ML20246F577
| ML20246F577 | |
| Person / Time | |
|---|---|
| Issue date: | 08/31/1989 |
| From: | NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| To: | |
| References | |
| NUREG-1365, NUDOCS 8908310061 | |
| Download: ML20246F577 (36) | |
Text
_ - - _ _ -
NUREG-1365 t.
1 i ..
Revised Severe Accident Research Program Plan
- FY 1990 - 1992 U.S. Nuclear Regulatory Commission Omce of Nuckar Regulatory Research p?* "'*%f
% ..... f
\
gS3SI! *' e'on 136b R l
I AVAILABILITY NOTICE i
Availability of Refe ence MaMrials Cited in NRC Publications Most documents cited in NRC publications will be available from one of the following i sources: 1
- 1. The NRC Public Document Room, 2120 L Street, NW, Lower Level, Washington, DC 20555
- 2. The Superintendent of Documents. U.S. Government Printing Offica, P.O. Box 37082.
' Washington, CC 20013-7082
- 3. The National Technical information Service, Springfield, VA 22161 Although the listing that follows represents the majority of documents cited in NRC publica- ,
tions, I! is not intended to be exhaustive. '
Referenced documents available for inspection and copying for s fee from the NRC Public Document Room incluco NRC correspondence and internal NRC memoranda; NRC Office of Inspection and Enforcement ' bulletins, circulars, information notices, inspection and investi-gation notices; Licensee Event Reports; vendor reports and correspondence; Commission papers; and applicant and licensee documents and correspcnderice.
The following documents in tha NUREG series are available for purchase from the GPO Sales Program; formal NRC staff and contractor reports, NRC-sponsored conference proceed-ings, and NRC hooklets and brochures. Also available are Regulatory Guides, NRC regula-tions in the Code of Federal Regulations, and Nuclear Regulatory Commission issuances.
Documents avo;!cb!e from the National Technical information Service include NUREG series reports and technical reports prepared by other federal agencies and reports prepared by tne Atomic Energy Comrnission, forerunner agency to the Nuclear Regulatory Commission.
Documents svillable from public and special technical libraries include all open literature items, such as t2ooks, joumal and periodical articles, and transactions. Federal Register notices, federal and state legislation, and congressional reports can usually be obtained from these libraries.
! DoNments such as theses, dissertations, foreign reports and translations, and non-NRC coraerence proceedirips are available for purchase from the organization sponsoring the publication cited.
Single copies of NRC draft reports are available free, to the extent of supply, upon written !
retuest to the Office of Information Aerources Management Distribution Section, U.S.
Nuclear Regulatory Commission. Washingtnn, DC 20555.
Copies of industry codes and standards used in a substantive manner in the NRC regulatory I process are maintained at the NRC Library, 7920 Norfolk Av9nue, liethesda, Maryland, and are available there for reference use by the public. Codes and standards are usually cooy-righted and may be purchased from the origuating organization or, if they are American '
National Standards, from the American National Standards institute,1430 Broadway. .
New York NY 10018. i
t
)
( ,- I NUREG-1365 L
ny. -
Revised' Severe ~ Accident L Research Program Plan FY 1990 - 1992 Manuscript Completed: May 1989 Date Published: August 1989 .
Division of Systems Research Office of Nuclear Regulatory Research U.S. Nuclear Regulatory Commission Wcshington, DC 20555 p** *%,
W.....J I
i
. i
Abstract For the past 10 years, since the 'Ihtce Mile Island acci- nical basis upon which decisions on important contain-dent, the NRC has sponsored an active research program ment pet formance issues can be made and the long-term on light-water-reactor severe' accidents as part of a multi-research needed to confirm and refine our underrlanding faceted approach to reactor safety.
of severe accidentr In developing this plan, the staff rec-This report describes the revised Severe Accident Re- pgnized that the overall goal is to reduce the uncertainties search Program (SARP) and how the revisions are de- in the source term suffic ently to enable the staff to make i
signed to provide confirmatory information end technical regulatory decisions on severe accident issues. However, support to the NRC staffin implementing the staff's Inte_ the staff also recognized that for someissues it may not be gration Plan for Closure of Sever: Accident Issues as de. practical to attempt to further reduce uncertainties, and scribed in SECY-88-147. 'Ihe revised SARP addresses some regulatory decisions or conclusions will have to be both the near-term research directed at providing a tech- made with full awareness of existing uncertainties.
1 i
l 1
1 I
I iii NUREG-1365
v 4s CONTENTS Page Abstmct....................................................................................... iii
- 1. I n zrod uction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
- 2. Goals........................................................................................ 1' 2.1 N ear. Term Wo r k . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 2.2 12m g-Term Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 2.3 Results of Previous Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
- 3. Mee t ing Nect. Term G oals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 3.1 Perspective s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 3.2 3evere Accident Phenomena-Core Melt Accidents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 3.3 Iss u es a n d Tasks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Iss u e 1 - S caling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Issue 2 Depressurization and Direct Containment Heating . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 Issue 3-BWR Mark I Containment Shell Meltthrough . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 l
1ssue 4- Adding Water to a Degraded Core . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 Issue 5-Use and Status of Severe Accident Models (Codes) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
- 4. M eeting 12m g-Term G oals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 Issue L1-Modeling Severe Accidents. . . . . . ... . ..... ................... ..... .. .. .... . 19 Issue L2-In-Vessel Core Melt Progression and Hydrogen Generation. . ................ ...... .. .. 19 Issue 13-Hydrogen Transport and Combustion. . . . . . . . . . . . . . . . . . . . . . . . . ...... .. . ....... .. 20 Issue L4-Fuel-Coolant Interactions. . . . . . . ........ . . ......... . ...... ........ ....... 21 Issue L5-Molten Core. Concrete Interaction (MCCI). . . . . . . . . . . . . . . . . . . . . . ................. .. . 21 Issue IE-Fission Product Behavior and Transport. .. ...... .......... ....... ....... .......... -
21 Issue L7-Fundamental Data Needs. . . . . . . . .. . ... .... ..... ............... .............. 22 APPENDIX A Background on Severe Accident Research Program and Relationship to Other Elements of SECY-88-147 . . . . . . ....................... .... .... .. ... .. ......... 23 APPENDIX B A Severe Accident Scaling Methodology (SASM) .. . . . . . . . 26 Figures 1 Severe Accident Program-Schematic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. .. . .. .. .. 6 2 Revised SARP Near. and 12mg-Term Programs . ....... . . .. . ..... . . ... . .. 7 B.1 Severc Accident Scaling Methodology (SASM) .. . ... .... . .... . . . . . 29 Tables 1 SARP Milestones and Estimated Costs . .. . . . . 2 v NUREG-1365 l l
i
L Introduction costs and significant milestones for implemen.ing the revised SARP are presented in Table 1.
In a memorandum dated July 8,1988, the Director of the i
Office of Nuclear Regulatory Research charged the Divi- 2. Goals sion of Reactor and Plant Systems (now the Division of Systems Research) to develop "a detailed plan identifying In developing the revised SARP plan, the staff recognized individual goals, discrete products, and anticipated sched- that the overall goal is to ieduce the uncertainties in ules for the Severe Accident Research Program (SARP) source term sufficiently to enable the staff to make regu-that clearly demonstrates the implementation of the latory decisions on severe accident issues. However, the SARP portion of the staff's Integration Plan for Closure staff also recognized that for some issues it may not be of Severe Accident Issues (Integration Plan) practical to attempt to further reduce uncertainties, and (SECY-88-147). [ Preparing such a plan will require) a regulatory decisions or conclusions will have to be made detailed review of the current activi;y of the SARP against with Iull awareness of existing uncertainties.
the framework of the Integration Plan and the detailed activities and whedules of the other five elements of the As the Severe Accident Research Program represented Integration Plan, [together with) an assessment of what in this plan is a goal-oriented program, it is critical that redirecting /reprioritizing/ rescheduling within the SARP such Foals be clearly stated at the outset.The staff expects may be needed to fulfillits portion of the Integration Plan to use safety goal policy and objectives, as appropriate, in and how and when such can be accomplished." His re- determining what potential improvements in the techno-vised Seve.e Accident Research Program plan reflects logical base are needed for closure of severe accidents the evolution of the sevcre accident research program and in order to ensure that the safety of existing plants are ,
and was developed in response to that July 8 memoran- reasonably consistent with the safety goals. The goals of I dum. It will be used to guide the formulation and conduct the revised SARP are as follows: ,
of severe accident research, addressing in the near-term l those issues pertinent to the implenantation of the Inte- 1. Provide the technological base for aesessing contain-gration Plan: viz., individual plant examination (IPE) and ment performance over the range of risk-significant containment performance improvement (CPI) activities. core melt events.
The plan also will be used to guide the formulation and conduct of the program oflong-term research needed to 2. Develop the capability to evaluate the effectiveness support the NRC's accident management activities and to of ger.cric containment per formance improvements.
confirm the Commission regulatory decision on severe ac- ,
cident issues.The plan provides a description of the gen- 3. Provide an understanding of the range of phenom-eral approach to the needed research. Naturally, many of ena exhibited by severe accidents that includes the l the details associated with specific experiments or analy- impacts of generic accident management schemes, ses have not yet been developed and can only be devel-l oped after the specific research is identified. Therefore, 4. Develop improsed methods for assessing fission for each near-term issue, a detailed research plan will be product behavior and availability for release in the developed, describing each experimental and analytical event of containment failure at various phases of program and how these programs fit together and lead to severe accident sequences.
a resolution of the issue. The plan may also give the im-pression that we are just now entering Ihe arena of severe Goals 1 and 2 above are primarily near term, while goals 3 accident research; the facts are quite the contrary. A tre- and 4 are longer term objectives that aim for additional mendous amount of severe accident research now exists, depth of understanding of both accident evolution and fi-and this plan is intended as a needed step to the critical nal consequences. These goals are consistent with NRC's and focused reexamination of our further needs and the Integration Plan for Closure of Severe Accident Issues regulatory questions involved. (SECY-88-147) and the gu: dance provided in Generic i Letter 88-20 for individual plant examinations. It should '
be noted that there are not two SARPs, a near-term and a nis plan is organized as follows. Section 2 presents the long-term. Rather the revised SARP draws from and fo-i goals of the SARP and a sketch of its structure. In Sec- cuses specific tasks of the continuing program of severe tion 3, the work directed toward achievement of the near- accident research experiments and analyses to address term goals articulated in Section 2 is described. Research the near. term implementation of SECY-88-147.
dealing with the longer term goals is covered in Section 4.
Appendix A te this plan provides background information 2.1 Near-Term Work on the existing SARP and the relationship of SARP to other elements of SECY-88-147, and Appendix B is an Subsequent to the Three Mile Island Unit 2 accident exposition of the role of scaling in the SARP. Estimated in 1979, the U.S. Nuclear Regulatory Commission 1 NUREG-1365
d2 e9 ~
t cO K K K K K e
p 0 0 C 0 0 K K K K K 0
S 7
0 xM 0 8
0 8
0 0
0 0 8 5 0
0 0
0 3 0
'4 4
7 T
EI( 1 1 3 1 2 6 8 5 6 n
o di t e e t l c p 9 9 pm e 1 9
8
/
0 9 9 1 0 9
0 9
0 9
0 9
8
/
x o / 2 / / / / / / 2 EC 6 1 9 9 9 6 8 4 1 s
lb e
a r
s e 9 t v 9 9 9 9 9 s : i i
9 9 9 S 8 8 o c 8 S
/ 8
/ P 8
/
8
/
8
/ / /
/
C d. D 1 3 6 9 I 9 9 8 9 6 d
e t
a -
m e s - d it e r s e rd a n s hu n n r e E ti t
u d uedn ito a co u? a t
e f
o ig d n m n dien ut r ds f n a m a omp u trca e l
nd ao os n H C
o s
e nre xe c la seh f rps yh rtn ia ot m rn D
n o
dt nea l coeo ct a e n emo a c
tt e it s l a oi o i
s ad n n thm1 ct fimo mir pi vr t n hr t t n
e cr l
oio ito fe t nak f et i f eo amr i ci i
e tatit a fi s el a op s ir er tn a giibz nsd emd lu ye eT v x ce iu iM e n su sr e g e ,r lt u wss P s maa cr damI t i r os ny tyu s e no er R
A lp y irf rgo c eo i
t u
wn c s f is l ops e io d r ic t
n n T
pep l
ae pf r aa t a e t
S p a xpe ru ts e ny exop prs sa oot se r m ul t t
s el e ud st d et n s c ola sed er qod d es emae e o ai sn 1 t e n n hnt et e a t nncnnc a n e ed a o n a o n l
vif l
b io a p bomi pr mr mluny ien so nfh ros dmba g afilagi u o scC io ea t t T
a ip r l o so sl o p i ifrn ei yt er csue n t lys . ero rgmjs m ii y idr c e f el vo s
e v
e eer svp n pd viv s ep o xnao a o n n l
e a t n no emto cn an or ro ti t D D Ada C Eact C ci Ai D itb t I a Pf l
i s
C t n d D e a ls c e is d
n h i m I I d s
c n r de ly a r o ep rh o a a n ec s
n o e it a xE e
wa m n on is it s
e l io H oer H iA t
y u t a) C at la a ntR c r d e izl m o
lB zf i n ir es i r a t
af o D ire A 1 u mge C la c u
son t
t o l ae e un se 2 g g s nne l a S si Ber irm d s eB E E e s r e r n t
a n n r iai u H Vl ei u r r/
U e i i la l U p nd t
t pe t aT g p
pst eo S en t
l t a S e oh a C Pa i
S iT c c SS S
- L : L 1
A 2 A ks e u T e u
T O 5 6 7 8 O
_I s 2 3 4 s 1 a s s T T I 1 T 2 2 2 2 2 2 2 2 ZC% 9 M" w
d2 t
e9 '
c K -
K K
- K eW p 0 K K 0 0 K K 0 K K x) 0 0
0 0
0 0
0 5
0 0 O 0 0 5 5 E (F. -
1 9 6 2 1 i
P 2 0 0 2
5 8
5 8
n
, o t
e e t l
_di cp pm e
x o 0
9
/
0 9
0 9
0 9 9 0
8 9 0 9
EC 9
/
6
/
6
/
6
/
1 9
/ /
3 s
l e
b a
r e
v 9 9 9 1 i
l 8 8 6 9 9 9 9 t / / /
e 2 8 8 8 8
. hD 1 2
1 2
1
/
9
/
9
/
9
/
9
)
s lt 4e l e n d yd m n nf b it 2Pi Vn k r
a i
s el s r s :c n d) t n
n o vl s o nl aa t e o sHl te s t ee n C
(
e dRm wn net io msg.
o wv it cs n mr t e
he ed pm at io he x t rl o 1
e t
uW#
sB s lat t c na i
e r e eod rndu t t
rn ee nma t
at e
e f oe osn v l l afoic e pnt e ii rd r r c r h t e b
a t n ust i mtn x ee ezrc t r
e p eoa f
t e nr n epsads T _
e t'i ri e r tspsp cf t ax ea l nie me miiver pe axpe a e n x nc mn c sCd o s
i xt we d imnr eRc t
irt c r o esai ca f r r .
t l e a e ct. c sNs pnr d w. o etd wta e ed ahb l s
~. n o
hg e shxe amnt c nlu ale ofh nn e n u
mmex cie c diy fo n o u l
t ai n hed e
- ts (mc r oa ai t
i u ' ate noelpig ami t a oc d I t 3 pe c n er c t f ic o cit c n st t
ek herime ip i t t r i3 s n o sti s yt isr i de d. aa aerr nr4eM yr t u ne e e ymao im1 pca r
c hglyi ulir ;yc Cirie r l l D
s e Mnt! AdfIewnn e c As e Eoxn c ei sad n se abc ai Atso na f a Ao n fe r t nt nc nAes v
ne e r e h o d i
S ,
C c c
t n
I c A e It d e m c a r e e i
n y ad u M n
t 3
r g n e o - S v
e r
e v
e t
c Dp nt u f S oS r m n
oU t s o ef s Cmu e ia t
r n u
r s
u u ro e ei o t
I a C n t td k n r e W g onru t e t a a t
f y
t o t
t a FeC u no Mr al P 't n Ci oi d IF l
a 3 We gnd r
ye o
t l i it laa icg 4
S d npnd m t 5 t
c e a kl E n e o ib itit S n aloe E le R e e rl U iGloic l i b u r s U a e ed U t
Wo w f r f p a he S d d 2e a r f c e ev S el vc s oec i
S ES MS i
S T AHR SC RI n S t
BL S I I URDA I 3
L 4 L : L k e A e A 5 A
s u T u T e u T T
a s 1 s
2 3 D s1 s
2 3 O s s O I 3 3 3 I
' i4 4 4 T I 5 T w y:mo=
d2 K' K t
e9 -
K K K K K K K 0 0 0 0 8 8 0 6 1 c0 0 0 0 0 6 3 0 0 8 e9 0, 0, 6, 5, 5 7, 2, pY 0, 0, 0 5 xF E(
6 7 3 3 6 4 1 1 5 6
)
d e g u e n int i
t dad n ni o tala C
(
sv r e e y, ar n 1 dg o le nos ulou e i
t a
b r o ns f
a e ei s n T )
s hme e r
a e
t poh G y de e n T w. n e
3 r h sl.o g e pt e o f
o n ht no cb r o t d
_ n itp r etde u
y o f ds l t r
it b I o
c ir c k cyr icisi d n p j
e s eaas l c n io s e ef ne a t n n m
p D Soad n u s o ar c T
_ o b it
_ h is m it s a r d w s e o te a n
t n r g C s n r s
_ m a
e or d nn o I io d e
_ r n id P a it e v e g
o soh ci c c li t c a
t er ha N t isc op r i
r p nce a r A eM e nc c a a r
e e l t itDse n o
_ t I y a r s C c D r
o r e v
er na I t
u M e
cryR Se Co t l t
a r
n e d a R m
n o er t
T r a o o n t E
Ua lo C rP em g
i r
f lut u l n se e n g o i f
ogu e or -C nte n o da G i
n l s hc !cx V r e eF N
_ C le Rd dy ulc is n o
lo si Fu lt
_ ( i n n r k n O
_ l s T Ria a h e I l
I F MF L a s o e G h m m r r e e T L L2 3 5 6 7 1 4 L L T. -
1 L L L L A A g g e e e c e e c n k s n u u su us u s u us T T o a o s s s s s s s
s s
s O O T T T L LI I I ! I I I 7Cy?{> ., s
- undertook a broadly based research program to develop 2.2 Long-Term Work an understanding of severe accident behavior. Over the past decade, major experimental and model (code)devel- Achievement of the longer term goals demands a broadly opment programs have been performed that provide a based severe accident research program. To be respon-greatly improved understanding of severe accident sive to regulatory needs, particularly confirmation of ck>-
phenomena, reflected in an ability to model those sure of severe accident issues, that program must both ex-
. phenomena. Now, as the Commission is preparing to piore severe accident phenomena and develop the meth-close on severe accident regulatory questions the odologies appropriate to quantitative assessment of se-immediate priority of the NRC's SARP is to support the vere accident risks. The plan to guide such a program closure process. That process
- is displayed schematically must reflect a balance among the risk importance of phe-in Figure 1. Hence, over the next 3 years, a significant nomena being investigated, the desire for technical com-portion of the revere accident research effort will be pleteness, the expected likelihood that the research will directed toward issues that relate to the dominant topic result in a s;gnificant reduction in the uncertainty of risk, area of the Integration Plan, viz., the accident sequences and the need to promote the maintenance of trchnical ex:
that lead to early containment failure: direct containment pertise and analytical capabilities. He features of the heating (DCH), BWR Mark I containment shell SARP that address the long-term goals are presented in meltthrough, molten fuel-coolant interactions in BWR Section 4 of this plan. Figure 2 depicts the elements of Mark II and Mark III containments, and hydrogen this revised SARP and the relationships among thom.
detonation in BWR Mark III and PWR ice condenser containments. Note that the hydrogen behavior issues for accidents in which the core is degraded but is not into the 2.3 Results of Previous Work severe accident domain have been extensively investigated, which resulted m, the resolution of USI Review of information available from previous and ongo-A-48, " Hydrogen Control Measures and Effects of ng research, including that sponsored by U.S. industry or Hydrogen Burns on Safety Equipment. The additional foreign organizations,is being endertaken by the staff and hydrogen research discussed in this' revised SARP is expecte'l to be completed by the end of FY1989.Tbc pur-pose of this review is to identify and redirect tLe rescarch intended to address the hydrogen threat from severely damaged : ores and will serve to reduce tmcertainties in program as appropriate to ercure that the needed infor-the threat to containment integrity resultmg from various mation to reach closure or resolution of severe accident hydrogen combustion mode (s) for a wide range of acci- issues important for regulatory decidons is developed, dent conditions. In the near term, severe accident re- 'IMs review will also ensure that the various ret,eir6h search will also support the resolution of broader, snore prnjects are consistent and well integrated a rnong them-general questions regardmg the response of various con- selves with a common goal of ultimate 4 leading to closure of severe accident issues and that the lorg-term confirm-tainment designs to severe accidents. Task descriptions have been developed for the near-term work on these is- atory research is properly focused.
sues and are pres;nted m Section 3 of this plan.The ma.
jority of the near-term severe accident issues are consid-cred relevant to evolutionary LWRs currently under re- 3. Meeting Near-Term Goals view by the NRC. Therefore, these near term efforts ,
are equally important to evolutionary LWRs. We are cur. 3.1 Perspectives rently not aware of any features of the evolutionary '
LWRs under development that are unique with respect to This section of the SARP plan presents those issues ofim-severe accidents and require special research programs. mediate priority that need to be and can be tddressed in As such, no funding has been allocated for work in this the next few years m order to achieve the near-term goals area. However, we will keep abreast with the develop. set forth in Section 2. Tbse are the issues that are perti-ment of these e evolutionary LWRs and notify the Commis- nent to the NRC's programs of individual plant examina-sion if our current perception changes. tions and containment performance irrprovements (di rect containment heating (DCH)and hydrogen transport and combustion, issue 2, and DWR Mark I containment she!! meltthrough, issue 3).The research to be performed
- ln S13CY48-147, the staff identified the steps that each Ikensee is ex- to address those issues is also presented. This research is pected to take to achieve a closure <m severe necidents Ic. its plant.
primarily directed to assessing containment performance I- mpletion of the individualplant examination (Ip10 and Hentifi. Under severe accident cond;tions and is constructed to en-cation of potentialimprovements, able as definitive a judgment as is possible about the l 2 - Development andimplementation of a franework for an accident potential threats and likelihood of early containment fail-management program that can accommodate new informatur. as ure in the event of a severe accident given the present 3 In e$n tNi n any generic containment performance nn. paucity of (or sparse) data describing core melt progres-provements with respect to severe accidents. sion. Also included in this section is a discussion of the 5 NUREG-1365
i^ &
s S S h N W1G O
I E
- )C T VE
A TN V N I
_ E EC
_ G MU E 'S /
N I GO T U AN T N N LPNG " N E OA AAI E D NSMD D I
I GC O
C - - -
C NCH C I A C U
ES T EE N VS VO OEE EL j - , .' CGR SC g"
'g i
. 3 ::
( ) [
)$'.
B T
D'/
i S
w[ -
TT NN E F 3
G N s MM ;g E t_
i 4 NE I V j s TM t s
(
RAO ::
I NE ETR I.. EG 4j -
C N 4 OCI M C A si AM (
8
_nj kr i .
< C N / f[ - li:!
l:
I3 i
- i AI L / .. s gii
- s$;$
l LI
- .3i
- iS
/
PS - :
l Y A f g
- qN l
FR I E T
- e3<
f<2:
N N L / ghf l E D U /
!!9 8si l I V
/ 3:
. xi :
$ $( '
[
i 5
5 I 5 S K DC E R S E AT V P -
O S M EN f
1 RH PC
[ ;
LM p
A I E j
t
_ ME I T 5,
p T V NO 5 ER 2;
PI T
_ S
(( / , N
_ .X:
..:* E l
/
' p" D
I C
OH i
_ T 1 f p" g
AC N
A
/N A S /
ER l
PS / PNGL Ni Ei d RA EE
- g.
~/
S VS D N
/ LATI
_ I EE O L U A%A S /
_ E SR :
I x
VT D :..s '
/ O Nt N N OA R h g.. 8
/
I HR VMR
/ D A E ET AZ /
_ ~ - .
_ PE I
_ P MI O / NIX EI T*.A N EH
/
l/j/// '
ZC O L. $
- 5 I
O T K U C MN T T
GNU R
S NC C EA T EI NIT RAA N A RE N
OM Y TNEF CI A ATA T
D NB Y
HWC M
TE O
C CN E
T L
O M
R E
T N
I E
R ONS C
E D
)! E
^ SRNOWG E
WVECC E P C 1 E
T C
E S
E I
A SCC T A
& G+ R L
E T
I V
S M UO E I OC T
T R M A R N:SC M &H HD I
E RN E YLA H OE T AECTE M GGO T ST FM -
RPDLA D AMOO RDLT N TSIN R E GNO SIE I
T C G D E RN N AEA T U GS OT ONMG O OAN I G
E HS R ALDT E I RTE SY G M
E T BO ANO N PT N N LEO U S PSM L C S I
- LE L RO ERL D A C ME OISH T B RNT MEN NA RDE A P U AG N S EI R M ENH AM R S CN A I I TR TZ A TE O TUP D T A NNS I NTC -
GH TE E N S EHS G RE A P RUMWE EA N NT NT O CEDN I ES SN OP E A D W T ES TSA G L E DID C I
' AO N UE SI EA TMGO R D C 4LC L S E VA PWDD 9
G S S NC A
NERN T I EODY R T
(
DCAH E E R H ERID O I
L ####
^ E D VD E E I
T EVP R VOP OE SC C UERU M A.
FSPS c
T DN E NA2.UR h #\
i A. O L l TO L CA I
l E F T MR M N L
~
N 2
HN CDM OE E TO L/O UNR E TT R U ETI DAO F N T I
L U EOR EMU I TOE V E D N VC L CT N RO O HAI OR SE M I
C l
O N I E
M I C P SB C I
H F
3 h: - - - -
issue of scaling of severe accident experiments and analy- reactor coolant system pressure. At low pressures, the ses (issue 1). Obviously, the question of verisimilitude in core heats up more or less independently from the iest of severe accident research is critical, and scaling is one cle- the primary system, and the process is acce9tated by ment of the same. Whether the correct physical and metal-water reactions. At high pressures, energy redistri-chemical processes are being investigated is another and bution by steam from natural circulation may be signifi-is appropriately dealt with on a case-by-case basis. But cant, transferring core heat to tin reinainder of the sys-scaling as an issue is common to all areas of severe tem, including upper vessel inornals and, for PWRs, accident research since full-scale experiments simply are steam generators. This transfer 01 energy affects the pro-i not feasible. For this reason, a systematic examination of gress of the accident in r,creral ways. It delays somewhat scaling of severe accident experiments and analyses is the onset of gross core degradation. It affects the steam included among the key elements of the near-term availability in the inctal-water reaction process, tending rescarch. This element is not intended to constrain to limit hydrogen generation. It affects radionuclides re-closure expected of the other near-term issues herein but lease and transport within the primary system. Perhaps is expected to t;c of benefit to decisions made on moct importantly, it also raises the possibility of compet-recommending further experiments over the near- and ing high temperature and stress-related failures of the long-term periodr, addrersed in this plan. primary system pressure boundary (e.g., RCS piping, steam generator tubes)in PWRs with attendant depres-In addition, this section contains a discussion of the issue surization, hence avoiding the possible ejection of molten of molten fuel-coolant interactions (issue 4)and a discus- core material from the reactor vessel while under high sion of the issue of the use and status of ceveic accident pressure. Current understanding suggests that for BWRs codes (issue 5). The molten fuel-coolhnt interaction issue natural circulation is restricted by the fuel assembly wall is included here because it is directly relevant to consid. channel boxes. Although natural circulation may affect crations of intentional reactor vessel depressurization to the core melt progression, failure of the reactor coolant cope with DCH, as well as low reactor coolant system system due to excessive heating from high-temperature pressure addent ceguences in general. Specifically, the naturally circulating gases would not be likely. However, ouestion of what happens when water is added to molten because of the enhanced reliability of the automatic fuel (in-vessel or ex-vessel) appears to be significant to depressurization system (ADS) suggested from the con-understanding accident management and some severe tainment performance improvement program, high-accident sequences. How to answer that quertion is not pressure scenarios are not likely to be dominant contribu-immediately obviot,s, ahd some initial work needs to be tors to early containment failure in risk assessment of done to allow development cf a rational approach to the BWRs. Hence, high-pressure scenarios in BWRs are not research. As for the issue of coder,, M is recognized that being addressed at this time.
code development is integral to the achievement of long-term gcals, since the understanding of severe Although the failure of the reactor coolant system (RCS) accident phenomenology ultimately is reflected in those by natural circulation appears to be possible, the ability to caicuhtional tool;. However, achieving the near-term predict its occurrence has not been established to the goals will neecsskate rcasoned judgments based pomt that confident conclusions regarding the likelihood essentially upon our present understanding of severe f RCS failure can be made. Moreover, while analyses to accidents and an essessment of just how well present determine the extent to which natural circulation oe-codes describe the key behavior of severe accidents.'the curred m the TMI-2 accident are under way, it must be confidence that can be placed in me severe accident codes noted that no evidence of high temperatures m the range also directly translates into confidence that can be vested f interest to cause failures in the primary system compo-in the near terre judgments tibout containment nents has yet been seen. Ihus, competing RCS pressure performance. Ihr tins Rason, the code assessment issue boundary failures from high temperature cannot be pre-is included in this section. sumed with a high del ree of confidence, and scenanos m-volving gross core degradation or core melt at high or in-termediate RCS pressure must be considered. Since such 3.2 Severe Acc.i dent Phenomena-Core Melt high-pressure PWL scenarios have the potential to pro.
Accidents duce ejections of molten core material into the lower con-tainment following lower vessel head failure, a variety of Before the issuci, and research tasks to addiess them are concerns regarding the possibility of containment over-discussed, a g6neral descr.ption of the important phe- pressurization havd been raised, in particular, direct con-nomena associated with core melt accidents to be dealt tainment heating. In fact, the high-pressure scenario with in the near term u presented.This is done to provide seems to influence risk in terms of the variety of mecha-a pe2spective and context for the issues and tasks. nisms with potential to overpressurize the containment and thus importantly contribAes to the uncertainties in !
In light water reactors, core melt acc; dents can occur preuct ed risk. In addition. core melt accidents at low RCS cither at low reastor coolant system pressure or at high pressures that lead to early containment failurc have l
MUREG-1365 8 m
been identified as potentially important to risk, in particu- system pressure scenarios are likely to dominate the lar, BWR Mark I containment shell meltthrough. probability of a severe accident but also give rise to early containment failure threats.
The following observations are intended to supplement the above general perspectives: 3.3 Issues and Tasks
- 1. At this time, competing primary boundary failure 10- Issue 1-Scaling cations, size, and timing are highly uncertain, par- . .
ticularly considering the role of relocated fission Full-size severe accident experiments are impractical or products and the possibility of loss of integrity of simply not feasible for many reasons, particularly costs.
steam generator tubes. 11cnce, severe accident research must rely on small-scale experiments that are carefully crafted so that the conclu-si ns drawn from those experiments about the phenome-
- 2. Core slump at high pressures may yield significant fuel-coolant mixing. Triggering of massive steam ex-n logy i severe accidents can be appropnately applied to plosions et such high pressures is generally consid- predict severe accident progression in typical nuclear, cred highly unlikely by the experts, but the potential p wer reactors. I'he confidence m this application is greatly enhanced through scaling analyses. A key element driving forces associated with such core slump inter-f the SARP, therefore, is a scaling methodology for se-actions may not be altogether ignored in the process of understanding fission oroduct relocation and the vere accident experiments. In those cases where an ac-potential for altering the' accident progression. ceptable methodology can be developed, it will be applied to relevant expenmental programs. As necessary, recom-mendations regarding revisions to experimental pro-
- 3. Natural circulation may, for example, cause a weak-grams will be made, and needs for new facilities will be ening of the upper internals and upper vessel head addressed in terms of the regulatory value in pursuing this (e.g., seals), with the concomitant result that com-research. A useful scalmg methodology should be capable peting relief paths may be formed that influence f providing the followmg.
both the fission product relocation and the potential for direct c >ntainment heating.
- 1. A scaling rationale and similarity criteria.
- 4. Irwer vessel head specific failure medes and the timing remain highly uncertain. local failurcs of in. 2. A procedure for conducting comprehensive review strument guide tubes could occur. However, melt at. of facility design, test specifications, and results.
tack of the vessellower head forgieg (in the absence oflocal penetrations) could lead either to a local fail. 3. A measure or index to indicate the applicability of ure of the forging or to creep rupture of a large por- correlations or models based on test data from sub-tion of the lowervessel head when heated to 600 to scale facilities to full-scale nuclear power plant con-700*C. Large failures at high PCS pressure can po- ditions.
tentially create thrust loads and threaten failure, for example, through overstress or tearing at piping 4. Quantification of the effects of scale distortion.
junctures.
- 5. Quantification of the effects associated with ex-
- 5. At high pressure, a large coherent failure of the trapolating correlations or models beyond their data lower head would likely be accompanied by prompt base. This information is needed for quantifying expulsion of melt, steam, and hydrogen. High- code uncertainty.
pressure expulsion has been postulated to lead to di-rect heating of the containment atmosphere and an Although scaling analyses have been performed for some increased potential for hydrogen combustion and of the severe accident research experiments, there asyet perhaps even detonation. is no systematic applicatica of a scaling methodoloD that would allow confident extrapolation of observations
- 6. 'Ihe ex-vessel course of accidents in which reactor made in intentionally scaled experiments to severe reac-vessel breach occurs with low pressure in the reactor tor accidents. The work to be done is to first determine coolant system is highly dependent on the mass, whether a scaling rationale and similarity criteria for se-thermal, and chemical properties of the raelt upon vere accident research can be developed to provide assur-exit from the reactor pressure vessel. Direct attack ance that the correlations or models deseloped ctm be of the Mark I containmer.t shell; molten fuel- properly scaled to reactor conditions. Second, if such a coolant interactions in the BWR Mark 11 and Mark methodology is developed, a systematic application of the III containments; and hydrogen combustion in BWR same will be made to appropriate experiments and com-Mark Ill containment. are examples of where low puter codes carried out under the SARP.
9 NURIIG-1365
a .&
e'-
P Task 1 Develop and apply a general setere accident this precess exists, and competer code predictions are at "
scaling methodology. best uncertain. Heuce, prior to dismissing crust failure on ,
the bottom portion of the crucible as not a plausible fail-Research Approach ute mode, this is.<ue will be assessed.The mam difference between these two failure modes cecms to be the amount Scahng methods will be reviewed. This review will be of of molten material that pours into the lower head and is '
scaling methods h. general and methods used in the ex- ultimateiy available to be released to the containmera. ;y periments of the SARP to date. Frorn examination of the Our current understanding indicate; anywhere from 4 nature of the phenomena being studied, the experiments around 25 to 6fl percent of .he core could relocate to the ', '
being conducted to study those phenomena, and the rc- bottom head upon crucible failure. Becauw of the diffi-view above, scaling methodology will be developed for the culty of conducting ap.propriate teste to better .
SARP.The efficacy of the methodology will be tested in- characterize these phenomena, questions associated with I itially by applying it to the direct containment heating ex- core relocation are likely e remain highly uncertain in the periments that have been conducted in t+ S ARP to date. near term, although modeling of these processes will continue under the .SARP. Thus, the question of lower This task will not address directly whether the correct head fai'.ure mode ctm be approached by considering a phenomena are being investigated in a particular experi- sufficiently broaci range of debris quantitics, composition, ment-that is done for each task throughout the SARP- temperatures, and relocation rates. Thc 1M1 lower head but rather w hether the experiment will indeed yield infor- inspenior, plays an important role in understanding the o
mation about the phenomena being investigated relevant rek) cation process and the threat to the lower head integ-to behavior at the scale of an actual reactor, rity. On the ether hand, depressurizatior of the RCS, either by structural failure of the pressure boundary (e.g.,
Usables and Use piping, steam generator tubes) result l Jag from natural cir- i culation or deliberate depressun7ation by operator ac- a, To the degree that this work is successful, the S ARP will tion, may alleviate concerns over direct containment have a scaling methodology that can be applied to the ex- heating (DCII). In this regard, any analysis with respect to periments of the SARP to provide a more rigorous and depressurization would be assessed in terms of the cur-systematic link between experiment and accident phe- rent state of operator training, including information nomena than is present today. The application of the available to the operator Fnstrumentatics), and it would methodology will help in the planning and conduct of the include the dominant accuent sequenccs anticipated for experiments directed at the longer-term goals, in the nc? r a given plant.
term, application of the methodology should help estab-hsh the level of confidence that can be placed in the pre- As part of the develor: ment of the reactor risk assessmertt ,
sent understanding of containment performance from document (NUREG-1150), the DCli issue was pre-scaled experiments and code analyses performed to date. sented to an expert panel that included experienced se-vere accident analysts, Monte Carlo analysis that inte-Issue 2-Depressurization and Direct Containment grated all the pertinent phenomena was performed.The lieating resultant likelihood of containment failure was found to be small, and the chance that containment would survive Present understanding of the in-vessel progression of a the loads associated with DCil is much greater than had PWR core degradation accident is limited but suggests been originally estimated and reported in the first draft of the fonnation of melting-solidifying fronts as the molten NUREG-1150. The experts based their judgments, in-core materials " candle down" into the lower, colder por- cluding the likelihood of RCS failure by natural circula-tions of tne fuel bundles and solidify.1his " frozen crust" tion leading to RCS depressurization, on the current body of previously molten core material may form in a layer of research evidence.
that inhibits steam flow from the lower plenum. Unsup-ported fuel pellet stacks may collapse onto the crust and Consistent with the above,in the near term the SARP will melt. As this process proceeds, a crucible-shaped crust address the DC11 issue by a two-pronged approach:
supporting a molten mass of core material may form in ,.
the central region of the core. Failure of this crust at a ra- 1. Before we can rigorously reach a conclusion on the dial kication (" sidewall") is expected from considering benefit of intentional depressurization of the pri-thermal convection of the melt within the crucible. In mary systems of PWRs, an improved und,trstanding fact, this is what was concluded to have occurred at TMI-2 of the challenge to containment integrity must be upon examination of the reactor core. Upcm crust failure, obtained. The existing research efforts will be criti-molten core material would flow downward, possibly cally examined with respect to the following; ablating through the core shroud and flowing onto the flow distribution plate, eventually reaching the lower a. W'nat is the nature and character of the in-head. Ilowever, very little expciimental information on formation most likely to be provided on NUREG-1365 10 -_ -
high-pressun core melt accidents in the next ex-vessel that will mitigate or eliminate it? (rasks several yea,s? Are continued commitraent of 2.4-2.7) time and resources iikely to result in a commensurate advance ~ ent of understanding Imgically an answer to this last question should precede of DCH phenomena? the rest. Hcwever, even with DCH apparently presenting much less of a risk than previously believed, there is not
- b. Are the existing research programs addressing yet sufficient confidence in thisL: lief that it warrants de-DCH optimized with respect to providing the lay in sechirg r.nswers from the other relevant tasks, espe-best information for closure of the DCH ques- cially in light of the 3-year horizon of f he Integration Plan.
tion in the near term?
Task 2.1 Evaluate current research program address.
- 2. The efficacy of PWR depressurintion will be exam- ing hlgh. pressure melt ejection challenges to ined. In particular, both beneDeial and detrimental containments.
aspects will be explored. Review of relevart emer-gency procedures, domhant accident scenanos, and Rcsearch Approach timing of required operator actions to depressurize also would ba, inclnded in this examination. Current research programs, both experimental and model development programs related to l)CH and hydro-Using the results from l and 2 nbove, an updated estimate gen production during high-pressure melt ejection of the risk associated with high-pressure core melt acci- (HPME), will be reviewed and assessed to quantify the dUrts will be made, along with an estimate of the risk current level of uncertainties to determine if any of the reduction that would be obtained by intentional questic,ns raised in the following tasks have already been depressurization, taking into account any increase in risk answered. In addition, the assessment will determine the associated with depressurization. Based on a careful nature and character of information likely to result from weighing of the net benefits of depressurization, a these programs over the next 3 years. Progress made to recommended course of action will be proposed. date, curren3 research topics, major uncertainties, and methods being used, bath experimental! and analytical, Some of the key questions that the SARP will attempt to will be examined.This task will be carried out prior to the answer in the near term in order to make the above deter- planning and conduct of any new experiments or analyses.
minations are:
Usables and Use
- 1. What is the likelihood that the RCS will fail by natu-ral circulation prior to lower head failure? If so, are Based upon the results of the above assessment, a judg-these failures of much lesser concern to overall risk? ment will be made as to whether a significant increase in (Task 2.2) understanding the threats to containment from high RCS pressure accidents will be achieved in the next 3 years. In
- 2. Is there a low-pressure cutoff below which there is particular, the question of whethera significant reduction no DCH threat? (rask 2.3) in uncertainties will be achieved only with a significant ad-ditional expenditure of time and resources will be ad-
- 3. If so, will this pressure be reacheci t>y natural dressed.
circulation-induced failure of the RCS, or through operator action to depressurize, or both? (Tasks 2.1 Task 2.2 Assess the likelihood of RCS structural failure and 2.5) by natural circulation.
- 4. If operator action is necessary, is there time avail- Research Approach able for this action? Given that operator action is to be taken, what are the hardware and procedure Calculations of RCS failure by natural circulation will be specifications that would enable successful depres- scrutinized and alternative models and assumptions ex-surintion actions? (Tasks 2.3 and 2.4) amined. Comparisons of calculations with available ex-periments (Westinghouse) will be made to identify poten-
- 5. Are there adverse consequences to early depres- tial weaknesses in both experiments and modcls, particu-surization? If so, what are these, and what is their larly with regard to scaling. An assesstnent will be made as potential significtmce in terms of causing either ear- to why there were only minor natural circulation effects in tier core damage or earlier containment failure? the upper vessel structure nnd none in the hot !cg piping (Fasks 2.4 and 2.8) at TMI.
- 6. What is the nature of the DCH threat, and what High RCS pressure PWR sequences will be selected by mechanisms (e.g., sprays) and configurations exist initiating event, characteristics of the early stages of the 11 NUREG-1365
sequence, and potential for evolving to intermediate or Usables and Use low RCS pressure sequences (e.g., potential failure of a .
Direct comparisons of contam. ment pressurization ob-safety valve in a station blackout dming core heatup).This served in tes:s at different scales (ex,i sting facilities) pro-selection should include loop seal clearing and pump seal vide insight on the dominant factors affecting the scaling failures. Codes predicting energy redistribution due to UP of the results. The data may be used to estimate melt natnral circulation will be applied to these sequences and ,
entrainment as a function of quantity of melt and ej,ection will be verified, validated, and examined for underlying rate. ne melt entrainment will then be used to calculate assumptions that could affect the results. Analyses and at what pressure DCH from core ejection does not comparisons of the selected sequences as to likelihood, threaten the contamment integrity (low-pressure cutoff).
location,)nd time of failure of the RCS boundary will be made. He effects di fission , product deposition and Task 2.4 Explore the feasibility of intentional RCS revaporization as p.n additional source of thermal energy depressurization over the relevant spectra of during natural cir;ulation will be considered.
PWR severe acendent scenarms.
Research Approach Unbles and Un The overall framework for analyses should be based on a comprehensive treatment of representative scenarios, De result o tnis task will be an estimate of thelikelihood including possible perturbations by operator actions of RCS fniure from natural circulation m high RCS pres- (based on the current state of operator training, sure sr.:narios for typical primary system geometries. It is Emergency Procedures or Functional Restoration possiole that this item will alleviate conce"n with the Guidelines. and timing for when the operator is required threat of direct containment heatmg. to depressurize) and equipment failures that may have a bearing on the issues addressed here. The research should consider, but not be limited to, the normally j Task'3 Investigate the influence of cavity and con. available power-operated relief valves (PORVs) as a !
tainment compartment structures on DCil, depressurization device. One such strategy is to decrease i
Iow9ressure cutoff for DCil, and hydrc ten the system pressurc sufficiently to allow the accumulators l proJuetion from HPME. to dump prior to significant cladding oxidation, so that l gross fuel degradation is approached at a pressure that !
will not threaten the containment integrity should the !
Research Approach vessel lower head fail The approach is essentially analytic, using thermal-hydraulic codes such as RELAP5 The scoping calculations performed for NUREG-1150 to orTRAC for the most part, although selective use of core estimate the DCH load from the dispersal of melt into the degradation codes may also prave useful. The use of containment correlated DCH load with melt quantity, models in these codes that have not been tested in ejection pressure, and ejection rate. The results of this circumstances peculiar to this application, e.g. reflooding analysis identified the key unknowns in the prediction of of a highly overheated core would be scrutinized to DCH containment loads from melt dispersal that need to facilitate judgment as to the meaningfulness of the be determined by experiment-particularly, the effect of results. The research will examine both beneficial and lower containment compartment geometries; the effect detrimental effects of early intentionM depressurization.
of water (co-dispersal with the c0rium or added by spray) on reducing the magnitude of DCH pressure; and the Usabies and Use l>roduction of hydrogen by HPME. The results of this work will be a defined relationship between depressurization stra.egies and final system Experiments at different scales (using existing facilities pressure at the time of gross fuel degradation, or the time of lower head failure, and an identification of any insofar as possible) with high temperature, thermitically generated melt and steam as a pressurizing fluid will be detrimental effects of intentional early RCS
, l performed. Measured quantities will be identified based depressurization. :
on the scaling analysis of Task 1. Variables believed at this time to affect the magnitude of the challenge to the can- Task 2.5 Determine the mode of bottom head failure of ;
the reactor pressure vessel in a high RCS !
tainment are melt mass expelled from the cavi:y and its distribution within the containment, containment atmos- pressure sequence.
phere pressure and temperature, melt metal content of Research Approach the dispersed melt heat transfer rates to structures, and hydrogen production in the dispersing melt-steam mix- Assess and determine the likely modes of PWR pressure ture and its transport and combustion in the ccmtainment. vessel bottom head failure considering the reasonable l l
NUREG-1365 12
range (rask 2.6) of quantities, composition, and timing of Usables and Use core melt that is arriving on the bottom head. From the .
above assessment, idcotify those experiments that may The result of this work will be the most likely characteris-need to be conducted to determine the values of parame- tics of melt to be considered m addressmg PWR bottom ters that are keys to the assessments and/or confirm those head failure (rask 2.5).
assessments. Application of the scaling methodology (rask 1) will be made in devising and interpreting the re- Task 2.7 Use the results obtained in Task 2.3 to up-suhs of the experiments. grade DCH models.
Resenrch Approach
%c TMI b ver head inspection plays an important role in understanding why vessel failure did not occur for this This task should proceed only if the results from the scal-accident. It is not sufficient to look for damage to the ng program (rask 1) and Tusk 2.3 indicate that further lower head. It is essential that the debris on the low:r analytical developments are necessary for calculation of head be carefully characterized m order to mterpret the DCH loading and closure of this issue. Analysis may be observed condition of the lower head. used for extracting detailed information from ex criments that can then be used to develop rrodels that sufficiently represent plant-specific geometries. De suitability of the models will be affirmed by comparisons Usables and Use of calculations to the results of experiments. The developed DCH models can then be used to predict con-This work will result in an understanding of the sensitivity tainment pressure due to direct heating phenomena in re-of reactor pressure vessel (RPV) lower head failure mode actor accidents as well as the mitigation of these phenom-to the quantity, composition, and timing of arrival of moi- ena by containment pressure control devices and procc-ten corium on the lower head. This understanding may be durer.
quantified in the form of system level codes that predict the mode and timing of RPV faHurc in a manner Usables and Use consistent with the predictcJ melt progression and ther' mal-racchanical loading of the RPV. The results of This work will result in DCH models that will be applica-this task will contribute to the work of Task 2.3. ble to the evaluations of containment performance, the review of individual plant examination submittals, and the assessment of the efficacy of depressurization and con-Task 2.6 Determine the likely range of qua ntity, compo.
"* I"""'" " *"" #"" E*
- sit. ion, and t.mung of molten core maten.a l DCII containment loadmg.
arriving on the bottom head of a PWR during a core melt accident at high and intermediate Tasm Depressun. tat.ien cosWenent analyu.s.
RCS pressures.
Research Approach The results of Tasks 2.1 through 2.7, and the results of Research Approach Tasks 4.1 through 4.3 dealing with fuel-coolant interac-tions at low system pressures, will be considered in per-Review and assess existing experiments and associated forming analyses of intentional PWR system depres-analyses to ensure that the analysis tools represent the surization as a mechanism to avoid DCil. The analyses important phenomena. A series of core degradation and will be focused on assessing the efficacy of depres-melt relocation calculations will then be made using surization to bring about a net benefit to safety forimpor-boundary conditions and initial conditions that span the tant core melt scenarios (e.g., station blackout and high-range of core melt sequences. For each set of conditions pressure conditions associated with loss of feedwater, and considered, the calculations should yield the quantity, small LOCA scenarios). The benefit is the estimated re-composition (proportion of metals and oxides), tempera. duction in risk of early containment failure from high-tures, and timing (rate of arrival)of melt reaching the bot- prcssure ejection of molten core materials into the con-tom head. Based on these calculations, best estimates will tainment. Other factors to be considered include equip-be made on the characteristics of melt reaching the bot- ment costs needed for intentional PWR system depres-tom head of a PWR in high and medium RCS pressure surization and also any detrimental effects that may result core melt accidents for assessing DCH. Iow RCS from intentional depressurization. It is recognized that pressure conditions will be used in evaluating the the results of Task 2.2 (Natural Circulation) may indicate potential effects of depressurization including the effect that the RCS will fail and depressurize without operator of molten fuel. coolant interaction on core melt action. If such is the case, whether to continue this task progression. will be reviewed.
13 NUREG-1365
Usables and Use TMI-2 accident with melt relocating into the water-filled lower head. In either case, following dryott, debris The result of this task will provide additional bases upon heatup will be accompanied by lower head heatup and which to recommend for or against re.luiring depres- eventual failure of the bottom head. At this time, the surization of PWRs during core melt accidents to avoid composition and characteristics of the debris arriving on DCIL the drywell floor beneath the reactor vessel is uncertain.
The debris could range from mostly solid oxidic with low issue 3-BWR Mark I Containment Shell Meltthrough superheat to mostly molten metallic with high superheat.
The nature of the debris exiting the failed bottom head of An accident sequence leading to carly containment fail- the BWit can determine the subsequent course of the ac-ure has been postulated for BWR Mark I containments. cident. However, the bottom head itself may play a signifi-This sequence involves the deposition of molten core cant role in determining the characteristics of that debris.
onto the dryweil concrete fkor at the time of vessel In light of the above, the SARP will address the questions breach and the subsequent spread of the melt to, and its of hcw sensitive (or insensitive) to the quantity, contact with, the drywch steel wall that is the containment composition, temperature, and timing of melt arrival is boundary, ultimately causing failm c of tbe wall. Some cx- the failure mcdc of the bottom head of a BWR and to perimental work has suggested that the addition of water what extent are the hydrodynamic and thermal properties in the drywell could provide a mitigative effect by causmg of melt arriving on the drywell floor determined by the fragmentation and quenching of the melt before it mode of bottom head failure.
reaches the wall or by providing significant cooling of the melt layer: The heat transfer processes from the rr. cit to The BWR Mark I containment shell meltthrough issue, the wall, the melt-concrete interactions, and the heat re- like the DCII issue, was presented to a panel of experi-moral processes from the melt and from the wall are not cnced severe accident anatsts who based theirjudgment completely understood, although some work has been on the current body of evidence. The panel was equally done at Sandia National laboratories on heat transi'er divided as to whether failure would occur or would not.
from steel melts to steel structures and on melt-concrete When the drywell floor is water covered, the expe rts were mteractions. There have also been experiments on bub- more confident that failure would be avoided. Similarly, bling heat transfer arid on meit spreading at Brookhaven when the floor is dry and superheat is high, the experts National laboratory, using simulants, and experiments had leu confidence that the ccmtainment shcIl would re-related to melt spreading and quenchmg have been con- main intact.The panel concluded that the degree of belici ducted by Fat,ske and Associates, Inc. la addressing this regarding containment shcIl meltthrough is 0.33 or 0.87; issue in the near term, the literature will be carefully re- the lowest failure probability corres;ionds to cases in viewed to clearly identify the critical questions to be an- which water is assumed to cover the drywell floor; and the FwerCd with respect to Mark I shcIl failure and the capa- highest corresponds to cases in which the drywell floor is bility cf the existing data to aoswer those geestions before dry and debris flow rate, debris superheat, and debris un-any new research efforts arc begun. oxidized metal content are all high. Nevertheless, it was generally agreed that the composition and temperature of One such question that has been identified is the sensitiv- the debris exiting the reactor pressure vessel and the ity of the time and mode of reactor vessel lower head presente of viater on the drywell floor are importani vari-breach to the quantitj, composition, and timing of arrival ables in estimates of wntainment shcIl meltthrough of molten core on the bcttom head of a BWR. The core probability. This issue is, of course, of intercs! where melt of a BWR may proceed much differently from that of ventmg might be accomplished after a severe core dam-a PWR. For example, one possibility that has been sug- age accident, but it does not bear on the recognized effi-gested is that the channel box and control blada geometry cacy of venting prior to such ar accident m order to pre-of a BWR suggests that stable crusts supporting the melt vent the same.
may not form (as in TMI) and that mo; ten fuel will more or less continuously flow onto the lower head. Further, Specific questions that the SARP will attempt to answer the lovcer head volume of a BWR is larger than that of a in the near term are the following:
PWR and is densely occupied by the control rod drive structures, and this may result in a different failure p.x- 1. What is the relationship of the BWR bottom head ess (e.g, less coherent failure) of the bottom head from failure mode to variations in quantity, composition, that of a PWR. temperature, and timing of arrival of the melt or the bottom head? (Task 3.1)
It has been postulated that a BWR core melt progression proceeds by gradually accumulating f ully quenched debris 2. What is the effect of water on the urywell floor when in the spaccs between the control rod guides in the lower the melt pours out from the pressure vessel? (rask head. Another postulated progression is simila- to lhe 3.2)
NUREG-1365 14
- 3. How does the answer to the above question depend Task 3.2 Deteanine the effect of water on melt cpread.
or. the initial conditions (melt ejection rate, melt su- ing, melt cooling, and melt. concrete interac.
perheat, melt composition, initial presence or ab- tions.
sence of water) and water addition rate? (Task 3.2)
Research Appronh
- 4. Under what conditions would the crust that forms at 'lhe results of melt spreading, cooling, and core-concrete initial contact between the melt and the shell be sta- interaction experiments will be reviewed to determine ble and for how long? What is the expected rate of the extent to which a consistent picture of critical vari-heat transfer between the core melt materials and ables and conditions can be drawn. Among the questions to be considered are the role of superhcat, the effect of the shell for various melt coc litions? (Task 3.3) ,
uncxidized Zr m the melt, crust formation and stability, the generation of aerosols and noncondensible cases In addition to addressing the questions related to attack of from core-concrete interactions, and the effect on the be-the Mark I contaimuent shell, the results of the following havior of the melt of the presence of water-both water tasks will provide initial conditions for a variety of ex-ves. beneath the RPV upon bottom head failure and water sel phenomena that depend on the quantities and dynam- added atop the melt following the arrival onto a previ-ics of spreading of the released corium. ously dry floor. Consistent with the results from the scal-ing program (Task 1), calculations will be conducted either to confirm the conclusions of the review (or clarify Task 3.1 Invutigate debris relocation phenomena into issues that were concluded to be ambiguous) and define the lower plennm of a BM11, including the fail. those near-term experiments that can be conducted to re-ure mode of the core plate.
s am gmues Mata.
Usables and Use Research Approach This work will produce refined estimates of meh spread-ing, hydrogen and other noncondensible gas generation Review and assess existing BWR core relocation experi. from melt-concrete interactions, and the effect of water ments and associated analyses. Analyze core plate and tuneath the RPV. Of caual importnnce, however, this RPV failure mode and timing, assuming a range of gaan, task will bring to focus those questions that may only be tities, composition. and timing of melt. Based on these addressable by experiment over the long term since this analyses, identify the key unknowns that can only be de. task goes to some of the basic chemical and process ef-termined by further experiments, and assess the feasibil- fects of high-teroperature interactions of the core-ity of those experiments. If such experiments are deter, concrete materials.
mined to be both necessary and feasible, design and con-duct properly scaled out-of-pile tests in which simulated Task 3.3 Determine the conditions under which the in.
moltt.n core debris is poured onto a scaled core plate or teract ons between the spreadmg melt and the bottom head to verify the failure mode as a function of the Mark I drywell shell will lead to containment nature of the debris (quantity, composition, and timing). shell failure.
The results of fhis task, including those for any additional experiments carried out under this task, will be used to Research Approach develop a model to predict lower head failure. All the Experimental results of heat transfer from steel melts to analyses and experiments should assume expected de- steel structures will be examined to determine their appli-graded BWR cott ccmMinr.3 (e.g., water in the lower cability to reactor accident conditions. In particular, ques-plenum t.t the time of melt arrival on the core plate and tions such as the effect of the inclination of the shcIl to the ADS operation.) drywell floor and the effect of concret e deconcoosition gas agitation of the melt will be addressed. Experenents will be devised and conducted to determine in the ntar term Usables and Use whether a stable crust can form and persist at the drywell shell. Unless otherwise indicated from the resub.s of Task 3.2, this task will assume the presence of water in the
'Ihis work will address the expected mode, timing, and drywell. The staff recognizes that it may well be that this temperature of core plate and bottom head failure and task might not be completed in tbc near term; however, passage of molten core debris through core plate and bot- available insights and data will t e used in the closure tom head for use in severe accident analyses. The results process.
of this task will provide crucialinitial conditions for a vari-ety of ex vessel phenomena that depend on the quantities 'Ihe presence of Zr metalin the melt arriving at the shell and spreading characteristics of released corium. depends on the extent of its oxidation both in-vessel and 15 NUREG-1365
while the melt is spreading over the drywell floor. If it is ' 2. What is the potential that reflooding a severely dam-determined that significant amounts of Zr remain un- aged BWR core will result in a recriticality? What oxidized before contact with the shell, then the impact of will be the effects as opposed to not adding water?
the presence of Zr wdl be determined. (Task 43)
Tash 4.1 Determine the effect of water injection on the Usables and Use generation of steam and hydrogen during reflooding of a degraded core.
The results from this task will enable the staff to deter-mine the conditions under which the shell of a Mark I con- Research Approach tainment may fail, and, in particular, determine the proxi-mate melt amount, composition, and superheat necessary Identify basic variables governing heat transfer and hy-for failure. urodynamics of melt-water interaction, including the ef-feet of water injection 01. debris reconfiguration. The work would be oricuted toward scoping the range of hy-Issue 4-Adding Water to a Degraded Core drogen generation that can be produced dering quench-ing and would assess the impact on containment perform-There is little doubt that in a severe accident situation the ance. In consideration of the results of the scaling analysis primary efforts of the operators will be directed toward (Task 1), establish whatever additional experiments may making water available to the reactor vessel. An impor- be needed or conduct analysis needed to determine the tant question that needs to be considered in view of such effect of adding water to a degraded core. We have al-likely efforts is what are the likely consequences of those ready performed a lot of TMI-2 analyses and will assess efforts. Given the uncertainties in core melt phenomena, these results.
the uncertamty associated with the operator's knowledge of the condition and location of the core during an actual Additionally, a benchmarking exercise, using available sevet e accident, and the intuitive drive to put water onto models (to be selected), against the TM1-2 accident be-the core in the event of an accident, it is not likely that an ginning at 174 minutes (when core was reflooded by the operator ever would be told not to put water into the reac- start of a reactor coolant pump and high-pressure injec-tor vessel should water become available during the tion) will be undertaken as part of this task.
course of an accident. However, along with the potential benefit of achieving a stable, coolable configuration, re-Usables and Use stonng water to a core that has been severely damaged can have effects of which the operator should be awarc. This task will produce estimates on the amounts and rates The operator also should maintain cognizance of possible of steam and hydrogen generation as a function of water symptoms and response of the plant to adding water in addition to degraded core geometries.The effect of add-such circumstances (e.g., molten core-coolant ir.terac- ing water to degraded cores has an important application tion, increased hydrogen generation, increased contain- to accident management as well as to improving the quan-ment pressure). tification of the resultant containment loads.
A related question that also needs exploration is what are Task 4.2 Stability of melt supporting crusts.
the circumstances in which a grossly degraded core can be prevented from melting its way through the vessel lower Research Approach head. Although important parameters can be identified as water availability, system pressure, and debn,s configu- fo io f dl npW ration, it is not likely that complete resolution to this issu predicted failure of the crust, and the eventual rekication will be achieved m the near term. Nonetheless, there are of the melt into the bottom of the reactor vessel depend questions related to the addition of water to a degraded stron81Y on the heat transfer coefficients at boundaries of core that can be addressed in the near term that will pro- the core melt crucible. The anticipated approach is to re-vide some insights into the longer-term issues of accident view and assess existing analyses for both the growth and management as wc!! as improve understanding of low demise of a melt-supporting crust and subsequently per-RCS pressure core melt accident phenomena. The ques- form a set of calculations of heat transfer along bounda-tions to be addressed are the following: ries of the crust crucible, analyzmg thermal and mecham-ctd crust stability as a function of melt accretion.
- 1. What are the amounts and rates of hydrogen and Usables and Use steam generation during the reflooding of a de-graded core and during relocation of the melt onto The work will affirm whether the heat transfer coefficknt the bottom head? (Task 4.1) and the code treatments of core behavior, including NUREG-1365 16
crucible formation and collapse and debris relocation into because the use of the detailed mechanistic codes to the lower plenum, are suitable. evaluate entire acciderit sequences is costly, impractical, and unnecessary.
Task 43 Investigate the possibility and consequences .
Any large code development program is always faced with of recriticality in degraded BWR cores. the difficulty of determmmg when the codes are " good enough." That is, when have they achieved an accuracy Research Approach sufficient for the application for which they were m.
There are two configurations to be considered with regard tended? At this time, an acceptable level of accuracy has to the recriticality problem. One is the possibility that not been defined. As a result, the current code develop-standing fuel pellet stacks will become critical upon ment program is continually iterative. Codes are devel-refk>oding (with control rod materials previously melted oped and assessed against selected experiments and out).'Ihe other is the possibility that core material will be. needed model improvements identified. Overlaid on this come critical upon relocation to the bottom head. process is the identification of new phenomena consid-cred important to typical plants that "must be"incorpo-The approach is to first review and assess existing analyses rated into the code.
of both the likelihood and consequences of recriticality of .
In addition to the above, a problem that is not unique to a damaged IlWR core. Where deficient, additional calcu- the severe accident program relates to the number os, lations will be performed assuming the existence of criti-codes that must be developed to have a complete analysis cal masses for expected degraded BWR core geometries capabihty. If one phase of the accident is difficult to reflooded with unborated water. model, or has mherently large uncertaiatics or variabilities, or perhaps is best understood as stochastic Usables and Use processes with a distribution of potential outcomes,it is not appropriate to model other phases of the sequence in The products of this task would be an estimate of the like-greater detail, since both the overa!! variability and uncer-lihood and effects of recriticality on accident progression tainty will be driven by the least precise and least certain and the alternative design mitigative measures (e.g., mini-P.rocesses, respectively. Currently, this appears to be thc mum baron concentrations in the reflood water) that s tuation with regard to late-phase core melt progression.
might climinate the prob!cm if one is found to exist, The degree to which the early phase is modeled, and those models tested and validated by experiment, must bc Issue 5-Use and Status of Severe Accident Models tempered by the degree to which the late phhse through (Codes) melt relocation and vessel failure can be understood and As noted earlier, the confidence in the analyses and judg. modeled and that understanding tested by experiment.
ments that will be made in implementing the staff's Inte-Therefore, as part of the code development program, a gration Plan will depend on how well the tools used in method must be developed that constantly assesses the performing those analyses and making those judgments state of code development from a number of standpoints conform to the realities of a severe accident. The tools that have been developed under the SARP, the physical in addition to how well a particular sequence or facet of an accident can be modeled.The fo!!owing questions will be i
and phenomenological models embodied in computer considered:
programs (codes), have had and continue to have a key role in furthering the objectives of severe accident re- How well do the mecham. .stic models reflect the phe-1.
search, viz., support of the U.S. Nuclear Regelatory Com, nomena beheved to be important to severe acci-mission's policy toward severe accidents. First, and per- dents? (Are the correct phenomena being mod-haps most critical, is the use of the codes in identifying the important sources of risk. In this capacity the impact of dedh the codes is important since the results of code calcula- 2. Ilow wc3 does the interactive program of code ad-tions support the evaluation of risk. The evahation of vancement and experimentation achieve the objec-risk, m turn, significantly determines both what is re-tive implied in 1 above?
scarched and the relative priority of what is researched.
Second, through an iterative process, exercising the codes 3. Is the level of detail in the codes appropriate to their helps define the experiments needed to improve the rep- use? (Are some codes more detailed than needed, presentation of physical phenomena by the codes. Third, others not detailed enough?)
the complex mechanistic codes provide a benchmark for less detailed but faster-running codes that are used in ap- 4. Which stages of an accident need to be modeled by pbcations that require extensive calculations, such a3 detailed mechanistic codes, coupled to adjacent parametric studies for evaluating accident management stages? (While clearly there is an interdependency strategies.This "two-tier" code strategy has been pursued among qes of a severe core melt accident, a 17 NUREG-1365
l detailed mechanistic code coupling all stages of acci- 6. What features (physics and chemistry) not presently dent phenomena from the onset of core uncovery in the code /model need to be incorporated to enable through containment failure is not needed to either the code to be a reasonable representation of acci-understand or make regulatory judgments about dent phenomena, severe accidents. Further, a detailed mechanistic model of all aspects of a severe accident may not be 7. To what quality assurance program was the code possible, let alone necessary.) subjected.
- 5. Is the level of precision needed for regulatory use Usables and Use being considered in the code development program?
The above information will be reviewed and scrutinized by the staff, and an assessment will be made of each code as to its present scaleup capability and usefulness in sup-
- 6. Is the level of precision from a given code more than p rtingachievementof thenear-termgoalsof therevised needed, and is it suitable to the expected ovemil A ased on dat assessment, recommenyons as to levels of uncertainty from an integrated analysis use of the present form of the code, further des clopment, package for applications as discussed above (e.g.,
or abandonment will be made.
accident management)?
. Criteria forjudging the suitability of each code are being Answering the above questions is a major undertaking developed. nese criteria will take into consideration the and is not likely to be achieved in the near term. intended end use of the code and the expected user of the Nonetheless, some measure of confidence in the tools code. Deciding when code development is completed that will be used in achieving the near-term goals must be would require a subjective judgment and a sense of pro-made. Herefore, the following task will be done. portion between the level of precision needed for applica-Although not a complete code assessment, this task is tion and the levels of uncertaintyand variabilityin models believed to be a workable compromise in the near term of severe accident phenomena.
and should provide sufficient information tc> enable the staff to gauge the appropriate level of confidence to be This latter task is applicable to both tl.e near- and long-placed in calculations using the codes. term research as is Task I on similitude. The conduct of Tasks 1 and 5 is beFeved logical and necessary to the criti-Task 5 Code documentation and review. cal reexamination sf SARP needs. However, neither is intended to constrain closure of the near-term issues de-Research Approach scribed above.
For each code developed under the S ARP, the developer will be asked to supply the followtng mformation:
- 4. Meeting Long-Term Goals The research described in this section of the SARP plan
- 1. Stages in a severe accident to which the code is in- consists of work primarily directed at reducing uncertain-tended to apply and the accident sequences in which ties in the estimation of the risk presented by severe acci-those stages are found. dents. This orientation is consistent with the direction provided in " Review of Research on Uncertainties in Es-
- 2. Current capability of code. (Of the intended acci- timates of Source Terms from Severe Accidents in dent stages and phenomena, which does the code Nuclear Power Plants," NUREG/CR--4883, known as the now model, which does it not.) "Kouts Report." That direction suggested a broad base approach, investigating manifold phenomena with the
- 3. Limitations of the code. (Portions of accident stages common goal of reducing uncertainties. Both the and/or special assumptions under which the code approach and the goal remain valid, as does the more operates.) specific direction in NUREG/CR-4883 that identifies individual technical questions as contribu; ors to
- 4. Statement of why the code needs to be or does not uncertainty in tisk estimates. However, there are sources need to be coupled sequentially to adjacent codes to of new information that bear on hcw to order research yield sufficiently precise calculations. and dis'ribute resources among the topics of the SARP.
These sources are the revised NUREG-1150, " Severe
- 5. The degree to which the relevance of the physics, as Accident Risks: An Assessment for Five U.S. Nuclear embodied in the code, to severe accidents has been Power Plants," and the guidance provided in tested by experiment. (This point questions not sim- SECY-88-147 that identifies accident management as a ply the conformance of code to experiment, but ex- strategy that the NRC will pursue in dealing with severe periment to severe accident.) accidents. In addition, the successes of the SARP itself NUREG-1365 18
over the years are a source of information providing a techniques. Priorities for additional development can be clearer picture of which phenomena are critical to under- determined through use of probabilistic risk assessment standing severe accidents, insight into what avenues of in- techniques. When risk is insensitive to the uncertainty in vestigation are likely to be most fruitful, and which topics the phenomena or variabilities in parameter values, there are likely never to be fully resolved in a reasonable time should be no need to attempt to further reduce the frame within realistic resources. Taking the above uncertainty or variability.
together with the needed balance discussed in Section 2, specific topics have been identified that characterize the Using the " documentation" provided by Task 5 of Sec-SARP cfforts directed toward achieving the long-term tion 3 of this plan as a point of departure, the approach of goals cited in Section 2-providing an understanding of developing mechanistic, deterministic codes for each the range of phenomena exhibited by severe accidents,in- stage of severe acim .. ll be reviewed. For each code cluding impacts of generic accident management development program / phenomenon bemg addressed, a schemes, and developing improved methods for assessing review of progress to date, magnitude of uncertainties the " severe accident source term." These topics are the remaining, and degree of validatien will be made. Based following: upon this information, a determination will be made as to whether an alternative approach might be appropriate for
- 1. Review of the SARP approach to modeling severe dealing with some phenomena. The efficacy of the accident phenomena SARP's "two tier" code strategy (fast-running lumped parameter system codes benchmarked against a
- 2. In-vessel core melt progression and hydrogen gen. combination of individual standalone and coupled eration mechanistic, deterministic codes) also will be examined.
This review will start at the beginning of FY 1990, and
- 3. Hydrogen transport and combustion recommendations will be made at the end of FY 1990.
The NRC staff assessment of the modeling program will
- 4. Fuel-coolant ir.teraction be assisted by various individuals outside the NRC and will be drawn from universities, nationallaboratories, and
- 5. Molten core-concrete interactions mdustry personnel who are expert in accident phenome-nology (have appreciation for the purpose of the analysis
- 6. Fission product behavior and transport and the uncertainties m all the parts of the analysis), regu-
- 7. Fundamental data needs 1 tory concerns, and programming and numerical tech-mques. Ihis effort would identify the strength or weakness of different modeling approaches, quantify the It is not at all clear that detailed mechanistic models can be, or need be, developed for items 2 through 6. Some uncertainties m code models, and identify additional research, if any, where uncertainties and risk importance phenomena may only be trt.ctable when treated statisti-cally as a population of related events with differing out- are significant.
comes. In recognition of this possibility, the first item Issue L2-In. Vessel Core Melt Progression and Ilydro.
listed is a sy stematic review of the approach taken m the gen Generation.
SARP to modehng severe accident phenomena.
l In-vessel core melt progression concerns the state of the Issue L1-Modeling Severe Accidents. reactor core from the start of core uncovery to reactor . ,
vessel failure. Included phenomena are thermal attack by Because of the difficulty in performing prototypic experi- the core debris upon the reactor structure and the reactor ments and the variety of scenarios possible, substantial re- vessel, in-vessel hydrogen generation, in-vessel natural liance must be placed on the development and validation convection and heat transfer, and in-vessel steam genera-of complex computer codes for analyzing severe acci- tion. The details of core melt also determine rates and dents. A number of " mechanistic" codes (e.g., SCDAP/ amounts of in-vessel fission product release and acrosol n -
RELAP, MELPROGm1AC) have been developed for generation and much of the fission product and acrosol ,
various stages in severe accidents, both in-vessel and transport (and retention) in the reactor coolant system.
ex-vessel, for both PWRs and BWRs. However, it appears (Some aspects of this issue relating to fuel-coolant inter-that there are practical limits to the feasibility of action were discussed in Section 3.) ,'
deterministically modeling all aspects of severe accident behavior. Moreover, it is not clear that all aspects of In considering melt progression, BWRs and PWRs need severe accidents need to be mechanistically reflected in to be treated separately, mainly because of their different deterministic codes. The resolution of some severe fuel assembly, control element, and lower plenum struc-accident regulatory issues may be achieved with bounding tures. For example, the effect of water injection on (1) the analyses alone. Further, some phenomena may be better integrity of upper in-vessel structures, (2) the mcde of addressed employing stochastic rather than deterministic core relocation, (3) hydrogen generation upon core 19 NUREG-1365 .-
relocation, and (4) the mode of bottom head failure and stoichiometric hydrogen-air mixtures and on all the details of the release of core and structural debris into hydrogen-air-steam mixtures.1:is believed that the steam the drywell following failure are questions common to inerting effect is reduced greatly at elevated both PWRs and BWRs. However, the answers are quite temperatures. There is a limited supporting data base for distinct arising from the physical and geometric differ- this phenomenon. Experi:nents are necessary to resolve ences between BWRs and PWRs. Bounding calculations the uncertainties associated with high temperatures and and a few small-scale experiments may be continued to steam concentrations typical of those likely to be identify which phenomena are critical for each reactor encountered in severe accident scenarios. It is also <
type and whether further reduction in uncertainties and necessary to extend or develop more mechanistic models I calculational variabilities related to melt progression and to predict, by either extrapolation or interpolation, the hydrogen generation are needed and achievable. temperature sensitivity of hydrogen-air-steam mixtures ,
thr.1 have not been tested.
Areas in which work will continue are the examination of TMI-2 data for whatever insights might be provided, the The small-break LOCA and station blackout scenarios NRU fuel melt experiments, and participation in the ex, are two cxampics of high-temperature hydrogen-air-isting cooperative program with the Federal Republic of steam mixtures that may exist below the autoignition Germany at the CORA facility. New experiments that temperature (550*C mmimum value for stoichiometric, evolve in the near-term research may continue as well. hydrogen-air mixtures). There are two aspects of the -
high-temperature /high steam concentration combustion i Issue L3-Hydrogen Transport and Combustion. problem. The first is the injection of high-temperature hydrogen and steam mixtures that autoignite upon The major concerns regarding hydrogen in LWRs are contact with preexisting and premixed hydrogen-air-that the static or dynamic pressure loads from hydrogen steam mixtures. The second aspect is the injection of combustion and detonation may breach containment hydrogen and steam mixtures at elevated temperatures integrity or that safety-related equipment may be that do not autoignite upon contact with preexisting damaged as a result of either pressure loads or high hydrogen-air-steam mixtures. This allows the possibility temperatures. To assess the possible threat to of a premixed condition to form and a subsequent containment and safety-related equipment, it is necessary deflagration or detonation. In both cases, the competition to understand how hydrogen is transported and mixed between chemicai reaction rates and physical mixing rat es within containment and to determine the likelihood of will ultimately determine the ensuing combustion mode.
various modes of combustion. The hydrogen behavior Data are needed to determine the hydrogen-air-steam issues have been extensively investigated since 1979, but flammability limits and the steam inerting criterion at cle-several important areas of uncertainty remain: vated temperatures if reliable predictions are to be made as to the likelihood and the potential threat resulting
- 1. High-temperature /higli steam concentration com. from various combustion mode (s) for a wide range of bustion, and accident conditions.
- 2. Deflagration-to-detonation transition. Therefore, the research approach is to determine hydrogen. air-steam flammability hmits, volumetric oxida-Research programs on these issues of combustion have tion (chemical reaction) rates, physical mixing rates, and been sponsored in the U.S. by the NRC and the nuclear the competition of these rates during the injection of industry and in the international community. However, high-temperature hydrogen and steam mixtures into the combustion processes are sufficiently complex that c ler preexisting and premixed hydrogen-air-steam many aspects are still not well understood. The resulting mixtures. A draft proposal is currently ,under !
uncertainties in the threat to containment integrity are consideration by the NRC that addresses feasibility; unlikely to be reduced significantly by the existing construction design, cost, and schedule; and an research programs. The intent of this research is to experimental test plan. It is our current estimate that if a reduce these uncertainties; however, in some cases facility is needcd, it can be constructed by early FY 1990 t
reduced uncertainties are not required to make a and that expenmental results can be generated by late FY near-term regulatory decision. Each category of uncer- 1991 r early FY 1992. Final data reduction and tormal tainty is discussed below. documentation should be available m late FY 1992.
- 2. Deflagration to. Detonation Transition (DDT) !
- 1. High. Temperature Combustion 1 Direct initiation of a detonation would require a concen-The Zeldorich-von Neumann-Doering (ZND) chemical trated high-energy source for insensitive, steam-diluted kinetics theoretical model developed under NRC spon- mixtures. This is not considered a credible mechanism of sorship predicts that increasing temperature has a strong initiation by almost all researchers. However, it is possible effect on the combustion and the detonation of off- to initiate a flame with a low-energy source such as a spark NUREG-1365 20 l 1
s
- 5. ,
^
n.
or a glowplug, and the subsequent propagation through (1) reflood of a damaged core or debris bed in-vessel, and orifices and around obstacles such as pipes can result in (2) fuel-coolant interaction in a suppression pool will be e flame acceleration that culminates in a transition to examined.
detonation.
In addition to providing final confirmation of the staff's -
The possibility of DDTin realistic and prototypic contain- position on the issne of the alpha mode of failure, this ment geometries and conditions needs to be resolved. work also will provide analytical tools, additional data, ne uncertainty in this area has increased because of and insights for use in evaluating the dynamics of molten recent experimental and theoretical results that indicate fuel-coolant interactions in-vessel and in various contain-an increased likelihood of detonations at high ment configurations. .
temperatures and large steam fractions as discussed earlier. The current data base on flame acceleration and issue LS-Molten Core."oncrete Interaction (MCCI).
DDT suggests that the mixture composition, obstacles, This is a subject fundan;:ntal to severe accidents. It has and venting are all important factors. Dese uncertainties received considerable research attention, both experi-result from a lack of experimental data on flame , mentally and analytically, mainly addressed at under-acceleration and DDT for conditions that mclude the standing the quasi-steady-state mteraction that would effects of steam dilution, elevated temperature, develop when the core melt materiak forte a pool above large-scale and prototypical obstacle types, and spacing. the concrete. However, some signiFc nt uncertainties remain. An important question is the long-term At present, no reliable model exists to predict or extrapo- coolability of initially molten corium pcols interacting late DDT results from small-scale expenments t w th concrete and flooded with water. Other uncertain-containment scale. Reactor safety studies and fundamen- ties involve the transient (early) stage of core-concrete-tal combustion research are being carried out in the Fed' water interactions, which include the spreading and relo-eral Republic of Germany and Canada.ncse data along cation of the melt over concrete and the associated with data generated for space shuttle application will1 e thermal-hydraulic characteristics of the attack. Also un-applied to reductng the uncertainty associated with DDT certain is the effect of pour rate on melt spreading.
as well as to assessmg the potential threat of DDT to con-tainment integrity. Speci'ically, this data base will then Specific topics that are addressed by the research include serve as the basis to improve or develop correlations and the key characteristics of molten-corium pools interacting models to allow extrapolation of experimental results to with the concrete basemat, the rate of melt cooling in the reactor scale and accident conditions. carly stages of MCCI, the energy balance of a corium-concrete interaction in the high-temperature regime, the 4 Issue L4-Fuel. Coolant lateractions. rate of fuel cooldown as it spreads over the concrete, the cHect of water on tae spreading behavior and associated Molten fuel contacting water can give rise to a range of heat losses of the con,um melts, the crosion and ablation phenomena. Very energetic steam explosions could re- of concrete structures (in particular, the reactor vessel sult in early containment failure (alpha-mode failure).
Less energetic interactions do not threaten containment pedestal), and the volatilization of fission produc:s and pr duction of acrosols during MCCI and estimation of directly but could change the course of the accident and the effects of these phenomena on the source term. Sev-the magnitude of the source term. Among these are sud-den coherent failure of the vessellower head, ex-vessel cral f these phenomena investigations are being carried ut under the existing cooperative agreement with the debris dispersal, and steam generation pressure pulses.
El RI ACE program and FRG HETA program.The de-(Core debris dispersal and steam and hydrogen genera-sign f any additional experimental and analytical pro-tion in some BWR Mark II and Mark III containments grams will follow from the work discussed in Section 3, also warrant further evaluation as to whether fuel-coolant
.Iasks 3.2 and 3.3.
mechanisms can pose a threat to containment integrity.)
ue s m uct Miador and hanspd The research to be conducted will be selective, confir-matory in nature, and focused to address some well- The NRC has had a substantial program to investigate fis-defined questions dealing with key parts of steam explo- sion product release from fuelin-vessel.Today PRA stud-sion phenomenology, such as degree of premixing, trig- ies no longer ider.tify this as an area of major uncertainty gering, and fragmentation within the detonation wave of a in risk assessment.The recent NUREG-1150 clicitations .
steam explosion for various premixture conditions. It is on a number of source term issues do indicate, however, clear that the program will necessitate limited scale ex- that late iodine rclease, revolatization,and fission product periments to test the predictive capability of calculations release from core-concrete interactions can have an of premixing and the fragmentation rate of melt drops in effect on the overall risk uncertainty. IIence, research in the explosion zone of a propagating explosion. Further, these areas is being continued. Nonetheless, because the utility of the calcu'ational tools to assess the effect of severe accident issues associated with phenomena that 21 NUREG-1365 ~-
have the potertial to result in containment failure have nection with core melt progression, but similar necds exist higher priority than modeling of fission product release with respect to MCCI. Among these needs are thermal and transport, source term issues will be addressed in the properties (including melting points, latent heats, and long-term program with lower priority, thermal conductivities) and phase relationships among the various constituents of the debris. A continuing pro-Issue III-Fundamental Data Needs.
gram to identify and measure (or calculate) the required For rome ranges of phenomena, fundamental data (physi. basic data and to incorporate the results in the codes will -
cal and chemical properties and constants) do not exist or be instituted. H0 wever, in supporting this basic data ef-are so poorly known that their use provides no confidence fort, we will attempt to strike a balance between the aceu-in the fidefity of experimental or analytic results. Data racy of the data and its significance relative to r y.:mo.y need3 have been briefly touched upon in Issue L2 in con- applications.
l NUREG-1365 22
l 1
APPENDIX A Background on Severe Accident Research Program and Relationship to Other Elements of SECY-88-147 For the past 10 years or so, since the Three-Mile Island The ex-vessel phenomena of melt ejection and direct con-accident, NRC has sponsored a research prcgram on se- tainment heating in high-pressure accident scenarior and vere nuclear power plant accidents as part of a multifac- molten core-concrete interactions in lower-pressure scc-eted approach to safety. Other elements of this approach narios, together with their associated threais to contain-included improved plant operations, human factor con- ment integrity, also are becoming be!!ct understood as a siderations, and probabilistic risk assessments. :n August result of the ongoing experimental and analytical efforts 1M5, the Commission issued a Severe Accident Policy under the S ARP. Understanding containment chr.llenges Statement (50 FR 32138), which concluded that existing due to steam and noncondensible gases is much improved plants posed no undue risk to public health and safety. as well. The "two-tier" code strategy of developing de-However, the Commission recognized that synematic ex- tailed " mechanistic" codes for understanding and faster-aminations of existing plants could identify plant-specific cunning " integral" codes for application has furthered the vulnerabilities to severe accidents for which further safety NRC 's ability to calculate the impact of accident se-improvements could be justified. quences and their associated risks to public health and safety. Both levels of codes have been used in f he risk as-In May 1988, the staff presented to the Commission an sessment document, NUREG-1150.
Integration Plan for Closure of Severe Accident Issues As can be inferred from the above, the thrust of the S ARP (SECY-88-147).The Integration Plan consists of six ma-jor elements: up to now has been to establish and refine the technical and scientific base of knowledge in the area of severe acci-Examination of existing plants for severe accident dent phenomenology, t o apply it at the scale of reactor ac-1.
cidents, and to reduce the uncertainties in this knowledge vulnerabilities (individual plant examinations).
base and the risk assessments that depend on it. The pro-Development of generic containment performance gram to date generally has been appropriate to the above 2.
motivations. However, because the needs for the SARP improvements with respect to severe accidents to be are changing as a result of actions being taken by the NRC implemented if necessary for each of the six contain.
to bring severe accident regulatory issues to closure, it is ment types.
natural to question whether the directions of the program Upgrading of staff and industry programs to improve should change accordingly. Obviously there will remain 3.
plant operations. practical and resource-related questions as to how many fronts can be pursued in understanding and characterize-
- 4. A severe accident research program. ing accident phenomenology and which of these fronts will prove the most fruitful ones to follow to achieve clo-
- 5. A program to define how and to what extent vul. sure of regulatory issues.
nerabilities to severe accidents from external events need to be included in the severe accident policy im- A.1 Development of Revised SARP plenientation.
To assist the NRC RES staff in this reappraisal and revi-
- 6. A program to ensure that licensees develop and im- sion of the SARP, a set of four expert groups was estab-plement severe accident management programs at lished to help the RES staffidentify and define the status their plants. of present SARP activities that relate to the Integration Plan, identify and focus research necessary for sound During the past few years, the SARP has generated a regulatory decisions to be made within the framework of large amount of insight into the progression of severe ac- the Integration Plan, and identify and rank the base of cidents. An extensive experimental program has led to a confirmatory research activities directed at achieving the vastly improved understanding of in-vessel core melt pro- long-terrn goals. Each group was composed of contractor gression and associated phenomena of hydrogen genera- personnel from the Department of Energy (DOE)labo-tion and fission product releasc. Models of core melt pro- ratories, consultants from universities and industry, and gression are being coupled to thermal. hydraulic in-vessel NRC staff. Each member was active in severe accident re-and primary cooling system models to allow more rigor- search. The groups consisted of three " working groups" oustreatmentof thein-vesselstagesof asevereaccident. that addressed the technical points of the research and an 23 NUREG-1365
" integration group" that dealt with programmatic consid- 2. Relationship of SARP to containment performa7cc erations. improvement program The working groups considered detailed technical issues, The Integration Plan states that the containment per-
& fining their status and needs for further research within formance improvement (CPI) program com91ements and the framework of the Integration Plan and confirmatory is closely integrat r.d with the IPE program and is intended reaearch. The integration group took the reports of the to focus on resolving hardware and procedural issues re-working groups and evaluated and synthesized them, lated to generic containment challenges.
along with other considerations, into an earlier draft of a recommended Severe Accideat Research Program plan. A detailed summary of the relationship of SARP and CPI ne NRC staff then used this as a basis for preparing the is given in Section 5.2 of the enclosure to the Integration revised SARP plan discuss:d in this report. Plan (SECY-SS-147).
- 3. Relationship of SARP to improved plant operations A.2 Relationship of Revised SARP To Other Elements ofIntegration Plan ne improved p' ant operations (IPO) program includes elements relating to continued improvements of the tech-In order to place the revised SARP plan in perspective, it nical specifications, emergency operating procedures is useful to first discuss in general terms the relationship (EOPs), expanding EOPs to include guioance for severe of SARP to the other elements of the Integration Plan. accident management strategies, and industry programs From the discussion it will be apparent that SARP pro _ to reduce tcansients and other cht.llenger to the engi-vides, or will provide, important and often essential data nected safety features (ESFs). These elements are obvi-to most of the otlier elements of the Integration Plan. ously closely related to accident management strategies, discussed below.
- 1. elat nshin of SARP to individual plant examina-The relationship of the SARP to these elemen:s of the IPO program involves providing analytical descriptions of The Integration Plan provides specific objectives for the severe accident sequences, including the effects of EOPs IPEs that each utility is expected to mect: and other mitigation strategies. These analyses would predict the effects of accident mitigation schemes, melud- l
- a. Reduction of it overall probability of core ing both a determination as to whether they will be effec- 1 damage and fission product releases by appro- tive and identification of possible un-les,rable t conse- ;
priate hardware and procedure modifications; quences This is particularly true with respect to the later phases ofin. vessel (core degradation) progression of a se- ,
- b. Development of an overall appreciation of se- vere accident and the ex. vessel and containment phe- !
vere accident behavior; and n mena. Accident sequence analyses, based on SARP re-su'.s, could be used to judge the appropriateness ard ef-
- c. Development of an understanding of the most fcctiveness of EOPs under severe accidem conditions.
likely severe accida.nt sequences that could oc-
- 4. Relationship of SARP to external events cur at its plant.
According to the Integration Plan, the evaluation of ex-It is envisioned that a principal tool for an NRC audit re- ternal events will proceed separately with a different view of IPE submittals will be a relatively fast-running,in- schedule from that of internal events. At the rooment, tegrated severe accident analysis code such as MELCOR. there is no part of the SARP that explicitly addresses ex-(Tbis code in tura is benchmarked and validated against ternal events. However, to the exter t that external events detailed and more mechanistic codes.) All these codes such as earthquakes or fires act as accident initiators or draw or have drawn heavily for their phenomenological constrain the availability of ESFs, safety-related cejuip-modeling on the SARP experimental and analyticel ef- ment, and other mitigation strategies. the accident p.0-forts (including the foreign and industry contributions un- gression can be analyzed with the severe accident codes der cooperative arrangements). Up to now, these codes developed under the SARP plan.
have been under contmual development and validation.
However,it is presumed that, when the staff begins its re- 5. Relationship of SARP to accident management pro-view of the IPEs, some versions of these codes will be gram i "frozca" for ese by the NRC staff.
The Integration Plan defined accident management :o in.
It should be noted that the new SARP addresses all the clude measures taken to prevent core damage; to termi-phenomenological uncertainties discussed in Appendix 1 nate the progress of core damage if it begins; and, failing to the IPE Generic Letter No. 88-20. that, to maintain containment integrity as long as possible NUREG-1365 24
.___-_______-_a
to minimize offsite releases.The NRC accident manage- detrimental).The staff is currently in the process of defin-ing the analytical tools that are available or should be de- j ment program concentrates on near term improvements 4 based on well-understood accident management proce- velopedand requiredforaccid. umagement purposes.
dures and strategies.The SARP will provide the data base We expect to complete this efh >y the end of FY 1990. l to allow the staff to examine procedures and strategies 11 is important that SARP should continue to address cer-whose benefits and adverse effects are not well under- tain important issues-in particular, the consequences of !
stood, adding water, both in-vessel (relsood) and ex-vessel, in at- l tempting to cool a severely damaged core or quench core ;
As in other areas, the role of SARP in accident manage. j debris-and that funher attention should be given to ment is to provide the tools (i.e., codes) to analyze the l progression of severe accidents and to evaluate the ef- evaluation of the uncertainties associated with the use of fccts of given mitigation strategies (both beneficial and the codes.
I l
25 NUREG-1365
n APPENDIX B A Severe Accident Scaling Methodology (SASM)-
B.1' Objective and Outline For reasons discussed below, both premises must be ad- -
dressed and evaluated before a high degree of confidence ne objective of this appendix is to outline a Severe Acci- in the analysis will exist.-
dent Scaling Methodology (SASM) that could be used to address in a systematic and practical manner questions B.2.1 Premise Related to Experiments concerning such scaling topics of interest to severe acci-In nuclear reactor safety research, experimental data are
. dents as:
used (1) to develop correlations and/or models for a par-ticular process or (2) to assess code capability to calculate
- 1. The adequacy of the design and operation of a such a process.
reduced-scale test facility to provide experimental
' data that can be used in safety analyses of full-scale Full-scale test facilitics capable of generating the experi-nuc! car power plants. mental data of interest are prohibitively expensive to con-struct and ;o operate. In severe accident research, diffi-
- 2. He appropriate initial and boundary conditions for culties and costs are compounded because some phases of experiments of interest. an accident scenario involve failures of the reactor vessel and of containment structures. Thus, large (or even
- 3. . The use of a set of experimental data or of a correla- smaller) scale integral facilities needed to provide experi-mental data on processes leading to or attending such fail-tion based on the data m nuclear power plant safety analyses. ure events become prohibitively expensive.
Consequently, in severe accident research, the majority
- 4. . The effects of test facility scale distortions (if pres- of experiments will have to be performed in reduced or ent) on physical and chemical processes of interest small-scale separate effects test facilities. The premise to nuclear power plant safety analyses. made in following tiiis approach is that experimental data from such facilities are applicable and relevant to nuclear S. The capability of a computer code to scale up physi. power plant conditions. nis implies that test facilities, as cat and chemical processes observed in reduced. well as the initial and boundary conditions of experi-scale test facilities to full-scale nuclear power plant ments, are properly scaled so that distortions (if and when conditions, etc. present) will not affect the evolution of physical and chemical processes of interest.
He role that scaling pl'ys in nuclear power plant safety Whether a facility and the initial and boundary conditions analyses is discussed in Section B.2 of this appendix. ne of an experiment are well scaled will have to be evaluated need for establishing a severe accident scaling methodol-for each facility and set of experiments because such an ogy (SASM) is discussed in Section B.3, together with the requirements that it would have to meet. The elements of evaluation wi!! determine whether the data can be used in nuclear power plant safety analyses of a postulated severe such a methodology are outlined in Section B.4, together accident.
with their rationale. He last section describes a program directed at demonstrating the methodology by applying it B.2.2 Premise Related to Codes to the DCH problem.
He reliance on computer codes to simulate the behavior B.2 Elements of Safety Analyses and I".nucIcar power plant during a postulated accident sce-Prem,tses nano ns predicated on three factors. First, it is too costly, <
and for severe accidents not even feasible, to subject a nu- ,
clear power plant to such an event. Second, very often one As was the use in studies of LOCAs, analyses concerned cannot directly apply results from test facilities to nuclear with severe accidents will have to be supported by a suit- power plants; this is particularly true for separate effects able experimental data base on appropriate models and test data.nird, the study of various plant recovery tech-computer code calculations, and, when appropriate, on niques can only be performed by computer codes.
baunding calculations. Implicit to this approach are two premises: one pertaining to experiments and the other to implicit to applications of computer codes to analyses of codes; both are concerned with scaling. postulated accident events in nuclear plants is the NUREG-1365 26
l l
prernise that these co&s have the capability to scale up It can be concluded from this brief discussion that if phenomena and processes from test facilities to full-scale meaningful safety analyses concerned with postulated se-plant conditions. For reascms discussed below, this prem- vere accidents in nuclear power plants are to be per- ,
ise must be evaluated on a case-by-case basis, that is, for formed, then questions icrated to scaleup capabilities of !
l cach postulated accident scenario or for a set of scenarios. experimental data and of computer codes will have to be
! addressed.
It is often stated that advanced codes are " mechanistic" B.3 Needs and Requirements j and are based on "first principles." Herefore, ipso facto, they have the scaleup capability. However, as a matter of The important role that scaling has in experimental and fact, for the followmg three reasons this is not the case. analytical investigations related to severe accidents was noted in the preceding section. This role establishes a First, because the conservation equations used in com- need for a scaling methodology that would be applicable puter codes are space averaged and bemuse of their de- to both experiments and computer codes.
pendency on numerous empirical correlations, computer codes are not based on *first principles." As the capability In order to meet the needs of research and development of a ,.we to model a particular process and/or phenome- activities conducted for a regulatory agency, the method-non is provided by particular closure relations, the scale- ology should:
up capability of a code will depend on whether or not the empirically determined closure relations have this capa- 1. Be systematic and practical, auditable and traceable.
bility. If the test facilities that generated the data are well scaled and the experiments are performed with appr.>pri-7tely scaled initial and boundary conditions, the empirical When applied to a specific severe accident scenario, the closure relatiens, and therefore a code, c:m be used in methodology should:
analyses of full-scale nuclear power plants. Otherwise limitations due to scale distortions must be assessed be- 2. Provide the scaling rationale and similarity criteria. J fore meaningful safety analyses can be performed.
- 3. Provide a procedure Ior conducting comprehensive reviews of facility designs, test specifications, and re-Second, because of discretization schemes used to nodal-suhs. j ize a nuclear power plant and perform calculations, the computed values are not local but are averages over very 4. Provide a measure or index to indicate the applica-targe volumes. Consequently, these averages are fanc- bility of correlations or models based on test data tions of node size and may affect the evolution or the ttm- from sub-scale facilities to full. scale nuclear plant mg of a physical or chemical process calculated by the conditions.
code. Furthermore, as nodalization used to model a nu-clear plant and small-scale test facilities differ (the latter 5. Quantify the effects of scale distortion.
have most often a much finer nodalization), the events calculated to occur in a full. scale plant may differ from 6. Quantify the effects associated with extrapolating those observed in test facilities. his problem becomes correlation and/or models beyond their data base.
ven more seriou if a test facility has some scale distor-tions that could affect a particular process of interest. 7. Relate in a systematic and unifying manner
- test facility design and operation, j Finally, because of " compensating errc,rs" that may be present in a code, there is no assurance that a code has the + test data accuracy and applicability, and capability to scale up processes observed m small-scale test facilitieno full-scale nucicar plaQ. " Compensating . correlations (or model) accuracy and applica-crrors" are generated most often durreg the code valida- bility to (1) computer code scalcup capability tion process. Since advanced codes have numerous pa- and (2) applicability to calculate the postulated rameters, coefficients, that is, " dials," improved agree- scenario for a full-scale nuclear power p'lant.
ment with experimental data is very often achieved by ad-justing some of these " dials." This " tuning" process of a 8. Provide a quantifiable and traceable procedure for code to a set of experimental data can introduce "com- specifying and prioritizing future experiments if pensating errors"in the code.The effect of such errors on needed.
the scaleup capability of a code becomes even more diffi-cult to assess if scale distortions are present in the facility It should be noted that requirements 4 through 7 provide or if the initial and boundary conditions of an experiment the information needed to quantify the code uncertainty are not properly scaled. to calculate the postulated severe accident scenarios.
27 NUREG-1365
A scaling methodology that rnects these requirements is 1. Structure the scaling process to follow the physically outlined below. based, hierarchical approach detaded in Element 1.
Such a structure is essential as it provides not only B.4 Eletnents of'SASM and Their Rationale the rationale for (establishing scaling and timilarity criteria but it ensures also that processes important ne proposed Severe Accident Scaling Methodology to the scenario are iaken into account in the simili-(SASM) consists of four primary elements as shown in tu:lc analysis.
8* *
- 2. Provide a structure and procedure that starts from a ne first element: Identification and Ranking off%cnom- gl bal, top-downviewpointandintroducescomplex-cna contains Steps 1-3. In this c'ement, scenario modeling ity and detail at each lower level. His is importar (and therefore scaling) requirements are identified by forconducting comprehensive reviews of facility de-sign, test specifications, and results.
(1) specifymg the scenano (Step 1), (2) establishing the event tree and selecting the accident path (Step 2), and 3. Provide a systematic and traceable approach to de-(3) identifying phenomena / processes along this path and rive and select similarity parameters. This minimizes ranking their importance (Step 3). Activities carried cut the arbitrariness, that is, the ad hoc approach used so in this element are designed to meet the first requirement often in facility design and test specifications, set forth in the preceding section. Dese activities are similar to those discussed and developed in References 4. Provide a method that can yield similarity criteria for B.1 End B.2, that is, they: processes that can limit the operational range of a system.
- 1. Provide a comprehensive, physically based frame-work for analyzing an accident scenario. This is es- The third c!cment: Quantification, contains Step 6, with sential to a scaling methodology that is systematic, activities designed to meet requirements 4 through 6, that auditable,;md traceable. is, to quantify the effects associated with scale distortion and/or with extrapolatmg correlations beyord their data
- 2. Decompose the scenarioin elementary components b se. Several methods can be used to achieve these objec-and provide a casual relationship approach. This is tives. One among them based on the funy-set theory of essential for identifying modeling (and therefore Zadeh (Refs. B.6 and B.7) appears very promismg m view scaling) requirements and understanding the role of of its successful application by Kubic and Stem (Refs. B.8 cach component (and therefore model)in the se- and B.9) to a problem concerned with quratifying model lected accident path. uncertainties and system rimilarity.
- 3. The fourth element: Application, which consists of Step 7, Identify and rank processes and phenomena along is designed to satisfy requirements 7 and 8, that is, to pro-the accident path. This is important to a scalmg vide the same methodology that etm be used to:
methodology that is not only systematic but also practical.The need for this screening process arises 1. Design and operate test facilities.
from the fact that it is not feasible either to desip an experiment or to develop a code that will scale prop- 2. Evaluate test data accuracy and'or applicability, erly all processes occurring during an accident. What is needed, however, is to ensure that physical and 3. Evaluate correlations and/or model accuracy and chemical processes important to the evolution of an applicability.
accident are properly scaled.
- 4. Evaluate computer code scalcup capability and ap-
- 4. Identify and prioritize research activities directed at placability.
resolving potential safety concerns related to an ac-It should be noted that treating experiments and code cadent scenano. This is essential to an expeditious closure of a safety issue. model/ correlation development in a systematic and unifying manner (by applying the same scaling The second element: Identification and Systematization of methodology to both activities) is a prerequisi;c for Scaling Criteria contains Steps 4 and 5. In this element, quantifying efficiently and more accurately code uncertainties ta calculate the postulated severe accident scaling criteria art (1) identified by performing a scenarios in a full scale nuclear power plant.
top-d own scaling analysis (Step 4) and (2) systematized by applying group theory methods (Step 5) discussed in B,5 Application and Demonstration References IL3, B.4, and B.5. Activities carried in this element are designed to satisfy the second and third re- A program has been initiated at the Brookhaven National quirements listed in the preceding section, that is, they: 1.aborato:y with the objectives to:
NUREG-1365 28
_ _ _ S DE
_ _ _ ND
_ _ AO
_ Y S C
_ _ _ L L P E R E
E H _ PDT O
E R
T A K) _ NWS _ Y _ AOMP U
T P NT R OS I
I R T ARI
_ DY Y
R
_ O _
_ E _
TO N N L H C E E D _ PA y E E
V DD NA _ ON l _ T _
C S
E NC I I ANE TA pT H Y H AC YM A pP S S A FO G AU -
F MN I
I I
TN _ O Y _
C E
L B
E R NE _ RU OA R _ Z _
S P A E EH F C G Z T
T DP _ U _
S S V
E I
_ R E S _ F _ Y N
E E S P L
_ _ OM P
P TIR A E P
X E
_ F _ _
_ DO _ _
D N N _
ANA _ _
A _
OI _ _
I R N N _ NTE O O
I
_ OAT I Z R I _ I T _ N T FA TI A 1 AONE _ 2 ATC _ 3 C _
- O T C CA I F 4 T I FNG M TC A I
FM G I
N I I O _ T I EI N _ T I
T E T N KN E NE N T TL N SA N N A NI LP
_ E E U E MDIAH RP _
M E
D YC I SS _
M Q _
MA P
E E L - _ E _ E E L L L
- E _ E _ E M zh
- 1. Develop a scaling methodology (SASM) that snects Plant Thermal-flydraulics and Operation (Seoul, Ko-the eight requirements listed above, and rea), November 14-17, 1988.
- 2. Demonstrate the methodology by applying it to the B.2 G. E. Wilson et al., " Quantifying Reactor Safety DCH problem. Margins, Part 2-Characterization of Important Contributors to Uncertainty," Proceedings ofthe Six-Furthermore, a Technical Program Group (TPG) has teenth Water Reactor Safety Information Meeting, been formed to assist the staff and provide guidance to NUREG/CP-0097, Vol. 4, pp.19-44, March 1989.
this work. The members wtre selected on the basis of their: B.3 H. A. Becker,"Dimensionless Parameters, Theory and Methodology " Appl. Sc. Publ., London 1976.
- 1. Knowledge of severe accident phenomena and issues, and/or B.4 S. T. Kline, " Similitude and Approximation The-
- 2. Internationally recognized expertness in modeling and scaling of complex systems and phenomena. B.5 A.12szlo, " Systema 0zation of Dimensionless Quantities by Group Theory," Int. of the Heat and In order to provide input from a broad spectrum of tech- Mass Transfer, Vol. 7, pp. 423-430,1964.
nical sources, the composition of theTPG was specifically designed to include technical talent from universities, na- B.6 L A. Zadeh, " Fuzzy Sets,"Information and Control, tional laboratories, and industry. Vol. 8, pp. 338-353,1965.
This wor k, that is, the development and demonstration of B.7 L A. Zadeh, " Similarity Relations and Fuzzy Or-SASM, is expected to be completed by the end of derings,"Inf Scr.. Vol. 3, pp.177-200,1971. j December 1989.
B.8 W. L Kubic and F. P. Stein, " Concept of System l Similarity in Chemid Engineering," AICIIE Y.,
References for Appendix B Vol. 33, pp.1986-1987,1987.
B.1 T. Theofanous, " Dealing with Phenomenological B.9 W. L Kubic and F. P. Stein, "The Application of Uncertainties in Severe Accident Assessments t.nj Fuzzy Set Theory to Uncertainty in Physical Prop-Probabilistic Risk Analyses," Proceedings of the erty Models," Fluid Phase Equilibria. Vol. 30, pp.
ThirdInternational Topical Meeting on Nuclear Power 111-118, 1986.
l l
l l
{
1 NUREG-1365 30 i
N C FORM 335 U.S. NUCLE AR REGULATORY COMMISSION 1.RE 1N ER
~^ ""~#~
IEE80/* BIBLIOGRAPHIC DATA SHEET isa anwuct.o= on one awa*1 NUREG-1365
'2. TITLE AND $UBTfTLE Revised Severe Accident Research Program Plan 3, DATE REPORT PUBLISHED FY 1990 - 1992 August I 1959
- b. AUTHORIS) 6. TYPE OF REPORT
- 1. VE RlOD COVE R E D uncharve D.orev 8 PE RF ORMING ORG ANIZ AT lON - N AM E AND ADDRE 55 II* NRc. prounk Dwnsnon. Otrar or Repron. U.S Nucwar Reputen *y Commouion. end mathno odoreu~ tr contractor, crowed-name ord maihne edarend Division of Systems Research Office of Huclear Regulatory Research U.S. Nuclear Regulatory Commission Washington,.DC 20555 9 ONSOR NG ORGANIZA*t?lN - fuAME AND ADDRE SS u!Nnc. ryne sanwasatere",iocontractor.provan NRc onosarm. orroce or Peooon. u s Nac+ar knouratory comnuuron.
Same as 8, above.
- 10. SUPPLEME NT ARY NOT ES
- 11. ABSTR ACT (200 words or out The revised Severe Accident Research Program has been orepared by the Office of Nuclear Regulatory Research to suonort the tasks and objectives discussed in the staff's " Integration Plan for Closure of Severe Accident Issues," SECY-88-147.
The revised SARP addresses both the near-term research directed at providing a technical basis uoon which decisions on im9ortant containment nerfomance issues can be made, and the long-term research needed to confirm and refine our understanding of severe accidents.
- 12. KE Y WOR D5/DE SCH:P10HS 4A ur words orparesen roer wei asssse reseerrners .n sornrens ene reporr. s 13 AvAstAaitii v 61*1tMtN1 Unlimited Severe Accident Research Program (SARP), Severe Accident Phenomenologi es .
Severe Accident Uncertainties, Phenomenoloaical Uncertainties, Direct Containment Heating (DCH) Liner Meltthrough, Hydronen """'"'Un cl a s si fi ed Transport and Combustion, Core Melt progression, Core Concrete Interaction, Steam Explosion """ *'""'bn cl a s s i fi ed Ib. NUMBER OF F AbE b 16 PRICE NRCfORM3JSC>'
UNITED STATES sveaat sounmaass aan 2 NUCLEAR REGULATORY COMMISSION "05'^Gjs*J!' S '^'" j WASHINGTON, D.C. 20555 ,,,,,,,o, m 9
OFFICIAL BUSINESS O PENALTY FOR PRtVATE USE, $300 12, 0 5, c.-- . I a., .. $
2332 2 I Ar:i 2 gs k,I VN;3C r ; y -o r ~. .t ,g Tf, p _ h,,[ g->q L I c t4 779g3 g pv a,
-c u N $ $ h ] Ij !} } 'y -
DC 2055-n M
0 M
C tr M
M M
M C
b e
M 2
m N
9 x
0 C
C h
=
=
L c
O C
s
_ _ _ - _.