct assemblies are energi/cd and cut-With regard to the llCOG's justification for the first sig-M Pi'iP E

ally' the !! COG eited rent / voltage measurements are performed and compared with similar measurements taken previously. lf more than the results regarding the qu rter scale testing facility three igniter assemblics on either cubsystem are deter-(QSlF) testing of a particular > :opmg test m which about mined to be inoperable, there is an inci case of the surveil-40 percent of -

r subsysti m was moperable in con-lance frequency to a 92 dayinterval. A second part of this junction witt subsyste n not functioning. Ilow-first surveillance requirement is the verification that in-cycr,the stan has ded that the 1-1C00 has not pro-y ded sufficient jusm uition to iclax the 13 to such a operable igniters are not adjacent to each other if more degree.The staff trade its determ nation because of such than one igniter on each subsystem is determined to be inherent uncertainties as the cytt, potation of the quar-inoperable. 'ihe basic for this requirement is the staff's view regarding potential hydrogen pocketing in enclosed ter scale results to full scale, various injection rates, ac-tuntion of different safety relief valves, and different pos-areas.

sibilities where 40 percent of the inoperable igniters could be located *lherefore, the allowable value for inop-

'the second type of surveillance is conducted at 18 month etabh igniters should be as low as practical and a 10 per-intervals to verify a surface temperature of at least 1700 unt salue appears to be a reasonable limit,

'l' for each accessible igniter and to verify by measure-ment sufficient current / voltage to develop 1700 'F sur-1hc second significant proposed change is to remove the face temperature for those igniter assemblics in inaccessi-surveillance that ensures that inoperable igniters are not ble areas. Accordingly, the bases section of the TS adjacent. This surveillance ensures at least one operable indicates that inaccessible nreas are defined as areas that igniter in each enclosed area and coverage of the have high radiation levels during the entire refueling out-azi nuthal-positioned igniters in the open regions. Essen-age; such enclosures include the heat exchanger, filter tially, the lilS TS are intended to prevent buildup of hy-demineralizer, and the pump room for the reactor water drogen in subvolumes of the Mark !!! containment, cleanup (ItWCU) system, thereby precluding the occurrence of large volume burns, liighlights of the HCOG's justification follow.

The current LCO allows no more than 10 percent of the e

The llCOG cvaluated the potential flow paths that igniter assemblics inoperable per subsystem. And if one subsystem is inoperable, the action statement requires could transport hydrogen in or near enclosed regions restoration to operabic status or to the required opem.

of the containment and determined that no poten-tional condition within 30 days (similar to the hydrogen tial hydrogen source exists, it is expected that ig-recombiner TS).

niters in open areas will function to preclude local hydrogen pocketing.

11yletterdated April 16,1986(IlGN-070) thc Hydropen e

Observations of the quarter se:de tests indicated Control Owners Group (11C00) proposed to revise se-that the released hydrogen will tend to mix with the lected portions of the existing plant specific TS as out-surrounding atmosphere and thus reduce the poten-lined above. Principally, there are two significant pro-tial of locally high hydrogen concentrations.

posed changes:(1) an increase in the number of allowable inoperable igniters per subsystem to about 40 percent, as The likelihood is low for inoperable igniters being e

compared to the current value of 10 percent, and (2) re-kicated in such a fashion as to create a large Appendix A 27 NUREG-1417

containment subvolume that would be without ip-assurance that the proposed 'lTi without the adjacent ig-niter coverage. Igniters would tend to fail in a ran-niter provision would not adversely affect the effective-dom manner, ness of the igniter system.

Currently, whenever at least one igniter is inoper-The staff finds the generic ills TS as documented in the able in each subsystem, containment entry is nor-IICOG letter dated April 16,1986, to be acceptable con-mally required to find the location of the failed ip-tingent on the following changes: the 40-percent value of niter.This would subject plant personnel to various a!!owable inoperabk igniters should be 10 percent in the occupational safety hazards such as radiation expo-a;3propriate locations of the text and the 60-day interval sures and the risks associated with the construction to restore a subsystem in the action statement should be of scaffolding.

'O days. IIach Mark 111 qwner that intends to adopt the neric lilS TS must confirm that the 11C00 assump-On the basis of these considerations provided by ths

. ans used in the development of the TS are valid for their ilCOG, the staf f has determined that there is reasonable plant specific configuration.

NURi!O-1417 28 Apl.endix A

APPENDIX 11 MARK 111 COMilUSTillLE GAS CONTROL EMERGENCY PROCEDURE GUIDEl.INE As part of the peneric program. the flydropen Control the ill Oli aJJtesscJ staf f conserns or pionJeJ suffo Owners Group dlCOlin has devehiped the combustible cient justificanon lot its povuon. lhese issues are dis-gas control emerpenes procedure puiJehne il PG) for cusseJ behm.

plants with h1 ark lit contamments The latest version of the ymdchne, w nh supportmp appenJices, was sent to the As one of the miual steps m the 1 PM the operator is m-Nuclear Reputator) ('ommission (NRC) staff by letter structcJ to sent the suppresuon thamber or drvwell dated July % IONA OlGN -12214 Ttus procedure m.

whenever caber of the respectne regions reaches the cludes operator action for the hydropen igniter system as mmunum detectab!c hydrogen concentratam (0.59 i, Well as other combustible pas etultrol syMems dcNipned iri pfovided the ollsite raJuucuvity release rate is expected the h1 ark 111 containment sut h as hydrogen recombmers to remam below the ofisne release rate hmitme conJiaon and the drywell naunt system in aJJiuon, the proposed for operation tl ( ( h 11 should be noted that dus step is prot cJur e provides guidance ior spray actuauon and con-umdar to the llWR I PC is for Ntark I and Ntark 11 combus-tamment venung. Tlus of fert is to supplement the overall tible gas control The staf f previously wmmented that ilodmp-Water React.o ()wners Group's (llWROG's) s entmp may not be necessan solely upon hydropen con.

I P(i propram.

centration abuse the nunmium detectab!c lesel anJ be-low flammabthis lesels; the use of recominners is valu-I apure 11.1 at the end of dus appendix lughhphts the op.

abk and shoulJ be unh/cJ where appnspn.ac.

crator actions wah r craid to hydrogen control m an emer-pency situation These act ons are determmed by a hydn, in responw, Ow I O K i Mated that smcc dissoh ed hydro-pen mncentration m the contamment and drvuell as pen n pr ewnt in du' r eacuu coolant gMem dunng normal mJicated by hydrogen monnors and or analyzers'that ob-opmuon and the 1 PGs are base ' on a symptomatic ap-tain gas samples from the contamment and drywell. The proxh a n the inknt of ths Mep i i winedy a hydropen sigmlicant inyper knut useJ m the i PG is'when the probkm dunny normal opennion and wahm the con-drvwell or contamment hydropen level reaches a concen.

sumnu, of tnhnical spaikanon lumts. The i R 'OG be-tration at u hich a global deflaerahon t oulJ threaten con-he thew a sukwnt puiJance to preclude dus acuon tainment or drvw ell internh IIom overpressun/ation, re-Inun bang unpkmenwd dmint a penume emtrpency ferted to as th'e hydrogen deflagranon overpressure hma canon. Ab o, the I WW i mmnuned us nnW3 the re-nene irk 111 containment prosedure at a later date, if (llDOI ) At dus hnut or w hen the contamment hydrogen concentranon cannot be determmed to be below the nnem to ; <unsiMent w nh the llWROG's approved llDOI and a cannot be determined that the igniters have mniMubk pas mnuol procedure ihe stall fmds the been contmuously operatmp, the hydropen igniuon sys.

subjnt procedun and the llCO(i approach acceptable.

tem (Ills)should not be used. lhe containment I(dol is Asone dw lau Meps u mnuol hy@open anurnulanon a cune of hydrogen mncentranon versus contamment pressure, whereas the dnwell llDOl is a smgle value dunny a pmpwmp wonemnp muanon, mntainment representmp a peak hydrogen concentration. The con-wnung g @nk \\ enung Ow mnuunment umpnuse tamment 11D01 n more knuting than the dnwc!I o dw oHse radmacunty wkaw noe wouM onh N mn-nd to NMon and manmun Ow mntamment hydro-M I IDul..

pen concentration below excessive hmits. Contamment fatture may follow if a huge deflagrauon were to occur.

The 1ilS, the hydrogen t ecombmers, and the drvw cll mn-Ventme the contamment may be the only mechanism that ing system are the key hydrogen mitigatmp systems. As m-wmami to present an uncontrolled and unpredictable dicated m I igure 11.1, these systems are acuvated at ap-breach of the containment. The controlled release of ra-propnate inpper pomis to deal with a progressmg dioacuvity to the environment is preferable to contam-hydrogen buildup. With the adJiuon of an mdependent ment fmiute, whereb) adequate core cooling also might power source to the lith, a is anticipated that for most se be lost and radio-acunn released wah no control. Tlus vere accident anyl derraJed-core situanons the resulting concept of venting is simiiar to the emergenes proccJures larpe amounts of hydropen can be accommodated.

for pressure control.

As part of the subject letter, llCOli haJ aJJressed staff Wah regard to this issue. the 11C0G pronded additional comments that were docussed dunny the ill 00 anJ mformahon respondmp to an N RC stall comment deahng N RU meetmg of October 22,14Sn. In it's latest version of with the hmited use of the drywell miung systems. The the Ntark Ill combustible gas control 1 PG tRevision 3L staff news contamment ventmp as a last resort to deal Appenda 11 4

NURI T b l417

with extraordinary conditions.The use of the drywell hy.

that the 111S would remain operational above the llDOL drogen mixing system may delay containment venting by if it can be determined that the igniters have been con-diluting the containment volume (at a higher concentra-tinuously operating. The addition of an independent tion of hydrogen) with the drywell volume (at a lower con-power supply to the 1IIS would further enhance the reli-centration of hydrogen).The llCOG cited various factors ability of the system. Consequently, the added reliability to demonstrate the drywell mixing systern is not beneficial would reduce the potential for containment venting to for hydrogen control inside the containment volume, control hydrogen inside containment.Therefore, the staff These factors include: dilution effects are marginal be.

agrees with IICOG that the inclusion of the drywell mix-cause the containment is significantly larger than the ing system would not provide significant benefits (in de.

drywell; the mixing system would re initiate a loss-of-laying venting) as compared to its disadvantages.

coolant accident signal and potentially interfere with event recovery; and implementing a modified procedure may induce conflicting direction, in addition, the design Overall, the staff finds the proposed Mark Ill contain-intent of the mixing system is to dcal with hydrogen in the ment !!PG (Revision 3)is based on sound technical judg-

drywell, ment and acceptable. Accordingly, each licensec with a Mark 111 containment should address its combustible gas The staff believes some of 11C00 concerns are valid, in control emergency procedure in the plant specific final addition, llCOG had modified its procedures to ensure analysis.

l l

l NURiiG-1417 30 Appendix B

C owbrit0W DPY waf t R L(Ytt E IAF DN CANNOf.a$(N D.L l( k u iN( D.

ACTVAT[ IGNIT($ $

D PC AND De M3 < HDDL 2 htaCHt$

PC Ch DW H WIN DCTICIADLL LIV $k ect/l*

"I,1,I,'gII Yt1 vtNT AND PURCt Dttow /

PC LCO NO Dw PC PC H2 RE ACN($ WIN CDNC FOR ow g g we,,

accousiNta o.rtaation

.on CritC1ADLt LIWrf win D(T(Cla$ll L(VEL P

" I *C PC H 4 LOWIII PC AND t$

CU" C NC FOR De H2 IE "8

'0"op't n at DE I C "' I'0"

< HDOL co#/2*

OPteatt ob8/3' N0 Hg

{

{

iCNITLk$

c D* "2 'E^CHII l'C AND ygg PC Hg g

Y(S CONC FDA L)Pwak0 DW Mg g gppg gggggg,,ggg (LAW ( PADP ACalpDN

< HD Ho k

epy Pr synt PC H RCact405 Wan CONC FOR gggggg g

htCOWOINCR OPC $ Atl0N Hg towtti cowe's on" otrt attatiow N,vfS optmait Deywrt PC H 2 Hg WiniNO sv$ltu.

, gg/g Ho u

CONilNULD at O

Figure 11,1 Operator actions for primary containment hydrogen control (See definitions at end of figure.)

Appendix 11 31 NURliG-1417

CONostsDN ACitON CD "2

St CURT ><

StACNES NDOL

, g g gg g,,,,IS PetNvRE vts e,tRatt PC SPR AYS,

> Cs ACC Ptawiftsho N0 scNett PS CONilNUOV$(T twtmoittp seNet vtt tacttDINO LOWElt N IOR OtFLAGRAftoM I

$! CURT N, ICNittPS AND Winiho STsitwS NO i

If PC H2 " E "N 3

• 0"/**

> HDOL VINI 'NU PV"0E I C I!

PC ppt URt OPERatt PC Sprats, q g IRRCSP OF ACC 1

PC H2 " '4004

• oR/6' IN DCT LCvtt I'

H2 DW H 2 WIXtNG SYSitw$

,,gn, RPV PRtssvRt C PCPL Ho k$

OPERAtt H 2 #CNiit R $

g l

PC H2 < WAk CONC rom RECOWSINCR OPERAT60H

• AND-PC H 2 < LowCSt CohC OPERaft NR "E00W8'"E"8 F Ost OtFLACR AtlQN

ANO

PC H2* 1%

Figure 11.1 (Continued)

NUlWO-1417 32 Appendix 11

1. 0 'D" COND1110N DR/t ND[vect RELC A$t R Att RE ACHt$ LCD vg DW H2 > HDOL

• LND~

4GNITE R $

H. IGNtttRS NOT CPERattNC SCCURt H2

• CONilNuoustY SINC E Dw H2 LRCttDED Lowt$f CONC FOR DtrLACRAtt0H PC H. > HDOL DR/3

.aND.

Ct!NuCU!LYs' INCL SCCURE H; WiklHC SY$ttws,

( I I

RtCowSINLR$ AND icN'itRS PC H2 t*CttDtD LowtSt CONC FOR (;trLACRATICH VEN1 AND PURCt PC PC H l$

SECURt PC

,, 2,3, vtN1 AND PVRGE OR/6 PC H RCACHts ttCVRC DW H 2

2 HDOL WixtNG SYSTtw$

Over ride OR DW

- Drywell Primary Containment l'C l' CPL - l'rimary Containment l'ressure limit CSil'L - Mark til Containment Spray initiation l'ressure limit liigure 11,1 (Continued)

Appendix 11 33 NURiiG-1417

Al'I'ENDIX C lilllLIOGRAl'IlY llydrogen Control Owners Group (11C00), ilGN-003,

-, ilGN-034, letter from S. !!. Ilobbs (llCOG) to R.

letter from J. D. Richardson (llCOG) to !!. R. Denton llernero (NRC), *Model for Ilydrogen Production l

(NitC),

• Report on llydrogen Control Accident Scenar-liquivalent 10 75% MWR." May 17,1985, ios, Ilydrogen Ocncration Rates arul liquipment Re-quirements," April 8,1982.

-, llGN-051, letter from S.11. Ilobbs (11C00) to R.

llernero (NRC), " Availability of Containment Spray Sys-

-,IIGN-006,letterfromJ.D. Richardson (llCOG)to tem," July 26,1985.

11. R. Denton (NRC)," Report on flydrogen Control Ac.

cadent Scenarios, liydro

-, IIGN-052, letter from S.11. Ilobbs (llCOG) to R.

liquipment Requirements, gen Ocneration Rates and Bernero (NRC), "llydrogen Release Time llistories,"

Rev.1, September 9,1982.

August 1,1985.

-, HON-009 P letter from J.D. Richardson (11C00)

-, liGN-055, letter from S.11. Ilobbs (IICOG) to R.

to i1. Denton (NRC). "CIASIX-3 Report," March 18, llernero (NRC)," Evaluation of Silo and ATWS Contri-1983.

butions to llydrogen Generation livents." September 27, 1985.

-, llGN-Oll NP, letter from J. D. Richardson (IICOG) to R. W. Ilouston (NRC), " Responses to NRC

-, IIGN-070, letter from J. R.1.angley (IICOG) to R.

Requests for Additional Information," May 11,1983.

llernero (NRC). " Mark lit flydrogen Ignition System Technical Specifications," April 16,1986.

-,llGN-012 NP, letter from S. II.llobbs(llCOG)tt R. Ilornero (NRC),"l/4 Scale llydrogen iest Program,,?-, llGN-072, letter f rom S.11. Ilobbs (llCOG) to R.

August 12,1983, llernero (NRC),"livent Scenarios Considered for livalu-ation of Drywell Response to Degraded Core Accidents,"

-,llGN-014 NP, letter from S. II. llobbs (llCOG)to

11. Denton (NRC), " Final 1/20th Scale Test Report,"

-,IlGN-073, letter from J, R. langley (llCOG)to R.

February 9,1984.

licrnero (NRC), " Justification for Manual A-tuation of Mark 111 Ilydrogen Ignition Systems," March 5,1986.

-,IIGN-017 NP, letter from S.II.llobbs(IICOG)to

11. Denton (NRC), " Mark 111 llydrogen Control Owners

-, llGN-084, let ter from J. R. langley (llCOG) to R.

Group Final Whiteshell Ignition Test Report," June 7, liernero (NRC), " Generic liquipment Survivability List,"

1984.

May 16,1986.

-,llGN-018, letter from S.11. Ilobbs (IICOG) to 11.

-,llGN-085,1etter from J. R.12ngley(11 COG)to R.

Denton (NRC) "l/4-Scalc Test Program to Define Ther.

)', Final Report Assessmg Adeguacy of II'"t Sink Modch.ng m the 1/4 Scale'l est Facility, May 5 mal Environment " July 6,1984.

39g6

-, IIGN-020, letter from S. I1. Ilobbs (llCOG) to

-,1IGN-089, letter from J. R. langley (HCOG)to R.

11. R. Denton (NRC),

• Transmittal of IlWR Cc.re llernero (NRC),"BWR Core IIcatup Code Responses,"

lleatup Code Manual," September 5,1984.

June 9,1986.

-,IIGN-027 NP, letter from S.11.Ilobbs(11 COG)to

-,1IGN-091, letter from J. R. langley (IICOG) to R.

R. Bernero (NRC), *l/4 Scale Test Facility 3D-Complex llernero (NRC), " Criteria for Existence of Inverted Dif-Calorimeter," February 13,1985.

fusion Flames in the Drywell," June 25,1986.

-, ilGN-031, letter from S.11.11obbs (11C00) to R.

-,liGN-092 P, letter from J. R. langley (11C00) to Bernero (NRC), *llydrogen Release llistories and Test R. Bernero (NRC), " Report of CLASIX-3 Generic Analyses and Validation of CLASIX-3 Against 1/4 Scale Matrix for 1/4 Scale *Iest Program," March 13,1985.

Test Facility Data," June 10,1986.

-,1IGN-032, letter from S. I1. Ilobbs (11C00) to R.

-, llGN-096-P, letter from J. R. langley (11C00) to Bernero (NRC),

• Submittal of Information on BWR R. Bernero (NRC), "BWR Core lleatup Code Report,"

Core Heatup Code," April 16,1985.

July 30,1986.

NUREG-1417 34 Appendix C

-, ilGN-098-P,1ctter from J. R.12ngley (HCOG) to Potential Impact of SBO Events on Operation and Per-R. llernero (NRC), *Scopin gTest Report," J uly 18,1986.

formance of the liydrogen Ignition System," January 8, 1987.

-,llGN-099 P, letter from J. R. langley (llCOG) to R. llernero (NRC), "Transmittai of llandouts from June

-, HON-115 NP, letter from J. R. langley (llCOG) 19,1986 ilCOG NRC hiceting" July 11,1986, to R. llernero (NRC), "l/4 Scale Test Facility Final De-sign Report," February 10,1987.

-,IIGN-100, letter from J. R. langley (llCOG) to R.

llernero (NRC), " Nevada Test Site Data Evaluation,"

-, HON-118.P. letter from J. R. langley (llCOG) to July 31,1986.

NRC, " Generic Equipment Survivability Analysis,"

August 7,1987.

-,IIGN-101,letterfromJ.R. langley (llCOG) toll.

Denton (NRC)," River llend Station Unit Coolers," July

-, HGN-119, letter from J. R. langley (HCOG) to 30,1986.

NRC, " Final Inverted Diffusion Flame Report " J une 10, 10.

-, HON-103, letter from J. R. Langley (HCOG) to R.

Ilernero (NRC), " Diffusive Combustion Thermal Erni.

-,IlGF 121 P,letterfromlangley(HCOG)toNRC, ronment Methodology Definition Report,"J uly 30,1986.

• Report of Ilymen C,mbustion Experiments in a 1/4 Scale Model of a hbrk Ili Nuclear Reactor Contain-

-,llGN-104 P, letter from J. R. langley (HCOG) to ment," July 22,1987.

R. Bernero (NRC),

• Evaluation of Emergency Procedure Guidelines Operator Actions Agamst 11C00 Assump-

-,llGN.122 P. letter from J. R.12nB eY (llCOG) to l

NRC, " Revision 3 to Mark 111 Combustible Gas Control tions for Analysis of a Hydrogen Generation Event,-

Emergency Procedure Guideline," J uly 8,1988.

. August 18,1986.

-,llGN-105 P, letter from J. R.12ngley(ilCOG)to

- ' IlGN-123, letter, from J. R. langley (HCOG) to NRC,, Response to NRC Questions on Station illackout R. Bernero (NRC)," Diffusive Combustion lleat Transfer and ATWS Sequences,,, September 9,1987.

Methodology Validation for Equipment Survivability in Mark Ill Containments," August 29,1986.

-, HGN-128. lettei from J. R. langley (llCOG) to

-,11GN-106-P, letter irom J. R. langley (HCOG) to

.)#["h8 g,

e 98 R. Bernero (NRC), " Supplemental Information on Sec-ondary Burning," September 29,1986.

-, llGN-129 P, !ctter, J. R. Langley (HCOG) to NRC," Revision 4 EPGs vs llCOG Scenarios," April 8,

-, llGN-109 P, letter from J. R. langley (HCOG) to 1988.

R. Bernero (NRC), "CLASIX-3 Generic Sensitivity Analyses," December 9,1986-

-, llGN-131 P letter, J. R. langley (HCOG) to NRC," Supplemental Discussion of PressureTransmitter

-, IIGN-110-P, letter from J. R. Langley (HCOG) to Survivability at Low Hydrogen Release Rates," April 5, R. Ilernero (NRC), " Combustible Gas Control Emer-

19gg, gency Procedure Guideline and Supporting Appendices,"

December 1,1986.

-, HGN-132, letter, J. R. langley (llCOG) to NRC,

" Final HCOG Hydrogen Release llistories," April 4,

-,llGN-111 P, letter from J. R. Iangley (HCOG) to 1988.

NRC, *CLASIX-3 Summary Report," December 15, 1987.

Institute of Electrical and Electronics Engineers, Std 323-1974, "lEEE Standard for Qualifying Class 1E

-, HGN 1 $NP, letter from J. R. langley (HCOG)

Equipment for Nuclear Power Generating Stations,"

to R. Ilerv 4RC), " Generic ilydrogen Controllnfor-New York City, New York, mation for o,v'R-6 Mark III Containments," February 23,1987.

International Technical Services (ITS), ITS/ LWR /BNL 85-1," Validity of the Use of a Temperature Cutoff for

-, HGN-113, letter from J. R. langley (IICOG) to R.

Zircalov Oxidation," by 11. Komoriya and P. Abramson, llernero(NRC)," Comparison of CLASIX-3 Predictions Septen.ber 1985.

to Nevada Test Site Data," January 8,1987.

-, letter from II. Komoriya (ITS) to L. Lois (NRC),

-, HGN-114, letter from J. R. langley (HCOG) to R.

" Review of Oxidation Modeling in BWR Core IIcatup

!!crnero (NRC). "Respom to isC Concerns Regarding Code," October 30,1987.

Appendix C 35 NUREG-1417

I c

j Mississippi INwer and Light Company, letter from 1., F.

Units 1 and 2." Supplement 3, J uly 1982, and Supplement Dale (hilssissippi Power and Lie;ht Company) to 11. Den-5, August 1984, ton (NitC), "li, Igniter 11nvironmental Qualifications Test Results," February 14,1983.

-, NUlti!G-0853," Safety f! valuation lleport Itclated to the Operation of Clinton Power Station, Unit No.1,"

Office of Federal Itegister, Title 10, Codcof red-Supplement 6, July 1986.

rral Regidations, _U.S.

Government Printing Office, Washington D.C.

-, NURl!G-0887,"SafetyI! valuation Report Related to the Operation of Perry Nuclear Power Plant, Units 1 Sandia National Laboratory, letters from S. !!. Dingman and 2," Supplement 6, April 1985.

- to A. Notafrancesco(NilC), June 26, August 10, Septem, ber 3, September 4, and December 23 (two let ters),1987.

-,NURl!G-0979,"Safetylivaluation Report Itclated to the Final Design Approval of the GliSSAR 11llWR/6 Science Applicat. ion Inc., SAIC/87/3114. "Ilydrogen Nuclear Island Design," Supplemer.t No. 2, November Generating I! vents for lloiting Water Reactors with hiark

1934, 111 Containments," by Ateft et al., San Diego, California.

April 15,1988.

-, NURl!G-0989, " Safety Evaluation Iteport Itclated t

e' n of River llend Station " Supplement 4, U. S. Nuclear Regulatory Commission, Generic Letter r S5' 88-20, Supplement 3,

Subject:

" Compilation of Contain-i ment Performance improvement Program and Fo,rward.

-, NURf!G-1150," Severe Accident Risks: an Assess-ing Insights for Use in the Individua1 I lant laammation ment for Five U.S. Nuclear Power Plants," Second Draft, for Severe Accident Vulnerabilitics,, July 6,1990.

June 1989,

)I hl G Ju e 4,1985.

-, NURiiG/CR-1659, Vol. 4 of 4. "Itcactor Safety Study hiethodology Applications Program (RSSh1AP),"

g

-, letter from R. W. Ilouston (NitC) to J.11.12mgley Grand Gulf Power Plant Unit 1. November 1981.

(llCOG), February 21,1986.

-, NURf!G/CR-4866, "An Asser.acnt of Ilydrogen -

-,NURl!G-00ll," Safety !! valuation Report Related Ocneration for the PilF Severt i uct Damage Scoping 1

to the Operation of Sequoyah Nuclear Plant. Units 1 and an'J 1-1 Tests," A. W. Cronenberg et al., liG&G, April 2," Supplement 6, December 1982, 1937-

-, NURl!G-0588,"Intcrim Staff Position on linviron.

-, NUREG/CR-5079, " Experimental Results' Per-i mental Qualification of Safety Related Electrical I! quip.

taining to the Performance of Thermal Igniters," Sandia ment," Rev I, July 1981, National Iaboratory, hi. Carmel, Sandia, October 1989, i

-, NURI!G-0831," Safety I! valuation Report Related Vinjamuri, K., et al., *Severc Fuct DamageTest 1-4 Data to the Operation of the Grand Gulf Nuclear Station, Report," EG&G, Idaho Falls, Idaho, September 1987, i

7 1

L NUREG-1417 -

36 Appendix C 1

t APPENDIX D ACRONYM LIST ADS automatic depressurization system MWR metal water reaction IlWR boiling-water reactor NRC Nuclear Regulatory Commission BWRCIIUC boiling water-reactor core heatup code PRA probabilistic risk assessment CFR Code of federalRegulations QSTF quarter scale test facility EPO emergency procedure guideline RCIC reactor core isolation cooling IICOG Hydrogen Control Owners Group RPV reactor pressure vessel (Mark 111 Containment)

RSSMAP reactor safety study methodology appli-HCU hydraulic control unit cations program HDOL hydrogen deflagration overpressure RWCU reactor water cleanup limit SBLOCA small-break loss-of coolant accident HGE hydrogen generation event SBO station blackout IIIS hydrogen ignition system HPCS high pressure core spray Sanda National Moratoy SORV stuck open relief valve IPE independent plant examination SRV safety relief valve LCO limiting condition of operation LOCA loss of coolant accident TAF top of active fuel LOSP loss of offsite power TS technical specifications l

Appendix D 37 NUREG-1417

i U.S. NUCLE AR fitGUL ATOAV CDeAMISSION

1. MtPORT Nuwet R f1CC posis $36 L"Zl12,". "Mldl'T' * ""'

!!.flLin.

mi. nu BIBLIOGRAPHIC DATA SHEET

<se,

,,ve

., rw --->

NUREG-1417

2. TITLE AND $USTITLE Safety Evaluation Report Related to Hydrogen Control 2

o^tt atroaT rui5"to Owners Group Assessment of Mark III Containment g

..a i e, October 1990

4. F IN OR GRANT N'JuSER 61YPE OF RWRT

6. AUTHOR ($l C. Y. Li Regulatory

7. Pt R 100 COV E R E D isseraw.me Deress S. 9 F A NG NIZ AT 60N - N AM L AND ADDR($5 fi8 *ac. amvan Devema. Ortwe er Asesea, v.s amurmwmapweeser, e-

-. mes mannip._ n_ de sowiennw. pewan, Division of Systems Technology Office of Nuclear Reactor Regulation U.S. Nuclear Regulatory Commission Washington, DC 20555

9. SPONSORING ORG ANil ATION. N AM L AND ADOR E SS ist tvac. erse *1 eau a **ew". # **"*erar. *'wa" * *CD'**'. 0""* * ****. U & 8a"*"

• • 8*'Y '

ena mensiv eness,J Same as 8. above.

10. SUPPLEMENT ARY NOTES Nnna

11. ASSTRACT (Japesem er mes Title 10 of the Code of Federal Regulations (10 CFR), Section 50.44, " Standards for Combustible Gas Control System in 1.ight-Water-Cooled Pwer Reactors," requires that systems be provided to control hydrogen concentration in the containment atmosphere following an accident to ensure that containment integrity is maintained. The purpose of this report is to provide regulatory guidance to licensees with Mark III containments with regard to demonstrating compliance with 10 CFR 50.44, Sections (c)(3)(vi)ano(c)(3)(vii).

In this report, the staff provices its evaluation of the generic methodology proposed by the Hydrogen Control Owners Group. This generic methodology is documented in Topical Report HGN-112-NP, " Generic Hydrogen Control Information for BWR/6 Mark III Containments."

In addition, the staff has recommended that the vulnerability to interruption of power to the hydrogen igniters be evaluated further on a plant-specific basis as part of the individual plant examination of the plants with Mark III containments.

st AvAnLAtiLaTV 81 ATit#ENT

12. KE Y WORDS/DESCRtPT ORS itee eeve er sasene sw ee seme sseeevaart m aureeme ene sesers.#

Mark III Containment Hydrogen Control unlimity,d

,,,, c,,,,, c

,,,,c,,,_

rra.,M.,

unclassified fram papertJ unclassified 16.NueettR OF PAGES

16. PRICE seRC Pome 336 faael

THIS DOCUMENT WAS PRINTED USING RECYCLED PAPER.

f y

UNITED STATES I

inciat,otarucuss we

! NUCLEAR RE20LAT@RY CEMMISSIEN '

• o*'^tet** Po l

WASHINGTON, D.C. 20tiO6

,,,,,y orrectAL sustNESS.

= PENALTY FOR PRIVATE USE,6300 -

1

)

1ANI 1

te ggLe'""ec 1

2 ->

ghw" B

I, il il y.

m s

.p s

i 1.i