ML20056B292
| ML20056B292 | |
| Person / Time | |
|---|---|
| Site: | River Bend  |
| Issue date: | 08/14/1990 |
| From: | Abbate C Office of Nuclear Reactor Regulation |
| To: | Deddens J GULF STATES UTILITIES CO. |
| References | |
| NUDOCS 9008280083 | |
| Download: ML20056B292 (2) | |
Text
m.
y m
.. S N > .; D .
, < , August 14, 1990
> Docket No. 50-458 1
Gulf States Utilities ATTN: Mr. James C. - Deddens Senior Vice President (RBNG)
-Post Office Box 220.
LSt. Francisville, Louisiana 70775 l
Dear Mr. Deddens:
SUBJECT:
RIVER BEND STATION FINAL ANALYSIS REQUIRED BY 10 CFR 50.44-The NRC staff has completed its- review of. the Mark III Conteinment Hydrogen Control Owners Group (HCOG) Topical Report HGN-112-NP. Enclosed is a copy of -;
the staff's Safety Evaluation Report (SER) transmitted to HCOG by letter dated' August 6, 1990. ,
1 By letter dated October 8, 1987, you committed to provide, within six months ;
ofl$ER issuance, the final analysis of the River Bend Station combustible gas ,
control system required.by 10 CFR 50.44. Accordingly, you are requested to :
provide'the final analysis by March-1,.1991, addressing each of the key elements identified in the staff's SER, Section 8.0.
The reporting and/or recordkeeping requirements contained in this letter affect fewer than ten respondents; therefore, OMB clearance is not required under P.L.96-511.
Sincerely, 1 J
OriginalSignedByi Claudia M. Abbate, Project Engineer
' Project Directorate IV-2 Division'of Reactor Projects - IIT, IV, V and.Special Projects j Of fice of Nuclear Reactor Regulation j
Enclosure:
DISTRIBUTION As-stated i, Docket iileA E.'Peyton NRC PDR .C. Abbate ,
cc: See next page Local PDR OGC l PDIV-2 Reading E. Jordan POIV-2 Plant File ACRS (10)
D. Crutchfield L. Constable B. Boger L. Kintner C. Grimes s
l l
l !0FC :PD4/LA :PD4/PM :PD4/D : :
g ..........
. g. , ............._....__ .....____ :........ ....
! ;NAME :EPeytoh :CAbba e:bj :CGrimes : . :
l _
D ATE :08/W /90 :08/l&/90 :08////90 : : :
l' RB LTR/
0FFICIALRECORDCOPYQ{G]7DocumentName:
L l
9008290Q83 900814 PDR ADUCK 05000458 h p PDC
- w. o 1,- _ .c ;r 4
' Mr. James C. Doddens < : o- cc:-
a.' Mark Wetterhahn, Esq.-- Ms. ' H. Anne Plattinger Bishop,^ Cook, Purcell-& Reynolds $456 Villa, Rose Drive-
+
1401-L' Street, N.W.- Baton Rouge, Louisiana 70806-Washington, D.C. 20005 1
Mr. Les England ..
Director - Nuclear Licensing
' Gulf States Utilities Company P. O. Box 220 w St. Francisv111e, Louisiana 70775
- Mr. Philip G. Harris Cajun Electric Power Coop. Inc.,
10719 Airline Highway LP.-0. Box'15540
, Baton Rouge, Louisiana 70895 Senior) Resident Inspector P.:0. Box 1051t St. Franctsv111e.. Louisiana 70775 e
Presidentlof West Feliciana-Police Jury P._0; Box 1921 St. Francisv111e, Louisiana 70775 Regional Administrator, Region IV x U.S.-. Nuclear' Regulatory Commission 611 Ryan. Plaza. Drive, Suite.1000
' Arlington, Texas .76011 .
Mr.'J. E. Booker Manager-Nuclear IndustryL Relations
> < P. 0.; Box 2951 Beaumont.-Texas 77704 Mr. William H. Spell, Administrator
-Radiation Protection Division Office of Environmental-Affairs P. O.-Box 14690 Baton Rouge, Louisiana 70898 Mr. J. David McNeill, III ib William G. Davis, Esq.
Department .c f Justice Attorney General's Office P.-0. Box 94095 Baton Rouge, Louisiana 70804-9095 A
.. . _ - -. - .. - . ~ . . . - . . . . . _ .-
)
' ' ? T .. %' UNITED sTATts ENCLOSURE'l l L#.
$ NUCLEAR REGULATORY COMMISSION wasme,ovow. o. c.aossa l
. ,.f l .!
h.b ..*
August 6. 1990
)
Mr. J. R. Langley Project Manager, Mark III Containment Hydrogen Control Owners Group (HCOG) c/o Gulf States Utilities .
North Access Road at Highway 61 St; Francisville. LA 70775
Dear Mr. Langley:
~
SUBJECT:
ACCEPTANCE FOR REFERENCING OF LICENSING TOPICAL REPORT TITLED,
" GENERIC HYDROGEN CONTROL INFORMATION FOR BWR-6 MARK III CONTAINMENT $", HGN-112-NP
- We have completed our review of the subject topical report submitted by your letter dated February 23, 1987.
I E We find the report acceptable for referencing in licensee analyses of hydrogen l' control systems for BWR Mark !!! containments under the limitations delineated y in the re enclosed.-port Theand its references evaluation definesand thethe associated basis NRC evaluation for acceptance .which is of the report. I Furthermore, each licensee should provide a plant-specific analysis.and an.
assessment of the need for an independent power supply for the hydrogen ignition system. The plant-specific analysis may use test data described in the topical report to confirm that the equipment necessary to establish and maintain safe shutdown and to maintain containment integrity will be capable of performing their functions during and after exposure to the environmental conditions created by the hydrogen in all. credible severe accident scenarios.
l.
L Recent risk studies reported in NUREG-1150 have shown that the overall core meltfrequencyforoneMarkIIIplant(theGrandGulfNuclearStation)is very low, i.e., IE-6/ year. However, a potential vulnerability for Mark III l'<- . plants involves station blackout (SBO), during which the igniters would be -
l inoperable; and this condition appears to dominate the residual. risk from '
severe accident in the Mark III plants. Under S80 conditions, a detonable mixture of hydrogen could develop which could be ignited upon restoration of power resulting in loss of containment integrity. On the basis of-a separate
.. evaluation of this possibility in the context of the NRC staff Containment.
Perfomance Improvement (CPI) program, the staff has recomended 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 l Examinations (IPEs) of the Mark III plants. The staff has requested that the Itcensees consider this issue as part of the IPE in Generic Letter 88-20 Supplement 3.
t
~,O q n; #t
. . . , s;i m.
.;' : e *:.
I.- -2 '
? Mr..h R.' Langley- ..
d to-repeat our review of the matters licensee described requests for in the report ' '"
zWe'do notL'inten and found acceptable when the. report is referenced tin- to ensure, l d
-approval ofifinal analyses d of the hydrogen control system,
!Our _ acceptance applies only to t$e matters escr references. m-In accordance with procedures established l inofNURES 0390 we req ths
, submit to the NRCThe accepted accepted versions versionsof this report should incorporatewith n three mon Thethis letter
- d e receipt of this letter.and the enclosed evaluation betwee as a basis for the staff's evaluation. accepted) following the report identification E mus6
, -A (designating'd include an application for withholding the proprietary t submittal shoul information accompanied by an affidavit meeting theedrequirements "l ~ of 10 a
version'of the proprietary reports referenced L2.790(b). lication. and inco !:
topical report and intended to be employed as a-part of a licensee app o clusions as to- i-
.Should'our criteria or regulations change such that/ our theclicensees n I i
. referencing the topical report will be expected to rev se ath i f the continued effective
. respective documentation, or submit justificat on orapplicabilit documentation.
Sincere ,
-- Ashok C'. Thadani, Director
> Divis on of Systesis Technology- 4 Office of Huclear Reactor Regulation 8
9 4
m
s.s
- . ,c ' r ..
~
THE GENERIC. .
SAFETY EVALUATION REPORT RELATING TO
. THE MARK III. CONTAINMENT HYDROGEN CONTROL i
1 4
t im
, >e
.4 n
~
?
, '4 m,r3 a,
jg .
L?. ;.
s
. w CONTENTS-Face; "1 INTRODUCTION AND BACKGROUND.................................... 1 2 GENERAL' DESCRIPTION OF THE HYDROGEN IGNITION. SYSTEM............ 3 4 3< COMBUSTION / IGNITER TESTING..................................... 6 3 .1 Sma l l - Sc a l e T e s t s . . . . . . . . . . '. . . . . . . . . . e . . . . . . . . . . . . . . . . . . . . . 6 3.2 Quarter-Scale Test Fac111ty............................... 7 3.2.1 Scaling Methodology.............................. 10 3.2.2 Quarter-Scale Testing Approach................... 10 3.2.3 .Qua rter-Scal e Test Resul ts . . . . . . . . . . . . . . . . . . . . . . . 11
.4 CONTAINMENT STRUCTURAL CAPACITY................................ 14 5- DEGRADED CORE EVENTS AND HYDROGEN GE.* RATION. . . . . . . . . . . . . . . . . . . 15 5.1 Introduction.............................................. 15 5.2 Evaluation................................................ 18 5.2.1 Acceptabl e HGE Sequence. . . . . . . . . . . . . . . . . . . . . . . . . . 19 5.2.1.1 The HCOG's Base-Case Scenario.............. 19 5.2.1.2 Station Blackout and NUREG-1150............ 20 5.2.1.3 TBU the " Acceptable Sequence".............. 20 5.2.1.4 Hydrogen Generation Profiles............... 21 5.2.1.5 Non-Mechanistic Hydrogen Release Profile... 22 5.2.2 BWRCoreHeatupCode(BWRCHUC)...................25 5.2.2.1 Introduction............................... 25 9 5.2.2.2- Phenomenol ogical Assumptions . . . . . . . . . . . . . . . 26 5.2.2.3 Steam Production........................... 26 5.2.2.4 Hydrogen Generation........................ 28 5.3 Summary and Conclusions................................... 29 6 CONTAINMENT RESPONSE - ANALYTICAL MODELING. . . . . . . . . . . . . . . . . . . . . 30 6.1 Localized Combustion /CLASIX-3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 6.2 Containment Pressure and Temperature Calculations......... 33 6,3 D rywe l l A na l ys i s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 3 6.4 Existence of Drywell D.iffusion Flames..................... 34
~~
Mark III SER 1
m
- . . ,(. ... ; y
- +;c .-
CONTENTS'(Cont'd)
Pace L7 SURVI VABILITY OF ESSENTI AL EQUI PME NT. . . . . . . . . . . . . . . . . . . . . . . . . . . 38 7.1 Identification of Essential Equipment.. . . . ... ........ . . ... 38.
7.2 Generic Equipment Survivability Analysis:(Localized Combustion)............................................... 39 7.3 Diffusion Flame Thermal Ent ironment Methodology........... 41
- 7. 4 S pray Ava il a bil i ty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 2 '
- 7. 5 Pres su re Ef fec ts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
-7 . 6 De t o n a t i o n s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 3 '
8: CONCLUSION..................................................... 43 APPENDICES A- GENERIC HYDR 0 GEN IGNITION SYSTEM TECHNICAL SPECIFICATIONS B MARK III COMBUSTIBLE GAS CONTROL EMERGENCY PROCEDURE GUIDELINE C. BIBLIOGRAPHY D ACRONYM LIST.
LIST OF FIGURES 2.1 General Hydrogen I gni tor As semb1y. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 3.1 Elevation Views of the Quarter-Scale Test Facility............. 8 3.2 Facility Configuration Tests S.01-S.11......................... 9 4.1' Ty pi c al Ma rk I I I Con ta i nmen t Co,nfi gu ra ti on . . . . . . . . . . . . . . . . . . . . . 17.
>5.1 Hydrogen Generation Rate (150 GPM Reflood). ... . . . . . . ........... 23 5.2 Hydrogen Generation Rate (5000 GPM Refl ood).. .. . . . ... .... .. .. .. 24
- 6.1 Pe r ry CLAS I X-3 Mode 1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 6.2 SORV Wi th- No Spray-Wetwell Temperature. . . . . . . . . . . . . . . . . . . . . . . . . 36 6.3 SORY Wi th No Spray-Wet vell Pressure. . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 LIST OF TABLES 4.1 Comparison-of BWR Mark !!! Containment Characteristics......... 16 I
5.1 Hydrogen Release Profiles...................................... 22 Mark III SER 11
. . q, . . ,. .
e i
1 INTRODUCTION AND BACKGROUND Following a loss-of-coolant accident in a light-water reactor (LWR) plant, com-bustible gases, containment as aprincipally (hydrogen, result of 1) may metal-water accumulate reaction -
inside involving thethe fuelprimary elementreactor cladding; (2) the radiolytic decomposition of the water in both the reactor core and the containment sump; (3) the corrosion of certain materials within the containment by sprays; and (4) any synergistic chemical, thermal, and radio-lytic effects of post-accident environmental conditions on protective coatings and electric cable insulation.
To provide protection against this possible h Title 10 of the Code of Federal Regulations 10CFR)Section50.44,"
(ydrogen as a result ofStandards an accident.
for combustible gas control system in light-water-cooled power reactors," and GDC 41, " Containment atmosphere cleanup," Appendix A to 10 CFR Part 50, require that systems be provided to control hydrogen concentrations in the containment atmosphere following a postulated accident to ensure that containment integrity is maintained. Conventional hydrogen control systems (e.g., hydrogen recombiners) historically have been installed to provide the capability to contM1 the rela-tively low rate of hydrogen accumulation (or oxygen accumulation n inerted containments) resulting from radiolytic decomposition of water, corrosion of metals inside containment, and environmental effects on coatings and insulation.
However, the net free volume inside containment (or inerting of the containment volume) is used to control the rapid hydrogen production resulting from a metal-water reaction of the fuel cladding. That is, the containment volume is large enough so that the hydrogen generated early would not reach the lower limit of flammability-(or the inerting would prevent combustible mixtures). The rationale for this approach is that the rate of hydrogen release resulting from cladding reaction was assumed to be too rapid (on the order.of minutes) following a pos.
tulated accident to allow for an active control system. Thus, hydrogen control systems (recombiners) would only be actuated later in the event to control the slow hydrogen /or oxygen release associated with radioly:;is and reaction of materials inside containment.
To quantify the metal-water source for design-basis accidents, 10 CFR 50.44, codified in October 1978, requires that the combustible gas control system be capable of handling the hydrogen generated from five times the amount calcula-ted in demonstrating compliance with 10 CFR 50.46 or the amount corresponding to the total , reaction of the cladding to a depth of 0.23 mils, whichever amount was greater. Typically, this would translate to a 1 to 55 metal-water reaction of the active clad.
However, the accident at Three Mile Island Unit 2 (TM1-2) on March 28, 1979, resulted in a metal-water reaction that involved approximately 45Fof the fuel cladding (i.e., about 990 lbs), which resulted in hydrogen generation well in excess of the amounts specified in 10 CFR 50.44. The combustion of this ,
hydrogen produced a significant pressure spike inside containment. As a result, it became apparent that additional design measures were needed to handle larger hydrogen releases, particularly for smaller volume containments and those with F -k 111 SER 1
4 I
lower de ign pressures. The Nuclear Regulatory Commission (NRC) determined that a rulemaking proceeding should be undertaken to define the manner and extent to which hydrogen evolution and other effects of a degraded core need to be considered in plant design. An advance notice of the rulemaking proceeding on degraded core issues was published in the Federal Renister on October 2,1980.
In addition, a new unresolved safety issue was insti".uted (A-48, " Hydrogen Con-
' trol Measures and Effects of Hydrogen Burns on Safety Equipment") to evaluate this new concern.
To formalize its requirements for additional hydrogen control sensures to deal with degraded core accidents affecting pressurized-water reactors (PWRs) with ice condenser containments and BWRs with Mark III containments, the NRC pub-lished an amendment to the hydrogen rule (10 CFR 50.44) on January 25, 1985 (50 FR 3498). The amended rule required that a hydrogen control system be pro- 4 vided and that the systes be capable of accommodating, without loss of contain-ment structural integrity, the amount of hydrogen generated from a metal-water reaction of up to 75% of the active fuel cladding. In addition, systems and components necessary to establish and maintain safe shutdown must be capable of performing their function regardless of hydrogen burning.
Pursuant to the provisions of the rule each utilit ment has installed a hydrogen ignition, system (HIS)y with a Mark III contain-and submitted analysis and a schedule for meeting the full requirements of the rule. The affected plants with Mark III containments are Grand Gulf Nuclear Station, River Bend Station, Perry Nuclear Power Plant, and Clinton Power Station. The staff's interim evaluations of these initial plant responses are documented in supplements to the safety evaluation reports for each of these plants (NUREG-0831, NUREG-0989, NUREG-0887, and NUREG-0853, respectively).
These responses were aided by the efforts of the Mark III Containment Hydrogen Control Owners Group (HCOG), formed in May 1981 by the utilities with Mark III containments to collectively perform testing and analyses to demonstrate the effectiveness-and reliability of the hydrogen ignition systems.
In addition, each licensee with a Mark III containment is required to provide a final analysis [10 CFR 50.44(c)(3)(vii)(B)] to confirm the conclusions of the preliminary analysis and/or, if necessary, to institute modifications to ensure compliance ~with the rule. The scope of this analysis is specified in 10 CFR 50.44(c)(3)(vi)(B). The generic findings from the HCOG's program will be utilized for this final analysis, supplemented by plant-specific design considerations as addressed in the licensee's IPE program.
The following staff evaluation focuses on the assessment of the completed gen-eric testing and analyses performed by the HCOG in support of the plant unique analysis. The HCOG activities have been summarized in a topical report trans-mitted by letter dated February 23 1987, correspondence identification HGN-112-NP, "GenericHydrogenControlInformatIonforBWR-6MarkIIIContainments." The topical report is a summary document of all of the individual generic submittals
-that have been sent to the staff by the HCOG. It should be noted that HCOG correspondence identification designators with a "P" suffix (HGN-XXX-P) are proprietary to HCOG. Whereas, those without a suffix or with an "NP" suffix are nonproprietary.
Mark III SEP 2
o . ..
e
=
As p"t of the review of the HCOG program, the staff obtained technical assist-ance Trom the Sandia National Laboratory ($NL), the principal contractor for the NRC research pronram on hydrogen control and combustion phenomena. SNL provided the NRC with an Independent assessment of technical issues contained in selected HCOG submittals pertaining to hydrogen behavior.
E The staff evaluation of the generic considerations of the hydrogen control system for the Mark III containment can best be understood if categorized as follows:
{
- - general description of the hydrogen ignition system combustion and igniter testing
(
- containment structural capacity
- degraded core events and hydrogen generation
- containment response-analytical modeling
- survivability of essential equipment
- overall conclusions b
Therefore, the following diseassion will follow this general outline.
{
2 GENERAL DESCRIPTION OF THE HYDROGEN IGNITION SYSTEM The regulation, 10 CFR Sn.44(c)(3)(iv)(A), states:
Each licensee with a boiling light-water nuclear power reactor with a Mark III type of containment... , shall provide its nuclear power reactor with a hydrogen control system justified by a suitable pro-gram of experiment and analysis. The hydrogen control system sust be capable of handling without loss of containment structural integrity an amount of hydrogen equivalent to that generated from a metal-water reaction involving 75% of the fuel cladding surrounding the active fuel region (excluding the cladding surrounding the plenum volume).
The concept employed by the licensees with a Mark III containment, and similarly the ice condenser licensees, is to intentionally ignite hydrogen generated inside containment. This method precludes buildup of relatively high concen-trations of hydrogen during degraded core accident scenarios.
To accomplish this early ignition a hydrogen ignition system (HIS) is installed
- in each of the four plants with Mark III containments. The HIS is a system which consists of approximately 100 igniter assemblies distributed throughout the drywell and containment regions.
T.ie main eles .t of tM igniter assembly is the Model 7G thermal igniter glow commonly used in diesel engines) manufactured by the General Motors AC plug Divis (ion. Each Mark III containment has an identical igniter assembly design.
Each igniter is powered directly from a 120/12-V stepdown transformer and de-signed to provide a minimum surface temperature of 1700'F. The igniter assem-bly (see Figure 2.1) consists of a 1/8-inch-thick stainless steel box that con-m tains the transformer and all electrical connections and is manufactured by the Power Systems Division af Morrison Knudsen. Igniter assemblies are Class 1E, seismic Category I, and meet the requirements of the Institute of Electrical j and Electronics Engineers (IEEE) Std 323-1974 and HUREG-0588 for environmental qualification.
5 l Mark III SER 3 F
e___
1 w.
>. +...
t
- t- -
= . i -..igj
- 1 _
1 L , ,n .
L 4 . 4 '"
y~
- ,,,. .j . v. . 4.. . . . ,,,j ,;y etuG
{ :'4,:3............,,<g
, !+ 3 vi. ,.,
i m
1 -j ;'
.i . !. ., ! f MITU M 1; ..
, i l' LA u.f ; ,' ' j4 u:1;;r sci 'I
, l t
jel PWL ,
jt ; l ,
( i !. ..
L j.
) .i
. e ..
l.l,4 '.
.,-e*;. 4 '.:, . l
- : .' .....4.......*'.*
- O O . 30tTP
_s, -- .
r i o. -
- j-1; .
- i c': 5j h I
4 l L g7 a
- s= :8 ,l l I.
4
, o
' 'l W w
\ .e a==
GLO. :'li.'
.: 3 PLUG g .,
- e . . .a ,
p ,r/n w a.s$ersi fg e 1R Ah$ FORT (R g . w.m.m m .
d L. .. i G su4 ;.
..............._t i i
l l g .
f Figure 2.1 General hydrogen igniter assembly Eark III SER 4
l
. I e ',
The igniter assemblies are divided into two groups; each group being powered from i a separate Class 1E division powv supply. The intent is to have at least two 1 igniters located in each enclosed volume or area within the containment that could be subject to possible h separate power division.ydrogen In openpocketing, areas within and theeach containment igniter bear.d powered drywellfrom a regions, igniter assemblies at the same elevation are designed to have alternating power division sources. I safety feature (ESF)gniters are separated power divisions by about are operable or by 30aboutfeet when both 60 feet wheneng'neered one power division it inoperable. Igniter placement is designed to be more widely spaced ih ,
the large open regions, such as a >ove the refueling floor, and in the lower regions of the drywell that are subject to flooding. Requirements of placement as well as '
other parameters of the system are contained in the technical specifications for l' the hydrogen ignition system, as proposed by the HC0G, and addressed in Appendix A of this report.
Each igniter power division has a corresponding e: Site emergency diesti generator.
Incorporation of the emergency diesels into the enign addresses the question of igniter power for many sequences, but not for station blackout conditions.
Under blackout conditions, the HIS would not be operable. On the basis of a separate evaluation of this possibility in the context of the NRC Containment Perfonnance Improvement (CPI) program, the staff has recomended 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 Examinations of the Mark Ill plants. (See Generic Letter 88-20, Supplement 3)
The HIS is designeo for manual actuation from the main control room. Actuation by the operator is required by plant emergency procedures when the reactor pres-sure vessel (RPV) water level reaches the top of active fuel (TAF). The proposed combustible gas control emergency procedures for plants with Mark !!! contain- ,
ments are further discussed b Appendix 8.
By letter dated March 5,1986 (HGN-073), the HCOG provided the following just*-
fication to support manual actuation. Actuation is linked to indication of the
, RPV water level, which is a key safety parameter and is closely monitored by the operators. Also, HIS actuation requires only the positioning of two hand-switches. Furthermore, operators should not hesitate to energize the system during accident scenarios in which the hydrogen threat is uncertain or marginal
' because there would be no adverse effect on the plant as a result of unnecessary igniter actuation. Time available to actuate the HIS is the other significant '
parameter. Based on the hydrogen events considered, the HCOG estimated thfF time to be approximately no less than 25 minutes; that is, after the water level t reaches TAF to the lower hydrogen flamability limit reached in the wetvell volume. The HCOG also noted that hydrogen would migrate to the upper portions -
of the containment before the wetwell reaches hydrogen combustian conditions. ,
This effect was seen in the quarter-scale tests. Therefore, the time intervs1 '
is expected to be somewhat greater than 25 minutes.
Because (1) manual actuation is a simple task, (2) the operator has sufficient time to perform the task (3) and there are no negative effects if the system is inadvertently or unnecessarily actuated, the staff finds manual opePator actuation acceptable.
- Mark 111 SER 5 5
- _ _ - - - - - - - _ _ - - - - - - _ - - _ _ . . _ _ , . . , . ,~...~,-,,,,,,-,.-.m., ,,e e,.,,,,- ,,.aw.,.,.m,n,,,..,,,,
a 1 H1s Design Assessment The staff finds these hydrogen igniter systems currently inst 611ed in the '
plants with a Mark !!! containment to be acceptable, with the covest that the vulnerability to interruption of power to the hydrogen igniters be further i evaluated on a plant-specific basis as part of IPEs of the Mark !!! plants. i (SeeSupplementNo.3toGenericLetter88-20) l 1
3 COMBUSTION / IGNITER TESTING i
Numerous research programs have been conducted since 1980 to better understand hydrogen combustion behavior and the performanc.t of ignition devices. Because )
these programs were varied in scope, Sandia National Laboratory ($NL) sumar12ed the findings of recent hydrogen combustion test programs in NUREG/CR-5079. This report also provides additional background information and insights related to hydrogen behavior.
The specific test programs considered necessary by HCOG to support the unique plant characteristics of a Mark !!! containment are discussed below.
3.1 Small-scale Tests Small scale hydrogen combustion tests were perrforwed at Whiteshell Laboratories and documented by the HCOG in a letter dated June 7, 1984 (HGN-017-NP). The program was intended to investigate ignition and combustion behavior of mixtures <
predominantly composed of hydrogen and steam, i.e., with limited available oxygen. l This condition may exist for a postulated drywell break event in which air is initially swept from the drywell and then later reintroduced into the steam-hydrogen environment. These tests confirmed that such hydrogen sir-steam mix-tures can be successfully ignited as long as the oxygeri concentra'. ion exceeds approximately 5-6 volume percent. ,
A 1/20th scale Mark 111 hydrogen combustion program was conducted by Acurex Corporation and documented by the HCOG in a letter dated February g, 1984 (HGN-014-Np). The objective of this program was to provide a visual record of hydrogen combustion behavior in a 360-degree model of a Mark !!! -
containment. Modelling included the suppression pool and major blockages in the annular region between the drywell and outer containnent walls. Hydrogen was admitted through simulated quenchers and/or vents into the suppression pool and ignited by prototypical ignition sources.
The most important result obtained from the 1/20-scale test was the confirmation
' of continuous hydrogen burning in the form of steady diffusion flames above the suppression pool. "he significance of this mode of hydrogen burning is the observed severe thermal loads that occur near the diffusion flames and could threaten the integrity of the containment ano squipment. Diffusion flames were observed when hydrogen flow rates of 0.4 lb/sec (full-scale equivalent) or P
Mark 111 SER 6
_ ' ' ~ ~ ' '
grgater were used.
Combustion was istitiated by the igniters and rapidly pro p- '
i agated to the pool surface and fonsed steady diffusion flames that were an chored released the at the hydrogen. surface of the pool and located above the submerged sparc,ers ,
- gen injection flow rate was increased, as evidenced by tall temperatures.
V Fscale single-sparger mockup was constructed.As part of the 1 predicted based on results of V20-scale tests was observe On the basis of these tests, it became necessary for the HCOG to pursue a larger-scale tes i program to containment. obtain thermal environmental data more representat1ve of a Mark III Subsequently, the HC0G undertook an extensive program to better define the conditions that could exist during a degraded core accident. Amajor element test program. of this effort was the quarter-scale Mark III containment combustion 3.2' Quarter-Scale Test Facility The research quarter-scale program. test program became the major element of the HCOG's hi The primary objective of this program was the investigation and on thecharacterization suppression pool in a ofMark the !!!environment containment. that could result from diffusive bu i
gathered from theofquarter-scale ing the survivability select equipment. test facility (QSTF) would be used in determ Test facility description and the 1986; HGN 115-NP, Februarycombustion test results scy be found in the i 10, 1987; and HGN-121-P, July 22, 1987. i The test facility in West Gloucester, is a quarter-linear-scale model of a Ma Rhode Island.
The test enclosure is designed to operate at pressures up to 40 psig and consists of an outside tank, 31.5 feet in diameter, 49.4 feet high, containing a smaller tank, about 21 feat in diameter by 23 feet high.
, and other Thelarspace between the two tanks is the test volume; which contains floors containments.ge blockages simulating the obstructions that exist in the actual Because of the unique features of the four plants studied modular construction needed. of the annular floors is used to modify the vessel interior,when At the bottom of the two tanks Several views of the facility are shown,in Figures 3.1 and 3.2.the suppression po Facility design features include:
simulated
- 48) loss-of-coolant-accident (LOCA) vents (top row only, numberin i simulated spar totalling 24) gers / uniformly spaced every 15 degrees azimuthally and
+ !
unit coolers (for the River Bend configuration) _.
The gas velocities facility gas is heavily instrumented to measure gas and surfare temperatur concentrations cameras are use,d for a visual recor,d. heat fluxes, pressure, and five video Mark III SER 7
- ), ,
)
D IE kl I .
[i.-h N ] '
\ l
/!
rio. "
b
.V h . v._h 8
!" V__N h w
d F e g j
<, , 3y g ;
g
......m i
W g b, % hY - h %
N. g a
Il m. .;
D .!
4 g
M \ '
i ' q *e i
< a
[ . ,
g_- 7 g
y_ >
e MTV -
>V. . . . .- .
t 61 I .
Mark 131 SER 8
l
.. .. j k k IC b kk t' d t
\
i C'.b.5lEJJ 88
\\
N \- \ \\ *
)\
1_(\ \ = +
1:v.(
\ A\ '
i
, n. " '
I- < .
\
- s '
. ...w
- . .. ..,: \ >
y $ \
\
n s 7 l
\
' /\ \
\A 4
L =
h #l.: , t g{ y,[, ,'.f r \ a E hyi i
~
l , 3 !
- f. ,
t R
-: 1
- 1 l I- %
. 4 u s ,hg l c b' l l l lb
?. I K )N I Et ,
i I
,MIa ~
/ /\ . - ~ ,/
y ili y.
\ .
,t;p;l4pMi Mi.gp,4;g. .
I ,j;
- .g . ige. Q
- ph
- g!;s.
M 4 a @ k,s ~.
); ,
=ge V/ / Y .
[ _g si Mari 111 SER 9
- : l 3.2.1 Scaling Methodology The thecretical basis for modeling hydrogen flames in the test facility is based on Froude scaling. The modeling assumes fully developed, buoyancy dominated, i turbulent 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 numerous experimental demonstrations in the field of fire research. 1 Using this type of modelling full scale was scaled tes quarter scale, a 4 to-1 )
linear scaling resulting in:
i 32-to-1 reduction in mass and volume flow rates i 64 to-1 reduction in total hydrogen released !
i 2-to-1 reduction in the time scale l
- 2-to-1 reduction in gas velocities !
1-to-1 relationship for gas temperatures and gas concentrations l
Flame heights and global flow patterns also were determined by Froude scaling. i Generally, Froude scaling was used to reasonably and practically design the
' QSTF blockages (e.g)., spray flow
. However, theand droplet following sizes, heat were discrepancies sink noted:
thermal characteristles, I (1) The quarter-scale tests revealed that the insulation used in the facility became wet and its thermal properties departed from dry insulation. J (2) The QSTF had only 30% of the mass as prescribed by Froude scaling.
(3) The scaling method did not rigorously simulate convective and radiative heat losses to structural heat sinks.
k assess the impact of these discrepancies, the HC0G provided a comprehensive analysis as documented by HCOG 1etter HGN-085 dated May 5, 1986. The HC0G determinedthatcompensatingeffectsexistedInthetreatmentofheatsinks.
Thus the data obtained from the QSTF does provide e reasonably accurate desciiption of the full-scale thermal environment when extrapolated by Froude scali3g.
Toassistthestaffinthereviewofthiscomplexmatter,SNLstudiedthesubject analysis and submitted various comments. On October 7, 1987, a meeting between the HCOG, SNL, and the staff took place to resolve the SNL's comments. Subse-quentl 1987. y,On HCOG documented the basis its responses of the additional in a letterSNL information, (HGN-128) concluded dated November and the staff 6, concurs that the application of quarter-scale experimental data directly to full-scale equipment survivability can be done conservatively in spite of the above discrepancies. SNL's assessment is documented in correspondeice dated December 23, 1987.
3.2.2 Quarter-Scale Testing Approach _,
Tests were performed in the QSTF for the four different plants with a Mark III containment; i.e., the QSTF was customized to reasonably represent the plant-specific characteristics of each Mark III crafiguration. During the tests,
)
Mark III SER 10 L
hydrogen was released through spargers used to simulgte the automatic depres-surization system (ADS) and a stue cepen relief valve (50RV) or through the '
simulated LOCA vents. Two h a low-reflood case (150 ppm)ydrogen release and a high-reflood profiles case were Aused (500 gpe). in the facility, discussion of the development of these profiles is contained in Section 4.
However before the plant unique or production tests were conducted, a series ofscopIngtestswereperformedtoassessdatarepeatabilityandthesignificance .
of various parameters. Results from these tests formed the basis for developing the final test matrix that were used in the production test program. Also, in the early development of postulated degraded core events spray availability was mcertain, therefore tests were performed with and wIthout sprays activated.
In tk production tests, each plant had its own specific array of tests, focusing on the 50RV locations, the combination of ADS spargers and LOCA vent releases, and the effect of sprays / coolers. The data obtained from these production tests formed a basis for determining the full-scale thermal environment and became a central element of each licensee's final analysis. This information was used as input to analytical evaluations of equipment thermal response for assessing sur-vivability of critical equipment. Further discussions on the use of this data are contained in Section 7.
3.2.3 Quarter-Scale Test Results The scoping' test and partial production test results are summarized and presented in the HCOG s correspendence (HGN-098-P and HGN-121-P). These results demon-stratedthatthedistributedglowplugIgnitersystemcanprovideaneffective means for limiting accumulation of hydrogen in plants with Mark III containments.
Hydrogen concentrations throughout the facility were maintained near or below 5 volume percent (dry basis) for all tests and steam concentrations were deter-mined to be about 10-15 volume percent for selected tests. Although low hydro-gen concentrations were maintained, difforent types of combustion behavior were observed during the tests, depending on the synergistic conditions. The various observed combustion modes are described below. ,
Diffusion Flames When hydrogen was released into selected spargers, hydrogen combustion would initiate as a mild deflagration or lightoff (pressure rise about 1 psi) in the wetwell region between the hydraulic control unit (HCU) floor and the suppres-sion pool surface and would persist in the form of standing diffusion flames anchored to the pool surface. This was the dominant mode of combustion and occurred for bulk oxygen concentrations of 8 volume percent (dry) and hydrogen injection rates greater than 0.15 lb/sec. It should be noted that the hydrogen flow rates are full-scale equivalent values (i.e., a 32:1 increase). In this regime of steady flames, combustion was essentially complete. Foraninjection rate of 1 lb/sec, a flame height of about 8 feet (full scale) was reached.
As the hydrogen injection rate was decreased to about 0.15 lb/secreombustion became less complete and the flames less stable. As the rate was further de-creased, diffusion flames on the pool surface could not be maintained. This point is known as the flame extinguishment limit. Moreover, it was observed that this limit was strongly influenced by background gas concentrations.
Mark III SER 11
l
\
To illustrate the various relationships, the following was extracted from,the ,
QSTFscopingtestreport:
(1) The flame extinguishment limit ranges from - 0.07 to - 0.15 lb/
see for ambient hydrogen concentrations below - 4.1 volume per- '
cent dry (high oxygen conditions).
l (2) The flame extinguishment limit de:reases withand increasing concentration to a minimum between - 0.025 - 0.03 lb /see at ahydrogen hydrogen concentration of ~ 4.5 volume percent (high oxygen i conditions.)
(3) For comparable ambient hydrogen concentrations, the flame extin-l guishment limit is slightly higher at low oxygen conditions. i
\
l Another effect that accompanied hydrogen burning was the formation of bulk l air currents. Horizontal air flow was c'ested above the pool surface allowing l
diffusion burning to continue by providing a source of oxygen. Another pattern ,
of circulation was the creation of chimneys, which provide for flow to and from ,
i.
the region of burning and exchange flow with the upper containment, that is, L hot (upward flow) and cold (downward flow) chimneys. .
I l
Localized Combustion Below the flame extinguishment limit, flames on the pool were not observed.
The prevalent burning mode at very low hydrogen release rates has been termed i i localized combustion. This phenomenon is characterized as weak flames or volume l burning through a marginally combustible hydrogen / air / steam mixture. This type l of combustion was detected only in regions at or above the HCU floor and concen-l trated mostly in chimney areas. This was evident by temperature measurements; >
I localized combustion was not observed by video recordings. Localized combus-l tion appeared to he relatively benign (i.e., less than 250'F at instrumented
! locations). Burning was more widespread and somewhat more intense at low oxygen conditions and was accompaN ed by slightly higher background hydrogen E concentrations (i.e.,near5%).
The STF was oriented to invest' gate the burning phenomena in the area immedi- >
atel above the suppression pool. As a result, the instrumentation layout above
, the CU floor was not sufficiert to provide a detailed mapping of the conditions
! in that region. Consequently, a rigorous investigation of localized combustion was not possible; however, the instrumentation that was present, along with HC0G's analytical effort (see section 6), provided a reasonable characterization
! of the phenomena. Localized combustion is discussed further in Section 6 as it relates to tHe analytical methods used by the HC0G. .
f Secondary Burning During the quarter-scale testing, an additional combustion phenomenon was ob-served late in one of the tests. When the bulk oxygen concentration dropped below 8 volume percent (dry), flames extinguished on the surface of the pool but formed at the HCU floor elevation. This type of burning has been termed secon-dary burning.
Mark III SER 12 l
l
i .
j l
l During a June 1986 meeting with the staff, HCOG revealed the presence of secondary l
! burning in one of the Perry production tests. This phenomenon was not observed 4 in previous production tests or in the scoping test phase. Until this particu- !
lar test, only in a scoping test did the containment oxygen concentration drop below 85. Oxygen concentrations were generally maintained above 8 percent due to a unique need associated with the video coverage. Each of the five video cameras used in the QSTF required a continuous air purge for the camera lenses i i
to prevent a condensation on the lens. This resulted 'n a continuous inflow of 1 oxygen in the facility, thus precluding the atmospheric oxygen concentration to l l fall below 8%. However, the camera air purges were not run continuously in the
! subject test; subsequently, late in this test the oxygen concentration fell below 8%. Additional information is provided by HCOG submittal HGN-106 P dated September 29, 1986, and also discussed in detail in the quarter-scale combustion test report.
To present its overall assessment of the significance of secondary burning, HCQG began by addre!tsing the limitations of the QSTF. The QSTF has various physical and practical limitations associated with the investigation of secondary burning. The instrumentation in the facilits was geared to define the thermal environment produced by diffusion flames an'.hored to the surface of the suppres-sion pool, which is the dominant combustio1 mode. Therefore, more instrumenta-tion would be needed to investigate burninp above ".he HCU floor. Since all l plants with Mark III containments have different containment volumes, simulation I of the expected oxygen depletion profile for each plant would be difficult. l Therefore, HCOG evaluated the need to further consider the secondary burning 1 phenomenon. The following identifies the various far. tors considered and their ,
relationship to the four plants with Mark III containments:
(1) Secondary burning is expected to occur over a narrow range of oxygen l concentrations, approximately from 6 to 8%. Based on the hydrogen generated by a 75% metal-water reaction, a Mark III containment would i experience this oxygen concentration. interval late in the transient or not ,
at all. Assuming the drywell air is not added to the containment inven-tory, a metal-w.tcr reaction of 55 to 67% would be reached before the oxygen concentration w expected to fall below 85. This range applies to three of the four plants. Pecause of the larger containment volume-to-power ratio, Clinton is not e m ected to fall below 10% oxygen; thus, secondary burning is not anticipated. When the drywell air inventory is included in the containment region, a metal-water reaction of 67% would be reached for Perry bef ore the oxygen concentration is expected to fall below 8%; River Bend and Grand Culf would have already consumed the equivalent -
hydrogen reqeired by the rule (i.e., 75% of the fuel cladding surrounding the active fuel region).,
(2) Considerable uncertainty is inherent in predicting the long-ters hydrogen profile, especially in the latter phase of the profile (refer toSection5). For example, an alternate accident sequa.nce such as a drywell break sequence, different hydrogen release rates, the.use of the drywell mixing system, or not even reaching a 75% metal-water-reaction value, could reduce or possibly eliminate secondary burning.
rk III SER 13
-<w - - - - - - - w~ ~ - - - .n
c .. .. .
(3) Also,~ not all conditions expected to exist in an actual plant existed in the test when secondary burning was observed. For example, sprays were not activated and the burning on the HCU floor was in a sector where the hydrogen release was most concentrated. This sector was in the 45-degree chimney in which the 50RV was located and the steam tunnel structure would reduce the upward cross-sectional flow area. Therefore. it is expected that these factors contributed to the locally high concentration of hydrogen that is required for secondary burning. It should also be noted that the overall shape of the flames occupied a relatively small area, forming a flame tone near the corner of the steam tunnel and drywell wall. On the basis of these differences, the following significant mitigating factors can be inferred to reduce the consequence of secondary burning:
(a) Increased turbulence inside containment through spray operation or unit cooling could potentially delay or preclude secondary burning.
This was evident to some degree in one of the scoping tests during which conditions were similar to those during the Perry test where secondary burning was present. This scoping test had sprays func-proximately 7.8L tioning andburning Secondary the oxygen concentration did not occur. Also,fell to ap/ unit coolers would sprays provide cooling to mitigate the consequences of secondary burning if it were to occur.
(b) Secondary burning appears to be extremely localized. It occurred in the reg$RVs) valves (ion above the spargers location released where three adjacent safety reliefFurther, secondary hydrogen.
occupied only a small zone. Because of equipment redundancy and separation, secondary burning is expected to affect only one train of equipment.
On the basis of its findings, the HC0G determined further experimental investi-gation of secondary burning was not necessary.
The staff's review of the evidence indicated that s'ocondary burning is not ex-pected to present a significant additional threat and, if this combustion mode were to occur, it is expected that the thermal zone of influence would be lie-ited. Therefore, the staff agrees that further detail study of secondary burning is unwarranted. However, since the redundancy of equipment (i.e.,
spatial separation of equipment performing the same function) is the most important element, the staff requests that the Perry River Bend, and Grand Gulf licensees (excluding Clinton) confirm tr.at sufficient separation (i.e.,
at least a 90' azimuthal displacement) &M sts between the redundant equipment expected to b,e affacted by secondary burning.
4 CONTAINMENT STRUCTURAL CAPACITY The burning of hydrogen inside containment has the potential to induce pressure excursions in excess of the containment /drywell design values. To determine the pressure capability of the containment structures, required by 10 CFR 50.44 (c)(3)(iv)(B), each licensee provided its plant-specific analysis for staff review. The details of the staff's evaluations regarding the containment and drywell ulumate capacities are documented in each of the plant's respective SER supplents. Rather than repeating these evaluations, a brief description ,
of the Math III containment will be provided.
Mark IU 14
In the Mark !!! containment design, the containment completely surrounds the drywell. At the bottom of the containment, a 360-degree annular suppression pool is located between the containment wall and drywell wall. Below the pool wrface, horizontal vents are constructed in the drywell wall. The principal l diffeiu ce between the four plants is in the characteristics of the containment l shell as illustrated in Table 4.1. For Grand Gulf and Clinton, the primary contaInmentisasteel-lined,reinforcedconcretestructureconsistingofa vertical cylinder and a hemispherical done top. For River Bend and Perry, the primary containment is a free-standing steel vessel consisting of a vertical l-cylinder and a torus-spherical dome surrounded by a concrete shield buiiding.
i The internal containment design pressure of 15 psig is the same for each plant.
I The ultimate pressure capacity was determined to be about three times design (i.e., approximately 50-60 psig) for each plant. Since the drywell structure is designed to greater pressure values than the containment vessel, the drywell ultimate capacities also are greater and are not limiting in the forward or reverse direction. The containment pressure capacity, taking into consideration limiting containment penetrations, is used as the limiting parameter when evaluating the consequences of hydrogon deflagrdtions inside containment.
Figure 4.1 is an illustration of a Mark III containment configuration.
5 DEGRADED CORE EVENTS AND HYDROGEN GENERATION 5.1 Introduction To determine the consequences of hydrogen burning, the hydrogen generation release must be addressed to establish a representative hydrogen generation event (HGE) and define representative hydrogen release profiles. l The regulation,10 CFR 50.44(c)(3)(vi)(B), specifically requires that the fol- l lowing be considered in the analysis: q
, I (1) large amounts of hydrogen generated after the start of an accident (hydrogen resulting from the metal-water reaction of up to and including 75% of the -
fuel cladding surrounding the active fuel region, excluding the cladding surroundingtheplenumvolume);
l (2) the period of recovery from the degraded condition; 1
i (3) accident scenarios that are accepted by the NRC staff and that are accom-l panied by sufficient supporting justification to show that they describe l the behavior of the reactor system during and following an accident result-l ing in a, degraded core.
The HCOG analyzed two degraded core accident sequences (HCOG transmittals l HGN-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 power, followed by reactor scram, isolation of both the containment and MSIVs, and power conversion system unavailability. One diesel generator fails to start and the relief valves cycle on high reactor pressure as a result of MSIV isolation. Relief valve cycling results in one stuck-npen relief valve (50RV). The second scenario models a small break in the drnall by using the same total hydrogen and steam release histories as the prev is case but the predicted hydrogen and steam release is cechanistically split tween the drywell and the containment.
Mark III SER 15 1
j
l... ,
. 1
~. 1 Table 4.1 Comparison of BWR Mark III containment Characteristics Characteristic Grand Gulf Perry River Bend Clinton l
Rated thermal output, MWt 3,833 3,579 194 2,894' Number of fuel bundles 800 748 ?4 674 l
Drywell structure:
i Design pressure, psig 30 30 25 30 External design pressure, 21 21 20 17 ,
, psid- !
l' Air volume, fts 270,000 277,685 236,196 246,500 l Suppression pool volume L .(includes vents), fta 1.3E4 1.12E4 1.3E4 1.1E4 Suppression pool surface area, ft2 ~553 482 522 455 i Holdup volume, fta 50,000 40,564 20,353 33,804 Holdup surface area, ft8 3,145 2,61* 2,564 2,490 Containment vessel:
Design pressure, psig 15 15 15 15 Ultimate pressure capacity, psig 56 50 53 63 External design pressure, psid 3 0.8 0.6 3 Total air volume, ft" 1.4E6 1.141E6 1.192E6 1.551E6 Air volume below hydraulic l control unit floor, ft: 151,644 . 181,626 153,792 173,000 1 Suppression pool volume, 8
ft- 1.24E5 1.06E5 1.28E5 1.35E5 j Suppression pool surface area, ft2 6,667 5,900- 6,408 7,175 Uppe;poolmakeupvolume, ft 36,380 32,830 0 14,655 Containment spray flow rate (1 train), gpm 5,650 5,250 0 3,800 j Number of loss-of-coolant-accident vents 135 120 129 102 ;
I l
l l
1 Mark III SER 16
. . , - - - . - . - . . - . - _ - . - . - . . . - - - . - . - - - . - _ - . - . - - ~.. ...
...o. . ..
c- >. ..
9 t
e l
i!
l-
/
=
y ,,,, 14tt LD H f, , sWi& DING
- j. . >; >3 i ., .
is -
t [! , d **'s. CONTAINMENT omvwsLL ;
,. m&AD .
h ., _ ^
h,,,s.WPPIR P0o)
, \An*** d
-(Am-! -p- j-
/ y -r (
.' .l ... ., E
........: j.,, SRV DISCNARGE
. IRLACTOR >g #
[. >
- * *' t 1
e ,,,, RSActon
'? ,
I '.
'1, gnag&p WALL .
- ,, . s
,. I
. s H 3, ; s a ,j +Rvwsu.
h ' ' -
? y' '
i , ,?
4 a, ,, ' , l
. ,. . t.: . .
()
j
's a . .j ,. P ws* w au. I N, c: .
.. ..... # ,, c, )
l y $y' ';f .
d 6. --- = < . ',,,, woamowfat
- . vi m <
J m .. .
........ =
......:.3< .,
r.
.... .,.. . ... .u,,Rs.. u
.... . .. .. .. .. .... ... .... .. ..w. . . 7- m reaL
,,..v...,.......
m.... p. .
. . . . . r l
... .. .. .. .. ..,. . .. : ...'...*,';., .. .,t....'. .. .<. . .
.*f l
. i l
l 1
Figure 4.1 - Typical Mark 111 t ,tainment Configuration l l
1 l
Mark 111 SER ' .
l l
l
- l. _ , _ _ _ . . . _ _ _ . . _ . . _ . . . _ _ _.--.1
The transients resulting in an 50RV were selected (1) to ensure a rapid foss of inventory and (2) to account for and create a limiting local thermal containment environment for analysis and testing. Small-break loss-of-coolant accidents (SBLOCAs) were selected as an alternate sequence to address the potential and consequences for hydrogen combustion in the drywell. Otherwise the SBLOCA sequence is identical to the 50RV sequence.
For the base-case scenario, all ac-powered reactor makt:,o systems are assumed to initially fail. According to emergency procedures, ine operator will depres-surize the reactor when water level decreases to the top of the active fuel or ,
when conditions requiring steam cooling are met. Following vessel depressuriza-tion, low-pressure system injection is assumed t> fail. The scenario continues with the core becoming uncovered and core heatup beginning at about 35 minutes into the transient. Limited hydrogen is produced during core heatup. At about 65 minutes into the event, the core is Mflooded before it becomes nonrecover-able (exceeding a 50% zirconium [Zr] melt fraction). During reflooding of the l core a significant amount of hydrogen is generated. This hydrogen is transported to the suppression pool through the safety relief ealve spargers and into con- '
tainment where it is ignited and burned.
The selection of the 50RV sequence was based on the reactor safety study methodol- i ogyapplicationsprogram(RSSMAP) stud In 1986, the staff questioned the absence of the station blackout (SBO) y. sequence (le tter dated February 21,1986).
HCOGheldtheviewthat$B0isnotalikely)HGEcontributorbasedonits relatively low core melt frequency (HGN-055 . This conclusion is based on the results of the RSSMAP study (NUREG/CR-1659), which assumed Grand Gulf to be representative of the.four plants with Mark III containments. The results of the GESSAR-!! PRA (NUREG-0979) also found that 580 is a dominant contributor to the probability of core damage, although the core damage probability is
! quite low. The PRA results were reinforced by the staff findings reported in l an NRC report (NUREG-1150). In view of these studies, the HCOG revised its I
submittal to account for SBO. These revis'ed results are contained in two reports transmitted by letters dated January 8 and September 9,1987. The staff's review focused on those revised results and also drew on information from previous submittals.
Additionally, the review effort focused on the hydroge production profiles that were derived using the BWR core heatup code (BWRCHUC) described in Science Application Inc. and International Technical Services ITS) reports to the staff; HGN-020. -031 -032 -034. -089, -096, and -132; and HCOG/NRC meeting August 28, 1984). The objective of this review was to ascertain the capabilities and the ro-acceptability files. of the BWRCHUC Particular BWRCHUC for(1)use concerns were in generating the Zircaloy oxidationthe model, hydrogen generat the transient simulation capabilities (3) the ability to predict the maximum expected hydrogen production rate, and (4) the ability to predict the total amount of hydrogen produced in each transient.
5.2 Evaluation ~~
The evaluation was divided into two parts: (') the estaolishment of an accept-able HGE scenario and (2) the acceptability of the BWRCHUC to estimate hydrogen production histories.
l l
l l
Mark III SER 10
p l .
5.2.1 Acceptable HGE Sequence The analysis required by 50.44 is to be based upon an accident sequence which is " acceptable to the staff" and is at the same time limited to " recoverable' events. The rule, however, does not
" acceptability
- or " recoverability.' provide criteria for the determination of The staff position with regard to recoverability is that there should be a reasonable expectation that the original core geometry is generally maintained.
However, a quantitative definition of a degraded core state that is recoverable is not required. The degraded core condition.is a condition in which the reactor core has experienced or is at the onset of experiencing damage from excessivetemperature(includingpermanentdeformationorlocalizedmelting).
Inherent.ly in the degraded core condition is an extended loss of coolant injection without a chance of immediate recovery. The purpose here is not to associate core recoverability with detailed phenomena of cladding or fuel meiting and relocation, but rather to provide a reasonable cut-off as far as the deterministic calculation of hydrogen production is concerned. The total amount of hydrogen production which must be considered is specified in the rule itself. It is in this limited sense that the term " recoverable
- is used in this evaluation.
- HCOG proposed a definition of recoverability in terms of the fraction of Zircaloy cladding which has reached or exceeded the Zircaloy melting tempera-ture of 2170 degrees K. The staff accepted a 50% Zircaloy clad melt fraction as the cut off point for " recoverability" based on HCOG's report that analyses indicate that-at this point significant fuel melting is in progress.
It is the staff's judgement that the maintenance of the original core geometry after damage to this extent is unlikely. Therefore, for the purposes of hydrogen rule considerations and hydrogen generation rate estimates, the 50%
Iircaloy melt fraction criterion is acceptable.
With regard to the " acceptability" of sequences, the staff considered two cri-teria: (1) the likelihood of a given sequence and (2) the contribution to risk from a given sequence. Based upon NUREG-1150, the staff concluded that (1) the most likely HGEs would occur with the reactor vessel depressurized, (2) the potential for greater consequences is associated with HGEs at high pressure, and (3) the risk from all HGEs is estimated to be extremely low.
In assessing which sequences should be considered by HC0G, the staff also con-sidered the uncertainties associated with low- and high-oressure events. For low-pressure events the requirements of the rule force t nditions which are physically unrealistic (e.g. that the core be recoverable yet 75% of the 2ircaloyisoxidized). This results in sequences which are somewhat artificial and therefore considerably uncertain. For high pressure events these uncer-tainties are further complicated by a further lack of experimental data.
The staff therefore judged that it is sufficient to consider only low-pressure sequences because (1) the overall risk from HGEs is believed to be low, (2) from a risk perspective the reduced likelihood of a high-pressure HGE is likely to offset the potentially higher consequences of such an event, and (3) the additional uncertainties associated with high-pressure event pre.ession.
5.2.1.1 The HCOG's Base-Case Scenario The base case scenario proposed by HCOG results from a transk aused by loss Mark III SER 19
g.. .. ,
of offsite power, subsequent reactor scram, MSIV and containment isolation, and one 50RV. All ac-powered reactor makeup systems are assumed to fail initially.
However, the reactor core isolation cooling (RCIC) system which is de powered and/or the fire truck diesel supply are available at Grand Gulf. Emergency operatira procedures require depressurization when the water level reaches the top of ti.e active fuel region and some low pressure injection system is avail-able. It is assumed that following depressurization, the low-pressure systems fail to inject. The core becomes uncovered and core temperature begins to rise at about 35 minutes into the transient. As the core temperature continues to rise some hydrogen is produced. At about 65 minutes into the transient, the core is assumed to be reflooded at a high flow rate. During the reflooding of the core, large amounts of hydrogen are produced and transported to the con-tainment through the safety relief valves. At this point in time, the core has reached the recoverability criterion (i.e., the Zircaloy melt fraction is at about50%).
A variation to the 50RV sequence is the SBLOCA scenario resulting from a hypo-l thetical drywell break. The drywell break is essentially the same as the 50RV sequence except that the hydrogen is discharged into the drywell as well as i
through the safety relief valves (SRVs). The staff considered this conservative
- sequence in order to evaluate the effect of hydrogen burning in the drywell w% re essential control equipment cabling is found.
5.2.1.2 Station Blackout and NUREG-1150 The results of the GESSAR-II PRA (NUREG-0979) also found that 580 is a dominant contributor to the probability of core damage, although the core damage probability is quite low. These results were reinforced by Grand Gulf findings docutanted in NUREG-1150. Subsequently, the HCOG submitted infomation to account for 5B0 to the staff by letters dated January 8 and September 9,1987.
The results of analyses of Grand Gulf documented in NUREG-1150 indicate that the most probable HGEs result from $B0. The.most likely of these sequences i
(designated as TBU sequences) consists of loss of offsite power followed by the failure of onsite ac power in divisions 1 and 2 the failure of high-pressure core spray (HPCS) and reactor core isolation cooling (RCIC), and depressuriza-tion of the reactor vessel. TBU represents more than 90% of all HGE sequences and more than g3% of all low-pressure HGE sequences. The development of the loss-of-offsite-power (depressurfred vessel) sequence begins by boiling off the entire reactor vessel coolant inventory. With the core dry but at pressure, the operator depressurizes the reactor to increase the length of time available
! to support core recovery before the initiation of core damage. Following vessel l
depressurization the core begins to hetop causing oxidation of incore Zircaloy and core damage. Before core damage progresses to a point where a nonrecover- .
able core geometry could develop, a reactor vessel reflood system is assumed to l be recovered. This reflood system then covers the fuel region with water termi-l nating the event with a degraded, but recoverable, core geometry._
5.2.1.3 TBU the " Acceptable Sequence" l
The HCOG had considered the applicability of the various significant sequeni.es i identified in the draft version of NUREG-1150 to the HCOG program. The err srios were divided into three categories; the short-tem (about I hr) damage sai i TBU, TBUX, TCUX, the intermediate (4-6 hrs), and the long-term (8-10 hrs) L -
quences TB, TBUI, and TQUX (HGN-123). Differences, however became apparen l Mark III SER 20
when the results of the revised NUREG were reviewed. The revised version of
' -NUREG-1150 estimates TBU to be the dominant HGE sequence, which accounts for i j 935 of core damage frequency. The phenomenology of the TBU sequence is ,
similar to that of the MCOG base case with regard to the 50RV. However, the 1 NCOG experimental tesd ng and analyses, which encompasses the T8U sequence, assumed the ignitors were continuously powered i .
of the transient when ac power was not available.ncluding during the portion .
I i
e i
In addition, the rule requires that the containment structural integrity and a safe shutdown be established and maintained. The ability to satisfy these l requirements depend on both the total amount and the rate of hydrogen production.
To estimate the maximum hydrogen production rate and the total amount of hydro- l gen produced, the rate of water supply in the recovery phase of the HGE is critical. For purposes of the hydrogen control rule, the T8U sequence as described in NUREG-1150 (which encompasses the 50RV as described by the HC00) is an acceptable sequence leading up to core recovery. In summary, the TBU is l acceptable for the time sequence of events and for the hydrogen production rate and total amount.
l (In probabilistic risk assessment notation, the tems TB, T8U, TBUX,f offsiteT8U1,TCUX '
and TQUX denote the following: TB - Station Blackout. TBU - Loss o power (LOSP) with failure of all high pressure functions. The SRVs are opera-tional and the vessel is depressurized. TBUX - LOSP with loss of all AC divisions and failure to depressurize. T8U1 - LOSP with failure of AC divisions 1 and 2 and of the high pressure core spray (HPCS). The reactor core isolation cooling (RCIC) operates for 6-8 hours before failing due to high pressure in the tubine exhaust. TCUX - ATWS with LOSP loss of AC and HPCS and RCIC fail-ures. TQUX - Failure of all ECCS functions except power.)
5.2.1.4 Hydrogen Generation Profiles Recovery of cooling water flow is effectively bounded between 150 gpm from a single control rod drive cooling pump to 5000 gpa from the emergency core cooling system (ECCS) low pressure, high-flow-rate core recovery system. The hydrogen generation profiles for these extremes are qualitatively and quantitatively different. The probability of a high-flow-rate recovery is expected to be higher than that of a low-flow-rate system, because there are more high-flow systems (or combinations of systems) to inject water into a depressurized vessel; hence, l it is reasonable to assume that the operator will attempt more often to recover one of the high-flow-rate systems. A high flow reflood rate is associated with a high, narrow spike of hydrogen release, while the low-flew reflood rate will yield lower hydrogen production rates but for longer times (see Figures 5.1 and
~
5.2) (HGN-132). These profiles have been estimated by the HC0G uslng the BWRCHUC l code, which i's discussed in Section 5.2.2. The total mechanistic estimated amount of hydrogen released in the low-rate reflood case is higher than that rf. eased in the high-rate reflood case. The hydrogen peak release of the high-l of the high-reflood case is about 35 seconds wide at half maximum, while for the low-reflood case significant hydrogen release lasts about 8.5~ minutes.
Table 5.1 shows a summary of the main features of both cases.
Mark III SER 21
Table 5.1 Hydrogen Release Profiles Length of Total Peak Width at Reflood Transient Hydrogen Rate Half Max Rate (gps) (sin.) (1bs) (1bs/sec) (sec) .
150 80.0 903.4 0.95 510 5,000 25.8 604.8 8.00 35 5.2.1.5 Non-Mechanistic Hydrogen Release Profile The hy".;rogen rule requires consideration of metal-water reaction (WR) for 75%
of tk.e Zircaloy cladding surrounding the active fuel region. However the esti-mated amount of metal-to-water reaction in either reflood rate scenarlo is far less than the required 75%. For Grand Gulf, the active core region cladding is 79,100 lbs. For the oxidation to proceed as: Zr + 2HsbIr0 + 2Ha, the amounts of Zr that correspond to the hydrogen released in Figures 5.1 and 5.2 are 20,450 lbs and 13 700 lbs, which represent about 26.0% MWR and 17.0% MWR re-spectively." (ThIs WR incorporates channel box and stainless steel oxidation.)
A mechanism was needed to increase the release profiles to 75% WR of the active core region cladding, as required by the rule. The 75% WR of the Zr in Grand Gulf is 59,300 lbs, which when oxidized will create about 2,600 lbs of hydrogen.
It must be pointed out that mechanistic models that account for 75% MWR of clad-ding oxidation result in a severely damaged core exceeding the recoverability criterion. There are many possible scenarios that can be hypothesized to yield 75% WR cladding oxidation; however, no attempt is made to estimate the phenos-ena associated with such an oxidation level because it would require an unreasonable recovery criterion.
the As discussed previously, after core quench'ing calculated maximum amount of metal-water reaction is limited to about 26%.
in the low-rate reflood case, To meet the rule requirement of 75% WR, the HCOG submitted a nonsechanistic model used to predict hydrogen production based on an energy balance in a severely damaged core. It assumed that such a core has energy losses at least adequate to remove decay energy in the core, the energy produced by continued oxidation and excess stored energy in the core. It also assumed that termi-of Zircaloy,idation nation of ox at 75% WR takes place by quenching of the core and removal of all excess energy (HGN-034). Considering the above, the cxidation rate will support a constant hydrogen release of about 0.10 lbs/sec. The staff finds that this release rate is acceptable for hydrogen rataase to 75% MWR, as required by the rule.
- Therefore, for the scenarios shown in Figures 5.1 and 5.2 the " tails" 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. _
The staff concludes that (1) mechanistic models can not predict the required 75% MWR of cladding oxidation in the active fuel region without core damage
- Zircaloy is assumed to consist of 100% Zr. The actual composition of Zircaloy-2 in weight percent includes .Sn: 1.2-1.7, Fe: 0.02-0.07, and Ni: 0.05-0.15 Mark Ill SER 22
.,_e.
2
. n.. '
I
}.
i e
i l o'
/ 'i -
1
- i .' -
g
.u . . E
. i .. g i n, ,
e.
5...f ~
P b 3 g e
)l ';. .NP- I E K- < '~.
g a hi - !
=
4i
-i i
y; a '
. -E -
a i
r .n g L" .
t i i i i i i i i _
N
- e. m o a e e e N = s
( wepql ) 21W No11 WIN 30 N2OOWONH Mark 111 SER 4f 23 ,
M _
1 c
. / :
l
, ~
c ,i L .
,t u. ,..
l w } t
- p. ;
l
<,..j i _
r*
. g _G y .
- ~? -
Z =>: <-
t >
l i -F Eli <
C:
i
, e. a 7:
- i g
j I
L 4
g W [. ..
E:
t y ,. l i
- (s.N.x m
,, ?!
r.
g g.
T Ly .i - i ;
Z E;i s
's o
r
- w. ,yi , .
N -
.R s,. .
~
a
- g._ .
C s
I.
1 .
\
. , , , , , i . ..
-. a g 3 *
- 3 ; i $ i 2 h
( we,wel ) 3J.YW NouYWIN'30 N3DOWQ.U4 Mark 313 SER @ 24 -
,.w..---.s-- .,-.%.,,y. . .._,.,- - . - ..g, ~.. . . - . . . _ _ . , , _ _ , . . _ , ,-._...-..,,,-,,_.,,,..y...
o . . .
beyond the recoverability criterion and (2) the use of a non-mechanistic release model based on heat balance is reasonable and acceptable. This leads to an oxi-dation rate producing 0.10 lbs of hydrogen per second .
about4.7and5.6hoursforthescenariosofFigures$.requiringanextensionof 1 and 5.2, respectively.
5.2.2 BWR Core Heatup Code (BWRCHUC) 5.2.2.1 Introduction The BWRCHUC has been used by the HCOG to calculate hydrogen rate release profiles for the hydrogen generation sequences described previously. The areas the staff considered in the review of the BWRCHUC is discussed below.
The BWRCHUC was not validated or benchmarked to global core experimental data, rather it relies on collective engineering judgment and understanding of the phenomena taking place in a core disruptive accident. The lack of benchmarking or, validation is due to the abser.ce of suitable experimental data. This lack of benchmarking prevents the wite af t% code from being used directly without appropriate consideration af s E Oc input parameters. The results of the BWRChdC should be seen as Tn engineering estimate of the anticipated phenomenn.
the code revisw was aimed at the reasonableness of the modeling, Accorcinglyl the ph(sica significance of the assumptions, and possible conservatises in the estima te. Reasonableness was assessed it, terms of models and hypotheses that have been advanced by other researchers in this field and any other evidence that cou'd be gleaned from whatever limited and partial experimental informa-tion was available. For code modeling, the TBU sequence for an HGE was consid-ered equivalent to the 50RV sequence (paragraph S.2.1.4) with respect to the depressurization and core uncovery time, thus similar as far as hydrogen generation is concerned. This sequence is the simplest and most straight-forward, thus having the highest probability of being modeled correctly.
The BWRCHUC is a well-written computer code in that (1) it faithfully represents the BWR geometric core design and (2) the models included in the code are ade-quate to cover the specific HGEs selected for analysis by the HCOG and repre-sented by the TBU sequence. Modular architecture has been used extensively, where each module (subroutine) in the code treats a different phenomenon or as-pect of the problea. The code is built by connecting the various modules with executive routines. The numerical solution technique applied in the BWRCHUC is apparently as good as any employed in severe accident codes. Numerical stability, as reported by HCOG, is evidenced by the graphs of code output and the fact that reflooding calculations can be run.
Some subtleties,of the The BWR core geometry is very complex.the potential' to Therefore, affect the prediction it of hydrogen is ap)ropriate that a best-estimate code contain a representation of the geome-try t1at is as complete as is reasonably achievable. This has been done in the BWRCHUC. Considerable attention has been drawn to the fact that the BWRCHUC allows for a different two-phase water level to be predicted in each fuel assem-bly represented. A separate level for the core bypass level varies according to the bundle power since the water in an assembly is assumed to be saturated at the system pressure. Water levels may also vary because the void fraction of Mark III SER 25
o, .
the water in an assembly is a function of assembly power. The bypass level cal-culation further assumes that water in the bypass is subcooled and thus corres-ponds to the collapsed water level in the core. There is a hydraulic connec-tion between the assemblies and the bypass so that the water level in the bypass is reduced as the core water inventory is boiled away by the decay heat generated within the assemblies. This representation closely corresponds to a partially covered BWR/6 core at low pressure before any structures in the core reach tem-peratures significantly greater than saturation.
5.2.2.2 Phenomenological Assumptions A model for channel blockage was included in the code, but has not been employed in the calculations since experimental results did not support total flow blockage. The blockage model assumed that the fuel rod cladding melts while the channel box remains intact. Molten cladding is then assumed to slump and refreeze within the channel forming a complete blockage, which prevents steam from reaching the Zircaloy surfaces within the assembly. In addition, steam generation below the blockage pressurites that portion of the assembly forcing the two-phase level in that assembly below the core plate. Since no steam enters the channel, all oxidation would stop. Experimental results from the PBF tests (HCOG presentation to NRC January 14,1985) indicated that a reduction in the flow area as a result of Zircaloy slumping did occur, but that complete blockage did not form. Without the channel blockage model hydrogen production is maximized all other conditions being the same. The lack of clad motion or channel blockage is a very conservative assumption with regard to bydrogen production.
It is assumed that the control rods will remain intact since it ::, consistent with the recovery criterion. Under certain conditions, experimental evidence (R. 0. Gaunt) suggests that BWR control rod blades could melt early in the core heatup phase of a transient. This would lead to the possibility of local loss of control; thus, when the core is reflooded, local criticolity could result in intense heat production and core damage beyond the limits of recoverability.
Therefore, control rod melt would be beyond the scope of this program.
5.2.2.3 Steam Production The modeling of steam generation within the reactor pressure vessel (RPV) can have a significant impact on the quantity of hydrogen generated. While some aspects of steam generation are accurately modeled in the BWRCHUC, other sources of steam are not modeled at all. The steam generation modeling generally is for the most likely HGE considered, the steam sources not incomplete; represented dohowever,ignificantly not s impact production of hydrogen.
Within the B9RCHUC, the following five sources of steam generation are modeled:
. deposition of the decay power from that portion of the fuel assembly celow the two phase level into the saturated water within an assembly
+ heat transfer (by nucleate boiling) from portions of fuel rods, channels, control blades, and the core shroud that are at temperatures greater than saturation when they are covered by the two phase level Mark III SER 26
- radiative heat transfer from portions of control blades that have tempera-l tures higher than saturation to surrounding channel walls when the two phase level within the channel is at or above the portion (at elevated tempera-tures) of the control blade
- flashing of water in the downconer and lower plenum as a result of reduc-tions in the RPV pressure (Pressure-time history is provided by user input.)
evaporation of core spray droplets entering the top of a fuel assembly during reflooding of the core Steam generation resulting from flashing of the water inventory within the fuel assemblies and in the bypass region is not modeled. If system pressure would
, decrease, flashing would occur. However, the selected HGE sequence does not involve changes in pressure vessel pressure after hydrogen generation has begun.
It is assumed that the RPV pressure is constant for at least 10 minutes, which is the time required to remove the bulk of the heat in the lower plenum struc-tures. Therefore, the lack of a flashing model is not a factor.
l Downward relocation of molten Zircaloy can have a large effect on steam genera-o tion. If the two phase level is above the core support plate, molten Zircaloy I can run into water. Quenching or relocating Zircoloy in water would enhanc6 steam and hydrogen generation. This phenomenon is not modeled in the BWRCHUC.
However, there is no water above the core support plate when Zircaloy melting occurs. Modeling of the melt relocation into the water would not increase the i quantity of hydrogen produced compared to that which will be produced in the core reflood because of the more favorable surface-to-volume ratio. l An oxidation cutoff temperature is used in the BWRCHUC as a surrogate for the i effect of cladding and channel box relocation and subsequent quenching thereby l removing the Zircaloy from the oxidizing environment (R.0. Gaunt and HCOG pre- !
sentation to NRC January 14,1985). The HC0G estimated Zircaloy oxidation vs. I Zircaloy temperature and concluded that 2400'K is a conservative representation to account for this effect (HGN-032, item 4). Based on the evaluation performed by ITS, the staff has accepted the 2400'K as the irreversible oxidation cut-off temperature (lettertoHCOGJune4,1985). I In the reflood stage, quenching of Zircaloy that is at temperatures higher than the saturation temperature is nonmechanistically estimated. This can lead to overprediction of the steam generation rate during the reflood phase. For nodes that are more than 100'K above saturation, quenching is assumed to take place in a single time step thus accelerating the heat transfer proce't and steam production. ,
Steam flow in the bypass region is underpredicted. However, the effect of this underprediction of the e/ pass steam flow rate on the overall prediction of hydrogen release is small. Overall steam generation rates in an overheated core could be underpreoicted for transients in which the two-phase. level is above the core support plate. Inthestaff'sjudgment,theextentofthis underprediction is small compared with the uncertainties associated with predictions of this nature.
Mark III SER 27 )
i
.. ...t .
. 1
. \
In summary, steam generation before reflood is reasonably well predicted provided ;
the RPV pressure has been constant for approximately 10 minutes. In the sequence '
considered, the RPV is depressurized and steam and hydrogen production take I place under conditions of constant pressure.
5.2.2.4 Hydrogen Generation l 1
As with the modeling of steam generation, the approach to modeling hydrogen generation is reasonable considering the difficulty of representing the phenos-ena to modelling techniques. The lack of models for a few relevant phenomena combined with some of the assumptions made for phenomena that are modeled, leads to some uncertainty with regard to the predictions of the hydrogen generation rate during the dominant HGE. This uncertainty is expected to be negligible (i.e.. nessess compensating effects) in the present context. Howsver, consider-
, ing the censervative assumption of no clad motion, we concluded that the overall hydrogen ganeration estimate is conservative.
, The considerations / phenomena that are related to hydrogen generation and are not modeled or are underpredicted are listed below. >
1
+
0xidation below the location at which melting occurs is not modeled.
Because of the underprediction of steam in the bypass channel, oxidation of stainless steel and the outside of the channel is probably underpredicted.
1 Ballooning of the cladding and localized failure resulting in simultaneous interior and exterior oxidation is not modeled, thus limited hydrogen underprediction may result.
Film boiling in a quenching mode is not modeled. This leads to higher rate of hydrogen production for shorter time periods. It is not clear that an overall underprediction will. result.
' +
In the reflooding stage, vaporization of droplets that enter the top of fuel assemblies by radiant heat transfer does not remove heat from fuel rods. This results in a conservative hydrogen production if the maximum temperature is below the cut-off and possibly not conservative if it is above the cut-off. It is not clear if the overall effect is nonconservative.
Reaction rates of Zircaloy and stainless steel with steam are calculated using the Arrhenius relationship. The reaction rate constants used in these expres-sions were derived by others from experimental results. This modeling of reac-tion rates and the associated heat generation is appropriate and consistent with what is used'in other severe accident modeling codes. A hydrogen blanketing fac-tor is included in the formulation of the Arrhenius reaction rate expression.
Hydrogen blanketing refers to the possible limitation of the oxidation rate from the diffusion rate of steam through the hydrogen emitted from the oxidizing sur-face. While the process represente6 by the hydrogen blanketing factor is real, a reduction of the oxidation rate is almost certainly not real17ed under the conditions expected during core damage in BWRs. Diffusion of steam through the oxide layer is the rate-limiting process. Therefore, the hydrogen blanketing effect was not considered in the HC0G calculations, which represents a slight conservatism.
Mark III SER 28
?
.- - . - - . . ~ - - . _ - . _ - _ . -.- - - . - -
~
Because the oxidation rate varies exponentially with temperature, the represen-tation of intact Zircaloy nodes reaching temperatures significantly higher than the melting temperature leads to hign.r oxidation than would be predicted
- if melting were explicitly treated. Therefore, this is a conservative assump-tion. However, because the oxide layer is generally thick at these times, the actual quantity of additional oxidation is considered to be small. One could view this enhanced oxidation es a r.cnetchar,istic approach to representing the initial enhancement in oxidation that probably accompanies slumping molten 21rceloy.
i reduces the tensile strength and increases ductility.
Heatingofthecladdinfthefuelandgaseswithinthecladdingleadstopres-Simultaneous heating o l surization of the rod from within. Bellooning of the cladding and localized L failure may occur before melting. Failure of the cladding would allow the in.
I terior surface to be exposed to steam. It is therefore entirely possible that both the interior and exterior surfaces of the cladding will undergo oxidation.
Since this possibility is not modeled in the BWRCHUC, hydrogen generation rates and total hydrogen generation could be underpredicted. However, in the staff's ,
judgment, the conservatism in the assumption that there is no clad slumping will '
, adequately compensate for this potential underprediction.
l Overall, it is the staff's judgment that the modeling of the hydrogen genera-tion rate in BWRCHUC is reasonable and the total and peak hydrogen production estimates are expected to be conservative.
5.3 Sumary and conclusions There was a twofold objective in this portion of the evaluation corresponding l to two requirements of 10 CFR 50.44 l l l
(1) "Use accident scenarios that are accepted by the NRC staff." Paragraph I (c)(3)(vi)(B)(3). ,
j (2) " Provide an evaluation of the consequences of large amounts of hydrogen '
i generated after the start of an accident...up to and including 75% of the l uel cladding surrounding the active fuel region...." Paragraph (c)(3)(vi)(B)(1).
The first requirement corresponds to the dominant a.ccident sequence that leads to an HGE. The information for such a sequence was derived from HCOG submit-tals and conforms to the revised (final) version of NUREG-1150. The TBU se-quence was found to represent 93% of all HGE sequences and consists of loss of offsite power followed by failure of onsite ac power in divisions 1 and 2 and failure of the HPCS and RCIC. The TBU sequence was found to be similar to the 50RV, which was initially proposed by HCOG. Thus, for hydrogen generation purposes the TBil sequence satisfies the requirements of 10 CFR 50.44.
The second requirement is limited to the acceptability of the BWRCHUC to estimate hydrogen generation profiles. The essential characteristics of such profiles are the peak rate, its duration, and the total amount of hydrogen produced.
l i
Mark III SER 29 i
_ __ _ _ _ _ _ _ _ _ _._ _ _ _ ~ _ _ _ _ _ _ _ _ _ _ _ ,
. The staff reviewed the SWRCHUC code on the basis of these requirements and information submitted by the HCOG. The staff finds the SWRCHUC code acceptable for ree in calculating hydrogen production profiles. Therefore, the staff finds that the profiles estimated by HC0G using the BWRCHUC are acceptable for use in demonstrating compliance with 10 CFR 50.44.
6 CONTAINMENT RESPONSE - ANALYTICAL MODELING In view of the quarter-scale test program, the emphasis on analytical methods for predicti g containment response to hydrogen N rning has significantly di-minished. This ccnclusion is based on the broad iange of hydrogen release
, rates in which diffusive combustion is expected to occur. The staff believes the evaluation of the survivability of essential equipment should be based on the QSTF data. Therefore, there is limited value in pursuing such analytical meth-ods as the CLASIX-3 code thus, the following evaluation addresses HC0G's of-forttoim1vetheCLASIX-3codeanalysisgenerically. As such, this effort is only relevant at low hydrogen flow rates that are near the flame extinguish-ment limit.
As documented by various staff evaluations perfomed before the completion of the quarter-scale test program, the CLASIX code has been the principal analyti-
- cal tool in predicting the centainment response as a consequence of burning
' hydrogen for plants with an ice condenser or Mark III containment. The CLASIX code or the CLASIX-3 code (which is the latest modified version that includes Mark III containment features) deals with deflagration (discrete-type) hydrogen burning. The code is a multivolume containment code that is used to calculate the containment pressure and temperature response in separate compartments.
Moreover, the code has the capability to model characteristics that are unique to Mark III containments while tracking the distribution of the atmospheric constituents (i.e., oxygen, nitrogen, hydrogen, steam).
The staff's desire to demonstrate verifica' tion / validation of the CLASIX code has been an extensive effort. CLASIX results have compared well with results of other NRC-accepted analytical codes and hydrogen burning experiments. Fur-thermore, the HCOG has performed additional code validation by comparing the more recent Nevada Test Site large-scale hydrogen experiments to CLASIX-3 code predictions. This is documented in HC0G's letter (HGN-113) dated January 8, 1987. With regard to hydrogen burning, the focus on code validation has been on pressure predictions because temperature comparisons are more difficult to predict due to their time and spatially dependent fluctuations.
As discussed earlier, the major element of the R0G's program is the quarter-scale test program. The data obtained from ti .; were used to perform equip-ment survivability analysis (see Section 7). These tests revealed that diffusion flames on the suppression pool surface can exist at a hydrogen injec-tion rate as low as 0.02 lb/sec under certain background conditions. As such, it is expected for a significant portion of postulated degraded core hydrogen profiles that diffusion flames will be the dominant combustion mode. Since CLASIX-3 does not model diffusion flames, these results have a significant bearing on the extent to which the CLASIX-3 code can be relied upon to predict containment temperature environments, which further emphasizes the importance of the test program.
Mark III SER 30
- - 6.1 Localized Combustion /CLASIX-3
, As noted in Section 3.1, testing perfomed in the QSTF revealed that combustion which occurred bel w the flame extinguishment limit was not global deflagration-type combustion; but " localized combustion." Localized combustion is character-ired as weak flames or weak volume burning through a marginally combustible hydrogen air-steam mixture. By letters dated June 10, 1986 (HGN-092-P), and December if,, 1987 (HGN-111-P), the HC0G provided various analyses to demonstrate that the CLASIX-3 model provided a boundary calculation for the combustion occurring below the diffusion flame threshold. This model employed a combustion mechanism that produces a more severe global thermal environment than has been i measured locally in the QSTF for locallred combustion.
The staff requested Sandia to review this approach. In its NGN-092-P submit-tal, the HCOG compared CLASIX-3 predictions for a quarter-scale model to exper-imental results of the corresponding test. The CLASIX-3 predictions of the wetwell volume showed that the temperature profile a::ceeded the volume-weighted
- average of experimental data. On the basis of this analysis, the HC0G conclud-I ed that CLASIX-3 yields conservative predictions of themal environments inside
, containment for very low hydrogen flow rates. The staff questioned the appli-cability of using experimental data pertaining to a local phenomena to denun-strate the capability of a lumped volume code. SNL shared the staff's concerns and recommended that the local combustion phenomenon observed in the QSTF warranted further evaluation.
In its HGN-111-P submittal the HC0G provided a comprehensive assessment of localized combustion seen In several quarter-scale tests. In these tests com-bustion activity, as evident b periods of low hydrogen Test flow.data y thermocouple responses, do not indicate that concentratedis widespread d flame energy deposition occurs at fixed locations. Energy deposition appears to be rapid and diffuse and is dominated by con?ective mixing, combustion-induced turbulence, plume influence, and background gas flows. Typically, peak temperatures recorded during localized combustion are relatively low and persist for short durations. The temperature responses are cyclic and return to relatively low background levels.
The HCOG analyzed five quarter-scale tests conducted during the scoping test phase of the program in an effort to better understand localized combustion.
At the low hydrogen flow rates, these tests demonstrated certain repeatable trends and the thermocouple activity observed recurrent and generally predict-able. Some of the findings resulting from the evaluation of localized combus-tion ari briefly described below.
l (1) In tests without sprays, combustion was generally widespread. Whereas, I when sprays were activated, combustion appeared to be suppressed in open l- chimneys (i.e., annular quadrants) as a result of cooling effects and l shifts in global flow patterns. Also, enhanced mixing resulting from I
sprays induced slightly higher temperatures in some areas, but not appre-ciably higher than those recorded when sprays were off.
, (2) Comparing the scoping test results, it appeared that the location of the 50RVs did not have as significant an effect as other parameters, such as variation in hydrogen flow rates.
Hark III SER 31
(3) Probably the most important finding pertained to combustion activity in the vicinity of the hydrogen igniters. The closest thermocouple located to e nearby igniter was about 15 inches laterally or about 6 inches later-ally and 15 inches above the igniter. Data indicated that temperatures close to igniters were generally no more severe than recordings several feet away. Also whe toopenflowregIons,ncomparingtheeffectsofblockagesaboveignitors a significant increase in temperature was not ob-served. Combustion energy dispersion was prevalent.
To further support its findings of turbulent sixing induced by combustion, the HC0G included a discussion of a test in which pool flames were observed. A thermocouple was placed about I foot directly above the pool sudce over an active sparger. Readings indicated that at low hydrogen flows the fismes at this location appeared to be intermittent and unstable. The temperature re-sponse did not exceed 425'F as a result of these unsteady pool flames. The HCOG contends that, because of the efficient mixing, one should expect local
' hydrogen concentrations elsewhere in the facility to be less than at the sup-pression pool surface. Moreover, this fact coupled with temperature readings (discussed above) and the absence of visual indications regarding flame forma-tions above the HCU floor, is not strongly supportive of a hypothesis that sus-tained high-temperature localized combustion zones will be established at very low hydrogen flow rates. In addition, H00G indicates that the temperatures generated from pool burning at low hydrogen flow rates (i.e., about 0.15 lb/sec) than from resultant localized combustion. hot plumes represent a more severe thermal environment As part of this assessment, HCOG provided 6dditional information with regard to the role of the CLASIX-3 code in its analyses and the conservatisms used for containment modeling. While global or large volume deflagration, as modeled by CLASIX-3, did not occur in the QSTF, the HCOG contends that the CLASIX-3 model-ling would conservatively bound the observed localized combustion environment.
To assess the severity of the environment from an equipment survivability per-spective, the HCOG compared the thermal loads created by the most severe local-ized combustion measurements at the OSTF to the corresponding CLASIX-3 temperature <
profile. TheresultsofthiscomparisonshowtheCLASIX-3profilegeneratesa significantly more severe environment than that produced by localized combustion.
The staff requested that SNL review this issue along with the consideration of scaling aspects, 5NL determined that the HCOG ader,uately addressed t M likely -
locations for localized combustion and identified teasonable bounds for thre most threatening thermal environment for equipmer.<. located near regions of lo-calized combustion or in the resultant hot plume . Moreover, the thermal local comparison in combination with the HCOG's discusaion of localized combustion provide adeqdate justification that the CLASIX-3 tnermal load would be more severe than that experienced in the QSTF for low avdrogen injection rates. In conjunctionwithdiscussionscontainedintheQSTFtestreport(HGN-121-P) SNL assertsthatthereisreasonableassurancethatthethermalresponseatfull scale will be no more threatening than that experienced in the QSTE. The staff has also evaluated this issue and concurs with SNL's assessment. Based on the modeling methodology used in the referenced submittals (e.g., low hydrogen flow rates and the focus on the wetwell profile), the staff finds that the CLASIX-3 predictions would be acceptable in determining the containment environmental conditions as a consequence of locrJized hydrogen burning. Accordingly, these profiles could be used to evaluate the survivability of equipment.
Mark III SER 32
*7* .. ,
6.2 Containment Pressure and Temperature Calculations Letters HGN-092-P and HGN-109-P documented the HCuG's calculations of the con- -
tainment pressure and temperature response based on postulated degraded core
- scenarios that are discu. sed in Section 5 using the CLASIX-3 code. To deter-i eine the adequacy of the hydrogen ignition system (HIS), the NCOG considered -
I two types of accidents in its generic analysis: a stuck-open relief valve l
(SORV) transient and a small-break loss-of-coolant accident (SBLOCA) in the
_drywell. The component of the hydrogen release history that is of interest in this analysis is referred to as the tail" portion and represents a nonsechanistically defined constant hydrogen production rate. As discussed above, the CLASIX-3 results bound the thermal environment that may be produced for low hydrogen release rates that are below the diffusion flame extinguish-ment limit.
.HCOG provided a generic sensitivity study using the Perry Nuclear Power Plant L containment characteristics for the LLMIX-3 model. In this sensitivity study, different parameters were varicJ to assess the effects on the calculated re-sults. The staff focused on the most important parameter considered, which was the assumed availability of the containment sprays.- HCOG had chosen to use the CLASIX-3 code predictions without sprays (i.e., for the 50RV case) in its ge-neric survivability study (discussed in Section 7). For the equipment surviv-ability analysis to t.? generic, it became necessary to consider the no-spray case because fan cooler: rather than sprays are part of the River Bend con-tainment design. -
For the 50RV case, all mass and energy releases were directed into the suppres-sion pool. The CLASIX-3 e 6 1 used in the generic analysis simulated four com-partments of the Perry contak: tant: the drywell volume, the wetwell volume (bounded by the HCU floor and the surface of the suppression pool), the inter-mediate volume (bounded by the HCU floor and the refue' ling floor), and the con-tainment volume (above the refueling floor). Figure 6.1 presents a schematic representation of the model. Ignition of hydrogen combustion was assumed to occur at a 6% hydrogen concentration with 65% combustion completeness. The CLASIX-3 50RV base-case model produce a transient in which the hydrogen was ignited in a series of burns in the wetwell volume. Figures 6.2 and 6.3 show the computed wetwell temperature and pressure profiles. The early portion of the transient resulted in the highest wetwell temperature. This is symptomatic of the hydrogen spiked release in the early phase of the release profile. Dif-fusion flames would be prevalent in this interval and would be beyond the range of use for the CLASIX-3 methodology. For the major portion of the temperature profile, the wetwell burr.s produce a peak wetwell temperature of d ove 800*F.
At the end of the hydrogen release period, the calculated hydrogen concentration in the containment volume did not reach the ignition criterion of 6%. In the CLASIX-3 calculation, the HCOG assumed a containmeni burn to occur at this lower concentration which resulted in the most severe presnre excursion, to approxi-mately 23 psig.
.6. 3 Drywell Analysis For the base-case analysis of a small pipe break in the drywell (DWB) the CLASIX-3 containment model was similar to the SORV case except the hydrogen / steam source MarP III SER 33
. .- 1
.V ..
r .
terms are directed to both the drywell volume and the suppression pool. The l DWB'scanario was chosen because of the potential and consequences for hydrogen .
combustion in the drywell. Of the cases studied, only the 2-inch DWB case had ,
conditions where a hydrogen burn was predicted to occur. Ignition in the dry-well was limited by lack of oxygen (i.e., below 55) because air is forced frem the drywell by vessel blowdown. The only burn predicted by CLASIX-3 for the .
2 inch DWB case resulted in a peak drywell temperature of about 1050'F and a peak drywell pressure of about 13 psig. For the events considered, where high steem flows are directed into the drywell along with the diversion of most of the hydrogen to the suppression pool, the hydrogen threat to the drywell appears l, to be relatively small. As indicated in Section 7. HCOG performed thermal re-l sponse analysis of selected drywell equipment. The results demonstrated that L the equipment would survive the drywell burn.
I 6.4 Existence of Drywell Diffusion Flames In the DWB case, air would be reintroduced in the drywell throu h vacuum breaker ;
ectuation or operation of the drywell mixing system. The drywe 1 environment '
is predicted to be e hydrogen-rich / oxygen-lean mixture. When oxygen is re-introduced in the presence of an ignition source, a diffusion flame may result in the vicinity of the oxygen source. This possible combustion phenomenon is referred to as an inverted diffusion flame. This is a concern since the poten-tial to establish a continuous inverted diffusion flame at the oxygen source may result in locally severe thermal loads.
By 1stters dateo June 25,1986(HGN-091),andJune 10,1987(HGN-119),HCOGeval-I usted;the potential impact of inverted diffusion flames in plants with a Mark III containment.- In the HGN-091 submittal, the HCOG discussed the criteria for establishing the existence of inverted diffusion flames. The HCOG indicated E
that flames will not occur in the drywell when conditions are outside the flam-mability curve. In the HGN-119 submittal,.the HCOG further discusses the low likelihood of achieving the necessary combustible conditions in the drywell based on the CLASIX-3 predictions. -
SNL reviewed the initial submittal and determined that the HCOG did not erovide sufficient justification to preclude drywell burning. Specifically, SF com-mented that the flammability limit merely establishes the limits that . 11 allow flame propagation; burns that do not propagate inte the mixture are not precluded by being outside the flammability limits. Furthermore, it was not obvious that the burning mixture should be expected to follow the path predicted by the HCOG.
Generally, there is a lack of experimental data to support the HCOG's position.
However, recdnt risk studies do not support the DWB case as a dominant core-melt / degraded-core event for plants with a Mark Ill containment; therefore, fur-ther phenomenological investigation may not be warranted. Drywell break events are further discussed in Section 5. In addition, the expected redundancy (i.e.,
spatial separation of equipment performing the same function) of the critical equipment should compensate for possible locally severe thermal loads.
The staff believes there is a reasonable level of assurance that the consequences of a drywell break event would not pose a significant threat to containment integrity and would not preclude safe shutdown of the plant. However, the staff believes, as part of IPE process that each licensee of a plant with a Mark III Mark III SER 34
g -. -
gr ., ,
1
- e. .,. .:
4 1 r- PERRY CL ASIX-3 MODEL ruuminmr-il
' C ONTAINMCNT
( v0L. 4 ) ' '@ 7 1
l.
3 ,
o , j l -l
'. . Seas, i
l casa, oney e- l
- 0. i i*
, I y ( 'l vacwww - estastas g ORYWELL Ni IN T ER MC0l&TE I i w
l l-~
-(VOL.ll ,
etpass Lasaaoy gygg,33 '
l - -
l i
_ tocs
- ,J l l 1r d L -6 6 j l ,
l g i
Isomat i ,I g WPPra poet staaet even j h ',' , cue, s y
- e s
s l
m< l l-
+
4 M
J ,_.
=,_
-l
- s. -
f^ :.
. I ;[; , , , . , I , : T*--
- d . WCTWELL J-On
- y
, C S'UN R E S $ ION e w - f__ .
gyot :3 .J CHlWNEY .;
, .y -
Poot 2 q,-,-- ,
I 5_ d.
(;,,,.[4 j;
. Y},_ .
I j:,
\t [; i p-r Plow sa tota san gersons
,:~
(
S
- r Flow sa Omt De a t e f t 0*
i eeaaea M IP A AT 8I A OINI 1 Figure 6.1 Perry CLAS3X-3 Model 1 Source: HGN-092-P t.
(. Mark 333 SER M 37 l
1,,- ... -
._{i? j- e
- ~ .,
.L y 7. r + ; ,_
vy = .
.R s .
' l' _
.-g
- a
'~
_ 68%
C' 1 E . t-
__ e-
= . g J h,.
k-i C4k.
- __ L.. I 3 a g ,,
e-3e
. . ..g 1 _ eq
__ w T.
= g
=* E i i i e a i a i < i : i O
e p) N w .- e e b e ers e M N == 0 J J J J 6 6 6 6 6 6 6 6 6 -
(spumanowA)
(4) 3WnAW3dW31 Mark Ill 55R # 3I.
. ~ - ._ __ . _ _ _ . . .. _
i [.
- 1 c5
, , . r. 1. _ , ; , . ..
< ;o: #
- - . i f , .
p .)
L}
j u . l
-1 L.
!! .K .
L a
a
.I
-1
-.R 1
-1 l
- l
, 1 i.
.e = ,
1 18 , -
~
i
. 1
... w .-
S -
. 4 tg
~
l I e
- :3*..
~* .
?
d
~
- .e l
1
. 1 e I i i i i i i i O l o e o e e- e .o e- a w n 6 N N -
~
(ese) zwnssawd I
1 1
I Mark 111 SER M 37 .
l s
._--_-_--__.L_---- -- A__.----_---------- -- . . - - - . - - - - - - - e a--e -- w~- ,,-vr ~"-e-,
i =i.. M 's; ,
l
- - i .
'l containment should confirm the 19eation of critical equiment with respect to potential oxygen sources through the drywell vacuum brea cars and a drywell mixing system to support the above conclusion for each plant.
7 SURVIVABILITY OF ESSENTIAL EQUIPMENT I As part of the analysis,10 CFR 50.44(c)(3)(vi)(B)(5)(11) states:
Systems and components necessary to establish and maintain safe shut- >
down and to maintain containment Integrity will be capable of perform-ing their functions during and after exposure "' the environratal i conditions created by the burning of hydrogen, including the ,,(fect of
's ,
- detonations, unless such detonations can be shown unlikely to occur.
Accordingly, each licensee with a Mark !!! containment is required to demonstrate that the essential equipment located inside the containment will survive the . ;
hydrogen burn environment._ To support this objective, the HCOG conducted two l programs to ' define the environment' that would result from hydrogen combustion. (
L As discussed earlier in this evaluation, the quarter-scale test data will be L-used to define the environment that would be produced by diffusive-combustion on the' suppression _ pool surface. In addition, the CLASIX-3 code analysis will be used to define a bounding environment for localized combustion below the-L diffusion flare extinguishment limit. .
7.1 Identification of Essential Equipment r
The equipment that has to survive hydrogen burning was selected on the basis of i function during and after a postulated degraded core accident. Generally, all the equipment located in the containment that was considered to be in one of the five categories listed below was considered to be essential for the safe l shutdown of-the plant.. .
(1) systems and ccaponents that mitigate the consequences of the accident (2) systems and components needed for maintaining the integrity of the l containment boundary L (3) systems and components needed for maintaining the core in a coolable
! geometry.
l (4) systems and components needed for monitoring the course of the accident and proViding guidance to the operator for initleting action in accordance with emergency procedure guidelines D
(5) components whose failure could preclude the ability of the above systems to fulfill their intended function -
Using these criteria, the HCC3 identified the equipment that would be needed to l-be evaluated for survivability. In its letter (HGN-084) dated May 16, 1986, the L
HCOG transmitted to the staff the list identifying the following system / components: ,
t-l.
L
. Mark III SER 38
^
- ;,/ , ..f
+
.a -
)
.* f' ~ (1.Scontainmentpenetrations:
w ,
l vent valves, and vacuum breakersair lock, hatch. seals, electrical penetrations.
V (2) 'drywe11' components: a a
' breakers air locks, hatch seals, and post-accident vacuum?
[> 4 L(3) hydrogen lignitersystem l .(4) combustible gas contro1: system:
- o. hydrogen recombiners o, .
system, and post-accident atmosphere sampling vahes , drywell mixing L
(5): containment cooling: ,
sprayisolationvalves,LPCIinjectionvalves,and unit coolers-y q (6) _ automatic depressurization system (7) containment and reactor monitoring:
- - containment and drywell- temperature instruments
- reactor pressure vessel wide-range pressure instruments !
reactor pressure vessel wide-range and fuel-zone level instruments J (8) junctionblocks. associated instruments, controls, cables, interlocks, terminal and
, f,
~
i equipment: essential for the mitigation of postulated d3 ,
conditions. :As part of the final analysis each licensee with a Mark III containmentshouldprovideplant-specificInformationcorrespondingtothe i
'the sele;ction criteria. generic 11st in conjunction with unique design features th!
e J .
. 7.2: J Generic Eouioment Survivability Analysis (Localized Combustion)
Lby using the CLASIX-3 code.As discussed in Section 6, pressure and tempe the containment environment or boundary conditions 'necess ment response analyses for hydrogen release rates below the diffusion flame .
i extinguishment limit where burning is limited to localized combustion. 1~
believes that equipment survivability can be established genericall The HC0G The su porting analyses was presented in its letter dated August 7. -1987 (y.HGN-11 .
- It was stated that a generic approach is sufficient b conditions used,in the generic equipment survivability analysis.
.The'HC0G has listed someidentified of the moreaimportant number of.conservatisms items below. in its generie analysis; the I
(1) . The constant hydrogen release rue o*f 'D.1 lb/sec, which is the nonmechan isticrecovery.
core " tail" portion of the rilease profile, is unlikely to occur after surface of the pool m y exist as lo,e as 4.02 lb/sec indicatin Mark III SER 39
.; f .: . . ~ ,
Nj ., {
combustion would not' occur. It is expected that the presence of diffusion flames would probably be the dominant combustion mode, possibly in combi-nation with localized combustion phenomena when the hydrogen flow rate is- ;
below the flame extinguishment -limit. The significance of these two dif- ;
forent combustion modes is tho' spatial shifting of thermal loads; as such, h r single piece of equipment would not continually be exposed to hydrogen burning resulting in a lower temperature profile. -
-(2) The CLASIX-3 wetwell temperature profile was used as the boundary condition j
for the ecuipmet response analysis although the most sensitive equipment-is locatec outside si t9 .t. ell volume. The wetwell has the severest l environment of the three containment volumes. The staff finds that based j
on limitations of the CLASIX-3 methodology used in the generic analysis, there is no choice but to use the wetwell volume. However, the staff does I recognize the selected profile is limiting. '
(3) In the selected CLASIX-3 case there are no active containment cooling mech-i anisms(i.e.,thelackofavaIlabilityofspraysorunitcoolers). Because L of the. type of event considered, a recoverable degraded core, the HCOG expects that sometime during these relatively long transient events, spray /
unit coolers would become available.
- A set of equipment common to each plant with a Mark III containment was compiled from the list of generic equipment. Subsequently, the most thermally sensitive D equipment, such as cables pressure transmitter, hydrogen igniter assembl l ADSsolenoidvalve,wereIncludedinthegenericsurvivabilityanalysis.y,and Based
! the drywell break l on the results equipment of the response generic analysis equipment showed favorablesurvivability results; whereas, analysis,it, the 50RV case,
- the thermal analysis response for the pressure transmitter indicated a 27'F ex-i coedance above its qualification temperature. The significance of this result is !
l "
assessed below.
O .
1 The equipment response analysis for the 50RV case used the wetwell CLASIX-3 tem- ,
- perature profiles, presented in Section 6, with some modifications. These mod- l l ified profiles exclude the few initial burns in which diffusion flames would l l exist'and the last induced global burn. the modified wetwell pro- l o file contains about 90 serial hydrogen burns. As a result,lculated The ca critical component l
- of the pressure transmitters exceeded its qualification after the seventy-first i hydrogen burn. However, the HCOG indicated that the pressure transmitter is 1 l expected to survive the hydrogen event because of various conservatisms in the
- j. analysis.
L 1he staff acknowledges conservatisms, as discussed earlier, are contained in these analyses which could compensate for the tem)erature exceedance over the qualification of the pressure transmitter. Nonet1eless the staff requested the c
'HCOG to provide additional data nn the qualification of,the pressure transmitter.
L' By letter dated April 5, 1988 (HGN-131-P), the HCOG indicated that during quali-
~ fication testing the transmitter had operated without failure at surface temper-atures approaching 380*F for several minutes (as compared to the qualification of320'F). As part of the HCOG's response, an equipment response analysis of i o the pressure transmitter was re-evaluated assuming containment sprays to be l
available. This res ture was about 70*F less ponse analysis than indicated qualification that the equipment temperature of 320*F.surface Thesetempera-results l.
L 1
E Mark III SE9 40 p
V l'
[ ,
demonstrated the impact of sprays to cool the containment environment thus L maintaining the function of essential equipment. RiverBendStationIsthe L only plant without containment sprays, but unit coolers are part of its design.
While specific analyses have not been performed to support quantification of the cooling effect provided by unit coolers versus sprays, the HC0G concludes that a reduction in background temperature would be adequate to reduce the l thermal loads on the pressure transmitter.
In summary, HCOG contends that further analysis is unwarranted because, with the u
- l. potential for active containment cooling and the conservatisms inherent in the analyses, the pressure transmitter will function as designed during recoverable .
degraded core events that progress to 75% metal-water reaction. With regard to , l these analyses at low hydrogen flowrates, the staff agrees with HCOG's position l that further effort in this area is not warranted. Moreover, the staff finds I that the determination of equipment survivability based on the data obtained I from the QSTF for diffusion flames is more appropriate than an assessment based
=
l on localized combustion conditions. I 7.3 Diffusion Flame Thermal Environment Methodolony
. 1 In its letter dated July 30 1986 the HCOG outlined the methodolo:y I tobeusedbyeachmemberIIcensee(HGN-103),inethefull-scaleplant-specific to determ
.ontainment thermal environments from the QSTF data. The full-scale environmen-l tal conditions would be used as boundary conditions in the HEATING-6 computer code to analyze the response of containment equipment during postulated diffu-l sive combustion events. As a result of these analyses, the survivability of essential equipment would be detemined.
From the production test series conducted for each Had. III containment, the test that produces the most limiting environment at ;le corresponding (to full l scale) equipment location is used. Thermal profile Are constructed by spatial mapping of the test facility data. Specific plant gofiles are developed from i average temperatures for time intervals of maximum hydrogen flow and low con-L-
stant hydrogen flow from the production tests. This allows determination of the plume locations and the effects of blockages and spargers.
Full-scale velocities are computed from the quarter-scale measured velocities using Froude scaling; test temperatures are used directly (scaling is 1:1). The convective heat transfer and radiative heat fluxes are computed using the scaled velocities and temperatures. Since this approach establishes an environmental l map, the heat transfer modes that should be considered are dependent on the location of the affected equipment.
To validate and assess the heat transfer methndology, a complex (three-dimensional geometry)imeter the calcr to different locations and different thermal environments.A calorimeter ass HEATING-6 model of the com)1ex calorimeter was constructed and the calculated response was compared to tie measured response to validate the methodology.
l This effort is presented in the HCOG's letter (HGN-105-P) dated August 29, 1986.
SNL assessed the submitcals and determined that most of the computed results were conservative from the standpoint of the equipment survivability. Therefore, the correlations and results presented are reasonable. However, SNL recommended 1
l Mark III SER 41
l-
. that when the generic methodology is used for plant-specific equipment evaluations, a review should be conducted to assure conservative specification of:the boundary ,
conditions.
In summary, the staff finds that the heat transfer methodology dealing with diffusive combustion as presented in HGN-103, provides an acceptable foundation to perform plant-specific equipment response analyses. Accordingly, each licensee with a Mark III containment intends to use this methodology as part of its final analysis, required by the hydrogen rule. The staff agrees with SNL's recommendation, that sufficient detail of input data should be provided by each licensee to ensure its analysis is conducted in an appropriate manner.
7.4 Spray Availability In the preliminary evaluations of hydro Gulf SSER No. 3. NUREG-0831 July the 1982) gen staff ignitercredit allowed systems (e.g., see Grand for operation of containment sprays in the analyses of the consequences of hydrogen combus-tion during degraded core accidents The validity of the assumption of contain-ment spray operability was premisu on several considerations. First, in the preliminary evaluations of igniter systems the staff and HCOG focused on the SORY transient and drywell pipe break accident sequences. These accident
-sequences do not necessstily imply loss of the containment spray function of the RHR pumps, RHR pumps may be operable but the LPCI injection path may be interrupted or lost. Further, at the time of the preliminar overall tone of the BWR emergency procedure guidelines (EPG)y evaluations was to focus on the containment integrity rather than adequacy of core cooling at an earlier point in a degraded core accident sequence.
Since the preliminary evaluations were conducted, additional information has been developed which raises questions regarding the validity of assumptions concerning availability of the RHR pumps in the containment spray mode. In contrast to the earlier focus on the SORV ' transient and drywell pipe break recent risk analysis indicates that SB0 is a signif,1 cant contributor to hydrogen generation events. For the 5B0, the loss of reactor makeup is tied to the loss of pumps, including RHR pumps, in either the LPCI or containment spray mode.
Thus for the SB0 sequence, the RHR spray function cannot be reasonably assumed to be available until ac power is restored. Finally, the earlier emphasis in the EPG's on containment integrity vs core cooling for containment spray operation has been reversed. In Rev. 4 to the BWR EPG's, (March 1987) the sequence of steps has been modified. Use of RHR pumps in the containment spray mode, irrespective of adequate core cooling, is now directed as the last step, to control pressure rather than before the decision to vent.
For the above' reasons the staff concludes the BWR Mark III owners should evaluate the containment and essential equipment response to hydrogen generation events assuming containment sprays are unavailable, consistent with SB0 assumptions and the EPGs. Spray operability can be modelled but should be trtated in the context of establishing margins for a variety of possible plant conditions.
Similarly, assumptions regarding availability of containment coolers should be consistent with the basic premise of the SB0 accident sequence.
Mark III SER 42 i
hj ., e i . f '
] ,
7.5IPres'sureEffects s In HGN 118-P, the'HCOG indicated equipment located inside containment is quali- .
fu to a pressure loading of at least 30 psig applied externally. -The CLASIX-3 predictions produced the most severe pressure rise of about 23 psig in the Mark III containment. The staff concludes that pressure is not a concern pending confirmation.by each licensee of the 30 psig capability. When the hydrogen ig-nition system is functioning,-various containment subvolumes till be randomly affected by hydrogen burning, however, a large pressure spike is not expected to occur.
-7.6- Detonations The HCOG believes that a detonation is not a credible phenomenon in the Mark III cont,ainment because (1) no rich hydrogen concentrations will accumulate inside containment since the distributed igniters will initiate combustion as the mixture reaches the lower flamability limit and effective mixing will occur and (2) there are nc regions of the containment with sufficient geometrical coa-finement to allow for the flame acceleration necessary to yield a transition to detonation.
The staff agrees with the HCOG position. As confirmed by the quarter-scale-test results, the atmospheric conditions inside the test facility was well mixed and burning at low hycrogen concentrations was prevalent. Thus, the potential for localized accumulation of significant concentrations of hydrogen is concluded to be unlikc?.
8 CONCLUS10t the staff finds the HCOG topical report, On the basis
" Generic of the Control Hydrogen above evaluation, Information for BWR-6 Mark III~ Containments,"
(HGN-112-NP)datedFebruary 23,1987, provides an acceptable basis for techni-cal resolution of the Mark III containment degraded core hydrogen issue.
rovide a plant-specific final, analysis, as required Eachlicenseeshouldp(vii)(B),whichwilladdresstheelementsspecifiedin by 10 CFR 50.44 c)(3)
'10CFR50.44(c)((3)(vi)(B). The HCOG topical report, or portions thereof, ray be referenced where appropriate, taking into consideration the staff recomendation as stated in this report. The plant-specific analysis will use
-test-data described in the topical report to confirm that the equipment necessary to establish and maintain safe shutdown and to maintain containment integrity will be capable of performing their functions during and after exposure to the environmental conditions created by the hydrogen in all credible severe accident scenarios.
An element of the staff's assessment for determining the adequacy of the HIS was the determination of whether or tot an alternate power supply was appropriate. An.important factor in this decision process is the level of risk associated with SB0 events leading to core damage. Recent risk studies reported in NUREG-1150 have shown that the overall core melt freqiiency for one Mark III plant (Grand Gulf Nuclear Station) is very low, i.e., IE-6/ year.
However, a potential vulnerability for Mark !!! plants involves station blackout (SBO),duringwhichtheigniterswouldbeinoperable;andthis i
Mark III SER 43
condition appears-to dominate the residual ~ risk from severe accident in the
~'_, Mark III plants. Under $80 conditions, a detonable mixture of hydrogen could
- develop.which could be ignited upon restoration of power resulting in loss I of containment integrity. On the basis of a separate evaluation of this possibility in the context of the NRC staff Containment performance Improvement (CPI) program, 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 (IPEs) of the Mark !!! plants. The staff has~ requested that the licensees consider this
. issue as part of the IPE in Generic Letter 88-20, Supplement 3.
With the caveat that the vul.ierability to interruption of power to the-hydrogen ignittes should be tarther evaluated on a plant-specific basis as part of the I Kt of the Mark III plants, the staff finds that there is reasonable assurance that the HIS installed in the plants with Mark III containments will act to centrol the burning of hydrogen so that 'here is adequate protection against centainment failure.
In summary, the staff conclu0es that the following key elements should be ad-dressed in each licensee's plant-specific final analysis to resolve the degraded core hydrogen control issue:
(1) hydrogen ignition system design (the vulnerability to interruption of power to the hydrogen igniters should be further evaluated on a plant-specificbasisaspartoftheMr.rkIIIplantsIPEs)
(2) confirmation of applir. ability to the generic effort (3) quarter-scale plant specific production testing results (4) primary containment structural survivability quarter-scale testing pressure capacity analyses for drywell and containment, for example, confirm previous plant-specific analyses ,
(5) survivability of essential equipment identify plant-specific essential equipment define thennal environment from quarter-scale testing perform equipment response analysis confirm that redundancy exists for that equipment affected by second-arf burning and drywell inverted diffusion flames confirm pressure capability of equipeent Mark III SER 44
7;p;;x..,;.'*i
,e
.*v .
1 1,. .t .
L
, g. ., >. a
., s,.
? (6) licensee's position'regarding the proposed NC0G'energency procedures-for-Jcombustible gas.-control ,, l y !(7)-:overall conclusions relating to conformance of the hydrogen rule--
.n,,
1 L e
i.
I L:
7 I
m,..
\;
i s
a
- O, f' p-h 1 i.
l L. .
(', ;c i
I i
1
' Mark III SER 45
L
}. .
4 APPENDIX A GEN [k]C HYDROGEN 3GNIT!0N SYSTEM TECHNICAL SPECIFICATIONS Generally, technical specifications-(TS) of a particular system consist of two distinct sections: . Surveillance Requirements to ensure system operability and Limiting Condition for Operation (LCO) to define the allowable operability range in conjunction with various plant actions when needed. Each of the four plants with Mark 111 containments have similar TS for the hydrogen ignition system (H15). The following discussion on this subject is focused on the proposed generic HI5 TS and its deviations from the current TS, Currently, TS on igniter systems in plants with Mark 111 containments prescribe two types of surveillance practice. At 184-day intervals, all the igniter assemblies are energized and current / voltage measurements are performed and compared with similar measurements taken previously. If more than three igniter assemblies on either subsystem are determined to be inoperable, there is an increase of the surveillance frequency to a 92-day interval. A second pcrt of this first surveillance requirement is the verification that inoperable igniters are not adjacent to each other, if more than one igniter on each subsystem is determined to be inoperable. This requirement is based on the staff's view regarding potential hydrogen pocketing in enclosed areas. The second type of surveillance is conducted at 18-month intervals to verify a surface temperature of at least 1700'F for each accessible igniter and verify by measurement sufficient current / voltage to develop 1700'F. surface temperature for those igniter assemblies in inaccessible areas. _ Accordingly, the bases section of-the TS indicates that inaccessible areas are defined es areas that have high-radiation levels during the entire refueling outage; such enclosures include the heat exchanger, filter domineraliter, and the pump room for the reactor water cleanup (RdCU) system.
The current LCO allows no more than 10% of the igniter assemblies inoperable per subsystem. And if one subsystem is inoperable, the action statement re-quires restoration to operable status, or be in the required operational condition, within 30 days (similar to the hydrogen recombiner TS).
By letter dated April 16, 1986 (HGN-070), the HCOG proposed to revise selected portions of the existing plant specific TS as outlined above. Principally, there are two significant proposed changes: (1) an increase of the allowable inoperable igniters per subsystem to about 40%, as compared to the current value of 10%, and (2) removal of the surveillance requirement in determining the location of the inoperable igniters after the requisite number of failed igniters have been attained Aho, es discussed above, the current action statement requires an allowable period of 30 days to restore the igniter sub-system to operable status. HCOG proposed to change this interval to 60 days because these events in which the HIS is required to be operable are less probable than design-basis accidents. The staff finds the proposed inoperable period increase is not based on sound engineering judgment since it relies l
Mark 111 SER 1 Appendix A
~
._',1' on being ceyend the DBA but does not provide a rationale for 60 days. Thus the 30 day interval is appropriate and should t,e maintained.
In its letter of April 16, 1986, the HC00 provided a justification for the other two significant proposed changes. Essentially, the HC00 cited the con-clusions observed regarding the QSTF testing (see Section 3.1) of a particular
~
i -
scop'ng test in which about 40% of an igniter subsystem was inoperable ia con-junttion with the other subsystem not functioning.
- Th's staff has concluded relative to the hydrogen aspects of the TS justifica-tions that the HCOG had not provided sufficient justification to relax the 15 to such a degree. The staff made its determination because of the inherent uncertainties such as extrapolation of the quarter-scale results to full scale, various injection rates, different safety relief valves actuating, and differ-ent combinations as to where the 40% of the inoperable ignitors could be located. Therefore, t se allowable value for inoperable igniters should be as low as practical; a If& value appears to be a reasonable limit.
The second proposed change is to remove the surveillance that ensures that inoperable igniters are not adjacent. Principally, this surveillance ensures at least one operable igniter in each enclosed area and coverage of the azimuthal-positioned igniters in the open regions. Essentially, the HIS TS are intended to prevent buildup of hydrogen in subvolumes of the Mark Ill contain-sent, and thereby prelude the occurrence of large volume burns.
The following considerations are highlighted as part of the HC0G's justification for proposing to remove the TS provision to determine that inoperable igniters are not adjacent:
The HCOG evaluated the potential flow paths that could transport hydrogen in or near enclosed regions of the containment and detemined that no
. potential hydrogen source exists. It is expecte'd that igniters in open areas will function to preclude local hydrogen pocketing.
e Observations-of the quarter-scale tests indicated that the released hydrogen will tend to mix with the surrounding atmosphere and thus reduce the potential of locally high nydrogen concentrations.
The likelihood is low for inope'rable igniters being located in such a fashion as to create a large containment subvolume that would be without igniter coverage. ]gnitors would tend to fail in a random manner.
Currently, whenever at least one igniter is inoperable in each sub-system, containment entry is normally required to find the location of the f ailed igniter. This would subject plant personnel to various occupational safety hazards such as radiation exposures and tha. risks associ.ted with the construction of scaffolding. -
On the *, asis of these considerations provided by HC00, the staff has determined that toere is reasonable assurance that the proposed TS without the adjacent igni',er provision would not have an adverse effect on the effectiveness of the ign'ter system.
A Mark Ill IR 2 Appendix A
- . . ._- .. .- - .. . . . . . .. . ~
w . >
A , 0, , ,, . J f;; c ,
q e ,
s -.
- (The staff finds the generic HIS 15 as eocumented in the NCOG 1etter dated V April 16. 1966, to be acceptable' contingent on the following changes: the'.40%
- value of allowable inoperable igniters should be 105 in the-appropriate locations of'the< text end the 60-day _ interval to restore a subsystem in the action state-sent-should be 30 days. Each Mark 111 owner that. intends to adopt the generic
~. 'HIS technical specification must confirm that the,HCOG assumptions used in the i
? development .cf the T5 are valid for their plant-specific configuration.-
o t
o N
,1 1 :-
l.
1' l
-c..
L.
i 5
i l
o p
I
~
1 Mark 233 SER 3 - Appendix A L .. . .
. _ _ _ _ _ . _. --- . -. -. - - - -_ - . . -. - . . = .
L
.APMNDIX B MARK !!! COMBUSTIBLE GAS CONTROL '
(- EMERGENCY PROCEDURE GUIDELINE L As part of the generic program, the HCOG has developed Combustible Gas Control-i Emergency Procedure Guideline (EPG) for plants with Mark III containments. The~
1atest version of the guidelines with supporting appendices were sent to the NRC by letter (HGN-122-P) dated July 8,1988. This procedure includes operator action for the hydrogen igniter system as well as other combustible gas control ,
systems designed in the Mark III containment, such as hydrogen recombiners and the drywell mixing system. In addition, the proposed procedure provides guid-ance for, spray actuation and containment venting. This effort is to supplement the overall BWR Owner Group's EPG program.
Fidure B.1 highlights'the operator actions in dealing with hydrogen in an emer-l gency situation. These actions for controlling hydrogen depend on a detennina-tion of hydrogen concentration in the containment and drywell as indicated by hydrogan monitors and/or analyzers that obtain gas samples from the containment I and drywell. The significant trigger limit used in the EPG is when the drywell or containment hydrogen level reaches a concentration at which a global deflagra-tion could threaten containment or drywell integrity from overpressurization, referred to as the hydrogen deflagration overpressure limit (HDOL). At this o
l limit or when the containment hydrogen concentration cannot be determined to be below the HDOL and it cannot be determined that the igniters have been con-tinuously operating, the HIS should not be used. The containment HDOL is a
' curve of hydrogen concentration versus containment pressure, whereas the dry- .
well HDOL is a single value representing a peak hydrogen concentration. The I containment HDOL is more limiting than the drywell HDOL.
The hydrogen ignition system, the hydrogen' recombiners and the drywell mixing system are the key hydrogen mitigating systems. As, indicated in Figure B.1, these systems are activated at appropriate trigger points to deal with a pro-gressing hydrogen build-up. With the addition of an independent power source to the HIS, it is anticipated that for most severe accident / degraded core ;
situations the resulting large amounts of hydrogen can be accommodated.
As part of the subject letter, HCOG had addressed staff consnents which were discussed in a October 22, 1986 HCOG/NRC meeting. In this latest version of the Mark III Combustible Gas Control EPGs (Revision 3), HCOG has addressed staff concerns or provided sufficient justification for their position as discussed below.
As one of the initial steps in the EPGs, the operator is instructed to vent the suppression chamber or drywell, whenever either of the respective regions reaches the minimum detectable hydrogen concentration (0.5%), provided the offsite radioactivity release rate is expected to remain below the'offsite release rate limiting condition for operation (LCO). It should be noted that this step is similar to the BWR EPGs for Mark I arid Mark 11 combustible gas Mark III SER 1 l
l
- - _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ _ - _ - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ b
i
'. control. The staff previously commented that venting may not be necessary solely upon hydrogen concentration above the minimum detectable level and below flammability levels; the use of recombiners is valuable and should be
-utilized where appropriate.
. In response, since dissolved hydrogen is present in the reactor coolant system during normal operation and the EPGs.are based on a symptomatic approach, it-
'is the intent of dis step to remedy a hydrogen problem during normal operation, and within the constraints of technical specification limits. HCOG believes there is sufficient guidance to preclude this action from being implemented during a genuine emergency situation. Also HCOG had c generic Mark III procedure at later date, if necessary,ommitted to be consistenttowith modify the the BWROG's approved combustible gas control procedure. The staff finds that the subject procedure and the HCOG approach to be acceptable.
As one of the last steps to control hydrogen accumulation during a progressively worsening situation, containment venting is directed. Venting the containment irrespective of. the offsite radioactivity release rate would only be considered to restore and maintain the containment hydrogen concentration below excessive limits. Containment failure may follow if a large deflagration were to occur.
Venting the containment may be the only mechanism which remains to prevent an uncontrolled and unpredictable breach of the containment.- The controlled re-lease of radioactivity to the environment is preferable to containment failure whereby, adequate core cooling might also be lost and radioactivity released with no control. This concept of venting is similar to the emergency procedures for pressure control.
Regarding the second issue. HC0G provided additional information responding to NRC comment dealing with the limited use of the drywell mixing systems. The staff views containment venting as a last resort to deal with extraordinary conditions. - The use of the drywell hydrogen mixing system may delay contain-ment venting by diluting the containment volume (at a higher concentration of hydrogen) with the drywell-volume (at a lower concentration of hydrogen). HCOG cited various' factors to demonstrate the drywell mixing system is not beneficial for hydrogen control inside the containment volume, which includes: . dilution effects are marginal, since the containment is significantly larger than the-drywell; the mixing system would re-initiate a LOCA signal and potentially interfere with event recovery; and implementing a modified procedure may induce conflicting direction. In addition, the design intent of the mixing system is to deal with hydrogen in the drywell.
The staff believes some of HCOG concerns are valid. In addition HCOG had modified its procedures to assure that the HIS would remain operational above the HDOL if it can be determined that the igniters have been continuously operating. The addition of an independent power supply to the MIS would further enhance the reliability of the system. Consequently, the added reliability would reduce the potential for containment venting to control hydrogen inside containment. Therefore, the staff agrees with HCOG that the inclusion of the dryw)
(indelayingventing ell mixing system as compared would not provide significant benefits to its disadvantages.
Overall,thestafffindstheproposed(Revision 3)MarkIIIcontainmentEPGs are based on sound technical judgment, and are acceptable. Accordingly, each Mark III licensee should address its combustible gas control emergency proce-dure in the plant specific final analysis. -
Mark III SER
., , - _ . . _ . _ _ . . _ . . . _ _ . _ ~ . = . _ _. -... . . . . _ . . - . _ _ _ . _ . . _ ._._ ~. _
- -e = , .
.~4 'e' sf. c' ' . ,
.s c J 'M , compitem attem 1
e.. ..,te 6teri e t. -
3 De gang;) et HitsemgD- attpatt surftes I!
.am>= -]
PC att De a y e etDL' .
I' .
l 4 I i
Pc De p. m yagatags um Mitttat.t 6f vit i
- 08/t*
etttatt ,n satt vtwt ame puest
- et60 Pc tco
" #C l _
D*
l.- PC
_ 9 t 1 I
et *2 88'C"El ** 88": '**
O. . .. g .. accoments opteatsom
.co.
fttic160.t tv* wee. ettttias.t 6tvte 3
i
'j Pc ama Pc m a saa L ,,gg3
,gg C C ,g,h,etoe De 62 t
- M006 f0e et ,p M, e Mfutestepa l ppg e *0* /2' WC i 98 teat! 'De/2+
- 3 i
q i __ f asserttes ;
e _ _
i
- 4- ,
l l 7 63 mg statuts 'I 8"E Ytl SPCeait SDeej-spe pe=4e3 De > $ PC "2 mg ;
r 8E88"8"I'8
- turt 9008a4414= l t #DD6
- l 8
eso i se i
i __
l= l
..* Ih3[I , Pt #2 statuts me Coat toe gg g g g etteset' ate optaatsem og -
t n ,e e ,.' a m es.,0., maa*=
I et a, vis _
e,tein ee=tu
. 6 ha asi=s stin=s so 4 - i i
- i.
toutmute 41 O
Figure B.1 Operator Actions to Control Hydrogen .
RRK'111 SER 3 APPENDIX B
, . <- ..t
^n: , ,,~,.
- e. . , ne -
i etws1es- et?ID" O- .
(
'I "2 atatsitt appa Stevet W E80MDutk -
1
.i:
': ' vtl Opteait Pt 98eav8.
$IN[,I ,
att Preafffset
" C ' ._
i
+
ecu'ttes t o= t a so.'s . '
f ate & ftt ligt! - 'll
' (8tttt"=t 60*f t' *3 f 08 9tf 6.nseate g.
Situtt W S m H et ame emand gfstges .
i l
ht -
9 g i
- @ / *
- PC mge lvant
> #90.
i t
et *ettsvet Optaatt PC smeats, -
- C I * ' mets
- Of Att F
s PC m2
- 88006 800/9' 8
to m 'I' 'I'I' 2 4818ue: TTSit
- v patsluet e PC'6
" ! ._ i
.{
(E ',"I[; -
$PttaTE N g starttes PC m2 e asas tout rea _
etComt.ute 0*Caateos .
- t8AM 'd 3 M C089mt91
~
Ptkga 151 cent f > S.tfLACR41som aao.
PC m2
- II FIGURE B.1 (Continued) Operator Actions to C .itrol Hydrogen
- RK 111,$ER 4 APPENDIX B
__ - _ . - ._ _ . _ _ . . m. . _ _. . . _ . _ . - _ __ ._ . . . . _ _ _
, e z Qf'-
. *. , . * **>e*
i ,..
' C0wMeeg - W84 1
De/*
.--.1
- 1Jalt tatt sta;w!I LCC W [anD Gt . i
,,,, l p *2 * "83*
.anD.
Wg sw*ttl mot gettatoms ggCURL M8 Surftel C0=tiasesl69 Seatt De "2 ts:ttMD 40will tout Fo* Mrpoestem i
l De/4 *! s'2.m*p."E8' m.somit*1 he* 98(tatm; Strutt W. asssut $tst!43.
I et 'E s CC DL ($1 kttenedas ame SW#$ .
C0=* rpe gruS8410m l
t I
Warf assp pues et
-l 0e/4
. pg W, tg Stevat PC' ett0e e@0.
Mm a4 PAGE .,
1
.I j 9e/$
pi m StatWtl 2
Sttutt tw M 2 I NDO. utsms gTgttus ,
l i
L l .
i i
Figure B.1 (Continued) Operator Actions to Control Hydrogen l MARK 111 SER .
5 APPENDIX ..
1 l
l
- ,. Q r. 7 i
APPENDIX C BIBLIOGRAPHY-Gaunt, R'. 0.'et al., "The DF-4 BWR Control Blade / Channel Box Fuel Damage Experiment " Sandia National Laboratories, to be published.
Hydrogen Control Owners Group (HC0G), HGN-003, letter from J. D. Richardson (HC0G)-
to H. R. Denton (NRC), " Report on Hydrogen Control Accident Scenarios, Hydrogen Generation Rates and Equipment Requirements," April 8, 1982.
-- , HGN-006, letter from J. D. : Richardson (HC0G) to H. R. Denton (NRC), " Report -
on Hydrogen Control. Accident Scenarios, Hydrogen Generation Rates and Equipment Requirements," Rev. 1, September 9, 1982.
-- , HGN-009-P, letter from J. D. Richardson (HAG) to H. Denton (NRC), "CLASIX-3 Report," March 18, 1983.
-- , 011-NP, letter from J. D. Richardson (HCOG) to R. W. Houston (NRC), "Re-sponses to NRC Requests for Additional Information," May 11, 1983..
-- , 012-NP, letter from S. H. Hobbs (HCOG) to R. Bernero (NRC), "1/4 Scale Hydrogen Test Program," August 12, 1983.
-- , HGN-014-NP, ' letter from S. H. Hobbs (HCOG) to H. Denton (NRC), " Final 1/20th Scale Test Report " February 9,1984.
-- , HGN'-017-NP, letter from S. H. Hobbs (HCOG) to H. Denton (NRC), "Mard -III Hydrogen Control Owners Group Final Whiteshell Ignition Test Report," June 7, 1984.
-- , 'HGN-020, letter from S. H. Hobbs (HCOG) to H. R. Denton (NRC), " Transmittal of BWR Core Heatup Code Hanual," September 5, 1984.
-- , HGN-027-NP, letter.from S. H. Hobbs (HC0G) to R. Bernero (NRC), "1/4 Scale Test Facility 3D-Complex Calorimeter," February 13, 1985.
-- , HGN-031, eletter from S. H. Hobbs (HC00) to R. Bernero (NRC), " Hydrogen Release Histories and Test Matrix for 1/4 Scale Test Prograa," March 13, 1985.
-- , HGN-032, letter from S. H. Hobbs (HC0G) to R. Bernero (NRC), " Submittal of Information on BWR Core Heatup Code," April 16, 1985..
-- , HGN-034, letter from S. H. Hobbs (HC0G) to R. Bernero (NRC), "Model for Hydrogen Production Equivalent to 75% MWR " May 17, 1985.
-- , HGN-051, letter from S. H. Hobbs (HCOG) to R. Bernero (NRC) " Availability of Containment Spray System," July 26, 1985.
s j 1 Appendix C
- c. ,....
,'HGN-052, letter from S. H. Hobbs (HC0G) to R. Bernero (NRC), " Hydrogen Release Time Histories," August 1, 1985.
'-- , HGN-055, letter from S. H. Hobbs (HCOG) to R. Bernero (NRC), " Evaluation
-of 580 and ATWS Contributions to Hydrogen Generation Events," September 27,
. , -1985.
( -- , HGN 070, ' letter from J. R. Langley (HCOG) to R. Bernero (NRC), " Mark III
. Hydrogen Ignition System Technical Specifications," April 16, 1986.
-- , HGN-072, letter from 5. H. Hobbs (HC00) to R. Bernero (NRC), " Event Sce-narios Considered for Evaluation of Drywell Response to Degraded Core Accidents,"
March 5, 1986. l
-- , HGN-073, letter from J. R. Langley (HCOG) to R. Bernero (NRC), "Justifice-tion for Manual Actuation of Mark III Hydrogen Ignition Systems," March 5,1986.
-- , HGN-084, letter from J. R. Langley (HCOG) to R. Bernero (NRC), " Generic l Equipment Survivability List," May 16, 1986.
-- , HGN-085, letter rom J. R. Langley (HCOG) to R. Bernero (NRC), " Final Re-port Assessing Adequacy of Heat Sink Modeling in the 1/4 Scale Test Facility," l May 5, 1986.
-- , HGN 089, letter from J. R. Langley (HCOG) to R. Bernero (NRC), "8WR Core Heatup Code Responses," June 9, 1986.
-- , HGN-091, letter from J. R. Langley (HCOG) to R. Barnero (NRC), " Criteria for Existence of Inverted Diffusion Flames in the Drywell," June 25, 1986.
-- , HGN-092-P, letter from J. R. Langley (HCOG) to R. Bernero (NRC), " Report of CLASIX-3 Generic Analyses and Validation of CLASIX-3 Against 1/4 Scale Test Facility Data," June 10, 1986.
-- , HGN-096-P letter from J. R. Langley (HCOG) to R. Bernero (NRC), "8WR Core
[' Heatup Code Report," July 30, 19E6.
-- , HGN-098-P, letter from J. R. Langley (HC0G) to R. Bernero (NRC), " Scoping Test Report," July 18, 1986.
l
-- , HGN-099-P, letter from J. R. Langley (HC0G) to R. Bernero (NRC), " Transmit-tal of Handouts from June 19, 1986 HCOG-NRC Meeting," July 11, 1986. 1
-- , HGN-100 Tetter from J. R. Langley (HCOG) to R. Bernero (NRC), " Nevada Test Site Data Evaluation," July 31, 1986.
-- , HGN-101, letter from J. R. Langley (HCOG) to H. Denton (NRC), " River Bend- I Staion Unit Coolers," July 30, 1986.
-- , HGN-103, letter from J. R. Langley (HCOG) to R. Bernero (NRC), " Diffusive l Combustion Thermal Environment Methodology Definition Report," July 30, 1986. j i
2 Appendix C
- s. .
Q,s;;.x Tt. W ,
.. s
. .z ,
a
~
- - ,1 HGN-104 P, letter from J. R. Langley (HC0G) to R. Bernero (NRC), "Evalua-tion of Emergency Procedure Guidelines Operator Actions Against MCOG Assumptions- '
for Analysis of.a Hydrogen Generation Event,". August 18, 1986. .
- - , HGN405-P.,1etter from J. R.' Langley (HC0G) to R.- Bernero (NRC) " Diffusive Combustion Heat Transfer Methodology Validation for Equipment Survivability
'in Mark III Containments," Augut t 29, 1986, s - , HGN-106-P, letter from J. R. Langle (HCOG) to R. Bernero (NRC), " Supple-mental Information on Secondary Burning,y' September29 1986.
e
-- , HGN-109-P, ' letter f rom J. R. Langley (HCOG) to R. Bernero (NRC), "CLA51X-3
. Generic Sensitivity, Analyses," December 9, 1986.
-- , HGN-110 P, letter from J. R.iLangley.(HCOG) to R. Bernero (NRC), " Combust-ible Gas Control Emergency Procecure Guideline and Supporting Appendices,"
' December 1, 1986. ,
- , HGN 111-P, letter from J. R. Langley (HCOG) to NRC, "CLASIX-3 Summary Re--
port,'? December 15,:1987.- .
$ - , HGN-112-NP, letter from J. R. Langley (HC0G) to R Bernero (NRC), " Generic
- Hydrogen Control Information for BWR-6 Mark III Containments," February 23, 1987.
-- , HGN 113, letter from J. R. Langley (HCOG) to R. Bernero (NRC), " Comparison of CLASIX 3 Predictions to Nevaca Test Site Data," January 8,1987.
h _-_- , HGN 114, letter from J. R. Langley (HCOG) to R. Bernero (NRC), " Response to NRC Concerns Regarding Potential' Impact of 580 Events on Operation and Per-
~
L l formance of the Hydrogen Ignition System," January 8,1987. . .
l-L -- , HGN-115 NP, letter from J. R. Langley (HCOG) to R. Bernero (NRC), "1/4 Scale Test' Facility Final Design Report, February 10, 1987.
--- HGN 118-P,' letter from J. R. Langley (HC00) to NRC, " Generic Equipment Survivability Analysis," August 7, 1987.
- , HGN-119, letter from J. R. Langley-(HCOG) to NRC, " Final. Inverted Diffusion l
Flame Report," June 10, 1988.
l
%- , HGN-121-P letter from Langley (HC0G) to NRC, " Report of Hydrogen Combustion l . Experiments in,a 1/4 Scale Model of a Mark III Nuclear Rear, tor Containment,"
July 22.1987.
-- , HGN-122-P, letter from J. R. Langley (HCOG) to USNRC, " Revision 3 to
' Mark III Combustible Gas control Emergency Procedure Guideline," July 8,1988.
-- , HGN-123, letter from J. R. Langley (HCOG) to USNRC,
- Response to NRC Ques-3 l
tions on Station 81aciout and ATWS Sequences " September 9,1987.
-- , HGN-128, letter from J. R. Langley (HC00) to NRC, " Responses to NRC/SNL Comments on Heat Loss Report," November 6, 1987.
l L - , HGN-129-P, letter, J. R. Langley (HC0G) to NRC, " Revision 4 EPGs vs HC0G Scenarios " April 8, 1988.
3 Appendix C L _ _ . , _
-a t "
j
(,n. ;; 3v v, -
q,,
- - , NGN-231-P, letter, J. R. Langley (HC00) to NRC, " Supplemental Discussion i
'of Pressure Transmitter Survivability at Low Hydrogen Release Rates," April 5.- '
l
.1988.
l
-- , HGN-232, letter, J. R.-Langley (HC00) to NRC, " Final NC00 Hydrogen Release.
Histories," April: 4,1988.
, 4 n -- ,~ HCOG/NRC meeting presentation', "8WR Core Heatup Code Results'," August 28,
-1984 I
-- , HCOG/NRC meeting presentation, G..R. Thomas (HC00), " Review of Basis for' "2ircaloy Oxidation Cutoff'in the BWR Core Heatup Code,"' January 14, 1985.
~
International Technical Services (ITS), ITS/ LWR /8NL 85-1,." Validity of the Use
'" .of a Temperature Cutoff for Zircaloy 0xidation," by H. Komoriya and P. Abramson, September 1985.
-- , letter from H. Komoriya (ITS) to L. Lois (NRC), " Review of 0xidation Model-ing in 8WR Core Heatup Code," October 30, 1987.
Institute of Electrical and Electronics Engineers, IEEE Std 323-1974, "!EEE Standard for.-Qualifying Class IE-Equipment for Nuclear Power Generating Stations."
. Mississippi Power a M Light Company, letter from L. F. Dale (Hississippi Power 1
- and Light Company) to H. Denton (NRC), "Ha Igniter Environmental Qualifications
' Test. Results," February 14,=1983. l
)
1 Sandia National Laboratory, letters from S. E. Dingman to A. Notafrancesco (NRC), ,
June '26, August 10, September 3, September 4, and December 23 (two letters), l 1987. .
, Science Application Inc. , SAIC/87/3114, "H l
-WaterReactorswithMark.III' Containments,gdrogenGeneratingEvents-forBoiling by Ateft et al., April 15, 19M.
. U.S. Nuclear Regulatory Commission, Title 10, Code of Federal Reculations,
.-10 CFR 50.44,'" Hydrogen Control Requirements." 1
-- , letter from R. Bernero (NRC) to S. H. Hobbs (HC0G) dated June 4, 1985.
-- , letter from R. W. Houston (NRC) to J. R. Langley (HCOG) February 21, 1986.
-- ..NUREG-0011. " Safety Evaluation Report Related to the Operation of Sequoyah Nuclear Plant', Units 1 and 2 " Supplement 6, December 1982. ','
, ,-- , NUREG-0588, " Interim-Staff Position on Environmental Qualification of Safety-Related Electrical Equipment," Rev.1, July 1981.
L
'-- , NUREG-0831, " Safety Evaluation Report Related to the Operation of the j Grand Gulf Nuclear Station, Units 1 and 2," Supplement 3. July 1982, and Supplement 5, August 1984. I i -- , NUREG-0853, " Safety Evaluation Report Related to the' Operation of Clinton L Power Station, Unit No. 1," Supplement 6 July 1986.
I 4 Appendix C
g .; . ;. . ..~ < '
'
- f ",, . ,
-- ,:NUREG-0887, " Safety Evaluation Report Related to the.0peration of Perry '
Nuclear' Power Plant, Units 1 and 2," Supplement 6. April 1985.
-- ,~NUREG 0979, "Sefety Evaluation Re M rt Related to the Final Design Approval. <
of the GESSAR 1* BWR/S Nuclear Island Design," Supplement No'. 2, November 1984.
-- , NUREG-0989, "S6fety Evaluation Report Related to the Operation of River
-Bend Station " Supplement 4. September 1985.
-- , NUREG-1150, " Severe Accident Risks: an Assessment for- Five U.S. Nuclear Power Plants," Second Draft, June, 1989.
L
-- , NUREG/CR-1659, Vol. 4 of 4 " Reactor Safety Study Methodology Applications Program (RSSKAP)," Grand Gulf Power Plant Unit 1. November 1981.
L
-- , NUREG/CR-4866, "An Assessment of Hydrogen Generation for the PBF Severe '
Fuel Damage Scoping and 1-1 Tests " A. W. Cronenberg et al., EG&G, April 1987.
l L- -- , NUREG/CR-5079, " Experimental Results Pertaining to the Performance of Ther-4
. mal Igniters," Sandia National Laboratory, to be published.
I Vinjamuri, K., et al., " Severe Fuel Damage Test 1-4 Data Report." EG&G Septem-l: ber 1987, i
l
^l l'
1 .
1:
I 1
l' l
l l
5 l_ APPENDIX C 1
.,bl1 gif ' .
. a' 9 % g h ;_; . ., '
We.w(*W: W
wws:,
R &. f '
\ '-
? ?q ', ,
-s
{
t <
} APPENDIX D' ;
o ;
ACRONYM LIST
' DADS automatic' depressuriration system dp BWR- .. boiling water reactor.
. f : SWRCHUC . boiling water-reactor core heatup code DWB1 : small pipe break,in the drywell y ,
l ': !PG emergency procedure guideline
> <ISF ~ engineered safety feature- ,
1 FMRC -- Factory Mutual Research Corporation:
- HCOG' Hydrogen Control Owners Group (Mark !!! Containment)-
HCU . hydraulic control unit .
'HDOL Jhydrogen deflagration overpressure limit HGE -hydrogen generation events
- HIS hydrogen ignition system
, 'NPCS high pressure core spray s
.o . ,
'3EEE Institute of Electrical and Electronics Engineers l q
c LCO . limiting condition of a <istion *-
LOCA' loss of-coolant' accident
~L ' LWR light-water reactor
- MiR . metal-water reaction .
1
+
NRC- -Nuclear Regulatory Commission I PWR pressurized-water reactor.-
' l L QSTF quarter-scale test facility n
.. iRCIC reactor core isolation cooling'
'; .m "RPV reactor pressure vessel
-l
,t. RS$ MAP reactor safety study methodology applications program -l M'3 RWCU reactor water cleanup a %- SAIC Science Applications Inc. O eport designation)
U SBLOCA small-break loss-of-coolant accident i_ _580 station blackout 7 5ER.
safety evaluation report g -5NL Sandia National Laboratory .
l l 50RV stuck-open relief valve l; 25RV' : safety relief valve L
I MarkIli SER- 1 Appendix D i
. ._, . - . _ . ~_ . . _ .
~ +
y,y y;;;th,
- i. :.. we,.= o 7, 1;.= .;s;,y;s y! f 7
= ,
g;9:p '
_ ::i C -i ; 6 "i%.
J , (g,', ._.;
i e w/i * * .g ,
! k S t. t 6- .. .
f"'",,_ -
'TAF- top of active fuel' le g;; 10-
.75.CFR Title 10 of. the Code ~of Federal Reculations - I e
- : technical specifications 2r- '2irconiur. .
4 n :a h '
l
$3 t I i j
' s q ';
k <,, ,
-l W!
l r -1 1
1 u.
l t j l
q i
w . El
/' .
e '(
1
-1 l
-l ,
-1 l
l
- q.
H
.I l
l 1
1
. 1 l
l Mark.'!Il SER 2 Appendix D 1
. . . _.