Text

Forwards Viewgraphs & Handout Matls of Listed Presentations to ACRS Class-9 Subcommittee 840111 Meeting Re Severe Accident Program & Final Policy Statement.Related Issue Papers Encl
	ML20127F110
Person / Time
Issue date:	01/20/1984
From:	Spangler M; Office of Nuclear Reactor Regulation
To:	Case E, Harold Denton, Mattson R; NRC
Shared Package
ML20127C753	List: ML20127C749; ML20127F101; ML20127F110; ML20127F126; NUREG/CR-3440, Further Response to FOIA Request for Five Categories of Documents Re Severe Accident Studies,Including NUREG/CR-3025,NUREG/CR-3440,issue Papers & Ltr Repts. Forwards App C Documents.App B Documents in PDR;
References
	FOIA-84-928 NUDOCS 8506250034
	Download: ML20127F110 (438)
	v • d • e

-

z ,

January 20, 1984 .

Note to: Distribution List From: M. B. Spangler

Subject:

HANDOUT MATERIALS RELATING TO SEVERE ACCIDENT PROGRAM AND FINAL POLICY STATEMENT Enclosed are a set of VUGRAPHS of the following presentations to the ACRS Class-9 Subcomittee meeting on January 11, 1984:

(1) " Regulatory Use of the Source Term", R. Bernero .

(2) " Severe-Accident Phenomenology and Containment Loading Technical Issues".

T. Speis (3) Eight Technical I,ssues Papers e LWR Core Heatup Phenomena and Models - J. Han -

e Steam Explosion Energetics in LWR Geometrics - T. Theofanous e Key Phenomenological Models for Assessing Steam Generation Rates -

M. Berman .

e In-Vessel Coolability and Non-Explosive Steam Generation - R. Wright e Pressurized Vessel Failure and Debris Relocation - D. Powers e Non-Explosive Steam Generation and Combustible Gas Fomation in Containment - W. Pratt e Summary Views on IDCOR Reports 12.2 and 12.3 (Hydrogen Distribution

' and Combustion in Reactor Containment Buildings) - J. Larkins e Summary of the Joint NRC-IDCOR Meeting on Accident Phenomenology i and Contaiment Loading - S. Burson -

As a result of discussions with the ACRS at this meeting on Dr. Mattson's presentation of the " Severe Accident Program for Nuclear Power Plant Regulation" (Draft NUREG-1070), the first paragraph of Section IV-F-1 (p. 31) is being revised and a supporting Appendix "X" is being drafted as per enclosed outline on the subject: "Information Bearing on the Need for Generic Design Changes

! or Further Regulatory Changes Affecting Nuclear Power Plants". .

Miller B. Spangler, Special Assistant for Policy D' velopment e

Division of Systems Integration

As stated Fo/A-84-11e 8506250034 850314 6j PDR FOIA SHOLLYB4-928 PDR

. i . - o DISTRIBUTION LIST H. Denton W. Minners E. Case G. Sege R. Mattson J. Rosenthal - -

R. Bernero B. Sheron T. Speis L. G. Hulman' F. Rowsome R. Barrett Z. Rosztoczy W. Butler D. Muller J. Henry' J. Malaro F. Miraglia M. Ernst R. W. Houston L. Rubenstein C. Neltemes E. Jordan D. Eisenhut .

H. Thompson R. Vollmer D. Ross, Jr.

G. Arlotto S

0. Bassett W. Shields J. Funches A. Thadani C. Grimes G. Meyer C. Thomas J. Han J. Larkins R. Wright B. Burson ___. .

C. Tinkler J. Mf1hoan -

R. Baer

. o. ..

s-OUTLINE M. B. Spangler (1-18-84)

APPENDIX "X" INFORMATION BEARING ON THE NEED FOR GENERIC DESIGN CHANGES OR FURTHER REGULATORY CHANGES AFFECTING NUCLEAR POWER PLANTS I.

Introduction:

The Need for Forward-Looking Policy Development II. Technological Maturation: The Outlook for Surprising DeveTopments III. Modifications Due to the TMI Action Plan IV. Modifications from Plant-Specific PRAs and ISAP V. Changes Resulting from Unique Plant / Site Characteristics (Indian Point, Zion, Limerick, Big Rock Point)

.VI. Modifications Dua to Construction and Operating Reactor Experience A. I&E Bulletins B. AEOD Reports C. Utility QA and Staff Reviews VII. Results of Completed SARP Projects and Preliminary Information on Others in Progress A. SARRP B. ASEP '

C. ASP D. SASA E. Others VIII. Conclusions

+e== e O

, . , + , - - . - ,,m.-, , _ - _.-_.---___-___,,-_,,---m,m.,,

.w-,,,, ,,

i_ e..

P i

HANDOUT MATERIALS PRESENTED BY THE NRC' STAFF AND CONTRACTOR PERSONNEL AT THE ACRS CLASS-9 SUBCOMMITTEE MEETING ON " SEVERE ACCIDENT PROGRAM AND FINAL POLICY STATEMENT" JdNUARY 11, 1984 4

.. 2, .' I .. t .

1 .. .

1- ,

i

. j. _

i

't REGULATORY USE OF THE SOURCE TERM ROBERT M. BERNER0, DIRECTOR ACCIDENT SOURCE TERM FROGRAM 0FFICE OFFICE OF NUCLEAR REGULATORY RESEARCH

.' (pwnsENTEbAT ACRS Cws3.SJ3 committee MEEmed JANUARY 11, 1984 e

a 3

I a

!=

-j i .

I e--...

. , - y........- ............ .. .. ..-.

..- . . - - - - . . - =

2.

} -

I %e Vn et9

.. ___ REGULATORY ACTIVITIES USINGORBASEDONSOURCETDIE

1. REACTOR SITING AND CONTAINMENT PERFORMANCE e TID-14844 e REG. GUIDES 1.3 AND 1.4

2. EMERGENCY PLANNING BASIS e NUREG-0396 WITH EPA e WASH-1400 RISK SOURCE TERM

3. EQUIPMENT QUALIFICATION e TID-14844-

4. FES CLASS-9 ACCIDENT RISK e WASH-1400 RISK SOURCE TERM

5. OTHER ACTIVIES e HIGH POPULATION SITE APPRAISALS

", (PLANT SPECIFIC RISK) a e COOLANT SYSTEM DESIGN BASES

- (1% FAILED FUEL)

.i ri '

-j .

-l 1

- .__..____...__..__._.1 - _ _ _ . . , .....____..___.__.______ .-

-1 .

$ tY A t t'O

~

i

___ . HAZARD STATE 1

.1

' S' TATE: MAJOR DAMAGE TO CORE NO MAJOR RELEASE OUTSIDE REACTOR VESSEL 14AZARD STATE 2 l

STATE: MAJOR DAMAGE TO CORE MAJOR RELEASE OUTSIDE REACTOR VESSEL NO MAJOR RELEASE OUTSIDE CONTAINMENT

, HAZARD STATE 3 STATE: MAJOR JAMAGE TO CORE

- - MAJOR RELEASE OUTSIDE REACTOR VESSEL MAJOR RELEASE OUTSIDE CONTAINMENT l

l l

. d..

1 *

-l

i

~. .

..e s

P h P P'MfT'O h SEVERITY ,

/

x SOURCE FWR :CM:: NAN" ACOIDENT SEOt.EN; E'S vs. RE 5.ASE CATEGORIES TERM CATEGORIES

/ _

.=un e.n-.iu . . . . .a T v .T . . i . . i . . .; .

/

f ** *" t . "a "* "

aa .*" .*

.***

aa' ...

aa

aa" i ACCIDENT - - =4 - =4 - ...

.: SEQUE>.tCES u .i . . . .a ..a .n i.a m a.. .

AND =

==~a. n un .n i.a .

. =< u.i. ..

PROBABILITIES ..,,, .

a un, ,

/ . -.u..g ua" j . .a* I wa-' m.* a* i.u ua* m.-' mi.-

E.uus ..u~n. a s~

i

m. v.=..*.. n..~..a s ma i - . .. . ~un vs ...

n e

i ua v .a~.u va . . ~ . . i.. s.w.

u i,~..

ua

.,~

m.

's s,-u.. .n y u -*. . i,-...

..~u... a a

,. i,*..

un i."s.a ua" ua" u .."

us.* ..n* .ma-

.,,.a..... 1 ua" lua' ua 5,

s.a . . s, ua .. v-.na.. 5,=d. a

. me i,ua.. i, w.

5,~

a y ma. i,-...

ua. i,~..

ua i,

ua y=a..

u .

i, a

. ,na.

=.. v. .a

. v..aa

. i,--..

w-.

v..a.a Y u+s

y. ua. vs.a i.a ' ua a ma a ma' l.a*

.,a.. m .-' ma- '

=~ -- .

"u -"". . .

INITIATING

"" o . "", .

- . a. "a'".

EVENTS = = . - .

1 =a....

.~..a. ua " . .a a l 1.a* l uu". , i.a a u,:a 1.a ' l

. -.a, . .

, ,- a . u. ." a* m." ua * -

l

=a.a.w. 2, =~ =-* . =a s,

..&. a.as . a. = ut-'.

"P deIa.T go. a* Tue-.

T.g..

n.a. * "

fug

3 4.

58WT

T 3 4.* asa.*

" 3.a..

- . ~

3.a.

T , um . &. .

n.a. ' 36. 4.* ' '.t.* 3 4.* ' l 3.at" b.a. *

.D B5 713 .P .s.& .CCIM em'Ps.is".* p eSL &f tafEE.r?

=, .... ..- .- .,

= ..- _...

- ..- ...... ..- .... . .- X .- g . m 0-s

= u ,.- ..a _.4 - _.- .. _.- .- .A.

1 a me a u ==sm ., c w . _ a. . ,.. . i , u reu_..3=.

e.4 , , au., .

't l

l l 11

.i i e

h

2 e e --

a w m .

. ~ * . . . _ . ~ . . ... . .. _o i

i *

bernero 5, O

m w m U m

~ <

5

wb w

. - .m ow g M

"J y z

. W= - m- o_ m o = ma - _m

_ u < m <<

WM uJ 6 = W E ec 2

9 - w= a.

et u.>

, , m m - . .o 1

r X

TABLE 5-1 St CF ACCI N . G CO

[ \

S / -r.0. . am J ' - r2: T ( 3E".AASS FRA ""I tst QF CCRE DN'Df"*Ntv RE*,JJ T I*I PE2mAa!1*TT y gy, lastatal (10 DtW) Be-Rt Org. I I Oa-m To-en mm.sr ha 14 (erl (tr)

CAhi haerter-Tr farl *#

3 520 W 0.9 6a10 0.7 0.4 0.4 0.05 0.4 3a10

- Pwn 1 910 2.5 0.5 1.0 25 *#

0.9 7a10

I 0.7 0.5 0.3 0.06 0.02 da10 Put 2 Aalt 2.5 0.5 1.0 0 170 *

3 0.2 0.2 0.3 0.02 0.03 3a10 '

2.0 0 6 0.0 6a10 Pwn 3 ea10" 5.0 1.5

0.6 2a10

I O.09 0.04 0.03 Sa10 3a10* esto Pwn 4 5 10 '

2.0 3.0 2.0 0 1

3 * *

3 0.03 9a10 Sa10 1m10 talo 7a10 '

2.0 4.0 1.0 0 0.3 0.3 2m10 PWR S 7 10 *3 *I * *

1.0 0 m/A 0.3 ta10 am10

en10 la10

9a10

Tal0 1a10* '

Pwn 6 I 6a10 12.0 10.0 *$

aA 6a10* 2a10'$ 2a10 lal0 2a10 La10 1alo 2a10 '

Pwn 7 da10* 10.0 10.0 1.0 0

0 0 a 4a10 0.5 0.5 m/h 0 m/A 2s1C Sa10 La10

Sa10"** 1a10

1m10 0 m/A 3a10 Tal0 ' la10 ' 6a10 1a10 la10' O O Pwn 9 4a10' O.5 0.5 N/h

I 130 1.0 Tal0

I 0.40 0.40 0.70 0.05 0.5 Sa10 m8m 1 1alo 2.0 2.0 1.5 25 *I J

0 30 1.0 Tal0* 0.90 0.50 0.30 0.10 0.03 4a10 ada 2 6al: 30.0 3.0 2.0

$ 20 1.0 7a10" 0.10' O.10 0.30 0.01 0.02 3:10' m8m 3 2a10 30.0 3.0 2.0 25

I * *

N/A 0.6 7a10*d ea10*d Sal 0 4s10 6a10** 6a10

1a10 BWh 4 2a10 5.0 2.0 2.0 25

14 tal0*' 6s10*II 4m10 ' ea10-12 g,gg 0 0 3.5 5.0 m/h 150 u/A 5s10 ,

se 5 lal:

' (a) A tascussson of t.ne tootcpes ased am the st.uty as found an Append;a V:. Bacaground on the isotope groups and release mechanians as found an Appendra V:2.

f.nl includes me. Ah. Tc. Co.

(c) includes ad. T. Oe. Pr.14 2 . Am. On. Pu. pp. St.

(d) A lower energy release rate ther this value appines to part of the parted over whach the radioactivity is teknq roleesed.

The ef f ect of lower energy release rates on consequences is found as Appendia VI.

1-I

(

l . . . . . . . . . . . . . . . . . .. . .

c I

i - - - - - - - - - - . _.-._ -____ _. _ _ _ _ _ _ _ _ _ _ _ _ _

i .

3er1ero_ .

b'

~

SOURCE TERMS WASH-740 (1957)

A FEW SOURCE TERMS TIED TO POSTULATED HAZARD STATES - A HAZARD STATEMENT WASH-lu00 (1975)

A SPECTRUM OF SOURCE TERMS TIED TO CALCULATED SEQUENCES AND THEIR PROBABILITY - A RISK STATEMENI NUREG-0739 (1980) . -

A SAFETY GOAL CONTAIN!i1G A FEW SOURCE TERMS TIED TO POSTULATED HA2. RD STATES WITH ASSOCIATED PROBABILITIES - A RISK STATEMENT NUREG-0772 (1981)

A REVIEW 0F THE SCIENTIFIC BASIS FOR ESTIMATING SOURCE TERMS - A METHODOLOGY STATEMENT l .

-{

. .i

=-....

Sernero 7.

t e

HUREG-0773 (1982) o A REBASELINING OR REEVALUATIOR OF WASH-1400 AND OTHER PRA'S - A RISK STATEMENT o A SYNTHESIS OF SITING SOURCE TERMS FOR TYPICAL LIGHT WATER REACTORS - A HAZARD STATEMENT o A COMPENDIUM 0F CORE HELT OCCURRENCE AND SEVERITY FREQUENCIES - A RISK STATEMENT BMI-2100 (1983/19814)

A SET OF SOURCE TERMS FOR SELECTED SEQUENCES IN REFERENCE PLANTS - A HAZARD STATEMENT NUREG-0956 (1984?)

o A REVIEW 0F THE SCIENTIFIC BASIS FOR ESTIMATlHS SOURCE t

TERMS - A_ METHODOLOGY STATEMENT

~

o A SPECTRUM 0F SOURCE TERMS TIED TO CALCULATED SEQUENCES AND THEIR PROBABILITY - A RISK STATEMENT o ?

5 u

) . .. . . . . . . . . . . . . . . . . . . .. . . . . .;... . . . . .

I' 3ernero .

4 SOURCE TERM RUIFMAKING l

A. SEVERE ACCIDENT RULE B. SOURCE TERM RULE TYPICAL SOURCE TERM FOR 3 HAZARD STATES INSIDE AND OUTSIDE CONTAINMENT IMPLICIT OR EXPLICIT CONTAINMENT PERF0P3ANCE (" LOW LIKELIHOOD OF HS-1 AND HS-2 WITH SUBSTANTIAL ASSURANCE THEY WON'T PROGRESS TO HS-3")

C. SOURCE TERM TABLES RULE A TECHNICAL REFERENCE USED WITH PLANT SPECIFIC ANALYSIS

D. TABLES AND METHODOLOGY RULE SIMILAR TO C i

. - E. SOURCE TERM METHODOLOGY SIMILAR TO APPENDIX K APPROACH WHAT PERF0PfANCE STANDARD LIKE 2200 F7

.)

l; b...,... .

., .x . _ _ ._.

k.

, %grnero .

~

A, SEVERE'ACCIDEllT RULE t

IECHNICAL BASIS e SEVERE ACCIDENT DETERMINISTIC ANALYSES e SEVERE ACCIDENT RISK APPRAISALS

, o RISK REDUCTION COST-BENEFIT ANALYSES e UNCERTAINTY AND SENSITIVITY ANALYSES COMMISSION FOCUS e HOW SAFE ARE PLANTS?

e IS FURTHER. IMPROVEMENT WORTHWHILE?

BULEMAKING STATEMENT e PLANTS ARE ACCEPTABLY SAFE IF...

~

OR '

, e SEVERE ACCIDENT RISK IS LO'W ENOUGH IF...

4 i

f e

e-dj f].

,e .

t

-]

l!

l .._.. . . .. ._ . . _..

-. ~

. '. a-

.[

b rnero ! g, 1

- TABLE - 10 CFR ?

SOURCE TERMS FOR REGULATORY ANALYSIS !

' I

- SOURCE TERMS i

HAZARD STATE 3 e i

HAZARD STATE 1 HAZARD STATE 2 I

{

I RELEASE CHARACTERISTICS

i. '

SEVERE CORE DAMAGE CORE HELT CORE MELT ACCIDENT TYPE

--- SIGNIFICANT LEAKAGE OVERPRESSURE 1

CONTAllMENT FAILURE MODE 11/ DAY LARGE LARGE .

CONTAllMENT LEAKAGE

}

TIME OF RELEASE (HR) 0.5 10 1.5 I, ,

1 2 2 ,

l ! RELEASE DURATION (HR)

WARNINGTIME(HR)

--- 4 0.5 ,

10 4

ItELEASE HEIGHT (METERS) 10 10 0 0 'O

.l e ItELEASE ENERGY l 1

1 : INVENTORY RELEASE FRACTIONS '

1 Xe-Kr GROUP 0.05 0.9 1.0 l l -0 -3 '

I GROUP 1x10 3x10 0.45

}

Cs-Rb GROUP 6x10'I 9x10-3 0.67 Te-Sb GROUP lx10-9 3x10'2 0.64 f -3 i Ba-Sr GROUP lx10 II 1x10 0.07 j 0 2x10-3 0.05 '

Ru GROUP

~4 j 0 3x10 9x10-3 La GROUP l

i

) .

e

<w. --

$9f n PYO. f C. SOURCE TERM TABLES RULE TECHNICdLBASIS s SEVERE ACCIDENT DETERMINISTIC ANALYSES e SOME SEVERE ACCIDENT RISK APPRAISALS COMMISSION FOCUS e WHAT RELEASES ARE CHARACTERISTIC 0F SEVERE ACCIDENTS?

e HAS STAFF APPRAISAL TO ESTABLISH THESE BEEN SOUNDLY BASED, COMPLETE, ETC.

RULEMAKINf STATEMENT e THESE TABLES DEFINE ACCEPTABLE RESULTS FOR FISSION PRODUCT TRANSPORT CALCULATIONS FOR USE IN PLANT SPECIFIC CALCULATIONS D

a

,4.,+- g*********-

..e.. * * " ' '

. u. ~._ . _ . . -

l !.. . . .. . _. m. _ . . , _ - .......2u.._.,.

h e

1 - 'q tg" g 5 I :

52 1

~

E$=5 g g .!

t :

3 2 -

. =

v 9

e.$

ga *

& K ,

s I 5.I

-ww

'o 2

~

se -

W w .

'e g a- ie

[=55 -

_c )'

" "g 3 21

. < Em

. .~ P' 3

g % g g m-

- , _e

= g

, , e *

,sY

'5~" rlEIg!" -c=3

e5-<I

1=.=e K. er 3

E! .

= .

E.

s' $$ ;

E 5!

-s s E

2e 3-1:-

3 zw E 23

!1 O'g 52 gE 2.

g gt

.t j ar_ g o,

. ~ EE El' cd

. c$,

. .x,

' --w-E. y * a l

- *2 .l

-!I I I ysO " 3

!w -E e se .W Eg _=

-a 5. ~=

Eu 2 g

NI k

[ !!

^ E!!

T-

+E!* 'P

)

m1 7 z1 1 55 53 ge W

{I c

!=I EI,

!I, m

Ec, oj y8g - EE:3 ," ,,,

8 i

=W 8 a 5 *sW -E =. E

=. *a. . g .gg s

g

= -

!,!N e. hm .

I! ;

iss

- * ! 9 ,5 E '

p -_ \ .,

h i

I

. 4 i

e l

l .- -

u =. .a .~ . . . . .

--.. .. - ~ - _...=.a,.,

s pg l7 8

'6 erne ro I

f:

(..

l

! Table 1. Limits on Occurrence of Hezerd States I

[

t Decision Rules on Mean Frequency Probability Goat Goal Level Upper I.imit i'

Heterd State

~4 J

Significent Core tiemage .

f g<lx10" 3

(> 10% of Weble gas inventory Less then 1/100 f*d( 3:10 '

l leaking into primary coolant) per reactor lifetine per reactor year e reactor year ;

I Large Scelo Fuel Melt - LSFM Less then 1/300 f <lx10

~4 f <5x10"# '{

(> 3M of oxide fuel becoming "

per reactor year l

molten) per reactor lifetime p,r reactor year i

Large Scale Uncontrolled Release fg 40.1 from Contstament given LSPM Small, given e Large f,7,< 0.01 !

I Scale Fuel Melt per LSPM per LSPM -

4 y

(> 10% of Iodine inventory ,

and 90% of noble gas) I t

1 is the frequency of Significant Core Damage po'r reactor year. i j* f,, i, is the frequency of Large Scale Fuel Melt per reactor year.

f, r I is the frequency of Large Seele Uncontrolled Release per Large Scale Fuel Melt, j i f,j, '

%e upper non-occeptance limits must be settsfied for extended operation of a new pla I' issuance of a construction permit, Once the risk level ;;

range for case-by-case consideration of uncertainties and competing risk. decis . g, schievable within the cost-effectiveness criterion of Table 4._- Iii '

I 1

l

i .. ,_

w m -- -- -- -

i E

i Berino .

'! Table 22 Risk insights versus design *

.t

.i Probability l- Probability *** of at least 1 fatality given 2 of core damage j per reactor year core damage Reactor Name Type 1/3 s

Peach Bottom BWR Mark I 2 x 10 5 j

PWR Dry 6 x 10 5 1/10 Surry PWR Ice Condenser 4 x 10 5 1/25 Sequoyah PWR Dry 3 x 10 5 1/100 Indian Point PWR Dry 3 x 10 4 **1/3 Crystal River PWR Dry 3 x 10 5 **(1/10-1/100)

Biblis B (FRG)

Calvert Cliffs PWR Dry 2 x 10 4 ""(1/4)

Oconee PWR Dry 2 x 10 4 **(1/4)

~

There are large uncertainties associated with the numbers in this

- table.

nn Very rough preliminary estimates, san -

Reflects median values point estimates.

O e

e l

l i

l l

'l

~

i i .i l

,l . . . . . ._ . .

-*-e.w -

,..w w+ -

. . ._..2._. .. . . . . _ _ . . _ _ . . _ . . . . . . . ___

l 4

t, fTP)PPO_

lb' ,

i

i 8

f l

'l Table 23 Siting source ters (SST) fission product releases 5l __

7.,.

j,

~

Released fractions (to atmosphere).

SST Xe, Kr I, Br Cs, Rb Te BaSr Ru La

I 1.0 .45 .67 .64 .07 .05 9x10 8 2 0.9 3x10 8 9x10 8 3x10 2 1x10 8 2x10 8 3x10 4 3 6x10 3 2x10 4 1x10 5' 2x10.s 1xio.s 1x10.s 1x10.s 4 3x10 8 1x10 7 6x10 7 1x10 8 10 12 NEG NEG

~

5 '3x10 7 1x10 8 6x10.s 1xio.2o 1xio.12 NEG NEG NEG = Negligible i

e

<l e

b

+

0 o

i

~

.2 ,

I ,*

4 g 1

2 Y

f 1

I k

m

. . , , s -

1 1

SEVERE-ACCIDENT PHENOMENOLOGY AND CONTAINMEi,T LOADING TECHNICAL ISSUES BRIEFING TO ACRS CLASS-9 ACCIDENT SUBCOMMITTEE T. P. SPEIS, NRR JANUARY 11, 1984

?

w w.-, . - + . . . , ,m. ., - __, .- .- - , _ . _ , _ . , ,

5 PIE ISt BACKGROUND NRC/IDCOR TECHNICAL EXCHANGE (T.E.) MEETING ON ACCIDENT PHENOMENOLOGY AND CONTAINMENT LOADING (11/29-12/1, 1983:

HARPERS FERRY, W. VA.)

OBJECTIVES OF NR'C/IDCOR TECHNICAL EXCHANG'E MEETINGS

PRELIMINARY RESULTS OF FIRST NRC/IDCOR T.E. MEETING EVALUATION OF IDCOR METHODS INCOMPLETE PENDING EXAMINATION OF CODES AND CALCULATIONS: 1.E., SPECIFIC TREATMENT OF THE PHENOMENA IN IDCOR'S COMPUTER CODES, AND FINALLY, APPLICATION TO SPECIFIC PLANTS e TECHNICAL ISSUE PAPERS; OBJECTIVES AND STATUS e

e

25.

!LPE.L5_

J PRELIMINARY RESULTS GENERAL AGREEMENT ON OVERALL APPROACH 'N) SEVERE ACCIDENT PHENOMENOLOGY AREAS OF DISAGREEMENT RELATE TO SPECIFIC MODELS AND ASSUMPTIONS RESOLUTION OF DISAGREEMENTS REQUIRES:

- BENCHMARKING OF IDCOR MODELS WITH EXPERIMENTAL DATA AND NRC SPONSORED CODES

- QUANTIFICATION OF IDCOR MODEL UNCERTAINTIES

- ADDITIONAL RESEARCH IN SELECTED AREAS

- EXAMINATION OF THE IDCOR CODES / CALCULATIONS AND THEIR USE AND APPLICATION TO SPECIFIC PLANTS

- AGREE TO DISAGREE G

'I

svets- 4.

ACCIDENT PHENOMEN0 LOGY AND CONTAINMENT LOADING TECHNICAL ISSUES SPEAKER

1. LWR CORE HEATUP PHENOMENA AND HYDROGEN GENERATION IN VESSEL J. HAN, RES

2. IN-VESSEL COOLABILITY AND

~

NONEXPLOSIVE STEAM GENERATION R. WRIGHT, RES

3. STEAM EXPLOSION PHENOMENA T. THE0FANOUS,.PURDUE U M. BERMAN, SNL

4. VESSEL FAILURE AND DEBRIS RELOCATION / INTERACTION D. POWERS, SNL

5. NONEXPLOSIVE STEAM GENERATION AND HYDROGEN FORMATION IN

~

CONTAINMENT T. PRATT, ENL

6. CORE DEBRIS-CONCRETE INTERACTIONS B. BURSON, RES

,- 7. COMBUSTIBLE GAS PHENOMENA J. LARKINS, RES C. TINKLER, NRR

S9E 15 5.

IMPORTANCE OF PHENOMENA AND CONTAINMENT LOADING ISSUES MEASURE OF IMPORTANCE - IMPACT OF ISSUES ON THE CHARACTER-ISTICS AND TIMING OF CONTAINMENT FAILURE AND THE TYPES AND QUANTITIES OF RADIONUCLIDES RELEASED FROM THE CONTAINMENT (AN IMPORTANT RELATED QUESTION: HOW MUCH IS THIS IMPACT A FUNCTION OF REACTOR / CONTAINMENT TYPE)

LWR CORE HEATUP PHENOMENA AND HYDROGEN GENERATION

- THREAT TO CONTAINMENT FROM H 2 GENERATION AND COMBUSTION

- AFFECTS THE SOURCE TERM IN-VESSEL C00 LABILITY AND NONEXPLOSIVE STEAM GENERATION

- DOES IN-VESSEL QUENCHING THREATEN THE REACTOR VESSEL OR THE CONTAINMENT?

STEAM EXPLOSION PHENOMENA

- POTENTIAL FOR EARLY CONTAINMENT FAILURE

- AFFECTS THE SOURCE TERM (TIMING, MAGNITUDE OF MELT RELEASE,

~

AND CHARACTERIZATION OF RELEASE)

VESSEL FAILURE AND DEBRIS RELOCATION / INTERACTION

- CEBRIS DISPERSAL DEPENDS ON IN-VESSEL CONDITIONS AT TIME OF VESSEL FAILURE AND SPECIFIC REACTOR CAVITY CONFIGURATION; DISPERSAL DYNAMICS AFFECTS PRESSURE GENERATION RATE AND POTENTIAL FOR CONTAINMENT FAILURE

SPE is 6e IMPORTANCE OF PHENOMENA AND' CONTAINMENT LOADING ISSUES (CONT.)

NONEXPLOSIVE STEAM GENERATION AND HYDROGEN FORMATION IN CONTAINMENT

- POTENTIAL FOR CONTAINMENT FAILURE BY SLOW STEAM OVERPRESSURIZATION

- ADDITIONAL H 2 ADDED TO CONTAINMENT; POTENTIAL FOR H2 BURN INITIATION CORE DEBRIS-CONCRETE INTERACTIONS

- POTENTIAL FOR CONTAINMENT FA.ILURE VIA BASEMAT PENERATION

- GAS GENERATION FROM CONCRETE DECOMPOSITION, ADDING TO THE POTENTIAL FOR CONTAINMENT FAILURE

- AFFECTS THE SOURCE TERM COMBUSTIBLE GAS PHENOMENA

-H 2 DURNING (COMBUSTION AND/0R DETONATION) COULD LEAD TO CONTAINMENT FAILURE (EITHER DIRECTLY OR VIA DEGRADATION AND/0R LOSS OF SAFETY EQUIPMENT) 4 9

u...

-w__,_.....__.__ _- _ i . .___._ __ __ . '- . _ _ . - - . . . _ . _ . .

1.

LWR CORE WATUP ROUER N0 ftMS

- A PREUMINARY NRC/CONTPACTOPS PESPONSE To IDCOR PRESENTATIONS AT llARPER'S FERRY, % ST VIRGINIA NOVBEER 29 - DECEMBER 1,1983 ENTRIPirTORS J.HAN,NRC-RES G. MRIND, NRC-RES R. BAPEIT, NRC-NRR J.1.0NG,NRC-NRR C.ALLISON,ITE.(SGMP)

P. CYBULSKIS, BCL (RR-FAUD S. H0DGE, OPIL GWR-FARCH) i l

PRESSi1E BY JAES HAN, NRC-RES ACPS SliBCCffilTEE 01 CLASS 9 ACCIDENTS J41AW 11, ITA O

M

~ ~ '

d ,8 . . . . . . . . . . . . . . . , . _ . . . * . . . . . . . ..

e ,

e 2'

NAo LWR CORE EATIP ROUBlA ATO fim S

- PRESB1 TAT 10tl OUTLifE

1. DESCRIPTI0fl 0F CDRE IEATl'P

2. PRELIf11tMR/ C0ttERS Oil IDCOR APPROACH

3. PROPOSED RESOLUTI0tl r.

" - ' ' w,- . n- _ _,_

OAM ,

3.

1. ESCRIFTI0l 0F CDPE lEATlP IfM%LOPEPATI&d v

ACCIIEIT OCCLES

- L.0CA,to EAT SliEETC.

RCS i.0SES C00l#!T

. INSUFFICIB,7 COOLANT Y*

CORE LNC0VERY EGIf6

' INSUFFICIENT ! EAT

, ,PM/AL CORE CORE TDPEPATURE RISES ETUP

<r COREDEGPADATIONBEGINS r

't

( POSSIPLE CORE tELTD0kN1 I .

BC IFFORTNff ISSLES: FISSIM1 PRODUCT RFIFASE RATE AND ODilCAL RWE AT HYDROGEN GBEPATI71 PATE.

_ - - . _ . . , _ . . _ r y __ _ ,..-_,--_r.y. ..- . _ ____m_ _ _ _ , --,_y -

.~. .-

. KAM 4,

1. ESCRIPT10fl 0F iDRE WATIP (C(NT'D),

"" s,s ...a 4 , n ta t

. ~ 3... v. __ __ _ -- ero, -.idt 7

, a np y _ . _ . -- - - - -

te " '3 !kb "' *' ~

4

. rk g ,, _ _ _ _ _ _ _ _. _ 4 . n.~ %

j' 1.os K k

con orde w w

-' , & T >, \ 00 0 K sless K -- -- - 39 cL1, n.g@.4 - - -

h2 o K <

( 22p pt.r.) ,

> [

n u.c cM

Time. Wg N=0-y

%63

__.. .. . . s WAN 5, UR COE EATUP P!BUEM #0 NDElS

FISS10fl PRODUCT EEASE ND TPKISPORT MEDER NE t0T ADDRESSEDAIDEED Ifi IDCOR'S C&RITER CODES - BR-EATUP #0 BR4EATLP. NDELItE ASSlWTIOFS

!%Y BE SIGilFIC#lTLY. Iff0RT#lT FOR FISSION PRODUCT ISSUE Ti'#1 FOR HYDROGEN ISSUE.

IDCOR Wlu. DISCUSS M FISSION PRODUCT ISSE Ill TE EXT NRC/IDCOR t1TG 01 FEB. 7 #0 8,1984. .

em

-m-- -

F Al\l be LWR CORE HEATUP f4SUBM AND fTYR

2. PREllillfMRY CatBITS ON IDCDR APPanACH AGREBE1TS

PODELIiG FOR HYDROGB1 GENERATION UP TO ZR ELTIfE IS GBEPALLY VALID.

DIFFEFECES tRC IDCOR e BLOCKAGE FORi% TION ROD BALL 00 NITE / RUPTURE ASSlfE AT ZR ELTING FREEZE OF P0L1Di CORIlE Za OR ZR0 2 P0VBET ATIETIfE t0T ALL0hED IMIL FUEL MELTIfE

!

FUEL LIQUEFACTI0il ATZRfETIfE NOT ALLOWED LNTIL FUEL ELTIfE l

BLOCKAGEUPONHYDROGBi CALQJLATED FURTER ENEPATION ASS'?ED GBDATION TOSTOP CORIlli (ZR,ZR0 , FUEL) CALO1ATED ASSlf G TO ACCLMJLATE ABOVE 2

ELDCATI0il LG G CORE StPPORT PLATE Y -

PWR SBLDCA WO ECCS TMI-2 PRODUCED 200KGs0FHYDROG91CALDAATED.

l l ~450 KGs 0F FM)ROGBi l

l

.,~. .. .

gAN 7.

LWR COE EATUP PIBITOlA ND t0DFIS

[.

3. PROPOSED ESOURIail e IDCOR SHCul.D OJNITIFY TE UNCERTAINTIES (F ITS CALQJLATICIS

( THIRD HRC/IDCOR MTG).

IDCOR #0 NRC SHOUll) DEFIE INPUT FOR SEECTED SASA CALCL'LATILES.

IDCOR VERSUS !RC CALOJLATI0t6 SHOULD E CWAED.

4 9

e

. . J l

1 .

i EX PLO dlOU tTE A M E U E R G E T / l' 1 LWR

~~

tw G EO N E7RIE S -

T' H E WEY f A/ G RE b lE A)TJ :

.

E wp Aosivify of 2eac/or- IbaMr.

Propog afien an d Eflicien cf n 0 UQnfibies In Pretniking '

. Inertia con.cfra.in t.

Ref. PA/E-RI-141 ; AJud Eey De.c g, sot (tst,s) 9 RE SE N TA TloN To A CR S SUB tom m rrras ou class-9 Acc t DEN TS J AWU6RY ll y 1980 f TG.Tlt'Of4nCqf

?urduel)my,

Thec4a noas i,

. 2 ~

l BROAD EXPLOSION CONCEPT STEP 1: COARSE MIXING

Guantities g4 BFA wikiJ STEP 2: TNGGER STEP 3 G p~

in .

Premixtng l

A Explosivi&"

lg,g'e 6

i STEP 3: COHERENT FINE-SCALE MIXING :

5, Y " Propagation -

,>- S efbcr ency" TO _WHAT EXTENT SHOULD EACH 5;1P BE CONSIDERED 7 i

3, ThAca m A T E gi e l.C 6 X PLost y/ TY OF RE ACTOR I

"*T couo ll 0

%poe 8fa n keis ei.\.s>

e.* g

4

Epe cial CondiNoru

' Y !!

n ,

y E x p t' brion.t founda me nfal GuelNOh! ll II

4 TLeemal E'y pro < ion) Q.

80 o> Vapor or (s) conRyurak of premixiure w.

Re quivel +riqq er- ?

R6/e o / E y pe cime n+i 4 (cale e

D e

-_ ---e .- . . - . _ _ _ , . - , , - _ , . , , _ _ -.e - _ , , -_ __ , _ , . . . . , _ . - - - - , - - _

s 4

Theohn*5 E x n otl VI T Y OF RE RCTOR G1 RTE 9/MLC Purdue : Jiave malniaine.l hard 40 fn'79er

/>u/ used Py ~ 15' due. fo uncerfn.tnWes and eack of funda nenfat undersfandry.

,fa ndia : Earty experinsenty thoued hard fo frigo Recenf exper/nents demonstrafe cponfa neous fripert f.e. PS ~ .l ,

Idcor : Inany defaifed aryumen/s lo' r

!?at canhof r of expfosfonc.a fuouf o non-expforitify'u/e Sfafus Reacfor Ma feria /t a et tofler expGbsire

.tstue CfoJe.l .

Iny f/co kons Exfend o f pre.mixty 1/mifed .

-,-m-- - _ - - - - . . - - - - _ _ - . - _ . , , . , . , , , , _ . _ _ , , _ _ _ _ _ _

l s -

EVALUATION OF P(leff; mg x m e )

l EXP. RXhl. PREMIXTURE ZONES A ,

, . +

LENGTH 7 SPEED 7 ATTENUATION?

P FEEDBACK 7-DISPERSION?

FRAGMENTATION?

HEAT / MASS TRANSFER?

p "f The ok hnotr5

-EUW" g m

Th e ohnoccs 6.

P R o p a c e'll O A? d E FFICl2 A?C Y Puedue : Opper fuel frayment tize. in preututure ute Upper bound conversior) eN7iciency [D% .

Boti b17 Afy un cerla./n .

~

J'o n dia : yery lov e/ftciency per experinenft i a r*g Recent up dafe : e ir 7,' .

Id cor : fuel frag inenf ir> perinIxfare ~l cm .

Very toto eWicfency per land /a.

Recenf update : D o not ca re .

I Slafui funda menfalt of fragomenfoNon onerfolo fu fdj eria Wish 'l cou /e nytX tcafev ' ce nd Mt "e ff/ciencies .p ting NPflCQ h0MC -

Fuefler sfudiet reyutred lo supjert lover efftclencie.t a n.t fie 40 penixfure conflyutRNoy, anJ fo de/ ermine effect af ofter var (cr4 Ver t.r.,Yotd fra cffon , inerfiq contfrulnf &c.

~

Falbia)2ciotukon : In ,,ed feu nonfat, theonftcaldy

+ie- dos n recen b deuefor,nenfs .

l -- .- ____: _ _ _ _ _. ------._ .___ _ ._ _ _ _ . _ . . _ - - _ _ _

2. , . .

7 777 e dO noud l G O R D T ITIE J IA) PREMIXTUEE m/ -

e~~

- t-;

v- _-

1(E x =_ -J J L, -

d 1

D$ sL'h tr y

=ms i .

frpproacter: o> Trontien/ brec hup 1 nixino I-p 6> E-S 1-b flermal 1imif.T

\ Le) S-S t .b +Lermal 4iinift l

G) Travaient S-b breakup eL J

l l~ Tiernal Umift .

- Scale Efhdr ere Crocial ifere.

1 I.

l

Meobnoud F.

4 GU A NTITIES IN PRE MIXIS? Q Purdue : Euel breakup 4 a rat'e proa c.r. ke.l on hydrodynaela anl peomefr(c constralnh wed ~ 3 yl o f con for in-react a 10 g o f core for ex-reaef.

~

Qualilaffvtfy acleneu Ped'ye./ teff fer.ffa/finf clarackr k de due fo heaf fr4nt "3% , ~ 10%

fondle: Scafe-up from experinenft ittiny tiy de drop apyroacict 1.e. //9 d r o d 3-ynoaict 5% , ~ 13 %

Idcor : Ideal /%ed,4imtfin ca.re to quantif i effect of sfet na . Exfencfou 44 y Cortedint* support ffar)l Rankoff and

-He genenti concept , but tHlf f sphG i dee dityd.( f-f , .L-b, efc) . Jo-foo lej fla de c .- Re .rotuHon

~

Reaeily confalnd LoM cowponenft . Re f/nenenft a te nesled . RnabpH newr ku wntU fadot.d by L ndi cafel ex pert e en fi.

Inyetca/icnt B conu appear fecciHe h demonstrafe fluf exted o f p re minfa re s *; severeAje 6mifed for Lv2 condiffon 4 ef Inferet6.

l

g '

Theok m,s 9.

.f I

CO .ucL U S 10 US fo e enerteHc laihues need:

o 3,o oo AJ => .107. of c.oe ab TOX efNcienu

- ~3m *I Aid "pakae.pn "

cumnt rows: 'a- Lue. ade uet4.'

" [o ncidersb6e. oyNsaV M ft.sf s'I UW perdWe k dou haf ihl Le nege1 conte m ,'

o f. 9 twe e

p res edel o k A tR s Lbcom,#ee. m C. l AII- 9 A cc.a cke Vlb COMMENTS ON IDCOR REPORTS 14.1A AND 14.1B i

KEY PHENOMEN0 LOGICAL MODELS FOR ASSESSING STEAM GENERATION RATES MARSHALL BERMAN SANDIA NATIONAL LABORATORIES

\ '

l JANUARY 11, 1984 f

S l

l l

_ = -

2.

@ E Fma %

OUTLINF OF PRESENTATION I. REACTOR SAFETY ISSUES II. KEY IDCOR FCI MODELS III. ' LIMITS TO C0 ARSE MIXING BASED ON QUASI-STEADY BOILING CONSIDERATIONS OR HYDRODYNAMICS A. DIFFERENT MODELS AND THEIR PREDICTIONS OF MAXIMUM MELT THAT CAN MIX B. TRANSIENT CONSIDERATIONS FOR COARSE FRAGMENTATION C.

SUMMARY

AND CONCLUSIONS IV. LIMITS TO MIXING BASED ON ENERGY CONSIDERATIONS A. THREE IDCOR MODELS B. THREE. ALTERNATIVE MODELS C. ENERGY LIMITS - ADDITIONAL EXPERIMENTAL DATA D.

SUMMARY

AND CONCLUSIONS V. NECESSARY TRIGGER ENERGIES

' VI. TRIGGER SOURCE ENERGY vs TRIGGER ENERGY

,VII. CONCLUSIONS

,,,c .-- - - - - - - -, - , . - , , - -- - , , -- .- - ,

S-

}ermsK REACTOR SAFETY ISSUES MOLTEN FUEL-COOLANT INTERACTIONS (FCIS) CAN STRONGLY INFLUENCE:

1. DIRECT OR INDIRECT FAILURE OF PRIMARY SYSTEM OR CONTAINMENT BY IN- OR EX-VESSEL STEAM EXPLOSIONS.

2. STEAM AND HYDROGEN GENERATION RATES.

3. ~ FUEL FRAGMENTATION, DEBRIS DISPERSAL AND DEBRIS BED C00 LABILITY.

4. ACCIDENT PROGRESS AND SOURCE TERMS.

5. ACCIDENT TERMINATION.

b e

Sumn '

FRAGMENTATION AND MIXING THE KEY QUESTION FOR FCis IS:

TO WHAT DEGREE, AND AT WHAT RATE, DOES THE -

MOLTEN CORIUM FRAGMENT WHEN IT CONTACTS WATER?

POSSIBLE ANSWERS

1. NO FRAGMENTATION: STEAM AND HYDROGEN GENERATION RALES ARE SLOW AND BENIGN. EARLY CONTAINMENT FAILURE IS UNLIKELY OR IMPOSSIBLE. -

. i

2. C0 ARSE FRAGMENTATION: (AKA " PREMIXING"): TIME SCALE OF ORDER OF SECONDS: SIGNIFICANT

- INCREASE IN STEAM AND HYDR 0 GEN GENERATION RATES AND MAGNITUDES.

INCREASED POSSIBILITY OF 6 , Y ,

AND c- FAILURE MODES.

3. FINE FRAGMENTATION: (AKA " STEAM EXPLOSION"): TIME SCALE OF ORDER OF MILLISECONDS: STEAM l

AND HYDROGEN GENERATION RATES AND MAGNITUDES CAN BE VERY HIGH. SHOCK WAVES AND MISSILES MAY BE GENERATED.

INCREASED POSSIBILITY OF ALL FAILURE

,' MODES, INCLUDING DIRECT FAILURE (a-MODE).

/

~

hermAW IDCOR POSITION THERE WILL BE NO SIGNIFICANT FRAGMENTATION, EITHER IN- OR EX-VESSEL.

THE MAXIMUM AMOUNT OF FUEL THAT CAN FRAGMENT IN-VESSEL IS

~

ABOUT 100 KG.

THIS LIMIT TO MIXING IS BASED ON HYDRODYNAMIC STABILITY ARGUMENTS (I.E., ON 1D COUNTERCURRENT FLOW AND CRITICAL HEAT FLUX) AND ON CONSIDERATIONS OF THE ENERGY REQUIRED TO FRAGMENT THE MELT.

lHE MAXIMUM AMOUNT OF FUEL THAT CAN FRAGMENT EX-VESSEL IS ABOUT 7 KG, BASED ON GE0 METRIC ARGUMENTS.

i l

l l

i

b' 3 f rm dPt KEY IDCOR FCI MODE!_S I '

g,

\

i

1. LIMITE TO MIXING BASO) ON STEADY-STATE,1D, CHF FLOODING ,

MODE!. :

Rp= MINIMUM PARTICLE RADIUS

s c 3M p [oE'Np 4 -T[),+H(Tp - Tc )]

= Eo. 3.7 IN 14.1A

(#F . 0/^..# CHF,SUB . Ay

2. LIMITS TO MIXING BASED ON ENERGY REQUIREMENTS: ,

pV 2 p

Eg (ONE STEP) = 3/8 C p Eo. 3.14 IN 14.1A i ,

y t

a

, I s' I ~, '

4 1

, _ , - - , _ - - . _ _ , , . . -) , y _,,

. . . . . ~

3f rnd a d - 7 LIMITS TO COARSE MIXING BASED ON OUASI-STEADY BOTIING CONSIDERATIONS OR HYDRODYNAMICS MODEL BASED ON WATER- MAXIMUM MIXED DEPTH FUEL MASS (KG)

DEPENDENCE IN-VESSEL EX-VESSEL HENRY-FAUSKE CHF AND NONE - 100 - 7 (200)

. CCF LIMITS (D - 1 CM) (D - 1 CM)

CORRADINI FLUIDIZATION STRONG - 3000-5000* - 13000-16000' 0F FUEL OR (D 10 CM) (D 10 CM)

WATER THE0 FAN 0US, HYDRODYNAMICS, THROUGH 2500-4000 - 13000 ET AL. JET BREAKUP MIXING (D 10 CM) (D 10 CM)

TIME

.

TYPICAL NUMBERS: ACTUAL MASS MIXED INCREASES WITH PARTICLE DIAME1ER AND AMBIENT PRESSURE: EVALUATED AT T = TSAT, P = 1 BAR, FUEL TEMPERATURE = 2700 K.

S e

t Ph14 h.

--~

BOILING LIMITS TO COARSE MIXING ADDI'TIONAL EXPERIMENTAL DATA e NOTE THAT CORRADINI MODELS AGREE WITH DATA AS WELL AS OR BETTER THAN IDCOR MODEL, FOR SOME INTERMEDIATE- AND

~

SMALL-SCALE TESTS. HENCE, DATA DISCUSSED IN 14.1A CAN BE USED TO SUPPORT A DIFFERENT MODEL. -

e EXPERIMENTS AIMED DIRECTLY AT CONFIRMING IDCOR MODELS BY A. T. CHAMBERLAIN AN7 F. M. PAGE, "AN EXPERIMENTAL EXAMINATION OF THE HENRY-FAUSKE VOIDING HYPOTHESIS,"

JANUARY 1983, DID NOT CONFIRM THAT HYPOTHESIS (ACCORDING TO THE INTERPRETATION OF THE EXPERIMENTERS).

O e

l L

hfFM'Id4 .

TRANSIFNT CONSIDERATIONS FOR COARSE FRAGMENTATION

TIME REQUIRED FOR STEADY FLOW DEVELOPMENT =

H G pg Rp

, 0/A)CHF. sus ' '

4 4 AI CHF, SAT 2apCoc(T p -Tc ) + H (Tp - TC )]

FOR A POSSIBLE REACTOR CASE WITH SATURATED WATER, Tp=

2500 K, P = 1.0 MPA, ap = 0.3, T = 40 us (IDCOR 14.1.A, P. 3-11) e EXAMINF DIFFERENT INITIAL CONDITIONS AND ASSUMPTIONS:

SINCE ap (FITS) - 0.01, FOR REACTOR CASE, WE CAN REASONABLY TAKE a p - 0.01 - 0.03 AND R p - 0.5 - 5 CM:

THEN .

t -

400 - 12000 MS

_ IF Tp = 2200 K, AND P = 0.1 MPA, THEN .

t - 7 - 200 SECONDS

l. TRANSIENT TIMES CAN BE LONG i.

l l

&FMb$

POSSIBLE LONG DURATION OF TRANSIENT CONDITIONS IMPLIES-

1. THE H-F STEADY-STATE MODEL IS INAPPROPRIATE, AND/0R

2. ALL STEADY-STATE MODELS ARE INAPPROPRIATE, AND/0R

3. A LARGE AMOUNT OF MELT CAN MIX IN THE LOWER PLENUM OR REACTOR CAVITY BEFORE MIXING LIMITS BECOME ESTABLISHED.

NOTE: FITS TESTS IMPLY TIME FOR MELT TO REACH VESSEL BOTTOM IS 200 - 400 MS PER METER DEPTH.

/ .

9

3 8 ($1d %

TRANSIENT CONSIDERATIONS CONCLUSIONS i

, STATEMENT ON P. 3-12 THAT "THE REACTOR ACCIDENT CASE WOULD DEVELOP FASTER THAN THE CORE MATERIAL COULD BE SUPPLIED TO THE WATER."

IS NQI TRUE, IN GENERAL, BECAUSE

1. TRANSIENT TIMES CAN BE LONG. .

2. SUDDEN CRUST FAILURES, LARGE-SCALE POURS AND MULTIPLE EXPLOSIONS ARE POSSIBLE.

i -

l L

b ( t'M$

BOILING CONSIDERATIONS

SUMMARY

AND CONCLUSIONS o THERE IS DISAGREEMENT BETWEEN SEVERAL CURRENT QUASI-STEADY MODELS OVER THE AMOUNT OF MELT THAT CAN MIX, AND THE RESULTING PARTICLE SIZES.

e THE IDCOR MODEL IS ONE-DIMENSIONAL, AND IS BASED ON STEADY-STATE CORRELATIONS. THERE IS STRONG EVIDENCE THAT

~

THE MIXING PROCESS IS VERY THREE-DIMENSIONAL AND TRANSIENT. ,

e THE DURATION OF THE TRANSIENT FHASE CAN BE VERY LONG, BASED ON THE IDCOR MODEL ITSELF, BUT WITH DIFFERENT INPUT ASSUMPTIONS.

e CONCLUSIONS: HQ LIMITS-TO-MIXING MODELS BASED ON BOILING CONSIDERATIONS CAN YET BE QUANTITATIVELY ESTABLISHED WITHIN CURRENT DATA BASE.

-e

,-- - - 9

i. . .

\3, b erma n LIMITS TO MIXING BASED ON ENERGY CONSIDERATIONS

~

THE ENERGY REQUIRED TO MIX HOT AND COLD MATERIALS, INITIALLY IN A SEPARATED STATE, IS MODELLED IN 3 WAYS IN REPORT 14.1A:

1. FRICTIONAL DISSIPATION, 1-STEP MIXING:

. E i=fC p Eo. 3.14

2. FRICTIONAL DISSIPATION, MULTIPLE STEP, " MINIMUM" VALUE:

2) 1/3

/v/3%2/

1 Rp Y

Eo. 3.17 Egyg = 1.81 Cp py p T gf1- 2/3j LN Rp

3. KINETIC ENERGY REQUIRED TO MOVE MATERIALS TO FINAL CONFIGURATION:

5

~

K.E.=hpv/3 g 2

Eo. 3.19 Tg FOR CD - 20. E IS ABOUT 2 ORDERS OF MAGNITUDE. GREATER THAN

~ K.E. (14.1A, P. 3-20).

l l

i s

l l

I

14 B erman ENERGY CONSIDERATIONS (CONTINUED)

ALTERNATIVE MODCL - 1

1. ASSUME IWQ MIXING STAGES, FIRST C0 ARSE, THEN FINE:

e C0 ARSE: Tg = 1 S, Rp = 1 CM, V = 3.1 M3 (M p - 22.6 TONNES),

. Cp=1 E

1C

= 0.4 MJ e FINE: Tg = 1 MS, Rp = 250 uM, GET 26 J PER PARTICLE, OR FOR 740,000 PARTICLES, E ip = 19.2 MJ

-ETOT = 20 MJ (0.07% OF TOTAL THERMAL E, ASSUMING 1.2 MJ/KG COMPARE IDCOR COMPARABLE SINGLE-STAGE CALCULATION IN TABLE 5.9 WHERE Tg = 1 MS, Cp = 20, n = 10%, Rp = 250 uM,

~

v = 3.1 M3 (Mp = 22.6 TONNES)

GET 8

MJ El = 2.63 x 10 2.6 7 DIFFER BY RATIO 0F 20

- 10 :1 FOR A PARTICULAR MODEL, THE REQUIRED ENERGY IS A STRONG

~

FUNCTION OF THE ASSUMFD PROCESS (NUMBER OF STAGES) AND THE ASSUMFD VALUES OF T g, R p, AND Cp .

, , - - - - - , , - - , -m ,m- , , , . - - ,- , -- , -- ,-

perm AN 5 FNERGY CONSIDERATIONS (CONTINUED)

ALTERNATIVE MODEL - 2

2. D. S0VARER AND M. C. LEVERETT, " STEAM EXPLOSIONS IN PERSPECTIVE," AUGUST 1983. USING SPHERICAL GEOMETRY, AND EQUAL VOLUME MIXING, GET RATIOS WHICH DIFFER BY:

E1 (H-F) - 500:1, FOR 1-STEP MIXING Ey(S0)

OR, FOR PROGRESSIVE MIXING, 3

[~500:1,FORV i = 1.4 M , Rp = 10 CM 3

~ 100:1, FOR V = 14 M , Rp = 2.4 CM 1 -

e a

~ - ,-,,um,,,,, _ _ . , ,, , ,_ , _ _ __ _

R TrVid h, .

ENERGY CONSIDERATIONS (CONTINUED)

~ ~~~

ALTERNATIVE MODEL - 3

3. N. E. BUTTERY, " FRAGMENTATION AND MIXING ENERGY REQUIREMENTS FOR STEAM EXPLOSIONS," 1981.

, HYDRODYNAMIC FRICTIONAL DISSIPATION ENERGY =

E gg=fC p h (ygf, WHERE R = PARENT DROP RADIUS (NOT RESULTING FRAGMENT RADIUS).

FOR 100 TONNES OF MOLTEN FUEL (14 M 3 ), MIXING TO MM SIZES IN - 1 S REQUIRES < 10-6 0F THE THERMAL ENERGY AVAILABLE.

"FOR C0 ARSE PREMIXING ON THE 1 S TIMESCALE, THERE IS SUFFICIENT ENERGY AVAILABLE FROM THE POTENTIAL ENERGY ASSOCIATED WITH THE POUR TO FRAGMENT THE FUEL."

4

. s 17.

T$ erms ENERGY CONSIDERATIONS ADDITIONat EXDERIMFNTAL DATA e EXTENSIVE ARGUMENTS ARE PRESENTED IN 14.1A, SECTION 4, INDICATING THAT " EXTERNAL" TRIGGER ENERGIES WERE SUFFICIENT TO MIX THE MELT, AND RESULT IN AN EXPLOSION.

HOWEVER, e MANY EXPLOSIONS OCCURRED SPONTANEOUSLY, WITH NO EXTERNAL TRIGGER.

e EXPLOSIONS HAVE OCCURRED IN UNMIXED, SEPARATED CONDITIONS: E.G., FITS 5A, CM7, ACM1, BROOKHAVEN TESTS, LNG TESTS, APPLEBY-FRODINGHAM ACCIDENT.

O e

Th ermy 19 ENERGY CONSIDERATIONS

SUMMARY

AND CONCLUSIONS e DIFFERENT EVALUATIONS YIELD A MODELLING RANGE SPANNING MAR 1 ORDERS OF MAGNITUDE.

e WITHIN EACH MODEL, ASSUMPTIONS CONCERNING VALUES OF DRAG COEFFICIENT, MIXING TIME AND MIXING GEOMETRY, CAN ALTER ENERGY REQUIREMENTS BY MAH1 ORDERS OF MAGNITUDE.

e CONCLUSIONS: RQ LIMITS TO MIXING BASED ON ENERGY REQUIREMENTS CAN BE CURRENTLY ESTABLISHED.

e. ,

e O

L

s krman 9.

NECESSADY TRIGGER ENERGIES

+ SINCE THE IDCOR MODELS ARE INTERPRETED TO:.XCLUDE LARGE-SCALE C0 ARSE FRAGMENTATION, AND SUCH FRAGMENTATION IS ASSUMED TO BE NECESSARY, THEY EOUATE TRIGGER FNFRGY TO MIXING ENERGY.

+ IF THE MODELS ARE WRONG, THEN NECESSARY TRIGGERS MIGHT BE SMALLER BY MANY ORDERS OF MAGNITUDE.

+ IF PROPAGATION IS POSSIBLE (1.E., THE EXPLOSION OF A SINGLE PARTICLE CAN INDUCE EXPLOSIONS IN MORE THAN ONE NElGHBORING PARTICLE), THEN TRIGGER ENERGY IS TOTALLY UNRFLATED TO C0 ARSE MIXING ENERGY.

G 4 O

5 e

6 a

O 4

~

b(lWlbh TRIGGER SOURCE ENERGY AS A MEASURE OF TRIGGER STRENGTH NELSON & DUDA SMALL DROP TEST NO. 11-74-1 TRIGGER SOURCE ENERGY (BRIDGEWIRE) = 14 J ENERGY TO FRAGMENT DROP (PROGRESSIVE FORMULA) = 7.3 x 10-3 J IDCOR: "0NE MAJOR RESULT FROM THIS ANALYSIS IS THAT THE TRIGGER WORK WAS APPROXIMATELY THREE ORDERS OF MAGNITUDE GREATER THAN. THE ENERGY REQUIRED TO MIX ALL THE MELT."

PRESSURE PULSE AT DROP: AP = 0.64 MPA ST = 18 uS ENERGY RECEIVED BY DROP = 2.9 x 10-5 J I.E., 4 x 10-3 0F THE NECESSARY FRAGMENTATION ENERGY AND 2 x 10-6 0F THE TRIGGER SOURCE ENERGY.

e TRIGGER SOURCE ENERGY IS NOT A GOOD MEASURE OF STRENGTH

P Dn A N CONCLUSIONS e IDCOR FCI MODELS ARE EXTREMELY IMPORTANT IN ESTIMATING THE RISK OF SEVERE ACCIDENTS.

e THESE MODELS EXCLUDE ANY SIGNIFICANT FRAGMENTATION OF THE MELT DUE TO AN FCI.

. . IN THE ABSENCE OF "SIGNIFICANT" FRAGMENTATION, THE RISKS DUE TO FCI PHENOMENA (STEAM AND HYDR 0 GEN GENERATION RATES AND MAGNITUDES, DEBRIS BED CHARACTERISTICS AND COOLABILITY) ARE SMALL OR ABSENT.

e CURRENT DATA DO HDI CONCLUSIVELY ESTABLISH THE VALIDITY OF THESE MODELS.

p

y I

I

) -

l IN-VESSEL COOLABILITY AND NON-EXPLOSIVE I

' STEAM GENERATION .

PRESENTATION TO l ACRS SUBC0fMITTEE ON CLASS IX ACCIDENTS '

JANUARY IL 19811

t i

l i

R. W. WRIGHT FUEL SYSTEMS RESEARCH BRANCH i 0FFICE OF NUCLEAR REGULATORY RESEARCH i .

e

l 2. ;

W r skU_

i f

IN-VESSEL C00 LABILITY AN!) NON-EXPLOSIVE STEAM GENERATION 6

I

TECHNICAL CONCERNS: ,i

! i e DRY-0UT C00 LABILITY LIMIT UNDER STAGNANT,REFLOOD

[

l e REQUIREMENTS FOR DRY-00T C00 LABILITY UNDER FORCED FLOW (TMI) 4 !

IDCOR: l e USE FLAT PLATE CRITICAL HEAT FLUX LIMIT-(CHF) AT TOP 0F BED

! e CONCLUDE DEBRIS BED APPEARS C00LABLE (STAGNANT) BEFORE ,

CORE-SLUMP INTO LOWER VESSEL f, e REFLOOD STEAM GENERATION WOULD NOT CHALLENGE RELIEF CAPACITY l'

0F PRIMARY SYSTEM l

4

~

~'

(AI t

AN }.

10 7

'I ' ' '

'"I ' ~

WATER

~

',, BED HEIGHT = 200 (mm) 't

. POROSITY = 0.4 -

-PRESSURE =1.0 (bars) .

e .

G G .

D y -

-[F F (Ipcot)- - - - - - - -p '

~

10" ~

e n

E

- O A +e Barleon, et al. :

g .

A -

O Hofmann 4 A Gabor, et al.

X

l3 - A A Stevens, et al. -

8

u. E E Squarer, et al. ~

g ~

O somerton, et al.

g ga A + ru ta. . ,

c: -

0 10 5 .

g ~

- - - - - H. < , z o nt.) P1.te

- C H F (IbcoR -

Dekras C.,l.kility Cede,;. ) .

Lipinski (19 8 3)

(NRC 3p.s ored )

4 ' ' ' ' ' ' ' ' ' '

10 0 10 1

O.2 10 PART!C!.E DIAMETER (mm) ii C,o\dilit Lim.h F 3ar. I. Meo nced bebri3 3

1 a-A ti,4. tea Crib iu 6

_.7____

-i M)rs,ld

~

[i

9. .. ,

kI e

IN-VESSEL COOLABILITY AND NON-EXPLOSIVE STEAM GENERATION !!

NRC

!I e DEBRIS SIZE IMPORTANT - USE DATA AND MODELS (LIPINSKI) !

e STRATIFICATION, INLET FLOW INPORTANT - NEED VAllDATION FOR LIPINSKI MODELS l e UNCERTAINTIES IN CHARACTERISTICS OF DEBRIS TO BE C00 LED EXCEED UNCERTAINTIES IN !,

DEBRIS C00 LABILITY MODELS DATA FROM PBF AND ACRR IN-PILE EXPERIMENTS j! ,

FUEL LIQUIFICATION EFFECTS i DEBRIS TO BE COOLED IS THAT RESULTING FROM REFLOOD QUENCHING, INCLUDING e jl FRAGMENTATION, BED DISRUPTION, STRATIFICATION, ETC.

e IDCOR CONCLUSION THAT FULLY-RUBBLIZED CORE IS C00LABLE IN PLACE WITHOUT FLOW ,

ONLY VALID FOR LARGE (71 MM), UNSTRATIFIED PARTICULATE FROM QUENCHING SOLID CORE I; MATERIAL !

TMI REQUIRED PUMP FLOW f, e STEAM GENERATION BY TOP QUENCHING 0F PARTICULATE DEBRIS (CDUNTER-CURRENT FLOW) [

FITS DRY-00T MODELING (LIPINSKI), 00-CURRENT BOTTOM FLOODING GIVES ABOUT SX 'l FASTER QUENCHING ;-

~ '

(A)es W Es

(

i IN-VESSEL COOLABILITY AND NON-EXPLOSIVE STEAM GENERATION l

PATHS TO RESOLUTION: l e IT IS HECESSARY TO USE DATA AND MODELS TilAT INCLUDE PARTICLE SIZE EFFECTS, NOT JUST SIZE-INDEPENDENT CilF. IT IS IMPORTANT T0 llAVE A VALIDATED MODEL FOR j IMPORTANT STRATIFICATION AND INLET FLOW EFFECTS l e DISCUSSIONS WITil IDCOR ON TilEIR USE OF CHF LIMIT WillCH REGULATES PARTICLE-SIZE l AND STRATIFICATION EFFECTS DEBRIS CHARACTERIZATION FROM ACRR AND PBF IN-PILE EXPERIMENTS, PARTICULARLY THE e l<

ACRR DEBRIS QUENCH (DQ) EXPERIMENTS I e NEED VALIDATION OF LIPINSKI MODELING OF STRATIFICATION AND INLET-FLOW EFFECTS ON DRY-OllT C00 LABILITY LIMIT FOR LWR ACCIDENT CONDITIONS. THESE DATA WOULD HAVE BEEN OBTAINED IN THE PLANNED ACRR DCC-3 EXPERIMENT e DATA AND ANALYSIS FROM CONTINUING EXPERIMENTS AT BNL AND ANL ON QUENCllING ARRAYS j OF HOT SOLID PARTICULATE ,

i I

\

Prennh fo % ACRS Class - 9 .

i %b Com', Hee on un w%rg Il,1924 PRESSURIZED VESSEL FAILURE

(~ DEBRIS RELOCATION D Awa T%wers , SANDI A PHENOMENA

DEBRIS DISPERSAL FROM CAVITY DEBRIS QUENCHED NO MEL T/ CONCRETE A TTACK .

l SMALL EX-VESSEL SOURCES

DIRECT CONTAINMENT HEATING

SENSIBLE & LA TENT HEA T CHEMICAL. HEA T

HYOROGEN RECOMBINA TION

'.

AEROSOL GENERATION

-- - -- ^ --- - -- ^

^--.

.c

? ewers 2.

I STATUS VESSEL FAILURE IS PRESSURIZED OF ACCIDENT A VERY NEW AREA ANALYSIS ONLY A FEW SCOPING STUDIES HAVE BEEN DONE.

1 EXPERIMENTS ANL 1

EPRI-SPONSORED WORK AT 2 Kg. MELTS SNL

NRC-SPONSORED WORK AT 10 Kg. MELTS 80 Kg. MELTS PLANNED l

l

3 755 .

AGREEMENTS i

PRESSURIZED FAILURE IS POSSIBLE POSSIBLE

'

DEBRIS DISPERSAL.IS

- SUBJECT TO GEOMETRY

EXPTS. YIELD SIMILAR RESULTS

- DIFFERENCES DUE TO SCALE

MANY AREAS OF RESEARCH THAT ARE OF MUTUAL INTEREST.

l

4 Powers ,

i I

J AREAS OF MUTUAL INTEREST

!

DIRECT CONTAINMENT HEATING l ,

DISPERSING DEBRIS REACTIONS j

HYDROGEN RECOMBINATION

IN-VESSEL PROCESSES THAT WILL DEPRESSURIZE VESSEL

AMOUNT OF MELT INVOLVED l

l

Zr & STEEL IN THE DEBRIS l .

F 5

Po wers RESOLUTION EFFORTS f

-

INFORMATION EXCHANGE UNDERWAY

- NRC IDCOR

- EPRI

- SNL I - FAUSKE & ASSOC.

I

NRC CONSIDERING PRESSURIZED VESSEL FAILURE IN ITS

- ACCIDENT SOURCE TERM

- REASSESSMENT CONTAINMENT LOADS WORKING

\ GROUP

FAUSKE & ASSOC. DOING SOME REANALYSIS

c.

~,. rf ., '+ .

r 1

,. 1.

NON-EXPLOSIVE STEAM GENERATION AND COMBUSTIBLE GAS FORMATION IN CONTAINMENT i t 4

PRESENTED BY i i 7

4 W. T. PRATT BROOKHAVEN NATIONAL LABORATORY UPTON, NEW YORK 11973 ,

PRESENTED TO ,

ACRS CLASS'9 SUBCOMMITTEE r

f8 i

JANUARY 11, 1984 7

(

l t i

f 2 ,

BROOKHAVEN NATIONAL LABORATORY l} ggl A5500ATED UNIVERSITIES, INC.(Illl 1 n

.p ... . ._ _ _ _ _ _ _ _ _ . _ _ _ . _ _ _ _ . . _ . .

Drak OUTLINE e DESCRIPTION OF THE ISSUE e SUB-ISSUES e STATUS OF UNDERSTANDING e APPROACH TO RESOLUTION i -

i-l i

I l~

j BROOKHAVEN Nail 0NAL LABORATORY l} g)]

! A5500ATED UNIVERSITIES, INC.(Illl s

-..-s.

1 3

.- YrAh !

~

DESCRIPTION-0F THE ISSUE e TO DEFillE IlllTIAL lilTERACTIONS OF HOT CORE ilA-TERIALS (RELEASED FROM TiiE REACTOR VESSEL) WITH ilATER AND C.ONCRETE IN THE REGION .lMMEDIATELY BELO',1 THE VESSEL '

e IMPORTANT FOR DETERMillING THE Til11NG AND MAGulTUDE OF CONTAINMENT LOADING (PRESSURE AND TEMPERATURE)

DURING THESE INTERACTIONS ,

s J

e- 4ly BROOKHAVEN Nail 0tul LABORATORY l} g}l ASSOCIATED UNIVERSITIES,'IfK.(IllI a j 6

_-.a , _ _

y M

SUB-1SSUES e . RELEASE OF CORE MATERIALS FROM THE PRIMARY SYSTEM:

e HIGH PRESSURE VS LOW PRESSURE RELEASE e LOCAL VS GROSS VESSEL FAILURE e COMPOSITIO:10F CORE MATERI ALS RELEASED FROM THE PRIt1ARY SYSTEM:

e HIGH TEMPERATURE (MOLTEN) VS LOWER TEMPERATURE (SLURRY) e RELATIVE QUANTITIES OF ZIRCALLOY, STEEL, AllD FUEL e FRACTION OF METALS OXIDIZED BROOKHAVEN NATIONAL LABORATORY l} g)l A5500ATED UN VERSITIES, INC.(Illl

f hPaki 6.

SUB-ISSUES (CONT.)

e WATER SUPPLY TO CORE DEBRIS:

i e IS WATER PRESENT PRIOR TO VESSEL

. FAI LURE?

e IS WATER RELEASED Oll TOP OF CORE MATERIALS AFTER VESSEL FAILURE?

e IS A CONTINU0US SUPPLY OF WATER AVAILABLE TO THE CORE DEBRIS?

l BROOKHAVEN NATIONAL LABORATORY l} g)l A5500ATED UNIVERSITIES, INC.(llll

r0 6.

SUB-ISSllES (CONT.)

e PLANT SPECIFIC CONSIDERATIONS e COMMllNICATION PATHS BETWEEN AREA BELOW REACTOR VESSEL AND REMAINDER OF CONTAIN-MENT e RELATIONSHIP BETWEEN AREA BELOW REACTOR VESSEL AND REMAINDER OF CONTAINMENT (RELATIVE TO SUMPS, SUPPRESSION POOLS, ICE CHESTS, ETC.)

e STRilCTURES IN AREA BELOW VESSEL AND IMPACT ON CORE / WATER / CONCRETE INTERACTIONS i

~

e EFFECT OF CONCRETE COMPOSITION (LIMESTONE VS BASALTIC) ON C0 PRODUCTION I

BROOKHAVEN NATIONAL LABORATORY l} g)l

! ASSOCIATED UNIVERSITIES, INC.(llll

.. 2 h f[

-__ STATUS OF UNDERSTANDING e IMPORTANCE OF THIS ISSUE DEPENDS ON REACTOR AND CONTAINMENT DESIGN UNDER CONSIDERATION e SIX REACTOR AND CONTAINMENT DESIGNS WILL BE DISCUSSED:

e PWR WITH A LARGE DRY CONTAINMENT e PWR WITH A SUBATMOSPHERIC CONTAINMENT e PWR WITH AN ICE CONDENSER CONTAINMENT e BWR WITH A MARK I CONTAINMENT e BWR WITH A MARK 11 CONTAINMENT e BWR WITH A MARK Ill CONTAINMENT BROOKHAVEN NATIONAL LABORATORY l} g)l A5500ATED UNIVERSITIES, INC.(Illl

p p 3 pL PWR WITH A LARGE DRY CONTAINMENT e SIMPLE AND RELIABLE CALCULATIONS HAVE BEEN MADE TO CLOSELY BRACKET POTENTIAL EX-VESSEL CORE / WATER / CON-CRETE INTERACTIONS e EXTENSIVE ASSESSMENT OF Z/IP FACILITIES e WATER AVAILABILITY LIMITED FOR SEQUENCES WITH LOSS OF ECCI AND CHRS:

e RAPID STEAM AND COMBUSTIBLE GAS GENERATION DURING CORE QUENCH e POTENTIAL FOR DEBRIS DRYOUT (WATER DEPLETION)

AND LONG-TERM STEAM AND NONCONDENSIBLE GAS GENERATION ERC0KHAVEN NATIONAL LABORATORY l} g)l A5500ATED UNIVERSITIES INC.(l(!!

l l

l

9

>pgL-PWR WITH A LARGE DRY CONTAINMENT (CONT.)

WITH RWST WATER IN CONTAINMENT, CAVITY WILL BE

~

e FLOODED:

e RAPID STEAM AND COMBUSTIBLE GAS GENERATION DURING CORE QUENCH e LONG-TERM STEAM GENERATION AND DEBRIS BED C00 LABILITY e MODE OF EX-VESSEL INTERACTIONS JPENDS ON CAVITY f ,

DESIGN AND ACCIDENT SEQUENCE r

l

[

BROOKHAVEN NATIONAL LABORATORY l} g)l A5500ATED UNIVERSITIES, INC.(Illl

- -~ -,m-, , , , .yp, -. y .m -<-

1

. . o l

74 -

5 IA'.t IX?'*S::N A

. .: 4 6#- -

a. a. e. a. . ..

a. a. .. . a. . . a... a. . a. a. a. '.'

r,. '.t..,n:.

e.

s. e. a. a .:. 4.

a. a. a. a. .: ...

a. a. a. a. a. . a.

, FF. ,1 SLr.y .

. . a. .:.:..:..;.a.a.a.aa. .. ; ~ r'. ""W'.'.'.,W.'t,."Lg n"*g2

~

............... wn y.r. .n

......C"

....:... ....Y

... ....... ..v .......

. . . ..... . sys.TU,1. . ._,,

..............s.

.: 1 ..

"I'si:ATCF.5N

m & .

g .: : .

. i ,

.o

. tti '.L, .... . ....a......

..e ...

l _L

',' r: ..a .

l Li \ /R._/p:.:u.T C

\ /u -

-~/,

~**.>"5 e _

, .7 ,sm. N..t-.I w 4 ....

l .= - ,

= .(

. r..

.:. . . .p. .

- a ' *'

. . . . .. .. '., ,' t-a e

s. T

! s=

, l ,' ,

-{

}- "

i

,. ...p .

.s., d;.-. .

- ..l,-,

e )

. .u

. . . . . .n

,, ,, , p .- , ,.

.. .'1 - -L ' tee-" *

. .: . 2 .

. r. .- , ,

j. [*4 Cl o. . .

.. F.T.T V.

)*., *' .}' ' , ' ..-. . , , ' . .

\ *.

,,...a.R

. .L,

.e . .. .w .

. , .:.'r~kr.:.

.1 R..*

.,f

., ... .J nr . Co.. TTY .

UC- .e.r. , 4 .

INS.T?t'iD,

. w, ...

.e iT - '.

9 C.... e.. .,. I F"'.a n,*.".".'.'..

PWR WITH A LARGE DRY CONTAINMENT

~

ll-

'> Q.

PWR WITH A LARGE DRY-CONTAINMENT (CONT.)

e WITHOUT CHRS OPERATION:

e CONTAINMENT BUILDING STEAM INERTED AT VESSEL FAILURE (COMBUSTION NOT POSSIBLE UNTil CHRS RESTORED) e QUENCH TIMES VARY FROM 1-60 MINUTES DEPENDING ON ACCIDENT SEQUENCE e COMBUSTIBLE GAS FORMATION COULD BE SIGNIFICANT DURING OUENCH (DOES NOT IMPACT CONTAINMENT PRESSURE LOADS) e CONTAINMENT TEMPERATURE LOADING FOLLOWS

. SATURATION v

e DIRECTING HEATING 0F CONTAINMENT ATMOSPHERE FOR HIGH PRESSURE FAILURE, IS AREA 0F UNCERTAINTY e BOUNDING CALCULATIONS INDICATE PRESSURES WELL BELOW THE ULTIMATE CAPACITY OF LARGE DRY CON-TAINMENTS (ZION TYPE)

BROOKHAVEN NAiiCMAL 1ASCRATORY]) g)l A5500ATED UNr/ERSITIES, INC.(1 ElI

rek PWR WITH A LARGE DRY CONTAINMENT (CONT.)

e WITH CHRS OPERATING:

e NON-EXPLOSIVE STEAM GENERATION RATES ARE NOT A DIRECT THREAT TO CONTAINMENT INTEGRITY e SNL EXPERIMENTS SUGGEST H2 EQUI. VALENT TO OXIDATION OF 1/3 0F METAL PARTICIPATING IN ENERGETIC FUEL-COOLANT INTERACTIONS SHOULD BE CONSIDERED e COMBUSTIBLE GAS GENERATION DURING CORE QUENCH COULD BE SIGNIFICANT s

e COMBUSTION OF IN-VESSEL AND EX-VESSEL COMBUS-TIBLE GASES CAN RESULT IN SIGNIFICANT CONTAIN-MENT PRESSURE LOADS e BOUNDING COMBUSTION CALCULATIONS INDICATE PRES-SURES BELOW THE ULTIMATE CAPACITY OF LARGE DRY CONTAINMENTS (ZION TYPE)

BROOKHAVEN NATIONAL LABORATORY l} g)l A5500ATED UNIVERSITIES, INC.(Illl

9)g

.h

)3, PWR WITH A SUBATMOSPHERIC CONTAINMENT e PLANTS EXAMINED (SURRY AND MILLSTONE-3) INDICATE RELATIVELY DRY CAVITIES AT VESSEL FAILURE e HENCE, EXTENSIVE CORE / CONCRETE INTERACTIONS WILL OCCUR e WATER INGRESS WILL OCCUR AT LATER TIMES e MODE OF CONTACT UNLIKELY TO PROMOTE RAPID. MIXING

~

NECESSARY FOR RAPID STEAM GENERATION e CONTAINMENT INTEGRITY WILL BE CHALLENGED BY LONG-TERM PRESSURE AND TEMPERATURE BUILD-UP BROOKHAVEN NATIONAL LABORATORY l} g)l ASSOCIATED UNIVERSITIES, INC. (I Lll

e O e

P y s-~- l4.

i

es0waBLE AACM FOR 00MC 4 MamTEhamCE al A(ACTOR

~

T8MALIf1 '

- >e ,g j . ll .l !

f L bN f a lanest h i $ f 4UC lilS[ #ULAR (24seg ll I

c. m '

,cs.. my, 8 1, r. ,*JL,_. s. . k?

-$ TEAM Cdb(AATCAM RI(W-,

A x -

l l SeeELD 1

n J '

$se(LO uu r T.ana -

. _ ; ,; l g.

\ x l . j a. l ip.7..17 lI i

I

,L....

m .s.- _ P _.._ w

. ,i ( ^v m,w , . .

.a , % .

1 li FC i

~

' f ' t,I' jy

(_li_ ----;j. t .---t*

/i *d..?i i u i .

e

. m

/

- Il r - , ....._

~ '

,.'2l11

~~

^j -? -

-33 o o [ 7__.__ -i -gS5u.a;2;a.gana -

l em_ J' -i

5 n a, I h A, 3 .. i ii. l .m "S1t*' -L g ' ';.. .. ' .o. .. .u.* - -

l ! l AC2!s settu T les(Cat awain.200u-- "'- M, t ,

, y)-r - -

,, . ,e esdgg a

e.2f m usan a

Contae wtut s\

Tafson l E/bf r7 4 '#

tL 2e6'

- $7m Tust g,.ign

[TW< G8E ,

.u_. 4 - e . ,, ~

~

"J:S o' j'~

petssumi2tR AELitf faaet b* Coat hAde pues* i=57 mas ./ stuffoe raa'a etred>-#% P tL 2** 4* casan muas71 2

T *=a LEA 0SMstLD -

~

PWR WITH A SU3 ATMOSPHERIC CONTAINMENT l

l

]p pg pl I F.

PWR WITH AN ICE CONDENSER CONTAINMENT e WATER AVAILABILITY IN CAVITY DEPENDS ON CAVITY DESIGN AND ACCIDENT SEQUENCE e IF WATER IS AVAILABLE, RAPID STEAM GENERATION,CAN OCCUR e IF ICE IS AVAILABLE, CONDENSER MAY BE ABLE TO ACCOMMODATE NON-EXPLOSIVE STEAM GENERATION RATES e THREAT TO CONTAINMENT VIA COMBUSTION OR LONG-TERM BUILD-UP 0F NONCONDENSIBLES IN UPPER COMPARTMENT BROOKHAVEN NATIONAL LABORATORY l)l)l A5500ATED UNIVERSITIES, INC.(I ElI

. e Y'b b.

t

.SNIELD 3 //LDING DCMEN v

- e

% e

, Q t

-Q L' r: ? ] ::

b /- I %$7!!L CCNTA/NMENT V!$3EL ~~

E 6 . ft A i

4-Syjggy ' ;'

l

,s-ITS TON PCL AR CRANE '$.,6 \

PCL A R CRANE /

SU/LD/NG h MLL m. \

I

?-

4 I 1 i.

^^ ejPRING L /?

/TCP OFCECK { #TCP CF CRANE RAIL W

,' EL 796 63 *

.' EL 797 /2 ,l EL 791 S

- l 'J <

a ,: J .

/CE L,OA0/NG , ,,, .A

. CN ygg7, ,7,Cy .I AC lh

}. ' z-BP/B E CAANE p

_lyyyy$

! .3 Jti st'.6-n UN/TK '

~1 ,,i TOR Or/CE SEo i,. f~

i

/-:tyt ER SAaRtER g yyg.g9 m t6 EL 730.0 ~( h ,.

,_p_w_ _d 'i.l; j.j

.. i.anii 6

i n< .

3

, .-CEANE g ygg.jg - '." l , _ l, ,

i.Y sn '

' mkm w%

~

$x .:n 1 ..; l " WALL

' s .s IC , -

t t ;

C:NCEN3ER I

I'l t CONTAINMENT ~ > e j

, .f l ff+ipi i , 7E, ,

M i;i ,

1J , .', ;\ j

' e c!N!iA r:R <

pj

i: \ C:NrRct RCo oRivE 1

. . t , . ,

,) ,

pyy:l

E MIS $/LE SMIELo

< > n r. es - ~. j z Q-}s[p i-

/CE CONO

., 1 ..i;i t j

-'l ,, is ' i

\ ...,...m-BA$!.!L 72J.0 *, . - k 1 . . . .t .+ .l

?

Q.

v .,3 1 1. \\. , , ' , ,, .yy,. ,J g

. 3, y. , . .

l. ,i s_--

j "

g L 1w x p,c ;.:

- w, .

J G, , E l_J L.. f.yg,jg

...~ -,

4 lla '

W fg ^C gy (3 *. C

d '.C

s BARR/ER ; ,, I PUMP pr A CCUHUL A TOR ~ ~~. .*! .

l n

f ,. .

t

> I

'A *. i ! -

l GRACE F

,. EL 7:5.C '.

-ra rms: .l .

- f s )f u

-*LCCR EL 7CZ.0 ::y' l

] EL 695.0 ~ : ' 1

,a*\W* - qy hf '.* { a cpr'"

. g,, --

EL 6SJ.0 ,3, .,

y3, , , , ,

4 ,t

l. C .. g'l!!'.

l -'

Ms l

.. *.. .is*

l r 3 *. 6 "

j l - ' , . EL67S73 ! ., lQ ikf l .

1. #

I gT ; ,

lg L 'EE^EY#

.) a

-l '

. . k, )

i r . === .4

> $ 37 EEL' LINER , -

l,;

. , M. . ,. .,

I'

% A NCt*CR SCLT i I. . . rIi 1

I j ff : C.304.e m

,, g ,j, EL 65/.69 f. i .

. r p l.

l $uup d - .,

I I

PWR WITH AN ICE CONDENSER CONTAIN.ENT M

17'

) n a.(.

BWR WITH A MARK I CONTATHMENT e CORE MATERIALS WILL NOT CONTACT SIGNIFICANT QUANTITIES OF WATER IN DRYWELL e WATER COULD REACH CORE MATERIALS BY CS OPERATION OR ECC RESTORATION e MODE.0F CONTACT UNLIKELY TO PROMOTE RAPID MIXING NECESSARY FOR RAPID STEAM GFNERATION e HENCE, CONTAINMENT INTEGRITY UNLIKELY TO BE CHALLENGED BY RAPID STEAM GENERATION e CONTAINMENT INTEGRITY WILL BE CHALLENGED BY LONG-TERM PRESSURE / TEMPERATURE BUILD-UP DURING

- CORE / CONCRETE INTERACTIONS e COMBUSTION OF COMBUSTIBLE GASES PREVENTED BY INERTING BROOKHAVEN NATIONAL LABORATORY l} g)l A5500ATED UNIVER$1 TIES, INC.(Illl

D

_ . ,,2 2 _ . _

j !l t,

g i p'9LOWOUT s

EtowCUT i:

"!IPANELS PANELSf

_l REFUELING BAY i

I ll I

m 1

O Os N i

- - n

h. a

, ,. T'-

p. . -q %

}: $ !t '}

au i J:

l "I RE ACTCR ::

U.. yr e c.r.i. ;i r* *a'T= n."" ;****' l i,.. .

l3 y . .

l

-s is .i

-i L,. . ,.

. .. .. 1 i.

n i. i

.A .

...: . a, p - t

= !

5 "! .4 %

y '

g ..gi

, :. . i. , .

e.

h I I

.: .. . .I r':

Ig

..) [... ,

N ). .: bl.

E, / -:.

......':;'.. .: : L,

~> ; , ,. .

.: .. :: E-_ .

^

-![. '

{;

...r .;'.j '[

4 m# -}}. . , .i [.'

. = ::::: _c -r::..:.:....

_ ==-: -

- .g-;- .;;;.. . 5 ;. !; ;; ..

a .. :: : : - .: .= : : . .. : . ...;:= : :.::::ct -

!E:N

e '! 35::.::e::r..;:c::::r.:.r EEEE:! % ESE425 -

: .: : . '? .: ~".. \-- J -; -,! -'i

= J i ' -/

' ~/ .

.=:n~ :tg 3.;:==::;:=s ::s::.- 3:m=i3 : 3 -

=.: =r ::: ~: :::::=:: . .c.::::. .c r ~. : r ::

=.r-: 22 ~ . :: . :::: : =. : :=n

-:. - :r :: :: .: = . .. . : = . .= '.,: :

.:: =' .,

. r : :: .

r:.r. :: ' ': r' ; . . .;. .

. . ~

k : . :"- : J f BWR WITH A MARK I C0:lTAIriMENT e- - _ _ _ _

~

,. , s ,

Pran 19-BWR WITH A MARK II CONTAINMENT e WETWELL (AND SUPPRESSION POOL) DIRECTLY UNDERNEATH DRYWELL (AND REACTOR VESSEL) t 4

e DIAPHRAGM FLOOR SEPARATE WETWELL AND DRYWELL e WATER CANNOT ACCUMULATE ON DIAPHRAGM FLOOR e HENCE, INITIAL CORE / WATER INTERACTIONS

. WILL BE LIMITED e SUBSEQUENCE ACCIDENT PROGRESSION DEPENDS ON HOW CORE MATERIALS PASS THROUGH DIAPHRAGM l

l-BROOKHAVEN NATIONAL LABORATORY l} g)l A5500ATED UNIVERSITIES, INC.(Illl f

s 1 i Y( $.

"fl.1 ;.~:n . : < :. rh

n.* a: ' , : G: p. . % s'@*. . .' s D:: . ; ,l i.G.- 1 v% ..<

, .1 3!1"T*v (, ,

13.1% ,,

' * :l 4:

r'.< .

.c,.,

...q eN

.:n r .. -

s*.' 1=-

r.' . ..

G

s-

,, .: 'm.d l. '.)

.e :..]. j 4 e' s *.". ]

.g .-4{ .i.e am

~e .;-

) I

c. ee I,'. .' S. i'. ' , ..;

?:i '.',/=e'i -

w.

-*g*.

.p.,}_

tt:,: p

}

t f" . .

.a

- I [: .'

. . p y .1-

'S ;

rs-

.j N.

? l /q p -'

?sl 1 .4

'.' '? . ,

6.- '3 34 f a.-; ,.' b f4 .

l 8.*

i.4! .-e 1; }l) . .\s .

. *y .s ..i. sL.. ... .

e. n

jN \ ',, L. . .

.r..

L'.

i.

.s

\b- *. ..g

,r /. 0/ 3e L.1 i \i i tl, ).i n:.

. . :l i -: .

i .s.i ld.;

. ..e r, - ~

p,e r.J.

g

,, \

Sj

. e.i

.,7, .

.s,, .r

..a.n. .

p V.1 d*

_f N. $.'.. I l (.*' .g 7...

k*d

r. ~ f.} a !

[ *ET' IL'-L.,.; 7 $.1 L.. .:

n 9..

yL d r.g <

i ....

.g:

- - .4 l,

l.

I-,9

. )'.'-Y'.e;,. ' . ...

i. .;. - e,.

.4 :.] r, l'

6. ,

e ,a a3 p:..l

)

s,

.J- La

.1 f r 3 fj ya r.1

1. H p..

P.-

s G.. .'A C..

+. . .. ,. .e. v..s. .. .,,.o..~

,.s.,.c.

. :.a ..a..r.: ..._... .,~, . ; . a. . . . , . .,

. . . ..<> ~ .y::s :

A .. . ~ . :. 2. . . ?. . w. . . < -

.2..,

u l l BWR WITH A MARK II CONTAINMENT Sta.m

3 ,

2. I .

'f Y YR' BWR WITH A MARK 11 CONTAINMENT (CONT.)

e IF MAJORITY OF CORE MATERIALS REMAIN ON DIAPHRAGM FLOOR, EXTENSIVE CORE / CONCRETE INTERACTIONS OCCUR A

e CONTAINMENT INTEGRITY WILL BE CHALLENGED BY LONG-TERM PRESSURE AND TEMPERATURE BUILD-UP e IF CORE MATERIALS RAPIDLY PASS THROUGH DIAPHRAGM FLOOR INSIDE PEDESTAL WALL, RAPID STEAM GENERATION WILL OCCllR e STEAM GENERATION RATES MAY CHALLENGE CONTAINMENT INTEGRITY e COMBUSTION OF COMBUSTIBLE GASES PREVENTED BY INERTING BROOKHAVEN NATIONAL LABORATORY l} l)l ASSOCIATED UNIVERSITIES, INC. (I ElI

prp .2 2.

BWR WITH A MARK III CONTAINMENT e REGION UNDERNEATH REACTOR VESSEL WILL BE DRY FOR

~

TRANSIENT INITIATED SEQUENCES e HENCE, INITIAL CORE / WATER INTERACTIONS WILL BE LIMITED e EXTENSIVE CORE / CONCRETE INTERACTIONS WILL OCCUR (SIGNIFICANT QUANTITIES OF COMBUSTIBLE GASES WILL BE PRODUCED EX-VESSEL) . .

. e THREAT TO CONTAINMENT VIA COMBUSTION OR LONG-TERM BUILD-UP 0F NON-CONDENSIBLES r

l I

BROOKHAVEN NATIONAL LABORATORY l} lj l A5500ATED UNIVERSITIES, INC.(Illl i

i_

Ora h 23.

BWR WITH A MARK III CONTAINMENT (Ct'NT.)

e FOR LOCA SEQUENCES, WATER WILL BE IN REGION IINDERNEATH VESSEL PRIOR TO VESSEL FAILURE e SUBC00 LED SUPPRESSION P0OL CAN ACCOMMODATE NON-EXPLOSIVE STEAM GENERATION RATES e EX-VESSEL COMBUSTIBLE GAS GENERATION LIMITED TO INITIAL CORE / WATER / CONCRETE INTERACTIONS PRIOR TO CORE QUENCH e CONTAINMENT INTEGRITY CHALLENGED BY COMBUSTION

~

IN WETWELL f

i BROOKHAVEN NATIONAL LABORATORY l} g)l ASSOCIATED UNIVERSITIES, INC.(Illl l

l

D 2,4, 1 _

i i i i,

,f ;

/ l

_.- .-- -=.. ;..:..-. --

,..-~ ..

/,

-i .

//

. ./

o i

'/ -

~~ ' /

1 1 .\

-' . i,_

i

_. - - --- 3

ll .

2. . . .

..! Lh. A.- '_ I

,/ -N. i 1

. . .. y . . .y.. v. . _.:.y i ;

i I I I 7

,g

-=,;

e

! ! !. I j'!

i t . !-

,l> t !

f...

i..

]' N. ..;i .

Ii i a.- -

i'

< = - .

(* f

m. ie -

.* , e .-s ::-

l .

m.:i :

i 1

i :_

m.=. ,

l.* -

j l 1 i : -

l

~t -

i s <. v. ..-.j : : i i ,

,y , . .

I.

l I SEDESTAL 1 ,1I

%,- s ,s si .

i ,25- ==.

l ._ - _ _ ,

c= . ..

==

t i A ,' ' -

. , .' p f, E L. 93 -0 l

BWR WITH A MARK III CONTAINMENT

, , --,,r -,,-,yy,..

Oyg J[,

APPROACH TO RESOLUTION e CONTAINMENT LOAD WORKING GROUP FORMED UNDER AUSPICES OF ASTP0 AND SARP MANAGEMENT GROUP e GROUP IS SYSTEMATICALLY ADDRESSING A NUMBER OF STANDARD PROBLEMS APPLICABLE TO EACH CON-TAINMENT DESIGN UNDER CONSIDERATION e THIS GROUP WILL:

e ESTABLISH STANDARD METHODOLOGY WHERE P0SSIBLE e PROVIDE A BROAD CONSENSUS VIEW 0F AREAS WHERE CALCULATIONS CAN BE PERFORMED WITH CONFIDENCE e IDENTIFY WHERE UNCERTAINTIES EXIST BROCKHAVEN NATIONAL LABORATORY l} gj j A5500ATED UNIVERSITIES, INC.(Illl

2rdi[ 25-APPROACH TO RESOLUTION (CONT.)

e STATUS OF CONTAINMENT LOAD WORKING GROUP EFFORT:

o CALCULATIONS OF A PWR LARGE DRY CONTAINMENT ARE NEARING COMPLETION e CALCULATIONS FOR BWR MARK I AND II CONTAIN-MENTS ARE UNDER WAY

-e BWR MARK III, PWR ICE CONDENSER, AND SUBAT-MOSPHERIC CONTAINMENTS HAVE TO BE ADDRESSED e EFFORT WILL BE COMPLETED IN FY 84 e CONTINUING EXPERIMENTAL PROGRAM AT SNL WILL PROVIDE DATA BASE FOR FURTHER ANALYSES AND MODEL DEVELOPMENT i

~

e MELCOR CODE UNDER DEVELOPMENT AT SNL:

e LONGER TERM CONFIRMATORY EFFORT l

BROOKHAVEN NATIONAL LABORATORY l} g)l A5500ATED UNIVERSITIES, INC.(Illl

.t. .

SUMMARY

VIEWS ON IDCOR REPORTS 12.2 AND 12.3 i

.UlYDR0 GEN DISTRIBUTION AND COMBUSTION IN REACTOR CONTAINMENT BUIBINESL Tills PRESENTATION REPRESENTS THE CONSENSUS OF THE FOLLOWING CONTRIBUTORS:

I WALTER BUTLER -

NRR/NRC ALLAN CAMP -

SANDIA NATIONAL LAB.

HN LA -

RES/NRC ROBERT PALLA -

NRR/NRC l

i ROGER STREHLOW - CONSULTANT - UNIVERSITY OF ILLINOIS ET ALS.

PRESENTATIONFORACRSSUBCOMMITTEEONCLASS-9 ACCIDENTS JANUARY 11, 1984

1

' ~

L a rkiks TECHNICAL ISSUE: DIRECT INITIATION OF DETONATIONS STATEMENT OF ISSUE: DETONATIONS CAN RESULT IN THE LOSS OF SAFETY RELATED EQUIPMENT AND/OR POTENTIALLY FAIL CONTAINMENTS.

STAFF GENERALLY AGREES WITH THE IDCOR POSITION THAT DIRECT INITIATION OF DETONATIONS IS UNLIKELY.

SOMEASSESSMENTSHOULDBEDONEFOR00ARTERNARYMIXTURES OF H2/C0/ AIR / STEAM WITH REGARD TO DIRECT INITIATION REQUIREMENTS.

STAFF NEEDS TO REVIEW THE ASSESSMENT OF IGNITION SOURCES IN VARIOUS PLANT TYPES THAT HAS BEEN DONE BY IDCOR.

o i

_ -, --- . y _ __,

-v-

~

La r kin s 3,

TECHNICAL ISSUES: DIFFUSION FLAMES AND DEFLAGRATIONS STATEMENT OF ISSUES: COMBUSTION 0F H 2 CAN LEAD TO LOSS OF SAFETY RELATED EQUIPMENT OR OF THE ABILITY OF THE CONTAINMENT TO ISOLATE DEFLAGRATIONS: 3 AREAS OF ASREEMENT (EXISTING DATA AND ON-G0ING RESEARCH WILL RESOLVE)

COMPOSITION REQUIRED FOR PROPAGATION COMPLETENESS OF BURNS IGNITION REQUIREMENTS PROPAGATION VELOCITIES EFFECTS OF AEROSOLS AREAS OF DISAGREEMENT EFFECTS ON ENGINEERED SAFETY FEATURES-l -

EFFECTS OF F0GS (FROM CONDENSATION) l -

PROPAGATION BETWEEN COMPARTMENTS.

1 -

EFFECTS-0F TURBULENCE

,- EFFECTS OF CARBON MON 0XIDE l

Nb YbriS Y-DIFFUSION FLAMES:

AREAS OF AGREEMENT AGREE THAT FURTHER WORK IS NEEDED ON:

CONDITIONS LEADING TO STANDING FLAMES COMPLETENESS OF COMBUSTION VERSUS COMPOSITION FLAME HEIGHT, LENGTH, AND TEMPERATURE AREAS OF DISAGREEMENT

.- IMPACT ON VITAL EQUIPMENT STAFF AGREES THAT FURTHER COMPARISONS OF CODES WITH EXPERIMENTS IS NECESSARY COMBUSTION MODELS USED IN MAPP CODE NEED TO BE REVIEWED IN ACCORDANCE WITH RECOMMENDATIONS IN 12.3 REPORT l

l I

l l

. . 1.

. , f.

k Ar ([PI A

~

n-TECHNICAL ISSUE: FLAME ACCELERATION AND DEFLAGRATION-TO-DETONATION TRANSITION (DDT) J ,

STATEMENT OF ISSUE: 2 BURNINGVELOCITIESCANBEACCELERATED H

TO THE POINT.WHERE SIGNIFICANT DYNAMIC LOADS MIGHT BE

~

GENERATED OR THE TRANSITION TU A DETONATION MIGHT OCCUR.

AREAS OF DISAGREEMENT ASSESSMENT OF PLANT UNIQUE CONFIGURATIONS THAT MIGHT BE

'. CONDUCIVE TO FLAME ACCELERATION. y

- EFFECTS OF ESF'S (E.6, ICE BEDS, FANS, ETC.) TO ESTABLISH OR MITIGATE CONDITIONS FOR FLAME ACCELERATION.

'1 ' -

EFFECTS OF GAS COMPOSITION AND IGNITION REQUIREMENTS ON

}4 FLAME ACCELERATION.

EFFECTS OF FLAME ACCELERATION ON THE SURVIVABILITY OF ESF'S AND THE COURSE OF ACCIDENTS.

POTENTIAL FOR HIGH DYNAMIC LOADS AND MISSILE GENERATION.

RESOLUTION: CURRENT FROGRAMS, BOTH COMESTIC' AND FOREIGN, SHOULD BE ABLE TO ADDRESS THESE'lSSUES '

/ 9 f

. 'f .

.4- ,

L aykins '-

TECHNICAL ~ ISSUES: DISTRIBUTION OF COMBUSTIBLE GASES STATEMENT OF ISSUES: THE STRATIFICATION OF COMBUSTIBLE GASES CAN POTENTIALLY LEAD TO CONDITIONS IN A CONTAINMENT WHERE DETONABLE CONCENTRATIONS OF H CAN BE FORMED.

2 AREAS OF DISAGREEMENTS ,

EXTENT OF MIXING DURING BLOWDOWN FOR PWRs AND BWRs.

BUOYANCY DRIVEN JET RELEASES MULTICOMPARTMENTS ,

EFFECTS OF BURNING On MIXING

- EFFECTS OF CONDENSATION, INCLUDING ICE CONDENSERS, ETC.

RESOLUTION: ADDITIONAL EXPERIMENTAL AND ANALYTICAL WORK IS NEEDED ON MIXING WITH 1) MULTICOMPARTMENTS,

~

2) WITH CONDENSING STEAM, AND 3) WITH COMBUSTION.

IMPROVED CODES TO COUPLE GAS DYNAMICS TO COMBUSTION.

l

_5_

l

~

l l

I t

. 1. _

SUMMARY

FOR ACRS OF THE JOINT NRC - IDCOR MEETING ON

! ACCIDENT PilEN0 MEN 0 LOGY AND CONTAINMENT LOADING NOV. 29 - DEC. 1, 1983 REVIEW 0F IDCOR SUBTASK 15.3-CORE - CONCRETE INTERACTIONS PRESENTED FOR IDCOR BY: ROBERT E. HENRY, FAUSKE & ASSOCIATES ,,

i NRC REVIEU OF IDCOR DOCUMENT BY: R. K. COLE, SNL l llRC C0-REVIEWERS:- J. E. GRONAGER, SNL D. A. POWERS, SilL NRC

SUMMARY

EVALUATION nND ACRS SYN 0POSIS PRESENTED BY: B. BU CONTAINMENT SYSTEMS RESEARCH BRANCH, RES

-T I

2 ?

hr5 ON

~

. c 1 .

CORE - CONCRETE INTERACTIONS PilEN0MENA TilAT AFFECT CONTAINMENT LOADING .

e NONCONDENSIBLE GAS GENERATION,11, 00, AND CO '

2 2

! e COMBUSTIBLE GAS GENERATION: H , AND C0 h 2

e STEAM GENERATION (PRIMARILY DURING QUENCH) e HEAT TO CONTAINMENT ATl10 SPHERE e TilERMAL AND / OR PHYSICAL DEGRADATION OF ESF'S e DEGRADATION OF STRUCTURAL COMPONENTS l i

e AEROSOL GENERATION e FISSION-PRODUCT RELEASE AND TRANSPORT l

i 9

, .r..

B(A rso n _ ,

y ,!

j CORE - CONCRETE INTERACTIONS IDCOR APPROACH - PHASES OF DEBRIS BEHAVIOR CONSIDEREDi i

l

. e MOLTEN-JET IMPINGE ENT AT TIME OF VESSEL FAILURE r

! e MOLTEN DEBRIS - POOL ABLATION [

9 e QUENCHING OF HIGH - TEMPERATURE CORE DEBRIS I

s

! e CONCRETE ABLATION BY HOT SOLID DEBRIS

! 1 i

I

i D Rr30rL 4

.c DECOMP - IDCOR CORE - CONCRETE INTERACTION MODEL e ONE - DIMENSIONAL MODEL -

e ASSUMES: THAT DEBRIS BEllAVIOR IS PRIMARILY CONTROLLED BY A I SELF-ADJUSTING TilERMAL BALANCE BASED ON CONSERVATION OF ENERGY e EQUAL DOWNWARD AND RADIAL ABLATION RATES l

e ll0MOGENIOUS MIXTURE OF ALL MOLTEN P00L MATERIALS, INCLUDING BREAK-UP OF CRUST ON P0OL SURFACE e NO CilEMICAL INTERACTIONS WITH GASSES RELEASED FROM CAVITY WALL e HEAT TRANSFER FROM P0OL SURFACE TO OVERLYING WATER CONTROLLED BY CRITICAL HEAT FLUX i

9

) ta r36n.

. p, ,

CORE- CONCRETE INTERACTIONS NRC

SUMMARY

EVALUATION PREPARED BY S, B BURSON IN CONSULTATION WITil K. D. BERGERON, R. K. COLE, AND D. A'. POWERS OF SNL, AND T. PRATT, BNL, .

AREAS OF NRC-IDCOR AGREEMENT e MANY PliEN0MENA INSUFFICIENTLY UNDERSTOOD TO MAKE FINAL DECISIONS:

l e INADEQUATE EXPERIMENTAL DATA BASE FOR' CODE VALIDATION e DETAILS OF THE EXPECTED CORE-CONCRETE BEllAVIOR, AS WELL AS ACCIDENT CONSEQUENCES ARE HIGilLY PLANT SPECIFIC e THE SIPPLIFIED ONE-DIMENSIONAL DECOMP MODELING 0F HEAT TRANSFER WOULD BE ACCEPTABLE IF SUPPORTED BY COMPARIS0N WITH NORE DETAILED ANALYSES

WS OO ,

g, .,

CORE - CONCRETE INTERACTIONS UNRESOLVED ISSUES e IDCOR APPEARS TO BELIEVE THAT Tile FINAL STATE IS A j C00LABLE DEBRIS BED: THIS DEPENDS CRITICALLY ON THE ,

! ASSUMPTION OF A IlYP0TilETICAL QUENCH MECHANISM WHICH IS UNSUPPORTED BY PHYSICAL OBSERVATIONS l ;

e THERE IS NO EVIDENCE THAT PHENOMENA OTHER THAN :

) CONCRETE ABLATION HAVE BEEN CONSIDERED, SUCH AS l THE INTERNAL TEMPERATURE OF THE DEBRIS AND GAS EVOLUTION FROM NON-ABLATED CONCRETE',

i e THERE IS INADEQUATE JUSTIFICATION FOR THE ASSUMPTIONS ON WHICH THE PARTITION OF HEAT AMONG VERTICAL ATTACK, 4 RADIAL ATTACK AND UPWARD LOSS IS BASED, 4

i e DISREGARD OF Tile EFFECTS OF UPWARD llEAT LOSS ON OVERHEAD STRUCTURES HAS NOT BEEN JUSTIFIED, l

l .

l '

3 nr3 6rt ,

7- \

CORE - CONCRETE INTERACTIONS APPROACH TO RESOLUTION e NRC AGREES THAT THERE ARE SIZABLE UNCERTAINTIES ASSOCIATED WITH IDCOR CALCULATIONS, PARTICULARLY IN THE QUENCHING RATE -- MORE STUDY IS INDICATED, ,

e IDCOR SHOULD PROVIDE A QUANTIFICATION OF THE EFFECTS OF THESE UNCERTAINTIES ON FISSION-PRODUCT RELEASE AND ON Tile VARIABLES THAT CONTRIBUTE TO CONTAINENT LOADING, e IDCOR'S APPLICATION OF THIS MODELING TO EVALUATION OF Tile RADIOLOGICAL SOURCE TERM AND OF CONTAINENT CHALLENGE, INCLUDING TEMPERATURE, PRESSURE. AND DIRECT STRUCTURAL ATTACK SHOULD BE

~

PROVIDED FOR VARIOUS CONTAINENT DESIGNS.

e THESE RESULTS WILL BE COMPARED WITH THE RESULTS OF NRC CALCULATIONS TO DETERMINE IF THEY LEAD TO THE SAE OVERALL CONCLUSIONS.

e MANY TECHNICAL DETAILS CAN BE RESOLVED BY EXTENSION AND CLARIFICATION OF Tile IDCOR DOCUENTATION.

___..._m_.__ _ ___

g f

a l

Issue Number : 1.1.1 7

Title : REACTOR C00LANT SYSTEM THERMAL AND HYDRAULIC BEHAVIOR A

Signature of Author:

Prepared by: J. Han v Draf,t i 1 Date! 2/16/84 C. Allison, INEL / W. Camp and J. Rivard, SNL Contractor / Consultants: P. Cybulskis, BCL Review Process (List here contractor who will supply a statement on issue relati toIDCORwork)

C. Allison, INEL te W. Camp & J. Rivard, SNL Initial Branch Chief N. /6!U Asst. Div. Dir.,

/7[M -

~. r Division Dir. 0 65 L/T/ M Tech. Series Div. Director M 2/22/@

Date Sent to ACRS Date Sent to IDCOR Manageme_nt Issues For near-term resolution, both NRC and IDCOR need to attention Management calculate the same accident sequences for the same plants.

is needed (a) to select plants and accident sequences, (b) to select codes, and (c) to authorize the work to do calculations and comparisons.

a llI 7,c 7

- i b

k 4

ISSUE 1.1.1 REACTOR COOLANT SYSTEM THERMAL AND HYDRAULIC BEHAVIOR 1

1. Description of the Issue f

The issue deals with the impact of the reactor coolant system (RCS) thermal-hydraulic behavior upon the important safety issues including the release of radioactive fission products and hydrogen to the containment during a severe accident. '

The thermal-hydraulic behavior of the RCS determines the temperature rise

<~} )

, JJ/ in the core, which strongly affects the timing and release rate of fission products. It also determines the availability of steam / water at the fuel cladding surface as well as the cladding temperature; as a result, the hydrogen generation rate in a degraded core depends strongly on the RCS thermal-hydraulic behavior. The fission products are transported from the core exit through a part of the RCS to the containment by following the pathway that is controlled by the RCS thermal-hydraulic behavior. Along the pathway, sone of the fission products are removed due to deposition or chmeical reaction with the vessel-internal structures, piping walls, and aerosol particles. Such depositions are strongly affected by the wall temperatures and flow rates and are, therefore, critically dependent on the thermal-hydraulic behavior. In conclusion, the RCS thermal-hydraulic behavior plays a major role in detrmining the release rates of radioactive fission products and hydrogen in the contain-ment during a severe accident by means of its control of core, cladding, and RCS temperatures.

i

2

s. 2. Implications of the Issue to Regulatory Questions Major implications of the issue are the thermal-hydraulics induced uncertainties under various accident sequences of (1) the release rate of radioactive fission products and aerosols to the containment, (2) the release rate of hydrogen to the containment, and (3) the resulting containment load and equipment survivability. ,

3. Subissues Fuel liquefaction and relocation before and up to fuel melting.

Effects of multi-dimensional flow patterns in the vessel upon fission product retentien and hydrogen generation.

Core heat generation during fission product release and fuel relocation.

~.

Radiation heat transfer and core debris temperature.

Surface heating effect on fission product deposition.

Hydrogen generation affected by blockage formation, oxidizable surface area, and hydrogen blanketing effect.

Coolability of degraded core prior to slumping of core below the lower core plate.

Effects of RCS water inventory upon fission product retention.

RCS system effect upon hydrogen generation (pressure, flow rate, etc.).

~

4. Status of Understanding

/ .

The MARCH code is currently used by NRC to model the thermal-hydraulic behavior in the reactor core and lower plenum, and the MERGE code is used to O

_ - . - . , _ , , ,w-, , - _. . _ - , - .. - - - - -

,. s 3

4 f~' . model the vessel upper plenum and ex-vessel region of the RCS. A number of 1

assumptions are used in MARCH / MERGE to greatly simplify modeling requirements; as a result, signiffcant uncertainties are believed to exist in the calculations for some accident sequences. These assumptions include: (1) one-dimensional hydrodynamics in the reactor vessel, (2) the shortest pathway 4

for fission product transport from the core through the break into the containment, (3) steam and hydrogen only are present in MERGE (i.e., no fission products or aerosols), (4) the decay heat generated by fission product deposition on structure surface is neglected in MERGE, (5) flows between connected volumes are calculated using the approximation that pressures equalize within a timestep rather than conservation of momentum, and (6) core slumping is treated using parametric models-that rely on user input options rather than a mechanistic treatment of the interaction between fuel debris and core support structgres.

The TRAC code is being used to investigate multi-dimensional flow patterns y in the upper plenum of a PWR. The result will be used to assess the first two

]

1 assumptions listed above by June 1984.

In the IDCOR program, somewhat detailed models have been developed to l

describe the progress of core degradation in-vessel. These codes, PWR-HEATUP and BWR-HEATUP, have been used as benchmarks for the less-detailed modeling of these processes in the MAAP code. The MAAP code, which has been the prinicpal thermal-hydraulic computational tool used by IDCOR in their integrated analyses of sequences for reference plants, is analogous to the MARCH code with regard to the scope of phenomena analyzed. Like MARCH / MERGE, the MAAP code also uses a number of assumptions which greatly simplify modeling requirements.

These assumptions include: (1) flowblockage, which terminates hydrogen prod-uction, occurs at an input specified temperature in the PWR version and at the time of zirconium melting in the BWR version, (2) molten zircaloy (or zirconium oxide) remains in a computational node until fuel melting occurs in the node, and (3) molten corium (fuel /zircaloy/ zirconium oxide) is instantly relocated above the lower core plate of the vessel, and (4) grid failure occurs at an arbitrary loading of fuel. The impact of those assumptions upon the code (w- calculations is yet to be determined.

-m-m--.----w---,._,_.,,. ,, y _. ._ _ . , _ , . _ . . , . , _ , - - , , .

_ m .,,,-., , . .-.-,--.,-,m ,y, . , , . - __-m__,_,,.-my-ws

l .

I 4

(- . 5. NRC Position

For near-term resolution, IDCOR is requested to quantify the uncertainties of its calculations. NRC and IDCOR should compare their calculations for the ,

same plants following.the same accident sequences selected.

NRC will continue to use the MARCH / MERGE code to study the RCS thermal-hydraulic behavior during severe accidents. The development of best-estimate computer codes, described in the following confirmatory wor, should be continued.

As those codes become reliably operational in 1985-86, they will be used to gradually replace the MARCH / MERGE calculations. ,

t

6. Continuing Confirmatory Work -

. Confirmatory research will continue beyond June 1984 to obtain best-estimate calculations with the uncertainties minimized and quantified. The work

/ '. includes:

s.O!

A. The extension of the existing best-estimate RCS thermal-hydraulic

codes, RELAPS and TRAC, to include modeling of (1) fission product and aerosol transport and deposition using existing models from the TRAP / MELT t

code, and (2) flow and temperature calculations for the primary system components under degraded conditions.

B. The development of detailed core behavior codes under degraded conditions to replace the core component in RELAPS and TRAC. Two codes under development include SCDAP for calculation of pin behavior up to a total loss of pin-like geometry, and MELPROG for melt progression in the lower plenum up to vessel failure after loss of core geometry. These codes will calculate (1) fission product and aerosol release from the core, (2) hy,drogen generation, (3) fuel liquefaction and relocation, and l (4) darnaged core cooling.

t'

%,/

e

e j

5

,. -s ,

I C. The integration of the above codes will be made via a system of coupled feeoback (if necessary) and loose linking where possible. In addition, the MIMAS code which was developed for TMI analysis with simpler modeling than SCDAP can also be used for a PWR degraded core analysis.

The linking and integration of SCDAP/MIMAS (modeling degraded core),

MELPROG (modeling lower plenum debris interaction with structure and water), the extended RELAPS/ TRAC (modeling upper plenum and ex-vessel region of the RCS), and TRAP-MELT will be made in 1985-86 so that a severe accident can be simulated for the entire RCS to realistically predict the release rates of radioactive fission products and hydrogen in the containment. The existing RELAPS and TRAC codes will be used to provide the initial conditions prior to the onset of fuel cladding deformation which generally occurs at around 1000K.

D. Best-estimate RCS calculations simulating severe accident sequences s

in a PWR will be performed in 1986. Similar calculations for a BWR will d! be performed in 1987. The results will be used to assess the uncertainties in the MARCH / MERGE codes and in IDCOR's HEATUP. codes for those accident sequences.

l l

l I

I l

l

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

w SISMAlt? 0F INIC/ CONTRACTOR /IDCOR I ER STATUS ASSOCIATED IDCOR REPORTS CtflRENT INIC/CONTRACTtut/IDCOR POSITIONS ISSUE PAPER TITit _ .IIRC/CONTR. LEADS J. Han - NRC 1. Technical Eleport 15.1A "In Areas of Antvement 1.1.1 - Reactor Coolant System Vessel Core Melt Progression Thermal and Hydraulic Behavior C. Allison - EGFA . Early heatup (T41000K) is we11 understood -

P. Cybulskis - BCL- Phencrena"

. Sufficiently accurate models and data exist

2. Technical Peport 15.1 " Analyst L for validation.

of In-Vessel Core Melt Pro- !

gression. Volume I. . Anreement that procer 9henomena are identified Phenomenological and Modeling to 2200K; e.g., clad selting, fuel Ifquefaction.

Background for the Pt81t and R5t Zr oxidation and 182 generation, clad ballooning.

Core 18eatuo Codes." "Voluee !!. '

il Users llanual and nodeling . IDCOR general approach in gedelling is reasonable ,

Details for the PHR Core Heat- and not unlike NRC aporoach. l

, i un Code."

[ "Volisme III. Users Manual Areas of Disagreenent !

and Modeling Details for the 8 i BWR Core Heatup Code" . IDC0lt assumes fuel remains in place to 3100 K.

9tC does not.

Areas Requiring Better Deffnftfon

. Inclusion of all lunortant phenomena during later heatup stages; e.g.. Ifquefaction and relocation n 22000K. flow blockage, flow f, blockage effects on Hp production, ballooning i effects on 112 production, pressure effects I (f.e., clad collapse irr lieu of clad ballooning) and its effect on earlier Ifquefaction (eutectic formation)andH2 generation (i.e., leiside j clad oxidation by 002 limits Tr available for

'. il2-producina steam reaction and Te retention).

. Themel hydraulics of Primary System under degraded conditions; 1.e. 3-D flow, upper plenue temperatures, and flow velocfties. .

, o ne

q. , ,

r < .

St# malty OF IIRC/ CONTRACTOR /IDCOR I PAPER STATUS _

CistitENT IIRC/ CONTRACT (pt/IDCOR P05fTIONS

.IlllC/CGIITR. LEADS A550CIATED IDCGt itEPORTS

, ISSUE PAPER TITLI_

'I ,

1. Technical 80eport 15.1A "In Areas of Agreement 1.1.1 - Reactor Coolant System J. Han - MRC C. Allison - EG7r. Vessel Core Melt Progression I Thermal and Hydraulle Behavior . Early heatup (T41000K) is well understood -

P. Cybulskis - BCL Phenonena" Sufficiently accurate models and data exist i

I e

2. Technical Peport 15.1 " Analyst i for valfdation.

of In-Vessel Core Melt Pro- . Anreement that procer ehenomena are fdentified I

{ gression, Volume 1 I

Phenomenological and Modeling to 2200K; e.g., clad melting, fuel liquefactfon, I : i Background for the PHR apd BHR Zr oxidation and H2 generation, clad ballooning, Core Heatuo Codes," "Volbee II, s i 1.~ Users Manual and Modeling . IDCOR general approach in modelling is twasonable

,' and not unifke MRC aporoach.

Details for the PHR Core Heat-uo Code." Areas of Disagroenent I

h , " Volume Ill, Users Manual j and Modeling Details for the 8 j BUR Core Heatup Code" . IDCOR assumes fuel remains in place to 3100 K.

i 'IRC does not.

Areas Requiring Better Deffnttlon l

l . Inclus1on of all innortant phenomena dur1no f ! later heatup stages; e.g., ifquefaction and relocation n 22000K, flow blockage, flow l

blockage effects on Hp production, ballooning

effects on 112 Production, pressure effects (i.e.. clad collapse irr lieu of clad ballooning) and its effect on earlier liquefaction (eutectic formation) and H2 generation (f.e., inside clad oxidation by U02 ilmits 7r available for ll2-Producina steam reaction and Te retention).

. Thermal hydraulles of Primary System under degraded conditions; f.e., 3-D flow, upper plenum temperatures, and flow velocities.

,i l

1 i I

I ,. .

g- . -

', 8 Issue Number :

1.1.2 Title

RATE AND MAGN'ITUDE OF HYDROGEN PRODUCTION IN FROM THE REACTOR COOLANT SYSTEM Signature of Author: . h 4*v Prepared by: J. Han v i

Draft i 2 Date 2/16/84 C. Allison, INEL / W. Camp and J. Rivard, SNL Contractor / Consultants : S. Hodge, ORNL / P. Cybulskis, BCL Review Process (List here contractor who will supply a statement on is C. Allison, IllELto IDCOR work)

W. Camp & J. Rivard, SNL Da

- Initial Branch Chief Y - /d kV 4 //7

~

) Asst. Div. Dir., N! '

D t./E 2 M4 Division Dir.

% 2/E2_/M Tech. Series Div. Director Date Sent to ACRS Date Sent to IDCOR l

Management Issues For near-term resolution, both NRC and IDCOR need to calculat accident sequences for the same plants.(a) to select plants (c) to authorize the work to do calculations and comparisons.

mow c/z l

'I(f

g.. -

.(

ISSUE 1.1.2 RATE AND MAGNITUDE OF HYDROGEN PRODUCTION IN THE VESSEL AND RELEASE FROM THE REACTOR COOLANT SYSTEM ,

1. Description of the Issue The issue deals with the rate of hydrogen release to the containment during a severe accident involving core degradation or melting. Two technical questions need to be addressed: (1)the hydrogen generation rate in the vessel, and (2) subsequent release of hydrogen from the vessel to the containment through a part of the reactor coolant system (RCS).

2. Implications of the Issue to Regulatory Questions Major implications of the issue are the potential problems caused by the presence of hydrogen in the containment regarding (1) containment loading and integrity, and (2) equipment survivability.

3. Subissues Effects of blockage formation and molten corium (zircaloy, Zr02 '

fuel, eutectic) relocation upon hydrogen generation.

Effect of cladding collapse (TMLB' sequence) upon zircaloy and fuel relocation.

Hydrogen blanketing effect.

Will corium enter lower plenum before lower core plate fails?

2 - - - - - , - - - , - - - - .

k . . .- .

2 f'.

~

Corium particle size distribution after contacting water in the lower plenum.

/

RCS system effect upon hydrogen generation.

Effects of multi-dimensional flow patterns in the vessel upon hydrogen generation.

Hydrogen generation due to water reaction with the steel structures in the upper and lower plenums of the vessel.

4. Status of Understanding The MIRCH code is currently used by NRC to model the hydrogen generation in the vessel during severe accidents, and the MERG code is used to model the transport of hydrogen from core exit to

,) the containment through a part of the RCS.

The HEATUP code (PWR-HEATUP or BWR-HEATUP) is used by IDCOR to model th generation in the core and the MAAP code is used to model the hydrogen generation in the lower plenum which is decoupled the core region for hydrogen calculation; the transport of hydrogen from the vessel to the containment is modeled by the MAAP code.

Hydrogen generation in the vessel is primarily due to chemica reaction between water and zircaloy cladding of fuelModels rods (zirca for shrouds of fuel assemblies may also be included for BWRs).

cladding oxidation and hydrogen generation for the temperature prior to Zircaloy melting (at =2200K) are generally valid. h gignificantuncertaintieseNst[nb!emofsusedinb and HEATUP codes to calculate hydrogen generation at high beyond zircaloy melting (= 2200K) and up to fuel meltin Ib -Be'th codet de - ? mechanistic modelt t o treat the relocation molten Zircaloy and fuel, which4 ect hydrogen generation in t ff vessel.

Furthermore N iiot c1 N regarding the hydrogen generat

[ due to oxidation of the steel structures in the upper and lowe D'

of the vessel during an extreme core meltdown accident.

-- -_ y

, .._.7___.

3 i

The transport process of hydrogen from the vessel to the containment is well understood. No significant problems are expected in the MERGE sad MAAP codes.

d A

5. MRC Position i

The BWR plants with nitrogen-inerted MARK I or MARK II containments are exempt from the hydrogen issue. For BWR plants with MARK III containments and for PWR plants with ice-condenser containments l with construction permit issued before March 28, 1979, hydrogen j

control equipment'(recombiners/1gniters) was proposed to be installed in the containment to handle the hydrogen generation equivalent to 75 percent of active fuel cladding oxidized

' (equivalent to about 2600 lbs. of hydrogen for a BWR/6 plant and 1500 lbs, of hydrogen for a PWR plant). ,For future plants the containment hydrogen control equipment wN proposed to have the

) capacity equivalent to 100 percent of active fuel cladding oxidized.

NRC IDCOR should quantify the uncertainties of its calculations.

and IDCOR should compare their calculations for selected accident se,quences.

6. Continuing Confirmatory Work Confirmatory research will continue beyond June 1984 to obtain best-estimate calculations with the uncertainties minimized and quantified.

I The work includes:

A., The extension of the existing best-estimate RCS system thermal-hydraulic codes, RELAP5 and TRAC, to include modeling of (1) fission product and aerosol transport and deposition using existing models fr the TRAP / MELT code, and (2) flow and temperature calculations for the

- primary system components under degraded conditions.

1

, - . _ - - - - , _ _ _ _ _ _ - - - - - - - -. ...---,._,_W-.-----,....

. , s , I. ) .

4

!p

~

B. The development of detailed core behavior codes under degraded Two codes conditipns to replace the core component in RELAP5 and TRAC.

under' development include SCDAP for calculation of pin behavior up to a total loss of pin-like geometry, and MELPROG for melt progression in the These lower plenum up to vessel failure after loss of core geometry.

codes will calculate (1) fission product and aerosol release from the core, (2) hydrogen generation (3) fuel liquefaction and relocation, and l (4)damagedcorecooling.

C. The integration of the above codes will be made via a system of.

In coupled feedback (if necessary) and , loose linking where possible.

addition, the MIMAS code which was developed for TMI analysis with simpler modeling than SCDAP can also be used for a PWR degraded core analysis. ThelinkingandintegrationofSCDAP/MIMAS(modelingdegrade core), MELPROG (modeling lower plenum debris interaction with structure and water), the extended RELAPS/ TRAC (modeling upper plenum and e

')

1985-86 so that a severt

, region of the RCS), and TRAP-MELT will be made in accident can be simulated for the entire RCS to realistically predict the release rates of radioactive fission products and hydrogen in the con-tainment. The existing RELAP5 and TRAC codes will be used to provide the initial conditions prior to the onset of fuel cladding deformation which generally occurs at around 1000K.

D.

Best-estimate RCS calculations simulating severe accident sequences in a PWR will be performed in 1986. Similar calculations for a BWR will be performed in 1987. The results will be used to assess the uncertainti-in the MARCH / MERGE codes and in IDCOR's HEATUP c sequences. .

l e

_.,..-nm

f' Q

!- ) .

\

.y

~. ~-

,. p

.g ,

- ' SISSIARY W MC/CSITMCTOR/ISCOR IS$5E PWER STAlWS ,[

'r.

. . 1 ISSME PAPER TITLE . W C/C M TR. LEARS ASSOCIATES IOC S rep 0RTS CINENT IEIC/ Mitt 3R POSITIONS !,

1 l, 1.1.2 - Rate and Magnitude of J. Men - latc . Technical Report 12.1. " Hydrogen Areas of Agreement l

Generation Durine Severe Core !

Hydrogen Production in the C. Allison - EG8G Demoge Sequences h S. Hodge - 0181L !

vessel and Release from the . Modeling(of

{

i Reactor Coolant systen - P. Cybulskis - 8CL meltino i.e..hwdrogen

~ 2209K) isproduction valid. before clad l

r

. Modelina and f .. --logy are sound for hydrogen productioninthevessellowerpienenregion.

' ___ v.

j*

Areas of Disagreement

j d

a g .

. None ffrely established, as y,et. {

I !! Areas Requiring Better Definition _

. Effects of fo11owing phenomena on H2 production i not now included in IDCOR calculations l

a)CladcollapseinIfewofballooning(TM.181 b) Cla*d ballooning (f.e., increase of Zr Il,'

surface area) c) Flow blockage and slupping (IDCOR claims g reduction in H2 I

d) Corium particle sfre in lower plenum (IDCOR ,I assumes large sire.71mm. thus no additional i H '** 1*" 'i'"""*

2 e) Effects of SRV cycling for DNR sequences not included in IDCOR analyses, i ;

~

I

\ ! ,

l

,i i

n l

i~

e' . O,,i* ,

s ;

l'

+i

'l -

.- . i2 m

W - SIMIARY OF 151C/CGITRACTOII/IDCOR IS$UE PAPER STATUS _

., [

' 5

. I

%'s '

ASSOCIATED TOCOR REPORTS CINIRENT Istt/ CONTRACTOR /IOCOR P05tTl0NS Ih- ISSUE PAPER TITtE .NRC/CONTR. LEADS ;

n j- Technical Report 12.1, "Hydrtmen Areas of Agreement :

1 1.1.2 - Rate and Magnitude of J. Han - IIRC i '

[ -

Hydrogen Production in the C. Allison - EG8G Generation Durfnc Severe Core S. Hodge - ORML Damage Sequences

VIssel and Release from the . Modeling(of meltints i.e., ~ hydrogen 2209K) isproduction valid. before clad [

Reactor Coolant System , - P. Cybulskis - BCL ?

. Modeling and phenomenology are sound for hydmgen

( production in the vessel lower plenum region.

, Areas of Disagreement l , % .

. None firmly established, as yet. ;

g. { ,

Areas Requiring Better Definition _

lf

. Effects of following p.a _ : on H2 Production !

not now included in IDCOR calculations !

a) Clad collapse in lieu of ballooning (TM.LB) l i b) Cla* d ballooning (f.e., increase of Zr surface area) c) Flew blockage and slu g ing (IDCOR claims j reduction in H2 I -

d) Corium particle sfre in lower plenum (IDCOR f -

asstates large size,>1 sus, thus no additional i 11 t'** T""'I *1'""

2 e) Effects of SRV cycling for BHR sequences not included in IDCOR analyses.

+. . '.

=

, s.

.- I

.a.

- - ~ . _ . _ ._

I -- - - - _ - . _ _

lV . ..

~

m.,

J h

Issue thsaber: 1.1. 3/1.1.4 CORE DEBRIS INTERACTION WITH LOWER PLENUM STRUCTURES, Title :

l REACTOR VESSEL, AND VESSEL PENETRATIONS signature of Author: [*.h w Prepared by: J. Han v.

Draft i 3/4 Date 2/16/84 i

W. Camp and J. Rivard, SNL Contractor / Consultants:

' Review Process (List here contractor who will supply a statement on toIDCORwork)

W. Camp & J. Riva,rd, SNL te Initial 7(< -

1 / ((

Branch Chief

~

Asst. Div. Dir., j , -

, /.N .

GOP) 2./e2 M Division Dir.

f% ,

'2/22/N Tech. Series Div. Director i

l Date Sent to ACRS Date Sent to IDCOR 1

fo/A -tY-121 ep t

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

/

I i

(

Feburary 16, 1984 f.

ISSUE 1.1.3/1.1.4 CORE DEBRIS INTERACTION WITH LOWER PLENUM STRUCTURES, REACTOR' VESSEL. AND VESSEL PENETRATIONS l

1. Description of the Issue The issue deals with core debris interaction with lower plenum structures, reactor vessel wall, and vessel penetrations during a severe accident involving core degradation or melting. .

2.

Imolications of the Issue to Reculatory Questions t I

?s Major implications of the issue are (1) reactor vessel integrity, i Cj (2) the timing and rate of core debris release to the containment thevesselorpenetrationsfail,and(3)steamandhydrogengenerationin the lower plenum.

! 3. Subissues 4

Core debris size distribution after contacting water in lower plenum Mass, temperature, and chemical composition of core debris.

Water inventory in lower plenum.

l 4. Status of Understanding '

i.

The MARCH Code is currently used by NRC Simplified to model the core l

l '

interaction assumptions are used:

with lower plenum structures, and reac ,

' /,

of core melting, (2) a user-input particle size for debris ih fragmen after contacting water in 1.ower plenum, (3) core debris interacti w,,--- ----....-,,n- , - - - - - - _ . - - - - - - _ _ - - - - - - - - _ - . - - - - - - , . . - - - - , - - - - - - . - - - - , , , - -

_ _ - . . .-,-------~,---,-~,----w,w--m-a -

c---,---

l s ,. e 2

ap lower plenum structures (excluding vessel bottom head) does r St begin until surrounding water has boiled off, and (4) vessel penetrations are lumped

' together with the vessel bottom head in the models. '

5. NRC Po'sition ,

~

For near-term reso1ution NRC and IDCOR should compare their cal-culations, MARCH versus MAAP, to determine the differences and to upper bound the uncertainties if possible.

The NRC will continue to use the MARCN code to study the issue.

l development of the best-estimate MELPROG code, described in the following l As MELPROG becomes operational confirmatory research shocid be continued.

! in 1984-85, its results will be used to replace the MARCH calculations.

6. Continuina Confirmatory Research Confirmatory research is being sponsored by NRC to develop a best-estimate mechanistic code, MELPROG, to study the core debris interaction

!- with the lower plenum structures, and reactor vessel and vessel penetrations.

Stand-The first version of the MELPROG code will be completed in 1984.

alone MELPROG calculations for a PWR will be made in 1985, using boundary conditions provided by a system code; similar calculations' for a BWR will be made in 1986.

! The MELPROG code will be linked with SCDAP/MIMAS, extended TRAC Best estimate RCS calculations simulating severe l and TRAP-MELT in 1985-86. Similar calculations _

accident sequences in a PWR will be perfonned in 1986.

for a BWR will be performed in 1987. The results will be used to assess the l

uncertainties,in the MARCH code and in IDCOR's MAAP code,for those sequences.

I f

e r a

f

\ <

_ _ _ - - _ _ _ _ _ - _ _ _ _ _ _ _ _ _ _ _- . , . . _ , _ . . - - - ..____--.c. _,..,, ,,, _-_,. _ . , , , , ,, . _ _ - .

~ D .- . _N.

.i f 4

1 L ,/ ~

l St90MRT OF ImC/CGITRACTOR/IDCOR ISSUE PAPER STATU5_ -

r f, f CURRENT NRC/ CONTRACTOR /l0COR P051T10NS NRC/CONTR. LEADS ASSOCIATED IDCOR REPORTS ISSUE PAPER TITLE _ -

1. Technical Report 15.1A. Areas of Aersament 1.1.3/1.1.4 - In-vessel J.Non- NRC R. Urfght - lutC "In vessel Core Melt . Initial core flow begins at Zry melting petet Core Melt Progressten Progression Phenomena" and Core-De6ris C. Allison - EGIS (flT3K).

U. Camp - SNI.

i Interaction with the ' 2. Techalcal lleport 15.29 . Molten Zry will freere in core areas less Reactor vessel and ..

Debris Coolability.

Structures Vessel Penetration and than 2173K.

8ebris Disperal"

- Areas of Disagreement

~

I .

g . IDCOR assumes U0, does not relocate untti it s reaches its melting point (3123K) - INIC models

11overaction" of UO2 bt ifquid Iry at flF3K.

alloutng for considerable UOr relocatten unch i

-l earlier in the transient. Ylits has been con-

firmed esserleentally in P9F and at Eft.

i~

l

.4 Areas Requirine Better Deffettfon l

l f

. IDCOR assumes sof ten 802 umediately relocates to lower core support plate. There is no evidence a

to suoport this.

I

. IOCOR essumes failure of LCP after a user-input fraction of core resides on it. Ils data is a

available to select proper input.

! . IOCOR assumptfens force the temperature of the

' cortus en the LCP to be ~3100K. thereby affecting l its failure time.

1I . Blockage and relocation effects on reducing M2 Seneration not proved. .

followine

. IDCOR assumes user-input failure vessel of LCP.fatture immediatelThus,nosofte I

corium/ structure interactions are modeled, pot proven.

e e

r. ?

~

V >

j St# MARY OF NIIC/ CONTRACTOR /IDCOR ISSUE PAPER STf 7U5_

ASSOCIATED 10CER REPtRTS CtRRENT IMC/CtutTRACTtR/IOCIR P0$ltitut$

intC/CGITR. LEADS ISSE PAPER TITLE _ -

1. Technical Report 15.lA. Areas of Aeroament_

1.1.3/1.1.4 - In-vessel J. Han - latC R. Wright - lutC *In vessel Core Melt . Initial core flow benfas at Iry melting point Core Melt Progression Progression Phenomena" and Core-De6ris C. Allison - Esas (2173K).

W. Cany - $NL i

Interaction with the ,

2. Technical Report 15.28, Reactor vessel and .

Debris Coelability. . Molten Iry will freeze in core areas less Strwetures Vessel Penetratten, and '

than 2173K.

Debris Disperal*

{ -

Areas of Dhagreement

\

s . IDCOR assumes UO, does not relocate esitfl f t '

reaches its melting potat (3123K) - HItC models

1touefactlen" of UO2 by ifguld Zry at 2173K.

alloutng for considerable relocation auch ecriler in the transient. T is has been con-firmed esserie:ntally in PSF and at Eft. ]

Areas Requirine Better Deffnftfon I ediately relocates to

. IDCOR Iouer coreassumes molten UO2 '"There is no evidence support plate.

to suoport this.

l

.10COR assumes fallure of LCP after a user-Input fraction of core resides on it. No data is a

available to select proper input.

. IDCOR assuetions force the tamperatore of the corfua on the LCP to be ~ 3100K thereby affecting its failure time.

. Blockage and relocation effects on reducing 182 l

generation not proved. .

l. . IDCOR assumes vessel failure immediately fellowine user-input failure of LCP. Thus, no molten -

cortimi/ structure interactions are modeled. Not W ne ,

l /

l 1

mamm, ..y, we ...e- . -

l

"**====_e-e. 4,

. t.

e Issue skaber 1.1.5 Title In-Ves'sel Steam Explosions

~

Prepared by: R. W. Wriaht Signature of Author: ]O Draft # 1 Date I h 2 / G4 T. Theofanous, M. Berman, M. Corradini Contractor / Consultants Review Process T. Theofanous, M. Berman

' Initial Date

~

Branch Chief-Asst. Div. Dir..

}

~

Division Dir. M 2/29A9d Tech. Series Div. Director M Z/2'1/1RL t

Date Sent to ACRS Date Sent to IDCOR o IDCOR claims direct containment or vessel failure is impossible, NRC impossible.

claims that failure is very unlikely, but cannot yet say it is o For NRC assessment need:

- Critical analysis of the data and conflicting models for pre-explosion mixing at full reactor scale to obtain the amount of interacting core melt.

(Purdue,Wisc/SNL)

- Critical review of the energetics and conclusions for the alpha direct failure mode of the reactor vessel and the containment. (SNL/Wisc, Purdue, LANL)

- Critical review of existing calculations (Swenson, Corradini) on the direct ~

impulsive failure of the reactor vessel by the steam explosion shock pressure.

1 (SNL/Wisc,Purdue,LANL)

- Assessment of whether the distributed core-support structure in the BWR lower

' plenum precludes a sufficiently energetic steam explosion to fail the ve b or the containment. (SNL/Wisc, Purdue, LANL) l Ford -t4 -iza c-H s

~' '

February 27, 1984 s

ISSUE 1.1.5 IN-VESSEL STEAM EXPLOSIONS R. W. Wright 4 j.

i

1. Description of Issue i This issue is concerned with the probability of occurrence and the magnitude of the steam ex,alosions (thermal explosions) resulting from the interaction of core melt and water, and whether or not such in-vessel steam explosions can directly threaten the integrity of the primary system .or even the containment. The steam explosions to be considered may be produced by either the collapse of a significant mass of core melt into lower plenum water or by accident recovery i ,,) action to reflood a partially molten (or liquified) core. Primary

system failure may be produced either by the impulsive loading of the hot weakened pressure vessel, or possibly the piping or steam-generator tubing, by the steam-explosion shock pressure; or by acceleration of an overlying slug of water and core debris above the steam explosion site leading to impact upon the vessel upper head. This slug acceleration and head impact mechanism, designated as the alpha failum mode in WASH-1400, in addition has the potential to fail the containment directly by ejection of the failed upper head as a large-mass missile through the containment.

In-vessel steam explosions also have the potential to affect the accident sequence in less direct ways, particularly when accident recovery is attempted by reflooding the core. Steam explosions can disrupt a partially molten core and widely redistribute the debris within the reactor vessel.

The very fine debris (size of the order of tens of microns) from fuel melt i that has actually participated (fragmented and delivered heat) in a steam explosion makes debris beds that have very low coolability limits when

~# reflooded. ,

1 1 l

_s A 2. Implications to Regulatory Questions

\.

This issue provides the direct input of the impulsive pressure loading 4

to the heated and weakened reactor vessel (and piping and steam generator tubes) from an in-vessel steam explosion for Issue 1.1.7, Primary System Failure from Overpressure. It impacts upon Issue 1.4, Fission Product Release and Transport, because it provides a potential mechanism for direct containment failure by an in-vessel steam explosion, namely the generation of a large-mass missile (such as the upper head) that penetrates the containment. It impacts Issue 1.1.6, Recovery Potential Prior to Vessel Failure, by providing the conditions, if any, for which accident recovery by reflooding a partially-molten core can worsen the accident by failing the reactor primary system or even the containment by an in-vessel steam explosion. Steam explosions also impact hydrogen generation in Issue 1.1.4, because of thewnetallic, component in the melt, in-vessel non-explosive steam generation in Issue 1.1.7, and in-vessel fission-product release from i the melt in Issue 1.4.

l .,',

3. Subissue Determination of the pressure-time impulse of a steam explosion in a core-melt accident and the kinetic energy and momentum generated by acceleration of an overlying constraining slug require answers to the following questions:

What are the initial conditions. in the significant core-melt accident sequences at the time of melt-water contact, both for core collapse l into the lower plenum water and for accident recovery by reflooding?

Important are the mass and temperature distributions of the melt, its l composition, the vessel pressure, and the configuration of the melt, the water, and the solid core debris and reactor structure.

What is the probability that the contact of core-melt and water will produce a steam explosion under these conditions?

c l .**

l .

2

_ .- _ . m n. _ , .

- =

7;- --

- - - - -- -- - -- - - - - - - --- ~ - -

What is the pre-explosion coarse-mixed configuration of the melt (s

4

( and the water, including the mixed masses of melt and water, the mixture void fraction, and the mean size of the coarse mixture?

Are there effective limits to the masses of melt and water that can be mixed?

At the existing vessel pressure, does a sufficiently energetic trigger occur and after how much premixing (timing) to propagate a thermal detonation (steam explosion) through the existing coarse melt-water-steam premixture?

What is the pressure-time impulse produced by the steam explosion?

Is this pressure impulse sufficient to fail the reactor vessel, the piping, or the steam generator tubing, allowing for the relevant temperatures (really Issue 1.1.7)? -

! } What is the conversion ratio of the total thermal energy in the melt

_7 into mechanical work, and how much less is this than the thermodynamic limit?

Does the overlying slug have sufficient kinetic energy and momentum to separate the upper head or other missiles from the vessel upon slug impact?

Do the separated upper head or other missiles have sufficient kinetic energy to penetrate the missile shield and fail the containment upon inpact?

1 Does the distributed core-support structure from below ty S Qcontrol-rod drive housings and the instrument tubes)in a BWR lower plenum preclude the development of a sufficiently energetic and coherent steam explosion in a BWR to threaten the reactor vessel?

/ Does the BWR ECCS water injection by sprays from above preclude the occurrence of a vessel-threatening steam explosion?

3

. , . , , - - -- , , . , . y -

r,. - - - - - - -_ _ . .- -- - - - - . ,. .-r ,,- - - - - - . - - --- , -

1

! 4. Status of Understanding l l

Steam explosions, as commonly named, are a special case of thermal l explosions, which are explosive interactions that can occur upon contact between a cold itquid and a hot liquid at a temperature far above the boiling point of the cold liquid. It is now known that the mechanism of thermal explosions, as proposed by Board and Hall, is analagous to that of chemical detonations, at least for the class of interactions in which a more dense hot liquid is poured into the cold.

In this free contacting mode, there is a pre-explosion phase of mixing and coarse fragmentation in which a detonable mixture is established.

This is followed by a trigger, the propagation or detonation phase with fine fragmentatrion and rapid heat transfer and vapor generation to sustain the propagating shock, an'd an expansion phase in which work can be performed on moving constraints. Thermal explosions may also be possible as a surface interaction between a hot and a cold liquid, and they are also known to be produced by transient heating and vaporization of a hot liquid previously dispersed in the cold liquid. This produced the steam explosions in prompt-burst power excursions in the SL-1 accident and the BORAX-1 destructive test, but the steam explosion in the SPERT-ID destructive test was different in that it did not involve fuel vaporization. -

f There must be a relatively low heat transfer rate between the hot and the cold liquids during the pre-explosion mixing if the temperature difference that drives the thermal explosion is to be maintained. Stable film boiling is a necessary condition for this low heat transfer in the pre-explosion intennixing and coarse fragmentation. Experiments show that the liquids ,

l subdivide and intermix with coarse fragmentation under the influence of density differences and vapor flow as one liquid is poured into the other, with the mixing zone cross section increasing with the depth of the pour.

In the absence af external triggers, thermal explosions appear to occur when the coarse fragmentation has reached the centimeter-to-millimeter

l' size range. Henry and Fauske have hypothesized that a flooding limit from

D .

4 s , , - - -

-. --;^ _ __

,- the upward vapor flow against the down-flowing dense hot liquid limits the surface area that can be created by fragmentation during the mixing process.

They have developed a steady-state, one-dimensional, counter-current-flow model for this ' process that predicts that for core melt masses in the 30,000kg range, the melt could only subdivide down to about 1 meter diameter, which is far too.large too large to sustain a steam explosion. Other mixing models that treat these processes (Corradini, Theofanous) do not predict such a strong limit on the mixing at full reactor scale. The predictions of the different models, however, do not differ significantly at the 30kg maximum size of the experiments that have been performed, and experiments at much larger scale (1,000kg or more) may prove to be necessary to resolve this difference. It should be emphasized that the question of the mass of core melt that can be intemixed with the water at sizes sufficiently small to give a detonable mixture under accident conditions at full reactor scale is currently the. p,rimary uncertainty in determining the consequences of in-vessel steam explosions.

A successful steam explosion trigger must provide sufficient momentum and

(-1 energy to fragment.at least one premixed fuel drop into fine enough fragments to start a self-propagating steam explosion. It is known from experiment that sufficient naturally occurring triggers exist at one atmosphere in premixed systems in the millimeter size range to produce spontaneous steam explosions with thermite melts in water. Increasing pressure makes triggering more difficult, but artificially-triggered steam explosions have occurred at l

system pressures as high as 3.0 MPa (450 psi). Whether there is an effective threshold pressure above which naturally-triggered steam explosions would not occur under reactor-accident conditions is not known.

Experiments at SNL with metal-oxidic thermite melts and water have given conversion ratios of thermal-to'-mechanical energy of a few percent, with a l

range extending up to 15 percent, which is about half the conversion ratio thermodynamica11y possible. There are questions about the direct application If I of such numbers to full reactor scale and to direct accident conditions.

the reactor-accident steam explosion occurs with larger mixing-zone particle j

sizes than in these experiments, the conversion ratios should be reduced, and l,

! (.' if a significant fraction of the core melt has not fragmented sufficiently in

(

5

.c -

m the mixing process to participate in the steam explosion, the conversion ratio would definitely be reduced. Mechanistic models are not available to calculate either the mixing process or the explosion propagation and expansion process to give quantitative answers to such scale and initial-condition questions. Under these conditions, the experimental conversion ratios can be used for conservative bounding estimates of the reactor accident conversion. ratios. The very uncertain and unvalidated mixing models can be used for estimates of the amount of core melt actually participating in the steam explosion, but better modeling and some experimental validatiop of the models is badly needed for this.

The primary concern about the consequences of in-vessel steam explosions has been the direct alpha failure mode of the reactor-vessel upper head and its ejection through the containment. In this failure mode the in-vessel steam emplosion accelerates a slug of water, melt, and solid core debris overlying and constraining the steam explosion, into impact with the upper head, and expels the failed head from the reactor vessel through the j missile shield and the containment as a large-mass missile. Slug impact may also expel the control-rod drives from the upper head, but calculations (Swenson and Corradini, NUREG/CR-2307) have shown that these would not penetrate the missile shield above the reactor vessel. In addition to the energetics of the steam explosion, there is a question abcut whether such an overlying slug would actually exist under core melt down conditions, and whether it would be accelerated as a lumped mass into impact with the upper head. Without su.ch a slug, the high-pressure steam generated in the steam l

explosion would vent into the reactor vessel void volume doing relatively little destructive work. Another important question is what fraction of the work done on the slug is dissipated by frictional losses to the internal f reactor structure.

Also of concern is the possibility that the direct pressure-time impuhe of the, steam-explosion shock will fail the lower reactor vessel or lower head, or even the piping or steam generator tubing. Some calculations (Swenson and Corradini, NUREG/CR-2307) have indicated that the steam-

[' explosion energy thres hold for such impulsive failure of the reactor

.V .

6 e- , w-e--- ,r . - - , . --- - - - , . , - - - - , e ,- - - - - - - - - , - - - ,

,r - vessel is lower than that for the slug acceleration and impact direct

$ failure of the upper head and the containment in the alpha failure mode.

This direct impulsive failure of the lower vessel or head would vent the steam-explosiori pressure, truncating the slug acceleration process and preventing the direct alpha failure mode. If these calculations are correct, this direct impulsive failure of the lower vessel or head acts like a safety valve to prevent the direct failure of the containment by an in-vessel steam explosion leading instead to a less serious ex-vessel core meltdown sequence.

In-vessel steam explosions need to be considered both for failure of the PWR core support structure with massive collapse of the molten core into the lower plenum water and for core reflooding in PWR accident recovery.

It is important to know in Issue 1.1.6 whether there are conditions in which recovery action by core reflooding can worsen the accident, as would be the case for the direct alpha failure mode of the vessel upper head and the containment upper head. With core reflooding in a PWR, the high-conse-

,) quence alpha direct failure mode of the vessel head and the containment would require that the steam explosion not be triggered until essentially all of the molten core region had been reflooded. A steam explosion triggered when the water just reaches the melt could disrupt the melt but would 'do little direct damage.

Direct failure of the vessel or the containment by an in-vessel steam explosion does not appear to be possible in BWRs. The distributed BWR core support from below by the control-rod guide housings along with the instrument l tubes should prevent propagation of a coherent large steam explosion in the lower plenum, although experimental confirmation to support this hypothesis i is lacking. In addition, the distributed BWR core support from below should

! prevent the massive coherent fall of a partially molten core into the lower ,

plenum water.

The BWR ECCS water injection by core sprays (if still functioning under core-melt conditions) should cool the melt surface and form a crust rather than producing a steam explosion, although later failure of the crust when

(' ' flooded might produc,e a steam explosion of relatively low energy.

7 l

*s

- NRC/IDCOR Position Our current understanding of the steam explosion process is qualitative, and validated mechanistic models that would give reliable assessments of steam-explosion consequences, particularly at full reactor scale, do not exist. Therefore safety assessment requires a mix of limiting calcu-lations and the application of a limited, relatively small-scale experi ,

mental data base to full reactor scale by the use of models of the scale-dependent processes. These models have large differences in their pre-dictions of steam-explosion behavior, and they also have little experi-mental validation. In this situation the consensus of the NRC staff and contractors is as follows:

Experiments have shown that melt-water steam explosions do occur frequently in experiments with metal-oxidic thermite melts, and therefore must be considdred likely in core-melt accidents.

Experiments indicate that steam explosions are more difficult to trigger at higher pressures, but it is not possible at present to set a threshold pressure above which naturally-occurring steam explosions would not occur.

In the alpha direct containment failure mode, a steam explosion accelerated slug of water and core debris impacts upon and fails the upper vessel head, ejecting it as a large mass missile through the missile shield and the containment. Analysis of this failure mode shows that a steam-explosion energy in excess of about 2,000 MJ is required. This would require an exceedingly energetic steam explosion, involving the heat from about 10% of a 130,000kg core as melt at maximum thermodynamic efficiency, or correspondingly more of the core at lower efficiency. On the basis of current understanding and information, such an energetic steam explosion is considered to be unlikely, but has not been demonstrated to be impossible.

i

/. There are a number of reasons, discussed below, why such an energetic steam explosion may not be possible. There are major questions about 0

l l .

-- - ^^- - -

e

/- the possibility of a pre-explosion coarse mixing of such large ,

$ masses of core melt with water with melt fragmentation down to the centimeter scale. An early explosion before the mixing is complete

would terminate the process and limit the melt mass mixed and therefore the energy of the explosion. There is experimental evidence that this limiting pr6 cess does occur, but no validated models for evaluation of this effect. There are also fundamental thermal-hydrodynamic questions (involving steam flow rates and melt and water fluidization) about limits on the amounts of core melt and water that can be mixed on the centimeter scale required for a vessel-and-containment-failing steam

! explosion. While there is much current work in this area, current models give vastly different results at full reactor scale, and larger scale experiments may be needed to resolve this difference. A current NRC mixing model (Corradini) indicates that mixing 10% of the molten core mass with" water with about 2 centimeter melt drops is possible. Currently there is uncertainty about whether an effective slug of water and core s

debris would overlie and constrain the steam explosion, under actual conditions, so that the alpha containment failure mode of slug acceleration and vessel-head impact with head ejection through the containment could occur. Otherwise the steam from the steam explosion would vent hannlessly into the partially-voided reactor vessel, although impulsive failure of the reactor vessel, the piping, or the steam generator tubing by the shock pressure must still be considered. The question of existence and accele-ration of a slug of water, melt, and solid debris is under intensive investigation, but definitive conclusions are not yet possible.

Experiments on steam-explosion conversion ratios have been performed at SNL with 30kg drops of metal-oxidic thermite melts into water. These I

experiments gave conversion ratios of the total thermal energy in the l melt into kinetic energy of typically a few percent, with results ranging up to a high of 15%, which is about half that thermodynamica11'y possible.

l Without validated mechanistic models of the steam explosion process, these and other data have to be used to estimate the conversion ratios l ~

under the different conditions of core-melt accidents and at full reactor scale. Use of the higher (15%) number is believed to be but cannot be 9

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

1. _ _ _ _ _ .

- - ^ ~

l 1_,L__ ..

proved to be conservative. Calculations (Swenson and Corradini,

\

NUREG/CR-2307) have indicated that the steam-explosion energy threshold for failure of the lower rector vessel or the lower head by shock over-pressure may be lower than the energy threshold for the alpha-mode slug-impact failure of the upper head with its ejection through the containment. Such lower head failure would truncate the. slug acceleration process and prevent the alpha mode upper head and containment failure, producing instead an ex-vessel

core melt sequence. These calculations combined a parametric treatment of the steam explosion with the dynamics of the lower head response and failure.

It appears that the extensive distributed BWR core-support structure in the lower plenum would preclude an energetic vessel-threatening steam explosion in the lower plenum, but this effect has not been demonstrated experimentally in the appropriate geometry. With this distributed core-support structure, the coherent collapse of a large fraction of the core mass as melt into the lower plenum does not

).

appear reasonable.

IDCOR The IDCOR position is that, while in-vessel steam explosions can occur, a number of fundamental phenomena prevent them from being sufficiently l

energetic to fail the reactor vessel or the containment. These phenomena l

I include the amount of melt and water that can be premixed to the necessary scale, limitations on the required trigger energy, and the absence of an f

effective slug to transfer the steam-explosion energy to the vessel upper l

head. The required energy to fail the upper head is taken as 1,000MJ, which is said to result from about 35,000kg of melt interacting with water with a 10% conversion ratio of thermal into mechanical energy (about 30% of that thermodynamically possible).

IDCOR uses the Henry-Fauske coarse-mixing model to argue that only about

/ 100kg of melt can undergo coarse mixing down to a one centimeter drop size, or that with 30,000kg melts, mixing would stop with meter-sized drops of melt that would be far too large for a steam explosion. In contrast, one NRC mixing model (Corradini), predicts that about 10,000kg 10

-...-----.-.__,_n. - _ _ , . , , , . , . , _ _ . , , .. -_,--~ ,,

~

}, of melt can fragment in mixing down to 2 centimeter melt drop size. l The Henry-Fauske mixing model is a steady-state one-dimensional counter- ,

current-flow model for melt break up by the steam generated by the heat flux from the fragmented melt surface area. i J

IDCOR also argues that sufficient trigger energy is not available to l

propagate a steam explosion in large scale, primarily because of the very large melt-drop size calculated by the Henry-Fauske mixing model.

It is the IDCOR position that an effective slug for efficient energy transfer to the vessel upper head will not exist because of the pre-explosion steam flow through the slug water.

IDCOR accepts the SNL experimental conversion ration results but does not consider them relevant for full reactor scale.

_., Path to Resolution

~i C'". ! '

It should be possible to narrow substantially the extensive technical differences between the stated positions of IDCOR and the NRC staff and consultants on this issue by a technical meeting to discuss the mcdels, the available data, and the reactor-safety-significant inter-pretations to be drawn from these. Work to narrow and resolve these technical differences with IDCOR, however, may not have high priority because of the limited safety significance of in-vessel steam expolosions, and because resolution of only one or two of these technical differences may be sufficient to resolve in-vessel steam explosions as a reactor-safety issue. The research currently underway on the processes of pre-explosion mixing and slug-ehergy transfer may significantly narrow these differences and also decrease the uncertainties in our understanding of the consequences of in-vessel steam explosions.

O

(. .

11

' ~ ~ ~ ~ ~~

., .--- - - - -- - - i_Z __ _

6.0 Continuing Confirmatory Work The continuing confirmatory steam explosion research is concentrating on the pre-expl'osion mixing process for both in-vessel and ex-vessel applications, with both experiments and analysis. Primary emphasis is given to application of the results of steam explosion research to full reactor scale. A major consideration is whether or not a large-scale (1,000kg melt) experiment will be needed to select among the very different mixing models (or to fonnulate new models) in analysis of the consequences of steam explosions at full reactor scale. Current models give similar predictions at the small scale (30kg) of current experiments.

Analysis is also underway on whether or not an effective constraining slug of water,4mit, and solid debris exists over the steam explosion

~

under the in-vessel accident conditions existing at the time the steam explosion occurs. The behavior of this slug in transferring the steam-

,) explosion work potential to the vessel upper head by acceleration and

~

then impact and the friction losses by the slug to the reactor internal structure are under analysis with advanced codes. The question of direct failure of the lower vessel or head by the shock impulse of the steam explosion also under analysis.

Confirmatory experiments will also be performed with more prototypic initial conditions for melt-water mass ratios and geometries and with better measurements of conversion ratios. Experiments on reflooding a pool of thermite melt are also underway.

k .

' 12

,O ] . :. . ._' ,

',; g..n, . . n . (.

. .~. . . . .' ;- -

k . .,, - ~ 3 r n. '

W' -

4 s :- -

Ae .*

es . ,

$ l v

f, ..s; .\

_ St9MAAV 0F liRC/ CONTRACTOR /tDCOR IS$ilE PWER STATUS _ ..:

1

) #.- 'lSSE P#tt TITLE -

WC/CONTR. LEADS _ ASSOCIATED 10CS REPORTS CMINitilf IIRC/ CONTRACTS /IOCOR POSITIONS tj 1.1.) In-Vessel Steam Explosions R. W. Wright, NRC Technical Report 14.1A, " Key Areas of Agreement:

l~

. T. Theofanous, Purdue Phenomenological Models for Direct failure of the containment or the reactor l '.

an in-wessel steam explosion is very d

(- , ,

M. Bertian, SNI. Assessing Explosive Steam "jjNy b 4 i.. M. Corradini, Wisconsin Generation Rates

l i-g 1 3 ,

Areas of Disagreement: I

'f' I I l y-N g But the been shown istCtoposition is that be lupossible. this

( , ka twgfailure o p (, has not'(.)

gl t .

.- Whether or not a significant a'nd fraction of the

-* core as melt can mix with water and tragment doun ,

1 to detonable size during the pre-explosion mixing i i.

l g. mixing phase. The validity of the different mixing l l:

g[ models is questioned. j

'i The IDCOR position that sufficiently energetic .p triggers do not exist to produce an energetic i.

steam explosion at full reactor scale. '

The maximum work and peak pressure that can be produced by an in-vessel steam explosion in core seit accident sequences. I I

Areas Requiring Further Definition: !

Whether or not an effective slug of water, melt, i ,

and core debris alli exist above the steam !,

~

explosion to provide an efficient energy transfer

- mechanism to fall the vessel upper head and to ,

. eject it through the containment. This is the alpha I mode containment failure mechanism.

j . .

e Whether or not direct tapulsive failure of the lower

! reactor vessel or the lower head can be produced by j!

i I

I an in-wessel. steam explosion. Such a direct lower vessel failure would went the steam explosion and j}

truncate the slug acceleration process to prevent :

the alpha-mode failure of the upper head and the containment, leading instead to an ex-vessel core- ,

' (Overi 4.4 j melt sequence. e I

f*)

. %s - . .

'g,',,,

.. M. e','3

. .s g .

4 . Z. . , .

F

'6

{

s Areas Reeutries Further Definittent (continued) l <

idhether or not the distributed Elst core support structure . . .

of the control-rod guide housings precludes an energetic !l Wessel-threatening steam espipston in the louer pienism. I ,

lAmther natural triggering of an energetic vessel I*

threatening steam explosion can be precluded at high reactor vessel pressure, and, if so, at what threshold pressure.

lihether or not a vessel or containment-threatening steam explosion can be produced by reflooding a' partially molten core, and, if so, under what conditions. .

i e i t

I s i.

l I

I I

t O %

1 e %

8

[

e

.. 8

. , .. ..il

, .. ... ..3,,

i

'. .* - ., .y 4

.. . s: . ;.,.51j;.,

.. . . . .4 '... . . : 4:. ,.e . .. . . '

- , .4 .. .:. .z.,n. .. ,w . ,

[' ,,

Y,,/

_SismARY OF NRC/ CONTRACTOR /IDCOR ISSUE PAPER STATUS _ .

f..

J

[ *-'

i

$w ISSIE PAPER TITLE ,.HRC/CONTR. LEADS ASSOCIATED IDCOR REPORTS CINIRENT NRC/ CONTRACTOR /IDCOR POSITIONS R. W. Wright, NRC Technical Report 14.lA, " Key Areas of Agreement:

1

, 1.1.) In-Vessel Steam Explosions 1 , T. Theofanous, Purdue Phenomenological Models for Direct failure of the containment or the reactor

' vessel by an in-vessel steam explosion is very

e M. Berman, SNI. Assessing Explosive Steam unilkely.

' i .. M. Corradini, Wisconsin Generation Rates"

, , 6 .l.

Areas of Disagreement:

g j n <

But the NRC position is that this failure has not

}- been shown to be impossible, j, .

j ,

l ,[. s 1 Whether or not a significant and fraction of the l (,.3 core as melt can mix with water and fragnent down

I j ,

to detonable si e during the pre-explosion mixin j

< l . mixing phase. he validity of the different mix ng ;

i models is questioned.

I** The IDCOR position that sufficiently energetic

l triggers do riot exist to produce an energetic I steam explosion at full reactor scale.

The maximum work and peak pressure that can be produced by an in-vessel steam explosion in core melt accident sequences.

Areas Requiring Further Definition: l' Whether or not an effective slug of water, melt, and core debris will exist above the steam "

explosion to provide an efficient energy transfer

- mechanism to fall the vessel upper head and to eject it through the containment. This is the alpha mode containment failure mechanism.

Whether or not direct impulsive failure of the lower reactor vessel or the lower head can be produced by

- an in-vessel steam explosion. Such a direct lower I vessel failure would vent the steam explosion and truncate the slug acceleration process to prevent the alpha-mode failure of the upper head and the C containment, leading instead to an ex-vessel core-

'd2 *

melt sequence. (Overl g

wa, - .. w . - - - .

-- .. -n--- -

.=

/.....,

(

o.,;s 3.- ,

T Areas Requiring Further Definition: (continued) hther or not the distributed BWR core support structure of the control-rod guide housings precludes an energetic .

vessel-threatening steam explosion in the lower plenum, '

hther natural triggering of an energetic vessel threatening steam explosion can be precluded at high reactor i vessel pressure, and, if so, at what threshold pressure.

l . Whether or not a vessel or containment-threatening steam I s

explosion can be produced by reflooding,a partially !

molten core, and, if so, under what conditions.

}1 i,

L r

')

e e

i 1 .

I I

I

- ~ ~

,,' . i

'g .,

W.

Issue Number 1.1. 6 ,

Title Recovery Potential Prior to Vessel Failure Prepa md by: R. W. Wriaht Signature of Author: ,

Draft i 2 Date 1/27/84 Contractor / Consultants Review Process (List here ceatractor who will supply a statement on issue relative to IDCOR work)

Initi 1 Da L Branch Chief ( - /

/7 !

Asst. Div. Dir..

Mb 21 7--

Division Dir.

Tech. Series Div. Director M DM i

Date Sent to ACRS Date Sent to IDCOR o Have no problem with IDCOR very-limited coolability claim (by which TMI-2 would not be coolable),

o For NRC assessment of the recovery potential, however, need:

- State of post-quench core from MELPROG calcolations (large uncertainty) (SNL)

- Coolability assessment including inlet flow (pump head), stratification, etc. (SNL) o Use Issue 1.1.7, Primary System Failure from Overpressure, for assessing this limitation to potential accident recovery by reflooding.

( -

k' - . fotA - s'Y-9u c,/r i'l grf I- . . . - - , - _ - _ . _, _ _ . _ __

~

Y 5. .

DRAFT 01/27/84 i

ISSUE 1.1.6 R'ecovery Potential Prior to Vessel Failure R. W. Wright Description of Isshe )

1.

This issue is concerned with whether or not a severe accident can be terminated by recovery action at different times in the higher-probability severe-core-damage accident sequences. Such action was taken successfully at TMI-2. In addition to reestablishing electrical power, potential recovery actions include core reflooding, reestablishing coolant flow, changing primary system pressure, and reestablishing heat sinks. A maje question is whether or not the damaged core is coolable following the fragmentation produced by reflood quenching .what the requirements on the emergency coolant supply are to produce coolability.

./ and the availability of a necessary heat sink. Another relevant question is whether or not there are accident conditions for which core reflooding or other potential recovery action can worsen the accident. This might happen if reflooding a partially-molten core causes a steam explosion or rapid steam generation (steam spike) that might directly fail th'e reactor vessel or even the containment (by missile generation), or if rapid i hydrogen generation from reflooding might fail the containment by a hydrogen explosion. As the PBF tests have shown, reflood quenching of hot fuel also produces a prompt puff release of fission products.

2. Implications to Regulatory Questions t

~

This issue is central to the subjects of potential accident recovery and accident management, as well as emergency response. It affects regul.atory' question 1, "What criteria should be used to detennine whether additional protection is needed or desirable?"; regulatory question 3, "How safe are existing' plants with respect to severe accidents?";

7

(,, regulatory question 4 "How can the level of protection for severe

- 1

[: .

a,ccidents be increased?"; and regulatory question 6. "Is additional l '

protection for severe accidents needed or desirable?"; all in the area of l It also affects potential acci, dent recovery and accident management.

l regulatory question 5. "What additional research or information is needed in tems of determining the coolability limits of damaged cores, including the requirements' on the emergency coolant supply to achieve coolability, and the probabilit'y and magnitude of steam explosions, and the magnitude o l

hydropngeneration,rapidsteam, generation,andfission-prou$trelease under reflood conditions. This ,infomation is needed to determine the conditions, if any, for which accident recovery actions such as reflooding can worsen the accident.

I

3. Subissues 4

What is the state (spatial permeability distribution) of the damaged core following reflood quenching (and fragmentation) for the relevant range of i -

reflood times for the higher-probability severe-core-damage accident i.' # sequences?

For these damage states, is the core coolable under stagnant reflooding.

and, if not, what are the requirements on the emergency coolant supply l

(pressure, flow, volume) tc achieve in-vessel coolability (which may not always be possible)?

1 Is one of the heat sinks for long.-tem coolability availabic, including the steam generators, the residual heat removal system, and feed-and-bleedoperation-?

Can accident recovery by core reflooding worsen the accident under ce Of concern are accident conditions, and if so, what are these conditions?

the possibility that reflooding a partially-molten core might cause a pressure-vessel or even containment threatening steam explosion, rapid hydrogen generation by reflooding might produce a containment-I -

threatening hydrogen explosion or a significant contribution to non-

,f explosive containment over pressurization, all in the presence of the

(- .

~ " "' - - - .

z. == --

3

.,'./ .

c' ,

i prompt puff release of fission products produced by the reflood quenching f g.

1 i . of the hot fuel.

//J. Status of Understanding The problem of assessing the potential for accident recovery by core l

reflooding is separable into component parts, as outlined in the previous Section 3, Subissues. These parts are: (1)characterizationofthestate l

of the damaged core following reflood quenching; (2) determination of the l

coolability limits (specific power) of this damaged core state, including l

requirements on emergency coolant supply; (3) the availability of a necessary heat sink; and (4) detemining under what, if any, accident l

conditions core reflooding can worsen the accident leading to more severe l

consequences. In general, our knowledge of the coolability limits (debris l

I specific powey in kw/kg U0 2 ) of a given state of core debris is much better

' than our knowledge of what the actual state of that debris is under given accident conditions, particularly after reflood quenching. What is needed for coolability analysis is the permeability distribution throughout the

'.3 core following reflood quenching at different times in the higher-probability severe-a:cident sequences. The details of particle size distributions, and l

I shapes are less important, but, because of capillary pressure, they do have a separate effect for fine debris and for stratified debris beds.

The mechanistic SCDAP and MELPROG codes, described in Issue Paper 1.1.3/1.1.4, are the best currently available tools for analyzing the state of the post-reflood-quench core. It should be emphasized, however, that the basis for the modeling in SCDAP and MELPROG of the state of damaged cores following core uncovery, significant oxidation, and reflood quenching, is very limited. Data are available from neutron tomography (MWe[aWigg$datU and post-irradiation examination (PIE) from the PBF Severe Fuel Damage (SFD) Scoping Test (ST). Tomographic data will soon be availab,le from the PBF SFD 1-1 test which had slow cooling, rather than quenching, in order to preserve the core-uncovery configuration of the debris The PBF ST test, with an unprototypically high steam-flow-rate for core-7

' uncovery conditions, had unprototypically-high oxidation of the Zircaloy

\

1-

4

, 'd.

- cladding and unprototypically low amounts of molten unoxidized metallic, l g .

Zircaloy to disolve UO2 and make liquified fuel. Nontheless, in the ST test, significant amounts of liquified fuel relocated downward and cemented fuel rod stubs.together to fonn large blockages, and these masses did not fragment into coolable particulate upon reflooding. Temperature records indicate that similar large uncooled masses may have existed in the TMI-2 core. There has not yet been significant analysis on the effect of such large masses and potential blockages of refrozen liquified fuel upon damaged core f

coolability limits. The relatively coarse cracked fuel pellets and oxidized cladding in the upper part of the ST test bundle should be readily coolable according to present models. Data from electrically-heated fuel-pin-bundle

! experiments at KfK with the low steam-flow rates prototypic uT core uncovery but without.reflood quench have shown greater liquified fuel relocation than the ST test. Data will soon be available from the PBF SFD 1-1 test, which had f Results will also prototypically, low steam-flow rates but no reflood quench.

- soon be available from the first two ACRR experiments on Debris Fonnation and ,

i Relocation which provide time-continuous cinematographic diagnostics of the

~

fuel-damage process and surface-temperature distributions under core-uncovery

'~

conditions. Some data will soon be available from the TMI-2 core examination The TMI-2 data will be of limited utility for coolability analysis because of '

l' and because

, the difficulty in representative sampling throughout the core vo ume the fine debris was swept out of the core by pump operation.

It is known in general from other experiments that quenching hot oxidic solid material (like fuel pellets) with water produces relative coarse debris in the few millimeter size range. On the other hand, quenching oxidic and even thermitt melts (that contain some metal) produces relatively fine debris in the tenths

(

of a millimeter range. Molten material that has gone through a steam explosion is much finer yet by roughly an order of magnitude. Bedsoffinhdebrisare much harder to cool than beds of coarse debris. It can be seen t characterization of the post-quench core debris by the permeability distribution in the cor,e is well beyond our current state of knowledge, and that only rough There are estimates of this penneability distribution can currently be made.

considerable data and rather sophisticated and partially validated models on

/' the dry-out coolability limits of well defined uniform debris bed under stagnant I b' .

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

-w - ,- m

. . . .- y e

j .- ,

1 conditions. Most of these experiments use beds of uniform metal spheres and

. [~ \..

electrical heating, but some validation data do exist with fission-heated beds of simulated debris with irregular particles and realistically broad particle size distribut' ions, and with stratified beds. Most of this work has been l

performed in LMFBR safety research, even when water was used as a coolant.

l l

Currently, fission-heated experiments are underway in ACRR to assess the validity j

of these LMFBR models for the LWR-specific accident conditions of very-deep beds, a pressure range up to 2500 psi, and the relatively coarse debris l

characteristic of LWR accident conditions, i

SCDAP incorporates the current Lipinski 1-D model of debris bed dry-out coolability limits, developed in LMFBR safety research, that gives the best l fit of all the coolability models to the world's data base on debris -

coolability. Recent experimental results have shown, however, that this 1

model strongly overpredicts the increase in coolability with increasing l pressure, so that some modifications for pressure effects will be needed.

! g By mid-March 1984, two experiments will have been performed in ACRR with -

*;) fission-heated simulated UO2 particulate debris. These experiments will furnish dry-out data for model validation for the LWR conditions of deep beds, l

high pressure, and for relatively coarse and relatively fine simulated LWR

)

i debris.

i There are no fundamental uncertainties in performing the systems analysis needed to determine the availability of possible heat sinks to maintain j Heat sinks

! damaged core coolability after accident recovery by reflooding.

i to be considered are the steam generators, the residual heat removal system, and feed and bleed operation. The latter will not generally be available for Combustion Engineering plants which do not have PORVs in the primary system.

l

' Concerns about possible worsening of the accident by core reflooding center on the possibility of producing a large steam explosion by reflooding a core which has a large fraction molten. There is a potential for failing the hot, l weakened reactor vessel by such a steam explosion, and this would also release the fission-product inventory in the :nelt into the containment. Experiments have shown that steam explosions do occur in reflooding a pool of thermite

{,,

. ~ , . . - ,--,,n - - .

_s

- 6 l '( ~

melt from above, but these experiments do not show how much of the melt and If a small steam explosion water would be, intermixed during bottom reflooding.

occurs to blow the melt and the water apart before they have become signi- ,

ficantly intermixed, the magnitude and the hazard from the steam explos' ion will be small. Reactor-vessel failure by a reflood-produced steam explosion is also treated under issues 1.1.5, Likelihood and Magnitude of In-Vessel Steam Explosions, and 1.1.7, Primary System Failure from Overpressure.

Also of concern is the rapid generation of hydrogen by reflooding with the This is more potential for a containment-threatening hydrogen explosion.

important somewhat earlier in the accident sequence when there has been less clad oxidation. Non-explosive rapid steam generation (steam spike) from reflooding a solid or a partially molten core along with the hydrogen generated by such reflooding needs to be evaluated as a quasi-static overpressure threat to the containment.

Information for this evaluation is currently available.

- with the principal uncertainty being knowledge of the state of the core.

The magnitude of the prompt puff release of fission products by the reflood l quenching of molten fuel, liquified fuel, and hot solid fuel needs to be i determined. The very limited available data on these processes under quench conditions comes mostly from the PBF Severe Fuel Damage Scoping Test '(SFD which has steam-flow rates and oxidation that were much higher than for prototypic core-uncovery conditions.

f"f. NRC/IDCOR positions Currently available models of damaged cere coolability (Lipinski 1-D model) with some adjustments at high pressures are adequate for assessing the ~

dry-out coolability limits for a given state of a damaged core, although better experimental validation and model improvement in some areas are desirable.' The Lipinski 1-D model also includes unvalidated modeling for the increase in the dry-out coolability limit with increasing pump head

,e (inletflow). Knowledge of the state of the core at different stages of severe-accident sequences, particularly follouing core reflooding with attendant debris fragmentation, is very primitive with almost no data base.

., *.. 1

.- . s The SCDAP and MELPROG codes model the state of the core during core uncovery and reflood quench, as well as debris coolability (Lipinski 1-D model), but they have almost no data base for validation of the modeling of the reflood-quench debris. Nevertheless they are the best,.

tools currently .available for such analysis.

Accident sequence and systems anlaysis need to be performed to deterinine '

heat sink availability for accident recovery by core reflooding, and also as part of the analysis to determine whether or not there are any conditions for which core reflooding can worsen the accident.

The BNL data on counter-current reflood quenching of an existing solid particulate debris bed from above give quenching steam generation rates that agree well with bed dry-out coolability data and models such as the Lipinski model. Co-current reflooding from below gives rates about a factor of five higher for the conditions. tested. However, reflood quenching

,a of partially molten cores can induce disruption and fragmentation so that the quenching and steam generation rates for stable solid particulate beds is exceeded. Analysis is needed on the primary system pressures generated by the relevant mass steam generation rates considering the available pressure relief areas and condensation surfaces.

IDCOR The particle-size-independent flat-plate critical heat flux is often

' used by IDCOR to assess the dry-out coolability limits of degraded cores, although a particle-size dependent IDCOR model with results somewhat similar to the Lipinski model is sometimes used. Two-dimensional effects Quenching of can increase the coolability by factors of two or three.

I particulate debris in the lower plenum from above would be too slow to prevent vessel melt through.

.)

e

r _ _. -_._ __ __

i 8 .

l

'f .

Steam generation rates during reflood quenching would not c'r allenge the In-vessel steam explosions would relief capacity of the primary system.

not exceed half the critical pressure 'and therefore would not fail the '

vessel by shock pressure.

path to Resolution-Although there are substantial technical differences between the NRC and the IDCOR analyses of the phenomena involved in accident recovery prior to vessel failure, resolution of the detailed differences does not appear to be necessary because IDCOR makes no claims for accident recovery if the core has not slumped into the lower plenum or the debris is relatively fine (<1mm). Under those conditions NRC analysis agrees that the core is coolable b9 reflooding, even without the pump flow used at TMI-2. Any significant differences here can probably be resolved by mutual discussions.

's It is the NRC position that assessing the damaged core coolability and accident recovery potential over a broader range of accident conditions CU requires use of the more precise coolability data and models such as a modified Lipinski model that include the effects of particle size including fines and stratification, and inlet flows.

On the question of the threat to the primary system and possibly the con-tainment from reflood-produced explosive and non-explosive rapid steam and hydrogen $s(Adg generation, reactor accident analysis of these threats is needed using currently available data and models, incomplete as these are with this analysis NRC would be in a position to evaluate indepen-dently the IDCOR conclusions that these threats cannot fail the primary system or the containment.

6.0 Continuing Confirmatory Work Results from the two fission-heated experiments in ACRR on the dry-out coolability limits of relatively deep beds (50cm) of relatively coarse and relatively fine simulated particulate core debris with representative

, 9 particle size distributions over the full pressure range to 2500 psi will be available in FY 84. The planned ACRR experiment to validate the effect on the coolabi,11ty limit of bed stratification and inlet flow in the Lipinski 1-D model is not currently funded. Data will be available from the Gerinan partners in the NRC Severe Fuel Damage program to validate '.

the coolability models for very-deep (100cm) beds of electricly-heated uniform spheres. 'Theofanous is also obtaining coolability data as a function of pressure for large-diameter (200cm), deep (100cm) beds contain-ing particulate debris between horizontal heating elements.

Continuing confirmatory work on steam explosions and on hydrogen generation that is relevant to the accident recovery potential is covered in the Confirmatory work on non-explosive respective issue papers, 1.1.2 and 1.1.5.

rapid steam generation under reflood conditions is continuing at BNL, in the NRC program 'and also at ANL.

e e l

[S e-

- O . W i ... *_ -

. . , , , . . , ~ .' } .. . -

. . !. >,' ) ,. .- - . 1 . , .

, \. . -

W ,

. i

fj

y. .-

i - .

_31setARV 0F NRC/ CONTRACTOR /IDCOR IS$UE P4PER STATUS _ . i a

5..t ,

t, .

1 i in u' IS$1E PAPER TITLE. ..llRC/CCNTR. LEADS ASSOCIATES IDCER IEp0RT5 CINUlFWT NIIC/ CONTRACTOR /IECSR p05ttleN5 i t

~'

' I) . , . i 0 ,

1.1.5.- Recovery Potentfal Prior R. Wrfght - HitC Technical lleport 15.25 " Debris Areas of Ayeement .

- N to Vessel Failure R. Lfpinski - SNI. Coolability. Vessel Penetration.

E. Gorham - Bergeren and Debris Dispersal" . Quenched cores of unstrettfled particles greater s' .' ,

than two millimeters att coolable in place. .' '

y p

~

y). '

Areas of Disagreement l

1. . j g ..

~ . IDCOR esos a particle-sire independent dryout modej {

s (flat-plateCHF). . '

1 . * ,

' (

$ . HRC uses Lfplaskt emperimentally-based dryout model l l

{y .

which nodels effects of particle sfre, pressere, and :

bed strattffcation.

d jV '

Areas Requirfne better Definition I

'l i

. HitC belIeyes IDCOR conclesfens en coolabi11ty under stagnant conditions do not soply for fine er strettfled f debris free aquenching partfally melted cores.

o.

. Effects of inlet flow should be, incorporated.

'l l 1

.q i 1 1 ,

i l -

I f, .

k I

I e

. 4 0

f?.f .

4.7

n

.n . (s e' -

! /

o , s. . .

, (

% %L.'

I ]/

StfMARY OF IRIC/ CONTRACTOR /IDCOR ISSUE PMR STATUS __ ..

y '

.g

.h !

v-' IS$1E PAPER TITLE _ .INtC/ CENTR. LEADS ASSOCf ATED IOCGt EPORTS CIMENT NRC/CWmtACTOR/10COR POSITIGIS i t'.1 D l~

1.1.6.- Recovery Potential Prior R. Wright - MRC Technical Report 15.28 " Debris Areas of Agreement l

1- to Yessel Failure R. Lipinskt - SNL Coolability. Vessel Penetration. !

E. Gorham - Bergeren and Debris Dispersal" . Quenched cores of unstrattffed particles getater I Y ,

than two millimeters are coolable in place. I Areas of Disagreement

! ,1 . . . IDCOR uses a particle-stre independent dryout model I p - (flat-plate CHF). ,

,1 1 {

'> , , MRC uses Lfpinski experimentally-based dryout model I' which models effects of particle sire, pressure, and i bed stratification.

Areas Requiring Better Definition l ,

l 4 l ; . MRC believes IDCOR conclusions on coolabillty under I '

stagnant conditions do not apply for fine or stratiflad l debris frtus quenching partf a11y melted cores. [i I

F. e

, . Effects of inlet finw should be, incorporated.

' 2 i: I I

. i.

i 1 ,

, . {

, . ?.

l

I t

i -

_.._.....-.m,.--... . . . - .. - - - - - - - - - *m

~ ~~~

,k k' ,

's Issue Number 1.1. 7

, Title Primary System Failure from Overpressure .

Prepared by: R. Wright /C. Serpan Signature of Author:, . A4 Draft # 2 Date 1/30/84 Contractor / Consultants Review Process (List here contractor who' will supply a statement on issue relative to IDCOR work)

- Initial te Branch Chief Mb .- [)

Asst.' Div. Dir., /$7</

CJd Division Dir. _ , G /2 2 ./

Tech. Series Div. Director M U22/W -

Date Sent to ACRS Date Sent to IDCOR o Need MELPROG analysis of core-melt progression and vessel and structure temperatures. (SNL) o Using existing data /models, need:

~

- P.S. pressure from reflood quenching, including steam generation rate, steam discharge rate, and cooling (BNL)7

- Evaluation of Steam Explosion shock pressure and potential to fail PJi. (SNL)

- Evaluation of quasi-static P.S. failure threshold (SNL?)

o Need check evaluation of failure calculations and material properties used is assessing vessel and P.S. failure from above impulsive and quati-static loads. ,

For A- t+-128 a.l t.

r2f/

..r. .

l 6 j

,i ' *. .

January 30, 1984 I

i

-[.

Issue 1.1.7 s- Primary System Failure from Overpressure R. W. Wright

1. Cascription of Issue This issue is concerned with whether or not the explosive or non-explosive, rapid generation of steam and also hydrogen produced by the interaction of hot dry or molten core debris and water can fail the heated and weakened reactor pressure vessel during a core-uncovery accident. This interaction can occur either because of a massive drop of core debris into water in the lower plenum, possibly following failure of the core support structure, or because of reflooding of the hot and potentially partially molten core during accident recovery action by cofre reflooding. If such a failure of the vessel does occur, the remaining coolant supply will be lost with prompt melting s of the core and entry of the core melt and solid debris into the reactor cavity.

2. Implications to Regulatory Questions This issue impacts upon subsequent system behavior as treated in Issue 1.2. Loading of the Containment, and Issue 1.4, Fission Product Release and Transport. This issue also impacts strongly on l

Accident Management strategy. It is important to know the range of l

core conditions, if any, for which reflooding a severely damaged hot, possibly partially molten core can worsen the accident by failing the heated, weakened reactor vessel by overpressurization.

J 3. Subissues What is the magnitude of the impulsive loading of the reactor vessel s'nd the primary system for the range of in-vessel steam explosions from massive drops of core melt into lower-plenum water 7

and from accident recovery by core reflooding for the relevant.

. range of accident , conditions?

..-., ..,y_- , _ , . , . _ - . , . , . . , . , ___ -.. - _ , _ . . . - _

.a , ,

s l

'E

.g.

Wiat is the magnitude of the vessel pressurization from non-explosive rapid steam generation and hydrogen generation from both massive drops of core melt and solid debris into lower plenum water and from accident recovery by core reflooding for the relevant range of accident conditions?

What is the pressure decay time constant from cooling and from steam discharge through primary system leaks?

What is the quenching time for non-explosive rapid steam generation for the range of conditions of interest?

What is the axial temperature distribution in the vessel walls for the core-uncovery sequences of interest, and what are the piping

} and steam generator temperatures?

What are the impulsive failure loadings for the reactor vessel, the piping, and the steam generators for the relevant temperatures?

What are thi quasi-static failure pressures for the reactor vessel,

! the piping, and the steam generators for the relevant temperatures?

For both impulsive and the quasi-static loadings, what are the modes of failure of the reactor vessel, the piping, and the steam generators?

,How do the probability of occurence and the magnitude of large steam explosions vary as a function of ambiant pressure?

4 4 4

e

, . , v. .-,, . -- - . . - , - . -- , . - - -- - . - . , - , , , , . . . , _ _-

,- ?' s 7' , (

Status of Understanding 4.

The technical question involved in assessing primary system failure from overpressure is separable into two parts. First, what are the pressure loads applied to the hot, weakened vessel, impulsively by steam explosions, and quasi-staticly by non-explosive rapid steam generation and hydrogen generation? These loads are required for the relevant ranges of damaged core conditions, water conditions, and geometries including leak areas in the reactor vessel and the primary system and also including pressure relief valve settings.

Second, what is the failure level for the hot, weakened vessel and the primary-system boundary for both quasi-static pressurization and impulsive loading. For assessing the consequences of the impulsive-loading from an in-vessel steam explosion that propagates j through the primary system, proper account must be taken of the low 2.) acoustic transmission when a pipe enters a large plenum, as at the entrance to the steam generators, and of elbow losses.

Both the pressure loads and the vessel and the primary system failure level are in large part determined by the state of the core, including the damaged core spatial distribution of mass and temperature.

This can be analyzed as a function of time for the dominant-risk accident

! sequences with mechanistic melt-progression analysis code MELPROG, for which the initial Mod 0 version is just becoming available. Currently, the models in MELPROG have little validation data, but such confirmetory data are becoming available from the results of the tests in PBF, MRR,

-l and later NRU in the Severe ' Fuel Damage research program. Mel t-progression analysis and the MELPROG code are treated under Issue 1

l'.1.3/1.1.4.

An evaluation of the impulsive loading from in-vessel steam explosions is to be made under Issue 1.1.5, Likelihood and Magnitude of In-

- Vese,el Steam Explosions, and this evaluation can be applied here.

+

- - - - - - - - - - -

y-.rs--- -wmw-----w- --+-----y- --w,er-- - - - * , - - - - - -,-----*----~w---+

w -- -- - - - - - - - - - -

i

.p .-4 b

Validated mechanistic models of the steam-explosion process do not exist, so par'ametric inputs are required. De limited and not precise available data, primarily from experiments in the FITS facility, indicate that the conversion of melt themal energy into work in steam explosions with corium melt dropped into ater ranges from a few percent downward, with corresponding low impulses from the short (millisecond) very-high steam-explosion pressure pul ses. There are major uncertainties in the application of current experimental results to full reactor scale, and in the pre-explosion processes of mixing the melt and the water that largely determine this application. Dere are also uncertainties as to what elevated system pressure levels are required to suppress naturally-triggered steam explosions under reactor-accident conditions. Experiments indicate that steam explosions become much harder to trigger at the higher pressures. De steam-explosion process involved in massive 3

Cil drops of melt into lower-plenum water and the process resulting from the reflooding of a partially molten core may be significantly different. In the reflooding case, in particular, an early steam explosion involving a small melt mass may disperse the melt away from the water preventing a large-scale steam explosion. His process can also occur in the dropping contact mode, but, because the time scale irr much shorter, introduction of a substantial I fraction of the melt into the water before a steam explosion is triggered is mere likely.

Vessel (primary system) ~ pressurization by non-explosive rapid steam generation depends on the fraction of the hot solid and molten core debris that is quenched on a time scale of the order of seconds and on the available supply of quenching water. For reflood quenching, the debris configuration to be quenched is not the particulate debris bed that results from the quenching, but the damaged core in

" the core uncovery configuration before the possibly explosive e.-- - - - , , - ,-4e - ,e -., n.,,----,,--,---,,------,,,,,,,--,--o,,,-,~,-,- . -

interaction involved in quenching. The quenching time for the fine particulate debris bed that results from quenching may not be i relevant. Sienching times significantly less than the pressure decay time cons. tant by wall cooling and by steam outflow through openings are effectively instantaneous. Even a relatively small in-vessel steam explosion can be a very effective mechanism for mixing the core debris, particularly the molten fraction, and the available vessel water.

I Detemining the impulsive or quasi-static failure loads of the reactor vessel requires, first a detennination of the temperature distribution in the vessel wall at the time of the steam pressurization, and second, the failure pressure or impulse for that temperature distribution. The temperature distribution in the vessel wall as a function of time for different accident sequences can be calculated with the new mechanistic melt-progression code MELPROG.

.,j MELPROG includes a quasi-static overpressure-failure model that utilizes the calculated temperatures of the vessel wall and penetrations.

In some cases this involves extrapolation of failure data beyond the available data base to higher temperatures. The failure criteria currently incorpo, rated in MELFROG Mod 0 are creep-fatigue interaction, using the Larsen-Miller parameter life-fraction rule, and melt f ,

i through. The failure modeling applies to both the vessel itself and to vessel penetrations. A brittle-failure model and a bolt-failure model are currently being added to MELPROG Mod 0 and will I

be available in late 1984. For cases near the failure thresholds calculated by MELPROG Mod 0, it would appear desirable to also perfonn a more detailed failure analysis. Such analysis should i

address the uncertainties in the material properties at the failure

~

wel1 temperatures, the fracture toughness of the mate'rfals (vessel, piping, steam-generator tubes, etc.) and, for the vessel especially.

l

'. the degraded conditions resulting from neutron irradiation and.the

/ effect of flaws. Analysis methods for predicting the failure e

N

.-,-f . . . , _ , .,_- _._. _ _- . ,,_,,-. . - . . , _ - .

' threshold of the vessel, the piping, and the steam-generator tubes are being developed in the MEBR/DET program, and should be available for later confimatory analysis and for possible incorporation in MELPROG Mod 1. , The Degraded Piping program and tests of the Surry steam-generator tubing should also furnish results for later confinnatory analysis.

For vessel failure under impulsive loadings, MELPROG Mod 0 uses creep-fatigue interactions with the Larsen-Miller life-fraction rule. Use of this criterion for impulsive loadings on a millisecond time scale has been demonstrated for the failure of stainless-steel LMFBR fuel cladding material in prompt-burst experiments in ACRR.

A separate brittle-failure criterion is currently being added to

! MELFROG Mod O. In the impulsive analysis for primary system failure from steam explosions, account must be taken of the acoustic transmission losses by reflection at area changes and of attenuation

. 4' down the pipe runs Wien analyzing failure in the sttsm-generator tubes.

5. IRC/IDCOR Pbsitions i Although primary system failure from overpressure is not currently considered to be ,a major risk, it is prenature to state an NRC technical position on this issue until appropriate analysis with ^

currently available tools has been performed. This should include MELPROG analysis on the state of the core (spatial mass and temperature distributions) and the vessel and support structures (temperature distributions) for the dominant risk accident sequences. Analysis l (from Issue 1.1.5) on the likelihood and magnitude of in-vessel steam explosions and the magnitude and uncertainties of the resulting pressure 1,mpulses for the conditions both of a massive drop of core

I melt into the lower plenum water and of accident recovery by reflooding a partially molten core. Analysis of the rate of quenching of the

9 e

       ,y-re         ,- ---         ,,,-c.%-,e..,--..-%---.-.       ,,c.---,--m,_m,--==-n,---,-,m--,--,,,wm-----

                =
  .t     ,        ,

j,.. , I

             ,                                                                                   7 core melt and solid core debris by the available uter with the analysis of the pressure-decay time constant by cooling and by
,                       steam discharge through openings gives the primary-system pressurization.

MELPROG analysis is needed for the reactor-vessel temperature distributions, with hand calculations for the piping and the steam-generator temperatures. From these temperabres and from the material properties, some of dich require extrapolations beyond the existing data base, estimates can be made about dether the 4 impulsive or quasi-static steam loads will fail the reactor vessel or the remainder of the primary-system boundry. It needs to be

recognized that there are many uncertainties involved in this analysis process, particularly because of the lack of materials-: properties data in the high-temperature regime.

IDCOR 1 The IDCOR position is that the steam generation rates during reflood quenching would not challenge the relief capacity of the primary system. It is also the IDCOR position that the maximum possible pressure in a reasonable-sized steam explosion is about one-half the critical pressure, so that the shock pressure from steam explosion cannot fail the primary-system directly. Fath to Resolution The analysis outlined here and the confimatory data should establish an acceptable techical posidon on this issue, which does not appear to be a major risk issue, however. The major aspect of the problem will be defining or limiting the range of uncertainty in the 3:onclusion.because of the' many fundamental uncertainties in the basic pr'ocesses and materials properties.

<~

t, -

6

             <e     -       ,=w- , -- ,v ewm -. , , , , -- -s   - , - - - -- - - - - - - - - , .   , - - - , - - - - - - - - - - - - - - -           - - - - - - - - - - - -

g. ; e l
i: m. i

    'w .

8 6.0 Continuing Confimatory Work 1 Confimatory data on'the basic models in the MELPROG melt-progression code will be available from the PBF and the ACRR tests on a continuing basis through 1986. Relevant research is also continuing on pplosive and non-explosive rapid steam generation for both the dropping and the reflood quenching modes. Relevant materials properties researt:h is also continuing on fracture toughness, steam generator tube properties, and degraded conditions from flaws and neutron irradiation, along with improved methods for failure prediction in the N/DET program. However, research on materials properties in the relevant very-high-temperature regime is not underway.

e e e s -l

   .l                                                                                                                         ,
       +
                                                                                                                                   ^6 9

N ,p ? .=. e. p r; 4 * .e e. .4.- y .; p.. . . a. . p p. pgy .. , , , ,9, , , ,

                                                                                                                             ,       .,    , ~ , , ,     , ,
                                                                                                                                                                                  /
             ;                       x                                               , .

N,. s - - y .

i.

                                 .                                    $1seunterult/teNinACTOR/10CORissWEPerfts7419s_                                                                                f.

l L s .- . ~. H' tssur PArta TITtt_ ..imC/Cem. LEADS ASSOCIATED 10COR REPORTS CNEWif NRC/Ce m RCTOR/IOCOR POSITIONS ->-

   !l                                                                                                                            Areas of Agreenent
   !     l4 1.1.n-PrimarySystem                            R. Wright - MRC                 Technical Report 14.18. " Key

Fallere from Overpressure W. Camp - SNL Phenomenological flodels for . Non-explosive steam generation from in-vessel '

         ,1                                                C. Serpen - NRC                 Assessinet flon-Explosive Steam          reflood quenching does not, by currset under-                            .

T. Ginsberg - IPil Generation Rates" standings threaten the reactor vessel ([ s s 4. O g Areas of Disagreement n . I {;, .: . - t d *<3- 4 Areas Retsfring Further Deflattfon x In . .. , .e ..

                                                                                                                                 . Effects of non-explestve steam generation require

[ { , more analysis s..t . p ' l, e ,

   *1 1                                                                                                                                                                                                 b l

L l 4 hI

     ?                                                                                                                                                ~

o l i i: !

 ?'                                                                                                                                                            q
              ?.

i .

       % J -e       , h ee ee. = +=.

I 1 - e e

           .e*.

I . I I *' ? ,1 *. *

                                                                                                                                                                                                             . ._ /                                                                      ~~/       . 'I D                                                                                                                                                       SIMIART OF IIRC/ CONTRACTOR /IDCOR ISSUE PAPER STATUS _                                                                        s r ..

\

                        *'                                         IS$1E PAPER TITLE _                                                                       .IIRC/CINITR. LEADS              ASSOCIATED 10 Cat KPORTS                    CINutENT IIRC/ CONTRACT (Bl/10COR POSITIONS                          ,
                                                                                                                                                                                                                                                                                                               .E l)$            i I

1.1.7.- Primary System R. Wright - NRC Technical Report 14.15, " Key Areas of Ay n _ .t t Fallure from Overpressure W. Camp - SNI. Phenomenological Pimiels for . Non-explosive steam generation from in. vessel

Assessing Non-Explosive Steam reflood quenching does not, by current under- I C. Serpan - NRC

        . j
                          ~

T. Ginsberg - BML Generation Rates" standing, threaten the reactor vessel !, j Areas of Dtsagreement i . . . None .

                                                                                                                                                                                                                             %     Areas Requiring Further Definition

s -

                ;                                                                                                                                                                                                                                                                                                 i
                                                                                                                                                                                                                                   . Effects of non-explosive steam generation require
               ,           1.                                                                                                                                                                                          ,

more analysis j

                ? I t          i.,                                                                                      .

el ( t ! ;* : j i j i' !l .

                  )  -

l l i e i

                                                                                                                                                                                                                                                          .                                                l t

I i e,. 6

                                                                                                                                                                                                                              - en          ~**          **                - * ' ..

h. -- amns i re w- . . . , *M. . - = = + . - .

( _ _- . -. .-

     . a         *

p C

I 9

                                                                    /                                                                              .D O

k on

           .' t a:

(.. . I V

e .. Issue Number 1.2.1 Title Containment Thermal and Hydraulic Behavior Prepare'd by: R.Barrett/J.Rosenthal Signature of Author: . Draft f- , Date 11/9/84 Contractor / Consultants M. Cunningham (NRC); W.T. Pratt (BNL) Review Process (List here contractor who will supply a statement on issue relative

                          .               toIDCORwork)

Initial Date Branch Chief nA //(( b Asst.Div.Dir.,!/ - ///dPP

                                                                                                    f Division Dir.

Tech. Series Div. Director h I f 8' Date Sent to ACRS , Date Sent to IDCOR

                                                                                .e e

e O

e FwA - t+ -41. 8 O'/7 4 . . _ _ . . ___ _ _ _ _ . . w _.

ISSUE PAPERS 1.2.1 Containment Thermal and Hydraulic Behavior

1. Description of the Issue This phenonmenological issue deals with our ability to predict thermal /

hydraulic behavior in containment given mass and energy sources. These sources are the subject for several other issue papers in series 1.1 and 1.2. An important aspect of this issue relates to the prediction of pressure / temperature histories in the various subcompartments of a con-tainment building. A more complex aspect relates to the prediction of fluid mechanics related phenomena and their impact on local concentrations of hydrogen-and fission products.

2. Implication of this Issue to Regulatory Questions The third regulatory question - how safe are existing plants with respect

'/                       to severe accidents? - can also be posed as: What are the magnitudes and likelihoods of environmental source terms associated with severe accidents?

In order to assess the(e items, the capability is required to analyze the ,

                         " products" directly resulting from the core damaging event (e.g., masses of radioactive and inert material, burnable gas concentrations, energy inputs) which enter into the containment building, the thermal-hydraulic response of the building atmosphere (in terms of pressure and temperature), the transport of materials in this atmosphere, and the response of the con-l                         tainment building itself to these conditions. At issue here is the l

accuracy (or extent of uncertainty) available in present models which consider the second of these items (the building atmosphere's pressure /

                     . temperature response) under severe accident conditions.

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

3. Sub-Issues

                        - Accuracy of pressure / temperature predictions for saturated conditions; in particular, the adequacy of condensation heat transfer models for steel shell containments
!                       - Accuracy of pressure / temperature predictions for the superheated conditions experienced due to hydrogen burns.
 .                      - Effect of convective flow patterns within a compartment and
,                           intercompartmental flow
                        - Extent of moisture condensation on aerosol particles I
                        - Effect of direct aerosol heating.

4. Status of Understanding i

l In general, it is believed that present capabilities to predict contain-rient thermal-hydraulic behavior provide rsasonable, conservative estimates, i However, there are a few areas in which improvement is needed.. I In most accident sequences the pressure buildup in containment is due to vapor formation, and aerosols are released to a saturated containment atmosphere. In these cases the models describing the condensation of water vapor on the aerosol particles are extremely important, since the aerosols are increased in size and change to spherical shapes by this process. These two factors strongly influence the attenuation of aerosols by agglomeration and settling. In the case of containments with smaller volumes, and for sequences in which pressurization is due to hydrogen burn, it is possible to achieve highly superheated atmospheres in which this co'ndensation process will not take place. The agglomeration and settling

3-will then be controlled by the aerosol shape factor. For highly superheated atmospheres, the temperature may be more of a challenge to containment integrity than pressure loadings. Flow patterns within a containment are not very well understood. These flow paths are driven by various forces (blowdown, core / concrete interactions, etc.) during the course of an accident, particularly in small containments, such as BWR drywells. In the final stages, just before containment failure, the flow will be influenced by buoyancy forces due to steam generation. This problem needs to be addressed in more detail. The fraction of fission products airborne at the point of containment failure will be a function of this flow pattern and could thus affect the fraction of fission products leaked to the environment. Current models of fission product settling are based on the assumption of a quiescent containment atmosphere. Finally, direct heating of the containment atmosphere by aerosolized core debris is a potential source of pressure loading for sequences in which the reactor vessel fails at high pressure. There is currently some uncertainty in estimating the magnitude of direct heating. The magnitude of the pressure and temperature rise in containment would depend on the fraction of core material that could be converted to aerosol and dispersed in the containment atmosphere, the efficiency of thermal energy transfer from the aerosol to the gases in containment, and the extent to which the aerosol reacts chemically with oxidants in the atmosphere. The efficiency of thermal energy exchange and the extent of chemical interactions both depend on aerosol size distribution, the mixing of core debris with the containment atmosphere and the residence time in the atmosphere. Resolution of this question will require a better estimate of the amount of aerosol produced, the aerosol size distribution and the residence time in the containment atmosphere. 4 4 6 4

                                                 --.+2_-      _ _ _ - - . - . . - .m_. __.,m- - --- _ . _ . - - -_ _ - . _ _ - - - --

' ~

_ _ _ _ _.____.___ _ _ ~ _ ._ 4-

5. NRC Position The NRC position is that the prediction of pressure and temperature histories in compartments within a containment building is sufficiently

, well understood. This conclusion is based on the availability of everal computer codes to adequately predict compartment average aistories. However, additional research is required to predict convective flow paths, local concentrations and non-unifom temperature gradients within compartments. These local effects are important when considering non-unifom hydrogen concentrations or aerosol behavior. The issue of direct heating has not yet been resolved. Scoping calculations have demonstrated that the failure pressure of a large dry containment could be exceeded if a large fraction of the core were aerosolized and chemical reactions of the core metal were complete. However, there is no current consensus on the extent of coriym dispersal or the completeness of chemical interactions. Furthemore, it is likely that plant specific details of containment geometry will be a determining factor. The NRC staff position is that direct heating is unlikely. Further technical analysis ahd experimentation is , required to resolve the uncertainties. t

6. Continuing Confirmatory Work Benchmark calculations using several containment systems codes against experiments at the HDR facility have been performed or are underway.

These benchmarks will pemit an assessment of the accuracy and precision of these codes. The programs at Sandia and ORNL now in l progress will contribute to our understanding of aerosol generation and behavior. - - I

l 3 5-The issue of direct heating is the subject of intense study by the Containment Loads Working Group.

e O O 6

    .-      -- -           . . .  . .   .      .   .--  ..--        . . .                                      . - -   . . - - . . . -     . . ~

i . . 9 i . StMtARY OF NRC/ CONTRACTOR /IDCOR ISSUE PAPER STATUS t , ISSUE PAPER TITLE NRC/CONTR. LEADS ASSOCIATED IDCOR REPORTS CURRENT NRC/ CONTRACTOR /IDCOR POSITIONS ;

    ~                                                                                                                                                '

1.2.1-Containment J. Rosenthal (NRC) The IDCOR reports dealing Areas of Agreement Thermal and M. Cunningham (NRC) with this issue have not , i Hydraulic R. Barrett (NRC) yet been received. Behavior Areas of Disagreement ] l ! Areas Requiring Further Definition ! l

i 4

i e ] . . . 3 i j I i i . 1

s i -

   ,l
 .                         Issue Number      1.2.2
     .                     Title Rate and Magnitude of Combustible Gas Production Ex-vessel Prepared by: R. Barrett                             Signature of Author:  '/ d j b ' w ,

Draft i 3 V 4/2/84 Date ' Contractor / Consultants W. T. Pratt (BNL), J. Long (NRC) ' Review Process (List here contractor who will supply a statement on issue relative to IDCOR work) Initial Date Branch Chief, I ff./' Asst. Div. Dir., // ' /Fy Division air. d 8R f/h/W , Tech. Series Div. Director ((1 14 vfE C h

       ' ' '.                                                                ~

Date Sent to ACRS . Date Sent to IDCOR

                                                                                          .e e

foiA - s +-en U c.ft e g

~.

( *. ' . . . (si 1.2.2 Rate and Magnitude of Combustible Gas Production Ex-vessel This issue covers the production of hydrogen and carbon monoxide resulting i from reactions of hot core debris with concrete, and hydrogen from core-coolant interactions in the reactor cavity, radiolysis of water and organic compounds, , and corrosion. It also covers the release of flammable gases from the decom-position of organic materials. All thesc gases must be considered, together with the hydrogen released from in-vessel reactions (1.1.2) to assess possible early or late threats to, containment calculated in 1.2.3 and 1.2.4.

1. Description of Issu' ,

Following vessel failure, unoxidized metal in the corium mixture can interact with water in the reactor cavity and produce significant T 3.s quantities of hydrogen as the melt is quenched. If the reactor cavity is dry, or if the melt does not quench for some other reason, interaction of the corium with the underlying concrete will lead to the production of large quantities of water and CO2, which then react with unoxidized metal in the melt, producing significant quantities of combustible C0 and H2 . In either case, the total production of combustible gases following vessel failure can be equal to or greater than the production prior to failure. In addition, smaller amounts of hydrogen can be produced by radiolytic decomposition of water and organic chemicals, and corrosion of metallic surfaces. Radiolysis of organic chemicals can also produce combustible hydrocarbons. The rate at which combustible gases will accumulate inside PWR and BWR containments will determine the probability of formation of flammable or explosive mixtures whose presence can threaten containment G

                                                                                                             - - - - - . - . - , - - , --   --- ---r n -.------
                                                                                         ,    . .-- -,w- .-

[, -7 integrity and safety-related functions of the plant. The rate

s. -

and amount of ex-vessel combustible ges producticn will depend on the mechanisms assumed and the accident scenarios.

2. Implications of the Issue to Regulatory Questions The magnitude of ex-vessel hydrogen production directly affects the likelihood and timing of containment failure by hydrogen burn or detonation. This in turn affects the magnitude of the fission product release and the warning time for evacuation, and thus has a direct bearing on regulatory question 3: How safe are existing plants with respect to severe accidents? Furthermore, for those types of plants for which the rate and magnitude of hydrogen producjion proves to be the dominant threat to containment integrity this issue also impacts. regulatory question 6: Is additional protection for severe accidents needed?

      .a
      . 3.      Sub-issues                          -

The presence or absence of water in the reactor cavity. The effect of concrete composition (limestone vs. basaltic) on C0 production, The cuantity of unoxidized zirconium and structural steel metal in the debris bed. The magnitude and distribution of radioactive materials to containment sumps.

                                                ^                           -

_..=- :- - --

                     ,                        e
   .        s.
       ..                                                                     \

4. Status of Understanding Major uncertainti.es in calculations of the hydrogen production due to quenching of. the molten debris include the initial temperature of the melt, the time required to quench and the quantity of un-oxidized metal in the melt. The IDCOR methodology yields minimal hydrogen hoduction due to this mechanism, because quench is cal-culated to! occur over an interval of less than one minute. However, j under certain conditions, quench times of up to 60 minutes are j possible. Recent Sandia experiments indicate that even during a rapid j quench, as much as 30% of the available unoxidized metal will react to produce hydrogen.

Existing calculations (using CORCON-MODI) of core debris-concrete

         ,g               interaction have yielded as much as 1000 pounds of~ H and 12,000 2

_l! pounds of CO. The major source.of conservatism in these calculations is the large quantity of unoxidized metal assumed to be associated ! with the melt. Recent IDCOR documents assert that the reactor ! vessel will experience local failure rather than a gross melting of the lower head. This result would lead to a prediction of less unoxidized metal in the melt. The atiility to predict generation of radiolytic gases is good as long as the amount of activity and iodine released from the core is known. Determination of these quantities fcr severe accident is a necessary step in calculating the rates and extent of radiolysis. The amount of hydrogen generated by corrosion of metallic surfaces and by decomposition of organic materials is small in comparison l ) b i

    .                                                                                     fn
,     \

with other mechanisms. Although calculations of production from corrosion and decomposition are highly uncertain, this has little impact on the overall prediction of hydrogen production.

5. NRC Staff Position Reasonable estimates of the amount of unoxidized metal in the core debris will be used to estimate hydrogen production. The basis for the IDCOR results indicating local failure of the reactor vessel lower head will be considered in these estimates. Based on its capability to predict the results of recent experiments, the CORCON-M002 code will be used to determine the release of H , C0 and CO from core 2 2 debris-concrete' interactions.

Sufficient information exists to bound the hydrogen production due (-

to quenching the molten debris. Less than 30% of the metallic

component is expected to be oxidized during ex-vessel quench. The best estimate of hydrogen production due to quench is still an j unresolved issue and will depend on the outcome of issues 1.2.5 l (steam explosions) and 1.2.6 (debris coolability). The NRC staff will base its estimates of radiolytic decomposition on the distribution of fission products, and iodine in particular, through the containment and primary system.

6. Continuing Confirmatory Work The CORCON-M001 a.1d CORCON-H002 codes have been developed to study the interactions of core debris with concrete and to determine the rate and magnitude of gas production (Issue 1.2.9). The code includes l

                   .                                  s t'

effects of slurries and crusts and an overlying pool of water. CORCON-N001 is ready for applications. CORCON-MOD 2 will be ready for some applications in 1984 With respect to hydrogen production during quenching of the core debris, there is uncertainty in the rate of production. Solution of that problem depends on the resolution of issues related to ex-vessel steam explosions (1.2.5) and debris coolability (1.2.6). There are no widely accepted methods of determining how much un-oxidized metal should be included in the core debris. Bounding estimates can be made based on accident progression issues (1.1.2, 1.1.4,1.1.7). , Another source of2 if is radiolysis. The distribution of radio-

           .,         nuclides in sump water will be determined by reference of PBF tests and by thermal-hydraulic aTialysis of the release and spread of source term materials in the primary system. Iodine distribution will be determined in source tem studies now underway.

Once these parameters are known reasonably well, the rates of radiological decomposition of sump water can be estimated from existing theory. Standards are available for the determination of the amount of hydrogen produced by corrosion of zine and aluminum. The aluminum rates are thought to be too high and the standard is being revised, but this source can be bracketed satisfactorily. Although no specific standards exist for detemining combustible gas gene. ration from organic materials, we are able to use state-of-the-art. methods to estimate the

amounts produced by thermal and radiological decomposition.

r., b~ Q D '

                                                                                                                                                               ) *
                                                                                                                                                                 .)
                                                                                                                                                                                , ,l j

{, i SUPetARY OF NRC/ CONTRACTOR /IDCOR ISSUE PAPER STATUS -l:

                                                                                                                                                                           .         t' CURRENT NRC/ CONTRACTOR /IDCOR                          i:

ISSUE PAPER TITLE NRC/ CONT. LEADS ASSOC. IDCOR RPTS. POSITIONS

1.2.2 Rate and Magnitude R. J. Barrett/NRC Tech. Rpt. 12.1: Hydrogen Areas of Agreement l' of Combustible Gas Prod. J. Long/NRC Generation During Severe

Neither party treats radiolysis Ex-vessel T. Pratt/8NL Core Damage Accidents. or corrosion as a significant source of hydrogen.

Tech. Rpt. 15.3: Core-

For an unquenched core debris bed, 4 Concrete Interactions virtually all of the metal is l

eventually oxidized via core-conc ' rete interaction. l l i 4 Areas of Disagreement i

IDCOR calculates corium quench over an i

interval of less than a minute, with minimal hydrogen production. NRC believes much longer quench times ] are possible, with a large fraction ,; of the unoxidzed metal reactions. j j i Recent Sandia experiments indicate >'

,                                                                                                                                 that even during rapid quench,                    l-up to 30% of the metal is oxidized.                l-l Areas Requiring Further Definition                     !
                                                                                                                             " An acceptable prescription is needed for estimating.the amount of
                                                                                                                      -           unoxidized metal in the melt at the time of vessel failure, j                                         .

as a function of accident sequence and plant type. i l I i

t '. .'

i p.

( Issue Number 1.2.3 Title Distribution of Combustible Gases, the Likelih od and Consequ ces of Flame Acceleration and Detonations - s Prepared by:J. Larkins/C. Tinkler Signature of Author: Draft i 3 d [ Date 1/27/84 J. Travis, LANL, A. Camp, SNL, M. Berman, SNL. R.Strehlow, Unh Contractor / Consultants of Illinois Review Process (List here contractor who will supply a statement on issue relative to IDCOR work) , Initial Data Branch ief $RMM N E- YJ /uT Asst. Div. Dir.. S-[M (1/ l Division Dir. '% 2./2 8 /go Tech. Series Div. Directo h'I Date Sent to ACRS f Date Sent to IDCOR Probl ems: Resolution of differences with IDCOR can be accomplished thru direct discussions with IDCOR project manager / contractors and NRC staff. A/AR. Coasses;uce : S<NU 'Vt 9ff9-W Barzas. (? O Yb a k r/ $ r f0/h -8+ '*f 2 8

     ..-                   ~R a s nw

[m[ x an,e-4/z3/g C./1 L- . --__ . - _ _ _-_ _ M

o.. , , 12 - p

     \       1.2.3                              Distribution of Combustible Gases, the Likelihood and Consequences of Flame Acceleration and Detonations

1. Description of the Issue Hydrogen gas released during an accident in an LWR containment can stratify, particularily in the absence of forced circulation and if there are significant temperature gradients in the containment. Hydrogen released with steam can form locally high concentrations in the presence of condensing surfaces. Should hydrogen form a locally high concentration, then flame acceleration and detonations could occur.

P

2. Implicction of the Isrues to Regulatory Questions The technical issues considered here arise in the regulatory assessment of containment integrity, or of failure of equipment important for main-taining such integrity. It is important to understand the likelihood for stratification of combustible gases, and the likelihood of flame accelera-tion and detonations, and to be able to assess the consequences of such phenomena. For example, in some accident sequences with the loss of engineered safety features, such as fans or sprays which enhance mixing, it is possible that combustible gases can stratify; with the assumption of a random ignition source, it may be possible to have a flame accelerating from one compartment of the containment to another and making the transition from a simple deflagration to a detonation. Even more simply, a detonable mixture may form in a subcompartment. The likelihood and consequences of such an event need also to be assessed. If flame acceleration or local detonations can be shown to be unlikely, or if the consequences can be shown to be minor, then this technical issue can be closed with no effect on the regulatory assessment. If however, for ,

certain plants and sequences these phenomena are judged relatively likely, and they have significant consequences in terms of plant integrity or the survivability of essential equipment necessary for the safe shutdown of i l 9

                -- - ---,,.,--e--m.,..,------,w-----.---g-             w       -m, , - - - - - -- - , -- w-- .% -     - --w   c r w- yrs.--wr.vm-------------,----

e,. s 1

     \. .

the plant, then the assessment of these effects on containment integrity plays a key role in resolving these issues.

3. Subissues Distribution of Combustible Gases Extent of mixing during primary system blowdown
Extent of Natural Convection in Containment and the Effects of f 3

Temperature Gradients in Containment Release Rates of Hydrogen and Steam From The Primary System and Composition Likelihood and Consequences of Flame Acceleration Effects of Containment Structures and Geometry on Burning Rates Effects of Engineered Safety Features (Ice Beds, Fans, etc.) EffectsofGasComposition(H/ 2Air / Steam)andIgnitionRequirements [' , ] Effects of Flame Acceleration on Safety related equipment Potential For Dynamic Loads From Accelerated Flame Potential for Transition from Deflagration to Detonation Likelihood and Consequences of Detonations Potential for Local or Global Detonations , , Effects of Loads on Containment Shell, Seals, and Penetrations Potential to Generate Missiles Effects on Safety Related Equipment Ignition Requirements versus Gas Composition ! Effects of Steam on Detonation Limits Effects of Geometry on Detonation Progagation 4

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

ps s

4. Status of Understanding Hydrogen mixing and distribution in a U*A containment is sensitive to the hydrogen injection rate, the availability of forced circulation or induced

           ^                                              '

turbulence in the containment (e.g. sprays), the plant geometry and the : initial containment temperature distribution. If fans are operational, i then there will generally be rapid and good mixing in the containment. Sprays also promote good mixing. Therefore, the mixing or stratification of combustible gases is more of a concern for accidents where the fans and sprays are not operational, such as in the case of a loss of electrical power (stationblack-out) accident. High hydrogen injection rate and significant temperature gradients in containment, particularly inverted temperature gradients, can produce significant volumetric differences in hydrogen concentrations. Although the experimental data base for code I comparisions is somewhat limited, such comparisons do afford some level of I confidence in our ability to predict' the distribution of hydrogen in a LWR containment during a postulated severe accident. Currently no additional experimental work is planned in this area, although multicompartment mixing tests, under varying flow conditions would be useful in the validation of analytical models. Additional analytical research for specific containment types will continue. The IDCOR position is that the current data base and analytical models are adequate for modeling hydrogen mixing for most l accident senarios. , i Over the past two years SANDIA National Laboratory has been conducting research to better understand flame acceleration phenomena and to assess the parameters involved in the transition from deflagration to detonation. Flame acceleration occurs during the propagation of a flame around obstacles where the flame front is distorted by the obstacle establishing eddies or vortices which broaden the flame front and yield a high volumetric rate of heat generation; more air is entrained and hence the , f burning rate increases and also the flame front velocity. Flame speeds l can be increased from a few meters per second up to several hundred meters

                           .                       ::=:2_:                    :: -     :::- _=.2
    =
                       .                                              j 1

i i per second depending o.n the distance the flame is allowed to travel and the number of obstacles. Turbulence induced by active fans can also lead i to flame acceleration as noted in tests performed during 1983 in ! Germany and Norway wherein a transition from deflagration to detonation occurred immediately down stream from a fan. The research at SANDIA is complemented by small scale experimental work being done at McGill , University in Canada and analytical or model development at the Combustion Research Institute at Livermore, California. We are rapidly approaching ) the point wherein, for a particular geometry, such as found in a l containment buildin'g, we will be able to predict whether or not flame acceleration would occur and to estimate potential pressure loads. Planned experiments at SANDIA will model specific plant geometries and test if flame acceleration occurs. These tests will assess the effects I of gas composition, ignition source strength and. confinement on flame acceleration. The analytical research should be at a level of development ) in the next year or so that we "can make predictions for some simple reactor compartments. The IDCOR position on flame acceleration is somewhat uncertain, however, their report supports additional research and specifically calls for coordination of current test programs to establish eno/or confirm scaling , laws. I Hydrogen detonations are sensitive to gas composition, ignition source ! strength, geometry, and, to a lesser extent, temperature and pressure. Data ! show a very small widening of the detonation limits with increasing I temperature and pressure. The effect is negligible for the range of pressure and temperature expected in containment during accident l conditions. Direct initiation detonations can occur only if there is l relatively high concentration of hydrogen ( 145) in air in a quiescent i environment with a strong ignition source. Transition from deflagration ! I to detonation can occur when the concentration is lower ( 13%) even

with a weaker ignition source. Diluents such as steam increase the

,.~

(.-

l - i

   .__,_..._.m.,           _ . _ _ _ _ _ _ _ _ _ ,      . _ , , _ _ , _ _ . .      _..           L _ _ _ _ . _ , _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ , - _ . . , .

(, t requisite hydrogen concentration for detonations and also impede the transition from deflagration to detonation. Early work by Shapiro and l Moffette indicates that detonations were impossible for hydrogen mixtures I in which the stea.m concentration is about 35%. Gibbey and Moore computed a much lower steam limit to preclude detonation of hydrogen in air. Using the theory of explosion limits, they conclude that detonations were prevented by only about 11% steam. SANDIA will experimentally verify a lower hydrogen concentration and hydrogen detonation limit in steam during FY 84. The effects of hydrogen concentration and geometry on the propagation of detonation has been studied by SANDIA and McGill over the past two years and pork in this area should reach a reasonable level of understanding in the next year. No significant model development is planned in the study of detonations since the current tool is judged to give answers which are'goed enough for predicting loads from detonations. There is no current plan for studying reaction kinetics associated with s dynamic processes. Analysis in this area will be directed at assessing the potential for local detonations, calculating pressure loads, assessing the affects on structures including the potential for generating missiles, and effects on safety related equipment which might alter the accident sequence and consequences. The NRC is in basic agreement with EPRI that the direct initiation of detonation is unlikely, however, other mechanisms leading to detonations must be understood. Also an assessment of potential detonation sources in various containrient needs to be completed.

5. NRC Position The distribution and mixing of cortustible gases in a nuclear plant containment building are functions of the availability of the fans and sprays which will enhance mixing. In those accidents which involve loss of electrical power (e.g., THLB') this phenomenon will be an important consideration because of the potential for combustible gas pocketing and stratification. Current analytical methods will handle mixing analysis for 'a large number of accident sequences, however, the models have not L

been validated for the effects of multicompartments, especially in cases with significant condensation. None of the current codes can be used to assess the effects of combustion on mixing, which could be significant in multicompartment' containments. In the near tem the current models and codes will be used where appropriate and as more data and improved models become available additional confimatory analysis will be perforined. Our evaluation of this issue relative to the source term, ASEP, and SARRP results in the current NRC position that more work is required in this area. IDCOR representati'.'es in this area should meet with the NRC staff and its contractors to discuss coordination of specific analysis and model improvements needed to close out this issue. a Flame acceleration is strongly dependent on geometry and gas concentration. In earlier licensing assessment of hydrogen control systems for ice i condensers and MARK III plants it was judged that it was unlikely that conditions could be established ~in the plants where flame acceleration posed a significant threat. The accidents covered in these reviews were recoverable degraded core accident scenarios. It was noted in the Sequoyah safety evaluation report that additional research would confinn these earlier decisions. The ongoing research at SANDIA and McGill University will include the likelihood and consequences of this phenomenon both for degraded core and core melt accident scenarios. There will be no near tenn resolution. However, these programs should in the next two years provide an answer to the questions of whether flame acceleration will occur and the consequences of this occurring in various containments. As noted in the IDCOR report, closer coordination with industry representa-tives is needed to discuss scaling and other testing considerations. As with the issue of flame acceleration, detonation were found to be very unlikely in the NRC's review of hydrogen control systems for ice condensers and MARK !!! containments for degraded core accidents. For large dry Y

                                                                   -      '     "* "--- - - -   -       -^   - ^ ^ -

. .-.-.: :: m :-~ - T2:2.-. .

 ; '.

s 1

      .                                                          !s containments a uniform detonable concentration of hydrogen requires l                              practically 100% cladding oxidation and a strong ignition source,
)                               conditions which are judged unlikely in large drys. The issue of
 ;                              detonations generating missiles or failing safety related equipment has l                              not been extensively' considered. For severe accidents arising from the
}                               loss of electrical power, the potential for forming detonable concentrations
)                               is higher than for degraded core accidents and will require a more detailed I                               assessment particularly for MARK III and ice condenser containments.
)                              Detonations can have a significant effect on source terms and accident

, sequences and the N,RC position is that some additional work is warranted in order to reach closure on this issue. IDCOR should work closely with , the NRC to assess potential detonation sources in various containments and

l. , ,

review ongoing work in terms of adequacy and timeliness.

6. Continuing Confimatory Work Work is currently underway at Los Alamos National Laboratory to modify an

,                              existing computer code to better model hydrogen transport in postulated
!                              severe accidents. This code will principally be used for benchmark j                        .      calculations and will include the coupling of hydrogen transport with j                               combustion. A preliminary version will be available in early 1984 and j                               will be applied to hydrogen transport and combustion analysis for MARK-!!!

containments. The work will support the Containment Loads Working Group l and regulatory issues associated with MARK-!!! containments. A few benchmark calculations will be made for the quarter-scale MARK-III tests being sponsored by the Hydrogen Control Owners Group, which will help to I

validate the code. ;

i SANDIA National Laboratory is' currently in the process of shakedown j i testing of the FLAME facility for flame acceleration studies, which in ! j the next year will specifically address questions of flame acceleration in

an ice condenser containment. The work will be coupled with on going i j p, analytical work to address questions of scale or other dimensional affects l

              /
!                                                                                                                      i i

l . i

[ that cannot be addressed in the experiments. There is also ongoing and planned research on flame acceleration in foreign countries which will contribute to a better. understanding of this technical issue. In the next two years the issue of flame acceleration should reasonably be resolved and will close out confirmatory requirements in this area. In FY 84 research on detonation limits in steam currently underway at SANDIA will be completed. Other additional work on the effects of carbon monoxide and dioxide should be completed in early FY 85. This will complete data requirements for analysis of detonation in various gas compositions. Assessments of potential ignition sources for detonations, and the capability of local and global detonations to generate missiles and effect the course of an accident are just getting started. With an accelerated schedule they should be completed in FY 85. t C.. '% . 4 9 0 0 Q ..

       .m                                                                                                       .-      ,' .*

J *

SumARY OF NRC/ CONTRACTOR /IDCOR ISSUE PAPER STATUS CURRENT NRC/ CONTRACTOR /IDCOR ISSUE PAPER TITLE NRC/CONTR. LEADS ASSOCIATED IDCOR RPTS. POSITIONS 1.2.3 - Distribution J. Larkins, RES/DAE 12.2 - Hydrogen Transport Areas of Agreement Combustitle Gases, C. Tinkler, NRR/DSI in Reactor Containment - Direct initiation of detonation ' the Likelihood & A. Camp, SNL is unlikely. Consequences of Flame M. Berman, Sandia 12.3 - Hydrogen Combustion Acceleration & R. Strehlow,U. of Ill. in Reactor Contairpent Areas of Disagreement Detonations J. Travis, LANL- - More work required.on mixing

                                                                              ' especially in absence of force circulation & with multicompart-ments.
                                                                                - Effects of burning on mixing needs to be assessed.
                                                                                - More work on flame acceleration needed particularly to assess plant unique configuration that might be conductive to flame acceleration.
                                                                                - Work to assess effects of engineered safety features to establish or mitigate flame acceleration.
                                                                                - Effects of gas composition &

ignition requirements on flame acceleration.

                                                                                - Effects of flame acceleration on ESFs
                                                                                  & the course of the accident.
                                                                                - Potential for high dynamic loads
              ~

Areas Needing Further Definition

                                                                                - All of areas of disagreement can be resolved with direct discussion with IDCOR & their subcontractors.

6 j , Issue Number 1.2.4 (~N Title Conditions Leading to and Resulting from Deflagration and Diffusion Flames Prepared by: CTinkler/JLarkins Signature of Author Draft i 2 Date 4/2/84 . f . Contractor / Consultants Review Process (List here contractor who will supply a statement on issue relative to IDCOR work) Initial Date Branch Chief U$0Y$ ffl9)$ % Asst. Div. Dir.,_ A h. //ec///' Divisiongir. d 7 )]h , Tech. Series Div. Director 4 ff 7f he , j t. t Date Sent to ACRS - Date Sent to IDCOR

                                                                    .e O

T (0/A 94-931

                                                                                    &l0 v

         >)-                         ,

s. 1.2.4 Conditions Leading to and Resulting from Diffusion Flames and Deflagrations

1. Description of the issue Concomitant with other phenomena associated with severe accidents are the generation and release of large amounts of hydrogen to the containment ,

atmosphere. Dependingonconditionsinsidecontainment(hydrogenandoxygen concentrations, ignition sources), hydrogen combustion will occur in one or more ways, e.g., deflagration, flame acceleration and detonation, or diffusion burning. Because the consequences of these phenomena are different and may significantly affect; the course of an accident, it is important to understand the conditions leading to these different combustion phenomena. Depending upon the conditions under which hydrogen or a hydrogen-steam mixture

            .       are released into containment'it is'possible that hydrogen may start to burn as a diffusion flame. A diffusion flame is one in which the burning rate is controlled by the rate of mixing of oxygen and fuel. For the jet to burn, it is necessary that at some locations the mixture be within flannability limits. Hydrogen deflagrations involve the reaction of hydrogen through the propagation of a burning zone or combustion wave after ignition initiation.

The combustion wave travels subsonically and the pressure loads developed are, for practical purposes, static loads. For certain core melt accidents the interaction of molten material and concrete may result in the production and release of carbon monoxide. The release of this additional combustible gas and its combustion may also influence loading of the containment. .

2. Implication of the !ssue to Regulatory Questions The initial assessment of this issue has led the NRC to the development of additional regulatory requirements for hydrogen contro1' applicable to various l

l

         ,-      [                        ,
                                                        - 22                                         t

(% s reactor containment designs. Because of their relative vulnerability to the containment atmosphere pressurization associated with hydrogen combustion, pressure suppression containments have been required to provide additional hydrogen control features to accommodate the large hydrogen releases associated with postulated degraded core accidents. Large dry containments are judged to be more capable of withstanding the effects of deflagrations or diffusion flames. For more severe accidents the NRC is currently evaluating hydrogen control options. The position on this issue can affect the acceptability of the various hydrogen control systems. If igniter systems were found to reliably ignite only relatively rich mixtures of combustible gas in air and steam under severe accident conditions, thesi other options for hydrogen control previously rejected by utilities, for example, post-accident inerting, may need to be considered. If the probability for random ignition is sufficiently high under post-accident conditions, then large dry containments may not need to provide C,3 any additional capability. The results of deflagration or diffusion burning of hydrogen when conservatively considered. may also require the modification or improvement of containment heat removal systems.

3. Subissues Def1 aeration

                      -       Composition requirements for propagation
                      -       completeness of burn
                      -       ignition requirements
                      -       propagation velocities
                      -       propagation between compartments                             *
                      -       effects of ESFs
                      -       effects of aerosols, steam condensation (fog)
                      -       effects of carbon monoxide                        -
    -.                -       impact on vital equipment L

,- .. o p .

\.

Diffusion Burning

                   -     ignition requ.irements versus composition completeness of combustion versus composition
                   -     conditions leading to standing flames
                   -    burning velocity as a function of composition and turbulence
                  -      impact on vital equipment                                       ,

4. Status of Understanding i

In order that substantial deflagration of hydrogen take place, the gaseous mixture must be flasamble, and an ignition source must be present. For a mixture of flammable gases such' as hydrogen and air, the fleemability limits are defined as the limiting concentrations of fuel, at a given temperature and pressure, in which a flame can be propagated indefinitely. Limits for j upward propagation of flames are wider than those for downward propagation. Limits of horizontal propagation are between those for upward and downward propagation. R - For hydrogen and air mixtures, the flammability limits of Coward and Jones are still generally accepted. Lower limit values for hydrogen flammability in air saturated with water vapor at room temperature and pressure are:

                  -     4.15 H for WPward propagation 2
                  -     6.05 for horizontal propagation
                  -     g.05 for downward propagation

, In reactor accidents the conditions inside containment prior to hydrogen combustion may include elevated temperature, elevated pressure, and the presence of steam. The flanzability limits widen with increasing temperature. J .

l , ,

At 212*F (100*C) the lower limit for downward propagation is approximately 8.85. Thus, in the temperature range of interest, the widening of the downward propagatton limits is small. If the containment atmosphere is altered by the addition of carbon dioxide, steam, nitrogen, or other diluent, the lower flammability limit will increase slowly with additional diluent, while the upper flasambility limit will drop more rapidly. With continued increase in diluent concentration the two limits approach one another untt1 they meet and the atmosphere is inerted. Approxi-mately 555 steam is Wquired to inert a mixture of hydrogen, air and steam. The effects of steam on flassability are seen to vary between test programs the variability of the results has not yet been adequately addressed. . The effects of aerosols and condensed steam in the. atmosphere are not clearly understood with regards to their impact on flasmability and flame J' propagation. Futhermore, there ils little or no data available to address quaternary mixtures, such as hydrogen-air-steam-carbon monoxide. While the research that has been previously conducted has provided con-firmation of combustion parameters, there remains some doubt concerning the applicability of flassability limits, as defined above, to accident conditions. There may be scale effects due to the large size of contain-monts relative to the small scale experimental apparatus. Turbulence, whether induced by, the combustion process itself or by external mechanisms such as fans and sprays, has a significant effect on combustion parameters. To date there has been little confirmation of anticipated offects of these phenomena which are applicable to reactor containments. Combustion can begin either because of an outside ignition source, or because the mixture temperature is above the spontaneous ignition temperature. A stable flame will develop at a distance from the oriff'ce such that the i v

(~ . t turbulent burning velocity is equal to the gas flow velocity. There is evidence to suggest that for a particular set of conditions (temperature, pressure, and composition), there is a minimum orifice diameter of a few millimeters or less. Of particular relevance to Mark !!! containment is the possibility of a hydrogen diffusion flame persisting above the suppression pool. Ongoing research is expected to provide data to resolve questions on the conditions leading to these phenomena as well as the resultant consequences.

5. NRC Position The NRC positio6 on the definition of relevant combustion parameters has principally been the result of evaluation of hydrogen' igniter systems installed in ice condenser and Mark !!! reactor containments. In that C 1) there has been no combustion

testing in a scaled facility representa-tive of a pressure suppression containment; 2) containment hydrogen analyses are perfonned using computer codes which are basically incapable of modeling the complex interaction of fluid flow and combustion process in a pressure suppression containment.

As a result of the above considerations the staff has required that - utilities performing containment hydrogen analysis address the consequences of combustion by perfonning numerous sensitivity studies. These studies investigate the effect of varying key combustion parameters such as flannability Ifmits, propagation speeds and burn completeness. The NRC expects the need to perform such studies will be diminished upon completion of large scale tests. With regard to diffusion flames in Mark !!! containments, the NRC position is that substantive data are lacking to resolve the issue at this time and that ongoing work is'needed to provide bases for the development of a position.

  .       .                       s In the case of dry containments, which do not have igniter systems, it
  ~

must be assumed that a random ignition event occurs under the least favorable circumstances. Thus, worst case pressure depends on identifi-cation of the most severe set of initial conditions.

6. Continuino Confirmatory Work The approach to resolution of this issue has been a coupled effort te define pertinent combustion parameters through testing along with the development of containment analytical nodels, which, using the prescribed parameters, may then be used to predict containment atmosphere pressure and temperature transients.

Much of the res6urces directed at resolving hydrogen control issues has been focused on combustion testing. Early in the application of ignition systems to nuclear plants it was necessary to test the proposed ignitors to detemine their perfomance characteristic ~s and durability in promoting a ignition of various mixtures. The tests, in general, have been carried out in relatively small vessels representing a single volume geometry. 3 Thelargescale(75,000ft vessel)hydrogencombustiontestprogramnow undemay at the Nevada test site will provide useful data early in 1984. This joint program will provide scale confimation and investigation of hydrogen combustion parameters such as 1) requisite hydrogen concentrations forignitionandpropagation,2)completenessofburn,and3) propagation velocities. The Nevada test program will also provide information on the nature of diffusion flames in confined volumes under certain post accident conditions. In addition to the investigation of combustion parameters, the Nevada test program will provide important large scale data against which existing containment analytical models, e.g., HECTR, CLA5!X, COMPARE, may be benchmarked. Testing at the Nevada site was teminated in February 1984 due O

s. to funding limitations. Data reduction and assessment is currently underway and will indicate whether additional large scale testing will be necessary. AdditionaltestingofdiffusionflamesisbeingcarriedoutatSNL(A1246). This testing, which was directed at understanding scaling parameters, was completed in December 1983. The effects of aerosols, generated during a severe accident, on hydrogen recombinaticn is being studied as part of the core melt-concrete interactions program at Sandia. ' Preliminary results indicate recombination will occur under conditions where steam would normally inert the mixture. a As part of the BI!R Hydrogen Control Owners Group Research Program, there are a number of experimental efforts for the near future. This research is designed to investigate combustion behavior in a geometry a'nd configuration peculiar to Mark !!! containments. The owners group has completed testing by igniters in hydrogen rich environments and testing in a 1/20 scale facility representative of the Mark !!! wetwell. Detailed preliminary results of this testing have recently (January 1984) been made available. Future work planned by the BWR Hark !!! owners group includes combustion testing in a i scale facility repre-sentative of a Mark !!! containment. This program is scheduled to provide results by the end of 1984. These tests should provide valuable data for a spectrum of combustion behavior in Mark !!! containments. g. (....

y' _ St#NWtY OF NRC/ CONTRACTOR /IDCOR ISSUE PAPER STATUS ISSUE PAPER TITLE NRC/CONTR. LEADS CURRENT NRC/ CONTRACTOR /IDCOR ASSOCIATED IDCOR RPTS. POSITIONS j 1.2.4 - Conditions C. Tinkler Program Apt. on Task 12.3 Areas of Agreement Leading to & Resulting J. Larkins " Hydrogen Combustion In freus Deflagration and Ongoing or recently completed Reactor Containment Build- research is probably sufficient ' Diffusion Flames ings" to resolve the majority of the s remaining technical questions. ' i Because the IDCOR document does not state explicit ' positions on most of the technical issues, a detailed comparison of IDCOR and NRC positions is not possible. Areas of Disagreement Areas Requiring Further Definition Additional technical exchange between IDCOR and NRC is needed to focus adequacy of available infonnation and the need for additional data. 9

                         '. ~ .

Issue Number 1.2.5 , Title Likelihood and Magnitude of Ex-Vessel Steam Explosions or eam S s Prepared by: J.Rosentha1/R.Barrett Signature of Author: WIL1_ 4 ,~ g Draft i 4

                                                                                 / *"ni .954

Date 11/9/84 Contractor / Consultants W. T. Pratt (BNL). T. Theofanous (Purdue)

Review Process (List here contractor who will supply a statement on issue relative toIDCORwork) tial Date Branch Chief' //!/ M Asst. Div. Dir. Mb o /s //3/p:/ Division Dir. = v ////f f Tech. Series Div. Directo i /(L ' / k - . s Date Sent to ACR$ Date Sent to IDCOR O

                                                                                               ,a 0          0 FotA-r+-1Le oltt 5

                                                        -   29 -

1.2.5 Likelihood and Maenitude of Ex-vessel Steam Explosions or Tteam 5 pikes

1. Issue Description This phenomenological issue focuses on the initial interactions of hot core materials as they are released from the reactor vessel with any water that might be present in the region immediately below the vessel. The most significant question associated with this issue relates to the magnitude of the containment loading during these interactions. This loading may be due to steam explosions or steam spikes. Steam explosions result from the pro-duction of steam by the efficient exchange of heat from the fragmented hot core debris to the surrounding water. The process is extremely rapid and, if mechanisms (geometric constraints) were available to focus the resulting pressure transients, could produce a shock wave or missile which might lead to containment failure. A steam spike results from quenching of the core over a period of minutes, challenging containment by overpressurization.

5 ~

2. Implication of the Issue to Reeviatory Questions

The likelihood, characteristics and magnitude (or extent) of the core debris / water interactions are essential ingredients for detemining the characteristics and likelihood of containment leakage resulting from steam ' spikes (Issue 1.3.2). If an ex-vessel steam explosion or steam spike results in a high likelihood of significant containment leakage shortly after vessel failure (within severalminutes)thenthishasimportantimplicationsregardingthepotential for fission product release from containment (Issue 1.4). If significant containment leakage occurs shortly after vessel failure then the time for deposition of fission products in containment due to natural processes (Issue 1.4.4)willbelimited. As these processes can result in significant reduction.in the airborne fission products, the characteristics and timing of containment leakage relative to vessel failure have an important influence on the quantity of fission products released to the environment, and thus 8 e 9

  .                                                                    directly relate to regulatory question 3: How safe are existing plants with respect to severe accidents?

3. Subissues Release of core materials from the primary system:

high pressure vs. low pressure release local vs. gross vessel failure Composition of core materials released from the primary system: high temperature molten release vs. lower temperature slurry release relative quantities of Zircaloy, steel and fuel fraction of metals oxidized Water supply to core debris:

                  -     water present prior to vessel failure water released on top of the core materials after vessel failure supply of water available to the core, debris (quantity, timing)

Plant specific considerations: }., ,

                  -     relationship between area below reactor vessel and

remainder of containment (particularly with respect to containmentsumps) -

                  -     communication paths between area below vessel and containment structures in area below vessel and their influence on debris / water interactions

4. Status of Understanding This phenomenological issue has been the subject of extensive analytical-and experimental investigation over the last several years. The importance of 'the issue depends on the containment design under consideratica.

4ege W '-mmm-m = ame w w

      *hs        ame   se&~WW'                       wg+
 .                                                                                                  Simple and reliable calculations have been made to closely bracket containment pressurization during the first one hour following vessel failure. In all cases the pressures calculated were well below the failure values of large dry containments (Zion type).

The key results for the PWR large dry containment are summarized as follows: for the high-pressure vessel failure case, the quenching will be close to complete.(~100% of the energy in the released corium). Further the quenching will be rapid; that is, tending toward a one minute quench time. Although small steam explosions are expected as part of the quenching process, these explosions are not expected to threaten containment integrity.

                                -       for the low-pressure vessel failure case, the core material will probably quench in a time period ranging from one minute to an hour. There is a possibility, however, that the water will not mix with the corium, and lower rates of steam production over longer periods of time will result.

for these conditions adiabatic as well as non-adiabatic bounding calculations of the containment pressure may be performed with confidence. ,

                                -       the temperature loading of the containment would closely follow the saturation temperature associated with the pressure buildup from the quenching process.

The importance of this issue to other PWR containment concepts depends on plant specific features. In subatmospheric containments the reactor cavity O e e

will be dry for almost all core meltdown accident sequences so that this issue will be unimportant. For PWRs with ice condenser containments, if water is available in the cavity, then the resulting steam spike during ex-vessel core debris / water interactions could be important particularly if there is limited ice available at vessel failure. This sensitivity is partly due to the relatively small volume and low failure pressure of ice condenser containments.A detailed assessn:ent of this issue for ice condenser containments will be undertaken. In general, boiling water reactors (BWRs) have less potential for contain-ment threatening ex-vessel steam explosions and steam spikes for two reasons. First, because of the in-vessel arrangement of core debris and structural supports, the potential for a large portion of the core exiting the vessel over a short period of time is less than for PWRs. In addition, 8WRs have less potential for core materials to contact water ex-vessel than PWRs. The BWR Mark I containment design ensures that core materials released from the reactor vessel will not contact significant quantities of water in j the drywell. Water could reach the core debris by restoration of ECC or by C5operationbuttheseconditionsareunlihelytopromotetherapidmixing necessaryforex-vesselsteamexplosionsokseveresteamspikes. For BWRs with Mark III containments, the region underneath the reactor vessel will be dry for the higher probability transient initiated sequences. In Mark !! containments, water cannot accumulate to significant depths on the diaphragm floor beneath the reactor vessel. Plant specific features determine the likelihood of steam spikes after the core debris has penetrated the diaphragm floor and enters the wetwell. The wetwell contains sufficient water to quench the core without producing a steam spike, but in some designs the core debris would be channeled into a portion of the wetwell confined within the pedestal walls. A steam spike would be possible in that region unless there is efficient heat exchange with the remainder of the wetwell. In other designs the wetwell region within the pedestal is dry, thus precluding steam spikes. i _ _ _ _2. _ _ _._ _ _ _

j. s

        . 9 l

5. NRC Position Ex-vessel steam explosions of sufficient energy to challenge containment j

integrity'are highly unlikely. This is partially due to the lack of a j credible mechanism for producing a missile to breach containment. Steam l explosions which have the potential to disperse corium may occur. l Bounding calculations of steam spikes may be readily and confidently-j calculated, given as input the melt progression, mode of primary vessel i fai fure and amount of water available for quenching. The importance of , l this issue to the six major containment types (BWR Mark I, Mark II, i and Mark III and PWR large dry, subatmospheric and ice condenser) 3 depends on plant specific features. This issue is potentially important to PWRs with large dry containments if water is available in the reactor cavity. However, the high failure pressure of these containments en- ! l

!                  sures that uncertainties regarding ex-vessel core debris / water inter-

) actions do not have a large impact. The importance of this issue to

!~                 PWRs with ice condenser containments remains to be determined. This I                   issue is not important to BWRs with Mark I and III containments. The                     '

./ issue could be important to BWRs with Mark II containments. However,

the importance of the issue depends on plant specific features.

1 i Containment overpressurization due to non-condensible gas generation or hydrogen phenomena is addressed elsewhere. f 1 . ! 6. Continuing Confirmatory Work ! l [ I Under the auspices of the Accident Source Term Program Office (ASTP0) and , the SARP Management Group, a containment load working group was formed. This group, composed of senior staft and consultants, is systematically address-j ing a set of standard problems which are chosen to illuminate strengths and ' j weaknesses of each containment type. Results of this effort will be < l pu'u11shed in NUREG-107g, " Severe Accident Containment Loads." l } ! i 1 l i . l

u_____ _ -_ . _ . _ _

r ___ _. -- __ _ . - __ _ _ _ _ . . . l !' i This effort will establish a standard methodology where possible and l provide a broad consensus view of areas where calculations can be performed with confidence and where uncertainties exist. l The MARCH-CORRAL set of codes is almost universally used by severe I accident analysts. These codes will ultimately be replaced by the integrated code MELCOR. MELCOR will provide a more physically

realistic model of the system than can now be modelled using MARCH-CORRAL. In turn, implicit and explicit uncertainties will be l reduced.

l l A continuing experimental program at Sandia provides a data base for 1 analyses and model development.

I 4 4 9 E e 6

s t . - 35 - - SUPMARY OF NRC/ CONTRACTOR /IDCOR ISSUE PAPER STATUS , CURRENT NRC/ CONTRACTOR /IDCOR ISSUE PAPER TITLE NRC/CONTR. LEADS ASSOCIATED IDCOR RPTS. POSITIONS - 1.2.5 - Likelihood and J. Rosenthal (NRC) -Tech. Rpt. 14.1A; " Key Areas of Arreement

                         - Magnitude of Ex-vessel                                   W. T. Pratt (BNL)                               Penomenological Models for

Ex-vesse' steam explosions which' Assessing Explosive Steam can threaten containment integrity are highly unlikely.

          ,                                                                                                                         Generation Rates"                                                                         ;
                                           ,  ,;                                                                                  -Tech. Rpt. 14.1B; " Key                                                                    l Phenomenological Models for
                                                                                        .-                                          Assessing Non-explosive
                     '~ '                   '

l , Steam Generation Rates" ' Areas of Disagreement

                                      'l                                                                                      ,

, J >

Areas Requiring Further Definition

                                                                                      ,    /                                      f           N
                                                                                                   ,,1 I
                                                                                                               --'          e
                                                                              /         . . .
                                                                     "                         ~'

j'

                                                                                         ~ **'

, e Ie

                                                                        ,                                                                   e
                                       '^

j s ._, s fs 1

   -__-          -        --                                              - h                                                                                                                  ---   ----.,--,,a

                                    . _ ._ .                        _~

O

            ^ ' '
              ,o
         .(.                            .

(. > -

                                                                                                                                   .}*
                                             . , p-                  ,,

ISSUE 1.2.5

                                                                                                                                  .F LIKELIHOOD AND MAGNITUDE OF EX-VESSEL XPLO'STORSOR[EAMSPIKES        D                          ,

t i

1. Implication of the Issue t'o Regulatory Questions

 -                                                                               k~ .~
 !                               This phenomenological issue d4a+s vtth the initial interact ns of t core materials (as they are released from the reactor vessel) wi any water that might be present in the region immediately below the vess                                  . The most significant question associated with this issue relates to the/ magnitude of the containment loading during these interactions.                              The likelih'ood, character-1stics and magnitude (or extent) of the core debris / water interactions are essential ingredients for determining the characteristics and likelihood of .

containment .,.w .leakige resulting from steam spikes (Issue 1.3.2). m i f ~ Joodhof significant containment leakage shortly after vessel failure (withi ' several minutes) then this has important implications regarding the potential for fission product release from containment (Issue 1.4). If significant con-tainment leakage occurs shortly after vessel failure, then the time for depo sition of fission products in containment due to natural processes (Issue 1.4.4) will be limited. As these processes can result in significant reduc-tion in the airborne fission products, the characteristics and timing of con tainment leakage relative to vessel failure have an important influence en the l quantity of fission products released to the environment. t - l In addition, the characteristics of the ex-vessel debris / water interac tions can influence the rate and magnitude of release of radionuclides fromin u[p, Jt' the fuel (ex-vessel) (Issue 1.4.3). Steam explosions were considered WASH-1400yto produce enhanced oxidation of ruthenium and tellurium. Also, if the core materials are not rapidly quenched by contact with water, then exten ,: sive core / concrete interactions would result. These interactions can result in further release of radionuclides from the fuel.

                          .n The initial interactions of core materials with water can influence the size of core debris particles that may be formed. The particle size distri button will in turn influence debris bed coolability in ex-vessel locations (Issue 1.2.6.). Also, the mode and magnitude of the core debris water inter actions will influence debris relocation following vessel failure (Issue 1.2.7). The relocation of the debris will inturn influence the potential for

[u. . 09/28/83 1 ISSUE 1.2.5 1 .- ~.. F*tA TL8 .

                                             ,..   . y..
                                                                ..c.      :
                   '           .                      .e    l_e_

Nf

s , interactions between the debris and the containment shell, floor and internal i r structures (Issue 1.2.8). All of the above issues can influence the long-term F potential for basemat penetration (Issue 1.3.7). However, this issue is (~ likely to be more influenced by the availability of water supply to the core debris. -

                                        }             ISSUE 1.2.5 (Continued)                                     4 All of the above issues can affect the answers to a number of regulatory questions. The most direct influence relates to Question 3, namely, how safe are the existing plants with respect to severe accidents.                   In addition, the issues can affect how the level of protection for severe accidents can be in creased (Regulatory Question 4). The emphasis regarding Question 4 is on "how" the level of protection can be increased . which requires as a base a re-liable estimate of the timing and magnitude of containment leakage. The is sues, and their influence on regulatory questions, will also provide signifi-cant direction as to what additional research or information is needed (Regu                         .

1 story Question 5). The relationship between the above issues and other regulatory questions is less direct. For example, recent assessments of core debris / water interac , tions have determined that plant specific features are extremely important when determining the mode and extent of the interactions. This implies that lant specific (Regulatory Question ions based on these issue .swsy

2. Sub-Issues Release of core materials from the primary system:

high pressure vs low pressure release P ~ localvsg/ossvesselfailure Composition of core materials released from the primary system: high temperature molten release vs lower temperature slurry release relative quantities of zircalloy, steel and fuel fraction of' metals oxidized Water supply to core debris: g ,,.g, 7 Is water present prior to vessel failure and mode of contact with core, __ l{w psuch Ludtr~ i.1 A VMfablt ?

 -(                  09/28/83                                            2                   ISSUE 1.2.5

5 - m

                                 , Is water nieased on top of the core materials after vessel failure?

Is a, continuous supply of water available to the core debris? ISSUEI.2.5(Continued) Plant Specific Considerations: - relationship between area below reactor ves'sel and remainder of con-- _ tainment (particularly with respect to containment sumps) *~ ~~ communication paths between area below vessel and containment " structures in area below vessel and their 1Efluence on debris / water interactions

3. Approach to Resolution i To be supplied by NRC/RES.

4. Status of Understandina This phenomenological issue has been the cafix and experimental investigation over the last several years. subject A signifiof extensive cant portion of this effort has direct application to pressurized water reac tors (PWRs) with large dry containment buildings (in particular, the Zion and Indian Point (Z/IP) facilities). However, applications to other containment designs is now underway. The importance of the issue depends heavily on the containment design under consideration.

) :: In general, boiling water reactors (BWRs) have less potential for core materials to contact water ex-vessel than PWRs. The BWR Mark I containment I

                . design ensures that core materials released from the reactor vessel will not                                       ,,

contact significant quantities of water in the drywell. Water could reach the core debris by restoration of ECC or by CS operation but these conditions are unlikely to promote the rapid mixing necessary for ex vessel steam explosions ' or severe steam spikes. Therefore this issue is not important for BWRs with Mark I containments. The same conc,lusion is also true for BWRs with Mark III containments. containment The region underneath the reactor vessel will be dry, in this l Water could be availablesign, for the higher probability transient initiated sequences. large break in this region if the accident is initiated by a OCA, however, lower prob 111ty than transient PRAs have determined these sequences to be of such initiated events. Consequently, this issue has releva ce to only the lower probability LOCA initiated events in BWRs with Mark III c ntainments. 09/28/83

3 BM l

;~

W W ISSUE I.2.5 abat-Mu sL L -

                                % :.e r & .c. ~ -                      ._

flr L- W

1

                                                              .j         -                         ,

g>- :- - - _ _ _ clearly showin in Figure 1. It should be noted that

           . The effect of heat
       '      the heat sink available into the concrete floor fromNthe corium/ concrete in c.uge           was not included and coyld be co                  syderable. f/wueefjw.t#a1 ef '/4e acti ra       c f . sed w=4 et Jaal muy A fwt We concluded that the temperature loading of the containment would closely

5) follow the saturation from the~' quenching prccess.

6) An additional uncertainty in the assessment is the potential for ~"dirIct" heating of the containment from fine aerosols of molten corium which are blown out of the reactor cavity and cool without quenching in large amounts of water. N

REFERENCES:

1. FITSC Ref.

o, 'An Assessment of Class-9 (Core-Melt)

2. T.G. Theofanous and M. ent Systems', Nuclear Energineering and Accidents for PWR Dry-conta

( NE-81-151, Purdue University, W. Lafayette, Design, Vol. 66, No. 3, 1982 an IN 47907, December 1980. Preliminary Assessment of Core Melt Accide at the Zion and Indian 3. Point' Nuclear Power Plants and Strategies for

                                                                                             'tigating Their Effects,        -

USHRC Report NUREG-0850, November 1981. t i '. l i g 6

                                                                                                          ~ - - -
                                                                                           ~         ~

___~ _ ___ l _1. i _- -

                                     ~                                    ~~
  ?'-,     _

Issue Number 1.2.6 Title Debr.f s Coolability in Ex-Vessel Locations Prepared by: T. J. Walker Signaturt of Author: 77442/M v' Draft i 4 Date 52/16/84 l Contractor / Consultants D. A. Powers Review Process (List here contractor who will supply a statement on issue relative toIDCORwork) Initial Date Branch Chief [ A[/7 /89" Asst. Div. Dir., 8M /1[ff*e Division Dir. _ 2.2,89-Tech.SeriesDiv.Directorb (ht. (f4 I I N' Date Sent to ACRS Date Sent to'IDCOR I Soth- N *18 b- C[t3 l I

                                                                                                                                     ^   -                               -           - - - - -
                                                 . :    -- . - ,. : : r : : :- : T - - - - '                            -
   .y 3                            ..
                                                                               /
                                                                                                             \..

1.2.6 Debris'Coolability In Ex-Vessel Locations

1. Issue Statement Given a severe accident in which a degraded-molten core has failed the i

reactor vessel and is deposited in the reactor cavity, management of the accident depends upon the ability of the containment configuration and plant response, including possible operator actions, to remove the decay heat from the degraded core and establish a long term stable situation which will not challenge the containment.

2. Implication of this Issue to Regulatory Questiens After a degraded reactor core has failed the vessel and the core debris is deposited in the . reactor cavity, the debris may be. cooled by the intro-duction of water from plant systems or by imersion in water that preceeded C.) .

the melt to the reactor cavity. The interaction with water may fragment the debris into relatively small particles. Assuming that an adequate water supply in available for continued cooling, there are two possible ways of challenging containment: ; (1) the debris bed is coolable but the steam generation and/or gas evolution is sufficient to overpressurize the containment or create a temperature or hot steam environment in the con-l tainment'that eventually fails the containment, or ; l l (2) the debris bed cannot be cooled by water in a stable manner : and the threat to the containment integrity would result from core concrete interactions and the resultant production of l noncondensibles. e i

  ,,w.
     ,                          -           u _ _ _ .u . ,
                                                                - - - -              -- -     -- ~-                - - - - - - - - - - - -

e.

s If the debris is coolable, it is possible to create a steam explosion or prompt steam spike of sufficient magnitude to challenge the integrityofsome' containments (Issue 1.2.5). Barring such an event, a coolable debris bed can gradually generate enough steam to challenge containment integrity for scenarios without active or passive pressure suppression. On the other hand, a coolable debris bed will minimize the interaction of In addition, low debris l the core materials with the basemat (Issue 1.3.7). temperatures prevent further oxidation of metals and hence limit the generation i ofcombustiblegases(Issue 1.2.2). Limiting the quantities of combustible gases generated'also decreases the likelihood of containment failure from hydrogen burning (Issue 1.3.2). Finally, by limiting core / concrete inter-actions, the rate and magnitude of release of radionuclides from the fuel,

               ',             in ex-vessel locations, will also be limited.(Issue 1.4.3).

This issue, therefore, is important to several other issues (1.3.7, 1.2.2, 1.3.2 and 1.4.3) which impact several regulatory questions. The most important impact of these issues is on Regulatory Question 3 (How safe are'the existing plants with respect to severe accidents). In addition, if water supply to the core debris is certain and coolability is assured a potential failure mode (basematpenetration)canbeeliminated. The issue of coolability can strongly influence the manner in which the containment will eventually fail, and this has important implications for Regulatory Question 4 (How can the level of protection for severe accidents be increased). All of the above issues will provide input to Regulatory Question 5 regarding the need for additional information or research. Finally, the conditions which will permit a coolable debris bed in ex-vessel locations are plant specific and most strongly depend on a supply of water. This implies that decisions for plant modifications or operator actions to mitigate this issue of debris coolability must be plant specific. l L _ _ _ . _ _ _ .. . . _ _ __ _ _ _ _ - _ _ _

D.

3. Subissues

                        -    Bed geometry and size.
                        -    Homogeneity of the bed - stratification of constituents and size and distribution of bed constituents.
                        -    Physical description: melt, solids, gases; fixed or moving; uniform or varying with time.

Interaction with concrete of other structure; dispersal forces. Chemical reactions and mechanical actions. Heat transfer:, radiation, conduction, convection; direction.

                       -     Initial mode of contact between the core debris and water.
                       -     Crust formation, strength and stability.

4. Status of Understanding Core debris that is deposited in a dry cavity reacts with the water in C],., ,

the concrete and . ablates the concrete with attendant gas and aerosol production. Limited energy is conducted downward into the concrete; the melt continues to heat and reacts witi.'the concrete. Much of the energy is carried off by the gases and aerosols. If, however, the melt enters the water (in a wet cavity) and fragments, the bed may be coolable provided that the particles are not too small and a supply of water continues to be available. If the particles are very fine and the bed reheats, the case is similar to the dry case ixcept for the decontamination effect of the water. This non-coolable case is similar to cooling water being introduced on top of a melt in which the water may boil on top of the melt but removes heat relatively slowly. If a stable crust fonns, little direct cooling by l the water results and the gases and aerosols again must provide the heat transfer mechanism. The aerosols and gases may have fission product constituents and transform a significant portion of the energy source to the upper containment. For all cases, hydrogen and other combustible gas generation, fission product release, and debris dispersal need to be known for further accident analysis. 4

                                                                 ~- -       - ~ ~ ~ ~ ~   ~ ~ ~ - ~ ' - ~ ~
                   ;__ M:1.1. ~ . -          1 --.: : 1 - ~                             -

j ,

 ,O
    \                  All of the above scenario components are based on limited experimental results with simulated melt constituents.      In FY 83, experimental large melt pours (200 kg) of corium were made at temperatures exceeding 2500*C. The Core Melt Technology Program will continue this work in FY 84 and provide an experimental base for corium pours onto concrete and establish the cooling          '

capability of pre or post introduction of water. The ensuing friability of corium particles and thus the coolability of the core debris will also be $ detennined. Instrumentation for measuring heat flux, upward or downward, and crust strength will be included as well as aerosol and gas generation. Models have been deyeloped and incorporated in codes to predict the coolability of debris beds. Verification of these models is based largely on steel melt tests. It appears that the coolability of a postulated debris bed can be predicted. However, the prediction of the bed description is uncertain as reflected in the discussion above. In the event of stable or self-healing

' p,                ' crusts, the formation of a coolable debris bed may require many hours. If the codes predict the corium test retsults reasonably well, minor improvements in heat transfer coefficients, energy partition, and particle size estimates may be all that are necessary for the application of the code to valid reactor plant calculations. The testing is generic, but for low pressure release of core melt, the plant specific effects should be readily calculable. However, 4

high pressure release of core melt and the subsequent dispersion of the core

debris may have a larger dependence on the specific plant geometry. A special

! series of high pressures ejection tests has been initiated and will be extended to one tenth scale to evaluate the dispersion effects of high pressure. This l latter effect is a major consideration under Issue 1.2.7 Debris Relocation Following Vessel Failure. l The development of all of the above information must be made as a function of time to assess the containment threat and the consequent radiological release. l A major difference in the NRC versus IDCOR time dependence is the result of the l IDCOR hypothetical quench mechanism for melts that is unsupported by physical observation. In addition the IDCOR position neglects internal temperature l l I

variation in the debris and gas evolution from the non-ablated concrete. The assumptions for partitioning heat among vertical attack, radial attack and upward heat loss need experimental verification as well as the effect (by IDCOR) of the' upward heat loss on overhead structures.

5. NRC Position For a debris bed of known geometry, composition and particle size, a reasonable estimate of the decay heat fraction in the bed may be made and the dry out cooling can be calculated with reasonable assurance. The uncertainties in the early transient portion of the accident lead to less reliable predictions of the debris distribution.

The phenomena involved in cooling a fully molten debris pool with a concrete boundary are sufficiently well ' understood,for severe accident analysis. As the molten pool surface cools, surface crust fomation and bulk freezing are more uncertai6 because crust strength, themal conduct-ivities and boundary heat transfer coefficients are not well known; Thus, the overall quantitative determination of the transients during initial melt quench as it reacts with cavity materials is uncertain. This uncertainty limits the precision of the time estimate for early contain-ment failure. Containment loading may possibly be managed to prevent containment failure or at least to minimize the source term release when the time dependencies are adequately evaluated.

6. Continuing Confirmatory Work The resolution of this issue is being approached by an experimental program to validate models that will be used in system code analyses to study full sized plants with applicable reactor and containment types. Various exper-imental programs are underway to obtain numerical values for analysis.

C;

41 - 4 .) First, the ability of water to cool particulate debris that is internally heated by fission products is being studied in the Annular Core Research

Reactor (ACRR) at Sandia. These tests apply to the debris coolability inside and outside of the reactor vessel. A model of debris coolability as a function l of particle size, bed porosity, pressure, and decay heat rate is being verified and quantified.

i l Second, the response of the degraded core is being studied in a series of out-of-pile tests for two temperature regimes in the Core Melt Technology - l Program: (1)moltencorematerialinteractingwithconcreteand(2) solid debris that is hot enough to melt and ablate concrete. Transient tests of

;                                                molten urania on concrete will start in the first quarter of FY 84. Thermite melts will be inductively sustained as melts interacting with concrete under
!                                                 a pool of water (FY 84) as a basis for similar sustained tests of urania (FY 85).

A separate series of sustained tests utilizing steel hot enough to ablate concrete will be conducted and repeated with 20 to 200 kg masses of UO2 (FY 84). l Geometrical effects to account for variability in plant designs are planned for i FY 86. i . i i i

.l                                                         .

.1 p)

                ',' ~                                                                                                        '3
                      ?

u SU MARY OF NRC/ CONTRACTOR /IDCOR ISSUE PAPER STATUS

                                                                                                                                )
                                                                                                                                            .i ISSUE PAPER TITLE                                                                      CURRENT NRC/ CONTRACTOR /IDCOR NRC/CONTR. LEADS       ASSOCIATED IDCOR RPTS.                  POSITIONS 1.2.6 - Debris Cool-             NRC: T.J. Walker       T.R. 15.3, " Core / Concrete  Areas of Agreement ability in Ex-Vessel           Contr: D.A. Powers     Interactions"                                                                  l
                                                                                             - For a debris bed of known geometry,            i composition and particle size, a reason-able estimate of decay heat in the bed may be made and dryout may be calculated. Debris distribution prediction is less reliable.
                                                                                             - Phenomena involved.in cooling a fully          1 molten debris pool with a concrete               !

boundary are sufficiently well understood. l i

                                                                                             - As the molten pool cools, surface crust formation and bulk freezing are more un-         I' certain because crust strength, thermal          :

conductivities and boundary heat transfer l coefficients are not well known. Areas of Disagreement ; t

                                                                                             - IDCOR invokes a hypothetical quench            '

mechanism for melts that is unsupported by physical observation. Areas Requiring Further Definition

                                                                                            - The above uncertainties limit the precision of the time estimate for early containment failure. But containment I
'                                                                                           loading may be managed to prevent contain-ment failure or to minimize the source term release.

i t

s .

~~

Issue f umber 1.2.7 Title DebHs Relocation Following Vessel Failure Prepared by: T. J. Walker Signature of Author: /8 T d#db Draft i 4 Date 2/16/84 Contractor / Consultants D. A. Powers - SNL Review Process (List here contractor who will supply a statement on issue relative toIDCORwork)

                                                   ,          Initial                    Date Branch Chief             N             2   /1/89"
           .. a Asst. Div. Dir.,       N                 k/7/h%f
             -                               Division Dir.       m.                    2 [4 Tech. Series Div. Directo                           5 4 Date Sent to ACRS i

Date Sent to IDCOR { l l hf ~ c- l t + t

e. .

                 ,    .                          s
     .._                                                                              1.2.7 Debris Relocation Following Vessel Failure

1. Issue Statement Given the severe accident in which degraded core material escapes from the reactor vessel, the potential for core debris relocation affects the coolability of the core and the ability of the plant systems and operator actions to mitigate the threat to containment.

2. Implication of this Issue to Regulatory Questions For certain core me'Itdown sequences the primary system may be at high pressure just before vessel failure. Consequently, as the core materials penetrate the verssel, they will be forcible ejected from the primary system into the region beneath the vessel. If the vessel failure'is local, then the dispersal forces from the primary system depressurization can be very high. In addition, the resulting interactions from molten

core materials contacting water can cause significant dispersal of the core debris. This issue is concerned with this potential for significant relocation of the core debris from the region innediately below the reactor vessel due to the above-mentioned dispersal forces. If the debris from the core is widely distributed, it may be coolable at a steaming rate that will not lead to early containment failure. This issue is most properly related to other issues as noted below (Issues 1.2.6 and 1.2.8) and those issues in turn bear the implication to Regulatory Questions. However, a case with high melt dispersal forces will result in the melt fragmenting into small droplets. Such droplets will heat the containment atmosphere rapidly causing a rapid rise in containment temperature and pressure. Furthermore, the droplets may react exotherr. ally with the containment atmosphere or catalyze the reaction of hydrogen and oxygen. Thus, debris dispersal could cause pressurization of the containment by blowdown of the pressure vessel, hydrogen combustion, or rapid heat transfer from the debris particles to the atmosphere.

         /                      This issue is strongly related to Issue 1.2.8, which deals with the potential interactions of the core debris with the containment shell, floor and internal

r . .

e. , .

 ,                                                            p structures. Issue 1.2.8 in turn impacts the response of the containment andotheressentialequipment(Issue 1.3). Issue 1.2.7 also impacts Issue 1.2.6, which deals with ex-vessel debris bed coolability. If the core debris is concentrated into a deep bed, then coolability is more difficult than for
                                                                              ~

a shallow bed. Significant relocation of debris from the region b'eneath the vessel may result in shallow and hence more easily coolable, debris beds. The eventual debris location could have a significant affect on the survivability of the ESP equipment.

3. Subissues ,

Dispersal Mechanisms: Effect of discharge pressure

                       -      Rapid react' ions with water Chemical reactions Hot liquid drill effect - particularly if the hot melt impinges
          ;                   on a wall.
                                                           ~

Water supply: Melt streaming into a water filled cavity Water released to the cavity after some or all of the melt is in the cavity. ,

4. Status of Understanding In the unlikely event of a general core melt accident, PWR reactor vessels with bottom entrance instrument tubes may fail while the system is pressurized, possibly at pressures approaching normal operating pressure.

Under these circumstances, the core materials will be subjected to extensive ex-vessel dispersal forces. The melt streams may impinge on the cavity walls and act as hot " liquid drills" creating large penetration rates at localized positions or the release may be energetic enough to sweep out through the keyway to the upper containment area. If a BWR plant were to have a core melt accident, both instrument tubes and :ontrol rod drive tubes offer relatively thin metallic surfaces for

  /                  pountfal melt-through while the system is still pressurized. In the Y.!

 ,                                                               (.

case of BWR Mark I containments, corium flow across the drywell floor to the liner would melt the liner quickly and fail the primary containment. Mark II containment designs are very plant specific and vary greatly in design, resulting in large differences in calculated time to failure. For either the PWR or BWR plants, the plant specific conditions must be considered on the basis of specific calculations while the experimental base must define mass and heat transfer, chemical and thermal melt / concrete inter-actions and provide model tests for mechanical transformation and mass motion, i IDCOR has not really addressed this issue but agrees that: (1)debrisdispersal is possible and (2) localized vessel failure is possible. They further agree that this issue'contains the following processes that require additional study and experiment (1) aerosol generation during dispersal, (2) chemical reaction

                          ' of the dispersing debris with the containment atmosphere and (3) the e'ffect l                  of the dispersing debris on hydr, ogen combustion. IDCOR notes that debris dispersal is governed by the configuration of the containment building and that detailed dispersal effects will be considered for the containments under current analysis.

! 5. NRC Position The plant design and accident scenario must be considered for either PWR f or BWR plants in evaluating the nature of debris relocation following vessel failure. The debris configuration is determined by iterative processes which are deterwined by the conservation of momentum and energy. l l Current research investigating key parameters that govern these dispersal mechanisms such as discharge pressure, water / melt interactions, chemical reactions and modeling of these processes will reduce the uncertainties for these mechanisms. Thus, the present position notes these dispersal l ! and recognizes early results from preliminary high pressure (to 2000 psia) ( ejection tests. The results of FY 84 tests will provide scaling information i for 1/20 to 1/10 scale, and aerosol and debris / atmosphere interaction data for scoping evaluations. FY 85 tests will provide better quantitative k

s s data for analysis of specific plants, while FY 86 tests should provide confinnation of geometrical effects.

6. Continuing Confirmatory Work Considerable uncertainty exists as to the operative chemical, physical and themal phenomena that exist as the melt streams from the vessel.

Considering first the melt discharge from the pressurized vessel, an experimental and analytical effort has been established to investigate the above uncertainties at pressures to 2000 psi. The experiments utilize thennitic melts with gas pressurization. Reactor cavities of 1/20 and 1/10 linear scales are used with nelts of 10 and 80 kg. Dry and wet cases will be included. Early results from these tests show that a considerable aerosol source tenn results from pressurized ejection. These aerosols are fonned by the mechanical process of the

 -                      jet steam impinging on the concrete and have essentially the composition of the melt. Thus, the aerosol particles provide 'a maximum contamination C;.).,                   load on the containment and, since the particles are heat sources, a distributed hydrogen ignition source. Equipment survivability within the containment may be challenged. Also steam inerting of the containment may be much less effective for limiting combustion. These tests utilize model keyways and discharges through the keyways will be evaluated.

i

                    /               47                                                      *
                                                                                                                                                ' D,      *
                                                                                              . _ .-)

SIM4AP.Y OF NRC/ CONTRACTOR /IDCOR ISSUE PAPER STATUS - CURRENT NRC/ CONTRACTOR /IDCOR ISSUE PAPER TITLE NRC/CONTR. LEADS ASSOCIATED IDCOR RPTS. POSITIONS 1.2.7. Debris Relocation NRC: T.J. Walker T.R.15.3 " Core / Concrete Areas of Anreement Following Vessel Failure Contr: D.A. Powers Interactions" - Debris d: spersal is possible. ,

                                                                                                                       - Localized vessel failure is             !

possible.

                                                                         ,                                             Areas of Disagreement        .
                                                                                                         ,             - IDCOR has not really addressed   ,

this issue to date. Areas Requiring Further Definition

                                                                                                                       - Aerosol generation during dispersal of debris.
                                                                                                                       - Chemical reactions of debris and containment atmosphere.
                                                                                            ,                          - IDCOR notes that dispersal is governed by the configuration of the containment and that detailed dis-j
                                                                                                  ~

persal effects will be considered for the 7 containments under current j , analysis. l I

     ..        i.

4 e s. Issue Number 1.2.8 Title : Fuel Debris-Containment Shell, Floor and Internal Structure Interactions Prepared by: T. J. Walker Signature of Author: 7N/#a/k

                                                                                .V Draft i     4 Date 02/16/84 Contractor / Consultants Review Process (List here contractor who will supply a statement on issue relative toIDCORwork)

Initial Date Branch Chief [b /7 [Ft/ Asst. Div. Dir., ( , ,

                                                                          / ./f,tf Division Dir.          %              Ah t/69 Tech. Series Div. Directc*
                                                             ~

(4 i h Date Sent to ACRS Date Sent to IDCOR fef k * $Y-jgg C, If

                                                                                              '/ft      _

           ,',                      e
                                                   ~%-

1.2.8 Fuel Debris - Containment Shell. Floor & Internal Structure Interactions

1. Issue Statement Given the improbable severe accident in which degraded core material escapes from the reactor vessel, the interactions of the core debris with the basemat, containment shell and containment internal structures must be understood and evaluated in a quantitative sense to provide realistic analyses of potential containment failures.

2. Implication of the Issue to Regulatory Questions This issue considers the interaction of core debris with the concrete floor or basemat and the containment shell or internal structures. Water may not be present to the extent that it is a major reactant except as the water content of the concrete. "

Analyses and some experimental work, since the Reactor Safety Study, indicate that the generation of noncondensible gases and aerosols from debris / concrete interaction could lead to containment failure before the basement is penetrated. The loading of the containment from the debris / concrete , interaction, could possibly lead to containment failure at interrediate times and consequently a serious external source ters, whereas basemat penetration causes a delayed source term to the environment. The delay in time of release coupled with the filtering effect of geological strata would tend to reduce societal risks. Possible interactions of the core materials with internal c6ntainment structures can have a number of adverse effects. The. interactions could weaken structural members that support essential equipment or provide additional strength to the conteir.ndr.t building. In pressure suppression containrient designs (8WRs and Ice Condensers), the integrity of the suppression pool or ice compartment is important to reduce containment pressure loads and scrub fission product aerosols. Inte'ractions between

   ,1      .

* . . i o the core debris and the walls separating the compartments in these contain-ments could cause a bypass of the pressure suppression function. Recent analyses, utilizing radial and vertical concrete penetration measurements from hot debris, have indicated the potential for such horizontal penetration and failures, depending on the detailed geometry of the containment. In addition, interactions between the core materfajs and essential equipment (such as fan coolers or spray systems) couW effect containment cooling.

Finally, interactions of the core debris with the containment floor can provide a potential containment failure mode (Issue 1.3.7). However, if water can be supplied to the cort debris, the potential exists to prevent these interactions andattendantfailuremodes(Issue 1.2.5). This issue 1,mpacts several other issues and they all impact several regulatory questions. The main impact is on Regulatory Question 3 (How safe are the existing plants with respect to severe accidents).- However, they are also p) (t important to Regulatory Questions 4 and 5 (How can protection for severe accidents be increased and what additional research is required). In addition, the potential for debris relocation and interaction with the containment wall, floor or internal structures is very plant specific (Regulatory Question 2.1).

3. Subissues

                 -     Heat conduction direction and amount.
                 -     Crust formation and stability.     *       *
                 -     Gas generation and effect on the melt pool.
                 -    Water migration in the concrete.
                 -     Concrete structural stability.
                 -    Magnitude and form of the dispersal forces acting on the core debris.
                 -     Communication paths between the reactor cavity and the remainder of the containment building.

4. Status of Understandine The hydrogen released in the melt / concrete interaction is estimated to equal that of the nominal half core zircoloy/ steam reaction. The noncondensible 9

5

    ...               gasesproduced(C0,CO2 and H2 ) could eventually overpressurize the contain-ment. The timing and release rates are not known and must be determined on a mechanistic basis to permit plant specific analyses. In the case of 8WRs the steam suppression pool is assumed to provide a large source term mitigation factor. . Collapse of the drywell internals of 8WRs by interaction with the melt could, however, create a breach in the drywell directly to we confinement butiding, bypassing the suppression pool.

A substantial body of data on melt / concrete interactions has been assembled since the Reactor Safety Study. These data have been used as the basis for the C0RCON model of, melt / concrete interaction and for the VANESA model of aerosol generation. These models have been based largely on steel melt / concrete irJteraction. Few tests have been made with prototypic melts involving significant quantities of U02 . Furthermore, the models do not describe the interactions once solidification of the core debris begins. Nor do the models include feedback of melted 3r failed plant structures p(, .- to the molten pool. The IOC0R work employs simplified models, which are claimed to be sufficient and emphasis is placed on the complex geometric effects of various specific plants. IDCOR is analyzing seven plants currently for geometric effects. The 10COR model for melt / concrete interaction assumes the same rate of attack horizontally and vertically. Experiments usually result in significantly larger rates of attack in the vertical direction. Thus, the IDC0R model could mitigate or enhance basemat penetration or collateral damage to containment structure as a functior, of specific plant design geometry. In addition, the IDCOR model permits heat loss from the pool to overhead structures but does not consider the effect on the overhead structures. Basemat penetration and subsequent solidification of a melt would result in a relatively small leaching rate of fission products to the ground water. However, a pool of water over a crusted melt, could leach out a large quantity

 ,-                of fission products and in turn release the large quantity rapidly upon basemat

f, penetration. The seriousness of such a release is plant specific and depends on the ground water motion.

5. NRC Position .

The interactions of. fuel debris with plant structure must be considered on a plant specific basis. The downward penetration of the basemat is signi-ficant and full penetration cannot be ruled out. Dispersal through the basemat has the advantage of being significantly slower than other models of containrent failure. Degradation of plant structures must be considered and failures of structures which in turn fail the containment system are possible. The degradation of safety equipment, containment paretrations, and instrumentation must be analyzed as a function of the specific plant geometry. Recent tests have shown that debris relocation following high pressure ejection of the melt may be quite extensive and although this aids long tem cooling, fuel debris relocation to the containment shell vicinity could result in ' localized shell failure. Tests which define the important mechanisms have been conducted. Analyses , have been conducted which define the significant interactions among: core melt / concrete interactions, gas ind aerosol evolution from the concrete, gas heating in the concrete and in the melt, gas and aerosol themal inter-actions with the containment atmosphere, radiation from the melt surface to the plant structure, and radial as well as vertical heat conduction and ablation in the concrete. Large scale melt tests must be completed in FY 84 for verified quantitative analyses. The analysis models need to be improved to include feedback from the plant structure, as it heats and melts, to the molten pool. Data for improved analysis models should be available by mid-CY 84. An experimental program is under way that utilizes a corium (U02~ Zr9 ) melt of about 200 kg to interact with concrete. These tests will 2 provide reliable measurements of gas and aerosol releases as well as directional heat fluxes. Com,u rison tests of molten steel will be run D

   .        '.     .                  s to enable careful comparison with the rather considerable body of steel / concrete interaction data. Fission product elements will be doped into the corium to study fission product retention in the melt or release to the containment. The capability to conduct large scale UO 2 melt / concrete tests with sustained heating is being developed.

Sustained tests of hot, solidified, core debris (U0 2 andsteel) interacting with concrete are under way. These tests will extend the data base for the CORCON model to an extent that the accident termination is predictable. The test variables include power, debris geometry and concrete type. A large scale well instrumented test of corium/ concrete will be completed in the second quarter of FY 84. Similar tests of corium/ concrete and water effects will be conducted in the 3rd and 4th quarters of FY 84. The results of these tests will be utilized to define a final series of generic tests for interactions of various concretes / core melts /and water in FY 85. Geometric effects that will permit analyses of specific plants will be evaluated experimentally in FY 86. M9 l

{,

                                                                            ~
                                                                                                               /            ;

SupmARYOFNRC/ CONTRACTOR /IDC'O[ISSHEPAPERSTATUS ' CURRENT NRC/ CONTRACTOR /IDCOR ISSUE PAPER TITLE NRC/CONTR. LEADS ASSOCIATED IDCOR RPTS. POSITIONS 1.2.8 - Fuel Debris - NRC: T. J. Walker T.R.,15.3," Core / Concrete Areas of Agreement Containment Shell, Floor Contr: D.A. Powers Interactions" - Substantial data have been

                 & Internal Structure                                                          assembled since the Reactor Safety Interactions.                                                                 Study and used as the basis for CORCON model of melt / concrete interaction and the VANESA model of aerosol generation.                     i
                                                                                               - The data base is steel melt /    '

l concrete interaction; little data with significant UO 7/ concrete.

                                                                                               - Models do not dettribe the               ,

interactions once solidification begins.

                                                                                               - No feedback of melted or failed
                                                                      ,                        structures to the molten pool.

Areas of Disagreements

                                                                                               - IDCOR employees simpitfied models and places emphasis on complex geometric effects.

Areas Requiring Further Definition

                                                                                               - 1he 200 kg corium melt tests (1984) are required to provide reliable measurements of gas and aerosol releases, directional heat
                               .                                                               fluxes and current strength.
                                                                                               - Fission product retention will be measured in these tests by doping into the corium.
                                                                                               - Sustained heating tests are necessary to predict accident' termination.

e

'g - .. Issue Number 1.2.9 Title Rate and Magnitude of Noncondensible Gas Production Ex-Vessel Prepared by: S. B.Burson Signature of Author: 2 Draft i 3 Date 51/35/84 Contractor / Consultants RCole SNL, KBergeron, SNL, D. Powers SNL, TPratt SNL Review Process (List here contractor who will supply a statement on issue relative to IDCOR work) Initi 1 Date Branch Chief M II s Asst. Div. Dir.. OIb /h [9t[ Division Dir. 2.k [Wf h Tech. Series Div. Director LC Sf4 i I Date Sent to ACRS Date Sent to IDC0R (m Jh. ~~ gb k~' - 50 t A -S +-12.s C.,

                                                                                                                        'lG lll*l

                   .                   4

f. .

1.2.9 Rate and Magnitude of Noncondensible Gas Production Ex-Vessel

1. Description of the Issue The production of. noncondensible gases during a severe accident follows as a consequence of. thermal / chemical interactions between (1) high-temperature core materials and concrete. (2) rretal-water reactions, or (3)thedegradationofcontainmentbuildingequipmentsuchaspaint, insultation, furniture, etc. Scoping studies have shown that only the first two interactions may be significant; the potential gas sources associated ud auilding equipment will not be considered further. The generation of combustible components such as H and C0 is a major factor 2

in assessing the containment inegrity (Issues 1.2.1,1.2.3and1.2.4). Moreover, the tgtal amount and rate of noncendensible gas generaticn does contribute significantly to other issues, such as containment failure due to overpressurizaticn by steam. The partial pressures of all the vapors and gases present in the atmosphere, together with possible inter-C/

    ..g actions between them, must be considered simultaneously in evaluating the instantaneous total containment pressure. Inthelongterm(days),the noncendensible gas by itself, even without the presence of steam, cculd become a significant factor.

2. Implications of the Issue to Regulatory Questions The existing level of protection of a particular plant can be assessed by a comparison between the specific risks to which it is sub'ected and its structural (and treragement) capabilities to survive an accident-imposed insult. f a given threshold for poten-ial, failure due to over-pressurization were to be exceeded, the plant could be considered to need additional accident prevention and mitigation measures. There are a number of weys in which protection against containment building overpressurization can be increased. The first, and most obvious, is to find methods of reducing the intensity,of the sources of overpressurization; another is to install systems which could Icwer the pressure if a given threshold is exceeded.

s

r = y - . 3

                    ,                                                                                                                                  s
      .                                                                                                                                                  +
                                                                                              ,/                                                                         .

i

               .           To make the determination of pressure loading, it is necessary to calculate the pressure trace thit is' expected to characterize any particular accident'                                                                                ,

scenario. To obtain this ;fnformation, in acdf tion > to the approprirte analytical tools.,a knowledge of all of the pressure sources, including non-condensible gases,'must be available.

.                                                                                       -                           1
<                                                                                       i                                    -

v . . , Cctrpared to the phsnomenological-uncertainties asio'ciated with many other safety issues, knowledge about the production and behavior of honcondensible , gases is relatively, advanced. The manner in which the CORCON hode treats the

production of gascs during contirete ablation is believed to be reliable. For the most part, the handling of. partial pressures of noncondensible gases in the atmosphere can be done ytth classical thermodynamic.t:ethods. The.CONTAIN code incorporate,s tKe' contribution of nencendensible gas corrponents in its ' atmosphere mcdele so that' the p.ressure histories of all of the compartments i throughcut the containcant system"can be followed. On 'the other hand, there is great uncertainty associated Qith the. transient processes that determine 3

              ..           the formation of the debribbed.,The shape            ,

i and r.ature of the pool are also strcngly infidenced by bie'irchitecture of th.e reactor cavity. Since these uncerte,inties; accompany the specific iinput data to be processed by the CCECON code, the output-(gas evolution rate's,'etc.), regardless of the reliability ofthemodelingandnumericalanalysis,cjnbenolessuncertainthan-the input. ~

{ If the problem of overpressurization is to be manegec by the installation of i filtered-vent systgms,'furEher research will probably be desirable to idertify acceptable systems to permit depressurization of the containrent building without excessive radfological releases 4 ,

3. Subissues '

3.1. Thermal-HydrnJic Behavior of ' Secondary Containment The most significant' concern related to the production of non-condens'iblq gcses is' their contribution-to the totai' internal

        /~                        pressure that'deielops within the containment system as the' result V.                                                                               d4
                                                                             \

k i s .

                                 - _ _ , . . - - - - - . - - - . . . . . -                             ;___---,_         -,b           . _ - - -           . _-         ,       ,,   ,-

                                                                                  ^
                                                                                                       ^         - ^ -                   -

of accident conditions. The gases of pricary concern are CO, CO '

       '                                                                                                                                   2 and H2 which are, produced by metal-water reactions and as a consequence of concrete ablation by high-temperature core debris.

i The primary source evaluation is derived from an understanding of these processes (treated in the CORCON code) and a knoweldge of the details of the accident scenario. The treatment of noncondensible j gases in the secondary-containment atmosphere differs somewhat among various codes (such as MARCH /NAUA and CONTAIN) so that thermal- ! hydraulic behavior of containment is included. 3.1.1 Hydrecen Production during Debris-Bed Formation Imediately following vessel failure, highly uncertain transient conditions exist. If the reactor cavity is flooded with water, I molten steel and unreacted cladding can decompose the water to produce an intense transient hydrogen source. The phenomenon would be enhanced by possible steam explosions. The process is highly 4 dependent on the mode of vessel failure, the structural details of the specific plant and the accident scenario hypothesized. 3.1.2 Hydrogen Generation from Core-Concrete Interactions The primary ex-vessel source of hydrogen is the chemical interaction between water vapor released from the concrete basemat and high-i temperature metallic components (fuel cladding and structural steel) present in the debris pool. The composition of the core debris, together with the water-release rate, thus determine the hycrogen production. i 3.1.3 Production of CO and CO, During Thermal Ablation of Concrete The production rates of these noncondensible gases depend upon the heat flux from the' overlying debris and the chemical composition of the particular concrete under attack. The latter, of course, is a function of the specific plant being considered. The relative y abundances of the gaseous components may in turn be influenced by U . t . A

    -,e-           ,-,
                                         -----9-.          ,r._, _ , , , .          _ - - ..-- , ~.        --v   - - - - - - ,_ __ -

                                                                                              - 57                                    .

f> the composition of the debris as the gases bubble through it. The CORCON code treats these phenomena mechanistically, but the influence of freezing and crust formaticn is not yet adequately modeled in CORC01. 3.2 Influence of Noncondensible Gases on Condensation of l!ater and Heat i Transfer from the Atmosphere to Containment Builcing Structures 4 The magnitude of overpressurization resulting from steam injected into the containment is reduced by condensation en walls and other internal structures. Condensation is, in turn, highly dependent upon the percentage of noncondensible gases present in the condensing

atmosphere in the CONTAIN code.

3.3 Direct Heating of Containment Atmosphere by Aerosols The high-temperature aerosols generated by the core-concrete interactions (including some radioactive fission products) will ~ heat the containment atmosphere and contribute to pressurization'.

          'h
  ~

The VANESSA code under development at SNL computes the details of

           #                   aerosol release from a moTten-debris pool and will contribute to resolution of this issue.

4. Current Status of Understanding Although other possible sources of noncondensible gases have been considered, such as the thermal decomposition of organic containment l

j building and equipment materials, only the gases produced by the inter-l actions between high-temperature core materials and concrete are l considered in this treatrent. All concretes contain a large amount of water; some is chemically bound in the concrete, while some is entrair,ed i l as free water in the interstices. At various temperatures during thermal - attack, both types of water are liberated so that a large amount of water I vapor accompanies the thermal attack. The water and components of the i molten-debris pool (i.e., core debris plus concrete) react to form

hydrogen, carbon dioxide and carbon monoxide. Depending upon conditions I (

   ~ N./
                    .,     ,    - - - - - - - , , , - - , , , , , - - -             --,-a -       ,-----=~w-~w   ,   
                                                                                                                            ' " ' ' '   ' ~ ~ '   ~"    ~

and the availability qf interactive substances, intact water vapor can al.so escape. At high temperatures, the concrete itself decomposes to produce CO and C0g . Hydrogen combustion aspects are treated in detail elsewhere. Hydrogen gas is included here only as a noncondensible component.

                              .             4.1 Description of the Reference Case Conditions used to provide the sample problem in the CORCON-MODI code are used to establish a conceptual frame of reference for further discussion. The case is selected because, however imprcbable, it represents , credible conditions which might be expected after a whole core meltdown of a large PL'R system. Clearly, the problems associated with potential overpressurization become more severe for those plants having smaller total containrr,ent volumes and/or lower design pressures. At approximately 2.5 hours af ter shutdown, the entire molten core of the reactor is assumed to fall on the dry ficor i[                                                      of the reactor cavity. The cavity is approximately 6 M in diameter.

No water is assumed to reach the debris bed. Thermal ablation of the concrete begins immediately and proceeds indefinitely. A detailed discussion of the pool behavior is not included here, since all that is of concern is the rate of evolution of the noncondensible l gases which leave the concrete. During the first 2-3 hours of attack, I' approximately 4000 kg of CO 2 and 2000 kg of CO will be released. The corresponding H 2 release is approxirrately 80 kg. If the volurae cf the ' containment building is assumed to be approximately 2.5 x 10 6 ft3 (which is taken as representativt of a large cry FL'R contairrr.cr ), the noncondensible gases produced, when corrected to standard temperature and pressure, will pressurize an evacuated containtrent i building to one atmosphere in 3-5 days. The gas evolution rate will fall along with the decay heat evolution; hchever, sufficient gas could eventually Le produced to threaten many containment structures. It is clear that the prettetien of noncendensible gases resultir.g from

.?

7

concrete ablation could, in some cases, fail containcent unless stdps are taken to pretent it. 4.2. Effects of Noncondensible Gas Production 4.2.1 Interfacial' Heat Transfer

It has been established experimentally that the bubbling of gas through the interface between two inmiscible fluids enhances the heat transfer between the layers. This results from two distinct effects. (1) disturbance of the surface, and (2) mass entrainment 4

from the lower layer upward into the. upper layer. The net result of these effects is to improve the transport of heat from the molten fuel to the upper layer of the debris pool. This has the , effect of increasing heat transfer to the upper cavity structures, thus lowering the rate of attack on the concrete at the bottom. 1

               ,                   4.2.2 Contribution of Potential Containment Overpressurization
    -            A' This effect has been cover,,ed in the previous paragraphs and must be
            .i given primary consideration.

f 4.2.3 Influence on Water Condensation One of the most significant accident mitigating phenomena is the condensation of water vapor upon the massive structures and equipment housed within the centainment such as shielding, primary pumps, overhead crane, etc. The rate at which heat can be transferred to these structures is strongly influenced by the arcunt of non-condensible gas present in the atmosphere. This is especially i true for horizontal surfaces. The CONTAlli code incorperates these , effects into its treatment of the thermodynamics of the containment atmosphere. j 8

                                                                                       - -   -.-,---.,.---,,,--m-
                                                                                                                     , - - - - - - - , - , - - +

                                                                                                    -     60 -              -                                                                           !

l 6' 4.3 Current Analytical Capability The CORCON-MOD 2 gede, which will include the treatment of slurries ar.d crusts as well as the effects of an overlying coolant layer, will be operational in early CY 1984. It will provide improved , l quantitative' data on the production rates of all noncondensible l gases that result from concrete ablation. Refinement, testing, ! and verification will continue for some time. The CONTAIN code l already adequately handles the thermodynamic and thermal-hydraulic

l. bchavior of the containment atmosphere.

5. NRC Position i For those accident scenarios which involve molten core-concrete interactions, noncondensible gases. (when combined with other sources of pressure) may be produced in sufficient quantity to pose a threat to containment integrity.

            .                                    The CORCON code provides a method for predicting the release rates of non-
          .[,                                    condensible gases from core-concrete interactions which is sufficiently well developed and validated for use in severe accident analyses. These l                                                 accident analyses must ccr. sider specific reactor containment designs to 4                                                 draw safety conclusions for core-melt accidents.

6. Review of IDCOR Positien & Overview of Future Research Activities The most sigt.ficiant ex-vessel source of noncondensible gases is the !
interaction between high-temperatue core materials and concrete in the reactor cavity. This issue has been treated by the Industry Degraded Core Rulemaking Program (IDCOR) in Technical Report 15.3, " Core-Concrete Interactions". The work was reviewed and evaluated by NRC and ilRC-Contractor staffs and discussed with IDCOR represer.tatives at the Joint NRC-IDCOR Meeting on Accident Phenomenology and Containment Loading, November 29-Decemar 1,1983. The Technical Report 15.3 focuses on a description of IDCOR's core-concrete model (DEC0!!P) and a number of laboratcry-scale simulant experiments used for validation of the code.

The principal points of agreement, disagreement, and unresolved issues

S

                                                                                                                                                                            ,, ~ - - , , - , . , -   --
                         , - , . - - - - , - ~ - - - - . . ,               ,,--.-.,--n.,.,   ,,,.,--.--n.----,--n,.       . , . . . - - - , , , -.--,-,,.,,n.

(~~ addressed by the NRC group are outlined below. Considerable insight was gained by both ID(OR and HRC groups which should previde significant guide.r.ce for current and future research programs. 6.1 Areas of NRC-IDCCR Agreement 6.1.1 Debris bahavior is primarily controlled by a self-adjusting thermal balance based on conservation of energy. The rate at which water vapor and noncondensible gases are prcduced deper.ds dir,ectly upon the rate of thermal ablation of the concrete. 6.1.2 The simpliffed one-dimensional DECOMP medeling of heat transfer would be acceptable if supported by comparison with more detailed analyses. 6.1.3 Details of the expected core-concrete behavior, as well as

            .,                accident consequences, are highly plant specific.
            .. .h
 *       'k[*                 .                              .

6.1.4 fiany of the phenomena involved in debris-bed formation and 4 behavior are insufficiently understood to fomulate unique models. 6.1.5 There is an inadequate data base for adequate validation of any of the core-concrete interaction models, especially in terms of l prototypical temperatures, matcrials and scale. Alsc, experiments conducted under sustained heating conditions are needed. 6.2 Areas of IIRC-IDCOR Disagreement 6.2.1 As a sole calculational tool, the one dimensionality of the DECOMP ncdel limits application of the code to a small number of accident scenarios. The predictions can only be acceptable where the ratio of the lieight to dieneter of the debris pool is small so that radial l ablation is minimal when compared to downward attack.

            ;.N'

                '.                                                /

! 6.2.2 The DCOMP model neglects chemical interactions with the gases evolvedduringaQlationofthecavitywalls. In some plants, a y deep debris pool could develop; because of this assumption, DECOMP could underestimate the ex-vessel hydrogen production by as much as a factor of tuo. 6.2.3 Except for the formation of an upper and lower crust, DECOMP assumes i the kulk pool to be a heterogenous mixture of molten core materials.

Calcu;1ations with both CORCON and WECHSL show that if the pool becomes strat'ified, accelerated ablation of the walls will occur where the metallic (steel and zirconium) layer comes into contact with the concrete.

6.2.4 The DECOMP'model assumes that the cnly source of gas is the direct ablation of concrete; in fact, significant amounts of water vapor and other gases are evolved before' actual ablation begins. -

           ?                                                                                                                      .

J 6.2.5 There is inadequate justification for the assumptions on which the ! partiticn of heat among vertical attack, radial attack and upward loss is based. 6.2.6 IDCOR appears to believe that the final state is a coolable debris bed; this depends critically on the assumption of a hypothetical quench mechanism which is unsupported by physical observations. 6.3 Areas Affecting Safety Research Goals and Planning The items discussed above primarily reflect the current levels of understanding and predictive capabilities with respect to ex-vessel noncondensible gas production. There are a number of possible avenues to mitigation of the threat of overpressurization, e.g.: reduction of the concrete ablation rate by core dispersal and ecclir.g. providing the cor.tainment with a suitable filtered-vent system, and the

      , . ,                                     installation of core-rotention devices. Keretreh in these area,s is 1                                                                                                 .

1 l

                           .           .                                     >                                                                                                                       l
             ~

not complete, and the relative merit of these approaches shouldbeevaluated. The documented and verified CORCON-M002 code will provide adequately reliable pre' dictions of the production rates of noncondensible gases resulting from core-concrete interacticns. At present, there is only one other core-concrete interaction model having a sufficient level of sophistication te previde reaningful ceniparative calculations in support of CORCON verification. The WECHSL code, under development at the KfK, FRG, v. ten corrplated and available, will provide a valuable source of independent computational predictions for comparison with CORCON results. The large-scale sustained heating experiments a.t the Kfr, Beta facility (scheduled to cocinence in mid CY 1984), together with the large-scale quasi-prototypical experiments underway at Sandia, will contribute to the data base needed for the validation g of CORCON, WECHSL and DECOMP. The details of the initial conditions UY , (dictated by the accident scenario), and the structural details of the particular plant must be provided to assess the nature of specific threats to containment. Candidate systems to reduce the containment pressure resulting from accunulated noncondensible gases must be evaluated. \ l 0 9

                                                     ,..,,,----..-----,...~.-..,~_,.,.,----..-+------,--,.r,.,.--,---.ey--..-r_-_..,_-w                                             - - - - - - -,
                   , - - - - . . . - .   .-.,e

[N .O 'l *

          .,.               StMtARY OF NRC/ CONTRACTOR /IDO'. ISSUE PAPER STATUS                                      ,        ,

CURRENT NRC/ CONTRACTOR /IDCOR ISSUE PAPER TITLE NRC/CONTR. LEADS ASSOCIATED IDCOR RPTS. POSITIONS , 1.2.8 - Rate and Magnitude NRC: S.B. Burson T.R. 15.3, " Core-Concrete Areas of Agreement cf Noncondensible Gas J. Long Interactions" - Many phenomena insufficiently : , Production, Ex-Vessel Contr: R. Cole, SNL understood to make final decisions. ! K. Bergeron, SNL - Inadequate experimental data base D. Powers, SNL for code validation. T. Pratt, 8NL - Details of expected core-concrete behavior, as well as accident con-

                                                                       ,          sequences are highly plant specific.
                                                                                  - The simplified one-dimensional
                                                                             -    DECOMP modeling of heat transfer             ;

would be ' acceptable if supported by 1 comparison with mofe detailed analysis Areas of Disagreement

                                                                                  - Assumption that most debris beds wil become coolable is not adequately justified.
                                                                                  - Consideration of only direct concret ablation is insufficient.
                                                                                  - Heat partition assumptions (downware upward, radial) are not adequately
     ..                                                                           justified.

Area Requiring Better Definition

                                                                                  - IDCOR's intended application of the DECOMP model.
                                                                                  - Plans for DECOMP verification and assessment.
     *                                                                            - Quantification of effects on radio-logical source term and containment loading.

    .. . 3       .
 ! =$      ,'                          s

0

( i ISSUE 1.3.1 CHARACTERISTICS AND LIKELIHOOD OF CONTAINMENT LEAKAGE FROM SHOCK LOADS I l 1. Description of the Issue Shock loading in containment building can be caused b'y detonation of , combustible gases or by steam explosions. The effect of shock loading on containment performance is very much dependent on the characterization of the load itself - magnitude and time dependent nature of the load, its local and global distribution 6tc. Shock loads are of very short duration (fraction of a millisecond) and high intensity. Largecontininmentstructuresarecapable of withstanding very large shock load intensities. The characteristics of

       ,)     containment leakage due to shock loads can broadly be categorized into two
     ~

types. If the shock load exceeds the threshold of containment capability, catastrophic failure of containment may occur and the fission product inventory in the containment atmosphere will start to flow into the environment. Second, if the containment cracks, penetrations open, or j similar small leakage paths form due to shock loading, there will be a more gradual leakage into the environment than for a catastrophic failure. (

2. Implication of Issue to Regulatory Questions l

The regulatory questions and the implication of the issue to these i questions are discussed as follows: l a) How safe are existing plants with respect to severe accidents? This regulatory question is directly applicable to the issue, and the two types of containment leakage identified in the preceding l section must be studied from the standpoint of the sequence of likely events and phenomenological considerations for existing plants, fod - t4 'l22 c.li7 w

     - . - .               _                                      =_-     - _ . _ _ -                                     .

. b) How can the level of protection from severe accidents be increased?

This question is related to the issue in that the measures to be j taken to mitigate the conditions generating a shock must be , ; l investigated. L i i. { c) What additional research or information is needed? ! This question is relevant to the issue. The likelihood and con-l sequences of the accident scenarios that lead to containment leakage ' due to shock loading must be carefully considered. i l d) Is additional protection for severe accidents needed or desirable? l If it can be shown that the likelihood of a damaging shock load is j very small, increased level of protection may not be necessary. i l

3. Subissues ,

l The following subissues should also be considered along with the main

issue

      .. }

1 a) Local and global distribution of shock loading.

l b) Amplification of shock loading due to multiple reflections inside

! containment. c) Damage to essential equipment due to shock loading.

Subissue (c) is separately considered as part of Issue Paper 1.3.9.

l

4. Status of Understanding So far only a limited number of analyses have been made to review the gross behavior of containment structures subjected to shock loading. For the Sequoyah containment shock loading was characterized by assuming a six foot '

diameter hydrogen gas cloud. Detonation of this gas cloud resulted in a peak overpressure of 180 psig over a local area. The rise time to the peak load was calculated to be 0.10 millisecond with an additional 0.5 millisecond required for the load to dissipate back to ambient. An analysis was performed I by the staff approximating the containment response in a breathing mode in the f-- circumferential direction. b

                                                                                        --------o-+-_     --, _- ___ --

o. . ,

b\ Damping was conservatively neglected. The results of this analysis indicated

     -          that the effective equivalent static pressure on the steel containment was approximately 14 psig. These simple analyses indicated sufficient reserve strength in the Sequoyah containment to withstand shock loading. However, better characterization of the load itself must await more detailed analyses.

letC Position 5. Consideration of shock loading must be approached from noth the I likelihood of the event, and the relevant phenomenological considerations. ; Containment response / leakage must be based on realistic characterization of the potential loads withidue consideration given to individual containment characteristics.

6. ContinuingConf[rmatoryWork .

p under the severe accident research program (SARP) two areas were \N highlighted in order to provide support for severe accident policy decisions. These two areas are: (1)containmentloadi1g/ challenge,and(2) containment failure / leakage. The characteristics of the loading / challenge will come from

   .          the studies associated with the first area and the containment response in                l terms of leak rates will be estimated under the efforts associated with the second area. Primary challenge to the containment integrity comes from slow pressurization and loads resulting from steam spike and hydrogen burning                  ,

(ref.IssuePaper1.3.2). The dynamic characteristics of containment structures are such that they behave in a static manner to all loads except shock loading. The detailed study of containment response for static loads currently underway will provide information on the overall characteristics of the containment and details of penetrations that influence local behavior. - The next step will be to analyze these details when subjected to shock loads to effectively estimate both types of leakage characterized in Section 1. Characterization of shock loads will come from the efforts under Issue 1.2.6

              " Conditions leading to and resulting from detonation". Several research
,.            programs that provide support for the resolution of this issue will continue j     beyond 1984, and will produce adequate bases.

l J $., t. , i . , . (~

 )    N l

ISSUE 1.3.2 l CHARACTERISTICS AND LIKELIHOOD OF CONTAINMENT LEAKAGE RESULTING - FROM STEAM SPIKES AND/OR HYDROGEN BURNING l 1. Description of the Issue A severe accident s,cenario, which results in significant core degradation, may lead to the formation of steam spikes and/or hydrogen burning in the containment. Containment leakage, if it occurs, is more likely to result from an ex-vessel steam spike rather than an in-vessel steam spike. Snortly after the core melts through the reactor pressure vessel, the molten ) . core may react with a significant amount of water in the cavity to result in a

,j large steam spike. Pressurization during core-debris-water interaction can occur within seconds or minutes depending on the mode of contact. The maxir.um: pressure pulse depends primarily upon the amount of water available for ;

interaction with the molten core debris. For a large dry containment, the dynamic pressure pulse could reach twice the design pressure of the containment. For certain accident sequences the primary system may be at high pressure during the core meltdown. For these sequences, high pressure

ejection of the core debris into the containment atmosphere may occur at vessel failure. Direct heating of the atmosphere by the core debris could l increase the pressure even further.

i In addition, during core melting a large amount of hydrogen may be

generated as a result of steam reacting with Zircaloy present in fuel j cladding, control rods and channel boxes, and also with steel. The hydrogen j may reach sufficiently high concentration to detonate or burn thereby increasing the temperature and pressure within the containment. Hydrogen 1 ,

burning is not an issue for BWR Mark I and II containments because the ,, l containments are inert during operation. The presence of ignitors in PWR , Foth - % -ut {

i ice-condenser and BWR Mark III containments is expected to control hydrogen burning although the local temperatures in the vicinity of flames continue to i be of concern. PWR large dry containments may experience high temperatures , j and pressures within the containment due to hydrogen burning depending upon ! availability of hydrogen released during the core meltdown. However, hydrogen l burns are unlikely to 'directly challenge the structural integrity of a large j dry containment. l i j The basic concern of the issue for either event (steam spikes or hydrogen i burning) is whether sufficient pressure will develop to cause a catastrophic j structural failure of the containment or to form leakage paths to release ! radioactive materials into the environment. i j The'following other issues are related to the subject issue: l i a) Issue 1.2.3, Distribution of combustible gases and conditions , l , leading to and resulting from detonations.-

~1 ~

] b) Issue 1.2.4, Conditions leading to and resulting from deflagrations, i diffusion flames and flame acceleration. i ! c) Issue 1.2.5, Likelihood and magnitude of ex-vessel steam explosions or steam spikes. i

!                                               d)           Issue 1.3.1, characteristics and likelihood of containment leakage
!                                                            from shock loads, i

t i e) Issue 1.3.3, Characteristics and likelihood of containment leakage from slow pressurization. i

!                                               f)           Issue 1.3.5, Characteristics and likelihood of containment leakage from thermal loading.

1 f 5

2. Implication of the Issue to Regulatory Questions

( The regulatory questions and the implication of the issue to these questions are discussed as follows: . a) How safe are'the existing plants with respect to severe accidents? The issue has direct bearing on this question. Since the formation of, steam spikes and hydrogen burning very much depends on the reactor type and subsequent leakage on the containment, this issue is expected to be plant specific. b) How can the level of protection for severe accidents be increased? c) What additional research or information is needed? d) Is additional protection for severe accidents needed or desirable? Implication of the issue and the answer to questions b, e and d depend

upon the result of current research programs related to the severe accident phenomena as described in Sections 1 and 6.

3. Subissues The following subissues should be considered with the issue:

a) What is the ultimate strength of different containment designs? b) Will penetations fail before the structure? c) Will the failure be energetic? d) How large is the leak ares? e) Where will failure occur? .

           ?              f) Would a leak reclose after the pressure dropped?          , , ,

7 j , l l ,

4. Status of Understanding

) (. ~ j Analyses of core meltdown accidents in BWR's with Mark I and Mark II . l containments indicate very high temperatures in the drywell region. As a 1' result, concerns were raised regarding the integrity of seals and penetrations l e.g., drywell head, electrical penetrations, equipment hatch etc. The same ; ! concerns were extended to PWR's because of the potential for high temperatures l above the design limit. l A matter of concern for the BWR Mark 111 containment design is the hydrogen burn behavior (diffusion flames) in the wetwell. Diffusion flames ( could jeopardire penetra, tion integrity, through long-term high temperature j effects. This matter is being investigated further by General Electric ! through 1/4 scale testing. Test results, however, will not be available until ! the end of 1984. l. i A related concern for Mark III BWR's is the availability / effectiveness of hydrogen ignition devices for hydrogen control ~under severe accident + conditions. The decision in this matter would influence containment loading considerations for assessing containment leakage behavior. l For PWR dry containments, hydrogen burning is less of a concern because of the lower possible inventory of hydrogen compared to the large containment l volume. l l It is considered that steam spikes alone may not be that much of a j problem. However, as discussed earlier, for certain accident sequences the j primary system may be at a high pressure during core meltdown and high l pressure ejection of the core debris into the containment atmosphere may occur

l at vesgel failure. The core debris could then directly heat the containnment I atmosphere and, in addition, mMal in the core deris could also oxidize, which j would add chemical energy to the containment atmosphere. Under these l circumstances, the 1,ntegrity of the containment could be challenged.

        )                                                            ,

l . .* ..

                                                             ,                                                                                                                                                                  1 I

i 5. NRC Position i j' . [ There is insufficient information to properly characterize the leakage '

}!                              behavior of a containment, and definitive judgements on the likelihood for such leakage are still contingent on the resolution of outstanding l                               phenomenological concerns regarding accident sequence analysis. Nevertheless,                                                                                                                  :

I some judgements can be made regarding containment leakage behavior under l severe accident conditions. Outstanding phenomenological concerns which ; r directly influence considerations of containment integrity will have to be conservatively treated through parametric analysis to assess the potential

;                               impact on the accident source term.

6. Continuing Confirmatory Work The study of characteristics and likelihood of containment leakage due to '

I the high pressure and temperature resulting from steam spikes and hydrogen ! burning, largely depends upon proper understanding and evaluation of the , s' accident phenomer.a which are being assessed as separate issues as discussed in

                 ,')           Section l'. Therefore, the leakage estimate due to such sudden and sharp rise

in temperature and pressure, awaits the phenomenological characterization of the accident source. However, the containment leakage behavior itself is being studied as discussed below and also in Issue Paper 1.3.3.

Under the auspices of the Accident Source Term Program Office (ASTP0), l the Division of Engineering (NRR) was given the task of developing containment leakage models (i.e., leak area as a function of pressure) to be used in containment analyses, for the purpose of determining, in a more mechanistic way, the leakage behavior of the containment under severe accident conditions. (i.e., leak rate as a function of time). The containment leakage behavior would then be factored into the accident source term estimations. The participants in this effort comprise the ] Containment Performance Work Group (CPWG). There is an on-going, joint effort by CPWG members from the Equipment Qualification Branch (DE) and the Containment Systems Branch (DSI) to develop the leakage models and calculate the containment leakage behavior. 4 I

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

l , i ** .'. , The EQB has engaged several contractors to evaluate the leakage behavior

. b '.

of different penetrations types and fluid line isolation barriers when subjected to increased pressure loads, and assess the integrity of containments under normal operating conditions to support estimates of pre-existing leak area. From this, containment leakage models are being developed. . The CSB has sponsored a program to perform containment response analyses for certain risk dominant accident sequences for represenatative containment . designs, using the containment leakage models. The MARCH code is being used for the analyses with updated information from members of the Containment Loads Work Group (CLWG) using a variety of analytical methods. In addition, confirmatory containment response analysis will be performed using the CONTEMPT code. The containment leak rate as a function of time is the end product of the efforf.. n.. .. . e D 6

e 1 b

ISSdE1.3.3 CHARACTERISTICS AND LIKELIHOOD OF CONTAINMENT LEAKAGE RESULTING FROM SLOW PRESSURIZATION e

1. Description of the Issue The basic design function of a containment is to prevent leakage during pressurization of the containment in a design basis accident. Severe accident conditions can produce pressure and temperature conditions that are much greater in magnitude', and thus provide a more serious challenge to containment integrity.

4 During the early stages of core melting and at the time of vessel failure, steam spikes and hydrogen burning may occur creating a sudden rise in containment temperature and pressure as discussed in Issue Paper 1.3.2. However, if the containment withstands the steam spikes and hydrogen burning, subsequent to core melting and vessel failure, another scenario can be conceived where long term decay heat removal is compromited. Under these circumstances, decay heat and noncondensable gas generation slowly pressurizes the containment until it fails structurally or leakage paths develop. In either case, containment integrity will be lost and radioactive products released to atmosphere.

2. Implication of the Issue to Regulatory Questions The regulatory questions and the implication of the issue to these questions are discussed as follows:

for A - f 4-928 (.) ~. e-M

a) How safe are existing plants with respect to severe accidents? The issue is directly related to this question so far as the slow pressurization due to a severe accident is concerned. However, there are very few containments that are alike. Not only does the thres-hold level of,the overall containment structure vary, but also the leakage characteristics of various containment penetrations vary - between plants. Therefore, the likely outcome may be plant specific improvements. b) How can the level of protection for severe accidents be increased? c) What additional.research or information is needed? d) Is additional protection for severe accidents needed or desirable? The finding of this issue bears directly on the above questions b, e and d.

3. Subissues The following subissues should be considered with the main issue:

a) Pre-existing leak paths and leak areas prior to initiation of accident. b) Degradation of seals and gaskets due to higher temperature and pressure. c) Behavior of seals exposed to steam vs. air.

1 d) Behavior of containment liner plate near discontinuities and l possible liner tear due to large strains across cracks in con-l crete containment wall due to pressurization.

l l p e) Leak paths outside the containment whether direct or through

     \,,                                           other structures or secondary containment.

f) Blockage of leak areas due to aerosol deposition. . g) Blowout of seals following significant structural distortion. h) potential for energetic failure. -

1) Localized versus global failure.

4. Status of Understanding 1

The response of containment building to static internal pressure is well understood. There h'as been some debate as to the level of strain at which the containment structure ceases to ,be functional. However, there is a general !

           .          ' agreement that for concrete containments'available strength is up to the point                                               '

of general yielding in the hoop dire,ction and 2 to 3 times yield strain for steel containments. Beyond this threshold value the containment deformation increases excessively for a very small increment of internal pressure, such as 5 to 10 psig. There may be some plants where specific components have a lower failure threshold such as personnel airlock door frames, purge and vent valves with inflatable seals, or butterfly valves procurred under American Water Association (AWWA) standards. Valves procurred under AWWA standard are currently utilized in Rancho Seco 1. Maine Yankee, H. B. Robinson, and Fort Calhoun. These valves may have a design rating of 25 psi or 75 psi. Valves built to this standard are required to subject the valve body to an internal hydrostatic pressure equivalent to two times *the specified shut off pressure of the valves, while the valve seats are to be designed to provide tight shut off at 25 psi across the disc for class 25 valves and 75 psi across the disc for the class 75 valves. Thus, under severe accident conditions with i l .

                                                                                                             ., . s ..

I

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

                  .                                                                                                l
       -             S i ';                 :                          -

l pressures well above 100 psi these valves will not be offactive. However. l [ .. , 1eakage behavior through penetrations of various types is not understood well l because of a lack of data. In many cases the leak area have to be determined ( on a mechanistic basis, such as calculating leak areas on the basis of maximum l

gaps and clearances between mating surfaces without any sealing material.

l Such leak area would then have to be reduced to account for partial ) effectiveness of seals. Another area that suffers from a lack of data is ] liner plate behavior across cracks in concrete and near discontinuities. Both l ] M C research and research sponsored by EPRI should shed some light on this [ j issue. .. l ! 5. M C Position ( j i 4 A realistic estimate of containment response during slow pressurization l l must incorporate pre-existing leak paths, account for leakage through various l l penetrations as internal pressure and temperature increase, and incorporate I other failure modes such as tearing of liner plate. Because of long lengths !

>                      of welds, stress concentration at various discontinuities and potential for leakage through various penetratons; large dry containments are more likely to              l i

leak prior to catastrophic failure due to over-pressure. 4 l l 6. Continuing Confirmatory Work i ! As discussed in Issue Paper 1.3.1, containment response in terms of leak rate is being estimated under SARP to provide support for severe accident j policy decisions. Six reference plants have been chosen for a detailed study i of leakage potential through various penetrations, and also to study other 4 failure modes such as gross failure, leakage through tearing of liner plate j etc. These reference plants are Zion (large dry containment). Surry (large i dry subatmospheric containment). Sequoyah (ice-condenser containment) Peach Bottom (BWR Mark I containment), Limerick (BWR Mark !! containment), and Grand Gulf (BWR Mark !!! containment). Containments in the reference plants are by no means representative of each family. They have been chosen, however, because of the extensive PRA studies available for them. It may become ,

.; ,. e necessary to study additional containments at a later date. Detailed studies of plant specific and component specific features may be necessary following a determination of desirable plant improvements. Containment leakage estimates , for reference plants due to slow pressurization should be available by the end of 1984. Several research programs that provide support for this issue will continue well beyond 1984, and will produce appropriate test data. Results of leakage behavior of purge and vent valves exposed to steam pressures should be available by the end of 1985. Test results of a 1/8 scale model of a steel containment with representative penetrations will be available by early 1985 from a test program sponsored by the NRC. 9 e 0 0 . s, ge Y b

                .-                                                                                             l ls-                       .

I l !D S. l l r i

                                               .             !$5UE 1.3.4 2
                                                                                                          . i.

CHARACTERISTICS AND LIKELIN000 0F LEAKAGE RESULTING FROM EXTERNAL EVENTS ! r I 1. Descriotion of the issue ! I l l It is possible to construct scenarios in which a containment building might be , affected by an external event. In all but one casa, however, the offacts are [ not espected to be significant. For esemple, it is possible that a heavy. I tornado-driven missile might impact a containment building. However, t 4 containment shield building is designed for a tornado-driven missile impact. Even after impact by a heavier missile, it is extremely unlikely that ; containment performance would be impaired by such a local effect. Another l external event could be a large flood. But the resulting hydraulic pressure

            "          would be below the capacity of the shield building or the containment

[

i structure. The scenario that is of concern involves a large earthquake, l l greater than that considered in the design of the containment, that initiates I

) a severe accident. The dynamic effect of an external emplosion or a tsunami [ is espected to be enveloped by a large earthquake. An earthquake will affect ; j the whole containment system, not just a local area. Here we encounter ( j another aspect of the diversity in containment designs. All containments were . } designed to accommodate the effects of a loss of coolant accident (LOCA) in L ! combinationwiththoseofthesafeshutdownearthquake(55E). No explicit ( i consideration was given to expected performance for more severe conditions. Consequently, different designs can be expected to have different , capabilities, with the differences mainly dependent on design details. The ' fundamental question is this: Will containment performance in a severe ; accident be significantly degraded if the accident was initiated by a large l

earthquake?

1 j (..... Fora- f 4 'l28 l c,/te

;                                                                                                             r

   - l . . ,' ' ,    .
   .            2. Implication of the issue to tenulatory Questions The regulatory questions and the implication of the issue to these questions are discussed as follows:

a) How safe are esisting plants with respect to severe accidents? The issue has 1,ndirect bearing on this regulatory question. As . discussed in Section 1, since a large earthquake is expected to be the only enternal event responsible for possible containment leakage, and since both such an earthquake and structural resistance to it are plant-tpecific, the safety of an existing plant should be evaluated based upon plant-specific parameters. The resolution of this issue will at most generically characteriae the effect due to the esternal event, but will not evaluate the safety of a specific plant. b) How can the, level of protection from severe accidents be increased? c) What additional research or information is needed? C.l.; d) Is additional protection for severe accidents needed or desired? The implication of the issue to questions b and d is espected to be plant-specific. Question c can be better assessed at the end of the current research program as discussed in Section 4.

3. Subissues '

The subissues that should be considered with the main issue are as follows: a) possible impairment of conta[nment function due to large displacements under earthquake loading. , b) Effects of aftershocks on containments undergoing severe accident pressures and temperatures. . l c) Capability of fixed penetrations to accommodate large displacements. l d) Buckling capacity of steel containments under large earthquake } - loadings. ! e) s Survivability of essential equipment under large earthquake loadings, i 9

L. . . . . . .

4. Status of Understandina Investigations into the behavior of containments under severe accident conditions have progressed in the last few years. It it new possible to select, with confidence, a pressure significantly beyond design at which a given c.ontainment will,not fail. predictions of actual failure modes are not currently within the state of-the-art, but significant progress is being -

made. gy FY 1984 it sho'uld be possible, utillains the result of letc and epa! research programs as well as analytical studies by IftC and industry, to predict failure modes under severe accider.'. conditions for any containment.

5. NRC Position The overall question of the risk associated with earthquakes has not yet been resolved. The probability of occurrence of large earthquakes and their potentials for initiating severe accidents are currently under study from both the seismological hazard and nuclear power plant risk perspectives. plant specific studies of probabilistic risk assessment (pAA) submitted to the llRC

   ~'        so far, have indicated that earthquakes can be an important contributor to core melt or early fatalities. These pAA studies have also indicated weak links that can be strengthened on a cost effective basis. However, these weak links continue to be plant specific. A stub to assess margins available in l            Plant seismic designs in order to resist large earthquakes, well beyond the range of the safe shutdown earthquake, is underway. Also, there is a research

![ program to stub seismically induced failure modes of containment. Should this l research program identify important containment failure modes that were not l considered before, these research results will be reviewed carefully and l necessary action will be taken. 1 . l .

l. -

l .* O ..

J

6. Continuing Confirmatory Work

 ,r~ s g                        The first step to resolution is the establishment of a baseline capacity under severe accident loading with no earthquake involvement. Then, the effects of severe earthquakes on different containment designs will be examined to determine which containment functions will be impaired at increasing earthquake' levels. The final step will be to merge the results of these two efforts to determine whether the occurrence of a large earthquake before the containment is subjected to severe accident loads will reduce its capacity or change its failure mode. As a parallel effort, another research project will determine the fragility level for critical equipment when subjected to earthquake loadings.

I e 0 0 en . t h 4

                                    .e 4

'N ,, " e s O en

r . t ISSUE 1.2._5

CHARACTERISTICS AND LIKELIHOOD OF CONTAINMENT LEAKAGE RESULTING FROM THERMAL LOADING: 1. Description of the Issue The effect of thermal loading .on containment response is to increase leakage characteristics due to internal pressure by reducing the effectiveness of seals and increasing ti'e potent 161 for liner cracks through buckling of the <

liner plate. A temperature rise overta short duration is predominantly related to hydrogen Burning whereas a temperiture rise over a longer period is related to decay heat and/or radioactive heating.

2. Implication of'tha Issue to Regulatory Questions The regulatory questions and the implication of the issue to these questions are discussed as follows:

a) How safe r are the existing plantsylth respect to severe accidents? It may well be that ti.ermal . effects are significant only in smaller volume containments. - b) How can the level of protection for severe accidents be increased? A detailed assessment of risk measures of early vs. late fatalities must be made before.the questien of increased pro-tection can be answered. 1

c) What additional research or information is needed? ' l This regulatory question car. be better dealt with at the"end-of the current research programs underway as discussed in Section 6. k- '3 f of A - U4 MI

i c,

1 Q k

               -     -                 /

d) Is additional protection for severe accidents needed or desirable? This question is expected to be answered with the resolution of the thermal loading issue.

3. Subissues The following subissues should be considered with the main issue: -

a) Thermal load due to deposition of radionuclides. b) Non-uniformity of thermal loading. c) High temperatur'e behavior of seal materials. d) Heat transpdrt mechanism between inboard and outboard seals. , e) Survivability of essential instruments.

       ~                                                .

4. Status of Understanding Effect of thermal loading on containment leak behavior is conceptually well understood. However, basic data are necessary to make reliable

! estimates. Until more data are available, thermal effect on containment leak rate would have to be qualitative and somewhat judgmental.

5. NRC Position l Thermal loading in low volume containments, particularly those of Mark I l and Mark II types, can be significant and they must be considered. Several accident sequences must be considered, particularly those that can produce higher pressure in addition to high thermal loads.

1 j ' C 4 t

f .

6. Continuing Confirmatory Work

, As discussed in Issue 1.3.3, containment respone in terms of leak rate is being estimated under SARP to provide support for severe accident policy decisions. Containment leakage estimates will be based on a detailed study of six reference plants discussed in Issue Paper 1.3.3. It must be recognized that thermal loading accentuates leakage rates due to internal pressure by causing degradation of seals, thermal stress etc. Both the internal pressure and temperature are time varying functions caused by a particular event sequency of interest. ,The longer the duration of the temperature environment, the greater is its effect on sealing materials. Also, the higher the temperature environment, the less is the pressure retaining capacity of a given sealing material. A full understanding of containment leakage likelihood due to thermal loading may require investigation of several pressure and temperatture functions of time. Several research programs are underway to address the thermal effects. Test results on electrical

          .           penetration assemblies should be available by the end of 1984. Results of C .,    ,

tests on butterfly valves should be available by the end of 1985. S y

t ISSUE 1.3.6 4 CHARACTERISTICS AND L1KELIH000 0F LEAKAGE RESULTING FROM INTERNAL MISSILES

1. Description of the Issue i

A containment failure in the early stages of a severe accident scenario has implications for both predictions 'of risk associated with the scenario and i possibilities for mitigating the effects of the accident. In both cases, the effects of an early containment failure are detrimental. As far as consequence predictions.are concerned, little benefit can be taken for source term reductions if the containment fails early. Also, in those cases where containment failure is assumed to be ess'entially simultaneous with core 5 failure, steps that might mitigate consequences after core failure are precluded. Containment failures, postulated as a result of internal missiles, fall into the category of early containment failures. In the terminology of the Reactor Safety Study, they were designated " alpha" failure modes. The possibility of these failure modes depends, mainly, on the expected energetics of steam explosions resulting from contact of molten core material with water. There is, currently, some technical disagreement about the nature and magnitude of both in-vessel and ex-vessel steam explosions.

2. Implication of the Issue to Regulatory Questions The following are the regulatory questions:

a) How safe are existing plants with respect to severe accidents? b) How can the level of protection for severe accidents be increased? Tota -SMM

e ...+ ;. . ,

      .-             c) What additional research or information is needed?

d) Is additional protection for severe accidents needed or desirable? The regulatory questions impacted by the issue of internal missiles are associated with specific safety problems pertaining to existing plants and . additional information as needed.

3. Subissues The following subissues should be considered with the main issue:

a) Failure mode of reactor vessel and potential for an in-vessel steam explosion (see Issue 1.1.5, Likelihood and Magnitude of In-vessel Steam Explosions). b) Failure mo(e of reactor vessel as a result of pressurized thermal shock or spontaneous crack propagation. c) Failure mode of reactor vessel'as a result of fuel debris-vessel interaction. ~ d) Missile production in an ex-vessel explosion (see Issue 1.2.7, Likelihood and Magnitude of Ex-vessel Steam Explosions or Steam Spikes). e) Missile production resulting from gas combustion.

4. Status of Understanding The potential thermal energy available in melting fuel is high and may be sufficient for missile generation. This is especially so for in-vessel explosions with the reactor vessel head being the obviously most massive and potentially damaging missile. Uncertainties about the generation of missiles and the effects on containments of missile impacts are small when compared to those associated with estimates of the energy released. For the dominant concern, that of the vessel head, calculations based on the energy absorbing 9

7 l E

      *       ;~'                                                                                )
  ;. s                               ,

capacity of the closure bolts are routine and reliable. For a given energy

/*

t - release there would be no najor argument about whether the head would be retained. For cases in which the head might be expected to become a missile, assumptions would be involved in estimating its velocity and trajectory and its effect upon impact. However, analytical and experimental work performed on turbine and tornado missile questions could be used to reduce uncertainties. A less sound basis exists for predicting whether or not smaller missiles might be generated by an ex-vessel explosion. But bounding estimates would likely show that damage leading to containment failure is j unlikely. .

5. NRC Position 4

The NRC staff has taken the position in assessments of severe accidents at the Zion and Indian Point plants that the probability of failures due to

  ,          , steam explosions is sufficiently low to be disregarded when estimating the consequences of severe accidents.

()M 6. Continuing Confirmatory Work . Resolution of this issue is essentially controlled by resolution of the issues on ex-vessel and in-vessal explosions. Therefore, the need for confirmatory research work on this issue will be decided in the future once the phenomenological characteristics of the potential energy source are properly defined and evaluated. I 1

        'r b'. )                                                                        ..--

' ~

it l 1 ISSUE 1.3.7 i POTENTIAL FOR BASEMAT PENETRATION , l

1. Description of the Issue

There are two intrinsic characteristics of nuclear power reactors that dominate the hazard. The first is the inventory of radioactive fission products that accumulates in the core. Secondly, the fission-product inventory constitutes an, uncontrollable heat source which, when established, cannot be interrupted or modified. Safe management of the core requires a system which prevents escape of the radioactive material itself, while at the same time provides a' pathway by which the thermal energy can be released in some benign fashion. The issue of basemat penetration is a direct consequence

           ,          of the heat-producing characteristic of the core. A severe accident scenario
             ,I can be constructed by assuming penetration of the reactor building concrete
         ~

basemat subsequent to a core meltdown resulting in the following primary hazards: a) Release of radioactive fission products to ground water, b) Release of radioactive fission products to atmosphere (possibly by pathways bypassing transport through the ground). c) Degradation of the containment structural integrity. l l The indirect beneficial effect of basemat penetration is the reduction of the possibility or prevention of direct atmospheric failure modes thereby providing more time to mitigate the accident consequences.

2. Implication of the Issue to Regulatory Questions The regulatory questions and the implication of the issue to these questions are discussed as follows:

a) How safe are the existing plants with respect to severe accidents? Since basemat penetration is one obvious pathway for radiological Fat A W c/13

             .,                                  release to the environment, the issue bears directly on plant safety.

Since construction details vary from plant to plant, a generic answer is not possible. The hazard depends upon the probability of basemat penetration at each plant and this, in turn, depends upon the details of the accident scenario. Thus, the magnitude of the threat depends directly upon the source term that accompanies the penetration. b) How can the level of protection for severe accidents be increased? c) Is additional protection for severe accidents needed or desirable?

                                  .             For this issue, both the above two questions can be contemplated simultaneously.            If the source term that accompanies basemat penetration exc'eeds an acceptable level, additional protection is needed. If the acceptable threshold is exceeded, the question of back-fit becomes relevant.

d) What additional research or information is needed? For a given plant design and a postulated accident scenario, the present analytical tools, t'aken together with the results expected from research in progress, can answer most of the technical questions.

3. Subissues i The following.subissues should be considered with the main issue:

a) Long term behavior of concrete attack after freezing of the core debris, b) Effectiveness of cooling the upper surface of debris. c) Debris bed coolability during concrete attack, d) Importance of liquid pathways.

4. Status of Understanding The physical processes involved in the corium/ concrete interactions are reasonably well understood through research in the USA and West Germany. The t . . . - _ - _ - . - . - - - - .

   ,,2         .

CORCON computer code (USA) and the WECHSL code (West Germany) can adequately ( model corium/ concrete interactions prior to significant solidification of the molten debris. Beyond this point, independent calculations may be performed to estimate solid corium penetrating concrete. Future experimental work is being directed to understand the way the solid attack occurs. The potential for btsemat penetration is extremely accident-sequence and - containment -design dependent. For example, high pressure ejection of the melt will disperse the core debris and effectively prevent debris concentration and minimize potential for penetration. For low pressure sequences in containment with a confined space under the vessel, the potential for extensive core / concrete interactions is greater, and under these circumstances, basemat penetration cannot be dismissed depending upon the thickness of the basemat.

5. NRC Position From the foregoing, it is apparent t' hat the basic position of the RC e with respect to the potential for basemat penetration can be specified at this time both for present and future cases. The analysis is based on first principles, the conservation of energy, and current state-of-the-art knowledge or sne tneraal ablation of concrete. It'can be stated that, unless the high-temperature core debris produced by a severe accident can be dispersed and cooled, the possibility of basemat penetration can not be
dismissed, and the consequences should be factored in severe accident l assessment analyses on a plant-specific basis.: 6. Continuing Confirmatory Work It is convenient to consider the issue of basemat penetration in two phases: 1) the transient, or debris-bed formation phase, and 2) the quasi steady-state phase. The core debris may or may not be coolable, and this question depends upon a large number of factors, such as the mode of vessel failure, the physical configuration and construction details of the reactor cavity, and the status of other plant systems at the time of the accident. A -

number of computer codes are being developed or advanced that are expected to l l ._ _ . ..

analytically predict the factors leading to, and the behavior of basemat

        'N s            penetration as briefly addressed as follows:

a) ELPROG computer code treats melt progression in-vessel up to and including vessel failure. b) MEDICI addresses the early as well as long-term cavity phenomenon, c) CORCON-M002 concernsthe concrete ablation rate, the heat partition of the debris bed, and the gas evolution rate. d) VANESA code computes the fission-product release from molten pooljs. MELPROG together with MEDICI will provide a mechanistic uescription of the evolution and subsequent behavior of most of the phenomena that determine the initial conditions f,or the second phase of the basemat penetration issue, the phase treated by CORCON. The experimental work on debris bed coolability and long term corium interactions with materials will continue to produce the data base needed for' validation of the above computer codes. The modeling and experimental research programs that are currently in place should be adequate to provide most of the decision-making gbidance that is needed. O . G .

i

                 =
     \..

ISSUE 1.3.8 RELIABILITY OF EARLY CONTAINMENT ISOLATION

1. Description of the Issue In the event of a reactor accident, the containment is the last barrier to protect against the uncontrolled release of fission products to the environs. Public risk depends significantly on how the containment performs its intended function. Failure to isolate containment will result in the

                    , release of radioactive material, creating a potential haned to publich health. One of the essential elements in assessing accident risk is the ability to predict the leakage behavior of the containment during severe C       j           accidents. The resolution of this 1,ssue is expected to contribute to this need.

2. Implication of the Issue to Regulatory Questions The following are the regulatory questions:

a) How safe are existing plants with respect to severe accidents? b) How can the level of protection for severe accidents be increased? c) What additional research or information is needed? d) Is additional protection for severe accidents needed or desirable? The issue of containment isolation reliability is not directly related to the above regulatory questions. The issue will affect the results of applica-C For A - n -318 ( 6 19

        .' s ..

[ tion studies (accident source term reassessment) which have a direct impact on the regulatory questions. However, the issue can ultimately affect the answers to the above four regulatory and the following related questions: e) What should be considered in this measurement of safety? f) What types of improvements are available? g) What are the information gaps bearing on the severe accident decision?

3. Subissues ,

The following subissues should be considered with the n.11n issue: a a) Existing leakage

1) t) Leak size distributions ,

c) Ability to detect and correct malfunction d) Receiver volume (direct leakage to the environment versus leakage to auxiliary buildings)

4. Status of Understanding Primary containments for comercial nuclear power plants constitute the final barrier for protecting the public from the effects of postulated accidents. They are designed to withstand the pressure and temperature concitions resulting from design basis accidents, while functioning as low leakage barriers against the uncontrolled release of fission products. The basis for current licensing practice is that the containment will perform its intenced function against design basis accidents, i.e., the likelihood for the design leakage rate being exceeded is extremely small. In this regard, ,

y e W-

                                                         /
 .           containment penetrations and fluid systems penetrating containment are equipped with leak testable isolation barriers (redundant barriers being provided for fluid systems). All non-essential fluid system lines are            ,

required to be automatically isolated, with air operated valves closing on loss of air; diverse parameters are sensed for the initiation of containment isolation. The control system design for automatic isolation valves prohibits automatic reopening of the valves on resetting the isolation signal. Since the TM1 accident, the Nuclear Regulatory Commission (NRC) and the industry have expanded ,their consideration of ~ postulated accidents. Greater attention is now being directed toward the more severe accidents, including core-melt events. i Computational techniques are needed to enable the industry and the NRC to determine the relative risk significance of potential nuclear power plant accident sequences. Although great strides have been made in developing the probabilistic risk assessment (PRA) methodology, the containment capability and its leakage behavior under severe accident stresses have not been

     )       sufficiently well defined to be realistically addressed with the PRA methods.      ,

In most PRA studies, the containment is initially assumed to be essentially leak tight, with large amounts of leakage occurring only as the accident pressure inside containment progresses to a very high valua. This is neither realistic nor necessarily conservative since the containment is comprised not only of massive concrete and/or steel structures, but also of numerous penetrations and active components. From a probabilistic viewpoint, the

containment can leak at varying rates, with a wide range of probabilities.

The risk to the public, then depends on the magnitude and timing of the release as well as the associated probability.

5. NRC Position

! It is expected that the program described in the following section for developing a more realistic containment model for use in PRA studies will progress sufficiently by the end of 1984 to permit definitive estimates to be _ _ - _ - _ _ _ _ _ _ - . 1.

l.*' i made of the extent and frequency of pre-existing containment leakage at [' i operating reactors. This information will be used in analyses of containment leakage behavior to support the accident source term reassessment effort. 1

6. Continuing Confirniatory Work d

The Containment Systems Branch (DSI) has sponsored a program to undertake a reliability study of containment isolation systems. The primary objective , of the study is to assist the CSB in monitoring the various efforts in , containment studies and,in consolidating the widespread experience data from i research, testing, and plant operations in the areas of containment integrity and leak tightness so that a realistic evaluation of the containment performance under various accident conditions can be conducted. Another objective is to develop the necessary information for assessing the containment functional performance and for quantifying the probability of partial sy, tem failure leading to a loss of containment iritegrity. The results of the study will be used to evaluate ths consequences of potential y containment leakage and to assess the adequacy of containment isolation system

design criteria.

  \. . .../

b l l

                              ,                       ISSUE 1.3.9 EQUIPMENT AND INSTRUMENTATION SURVIVABILITY

1. Description of the Issue Present regulations re' quire qualification of equipment important to safety for design basis, accident (DBA) conditions. The issue relates to evaluation of the safety-related equipment for survivability and functioning in a further degraded environment resulting from a severe accident. The equipment of concerrf are primarily those that are required to perform the following functions:

a) Monitor and implement actions to arrest a degraded core accident,

and maintain the plant in a safe shutdown condition, b) Monitor and maintain functions that protect containment integrity. The environmental conditions resulting from a core melting accident are discussed in numerous issue papers of the series 1.1. 1.2 and 1.3. Among others, the following accident events are basic concerns for equipment survivabilty: a) Diffusion flames b) Deflagration c) Detonation d) Aerosol loads e) Clogging by accident debris f) Internal missiles g) Radiation exposure (L; . For A - 1+-W c.l15 0

k. ' :' , .

4 h) Humidity or flooding _

1) Chemical attack These events lead to high temperature and pressure conditions inside the containment.

l Implication of the issue to Regulatory Questions

2. -

a) How safe are existing plants with respect to severe accidents? Since numerous. instruments and equipment are required to be functional to control a core melting accident, the issue has direct bearing on the question. For existing plants, the obvious related questions are whether the DBA based criteria used to qualify the existing equipment are adequate to protect against sevpre accidents environments and, if not, whether

                  .                    specific new requirements should be imposed on existing plants.

b) How can the level of protection for severe accidents be increased?

         /                             The issue not only relates-to the evaluation of the existing safety of plants, but also how they can be made safer.

c) What additional research or information is needed? In order to respond to the concerns raised by this issue, the i physical environment within the containment during and after a severe accident must be evaluated and defined. I d) Is additional protection for severe accidents needed or desirable? This question should be considered once the safety of existing plants are evaluated. ~ f

                 , , - , , -  ,m-w      -,,,y-- ,-- ---- ----- , - ,            , --r,     r-------   - - - - , , - , - - -

        * :' ,.       ,                                                                          l

\. ., .

3. Subissues

\,

a) What are the expected environmental conditions for the various severe accident scenarios? b) How much do titey vary from basis accident conditions? c) What equipment are needed for the various severe accident scenarios? d) What is the lik,elihood that the essential equipment would survive the expected conditions? e) What improvement's are needed to assure survivability of essential equipment during and following the severe accident?

4. Status of Understanding So far the industry and staff are expending a significant amount of effort to qualify equipment for design basis accidents. The staff has also evaluated the equipment survivability questions with respect to degraded core accident scenarios for Sequoyah and McGuire. The staff was in agreement with the licensee that equipment will probably survive the event with some exceptions. Additional investigation is underway regarding a few confirmatory items, e.g., scaling effects and local detonations. However, since the bounding environment and the essential equipment. required are not known at this time, it is not possible to determine what additional work, tests and/or analysis, will be required to satisfy the equipment survivability requirement and how long that effort will take.

5. NRC Position The staff is not able to establish any position on this issue at this time. However, when all the information on the bounding environment and equipment required becomes available, the staff would be better able to take a -

f

\/

position on the issue.

r~,,;. . ,

  ,...    .,                        /
  ' f-          6. Continung Confirmatory Work Phase 1 Tasks. Various research activities concerning severe accident scenarios have been initiated by the irdustry as well as NRC. Research is being performed on defining various severe accident scenarios and the environmental conditions associated with them. Based on these, bounding environmental conditions, e.g., temperature, pressure (static as well as             '

pulses created by steam explosion or detonation) and radiation, should be developed. In addition, effort by the system branches of NRC is required to develop lists of essent,ial equipment required for the various severe accident scenarios. These equipment should include all the equipment required to bring the plant to and maintain' it in a safe shutdown condition and those required for containment integrity. These equipment will also include all monitoring equipment needed by the operator to perform his function. e Phase 2 Tasks. Once the phase I is completed, effort must be expended to determine if these pieces of equipment will survive the environment they will j be exposed to. This could be based on past research efforts by the industry and the NRC's research programs. If'the equipment cannot be demonstrated to survive at its present location, then analyses must be done to evaluate whether it will survive if it is relocated, shielded or replaced with other equipment which can survive the environment.

  \ s .e l

a.. , . e He, ,. / ISSUE 1.4.1

RATE AND MAGNTTUDE OF RELEASE OF FISSION FRODUCt5 FROM FUEL (In-Yessel) j

1. _ Description of the Issue This issue deals with how core asterials can become " airborne" (i.e.,

1 entrained in the steam / hydrogen oarrior goo) during severe reasser accidents. l 1 It is intended that gas phase entratnment of fuel materials, core structural ' materials, and associated fission products will be included when referring to " release from fuel." on the other hand, processes which involve escape of fission products from the core sacerials to the condensed coolant (e.g., the leaching process) but which de not involve a subsequent nachanism for / the escaped material to become airborne are not included. Because releases from fuel followed by subsequent transport ("sirborna") g of the radinactiva fission products to Incations external to the plant are

           , .. )
              /           essential steps for reactor accidents to have significant radiological source terms, the scope of this issue is interpreted to include the follow-ing

1. The rate of release of individual fission product isotopes from the .

fuel.

2. The properties of thC teleased fission products and/nr the saferials with which they are associated and which determine their subsequent transport behavior. For fission products released as vapors, this means that-the chemical forms of the released species can be important.

For fission products released as or in particulate material, the par-ticulat,e properties, e.g., sise, shape, density, and perhaps composi-tion which dictate the subsequent transport behavior can be important. n 1 Fo/A - f Y-92 # J c We f 3-

1 , . . :. o .. .

                                                                  ,,g_

i.

3. The release of nonradioactive core materials either as vapors or aerosols that can influence the transport behavior of the fission products or.their carriers through vapor phase chemical reactions which change the mol' acular species involved or through processes such as agglomeration, condensation, adsorption of vapor species on par-ticulates, etc.

i 4

2. Irelication of the Issue to Regulatory Questions Nuclear power plants could advarsely affect public haalth and safety through release of radioactive material to the environment. The extent to

                       , which such a release could occur in postulated severe accidents is referred to as the " source ters" question. Ifhile the process and crit..ria that will be used for making regulatory docisions on questions of reactor safety with respect to severe accidents have not been fully defined. it is clear that any reasonable decision making process will involve consideration of severe secident source term information. Unfortunately, our knowledge of severe accident source terms is incomplete and, while it can and is being improved through research, will remain somewhat incomplete in the time frams that has been established for decisions on the regulatory questions.

At present summary technical input on our knowledge of severe accident source terms can be provided to the decision making level by providing best estimate source terms with associated uncertainties and/or conservative bounding astimates. This requires analysis of how fission products can be released from the fuel and subsequently transported along pathways that result in their eventual escape to the envirooment. Unfortunately, there is not a consensus of the informed technical opinion on the choice or adequacy G' .

i wh of the available Analysis methods. As a result, it can be expected that the decision. making technical information environment will include:

1. MRC's own "best estimate" source terms for selected plants and accident sequences with incomplete information on the uncertainties associated with these estiastes.

2. IDC01's "best estimate" source terms for selected plants and accident 8

sequeness again with incomplete information on the uncertainties asso-eisted with these estimates. Because of the different perspectives involved, it will not be surprising if IDCOR's best estimate source terus disagree significantly from MEC best estimates for similar plants and accident sequences.

           ..                 3. Questions conearning the validity and/or relative merits of the analy-

[ sia approsebes need in the NEC and IDCNt analyses. Ration lecision asking in this environment will require that the decision-makers be informed on the state of knowledge concerning the unresolved underlying technical issues. ' Reesuse analysis of the " Rare of Magnitude of Release of Fission

Prodnets from Puel (In-Yessel)" is an essential and first step in the analy-eis of source terms for existing plants and existing plants with proposed safety improvements, $t follows that technical issues that arise in this portion of the source term analysis process will have implications for deci-sien asking on the questions

1. Now safe arp existing plants with respect to severe accidentet

2. Eow can the level of protection for severe accidents be increased?

e a.__

!, ' i . . ' - . ) j*5 e

i s :x Deficiencies in current or past analyses and analytical methods or in the supporting experimental data base that preclude adequate resolution of important seresolved issues have direct impact on the regulatory question "what additional research or information is needed."

3. Subissues

.l

l j

8 The subissues that we have identified are the following: 1 1 ! 1. Chemical effects - The transport behavior of the released fission pro-ducts depends strongly on their chemical forms. Thermodynamic calcula-tions show the importance of the 2I /82 0 ratio sad predict the formation of certain volatile hydroxides at high system pressure.

2. Structural, eladding, and control rod material effsets - Some fission 2 prodnets react with these meterials. When the fission prodnets react

with the solid surfaces they tend to be fixed (e.g.. To on unozidised 11resloy), but when they reset with vaporised or serosolised materials they tend to be transported further.

3. Specific important release-from-fuel mechanisms - Two such mechanisms are liquefaction (occurring at 1900-2100 K when airconium begins to react with the D0 2) and quenching (the enhanced fission product release observed with rapid cooldown). Quanching with water traps much of the soluble material in the water. CORSon at present averages out these anchanisms over the complete bestup-cooldown eyele except for the absence of rapfd quenching in the data base. The IDCOR model treats release only as a combination escape to the fuel surface followed by vaporisation.

,L .

l

   =              . .

Y _ .. .

o

  .              2 s                                                   5-

4. Geometry changes and scaleup - All fission product release tests have been conducted at relatively small scale. Changes in exposed surface-toblumeratiosasfuelbundlescollapseormaltanystronglyaffect thereleaseo[remainingfissionproducts.

5. Aarnan1 freetion - N feasties of vaports.a risaton products norming or become attached to other particles is importsat. Yapor-form species that condense on or react with fixed surfaces (as in the upper rela-4 tively cool portions of the core or plenum) remain fixed unless later heatup from the core or decay best causes revaporisation. Particle-associated species tend to travel greater distances without significsat delay. N steam hydrogen flow rate can vary from naar zero to very high rates and therefore effects fission product transport from the s

             . ';              core.

a

6. Adequacy of simple sedels - Simple models such as coRs0K and the IDCOR
model attempt to average ont individual release / behavior anchanisms or i

attribute release to calculations one or two controlling mechanisms. Simple models have distinct advantages such as short computing time, ease of understanding of the model by the user or reader, and the opportunity for the casual user to perform his own calculations. The need for scoossiedation of individual effects such as burnup, heatup rate, etc. is not clear.

4. Status of Understanding --

The status,of the surrent technical understanding as it pertains to resolution of each of the subissues identified above is outlined below. An attempt is ande based on a rettew of the ongoing resaarch and the saticipated i V .

                                                              .- f -                                                1

s i - i !f% ~6= i

i timing of pertinent results to present the status as it is likely to exist l l in the time frame of the mid-84 decision. l

1. Chemical effects - Tha knowledge of. chemical effects depends a great I deal on equilibrium thermodynamic calculations. Isamples of experiaan-tal observations are the enhaneament of Ba and Sr release with strongly I

reducing atmospheres, the reduction in Te and Eb release when trapped 4 by moxidised Eircaloy cladding, and occasional slow decomposition of Cs1 under oxidising conditions. The effect of pressure has not been i explored ezeept for a few tests with transient bested fuel conducted in

TREAT. No prpsaura effect was observed.

I

1. Structural, cladding, and control rod asteriale effects - These

! s materials have been shown to reduce the release of some fission pro-ducts (Te, Sb). Most fission product release tests have not included any of these materials except for cladding, so little is known espe-etally about transport with the aerosols formed from these materiale. There is currently a major disagreement concerning the amounts of Ag . and in vaporised. There is also complete disagreement sa to the amount of boron species volatilised and ava'ilable for reaction with easium and other anterials.

3. Specific important release-from-fuel mechanisms - Evidence suggests 4 that liquefaction is important with low burnup fuel but not high burnap l fuel. Quenching appears to be important but the amount of fission pro-1 l ducts not , trapped by the water is unknown.

4. Geometry changes and scaleup - Little data are available, but good engineering practice is being applied to treat certain situations.

I h

f,. 7

3. Aaessel fraction - This important aspect of fission product transport is under investigation het relatively littia asperimental data are

available.

6. Adaquacy of simple models - There is little agreement on this subject although a variety of models all tend to give similar high releases for the volatile fission products. Ic is recognised that artificially early and escassive releases from the core are not acceptable since this may deplete the source of fissica products that should rightfully j remain with the fuel for possible release after asltthrough of the j pressure vessel.

5. MRC Fosition i

1 . Despite their limitations, it is the NEC position that the applications

   \s                studies. Which incorporate use of the empirically based CURSOR oods for in-j vessel release estimates, provide a reasonable best estimate basis for sup-j                    porting the mid-84 decision. The chemical and physical forms of the released materials are likewise sufficiently well established to support the     -

best estimate transport caleslations. Although the uncertainties in the in-vessel releases are large and the implications of these uncertainties for the overall scores tera estimates are not fully established, the overall under-standing is adequate to permit the 3RC to proceed with the mid-84 decision while developing an laproved data base and confirmatory analysis anthods through SARP Phase II.

6. _ Continuing tonfirmatory Work Efforts to identify the chemical forms of the released fission products will continue is existing experimental programs, mainly the severe fuel

  .~,           .-                                                                                                    -
    ~

e r -e-4 damage tests at FFF and the high temperature fission product release tests at Ook Ridge. CsI, Ca05, and To are the chemical forms adopted for the ele-ments Co. I, and Te in the on going term calculation. The effect of pressure on fission produce release will be investigated in the severe fuel damage tests at FBF. Data is also expected from ' DEI-2 cote esamination and from testa pisaned for the ACER. The NUREC-0772 release a rates which are based on results obtained at low pressures are being used in the current calculation. The behavior and effects of structural and control rod asterials are being investigated, in the expanded Core Melt tests, the new SASCRA tests, the bundle tests of Regen (EfK), the bot call tests et ORNL, and'the test to be run in the ACER. Results from 'DtI-2 emanination will also be infor-

           #           motive. Results from older AASCRA and Core Walt tests are being esamined more critically.

f. Inforastion on the release of fission product from the 502 -tirealoy interaction is expected from the severe fuel damage tests at FBF and the ,

t separate effects esperiments at Ook Ridge. The 'DL1 core esamination program will also be able to provide information in this area. The difficulty of scaling results from laboratory scale esperiments will be addressed when results from the FBF tests, the full length MEU esperiments, the TMI-2 core examination, and the 10-kg Core Melt experiaants at Oak Ridge are available. Results from the 10-kg Core Melt experiments een be compared directly to those determined previously in the corresponding 1-kg furnace. I l . 1 \ 8 1 =

i (-..

' \,                                             Correlation of data from the various arperiments containing fission produets or simulh ts should provide information toward resolving the parti-tioning af vapore between deposition on fixed surfaces and association with                                                                            !

4 serosol particulates. Calculational studies such as QUEST. MELCOE, and ISLFROG will provide guidance as well as improved results. Detailed studies of specific accidents as done by BCL and the R&SA program are similarly important. t 4 e e*g ;

                ' 't i
               ' '/

i e 1 4 e ( . I i . I i

                                                               .R                                                                                  !

r ) . , Q' bs / Aman,f mead 1928 B C Centracter Reed Esports Cemeeste EC/Costracter/IBCOE meettiene Isees Paper Title ! i

                                                *Estiesties of Piastos                      e ,*-ese of Agr=====e                                                         l Technical Isome 1.4.,1  CEPL (R. A. Ierema/                                                                                        '                                     ,:

Bete end IIngnitude R. F. II1 cheer) Fredact med Core Iloterial 'f 5eerce therectoristice" Enlative to the IBCOR Steam Osidotten of Fiesten Fredect IBCDE Technical Bayerts Model antamme from Foal , 11.1. 11.4, and II.5 , s 1. enadatiam of 502 resulte.ie ' ll enhanced fleeien predeet release, l, but it la only see of eeny facters !!

           .                                                                                                     that effect rolesse.                                       ;;

e Areas of Disagressent j t Boletive to the Steam oxidaties IIndet

                                                                                                                                                                            .J s                                                                                                        1. The model should be capirically D                          '

edjusted to give leser rolesse ' rates for Ie, Kr, Ca and I. i

2. The reteetles of To by emeusdised Zirealoy cladding appears to be a j real (but asseguilibries) yhenome-ese that ebeeld be takes inte com-siderettee is the esebleed steen !

oxidatios/chemicol theg_-ff ;ste mad =E. t { j

                                                                               ,O                                                              .

7 .

p. g

                                                                                                                                    ...].
                                                                                  . w'      .
                                                                                                                                                          ,       I Aseeciated ID00E Reports                  Comments ELC/ Contractor /IDOOR positions Issus Paper Title         WBC Centreeter Esed e  Areas of Disagreement (Cost.)
                                        '                                                   Relative to the IDQNL th=ical' hrmo-dynamic Model

1. h SOESASK11 celesistions should be s checked for accuruy. .

2. The effect of alleying (diluties) obould be aceoasted fer by using Reuelt's las as an appresimation. >

3. Complete runoff of the control rod

                                                                         ~
                                                ~
 ,                                                                                               alley at it selting point is not cor-

" rect. N retenties of a fraction as " an alloy with Zirceloy for inter

vaporisation at high tcuperature is probable. ,

4. h reaction of fissiop products with control rod med structural materials components should be included in the v.4por pressure calesistions.

5. b plan was presented for the calemle-tions of release of fission, products ,

other than Ie, Kr, Cs, I, Sr. and Rs. If computer space is limited, the sub-

                                                                                                 -Jtitution of Be .for Xe or Er would be more inforastive.                -

6 e

   - - - - - _ _ _ _ _ - - - -        w      =_                            u
                                                                                                           ,          \

e 9

e J

*\

1 ^a l-l; r,, i '1 1 I J 3 l' Ib 1 1

                             ] illi ab]
                          'd i Ilh
                     }    }  i .:

a

                   }l 1                              .

E E 1 e e

e. W

/                      se y             :

i t

f% D.- 1 -Q - - ( a'

x. . -

Aseeciated IBCER T==== Paper Title ERC Centractor Esod Reports Commente IWtC/ Contractor /IDCOR Bositions e Areas Requiring Further Definition'

1. Fission product chemical reactione

                                                                                                             '                that might alter flesion )roduct         s treesport - e.g., those that might               j' effect the stability of CsI.                     ;

2. Oseach release is that which occurs as the result of fragmentation and leechlag whom a het core is reflooded with unter. We need to know how each release necers to the gas phase and how much is effectively trapped la the water.

3. WO 2 -Er Chemical Resetion (Liquifaction) - Evidence exists Q (FSF scoping test) that liquifacties enhances fission prodect release at least for low barnap fuel.

4. Fission Fredact-Aerosol lateractica -

Although this is primarily a concern for transport beyond the core, the potential for condensaties on surfaces or existing particles, or the emelee-tion of particles in the cooler apper portion of the core exists.

5. Distribution of Decay East - Proper calenlation of fission prodoet release and transport will enable accurate tracking of decay heat.

                                                                                                        ~
                          ~~         ~

{ . s ISSUE 1.4.2

                                               .      RETENTION OF FISSION PRODUCTS DURING IN-VE5SEL TRANSPORT

1. Description o'f the Issue This issue is the extent to which fission products are retained within the reactor coolant system during severe accidents. Its scope includes:

(1) The rate of transport of airborne fission product vapors and aerosols along the segments of the escape pathways that are within the primary pressure boundary (2) ,The deposition processes that attenuate the fission product vapor and aerosol inventory as they are transported along the escape path-ways segments within the primary pressure boundary (3) The processes that the suspended vapor and i aerosol species undergo that affect their ! subsequent transport and deposition behavior in the containment and the plant environment (4) Processes which deposited materials undergo which can result in their re-evolution and release.

2. Implications of the Issue for Regulatory Questions Because analysis of the " Retention of Fission Products During In-vessel Transport" is an essential step .in the . analysis of source terms for existing plants and existing plants with proposed safety improvements, it follows that technical issues that arise in this portion of the source term analysis process will have implications for decision making;on the questions:

                                       -(1) How safe are existing plants with respect to severe accidents?

Fo t A %8

          ~                                              -

4 c.l 27

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

2 (2) Now can the level of protection for severe accidents be increased? Deficiencies in current or past analyses and analytical methods or in the supporting experimental data base that preclude adequate resolu-tion of important unresolved issues have direct impact on the regulatory question: "What additional research or information is needed?" e

3. Subissues !

The subissues which we have identified are the following: (1) Completeness of the available analysis results (in terms of the plants and accident sequences snalyzed) for supporting generic and plant specific conclusions (2) Influence of radioactive decay effects (e.g., on

     - ,                                   re-evaporation of condensed material due to i?                                    local decay heating) on the in-vessel transport and deposition behavior of fission products I                                   (3)     Importance of in-vessel deposition and resuspension phenomena not modeled in the TRAP-MELT code
                                           --   Chemical transformations (due to homogeneous-gas phase and heterogeneous-surface reactions) including radiation effects in the primary system, e.g., Cs1 reaction with boric acids in BWRs and possible formation of volatile species
                                           --   Long-term release from the RCS surfaces after the primary pressure boundary failure
                                           --   Chemical effects of intermediate vapor pres-sure PWR control rod component materials (e.g.,

l

Ag,Cd)

(4) Adequacy of the TRAP-MELT description of those processes for which it does include models

                                         ' ,-- Aerosol deposition processes -- the rates are assumed to be mass transport limited (i.e.,

surfaces are treated as infinite sinks)

      /                                     --  Th.e vapor deposition velocity correlations are based on a fairly ifmited empirical
                         .                       data base and are not supported by a

        ,   ,     _. .            _      n        1 1

3 microscopic understanding of the reactions involved and the surfaces are treated as , infinite sinks for'the chemical reactions.

                                           - , The aerosol composition is not followed for each particle size bin                                              1 1
                                           --    Vapor / aerosol interactions including analysis of both heat and mass transfer effects.

(5) Validation of TRAP-ELT with respect to alternative analytical approaches and experimental data available (6) Sensitivity of source term uncertainties to uncer-tainties in the in-vessel transport and deposition i behavior (7) Uncertainty associated with available analyses c'. ' the in-vessel transport of fission products.

4. Status of Understanding O^ The status of current technical understanding as it pertains to resolution of each of the subissues identified above is outlined below.

An attempt is made based on a review of the ongoing research and the I anticipated timing of pertinent results to present the status as it is likely to exist in the time frame of the mid-84 decision.

   -                                  Subissue 1. The SARP model applications project Fission Product Transport (B6747), which is providing best estimate severe accident source terms, has addressed this subissue by judicious choice of a representative subset of LWR plants and accident sequences for analysis. The choice of the plants analyzed (Surry, Zion, Peach Bottom, Grand Gulf, and Sequoyah)                ~

i is intended to reflect the effects of a broad range of plant designs. Likewise, the accident sequences chosen for analysis (i.e., Surry AB, S20 V, TE8', Sequoyah S HF, 2 TM.B', TM., Peach Bottom TW, TC, AE; Grand Gulf TPI, TQUV, TC) cover a wide range of the expected severe accident , parameter spac'e. The selection of basic accident sequences and plants for analysis by IDCOR has a significant amount of overlap with those used in ! the, Fission Product Transport project. There are, however, important

4 . i

l t i . , _ , 4 . f r. differences in the anticipated progress of the sequences in the two sets of analyses, these differences will help define the completeness of the available analyses. l Subissue 2. Radioactive decay effects on fission product trans-port (e.g., precursor transport and decay, decay product recoil, local decay heating) are not explicitly considered in the TRAP-ELT code. Resolution of this subissue is being pursued through the SARP model development projects: ELCOR Development (A1339) and Fission Product Transport (86747). In the IDCOR analyses, the effects due to radioactive transmu-tation of species are not considered. The most significant potential l effect of decay, however, appears to be the heating of surfaces upon ! which fission products have deposited, leading to potential for re-evolution of, volatile fission products. This is included in the IDCOR hi analyses in a manner which is not incons'istent with the current high O level of uncertainty regarding this phenomenon. Subissues 3 and 4. Resolution of the specific topics associated with these subissues is being pursued through the SARP experimental and model development projects. The specific projects involved include High Temperature Fission Product Chemistry (A1227), TRAP-ELT Verification Tests (80488), Fission Product Deposition on Aerosols (80815), the Marviken Project, ELCOR Development (A1339), and Fission Product Transport (86747). The IDCOR analyses shed no new light on these subissues. The lack of available information regarding the phenomena not considered in TRAP-MELT and RETAIN severely limits the extent to which such uncertain-ties can be resolved in current analyses. The QUEST project may provide insight into the adequacy of the modeling of certain phenomena within these.two c6 des for specific conditions. Subis' sue 5. Although many of the individual process models in

    /             the TRAP-MELT code have been validated, integral validation data for the overall code is not yet available. Resolution of this subissue is being pursued through data obtained from the SARP projects Fission Product e

6 9 4

o 5 ( Deposition on Aerosols (80815), the Marviken Project, and TRAP-ELT Verification Tests (80488). However, this subissue will be only partly resolved in the mid-84 time frame and more complete validation of TRAP-

ELT will not occur until the SARP Phase II time frame. The RETAIN code used in the IDCOR analyses lacks integral validation also. While many of the physical processes are modeled in very similar fashion, the use of the log-normal assumed aerosol size distribution is in need of validation. Subissues 6 and 7. Because of the status of the current under- ! standing with respect to resolution of Subissues 3, 4, and 5, it is clear i that the uncertainties associated with available analyses of the in-vessel transport and deposition behavior are large. However, analysis of these uncertainties and their implications for the source term uncer-tainties will not be completed. The SARP Phase II experimental and model l development work is focused on reducing these uncertainties and eventually p providing for quantification of the residual uncertainty. The specific SARP projects involved are Fission Product Chemistry (A1227), TRAP-ELT Verification Tests (80488), Fission Product Deposition on Aerosols (80815),

Marviken Test Results, ELCOR Development (A1339), and Fission Product Transport (B6747).

No systematic quantitative assessment of the sensitivity to uncertainties in the in-vessel transport and deposition are presently ' included in the IDCOR analyses. L l Overall Status of Understandina. The TRAP-ELT 2.0 code has l been used for specific plant and accident sequence analysis in supoort of resolution of this issue. However, remaining questions concerning the completeness of the set of models for describing potentially important ' phenomena ('e.g., resuspension, chemical reactions, energetic events) mean that considerable uncertainty must be associated with current results. The actual known sources of uncertainty have not been formally propagated l ' through the code calculations. The results of validation studies which will become available in SARP Phase II should permit removal of some of the,se uncertainties. o 9

6

5. NRC Contractor Position Despite their possible limitations, it is the NRC contractor

position that the' SARP applications studies, which incorporate use of the TRAP-MELT 2.0 code for analysis of the in-vessel transport and deposi-tion, do provide a reasonable best estimate basis for supporting t:te mid-84 decision. Although only limited consideration has been given to the uncertainties associated with these best estimates, the overall under-standing reflected in the current analyses is adequate to proceed with the mid-84 decision while developing an improved data base and confirma-tory analysis methods through SARP Phase II. i Further, based on our evaluation of the existing IDCOR report on in-vessel transport and deposition, the current NRC contractor position

is that the adequacy of the log-normal aerosol size distribution ass'umption used in RETAIN needs to be demonstrated. Also, an assessment of the sensitivity of the predicted extent of re-evolution of deposited fission [ products to the recognized uncertainties relevant to their phenomenon should be performed. l

6. Continuino Confirmatory Work Several current programs have been cited above in relation to

! reducing uncertainties connected with this issue. The TRAP-MELT verifi-cation tests (B0488), the Marviken tests, and Fission Product Deposition on Aerosols project (B0815) will provide information of great importance to the analyses being conducted. These projects are expected to provide useful input to the validation of existing modeling approaches. There . is also a need for coritinued generation of information which will aid in understanding processes currently excluded from the analyses. Work is ' needed-along the lines of High Temperature Fission Product Chemistry (A1227) to help fill in the significant gaps in our current knowledge l regarding proc' esses and species which are potentially important. > /. .

V e e 9

e

     ,e,-,--       - , - - _ - - . - - , _ - - - , , , _ - .,, -
                                                                                  ,,,___.-__,,,_,,,r_,,          , _ . _ _ _   _ , . _ , ,   _ , , , , - . , -, ._,_m__-._,-__   -.-m., , _v- ,

~'-

1 l, Issue Paper Title

7 .-

f

t 1.4.2 Retention of Fission Products During In-vessel Transport

NRC/ Contractor Leads NRC: M. Jankowski 4 BCL: M. Kuhlman Associated IDCOR Reports Fission Product Transport in Degraded Core Accidents. IDCOR 11.3 Report ! Current NRC/ Contractor /IDCOR Positions Areas of Agreement:

1. Physics governing vapor and particulate deposition within the RCS

) ' 2. Re-evolution of semi-volatile species due to fission product decay

heating is potentially important in determining ultimate fate of

         .y               released fission products

[i 3. Significant retention of aerosol material is possible in the RCS, its extent largely governed by source rates and residence times l

4. Base-case comparisons of RETAIN and TRAP-MELT 2 are desirable Areas of Disagreement:

! 1. The adequacy of the lognormal assumption for the aerosol size distri-bution used in IDCOR's formulation of aerosol mechanics is suspect

2. The potential importance of resuspension of particles deposited in the RCS is doubted by the NRC contractor based on the expected large size of this material, and the lack of this phenomenon's demonstrated effi-ciency for deposited fine particles

3. The formulation of aerosol turbulent deposition differs in the two groups' treatments Areas Needing Further Study:

1. Further work is needed to assess the potential effects of heatup due to fission product decay. The sensitivity of the amount of revolatili-zation to such items as the mass of structure requiring heating. the poss.ibility of reactions of the deposited vapors with the surfaces, and the possible flow patterns within the RCS which may lead to redeposition require investigation.

     ...            2. Sensitivity of code predictions to nodalization of the RCS requires
  /                       examination V                 3. The flow field within the RCS for conditions in which natural convection might be expected to dominate the circulation pattern requires better definition in all analyses

s ISSUE 1.11.3

RATE AND MAGNITUDE OF RELEASE OF RADIONUCLIDES FROM FUEL (Ex-Vessel)

1. Description of the Issue This issue deals with how core materials can become airborne during severe reactor accidents after the core materials have penetrated the

reactor vessel. It is intended'that gas phase entrainment of core materials and associated fission products will be included when referring to " release from fuel". On the other hand, processes which involve escape of fission products to other condensed phase materials (e.g., the leaching process) but which do not involve a subsequent mechanism for this material to become airborne are not' included. Because release from fuel followed by subsequent " airborne" transport of radionuclides to locations external to the plant are essential steps for reactor accidents to have significant consequences, the scope of this issue is interpreted to include the following:

1. The rate of the ex-vessel release of radionuclides.

2. The properties of the release radionuclides and/or the materials with which they are associated and which determine their subsequent transport behavior. For radionuclides released as vapors, this means that the chemical forms of the released species can be important and for fission products released as or in particulate material, the particulate properties, e.g., size, shape, density, and perhaps composition, which dictate the subsequent transport behavior can be important.

3 The rel, ease of inert (nonradioactive) materials either as vapors or as aerosols that can influence the transport behavior'of radionuclides or their carriers through vapor phase chemical reactions which change the molecular species involved or through processes such as agglomeration,

          '                                      ~

Fos A - N -%i

                                                                      .         M

  ,     ,                               i cocondensation, adsorption of vapor species on particulates, f

et..

2. Implications of the Issue to Regulatory Questions Nuclear power plants could adversely affect public health and safety through release of radioactive material to the environment. The extent to which such a release could occur in postulated severe accidents is referred to as the " source term" question. While the process and critieria that will be used for making regulatory decisions on questions of reactor safety with respect to severe accidents have not been fully defined, it is clear that any reasonable decision making process will involve consideration of severe accident source term information. Unfortunately, our knowledge of severe accident source terms is incomplete and, while it can and is being improved through research, will remain somewhat incomplete in the time frame that has been established for decisions on the regulatory questions. ,

j ) At present, summary technical input on our knowledge of severe accident

    "       source terms can be provided to the regulatory question decision making level by providing best estimate source terms with associated uncertainties and/or conservative bounding estimates. This requires analysis of how fission products can be released from the fuel and subsequently transported along pathways that result in their eventual escape to the environment.

Unfortunately, there is not a consensus of the informed technical opinion on the choice or adequacy of the available analysis methods. As a result it can be expected that he decision making technical information environment will include:

1. NRC's own "best estimate" source terms for selected plants and accident sequences with incomplete information on the uncertainties associated with these estimates.

2. IDCOR's "best estimate" source terms for selected plants and accident sequences again with incomplete information on the uncertainties associated with these estimates. Because of the different perspectives involved, it will not be surprising if IDCOR's best estimate source terms disagree significantly from NRC best estimates for similar plants and accident sequences.

.g.

                                  . . - -               -:.= -=    . a.:  --                        -            '- --
,    e-                        3          Questions concerning the validity and/or relative merits of the analysis approaches used in the NRC and IDCOR analyses.

Rational decision) making in this environment will require that the decisionmakers be informed on the state of knowledge concerning the unresolved underlying technical issues. Because analysis of the " Rate and Magnitude of Release of Fission Products from Fuel (Ex-Vessel)" is an essential step in the analysis of source terms for existing plants and existing plants with proposed safety improvements, it follows that technical issues that arise in this portion of the source term analysis process will have implications for decision making on the questions:

1. How safe are existing plants with respect to sev'ere accidents?

2. How can the level of protection for severe accidents be increased?

s Deficiencies in current or past analyses and analytical methods .in the, ! supporting experimental data base that preclude adequate resolution of important unresolved issues have direct impact on the regulatory question,

., "what additional research or informaton is needed".

I.! 3 subissues The subissues which we have identified are the following:

1. Completeness of the available analysis results (in terms of the plants and accident sequences analyzed) for supporting

.{ . generic and plant specific conclusions ! 2. Importance of possible release associated with ex-vessel phenomena which are not currently modeled, e.g.,

                                         - release associated with melt ejection from the reactor vessel in high pressure accident sequences
                                         - release associated with ex-vessel steam explosions 3         Validation of the VANESA code results for release rates and physical and chemical forms of the released materials

4. U'ncertainties in current best estimates for the ex-vessel

, release and their effects on the uncertainty in source term

   ,.                                    estimates.

L 4 l

                 . . . _            .                 = - - =      =              ~              -               --                       - -
          .l '
                         ,                                      s

4. Status of Understanding 4

The status of technical understanding of the subissues of Rate and Magnitude of Release of Radionuclides from Fuel (Ex-Vessel) identified above is outlined below. An attempt is made to indicate the status that will exist at the end of FY 1984 in view of ongoing research. Subissue 1: A series of analyses for specific accidents and plants (SURRY AB, S D, V, TMLB'; SEQUOYAH S HF, S D, TMLB', TML; PEACH BOTTOM TW, TC, AE; GRANDGULF TPI, TQUV, TC; ZION S D, TMLB) will be completed as part of the SARP model applications project - Fission - Product Transport (B6747). These analyses show that the ex-vessel release of non-radioactive aerosol can mitigate the available

                                     . inventory of fission products released during in--vessel phases of accident. Radioactivity from the ex-vessel sources can include h                      refractory isotopes. Long term low-level releases of moderately volatile fission products such as tellurium can occur. All these effects can be mitigated substantially by a water pool overlyling the debris.

The QUEST program at SNL (A1227) has identified significant sources of uncertainty in the ex-vessel source term estimates. These ! include predictions of the initial composition of the fuel at the start of ex-vessel interactions as well as descriptions of the interactions and source term generation processes. Research to reduce the current uncertainties is underway in tt SARP programs Molten Fuel-Concrete Interactions (A1019), Core Melt / Coolant Interactions (A1030), Core Melt Technology (A1218), Containment Analysis (A1198) and MELCOR Development (A1339). Subissue 2* The QUEST program'(A1227) has attempted to estimate ! the effects of ex-vessel source term effects not currently modeled. These estimates, done in an environment of little or no data, indicate l# ' that the importance of aerosol generation from high pressure melt 4 ejection or steam explosions depends critically on the timing of j containment failure or significant leakage. Though both processes can 1 l -=- ..

 . - , , -                 ,-.,.~., -... . - __.                                                                    .__..---.--,--..,--,n

( . ., - s I . produce quantities of aerosols comparable in magnitude to those

      %,                                 produced by modeled phenomena, these sources are of short duration.

The time aerosols pr,oduced by these processes are potentially releasable from containment is small.

Research underway to clarify the nature of these phenomena that are not currently mode'l ed is underway in the SARP projects cited in connection with sub-issue 1. Subissue 3 Ressarch to validate the VANESA model of fission product release during core debris interactions with concrete is underway in the SARP project Molten Core Technology ( A1218). Large-scale tests using prototypic melts of UO,/2r0 3 /Zr and realistic concretes will be completed at the end of FY84. Separate effects tests and tests of the effects of water are also being planned. In addition to these experimental studies, a direct comparison of the IDCOR model and the VANESA model for a generic accident situation is being conducted. Early results indicate that the philosophical

                 ~)                    bases of the two models are similar. The quantitative discrepancies in the models are caused by (1) choice of the appropriate melt temperature for vaporization calculations, (2) the approximations of

^ melt and concrete. compositions, (3) the descriptions of core ! debris / concrete interactions that drive the aerosol generation processes, and (4) the differences inherent in the thermodynamic model of IDCOR and the. kinetic model VANESA. i j Subissue 4 The first, formal examination of uncertainties in the Rate and Magnitude of Release of Radionuclides from Fuel (Ex-Vessel) was conducted in QUEST as described in conjunction with subissue 1. This examination is restricted to the TMLB' accident at SURRY. , Similar analyses for the S:D accident at SURRY and the TC accident ! sequence at GRAND GULF should be availab'le by the end of FY84. Continued ana' lyses of the uncertainties will be conducted in conjunction with the QUEST program and as part of the MELCOR I Development program (A1339). i . . 4 a i

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

7

. .. e; 5. NRC Contractor Position The NRC Contractor Position is that the,VANESA model represents a '
first step-in the mechanistic modeling of fission product release and aerosol generation dur;ing ex-vessel core debris interactions with concrete. These releases can be of significant importance to the overall accident source term. Uncertainty analyses of the predictions and the comparison of VANESA to the IDCOR model suggest uncertainties stem from descriptions of the melt / concrete interaction, the initial conditions and compositions of the fuel at the start of the interact!ons as well as the uncertainties in the physics and checistry of the release processes. Despite these limitations, predictions from the VANESA model can be a reasonable best estimate for supporting decisions at the end,of FY 1984.

The NRC Contractor Position is also that other sources of aerosols and fission products generated ex-vessel by melt ejection or

       ]         steam explosions have been neglected because of a lack of data. IDCOR
       ..        has neglected these sources as well. Uncertainty analyses suggest
!                this approximation may not be serious if containment failure or leakage is not coicident with these sources. The significant

, uncertainty in the timing of containment failure as well as the uncertainties concerning these other sources must be recognized in decision-making at the end of FY 1984. The high levels of uncertainty that will remain at the end of FY 1984 suggest the conduct of confirmatory work described below. ! 6. Continuing Confirmatory Work 1 Continuing work in the SARP Phase II programs Molten Fuel-i Concrete Interaction (A1019), Core Melt Technology ( A1218) and i Containment }.oads+ Mitigation (A1247) will provide data to validate the VANESA model and the CORCON model of core debris / concrete interactions. Of particular interest are sustained tests involving f. UOn-Zro, melts, doped with fission product elements, and in contact (,. with realistic concretes. , Continued uncertainty analyses in the QUEST e s [ program (A1227) will demonstrate when enough research has been done to

adequately resolve this issue. Continuing work in the Core Melt / Coolant Interactions program (A1030) and the. study of high pressure melt ejection in the Molten Core Technology program (A1218) will reveal any significant errors made in omitting other sources from best estimates of the ex-vessel source term. Again, uncertainty analyses by QUEST and accident analysesintheMELCORDevelopmentprogramYA1339)willestablishthe e effects of findings of the experimental programs on the overall source term. f WO

9 O 9 Y e e.. s.,, - . _7

p SupMARY OF NRC/ CONTRACTOR /IDCOR,7 PAPER STATUS q-

o- -

i ( ..

                                                                           )                                                   I

) ISSUE 1.4.5, RATE AND MAGNITUDE OF RELEASE OF RADIONUCLI.DE FROM FUEL (EX-VESSEL) NRC CONTRACTOR ASSOCIATED IDCOR CURRENT NRC/ CONTRACTOR /IDCOR LEAD REPORTS POSITIONS

; D. Powers /SNL           No IDCOR report is specifically devoted           1. Areas of agreement:

to this area. .

i o Both IDCOR and the NRC contractors have l Applicable reports are: neglected the source term from ex-vessel (1) Technical Report 15.3 Core-Concrete steam explosions - largely because of lack ! Interactions data.

r s j (2) Technical Report 11.1, 11.4 & 11.5. o Both IDCOR and the NRC contractors believe Estimation of Fission Product Core- vaporization is the dominant mode of fission Material Characteristics product release during core debris interactions ;

with concrete. Vaporization rates are enhanced i by gas sparging through the melt. i ! o Both IDCOR and the NRC contractors have l 3 neglected release associated with high pressure melt ejection. Sensitivity analyses by the NRC l contractor indicate the importance of this source term depends critically on containment failure time.

2. Areas of disagreement:

o There is a large quantitative difference in the 1 estimates of the ex-vessel source term by IDCOR j and the NRC contractor. These differences are i largely caused by differences in the initial " l conditions for core debris / concrete interactions i used by 10COR and the NRC contractor. Specific areas of disagreement include l i o presence of Te in the ex-vessel core debris, l , o formation of a stable, solid, crust over core debris ex-vessel, and l' i o vaporization rates at given bulk and surface debris temperatures. I

      ,. ^      ,,

l. '

ISSUE 1.4.4 DEPOSITION OF FISSION PRODUCTS IN CONTAINMENT DUE TO NATURAL PROCESSES 1 I 1. Description of the Issue j This issue concerns the deposition of fission product <s within the containment resulting from natural (not to include engineered safety features) processes. Because the issue is interpreted as involving the evolution of a source of airborne radionuclides for possible leakage into other buildings or to the environment, natural processes leading to . . re-evolution of deposited materials as well as those affecting in- and f ex-containment transport are to be considered. Therefore, the scope of this issue is interpreted to include: a ., Depositio'n of fission products in the containment due to natural processes

b. Other natural processes occurring in the containment that can influence the quantity ~

and characteristics of fission products moving through the containment to the environs.

2. Implications of the Issue for Regulatory Questions The major technical input to the establishment of positions on the regulatory questions will occur through analyses (implemented through applications studies) to support decisions on the questions:
(1) How safe are existing plants with respect to severe accidents?

(R) How can the level of protection for severe accidents be increased? Since such ana,1yses include evaluation of the " Deposition of Fission ) Products in Containment Due to Natural Processes" as a key step, it is f cl' ear that subissues raised by the approach or approaches used in this segment of the analysis process will have implications for the positions Foi A -N-%8 Ll2,*\ g

g- 2 ( taken on these questions. Also, the outputs from the overall analysis process that will be used most directly in consideration of these ques-tions will be:

a. Estimates of some measure of plant safety

b. Estimates of the uncertainties associated with the plant safety estimates.

Deficiencies in current or past analyses or analytical methods have direct impact on the regulatory question, i.e., what additional research or information is needed? Subissues arising in connection with current estimates and , associated uncertainties for the rate and magnitude of release of fission products are identified below.

3. Subissues

() L The subissues derive from consideration of the completeness of the analyses that have been conducted, the realism of available analysis techniques, and the uncertainties associated with current estimates of the deposition of fission products in containment' due to natural processes. l The subissues which we have identified are the following: (1) To what extent do current source term estimate include realistic models for the most important processes affecting fission product behavior in the containment? (2) To what extent do current containment deposi-tion analyses incorporate consideration of possible energetic events in containment (e.g., hydrogen combustion events, steam explosions, i and containment depressurization). (3) Are the potentially important sources for

                               -    aerosols in containment identified (e.g.,
                        ~

nucleation, resuspension, flashing of liquid pools)? Is sufficient experimental and/or

             .                  . . analytical data available to allow quantifi-cation?                                                                                               ,

f !b . (4) What is the rate of production and depletion of volatile species (e.g., radiolysis products, organic iodides, and Ru04) in the ~ containment atmosphere?

                                                      .__,.,_,_,,._..,__,.._m.___.,,-_-m,,.,y.,

__-.e,,.,.e_.-,.y_-, ,,_._,_, -

(-

3 (5) What is the long-term behavior of fission products in the containment (e.g., fission product aqueous chemistry, surface sorption / desorption)? (6) To what extent does attentuation in con-tainment leakage pathways affect source terms? (7) To what extent have the models used in current analyses been validated?

4. Status of Understanding The status of current technical understanding for each of the subissues and the underlying data base are outlined below. -

Subissue 1. Reasonably mature analytical methods are available

to quantify the major natural fission product transport and depletion

            ^
       --         mechanisms in the containment, i.e., aerosol agglomeration, settling, and deposition. Comparisons between the NAUA and CONTAIN codes will provide a measure of confidence with respect to the adequacy of major models. Validation of these models has been accomplished with small-scale or dry containment experimental data. Validation with larger scale data in wet steam environments typical for. LWR containments is currently in progress and can be expected to provide substantive (although prelimi-nary) data in the mid-1984 time frame.

Subissue 2. Since analysis of the natural deposition in containment buildings does not now incorporate explicit analysis of the effects of energetic events on fission product behavior, it is clear that such omissions must be recognized in assessing uncertainties associ-ated with the current estimates. However, the effects of the uncertain-ties so introduced on the overall source term estimates have not been

                -  analyzed. This issue will be addressed in SARP in the MELCOR project
   .               (A1399) but these results will not be available in the mid-1984 time

( frame. Therefore, parametric analyses will be necessary to assess the - significance of the energetic events.

Subissue 3. Experimental and analytical investigation of core-concrete and co're-coolant interaction (Sandia), investigation of iodine volatility (Canada), and preliminary results of experimental programs addressing vapor deposition (Sandia), melt ejection and dispersal (Sandia), aerosol resuspension (ORNL), and aqueous fission product chemistry (ORNL) can provide reasonable confidence that major sources of fission products in the containment have been characterized. Completion of these programs provides confirmatory results for these major processes. The effects of radiation, desorption of surface deposits, and flashing of liquid pools are considered secondary in importance, and will not be quantified in the mid-1984 time frame. Subissue 4. Processes which involve production of volatile species in the containment atmosphere are currently believed to make a

          ,    small contribution to the overall release after' containment failure.

Resolution of this subissue is considered to have relatively low priority as long as aerosol releases dominate the source term. Re-evaluation of this approach becomet necessary when very high attenuation of aerosol l releases are predicted. Subissue 5. Long-term release of fission products, even l following containment failure, is believed to be relatively unimportant from a public risk perspective. Information pertinent to the resolution of this subissue in the long term (1985 and later) is being generated through the SARP project B0453 to aid in understanding in-plant accident response and recovery. Subissue 6. The SARP is currently pursuing characterization l l of containment failure modes through the projects A1249 and A1375. Because of the timing for completion of these studies, it will not be possible to resolve this issue. The mid-1984 decision will have to be based on parame'tric and sensitivity studies. ,( _. , Subissue 7. The codes used in the analysis of containment natural deposition processes (CONTAIN, NAUA, CORRAL 2) have been compared 1 e

                                                              #                         gG e

W .

   .                              i e

b 5 with the results of experiments (CSE, NSPP, ABCOVE) conducted under However, simulated quasi-equilibrium containment accident conditions. the conditions associated with some potentially important energetic (short-term) containment events or multiple species effects have not been simulated. Preliminary results of large-scale experiments in progress (e.g., Marviken) will be available by mid-1984. Completion of these programs (including EPRI and internationally sponsored experiments) will provide confirmatory results. Overall Status of Understanding. It is possible that the neglect or underestimation of several natural fission product attenuation processes in pa,st analyses has resulted in overly pessimistic assessments While this conservative approach provides a of accident source terms. safety margin in the assessment of potential accident consequences, this approach results in skewed (or biased) results when applied to such

     -         regulatory questions as "how safe are existing plants" or when used in Therefore, a the assessment of the cost effectiveness of features.

correction of the assessment of the contribution of natural processes to the attenuation of source terms is necessary. A reassessment of accident l source terms and the necessary methodology is currently under way by the Accident Source Term Program Office. Preliminary results of this effort indicate that the degree of source term attenuation by natural processes is strongly dependent on While the residence time of the fission products in the containment. several orders of magnitude attenuation (by agglomeration and settling) of aerosols can be expected if containment integrity is achieved and j maintained for time periods exceeding one day, little attenuation is accomplished by these accomplished by these mechanisms when containment lI failure follows the release of fission products in a few hours or less. It is' apparent, therefore, that this issue is closely related to the issue of containment failure more and time.

  ~~        '

Ther'e are several natural processes identified in Subissues 2 For these

./

th' rough 6 which are not modeled in currently available codes. processes, the levdl of understanding is significantly lower than that Current estimates for,the aerosol agglomeration and settling processes. L

y, ._. 6 [ 1 indicated, however, that these processes are of secondary importance whenever the accident sequence permits sufficient time for the primary

aeroso1 deposition processes to become dominant, i.e., result in fission product attenuation exceeding an order of magnitude. In contrast, acci-dent sequences if.volving early containment failure do not include suffi-cient residence times for aerosols to agglomerate and deposit in the e containment. For these sequences the large attenuation factors associated I with long-term aerosol processes do not develop, and source term attenua-tion by natural processes may be limited to that associated with energetic containment events or other processes or secondary importance in compar-ison with long-term aerosol behavior.

5. NRC Contractor Position It is the NRC contractor position that the ASTP0 efforts in

( ')' ' predicting source terms using the modified NAUA and SPARC codes as part of the BMI-2104 suite of codes provides are reasonable best estimate basis for supporting the mid-1984 decision. Work under way will provide estimates of the range of uncertainties in such predictions but the current understanding of the technical issues is sufficiently well developed to proceed with the mid-1984 decision while recognizing that an expanded data base and confirmatory research is under way through SARP-Phase II. Based on a review of the IDCOR 11.3 Report, it is the NRC contractor position that either a discrete size range (sectional) approach to the size distribution should be adopted or the adequacy of the IDCOR log normal size distribution approach used in RETAIN should be demonstrated. Further, it is expected that water condensation on aerosol particles in the containment may have a significant effect on attentuation and its

neglect in the RETAIN approach should be supported.

6. Continuing Confirmatory Work

                           -        The resolution of this and the associated subissues is pursued
                     ,through research performed under the SARP together with the information obtained through review of the IDCOR and other pertinent work.
                                        - - - - - -      , ---       ,           O

p- - , [

7 SARP Model Application Projects. The SARP model application projects which will and are generating information pertinent to the

res31ution of this issue: Containment Analysis (A1198) and Fission ProductTransport(B6747). SARP Model Development Projects. The SARP model development projects that will support resolution of this issue are CONTAIN Develop-ment (A1198), MELCOR Development (A1139), and Fission Product Transport Model Development (B6747) will be confirmatory to the simplified models and assumptions employed in evaluations available by mid-1984. SARP Experimental Profects. The SARP experimental projects that will suppo'rt resolution of this issue are Post Accident Fission Product Chemistry (B0453) and LWR Aerosol Release and Transport (B0121). Significant results from these projects are now available and can support

     )'      resolutio'n of this issue by supplying data that can be used to test the containment analysis codes. Additional experimental data on aerosol behavior (although project schedules may make them confirmatory to the mid-1984 time frame) are expected from several cooperative programs with EPRI and foreign countries.

I In addition to work currently scheduled, it is believed that the behavior of mixtures of aerosols of several materials should be studied as well as aerosol behavior in relatively large volume vessels under conditions with steam condensation. l ' e

                                                #              S

4 *

             ,                                                        Issue Paoer Title Issue 1.4.4 Deposition of Fission Products in Containment due to Natural Pro, cesses
                                          ~

NRC/ Contractor Leads NRC: W. Pasedag , 8CL: J. Gieseke 4 . Associated IDCOR Reports Fission Product Transport in Degraded Core Accidents. IDCOR 11.3 Report i Current NRC/ Contractor /IDCOR Positions Areas of Agreement:

1. Specific sequence / plant analyses are preferred method for estimating

                         .         source terms

2. Mechanistic analyses of radionuclide transport lead to understanding i

C~. ,, ., of factors controlling release

3. Thermal hydraulic factors are important to release and their analysis provides crucial input for release calculations i

4. Volatile iodine species should be considered at long times Areas of Disagreement:

1. NRC contractors believe discretized particle size representation and water condensation on particles are crucial to proper aerosol transport predictions

2. NRC contractors believe that mechanistic representations of pool scrub-bing and deposition in ice compartments are necessary Areas Needing Further Study:

1. Decay heating of deposited fission products and resulting thermal i hydraulic effects need further evaluation in their effect on fission product re-evolution and transport p ..

6 sd .

7,"j . e

 'T ISSUE 1.4.5 EFFECT OF ENGINEERED SAFETY FEATURES ON FISSION PRODUCT RETENTION

l. DESCRIPTION'0F ISSUE This issue is the extent to which fission products are retained by Engineered Safety Feature (ESF) systems during postulated degraded core and core melt accident sequences. Systems under consideration include those specifically provided for the control and removal of fission products (e.g. filtration systems) as well as those that can provide retention even though installed for other purposes (e.g., pressure suppression systems such as pools, sprays, and ice compartments).

2. IMPLICATION OF THE ISSUE TO REGULATORY 00ESTIONS EngineeredSafetyFeature(ESF)performanceintermsoffissionproduct

            . retention can have a significant effect on the outcome of severe acci-dents. Therefore, in order to estimate and improve plant safety during severe accidents thorough analyses of the ESF themselves are an inte-gral part of any overall accident analyses. The Severe Accident Re-
      +

searchPlan(SARP),NUREG-0900,explicitlyaddressesquestionsthat relate to overall plant safety and ESF performance: Is additional protection for severe accidents necessary? What is the likelihood that performance could be improved?

. Are improvements cost effective?
. How safe are existing plants with respect to severe accidents?

! . What additional research is necessary? l ! Developing a technical basis for answering the above questions for ESFs is a primary aim of the fission product control program that is being conducted in support of the reassessment of regulatory assumptions of severe accidents. WASH-1400 had only minimal technical bases for nearly all ESFs except sprays. This is especially true for conditions exceed-ing the design basis. Following the development of appropriate tech-nical bases, a complete evaluation of ESF performance can be made in the context of the severe accident reassessment.

3. SUBISSUES Several subissues derive from the limitations of the scope of the anal-yses of ESF retention estimates, from the uncertainties of the esti-mates, and from technical questions raised by the use of available analyses techniques. Present subissues are identified as follows:

                                     ,                            [Of$*N c 30 l,S[

              ^
 ;- h-[

l .- i1 4 [-

1) To what extent have the effects of engineered safety features been evaluated in available analyses of fission product retention in LWR containments?

l

2) How should'the effects of engineered safety features be analyzed i for

                                . sprays,'

i . water pools,

                                .      ice condensers,
;                               . filtration systems, and
                                . reactor building coolers.

l 3) Does removal of fission products from the containment atmosphere J imply that they are " retained"? This requires study of

                               '. aqueous chemistry and gas / liquid partitioning, i

j

                               . water flashing on containment failure, and
                               .      resuspension of fission product materials associated with degradation of ESF effectiveness, e.g., filter igniti~ due to local fecay heating and water droplet aerosol formativ from -

recirculating spray water. -

4) To what extent are ESF models valid under severe accident conditions?

This requires study of l l g ., (.. ESF equipment performance under accident conditions and the effects of energetic containment events, e.g., hydrogen com-l bination and steam explosions. i

5) To what extent do uncertainties in ESF effects influence overall uncertainties in the evaluation of severe accident consequences?

l 4. STATUS OF UNDERSTANDING The status of NRC contractor understanding which will be available for the mid 84 decision can be sumarized as follows:

1) The effectiveness of spray removal of iodine and aerosol species j is believed to be fairly well established. Because sprays are i

thought to be very effective, uncertainties in the actual effec-tiveness will probably not make a major contribution to uncertain-l ties in accident consequence evaluations. 1 { 2) The effectiveness of water pools has only recently been evaluated inamechanisticfashion(SPARCcode). Because the severe accident j source terms for some sequences in BWR plants are known to be very i l f

y- . . t

                                                  /
              ;     .-            .             ./

it . sensiti e to the water pool scrubbing effectivenass, it is clear , that uncertainties in this area are very impcrtant. Only now are 1 data appearir.g that will help validate the pool scrubbing models. Since parti,cle size is so critical to pool scrubbing performance, the definition of the aerosol entering the pool is paramount to ptedicting ' pool performance.- The data supporting the input particle parameters are just surfacing and Overall pool performance fs still uncertain.' ' '

,j

3) The basis for the evaluation of the effectiveness of other ESFs, e.g., retention characteristics of ice compartments of ice conden-

ser containment systems and coolers, failure modes of filtration systems, is limited. Hence, any current evaluation of the effects of these other systems on fission product retention must be assoc-iated with large uncertainties. An analytical model has been devel-oped to predict retention in ice compartments, the ICEDF code, Although this code is based on well understood particle capture

, mechanisms, there are no data for validation. Uncertainties are primarily due to poor understanding of gas flow patterns through i the ice beo' and and of the availability of ice surfaces for aerosol depletion. , The status of IDCOR understanding appears to be based on pragmatic , decisions concerning bounding accident sequences. As a result, ESF ' p effectiveness ~is limited by the conditions predicted for these se-4 quences. For example, it is predicted that a significant portion of lCe the fission product release will bypass the suppression pool, no active ESFs are operational, and ice is depleted prior to release via ice compartments. In general, when ESF systems are considered, non mechan-istic treatments are used. < For active ESFs:

a. Compartment filters use a removal rate constant

b. Compartment sprays use a removal rate constant as in WASH-1400 4

For passive ESFs:

a. 8WR suppression pools are modeled using a single user specified decontamination' factor (DF). The preferred value is DF = 1000. ;

b. Ice condenser systems use a settling only mechanism (onto struc-tural materials) when the ice is depleted. The details of particle capture when ice is present are'undeiWed. ,

                                                                          ^
                                                                    ^

i 5. NRC CONTRACTOR POSITION l The current NRC coatNctor position rehtive to the source term can be i summarized as follows:

       , .         . . . _                                              _ _. _ .;n c r -                                 ~~~ -                    - ' ' '  - -  ~

9

                            .                                        s

a. ESFs can be extremely important in terms of fission product reten- '

tion. This does not mean that all ESFs are equally important nor does it mean that each ESF performs in a predictable efficient manner under all conditions. These reasons generate the next por-tion of the position statement. ,

b. Validated mechanistic analytical models are necessary for a'dequate source term analyses. The Reactor Safety Study (WASH-1400) used simplistic assumptions for ESF behavior (except for sprays). To upgrade the source term analytical methods, mechanistic models are now being developed for suppression pools, ice beds, coolers, fil-ters, and even sprays.' .

c. The contractor's evaluation of the use of a constant DF for sup-pression pools is that this non mechanistic approach is too sim-plistic. We suggest IDCOR adopt the use of one of the mechanistic codes currently being developed. We find that the IDCOR position on sprays is adequate and that the use of settling only for the ice condenser is adequate only for the cases when the . ice is depleted.

d. Since full understanding of ESF performance is not complete by both A

the NRC contractor and IDCOR, it is suggested that the confirmatory work discussed.below should be conducted.

6. CONTINUING CONFIRMATORY WORK

      ,  3                      The overall approach to the resolution of this issue and the associated
      "~

subissues is through the research performed under the SARP together with the information obtained through review of the IDCOR and other pertinent work. Here, the SARP projects that will support resolution are identified together with the timing of the availability of key results. This is followed by an attempt to identify those SARP proj-ects that will satisfy the information needs for the subissues identi-fled above. In addition, an attempt is made to identify information needs that are not expected to be satisfied by current SARP projects. SARP Model Application Projects. The SARP model application projects that will have input to the resolution of this issue are Containment Analysis (A1198), and Fission Product Transport (B6747). Because of the project schedule, the analyses conducted using CONTAIN will not be l complete but the NAUA analyses with ESF models will be available in ! time for the mid 84 decision. SARP Model Development Projects. The SARP model development projects which will have input to resolution of this issue are: Containment Analysis (A1198), and Effectiveness of LWR ESF Systems Under Severe Accident Conditions (B2444). The output from these projects is being reflect ~ed in the model applications projects cited above and through this will have input to the $id 84 decision. l , 4 l

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

            ;            .                              s 1

SARP Experimental Projects. The SARP does not currently have any expe- {, rimental projects that will support resolution of this issue. However, work performed as part of B-2444 has already identified the need for experiments to assure that validated models are ultimately available. Areas.of experimental investigation idantified include particle reten-tion _in ice condenser systems, deposition of aerosols in containment t coolers, and system reliability under severe accident conditions. At this point the approach to and anticipated timing of the availabil-ity of information sufficient for resolution of each of the subissues identified above are discussed. (Note: it is believed that the issue resolution will proceed through resolution of the specific subis:;ues.) Subissue 1. The source term reassessment project (B6747) involves use of a complete set of models to evaluate the effects of ESFs. Models developed under project B-2444 have already been used to predict the extent of fission product retention in BWR suppression pools and PWR ice compartments (SPARC and ICEDF computer codes, respectively). Subissue 2. Validated analytical models should be used to analyze the retention due to each of the engineered safety features identified above. This allows the effectiveness of the ESFs to be evaluated on a plant and sequence specific basis. Subissue 3. Current ESF models do not explicitly model the loss of

       - 3                         trapped fission products associated with the degradation of ESFs. The
      ..'                          question of resuspension therefore needs to be examined in SARP Phase II.

Subissue 4. The effects of energetic events on the resuspension of fission products removed by ESFs.is not currently being evaluated in SARP. It is, therefore, necessary to develop a basis for resolution of this issue and/or to determine how to deal with the uncertainties that derive from omission of its resolution. Subissue 5. Resolution of this subissue is not currently being pursued in the SARP. The focus of the SARP project seems to be on minimizing the importance of the issue by developing improved methods for the analysis of ESF effectiveness and thereby reducing the uncertainties associated with evaluation of ESF effectiveness. L _ e s' (.- .

                   '(                                                               ' . a.)                                                    '
                                                                                                                                                           ~

SUPetARY OF NRC/ CONTRACTOR /IDCOR ISSUE PAPER STATUS ISSUE 1.4.5, EEFECT OF ENGINEERED SAFETY FEATURES ON FISSION PRODUCT RETENTION NRC/CONTR. LEADS ASSOCIATED IDCOR REPORTS CURRENT NRC/ CONTRACTOR /IDCOR POSITIONS P.C. Owczarski/ Fission Product Transport in. . Areas of Agreement - s ,. PNL Degraded Core Accidents -

1. Treatment of spray removal of. particles is l-IDCOR 11.3 report similar; however, IDCOR method is less mechan- !:

                                                                                            *istic                                                         l

2. Mechanisms of particle removal on structural s

materials of ice compartments of ice condenser containment systems are essentially the same.

                                                                                  . Areas of Disagreement IDCOR use of DF = 1000 for pool scrubbing is not adequate. Mechanistic model for pools preferred.

O l

s .. , ISSUE 1.4.6

DEPOSITION OF FISSION PRODUCTS IN OTHER PLANT STRUCTURES

1. Description of the Issue This issue concerns the deposition and retention of fission products along pathways and in structures beyond the containment barrier up until the point of their release to the environs. Because this issue is interpreted as involving the evolution of a source term to the environ-ment, natural processing leading to re-evolution of deposited materials as well as those affecting transport along these pathways and transport after release are to be considered. Therefore, the scope of this issue

            ,is interpreted ,to include:

a. Deposition of fission products in other

buildings and along pathways outside the containment up to the point of their release to the environment.

b. Other processes occurring along these pathways that can influence the quantity and characteristics of fission products.

Note that in some cases these are imposed as well as naturally occurring processes involved (e.g., stand-by gas treatment system).

2. Implications of the Issue for Regulatory Questions The major technical input to the establishment of positions on the regulatory questions will occur through analyses (implemented through application studies) to support decisions on the questions:

(1) How safe are existing plants with respect to severe accidents? (2) How can the level of protection for severe

          -              . accidents be increased?

Since such analyses include evaluation of the " Deposition of Fission Products in Other Plant Structures" as a key step, it is clear that sub, issues raised by the approach or approaches used in this segment of FolA44-98 c al

                            ^           -             ^                                        -          ^

+ - - - - - - - . -

i ,' - t 1 . 2

(^

     \
         ~

the analysis process will have implications for the position taken on

                              .these questions. Also, the outputs from the overall analysis process that will be used most directly in consideration of these questions will be:
'                                                       a.      Estimates of some measure of plant safety

b. Estimates of the uncertainties associated with the plant safety estimates.

Deficiencies in current or past analyses or analytical methods have direct impact on the regulatory question, i.e., what additional research or information is needed? Subissues arising in connection with current estimates and associated unceytainties for the rate and magnitude of release ,' fission products are identified below. 4

               ,                      3.        Subissues
          '                                              The subissues are derived from consideration of analyses that have been performed, realism of available analysis techniques, and uncer-tainties associated with current estimates of the deposition of fission products in other plant buildings.
                                           -             The subissues which we have identified are the following:

(1) For which release pathways of which accident sequences is deposition in other buildings potentially important? Resolution of this i will prevent the scope of analysis from being needlessly broad. However, this resolution will probably have to wait until identifica-tion and description of risk-dominant accident sequences is nearly complete. (2) What is the best approach to modeling the reduction of releases via deposition in other structures? (.3),.What is the potential for releases becoming airborne outside of containment through the flashing of leaked reactor coolant? (.- (4) What are the temperatures, pressures, and humidities that affect deposition in struc-tures outside containment?

. i !r

t *

(5) What LSF andis the effectiveness normal ventilation)offor filters (both attenuating j _ the release of iodine and aerosols to the i environment? Are filters outside contain-ment likely to fail due to aerosol over-l loading or fission product decay heat? j (6) To what degree does containment behavior ' (e.g., failure mode) affect the thermal hydraulic environment, and potential structural failure of other plant structures. , 4. Status of Understanding The status of current technical understanding for each of the > subissues and the underlying data base are outline below. Subissue 1: Containment Analysis (A1198) and the several severe

accident sequence analysis prograns will provide technica1 input to the l resolution of this subissue. Programs that may identify penetration failure locations (A1364 and A1375) may also point to release pathways 1 for which deposition would be significant. l Subissue 2. Deposition phenomena in structures outside contain-ment are included in the phenomena inside containment, so this subissue l can probably be resolved by applying the models developed to describe l aerosol and vapor behavior inside containment. For some models of ' containment fission product behavior, analysis of deposition in other structures can be obtained by simplying extendir.g the spatial modeling to include other structures. Subissue 3. The staff will attempt to resolve this subissue j by us.ing efisting technology and methodology previously developed by a i consultant.

i . . .-

   '                          ,         'Subissue 4. Existing methodology and the results of the several accident sequence p,rograms may be sufficient to resolve this. If further refinement is required, thermal hydraulic codes can be applied to plant st[ucturesoutsidecontainment.

6

    .-                    -..-_-.:                                             =           --                    - ---

4 Subissue 5. The information generated from resolving Subissue 2, together with the information from resolving Issue 1.4.5, " Effects of Engineered Safety Features on Fission Product Retention", can be used to resolve this. However, for most accident sequences, the experience gained in design basis accident evaluation will suffice for a reasonable estima-tion of filter effectiveness. Subissue 6. Present thermal hydraulic models are adequate to assess the thermal hydraulic response, including the likelihood of struc-tural failure in other plant structures. However, the great variability of structural and system detail from plant to plant make a generic assess-ment of the effects of other plant structures extremely expensive. Parametric studies, therefore, may be used to assess the range of poten-tial effects of other plant structures on the estimated release, so that - more detailed assessments will be limited to those sequences where this {A . O.., a . effect may significantly impact the overall risk. Overall Status of Understanding. Because some accident sequences (for example, V sequences and containment failures which do not result in a direct release to the environment) involve " Deposition of Fission Products in Other Plant Structures", the understanding of this issue will affect the understanding of.results of appli:,etion studies directed toward quantifying releases to the environment from potential severe accidents. A conservative position would be that deposition in other plant structures cannot be quantified with any assurance and, therefore, should be assumed not to happen. Such a position seems unwarranted since a more mechanistic understanding of the physical processes affecting deposition exists. A more mechanistic position would be: for accident sequences for which releases to other plant structures are predicterd, deposition could occur via fallout, impaction, or sorption onto large, cool surfaces; and that this could reduce releases directly

              -           or reduce the'llkelihood or filter failure from overload or decay heat.

l , !/ - Deposition phenomena are fairly well understood. Under condi-b tions expected to be. typical in other plant structures (moderate temperatures, and pressure and relatively low aerosol concentrations),

(. 5 the body of information that can be used to validate codes is large. Based on current source term methodology, the reduction of releases to the environment via deposition in other plant structures can be estimated using existing mo'dels and analytical procedures.

5. NRC Contractor Position 1 It is the NRC contractor position that the ASTP0 efforts in i

predicting source terms using the MARCH, modified NAUA, and SPARC codes as part of the BMI-2104 suite of codes provides a reasonable best estimate basis for supporting the mid-1984 decision. Work under way will provide estimates of the range of uncertainties in such predictions but the current understanding of the technical issues is sufficiently well developed to proceed with the mid-1984 decision while recognizing that an expanded data base and confirmatory research is under way through SARP-Phase II. ,

          $                       Based on a review of the IDCOR 11.3 Report, it is the NRC contractor position that either a discrete size range (sectional) approach to the size distribution should be adopted or the adequacy of the IDCOR log normal size distribution approach used in RETAIN should be demonstrated.

Further, it is expected that water condensation on aerosol particles in other buildings may have a significant effect on attentuation and its In conjunction with neglect in the RETAIN approach should be supported. predicting the effect of condensation, detailed thermal hydraulic predic-tions should be made for other buildings in the same manner as is done for the containment.

6. Continuing Confirmatory Work l

The resolution of this and the associated subissues is pursued through res'earch performed under the SARP together with the information obtained through. review of the IDCOR and other pertinent work.

  ,e                               '
                         '           SARP Model Application Projects. The SARP model application p'ojects r        which will.and are generating information pertinent to the

           - ~~                         __                               _ __ _ ,_

g (.

resolution of this issue: Containment Analysis (A1198) and Fission ProductTransport(B6747). SARP Model Development Projects. The SARP model development projects that will support resolution of this issue are CONTAIN Develop-ment (A1198), MELCOR Development (A1139), and Fission Product Transport Model Development (B6747) will be confirmatory to the simplified models and assumptions employed in evaluations available by mid-1984.

l 1

r ! SARP Experimental Projects. The SARP experimental projects that will support resolution of this issue are Post Accident Fission Product Chemistry (80453) and LWR Aerosol Release and Transport (80121). Significant results from these projects are now available and can support resolution of this issue by supplying data that can be used to test the

      '                               fission product transport code. Additional experimental data on aerosol C, , ;
                                                     ~

behavior (although project schedules may make them confirmatory to the mid-1984 time frame) are expected from several cooperative programs with EPRI and foreign countries. In addition to work currently scheduled, it is believed that aerosol behavior in relatively large volume vessels under conditions with steam condensation should be studied. m I ^ \: . o e

Issue Paper Title i'- Issue 1.4.6 Deposition of Fission Products in Other Plant Buildings _ NRC/ Contractor Leads NRC: W. Pasedag BCL: J. Gieseke Associated IDCOR Reports Fission Product Transport in Degraded Core Accidents. IDCOR 11.3 Report Current NRC/ Contractor /IDCOR Positions Areas of Agreement:

1. Specific sequence / plant analyses are preferred method for estimating source terms

2. Mechanistic analyses of radionuclide transport lead to understanding of factors controlling release f 3. Thermal hydraulic factors are important to release and their analysis

~~

provides crucial input for release calculations Areas of Disagreement: -

1. NRC contractors believe discretized particle size representation and water condensation particles are crucial to proper aerosol transport l predictions

2. NRC contractors believe diffusiophoretic deposition may be an important mechanism and should be considered

3. NRC contractors believe that particle removal in SGTS should be considered

4. NRC contractors believe that thermal hydraulic analyses of other buildings should be at same level of detail as those for containment Areas Needing Further Study:

1. The effect that failure modes may have on fission product behavior l

in other buildings l(- {

l - f//f7 llz[//.5 l.'

      .!                                                          ISSUE 1.5.2

( FOOD CHAIN TRANSPORT i

      ,                 1. Implication of the Issue to Regulatory Questions
       ?

This is a political rather than a technical issue. Well before crops i wouTd be impounded because of excess crop contamination, land would be

      ;                    interdicted because of external gama and beta dose rates. Interdicted 1and and crops are written off as part of the estimated costs of severe i

accidents. From a technical standpoint, generic environmental transport models and parameters are developed sufficiently for risk analyses and

environmental statements for hypothetical accidents. Site specific issues

      !                    can be resolved as individual specific issues are raised in individual
   ;                       licensing cases.
      ,                    In the event of a real accident, concentrations of radionuclides in food would be measured. The measured values would be compared with i                     the emergency reference levels for contaminated food recommended by the FDA (47 FR 47073-47083).

Politically NRC must interact with FEMA EPA and HHS (BRH/FDS) at the

' Federal level and with State and local agencies on site specific issues.

The potential for contamination of milk and water is a visceral issue, especially as children are associated with the milk pathway. ( 2. Subissues Reentry criteria for interdicted land, crops, reservoirs, milk sheds and watersheds. - Insurance (i.e. Price-Anderson) issues s Protective Action Guides for food chain

3. Approach to Resolution

'i

1. Apply current models for all generic accident consequence estimates.

. ;.$ 2. Address site specific licensing issues individually. i 3. Maintain a low-level .(cost and effort) program to improve the current models. i 4 Status of Understanding As stated above. [ 5. NRC Position This is a political rather than a technical issue. It is a minor issue. f c 4 l l- . -_.. ... l . . . . .. , . , . . , . , , . . .

                                                                                 / kdb 8v f                                           ISSLE 1.5.2                        8M AN
                                                                                           ~ 1/s.o/P3 FOOD CHAIN TRANSPORT

1. Implications of the Issue to Regulatory Questions Following a severe accident, the radionuclides released may deposit on vegetation or crops. These in turn could be consumed by humans or by animals furnishing food for people. Food chain transport is then one of the processes which contributes to off-site dose and therefore affects the answers to: How safe are the existing plants with respect to severe accidents? How can the level of protection for severe accidents be increased?

The food chain transport issue is common to all types of plants, however, the local agricultural practices and food consumption patterns would affect the population dose and individual doses from ingestion would be determined by the individual's particular eating habits. The importance of the food chain transport issue may therefore vary at different sites. Also, k food chain transport depends on the environmental dispersal of the deposited radionuclides which is determined by the weather at the time of the accident and other site specific factors. The time of year at which the accident occurs also affects food chain transport. Calculation of the dose from food chain transport for any particular accident scenario requires knowledge of the dispersal of the radionuclides to the environment, local agricultural practice, food consumption patterns, and dose-exposure relationships. The cost associated with interdiction of vegetables, milk, milk products, and meat is affected by the transport of radionuclides within the food chain. ; l l Since food chain transport occurs outside the containment, it is ; subject to seasonal, daily, and hourly variations in the weather and in l seasonal variations in agricultural practice and food consurnption patterns. All aspects of consequence analysis (including food chain transport) currently l handle these variations using a Monte Carlo analysis of a large number of weather scenarios. Th!s practice simulates a severe accident which is expected to occur randomly in time. - l ( 4tA-t+ '123 I c,, 33 1

2 s

2. Subissues Direct Contamination of Vegetation Activity level for interdiction Fraction of deposited material retained on the vegetation Reduction of material on vegetation due to weathering Ingestion paths o milk o milk products

  .                                            o           vegetables o           meat
  ',                                           o           other foods
 .                            Incorporation of Contaminants from Soil into Vegetation Liquid Pathways.

3. Approach to Resolution Phase I (not SARP) -

International Benchmark Committee survey on chronic effects should provide insights regarding uncertainties in food chain transports. Phase II

a Resolution of the food chain transport issue is based on the

  ,                           development of improved models to be incorporated into MELCOR. MELCOR is

,i intended to model the essential features of population dose due to exposure to 'I radionuclides deposited on the ground and to ingested radionuclides including j food and drinking water pathways. ELCOR will not be available until Phase il II.

t

4. Status of Understandina

,a Currently most consequence modelinn is done with CRAC2 or codes with

   ,                          similar food chain transport routines. These codes use simple models to I!

!1 l1 !l ll

                                                                                       ._      ..       . . . ~ .. . . .

       . y . *.

3 l . (

l. determine which areas of crops are interdicted and the length of time the areas are interdicted. Simple models also estimate the population dose from i ingestion of milk, milk products, meat, vegetables, and other foods. These

,j mCdels generally are based on averaged values for food intake by animals and

  ,                       man. Model'ing of the dose from soil contamination and subsequent uptake by plant roots and consumption by humans is thought to be an inefficient mechanism i                       of radiation exposure.

,f , Currently liquid pathways are not treated as completely as j atmospheric pathways because liquid pathways are shown to be the less l- significant contributors to accident consequences. t . .j 5. NRC Position ,( , O e h i I

  /
 .)

4 L i

I 3 l i [ 'f (i l' 1 i

      ^
      ..I     1*gyy. i,* tit,^~:f P -         _

f Iwk s. i (' BMT Adf ISSUE 1.5.3

                                                                                                             ~f/N[G 00SIETRY AND HEALTH EFFECTS

1. Implications of the Issue to Regulatory Questions Since the dose / health effects issue affects the predicted consequences of an accident the position on the issue can affect the answers to: How safe are the existing plants with respect to severe accidents? How can the level of protection for severe accidents be increased?

The dose / health effects issue is generic to all LWR's. All issues related to the calculation of the disposal of radionuclides and the transport paths to man are related to this issue. Also related are the decontamination and interdiction issues since long tenn dose and therefore health effects are at least partly regulatable following a severe accident by interdiction of crops, relocation of the populatiori and decontamination. (

2. Subissues s

Linear health / dose relationship

  ,                       Threshold s

Biological repair. ,, 3. Approach to Resolution The approach to resolution of most of the health / dose response j issues is based on development of improved models to be incorporated into jj ELCOR. An improved radiological health effects model is being developed by . 'i I Harvard for Sandia. ELCOR will not be available during Phase I. 4 Status of Understandina Exposure to radiation can result in various health effects; fatalities, cancers, genetic damage, etc. The type and extent of the health effects is related to the dose received. The dose can be from external or Fora-1A 4 8 . C.f 3h

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

g. .

  ,:; .l $

f internal sources. The dose can be to the whole body or concentrated in a

particular organ such as the thyroid. The health effects for a given dose are f, not clearly established. There is particular disagreement regarding the health l

effects of low doses or low dose rates. Further, there is considerable

!                  uncertainty in the dose received even if the radionuclide concentration in the arta is known. Dose can result from external exposure to a radioactive cloud or radionuclides deposited on the ground or building or it can result from
 ,                 internal exposure due to inhalation or ingestion. The calculated dose depends on the radionuclide concentrations, the time of exposure, the breathing rates.

eating habits, biological half lives, and other factors. Current predictions in CRAC2 are made using concentration factors, assumed breathing rates and ingestion factors which are based on a large number of assumptions regarding the reference manual.

5. NRC Position

( - 1 e t G

    - p- -- - -- - r-  : ,                ., ;; s , .              . . .
                                                                         . ., ,..y;ms.    < ,; t_. ;..,-;;y
                                                                                             ,              .ur. .q g . - .-

= :- = - -

        .".     'n DAkFT1/zf/n W     [.                                               ISSUE 1.5.3 f~

1 j[ Dosimetry and health effects

l. 1. Implications of the Issue to Regulatory Questions i Adverse health effects are the (perceived) bottom lines of risk j analyses. The Commission's adopted and trial use safety goals are phrased in tems of the risk of adverse public health effects.

t i These are highly visible outputs. Dosimetry is a means to the i' endpoint calculation of health effects. All health effect results

     ,                      are sensitive to the source term issue, but early fatality and j                        early injury estimates are very sensitive because of the high doses i                       required to induce these effects. Early health effects are also i                       very sensitive to source term and emergency response assumptions.

1 A reduction of current core release fractions by a factor of ten (for the largest plants) would result in few, if any, calculated early fatalities assuming only leisurely emergency response. In contrast, latent cancer fatalities accunulate to hundreds of miles and are insensitive to assumptions regarding emergency response within rational distances (10 to 50. miles). Estimated latent cancer I} fatalities are directly proportional to source tem, and re-entry criteria, and inversely proportional to interdiction criteria. Dosimetry is a current issue as regards current health effect models. D00 and DOE are currently sponsoring an international {' program to re-evaluate the atomic bomb dosimetry and health effects. Revised dosimetry is expected, which will affect dose / health effect relationships. This will evolve as a major issue over the next one through five years. Major changes are not expected, but any 1 change could be perceived as major in this highly visible, emotionally

   ;                       charged area.                                                                 i 1

Except for one issue, recently identified, CRAC 2 dosimetry, per se, is I not a current or forseeable major issue in PRAs. The new issue is dosimetry 1 1 for, and health effects of, extremely high skin doses due to beta emitters deposited on the skin. This is a new concern because of the

potential for severe skin burns, especially in conjunction with damage to the bone marrow. The concern applies to SST2 as well as SST1.

t i 1 2. Subissues . 3 Emergency Preparedness and Response 1 c Environmental Statements (Accident risks)

  ;                        Accident insurance (Price-Anderson).

j Atomic bomb dosimetry and health effects. i i i( Fo I A - W *>8 C. ; 1s i k

     -, --.c:r.
             ,-     . y .,;.7,g    g, g g;;       g c-    .

_.. _ _ 11 ___ _ _ _ _ _. _ m_.. ._ - -

    . 7, . N:

s, 4

3. Approach to Resolution RES has formed an Advisory Group of noted experts on dosimetry and' health effects to develop state-of-the-art health effect models.

Major revisions of current models are not expected. However, the introduction of a new concept into the risk perspective is expected. This concept will result in a new measure of late health effects, i.e. life shortening, as compared to fatality counts. Even though health effect models are not expected to change much (the technical perspective), the political aspects of this issue demand a visible, well funded program for the foreseeable future. Revised health effect models are expected in mid-1984 (FY). Ul ti-mate resolution is not forseeable in a useful time frame (ca 2020AD or later).

4. Status of Understanding The status of the subject matter is well understood - it is evolving continuously as the Japanese survivors age. Data on health effects in humans is sparce and the health protection objective is to keep it sparce. The principal thrust'of current efforts is to bound the uncertainties.

5. NRC Position

  ' -.                      Dosimetry and health effect models currently in use for PRAs are acceptable. If the NRC Advisory Group proposes new models in a timely fashion for application before mid-1984, they will be used; if not, they will be used as they become available.

a S I 4 e

s e

                .. e       9      *. - .

a [7 ,, , , ,j , ,. **

pgj7p.3- glzy}/ 5

                                                                                                   ~
  ?

(- ISSUE 1.5.4 MODELING OF (OFFSITE) EMERGENCY RESPONSE

1. Implication of the Issue to Regulat6ry Questions Emergency response by the public is the last factor of the defense-in-depth equation. Of the various possible protective actions available to the public, evacuation and shelter are the most visible

  ;                       and visceral, with the issue of prophylactic drugs (e.g. potassium iodide, KI) closely following. Many other protective actions could
   ,                      provide substantial protection, such as ad hoc respiratory protection, but these are nomally not accredited in risk analyses.

Key parameters in the evacuation / shelter issue are distances and times: distances to which evacuation and shelter would be required, within what time frames. These are central to siting and emergency planning regulations. Policy and politics are intimately bound to the protective action issue. FEMA. State and local agencies and other Federal agencies are involved with the NRC. Conceivably, an adverse action or ruling by a third party (FEMA or a State or local agency) could cause the NRC to withhold or withdraw a license to operate. (- Central to the protective action issue is the question most often unstated; what is to be accomplished by the protective action (s)? Avoidance of exposure (dose)? Reduction of consequences (health effects)? Reduction of risk? If the latter, what level of risk reduction is acceptable? In a site specific analysis, what risk reduction is feasible under what circumstances? Further, is it acceptable to have a great disparity in risk reduction potential from site to site? Finally, how

 .                        much credit can be attached to emergency planning and preparedness (there is no data base, and none is expected)?

2. Subissues In-Plant Emergency Action Levels (timing of offsite emergency response)

Protective Action Guides Siting Emergency Planning Zones i Safety Goal . i Reactor Power Level i l Severe Accident Sequances r: 1 'i j FO/A- U dM llI. c,/%

 "...            .......,......_.;,..,_.z.,                 ,g  ., ,.,_,,..,.....,.(._yz,,

u. a- ,

3. Approach to Resolution

1. Revise Emergency Planning Rule (App. E to 10 CFR Part 50) to provide reasonable assurance that early fatalities and injuries can be avoided in the event of a severe accident.

2. Adopt an ALARA approach with regard to other health effects.

4. Status of Understanding All feasible protective measures are well understood in principal.

Many feasible protective measures are not accredited in risk analyses. In particular, early evacuation (before a release) by populations most at risk is not currently accredited. This is a political as well as a technical issue.

5. NRC Position Immediate precautionary evacuation of nearby areas (2 to 3 miles) in the event of a core melt accident, or the clear and imminent threat thereof, would provide substantial reduction in the risk of early fatality or injury. Immediate shelter elsewhere would reduce the risk of latent health effects. In the event of an actual major release to the atmosphere expeditious relocation by the public from i highly contaminated land would substantially reduce individual risks t, further.

In the near term (FY 84/85) the NRC, working with FEMA, will establish a preferred emergency response strategy. The potential benefits will be illustrated in future Environmental Statements. It is emphasized that in the absence of data it is imperitive that the rule provide the assurance that the preferred emergency response can and will be accomplished. The probability of this will have to be estimated for various accident sequences. There will be no data base for this, so a high level policy decision will have to be made to set the probabilities for risk analyses. However, the probabilities are not needed to illustrate the potential benefits. i

i

                                   ~~;                                                        ' "               ""*

a - a .~ e: :{. _ . , . "'S-

                                                                                       ,, { '  . T *  C * ~ '     --

VGrsion 100683

           .1 Issue 2.1.1 Plant Categorization -

'i

I 1. Implication of the Issue to Reculatory Questions i

i j The ability to develop plant classes has great effect on the NRC regulatory decision process. Since it is desimd to make severe accident decisions by i mid-1984 on as many existing LWRs as practical, the issue of developing plant classes by evaluating the results of the existing PRAs (12) is highly important.

   ,               There is o jirect relationship between the issue and tne affected regulatory
   +

j questions. R e position of the issue can effect the answers to:

2. How should the Commission decide the severe accident question?

2.1 Should the decisions be generic or plant specific?

3. How safe are the existing plants with respect to severe accidents?

3.3 Which accidents are to be considered and which ones can be ruled out? . 3.4 Using these measurements, how safe are the existing plants?. 4

4. How can the level of protection for severe accidents be increased?

4.2 How effective are they?

  ?
                                                                                                             ~

j 5. What additonal research or infomation is needed? l 'j 5.1 What are the information gaps bearing on the severe 4 . t accident decision?

 .1                           5.2 Are data necessary for implementation of the decision?

j . Fora- N-W 7 c, 27 a:J.[ K

h" ~ ~ ' :

                                            ~~
                  ~~ Y                            *^

IV -

                                                                               '                       ~
                                                                                         '8..._.' T-

l l e .. 1 i

     ;                      5.3 Are data necessary for conformation of the decision?

f 5.4 Are there specific issues that require more data before they

     !                            are decided?

i:

6. Is additional protection for severe accidents needed or desirable?

6.2 What is the likelihood that perfonnance could be improved I by the alternatives available? i 6.3 Are the likely improvements worth tiie cost?

2. Subissues It is possible to group plants having similar dominant accident sequence characteristics?

Do the existing PRAs (12) provide enough information to make inferences on all existing plants? Can insights drawn from plant categorization be used in severe accident nalemaking? If so, how should uncertainties be considered? 4 '{ - Given the associated uncertainties, can plant categorization be used to identify weaknesses in plant design? What level of information is appropriate for plant categorization-- { function, systems, or component level? Can event trees for plant classes be fonnulated? If so, how l useful are they? 4

               .          e

_Y 4
E ^ *
i 3

il' Plants are categorized based on dominant accident sequences and

   .                    plant systems needed to mitigate the accidents, consequences resulted   j from these incidents are not considered. Does this have any impact 9

on regulatory decisions?

3. Approach to Resolution

  .          The Accident Sequence Evaluation Program (ASEP) is the primary NRC research
  }          to support the issue. ASEP is funded and managed by the Division of Risk Analysis of the NRC and the primary contractors are SNL and INEL.

The NRC approach uses the twelve existing PRAs as the starting point to develop plant classes. Dominant accident sequence classes are fonnulated based on the "like" characteristics of initiating events and system failures of the PRA dominant accident sequences. The plant systems and support systems needed to mitigate the dominant accident sequences are identified and a plant survey is

undertaken to obtain accident mitigation fluid and electrical system drawings

 ,           for as many plants as possible for all systems and support systems. Concurrent to the plant survey, the plants are grouped into appropriate containment types.

After the system drawings are collected, simplified drawings including the

 .'          important components and dependencies are drawn. The simplication process is based on simplifying Piping and Instrumentation Drawings (P&ID's) and Electrical
 ;           One Line Drawings.

Once the systems have been simplified, comparisons are made of the systems

 -'          (e.g., compare all AFW systems) considering significant differences in the i

[ redundancy, diversity, or support system dependencies for each system of interest. o I Emzwwm

L 1 : 11 -_ : - o

For each of the selected accident sequences to be examined, system designs for systems used to mitigate the accident in each plant are grouped, depending on similar system characteristics. These groups are used to form combinations l of known system configurations that are possible for mitigation of the selected ! accident sequences. These combinations are defined as initial plant classes l for the particular accident sequences of interest. For example, suppose all AFW systems could be grouped into three configurations; AFW l AFW 2, and AFW 3. In addition, suppose all high pressure injection systems (HPIS) could be grouped into two categories; HPIS1 and HPIS2. Furthermore, suppose that the known combin-

                     . ations of these system groups are only; AFWl/HPIS1, AFW2/HPIS2, and AFW3/HPIS2.

Then, for calculating the sequence frequency for a sequence involving the failure

          .-               of these two systems (e.g., TMLU sequence), three calculations are made. One calculation for each of the above known combinations, which are now defined as initial plant classes for this sequence. After the deductive modeling, quanti-fication, sensitivity and uncertainty analyses tasks, the next step is to use the quantitative results and important insights to collapse the initial plant classes to the final plant classes. The approaches on modeling and quanti-l' l

fication are covered under Issue 2.1.3, Quantification of Sequence Likelihood. Up to now, the results are presented in the following heirarchial format; by containment type; within con'tainment type, by sequence class; and within l ' sequence, by initial plant class. Figure 1 is a illustration of the ASEP approach to plant categorization. The next step is to collapse the initial plant classes, where possible, based on similarities in the dominant con-tributors, the sequence frequency and its bound for each plant class.

. Next, the results are arranged in a different hierachial format with the I containment type as the first level. The second level the plant class l .

j*

pggypg j l., APPROACH TO PU ATEGORIZATIONS ,

). !

Sequence Sequence information [ Classes Dominant Collapsed ** Final *** Under Final Frequency Upper and Initial Plant .(Best . Lower Dominant : Sequence Plant Plant Plant . Contibutors _l l Containment Classes Classes Estimate) Bounds Type Classes Classes Classes l

                                                         "A"-2 Trains                                                         3E-4                                              l

' of ARG 5 E -5 Hardware /T 8 M

                                                                                                   .T*PCS AFWS l

w/o F & B 1 E -7 Failures of AFWS , (Sequoyah, AM - Watts Barl i "F"-2 or 3

                                                      /               Trains of i

lT=PCS.AFWS AFWS w/o ! FAB i i g' (Sequoyah, "B"-3 TrainsWatts Bar, of AFWS kQuire) w/o F & B (McQuire)

                                                                                     "G"  2 or 3 Trains of SI Ice Condenser--

Free Standing AFWS w/o j I F & B and . .r; Still ains of !71 , v.

.. lSLOCA.HPISl \ 8 E -4 Operator Procedures J. .

lSLOCA*HPIS]1E-6

   .:.                                                \,                                                                      1 E -7        for Identifying 1;                       '                                            _                                                                   SLOCA and
    ;                                                     "C"-3 Trains "C"-3                                                                Initiating HP15
      '                                                   of HPIS        Trains of (Sequoyah,     HPIS Watts Bar,      (Sequoyah, McQuire)      Watts Bar, McGuire) 5.-l 1                  *This is an illustration only
                    ** Collapse classes by examining sequence frequencies, their upper and lower hounds, and their dominant c
                  *** Redefine collapsed plant classes by examining specific plants.

i . . . i.. ,'

     ;,                                                                        -s-1
     -                   and the third level is the sequence class. By the mfomatting the plant classes are redefined. The collasped plant classes intrinsically contain specific plants within their description, by venture of the specific plant's related system characteristics. After all sequences within'containmet type are identified as contsening similar combinations of plants for those
   .                     sequences, these similar combinations define the final plant classes and their associated character 1.stics. For each final plant class, the dominant accident sequence, the associated likelihoods, the relate.d range of values and the dominant contributors to the sequence likelihood are given.
   ;                     4. Status of Understandina ( As of 9/30/83)

A better understanding of the status of plant categorization can not be made at the time. Initial plant class identification is progressing

                                                                                                                   .4
    -                     and will be completed by mid October,1983. The number initial plant classes              f will be more than anticipated due to the large variability in plant design.

'i

< 5. NRC POSITION Given the large variability in plant design and the number of attributes i

that makes the accident sequences important, which include hardware characteristics, functional capabilities (e.g., for feed and bleed), and i special vulnerabilities to common cuase failures, the NRC believes the i' ) potentially large number of plant classes can be reduced to a more . manageable size by two means. I First, the NRC expects that some plant classes that differ with respect to less 1

    }                      significant attributes can be consolidated.                                Further, and more importantly, f

the NRC believes that plant classes can be combined if regulatory action

 ,i                                                                                                                                    >

l - i I

1 l can be proposed to eliminate their differneces in attributes. That is. plant classes that are different because of, for example, a particular HPIS dependency on component cooling water, could be merged if one SARP

   ;                         regulatory recommendation is the elimination of such dependencies. Such regulatory actions could be relatively narrow (for a specific problem or j                         a specific set of plants) or more broad (e.g., requiring all plants to l                         perform reliability analyses). Through the development of relatively

{ detailed initial plant classes and their subsequent collasping and the associated identification of regualtory recommendations, the NRC believes

 .                           that a defensible and manageable set of plant classes can be developed.

e . e o I a

w. _ ;

3 ., t i 7

V8 h ji J/de Issue 2.1.2 , gjQ(e I

                                                                                                                   'n')

TITLE: Identification of accident sequences ; , [v I ,

1. Implications of the Issue to Regulatory Que:.tf ons d ,

    ,                  The Severe Accident Sequence Analysis (SASA) progra,m 13 an integral               ;

part of Severe Accident Research Plan (SARP) and is interfaced directly with programs in the Division of Risk Analysis (DRA). Office of Regulatory Research. Risk assessment programs such as th'e Interim Reliability Evalu-ation Program (IREP) define high risk sequences for specific plants. The Accident Sequence Evaluation Program (ASEP) reviews the event tree from plant specific risk,assesssment. The ASEP program identifies dominant

   .;                  accident sequences bued on past PRA's and related studies which are then                           .

analyzed in detail by the SASA program. Detailed analysis of these high- > risk sequences is neeoad to determine appropriate operator actions and any need for specic1 instrumentation. Specifically recommended operational . techniques for managing accident recovery and algorithms to be used by the operator to prevent, diagnose, and respond properly to accidents are to be 4 provided as a basis for appropriate regulatory actions. The findings from the SASA deterministic studies are then fed back into the evaluation and modifications process in ASEP. Presently external events and sabotage,have not been considered in ASEP.

2. Subissues

[ Does one need pient specific W s)

an one look at this question , l

generically and make decisions for the existing plants as to whether or not we need additi nal features?

    .                  PRAs should be used as a tool for decision-making but it should not be used to give bottom line on, for example the probability of core melt.

hl H Y 10/04/83 Ost I 1 ISSUE 2.1.2 l 1 J l

j' 3. Status of Understanding i i i Attempt has been made to identify all assumptions, and their bases, in the , ,f SASA report. The effect of operator actions to the extent considered

  ;,                        practical are included. The probable operator actions from (a) review of
   ;                        the existing emergency procedures; (b) discussions with utility operating

( personnel; and (c) rehearsal of the accident sequence at the control room I simulator are determined. Operator actions that would aggravate the consequences wherever it was felt there was areasonable chance that the operator might take such action and the action would have a significant effect on the outcome of the sequence that was considered. It is intended

 .[                         that consideration of operator actions will be expanded in future studies through cooperative efforts with the Human Factors Branch of the Division 5                         of Facility Operations.

Considerations of the availability of systems and equipment under the

  ;                         conditions that would exist in the surrounding environment during an
  ;                         accident has been included in each SASA report. An example where the interface of ASEP and SASA program was shown beneficial is the following:
  ;                         PRAs generally ignore the effect of Control Rod Drive (CRD) hydraulic system in BWR Accident Mitigations. It was shown via the SASA )rogram that the CRD hydraulic system plays a major role in preventing core 4                          uncovery. The SASA analysis indicated:

(1) Effect of CRD hydraulic system should be factored into risk assessment and operator training (ii) Dominant loss of DHR sequences must be reevaluated. References to reportsinclude: - (1) D. H. Cool et al., " Station Blackout at Browns Ferry Unit One-Accident Sequence Analysis," Vol. 1. NUREG/CR-2182, November 1981. 10/04/83 2 Issue Paper Sec 2.1.2

          ?-     - ... -.             .

3.. . . ,,.

                                                                  ..  , , .  .,,..g         .,n. e y- .

l..... l l (2) D. H. Cook et al., "Less of DHR Sequences at Browns Ferry Unit One-Accident Sequence Analysis," NUREG/CR-2973, May 1983. (3) R. M. Harrington et al., "The Effect of Small-Capacity High Pressure Injection Systems on TQUV Sequences at Browns Ferry Unit One," l NUREG/CR-3179, September 1983. (4) N. DeMuth et al., " Loss of Feedwater Transients for the Zion-1 Pressurized Water Reactor," NUREG/CR-2656, May 1982.

 -           4.       Approach to Resolution The methodology for the deterministic analysis of severe accidents has been developed as far as possible using the existing analytical tools, and available funding and efforts are underway to obtain the improved tools (such as MELCOR) to progress to a higher capability.

The selection of accident sequences for analysis has in general depended on the identification of the dominant (most probable) sequences by previous _ probabilistic risk assessment (PRA) studies conducted by utilities or under the auspices of DRA (the IREP program). SASA is contributing to the development of probabilistic methods by supporting ASEP.

5. NRC Position The NRC position is to use both probabilistic and deterministic analysis to identify dominant severe accident sequences.

t 10/04/83 3 Issue Paper Sec 2.1.2 i i 1..-_...- , . . . . s . . . . .

I 4 1 I

   ;                                                              Issue 2.1.2 i

TITLE: IDENTIFICATION OF ACCIDENT SEQUENCES i i 1. Issue r

   ;                               The issue is the identification of accident sequences using deterministic and probabilistic studies.
    ,m
   ,'                   2.         Implidations of the Issue to Regulatory Questions
   ,                               The Severe Accident Sequence Analysis (SASA) program develops information needed to develop guidelines for regulation of accident management proce-dures and to develop technical basis for declaration of emergency action levels during an accident. SASA predicts the time to failure of contain-ment and important intervening events.

The Severe Accident Sequence Analysis (SASA) program is an integral part of Severe Accident Research Plan (SARP) and works directly with programs inthe'DivisionofRiskAnalysis(DRA). The relationship is: Risk assessment programs such as the Interim Reliability Evaluation Program (IREP) define high risk sequences for specific p* ants. The Accident

   .                               Sequence Evaluation Program (ASEP) reviews the event tree from plant i                                specific risk assesssment. The ASEP program identifies dominant accident sequences based on past PRA's and related studies which are then analyzed in detail by the SASA program. Detailed analysis of these high-risk sequences is needed to determine appropriate operator actions and any need j                                for special instrumentation. Specifically recommended operational tech-niques for managing accident recovery and algorithms to be used,by the operator to prevent, diagnose, and respond properly to accidents are to be provided as a basis for appropriate regulatory action.

1 - I fo/A-14 -128 d I 11/15/83 h( M 1 ISSUE 2.1.2 i

   .I i

m-- ~" ~' ' r < n - a ?. p* ,..s .: ' _ g __ -_'

i . ..

i . j The findings from the SASA deterministic studies are then fed back into j the evaluation and modifications process in ASEP as well as to the regula-j tory offices, NRR and IE. Currently, external events and sabotage have L not been considered in ASEP. i ,) Through projects co-sponsored with DF0, an attempt is also being made to j extend ther scope of human factors studies to severe accidents and to j develop the " Accident Management" element. i 1 j 3. ' Subissues - 2 4 Completeras of the PRA sequence description is a major subissue.

     -                  4. Status of inderstanding An attempt has been made to identify all assumptions, and their bases, in

the SASA report. The effect of operator actions to the extent considered practical are included. The probable operator actions from (a) review of the existing emergency procedures; (b) discussions with utility operating personnel; and (c) rehearsal of the accident sequence at the control room ,

simulator are determined. Operator actions were considered that weald i aggravate the consequences wherever it was felt there was areasonable chance that the operator might take-such action and the action would have l- a significant effect on the outcome of the sequence that was considered. It is intended that consideration of operator actions will be expanded in

                    ,         future studies through cooperative efforts with the Human Factors Branch of the Division of Facility Operations. A trial program is now underway.

s Considerations of the availability of systems and equipment under the conditions that would exist in the surrounding environment during an

   ;                           accident have been included in each SASA report. An example where the interface of ASEP and SASA program was shown beneficial is the following:
      ;                        PRAs generally ignore the effect of Control Rod Drive (CRD) hydraulic system in BWR Accident Mitigations. It was shown via the SASA program lA
  .i:

1 11/15/83 2 ISSUE 2.1.2 1 l1 li '1 : - q v.y~ m :q p p ~,= p y q w:, ; -

that the CRD hydraulic system plays a major role in preventing core { uncovery. The SASA analysis indicated: l (1) Effect of CRD hydraulic system should be factored into risk j assessment and operator training j (11) Dominant loss of DHR sequences must be reevaluated. References to report include:

 ;                 (1) D. H. Cool et al., " Station Blackout at Browns Ferry Unit One-Accident

{ Sequence Analysis," Vol. 1. NUREG/CR-2182, November 1981. m (2) D. H. Cook et al., " Loss of DHR Sequences at Browns Ferry Unit One-Accident Sequence Analysis," NUREG/CR-2973, May 1983. (3) R. M. Harrington et al., "The Effect of Small-Capacity High Pressure Injection Systems on TQUV Sequences at Browns Ferry Unit One," NUREG/CR-3179, September 1983. (4) N. DeMuth et al., " Loss of Feedwater Transients for the Zion-1 Pressurized Water Reactor," NUREG/CR-2656, May 1982.

5. Approach to Resolution l

The methodology for the deterministic analysis of severe accidents has been developed as far as possible using the existing analytical tools, and available funding and efforts are underway to obtain the improved tools (such as MELCOR) to progress to a higher capability. The selection of accident sequences for analysis has in general depended )

 .                       on the identification of the dominant (most probable) sequences,by previous 3,                         probabilistic risk assessment (PRA) studies conducted by utilities or
 !                        under the auspices of DRA (the IREP program). SASA is contributing to the development of probabilistic methods by supporting ASEP.

G f 11/15/83 3 ISSUE 2.1.2 i I i

     - y y 7-..;.y .y y p g 9.q;;p:. _

r- - -. - - . - --- --- - - - - - - - - - - - - - - - - l

         . .        ..e

6. NRC position

t . t The NRC position is to use both probabilistic and deterministic analysis E to ide.ntify dominant severe accident sequences. 0K ' identified, a tech-nical basis can be prepared to evaluate proposed accident management procedures and to indicate when different emergency action levels need to be declared. 4 V w

.'s

'i 4 s 4 i 11/15/83 4 ISSUE 2.1.2 e i 3'*.'***{'""'" ,

                            -****9,-

_ . , * * * ' g * ** *.* * ((A,a

                                                                         '**) . .,
                                                                                +

p , 4 * . .- t,, s* . * * * . - = = a,.

                                                                                     .--       - .u-                        : t . -_ *:     .          . _,___               _ _ _ _ _ _ _ _ _ _   _ _ _ _ _ _ _ _ _

                                    = _ . .           .    . . . =

b y- Versien 100683 I Issue 2.1.3 - Quantification of Sequence Likelihood 4 ['

1. Implication of the Issue to Regulatory Questions i

Qur.ntification of the sequence likelihood is an integral part of a PRA to prod;;c.e the accident probabilities, which are combined with associated i

j consequences to produce the overall plant risk. The position of the i

   ;                      issue can effect the answers to the following regulatory question.

t 4 j 3. How safe are the existing plants with respect to severe accidents? I

                                  ' 3.1 What should be considered in this measurement of safety?
   ;                                 3.2 How do the terms of measurement compare, including uncertainties?

I 3.3 Which accidents are to be considered and which ones can be

       ,.                                       ruled out?

3.4 Using these measurements, how safe are the existing plants? 3 4. How can the level of protection of severe accident be increased? 4.2 How effective are they? I

5. What additional research or information is needed?

'j 7 l 5.1 What are the information gaps bearing on the severe accident decision? U)

l 5.2 Are data necessary for implementation of the decision?

1

 't J                                  5.3 Are data necessary for conformation of tne decision?

5.4 Are there specific issues that require more data before 4 I j they are decided? , ,Og c,l 4* I i,

      ~ ~ * * * * ~ ' " -
                                      , , , j .1 : *  ,-.,..    . 38  g...--   g -      *y
                                                                                               ,8_ ,*    ,

                                 -.       . -- - - - - -                  -                   ~

l J \ r 3

    ,'                                                                                                       (

j 6. Is additional protection for severe accidents needed or desirable? t l

   ,                                                     6.2 What is the likelihood that performance could be improved by i
  !                                                            the alternatives available?
  ?

8 i j 6.3 Are the likely improvements worth the cost?

 )                                  2. Subissues How is sequence likelihood quantified?

i { What are the attributts of the uncertainties and which one quantifiable? What statistical meaning can be associated with the generic estimates I of core melt frequencies and associated uncertaintles?

                                           -             Can decision be made based on the generic estimate of core melt frequencies and associated uncertainties?

3. Approach to Resolution The Accident Sequence Evaluation Program (ASEP) is the primary NRC research
+ to support the issue. ASEP is funded and managed by the Division of Risk l' Analysis of the NRC and the primary contractors are SNL and INEL.

l The NRC and its contractors use two approaches to quantify accident sequence 'i likelihood. The first approach is an interim reassessment of the core melt frequencies of selected PRAs. It is a limited rebaselining of the core melt ', frequencies of the RSSMAP/ WASH-1400 plants to be used by the NRC SARRP program for its initial risk reduction research. This quantification approach will be i* superseded by the second approach that is more compretensive in nature. 4

    * **

eepop-9,a+a e me a &*e-. e e
f* *p** * *-9 * * *e q e oq ****-M,*4 -

                                                                                                                         ,,e p  ,,   e e ,.
                                                                                                                                             **e_*  ene * * *
                                                                                  . - - - -                =         - - - - . = . . .~. x:-   -- -

1 -l i : - lf . ( i

     .;                           The rebaselining approach is a qualitative assessment of the sequence likeli-                                   )

I

 .:                               hood of the RSSMAP/ WASH-1400 plants (Oconee Sequoyah, Calvert Cliffs. Surry.
     !                            Grand Gulf and Peach Bottom) using the insights gained from existing PRAs
  ,  I 3

and other information sources including Station Blackout Study, Accident Pre-4 l cursor Study, human factor work and safety analysis reports. The rebaseline 4

     .                            steps are as follows.                                                                                           ,
 !/

a. Note the sequence class in which the particular PRA dominant accident sequence being examined, is categorized.

I

    )                                     b.      Review the insights applicable to that sequence class and determine if a particular quantitative value in the PRA (e.g., initiating g                                        event frequency or system failure probability) should be changed to reflect a new insight.

l j c. If the quantitative change applies to all the major cutsets for the ii failure expression in the PRA denoting that sequence, multiply the

     ,                                            PRA sequence frequency by a proportionality factor consistent with
    ,                                             the proportional change when one compares the old quantitative                                ,

, ', value with the new insight value. Apply this method to all changes lt li l of this type. I i i d. If the quantitative change applies to only certain cutsets of the j entire failure expression in the PRA denoting that sequence, apply 1 4 j the proportional change to just those cutsets and adjust the sequence frequency accordingly. t { I I

' 3 i

l j =- en e - - .. e *e,. .. g , . . . , . -

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

                 -      _         s                                         ,                   ,
           .                                                                      (

t e. Then, if an insight excludes certain failure modes in the PRA or additional failure modes should be added, delete or add cutset expressions by hand representing those failure modes. Then subtract out the cutset probabilities for those that are deleted {

  '                                      and estimate the additional probability for any cutset expression added to the PRA using generic data for that insight. Ad,just the sequence frequency accordingly.

j f. The result is the new rebaselined sequence frequency for the sequence being examined. Perform the above approach for all the dominant sequences as defined in the PRA. Finally review non-dominant sequences, to the extent possible, from the PRAs to make I sure these sequences do not become dominant after applying the new insights, or, if they become dominant, add them to the list. i The second approach is more comprehensive by incorporating event tree / fault tree modeling and using computer codes such as SETS. The data will consist of best estimate values and upper and lower bounds that represent the range i. of values to be used for basic events in the fault tree models. Selection of the best estimate value and the upper and lower bounds are based on an examin-ation of various data sources. The data sources include PRAs, NUREG/CR Data Sumaries LERs NPRDS and others. The best estimate values selected will be representative of the " typical" values used in these data sources. The cor-i i responding upper and lower bounds will bound the range of values used for the same event by the different data sources. The range of values defined by these upper and lower bounds implicitly includes statistical uncertainty in the best estimate value as well as variability in the best estimate value due to such i l l

      . ;- . ~ . . . . . . ; ;. .

4

                                                                                                  ,u. , 7 ,3,   ; g-       .

e j. ,, j i

                           . :.. :_ n x _ : 2 _. 2 1 _ _ ._ .

t i j< i.( i factors as subtle design differences applicable to the event of interest. l 1 The quantification includes propagating the best estimates and the upper and I lower bounds for the basic events by using computerized Monte Carlo simulation i techniques. By examination of the sequence cut sets and their values, sensi-i tivity analyses will be performed on the basic event importance, basic event [ uncertainty, and design differences. The results of the quantification and sensitivi y analyses will be used to derive the insights. j 4. Status of Understanding (9/30/83)

                 . The rebaselining approach has been completed and the more comprehensive approach will be completed in early 1984. The rebaselining effort showed, in general, i

i the sequence frequencies of the examined plants are slightly lower than the i original PRA estimates. For the PWR plants, the small LOCA with loss of coolant

injection sequences have slightly higher frequencies due to the examination of ,

operational experience that showed the smali break frequency is about 10-2 instead of 10-3 . For both the PWR and BWR plants, the total loss of AC power i type of sequences have slightly higher frequencies based on the insights gained i from the long tem blackout analysis perfomed under TAP A-44. A better under-1 standing of the more comprehensive approach can not be made at this time. l 5. NRC Position l: A fim NRC position can not be made at this time because supporting research

    .                   is still in progress, but the NRC feels that the generic quantification of i                   sequence likelihood can not replace plant specific quantification of sequence

) likelihood. Therefore, the generic approach can not be used to measure plant t performance against a specific criterion or safety goal. However, the NRC q r ~7 ;".y v- p. w - . . -

                  .                                                                                       l 1

l

.(

feels the insights gained from the generic quantification of sequence likeli-

       .               hood are valuable information to the NRC. The application of the generic

[ insights to a specific plant does not necessarily mean that the plant requires modification or has a significant safety problem. The generic insights only i flag an issue for further study on a plant specific basis.

     }
     }

4 i t e es t 1 l 4 i I l l l

                                 ~ y ,1,8*r';;si:- a ,
                                                       .'s,. .   * : ': * -
                                                                                ',~  *               ' 
                                 ,  ,, .      ,          ls-                  *

                                                                                                          '-               ^ - *
 .-.          ..                           . . -                . - ..-.- - - - .: L-                                   .
   *             ~'

g

  ', (                                                                     ISSUE 2.1.4 EQUIPMENT PERFORMANCE AND SUCCESS CRITERIA

1. Implications of the Issue to Regulatory Questions The light water reactor (LWR) systems engineered, constructed

  ;                                           and sited in this country followed a certain design philosophy for ensuring a very low level of risk to the public. Briefly the plant systems are engineered to withstand, with a single failure, the effects of (1) the design basis accidents and

incidents and (2) the effects of specified external events,: e.g., earthquakes, fires, floods, etc. In addition, a strong

! containment system is provided to contain any fission product activity produced even in the beyond-the-design-basis accidents. i The containment structures are designed and engineered to

  -                                            withstand pressures much beyond those imposed by the energy affluence during the design basis large loss-of-coolant accident (LOCA). Mitigation systems are provided in the containment buildings, e.g., (t) the suppression pool in boiling water reactors (BWRs), the ice condenser in pres-surized water reactors (PWRs),'and spray systems (BWRs and f                                            PWRs) to absorb the initial thermal loading and condense steam, and (11) fan coolers (in PWRs) and suppression pool l                                            coolers (in BWRs) for long-term heat removal from the l                                            containment buildings. The objective of these containment

,j safety systems is to keep the pressure low and protect the j integrity of the containment even in the beyond-the-design-I basis accidents. d fe&lV9d

,!                                                                                       C/4/
    ,                         10/04/83                                           1                             ISSUE 2.1.4 i N . ----            . .        ... ..                  .. .       ., , - . . . . .,       . , -  ..;--,-.     .  .  .           .
 ..         .       . . . . -                         ==            .  =w                                    .
                                                                                       . . . = . _ . - - - _ - - - - - . . -

                                                                                       -   ~--         --

The TMI-2 accident progressed from a transient with a small LOCA and with loss of emergency core cooling, to a degraded core accident in which the core was recovered to a safe coolable stable state, within the reactor vessel. It is significant, however, that the (defense-in-depth) safety design and regulation philosophy pursued by the reactor vendors and the Nuclear Regulatory Commission (NRC), respectively, was effective, since the radiological burden on I the public was insigificant. The most prominent aspect of the TMI-2 accident

;                       was that the containment integrity was not violated (except for temporary loss 4
!                       of isolation) in spite of the thermal and pressure loads imposed by the i

combustion of the hydrogen produced. The containment spray and heat removal systems functioned throughout the accident and the primary system heat removal through a steam generator (rest 0Ved relatively early in the accident) functioned and the heat generated in the core was removed in spite of a debris bed formed in the core region.

2. Subissues r What level of ESF operation is required to prevent severe core damage and to arrest core damage prior to complete core meltdown?

i What are the limits of core coolability and the implications of THI-2 regarding ! core coolability? How much credit can be assumed for the performance of nonsafety grade equipment? l How much credit can be assumed for innovative operator actions in controlling i core damage? i

3. Status of Understandina As a result of the TMI-2 accident, several research, development and evaluation programs were initiated to understand the course of such accidents, which have been j called either as degraded core or severe accidents. The ii

10/04/83 2 ISSUE 2.1.4 t

                                                                                                           ,p . . .

i. .

4

               - ,,                  main objectives of these programs are (1) to reevaluate the                                          ,
        ' {'                         risk to the public of the postulated degraded core (core                                             '

e aaltdown) accidents in LWRs, inhght of the knowledge I gained since the time that the Reactor Safety Study was j completed, (11) to assure that any threats to public health l and safety have not been neglected, (iii) to ensure that i any deficiencies found in the LWR systems are corrected and that the safety of the plants is improved, and (iv)'to l, develop procedures and training programs for preventing any j future occurrence of a degraded core accident and to prevent

  ,                                  the progression of an incident into a degraded core accident.

l' The research and development programs sponsored by the U.S. NRC, the Electric Power Research Institute (EPRI), the

U.S. Department of Energy (DOE), the German government and j others; the evaluation and the analysis / development efforts j performed with the industry degraded core rulemaking (IDCOR) program; and the probabilistic risk assessments (PRAs)

, completed for a number of plants are all part of the hectic activity pursued by the reactor safety community after the TMI-2 accident. i While a great deal of emphasis has already been placed on programs to improve the operator's capability to keep the power plant within its operating limits, j; there has been little basis for answering the basic question of what information s; the operator requires and how can such information best be used when faced with !' the multiple failures characteristic of severe accidents. The Severe Accident l Sequence Analysis (SASA) program utilizes state-of-the-art methods of analysis 1; to investigate in detail the progression of events in risk-dominan't accident i sequences, and thereby develop guides to evaluate possible operator procedures. Thus, the SASA program analyzed the TMLB' sequence (of which THI-2) is a representative case) and recommended a set of guidelines for the Zion nuclear j power plant. i i . l'

;                       10/04/83                                                  3                                 ISSUE 2.1.4 1                                                                                                                                       ,

i. :

(

     . .. . . . . . ,                 . ...                    ,a
      --                               - . _ . . - - , . - - - . _ _                 __- - __ :_ u _-_._.__ - . 2 s ..-               -

_:- . : - -- -. =. _: - . s . The minimum critical procedures for recovery include: f-o 15% auxiliary feedwater flow, or 70% emergency core cooling system (ECCS) flow.

!                      Thus, appropriate instructions at TMI-2 would have been to maintain at least 70% ECCS flow (assuming similar thermohydraulic analysis results).

I The accident management strategies include:

   >                            o       Early manual ECCS initiation.

o " Feed and bleed" recovery with ECCS and PORVs. o Primary depressurization using atmospheric relief valves (ARVs) recovers in 33 minutes. In the loss of decay heat removal sequence, SASA work revealed for Browns Ferry F that the nonsafety grade equipment, the condensate storage tank (CST), has sufficient water to last for 40 hours at post-scram decay heat levels and that numerous injection systems can take suction on the CST. ] In the study of TMLB' for Zion, one of the chief conclusions reached by the SASA program was that, if there was no containment cooling (e.g., sprays, fan f coolers), the containment would fail within 6 hours, but if there was contain-ment cooling then it may be a few days before containment would fail. l l Snould accident prevention measures fail, SASA has indicated levels of response corresponding to Commission rules. For example, in the case of station blackout at Browns Ferry, an alert is issued when the offsite grid fails and the diesels fail to start and load. Failure of the auxiliary feedwater system would be i noted in less than 15 minutes by virtue of indication of no flow and low or l' zero discharge pressure from the turbine-driven pumps. In this case, a site i emergency level is required. l l 1 - 10/04/83 . 4 ISSUE 2.15 1 1

                             ~ . . . . , , . . .      . . . . . . .      .    ....    .

             ,_   ..__..---3-_...-:----.                                                                  -

Core uncovery would occur in 1 to 2 hours after the batteries fail (about 6 hours) and would be indicated by high temperatures in the core. Thus, a (. ' general emergency would have been reached 7 to 8 hours after the blackout started, if power had not yet been restored. If cooling is restored in the following 2 hours (by restoration of power), the core can be recovered and cooled, although probably with severe damage, and the emergency procedures deescalated. Otherwise, meltdown will proceed over the next 5 to 8 hours with containment failure predicted to occur about 15 hours after th'e loss of all 'i power. 2 i

4. Approach to Resolution i

Accident management strategies and analytical studies like SASA are the key to

resolution for this issue.

5. NRC Position The position is to develop a consensus in March 1984 after reviewing IDCOR 7

results. e i L O D i i t l 10/04/83 5 ISSUE 2.1.4

                    --------g--,--,w--.,.-,.-..w-,-.--,.---,---.m.,m.,,-

.-...-.2.-. . . = - . - z...._.

=.

(. Issue 2.1.4 TITLE: EQuiPMENTPERFORMANCEANDSUCCESSCRITERIA ,

1. Issue The issue for the containment safety systems is to keep the pressure low and protect the integrity of the containment even in beyond-the-design-basis accidents.

2. Implications of the Issue to Regulatory Questions The continued operation of certain engineered safety features such as fan coolers and spray systems is crucial to the long-term capability of containment to withstand loads from a severe accident.

I

3. Subissues What level of ESF operation is required to prevent severe core damage and to arrest core damage prior to complete core meltdown?

What are the limits of core coolability and the implications of TMI-2

    -                             regarding core coolability?

How much credit can be assumed for the performance of nonsafety grade

    ,                             equipment?

i How much credit can be assumed for innovative operator actions in l ^. . ij controlling core damage? , FotA 939 Cl$2 11/15/83 1 ISSUE 2.1.4

            < - .        ,                      s            - -               -

y , , .e_.

   .___1._       . _ _ -               -_._                                                  _ ,

_ ._ _ _ _ __n .n . _ _ ._ 1

+,.

l 4

4. Status of Understanding ,

          .                       While a great deal of emphasis has already been placed on programs to          ;
 .                                improve the operator's capability to keep the power plant within its           1 operating limits, there has been little basis for answering the basic          :

question of what information the operator requires and how can such inform- i

 .                                ation best- be used when faced with the multiple failures characteristic of severe accidents. The Severe Accident Sequence Analysis (SASA) program utilizes state-of-the-art methods of analysis to investigate in detail the progression of events in risk-dominant accident sequences, and thereby develop guides to evaluate possible operator procedures. Thus, the SASA program analyzed the TMLB' sequence (of which TMI-2) is a representative case) and recommended a set of guidelines for the Zion nuclear power plant.

The minimum critical procedures for recovery include: o 15% auxiliary feedwater flow, or 70% emergency core cooling system

 .                                      (ECCS) flow.
  .f Thus, appropriate instructions at TMI-2 would have been to maintain at least 70% ECCS flow (assuming similar thermohydraulic analysis results).

l . The accident management strategies includ_e: o Early manual ECCS initiation. o " Feed and bleed" recovery with ECCS and PORVs. l l! o Primary depressurization using atmospheric relief valves (ARVs) i recovers in 33 minutes. In the loss of decay heat removal sequence, SASA work revealed for Br, owns Ferry that the nonsafety grade equipment, the condensate storage tank (CST), has sufficient water to last for 40 hours at post-scram decay heat levels and that numerous injection systems can take suction on the CST. l

  !                   11/15/83                                           2                         ISSUE 2.1.4 6

a_a2., w. A - ,.m,4 .--

                         ._ a

m -_. .- _

       *e .

In the study of TMLB' for Zion, one of the chief conclusions reached by the SASA program was that, if there was no containment cooling (e.g., sprays, fan

(~ coolers), the containment would fail within 6 hours, but if there was contain-t ment cooling then it may be a few days before containment would fail. 4 Should accident prevention measures fail, SASA h'as indicated levels of response corresponding to Commission rules. For example, in the case of station blackout

       !             at Browns Ferry, an alert is issued when the offsite grid fails and the diesels
       .             fail to start and load. Failure of the auxiliary feedwater system would be l              noted in less than 15 minutes by virtue of indication of no flow and low or i               zero discharge pressure from the turbine-driven pumps. In this case, a site j                   emergency level is required.

I f . Core uncovery would occur in 1 to 2 hours after the batteries fail (about I 6 hours) and would be indicated by high temperatures in the core. Thus, a

  , 3
      ;              general emergency would have been reached 7 to 8 hours after the blackout
    ;;               started, if power had not yet been restored. If cooling is restored in the
 ,   ;               following 2 hours (by restoration of power), the core can be recovered and
;,(                  cooled, although probably with severe damage, and the emergency procedures deescalated. Otherwise, meltdown will proceed over the next 5 to 8 hours with
      .              containment failure predicted to occur about 15 hours after the loss of all

!I power. t i

    ;I               5. Approach to Resolution i                       A plan for developing coordinated data on the "cruciality" of key equipment

!j is needed and then an approach developed to resolve the issues of equipment

I survival. The SASA studies determine the types of equipment (some I

nonsafety grade) that can be used by the operator. Where no adequate '{ substitutes exist, the equipment involved is obviously crucial and its reliability in a severe accident requires a high degree of assurance. l{ ,l , l

6. NRC Position l

l i None at this time. i

i - 11/15/83 3 ISSUE 2.1.4 I i

                        - 7'                                                                    '~        *           ' ~
                                                                                     ~

l W{C-~,qA*-- - - - . - . ' - - - . , .

ls , . '.. i k 1-

!(

i I, Issue 2.1.5 -- Influence of Operator Action on Accident Seauence 4

1. Implication of the Issue to Regulatory Questions i The human factor, and specifically the role of the operating crew

{ dur,ing severe accident sequences, has important and direct implications regarding Regulatory Question 2 -- how safe are existing plants. Regardless of the measure used to characterize plant safety, implicit in it is a contribution l attributable to the perfomance of the operating crew. A misunderstanding i of this contribution and its associated uncertainty can lead directly to inaccurate assessments of plant safety.

  '                                             It follows that there are also implications for the remaining regula-
 ;!                         tory questions. The perceived need for additional protection against severe accidents will be influenced by the perceived level of plant safety. Alter-natives for improving safety by addressing the human side of the human-machine k                       interface are sure to be identified. A reasonably accurate assessment of the operators' impacts on risk is important if one is to evaluate trade-offs ,

among operations and design modifications defensibly. 1

2. Subissues Can the operators adequately diagnose severe accidents?

,' Are the operators adequately equipped with training, procedures, instrumentation and control to restore the plant's stability during severe accidents? ' ' - r I * '- -s

                                                                                                   -l l ' v ' d O* '

l* How much credit should be given for corrective or innovative action? f Can an " excellent" operator compensate for a " poor" design? and I} vice versa? j Arethereconflictingstrategicgoalsduringsevereaccidents(e.g., ! -j preserve the core vs. preserve containment) and how should they be addressed? k

  )i                                                                                                                             F#A-f 4 -913 J                                                                                                                                c/4.s i

1 1

       . e.y .e ,.. . -          .. .y       , . .p . .., .; . . , , ,       ,
              -   4- - -,-r       ,    - .__          _ ,_ ... _ _ ,..____ _ _ _ , . , _ _ _

_-~_ __ _ __. _ _ . . _ _ . . p.. .. 1 !- 2 J

.(                            3. Approach to Resolution The principal approach to resolution has been to analyze the operators' role in severe accidents. The Severe Accident Sequence Analysis (SASA) program has been the major vehicle for these studies. Similar efforts

.f j are performed in the ORNL programs on Pressurized Themal Shock and on severe accident precursors. he studies performed to date select what are believed to be risk i dominating sequences for specific plants and then calculate the physical response of the plant for each sequence using an appropriate thennal hydraulics code (e.g.,RELAP. TRAC. MARCH). Each sequence is characterized by an " accident signature", that time-dependent set of observable parameters which represents - the infonnation available to the operators during the accident. Using their knowledge of plant systems, the analysts identify time windows during the sequence in which specific actions taken by the operators could restore the ,\ plant to a safe stable state. They also review the extent to which procedures, , , *, - I instrumentation and control are available to take such actions. There is no ) i reported effort to quantify the probability of those success paths. ,, At the time of the mid-84 decision, SASA analyses will have been

 -                  underway for almost five years. It is appropriate to review the results of those and similar analyses prior to the mid-84 decision to garner generic insights related to the technical issues. Beyond mid-84. SASA analyses will incorporate
 ;                  the best available analytical methods and focus on those combinations of plant, sequence, operator role and proposed modifications which can provide direct

- input to specific regulatory decisions.

4. Status of Understandino

? The operators' role in reactor safety is reasonably well understood in qualitative terms but is difficult to quantify accurately and precisely. For that reason uncertainty in operator performance is a (perhaps "the")

  -                  major contributor to uncertainty in risk analysis. Significant strides have t

I i

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

            --        - - ~ ~                                                        -    ~         -    -- -

n- . _ _ _ 1 -- : - - i.. ' ,. i 7 3 been made recently to model human performance during stressful (i.e., accident) j conditions in nuclear plants. These models are seldom quantitative, however, and those that are quantitative are just now beginning to be validated. A model in favor now is Rasmussen's hierarchical model of skill-rule-l knowledge-based behavior. Its applicability to operators' actions during , ,l severe accidents is in the increasingly sophisticated level of response reauired of the operators once they are cast into a situation for which they have little relevant experience, training or procedural guidance. Past human reliability 3 analysishasfocusedmoreonskill-basedbehavior(readinggauges, activating ll controls)andrule-basedbehavior(followingprocedures)thanonknowlege-basedbehavior(diagnosingtheevent,selectingaprocedure, innovating). l- It is the latter where greater attention must be focused if we are to ade- !, quately understand the operators' role in severe accidents. Relevant informa-

      .                       tion is being gathered from task analysis, simulator experiments, and controlled j1                             laboratory experiments which are currently outside the scope of the Severe
      .                       Accident Research Program.

.ip

    .i                                     5. NRC position i

In the current spectrum of Design Basis Accidents, no credit is f' given for operator action during the first twenty or thirty minutes after the initiating event.* After that time credit is assigned judgmentally based on justification supplied by the licensee. For postulated severe accidents beyond the design basis, the appropriate regulatory position is less clear. ,< Operating crews are generally recognized to be dedicated, well-trained and competent to handle a broad range of upset conditions. Post-TMI l J. requirements related to staffing, education and training are presumed to

I have yielded unquantified improvements in the operating crews' capabilities. s ji The Safety Parameter Display Systems now in most plants are viewed as a better ii way to alert the operator and monitor the plant's performance during abnormal

   )                          events. These need to be integrated with emergency procedures and more ex-perience reported on their value. In spite of these measures, it would be
    }

difficult to accurately assess performance during severe accidents in the j . absence of more and better evidence.

'An AN5 Standard is currently under revision to reexamine this position.

         .,..,;.........z...;.,,...

4 5' . , ( It is quite likely that if the NRC required greater protection i against severe accidents in existing plants, the industry would argue strongly 1

      ;                  for downplaying hardware fixes in favor of improvements in operator response.

j It is unclear at this time whether either approach to increased protection can be supported by objective, quantitative infonnation. NRC's position should i j not reflect a preexisting bias toward either approach. Rather it should be j shaped by its review of the products of SASA and other relevant projects and I, by its review of operating experience since the implementation of post-TMI j human factors requirements. 1 3 t e t

    .       \

O 2 W i l

  'i
 't 1

( d

             **o *sepe-      e- . ,e a ,   e .
                                                ;,          e e g.,,   ._y.+,   ,y..--.s_ y.a, , m ._ ..y, _.,.,    , , x. ,. -

hL W [

  !                                                             Issue 2.2.1 I

I

TITLE: Response of reference plants to selected severe accidents

1. Implication of the Issue to Regulatory Questions, j -

   .                       For many years, reactors have been designed, evaluated and licensed by the philosophy known as the design basis accident (DBA). As a result of using this
  !                        design and licensing method, the containment that is required for every major
 /,                        U.S. reactor is really there as a defense-in depth. However, the containment is designed to accommodate only limited hydrogen production in accidents such
 }                         as loss-of-coolant accidents. In combination with these LOCA or steam-line break thermal-hydraulic loads, the regulatory requirements has long included a specification that fission product load characteristics of core melt accident should be postulated to be present in containments. By severe accidents is meant those that go beyond the DBA, those that lead to severe core damage and I                  even to core bulk melting. The containment then becomes the front line of defense, and it will be exposed to very severe challenges. Hydrogen is generated in these accidents to a very great extent by reactions of the zirconium cladding with water, and by the reaction of molten core materials with the concrete in the reactor cavity. A spectrum of fission product load
  .                        will then occur, each characteristic of a different severe accident sequence.

Those accider.ts beyond DBA when analyzed for the risk to the health and safety of the public are shown to be the dominant risk. They are the source of what residual risk there is in a large nuclear reactor.

, 2. Subissues 1

What are the magnitude and characteristic of the source term available for i release when (or if) the containment fails? How credible are early containment failure modes for reference plants? 'What is the loading of the containment? What is the response of the containment and other essential equipment? * ( 10/04/83 h/A-n-W 1 ISSUE 2.2.1

C)

i

       . ~ ,        . .

[ ,f, 3. Status of Understandine !l ( The philosophy of the severe accident research program by the NRC was that i j data alone and analysis alone are not sufficient. It was determined that

   ,                                  new data and new analysis were needed in order to study the problem

)i . . thoroughly and adequately to make the regulatory decisions. Analytical,< model and code development and verification work is the heart of the 1e severe accident research program. . A major activity in the severe accident research program includes the fi ' reassessment of accident source terms, what actually gets out into the

biosphere in the event of a severe accident and consequence containment i

failure. In the source term reassessment, which is underway right now,

five specific plant designs are being evaluated for the full range of accident source terms. The plant designs include two large dry PWRs,

;. .                                  Surry and Zion; the ice condenser PWR, Sequoyah; two boiling water reactors, I'

Mark I containment, the Peach Bottom plant, and a Mark III containment, ' j the Grand Gulf plant. These plants were selected to span the range of ,

   .k                                 designs that are operating now or are under construction in the United

! States. The computer models that are used to predict both the heatup of

the core, its melting, and tne fission product transport and release include the MARCH 2 code, which has been developed for thermal-hydraulic ll
i analysis of this process; the CORSOR code for release from the fuel; VANESSA code for vaporization release during concrete attack after the

!, core melt has fallen into the reactor cavity; the TRAPMELT code for fission

,j                                    produ:t transport within the reactor coolant system; and the NAUA code Il                                     developed by KFK in Germany for fission product transport in the contain-ll                                     ment. Many of the source term results have been completed and are in the jj                                     peer review process right now. One of the calculational volumes on Surry

,j has already tieen released for broad distribution. ' l! All of 'the elements of severe accident reseach come together to develop a

j. support base of methods and data for a much improved assessment of plant safety. By the use of greatly refined probabilistic risk analysis,,it 1,s intended to evaluate the current level of reactor plant safety. The best i 10/04/83 2 !$$UE 2.2.2 1

1 7,..... .. .. . . . . .. , . . g.. ... ....;,r. .,.....,...y...- 7 y .. .. ,, .. .

l estimates of accident sequence probability and the consequences of severe

        ![                      accidents will be considered, as well as to evaluate the ways or means to l*!                              reduce risk in cost-effective ways, selecting the most judicious combina-

$j tion of features that might significantly reduce reactor risk. Given these

i two bodies of information, the statement of reactor risk and the method

' f. and costs for risk reduction, it will be evaluated and determined what , 3 changes, if any, are necessary in reactor regulation. A similar program j to this is going under the auspicies of IDCOR. ! 4. Approach to Resolution

j Three Mile Island did demonstrate that the Class 9 accidents can occur.

, j The NRC and IDCOR are both looking very intently at this. Many discussions between these two groups and with others has taken place and have identi-1 fled the key technical issues.that must be faced in order to appraise the - i risk of severe accidents. We must face these source term issues in which

fission products are released. Do we have accurate models of them? What

] is their distribution fo11owi.1g release from the fuel within the reactor [

pressure vessel--in the primary system--and release into the containment? How do they behave once they get out of the containment? In fact, how do ' t they behave once they get out in the environment? i z The coolability of degraded cores, the ability to turn accidents around as was done in Three Mile Island must be looked into. Containment failure is a major thing. The contsinnent is a defense-in-depth in the design i basis accident, but it becomes the first line of defense in the Class 9 j, , accident, in the severe core damager or core melt accident. It is very i

difficult for the reactor coolant system or the fuel barriers to hold all l of the fission products of the core. The containment is the major line of i defense.

1, .

      }                  The analysis of accident progression requires models of relevant plant charac-

,{ teristics and phys'ical phenomena. Inputs for these models must be specified 'i based, on e<1 sting knowledge. Uncertainties about the inputs, the models and i4 . l r 10/04/83 3 ISSUE 2.2.1 1

  ,......,.......g
                                                             ..,.--...-.--.-..-3.7..-.-.~,.-..---.           ,.         ..        ,    .-
                      .-__..____..a           .

___ w _ _...__.._ __ -

                                                             . - - . . - = ^            ..

, ; . .. ' a ; l I l P the phenomena lead to uncertainties in the predicted results. These uncertain- l l(, ties and their potential magnitudes must be recogni' zed if meaningful rpnclusions ! are to be drawn.

   }                                                                                                                         .

The appropriate operator actions and their timing to prevent or delay contain-

   ;                   ment failure or otherwise mitigate the release of radionuclides must he looked I

into. l l A containment loads working group has been formed by the NRC in an attempt to .-

 ;                     develop some consensus on containment loads resulting from specific events j                     (e.g., the magnitude of the PWR steam spike).                                          *
  .                                                                                                                                     I

5. NRC Pos-tfon The IDCOR reports and their evaluation will be conducted by NRC. N8tC's

 .                           source term assessment will be evaluated and recommendedation to the Nuclear Regulatory Commission based on IDCOR and NRC's results for what,-

ever action is appropriate for severe accident regulation is expected in

k. September 1984. ,

e e j s i

 ;                                                                                                                                    ~

4 i '

1

 !                                                                                                     ~,)

i

o J 10/04/83 4 ; ISSUE 2.2.1 ,j

      . , ,... g . s.. , '.
                                                           ~

Q: , ,

   .i       .

M l( l Issue 2.2.1 1, . . j TITLE: RESPONSE OF REFERENCE PLANTS TO SELECTED SEVERE ACCIDENTS

1. Issue A Containment Loads Study Group has been formed to study the containment loading that might result during PWR and BWR severe accidents. The i j

issue is, "How does the containment respond to severe accident loads and what are the environmental source terms for accident sequences in the l reference plants?" The potential for early containment failure and the associated potential for large radionuclide releases, due to " steam spikes"

has been under reevaluation by a group of phenomenological experts both for i PWRs and BWRs. From the standpoint of the Containment Loads Study Group, whether or not the containment pressure reaches the failure point is not

( at isssue. The question is as to the timing.

2. Implication of the Issue to Regulatory Questions There are many challenges to containment integrity generated during the t

progression of the degraded core accidents. Since TMI, the NRC and nuclear industry have given considerable atterition to the possibility of containment failure by overpressurization and overtemperature from the burning of hydrogen, the collection of noncondensible gases or the loss of heat removal capability studies for prevention of degraded core and for mitigation of degraded core consequences (containment failure and subse-quent radioactivity release to the public) are ongoing. It is important [. to know how the containment will fail and the location and extent of its failure. The currently active research development and evaluati,on pro- _ grams, e.g., on " source ters," containment loads, hydrogen etc., are all 7 oriented towards minimizing public risk. Equivalent research efforts, are q

( [0/b8WS 11/16/83 b 1 ISSUE 2.2.1 1

1

                                                        . , z .. ., a , t              ,

            . . _ . _ .      _ _ . . . . . . . _ _ . . . ~      . . _ . . . . . . _ _ . .               ..         =. ......    . . . _ . _ ._ .

i.- 4 underway, whose goals are the development of the information and experience

    ,     (                         bases for recovery from severe accidents as part of the SASA program.

Theobjectivewouldbetodevelopproceduresforoperatoractionswhich would achieve a safe and stable state (as soon as possible) along the progression path of an accident and limit the damage. The equipment survivability questions arise when one considers recovery actions during postulated severe accidents. Some of the equipment could be subject to adverse environmental (e.g., during hydrogen burn) or operational (e.g., pump cavitation) conditions. Some programs to test the critical equipment are already in place; (e.g., during hydrogen burns); othersneed to be implemented. Considering the dominant postulated accident sequences for the PWRs and BWRs R&D should demonstrat that: (1) the PWR containments do not fail early and catastrophically; (2) sufficient retention of fission products is achieved in the PWR primary system and the ECCS piping for the V sequence; and (3) scrubbing of fission products in the BWR suppression I pools is highly effective.

3. Subissues What are the magnitude and characteristic of the source term available for

, release when (or if) the containment fails? How credible are early containment failure modes for reference plants? What is the loading of the containment? What is the response of the containment and other essential equipment? What core damage would result for selected severe accidents? Under what conditions would a core melt result?

e 4. Status of Understandino l

A major activity in the severe accident research program includes the J reassessment of accident source terms, what actually gets out into the biosphere in the event of a severe accident and consequence containment failure. In the source term reassessment, which is underway right now,

'(

11/16/83 2 ISSUE 2.2.1 I

                                                                                     , , .:* p,p' q -
                                                                                                    .           .-           .:'~

s- , five specific plant designs are being evaluated for the full range of ( accident source terms. The plant designs include two large dry PWRs, Surry and Zion; the ice condenser PWR, Sequoyah; two boiling water reactors, Mark I containment, the Peach Bottom plant, and a Mark III containment, the Grand Gulf plant. These plants were selected to span the range of designs that are operating now or are under construction in the United States. Many of the source term results have been completed and are . in the peer review process right now. One of the calculational volumes on Surry has already been released for broad distribution. The results of the Source Term Reassessment Study, NUREG-0956, which will be completed MJuly 1984, provide the basis for near term resolution of this issue. Sandia's Severe Accident Uncertainty Analysis (SAUNA) working group wil shotily(W83 dishbute a draft report entitled " Identification of Severe Accident Uncertainty." SASA also provided input to the document. A containment load study group has been formed to study during a PWR severe accident in which a molten core dropped from the reactor vessel into a water filled reactor cavity below. The corium-water interaction would generate large ( amounts of steam and hydrogen and the large dry PWR containment might fail on the pressure surge. The BWR containment loading is totally different from those in a PWR and the two are being addressed separately. There is no large pool of water underneath the reactor - vessel in a BWR and although there is a pressure increase as the RV fails and blows down, there is no " steam spike" that might threaten containment immediately after RV failure. Nevertheless, the BWR containments have other failure modes that are of concern. The SASA work both at Oak Ridge and Sandia are providing important contributions to the Containment Load Study. At the request of the Senior Severe Accident Research Plan (SARP) Review Group, a working group was established to develop containment leakage models for use in severe accident source term estimates. These leakage models will serve to quantify leakage areas as a function of containment pressure, and possibly temperature loading, for various containment , types. The leakage models will be incorporated into existing containment t 11/16/83 3 ISSUE 2.2.1 4 4

        ' - ~ . <     .     .
                                                                   . rgg ;.y                   .

                    ,                     computer codes to permit a more realistic assessment of containment

( behavior for severe ac:idents; specifically, the consideration of containment leakage as a function of time and the impact of containment pressure relief (due to leakage) on the mode and timing of containment failure. As part of SNL, SASA program, structural analyses of the Watts Bar, Maine Yankee, and Bellefonte containment structures were performed with the objective of obtaining realistic estimates of their ultimate static pressure capabilities. The investigation included analyses of the containment sh, Ell, equipment hatch, anchorage systems, and personnel lock. In the longer term after the decision there is an important role to be played by the mechanistic codes, by SASA, and by MELCOR in the response of reference plants to selected severe accidents. The MARCH, CORRAL, and CRAC codes have been applied in a variety of LWR severe accident studies. It has become clear via experience that further improvements in severe accident modeling codes are essential. The NRC is pursuing the ( development of the MELCOR integrated risk analysis code to address these problems.

    ,                                   One of the intended applications of MELCOR is the addressing of recent
   ',                                    severe accident source term controversies. The code is also intended to
  .                                      provide a structure which can be readily modified as new data becomes available. The results of this effort will" be used for developing new regulatory and licensing perspectives which could then evolve into new l    ,

regulations. The assessment of severe accident modeling needs for BWRs, was undertaken by the SASA program at ORNL. All of the elements of severe accident research come together to develop i a support base of methods and data for a much improved assessment of plant safety. The best estimates of accident sequence probability and the consequences of severe accidents will be considered, as well as to evaluate the ways or means to reduce risk in cost-effective ways,

     .                           .       selecting the most judicious combination of features that might
,,                                        significantly reduce reactor risk.

l 11/16/83 4 ISSUE-2.2.1 l-

          , . . . - . . .     . . .              .       . .                                                            ..,..y,..  ..

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

p.-

5. Approach to Resolution l ; f, 4 '

Three Mile Island did demonstrate that the Class 9 accidents can occur. The NRC and IDCOR are both looking very intently at this. Many discussions between these two groups and with others have taken place and have identi-fied the key technical issues that must be faced in order to appraise the risk of severe accidents. We must face these source term issues in which i i fission products are released. Do we have accurate models of them? What is their distribution following release from the fuel within the reactor pressure vessel-in the primary system--and release into the containment?

 -                           How do they behave once they get out of the containment? In fact, how do they behave once they get out in the environment?

The coolability of degraded cores, the ability to turn accidents around as was done in Three Mile Island must he looked into. Containment failure is a major thing. A severe accident management task in PWR containments is presently under study by the SASA program. The objective is to provide a comprehensive analysis of strategies available to manage the i containment and its associated systems to mitigate the riease of fission products to the public in PWR core meltdown accidents. i ; Under this task, the following types of information are determined for

representative plants: (1) time-dependent pressures, temperatures, and radionuclide concentrations in containwent during accidents which are ; dominant contributions to risk, (2) realistic thresholds for containment failure due to static pressurization, (3) minimum systems performance necessary to prevent containment failure, and i i 11/16/83 . 5 ISSUE 2.2.1 I

      ..-. . .,. ,, .    . ,    , .       . . ~     . . . . . . . . .. ,                                   .. s .-. . , ; .

                                      ~
                   ,,                      - . . _212_.1_i                                  _.
                   ~

y.. i

)!

(4) appropriate operator actions and their timing to prevent or delay [ containment failure or otherwise mitigate the release of radionuclides. Many of the BWR severe accident sequences, notably loss of injection

      ,                                 (TQUV) involve loss of core cooling and failure of the reactor vessel at                                   before primary containment failure. These sequences always involve a
   }                                    boiloff of the water in the reactor vessel and after the core is uncovered,
   ?;                                   it heats up and melts. The available models for the subsequent events are
>j                                      totally inadequate for BWR analysis. For example, the MARCH code assumes that the entire core is moved into the lower plenum, boils away the water
   ]

i there, and begins attacking the reactor vessel bottom head. This might be reasonable for a PWR but is totally unrealistic for a BWR core which consists of hundreds of four-assembly, one-control-rod modules, each individually supported from the reactor vessel bottom head. SASA program subcontracted work is underway at RPI to correct these serious modeling deficiencies. , In these BWR severe accident sequences in which injection is lost, the core

    -P                                  is uncovered, melts, falls into the lower plenum, penetrates the reactor vessel bottom head, and fall,s onto the concrete floor of the drywell. The pressure suppression pool temperature rises only a few degrees in the process and pressure in the containment is not threatening. Results so far suggest the BWR MKI containment failure by overtemperature.

The analysis of accident progression requires models of relevant plant characteristics and physical phenomena. Inputs for these models must be specificed based on existing knowledge. Uncertainties about the inputs, the models, and the phenomena lead to uncertainties in the predicted results. These uncertainties and their potential magnitudes

   ,                                    must be recognized if meaningful conclusions are to be drawn.

6. NRC Position The IDCOR reports and their evaluation will be conducted by NRC. NRC's source ters assessment will be evaluated and recommendations to the 11/16/83 6 ISSUE 2.2.1

         . . . , .                  ~..           .:..                               .,. ..    .         .                     ~

.=

Nuclear Regulatory Commission based on IDCOR and NRC's resultsy for what-( ever action is appropriate for severe accident regulation is expected in September 1984. l . m O o 4 I e 11/16/83 7 ISSUE 2.2.1

                     ~
                        - ~ ' = ; 4. .. ,                ...                  .    . . .

                                 . - . a      a.--..= .:... - -. _ .. .           ..    .

i, ' ,

     .y                                                                                          Verston lor'13 j.

1 Issue 2.2.2

![.

Qualitative Assessment of Severe Accident Likelihood (Insights)

I Implication of the Issue to Regulatory Questions 1. 5 Qualitative assessment of severe accident likelihood can provide insights for the NRC decision making. These insights can be gained from examining past PRAs and safety studies, operational experience and performing the PRAs. i The position of the issue can effect the answers to the following regulatory

questions.

3. How safe are the existing plants with respect to severe accidents?

3.1 What should be considered in this measurement of safety? 3.2 How do the terms of measurement compare, including uncertainties? h 3.4 Using these measurements, how safet are the existing plants?

4. How' can the level of protection for severe accidents' be increased?

j 4.2 How effective are they? l .$ 5. What additional research or infonnation is needed? '1 j 5.1 What are the infonnation gaps bearing on the severe accident i decision?

5.2 Are data necessary for implementation of the decision? l] l .; i i 5.3 Are data necessary for confirmation of the decision? 5.4 Are there specific issues that require more data before they I ii are decided?

        ;(                                                                      Fo t A in c.lak 1
                                                                         .   ..      .s   .    . .             .

t,' o

                                                                                   -2
    !(-

l

6. Is additional protection for severe accidents needed or desirable?

f 6.2 What is the likelihood that perfonnance could be improved by the alternatives available? I 6.3 Are the likely improvements worth the cost? i 2. Subissues i What are the characteristics of plant design that contribute significantly to risk? -

   '                    ~
                                 -         How do the PRA insights compare to a review of operating experience (e.g., dominant accident sequences and failure I                               modes)?

What are the accident sequence areas that have the highest uncertainties (e.g., success criteria, human errors)? How important is it to reduce these uncertainties? l J,

  !                         3. Approach to Resolution t                        The Accident Sequence Evaluation Program (ASEP) is the primary NRC research a

to support the issue. ASPE is funded and managed by the Division of Risk j Analysis of the NRC and the primary contractors are SNL and INEL. .

t-Many insights that possibly affect the accident sequence like11 hoods have 1 l evolved in the field of PRA. ASEP has formulated a list of insights based l on the review of existing PRAs as well as operational experience, and special studies, including Station Blackout Study, kccident Precursor Study, industry l4 i' j, ATWS work, and human factors work. ,I li,! ,i..,...-....-..

                                                                                                           ~
                     ~    ~

u__._ ;__. . ,_ i _2. 2 m ;. w.w ,,, w _ ___,[.....,,..___._..,.....T 4 ' i ! (,. l l Additional insights will evolve from the quantification and sensitivity analyses I j tasks of the ASEP research. These insights will assist the fomulation of the final plant classes. i

  ;                    The insights gained from ASEP will be compared with those from operating

, j experience. A qualitative comparison of the dominant accident sequences i j and failure modes of ASEP and the Accident Precursor Study will be made. l Areas of agreement and disagreement will be highlighted.

  )                    4. Status of Understanding (9/30/83)
  !                    Insights gained from ASEo are tabulated in an ASEP interim report entitled
  !                    Accident Sequence Likelihood Reassessment, August 1983. The insights are jp                   grouped by in,itiating events, plant systems and overall sequence considerations.

i These insights represent a broad spectrum of findings that can effect the i1 ,i perspectives regarding accident likelihoods. For this reason, these insights have been used to perform a limited rebaselining effort to determine how the { accident sequence frequencies in some existing PRAs would change by applying ) these insights. Ths. rebaselining process and the results are described in if the interim report. In addition, %<. tabulation of insights can provide a

 )                     useful single source of accident likelihood insights for use by other programs and studies.

1-

  !                    The list of insights is not necessarily complete since an exhaustive literature survey has not been performed. New insights are being realized at this time.

However, the interim report provides a valuable start at collecting important i infomation which could change the NRC perspectives regarding the frequencies of accident sequences. o

        *$  Y                  ,        _ _               __     __                              __    _ _

            . _ . --              _ . _.._.. _ .... _ . _. . .: . ... __ __. .~.. _
                                                           .                            . . . . ._...._.i..~._____,__.

1 '. . ~ . I I l? j f. t -- ,g ( l Insights gained from the ASEP quantification and uncertainty analyses task l f and the comparison with the operating experience will be formulated in early 1 i 1984. i The results of the comparisions will be documented and where significant 4 . differences exist. ASEP will attempt to resolve those differences.

lI

)y

5. The NRC Position The insfghts can provide additional information to aid the NRC in rulemaking

    -                        and the development of standards.             Insights from PRA can identify features I

l of the plant, previously unrecognized, that has a measurable impact on the

   !                         severe accident likelihood or on the risk to the public associated with the j[                          facility. The NRC can use these insights to identify gaps in the present
   -                         regulatory concept of defense--in-depth or in the detailed application of.

that concept. It is possible to examine these gaps and develop deterministic I criteria to update or eliminate these undesirable features. 'i 4 t 3 f

 'I 1

J

'}

1 -l k I k 1

                . g ~ 7:;; ; q1:q 7; 7:,=-== y m .              --

                ., = .          .:             ^ -:.                       _   __

f Versicn 103583 - I. I Issue 2.2.3 l( Integrated Probabilistic Risk Assessment for Reference Plants 1 i j 1. Implementation of the Issue to Regulatory Questions I j The integrated probabilistic risk assessments for reference plants (plant classes) is a generic research approach consistent with the overall NRC severe accident rulemaking. It is a feasible and practical alternative to the plant specific approach given the time and funding constraints of the severe accident research. The approach will draw important infomation from I

 ,                    the existing PRAs and safety studies to make inferences of potential risk of f

all existing LWRs. The position of the issue can affect the answers to:

2. How should the Comission decide the severe accident question?

I 2.1 Should the decisions be generic or plant specific?

!(

1 -

3. How safe are the existing plants with respect to severe accidents?

3.1 What should be considered in this measurement of safety? 4 i 3.2 How do the terms of measurement compare, including i e, uncertainties? t i 3.3 Which accidents are to be considered and which ones can be i j ruled out? i a 3.4 Using these measurements, how safe are the existing plants? .y { 4. How can the level of protection for severe accidents be increased? i 4.1 What types of improvements are available?

!I 4.2 How effective are they?

Foi A- %-%f i 1 C-/'? .

                             *?
                                   '. f .I5.

i I. f' ' ,^bfd . *

*~

     .       o f0
       .                                                                                                        1 i

*(

I l 5. What additional research or information is needed? l 1 5.1 What are the information gaps bearing on the severe accident decision?

 '                                       5.2 Are data necessary for implementation of the decision?
}                                        5.3 Are data necessary for confirmation of the decision?

5.4 Are there specific issues that require more data before they are decidedi

6. Is additional protection for severe accidents needed or desirable?

6.1 What conclusions do we reach when we assess available

-('

i information in light of severe accident decision criteria (from question 1)? (' 6.2 What is the likelihood that performance could be improved by the alternatives available?

 .                         2.       Subissues
 !                                  -    What are the plant classes?

-1 4 1

                                    -    Which accident sequences dominate the likelihood for each            .

1 4 plant class? i i 1 - What is the core melt frequency and associated uncertainties of these accidents for each plant class? js k - What are the dominant contributors to these accidents by plant { class? E", *-"'*. M* '* O 'N 9%***9' P

O * # #4
S "
r F *9 * * ' %

3 g [ ,

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

O

(.

                                     -       What are the generic insights derived from the " integrated PRA for reference plants"?
                                     -       Does the core melt frequency and associated uncertainties for
  .                                          each plant class have any statistical meaning?

3. Approach to Resolution The Accident Sequence Evaluation Program (ASEP) is the primary NRC research f

 !                           to support the issue. ASEP is funded and managed by the Division of Risk
 ;                           Analysis of the NRC and the primary contractors are SNL and INEL.

The technical approach is partly described under both Issue 2.2.1, Plant i Categorization, and Issue 2.1.3, Quantification of Sequence Likelihood. k Figure *1 is a flow diagram of the ASEP approach to address this issue. 4 Status of Understanding (9/30/83) The research is in progress and the re,,sults will be available by March 1984.

5. NRC Position The NRC does not have any research results to support a decision. Therefore, a NRC position will be made by mid-1984.

t i I

,(.

1

       * = = - - -
                   ~+4 ; e -  ._g      .,,,              .-.      -.;y   y....- p ..g., .; t, ,- . .. ,, ,- ,.,g;

n.. = ... - . .... .... =. u u u _. .

                                                                                                -       -                                                                                           ~
                  '         9                                                                                                                                                            .-

Figure 1 ,. ,,o. R EP zuTEntM Acc DENT S E G VwCE. E vnL VA TIC /V rgcrosacAL /t?PtoncH fl.D4%/ O/Ak A8M AsE P i J. O 4 ' L C o**m entr Assm.7 i l I ZwTW8mi Ast i I ' l l T I I I I T g,g l $4sas su Thtf Averzwar e# CowrameWWerr* gTwergsj lgigy ,p Ca.a* *= e pgpq g , A.,0 pt AJ. Aaswoveg pe, g' /ea r4 s l cowfp.we.g.sv-

                   .    . wr eas c.ce)                                                l A.ap 7 deva Camr** amter TYes3                                                   I TT ' * * #*U l                                                                                         C W 8 m#C** * **bCJ V

i _s,. ,,. paws i z ovvein muu e6a.7vs vuoss mac.4 ,,,,, i c we.. n Tx et , p g. 7, - j..c- . l lco-e. - twe ______ p... _. . _. .

1 w - v,- nso rs.c a.. , ,e is isnm eaas .,, . .,,

g., ..g, l sa

                                                                                              * - . .          .m,,         . n,e... o.           . is e ra I , ,, , , ,,,,

[ < l ..,. c.... .-e ... 4. we

                                                                                                                     *. e wv i.y
                                                                                                                                     .        c.u a .

we. , con sea wee i l r tv ca asev eu.wss.(s P~a aun co-a) P.S. EOL41 DO Mi nish eM ' Ar$4l C gh#t" SEGo.Ae&E , bifl' OP

                                                                                               .z os.re. ,f,o a 6 v.is. ecaur                  x.n-s umsono                      M ** "5          **o j                          ,_ ,, m                                                                                                                 ' '"3
                                      !                                                         To mmegaTE THE Actics wr. T Mts                         z.,.i,uo e o      !
                                                               ****#'s.s awav                                                                                                             ===ous v.

pgi.vte.mfE 'The I l T*.E aug poe, 54& TEM s

P.d FOAMS.O Fet =t=AE l I **'~ **'

eo *.Mav.ucs.5 s em'" es '"*" l _ _ _ m _- l i U spawsweas maww pi.nwes as poss,gi,g .ro S MS T* *' s e 6 aa" D CSTAeaJ ACtaOS.8T Pter16#rted FL4 0 Aae E*s.4CT OC.4 6 O **" CF AL'

                                             * " E "**3                                        sysarv.M O A *w.=4               Fetc A n.L Facor 6.s.as A o                l I bua ANtp sis (Psay gggggmg ugtDED TD MitetPTE WC T*ds     l Ps.a sts
  ,                                                                                      l zorm.#r F.EO SE4wtwCE&. ASO.Fr #10 7e *F
    .                                                                                           ev.wr, ~.- r ,.                       .

awo < =r..c,c ,.

u s-,.el r.zo ,* W Fu P'*8 f S see ra a P M Lowe DEAvvenssis 4Asco ou PAsf Paa eestearf m

    ,                                        ;*P.r,,                                      . . c      ..... .is o -.T.*               r    </.           r e. ..;           [ 5xs=oca~~5 =
    ,                                                                                     i' z w :..                         . - ,. or . ; o 2 - - r - r                    ,
                                                                                                'rawus c.um ; % 4 me.evaur 3                                                                                            Soercat wsram.s .
,   ;                                                                                      i i .

l [

    !                                                   /                                                                                                                    I i                                                                                                       r ed' +

I-i

            ~***--8..,-.
                      ,                                             =;                   .              *           .,_-g ,t,        ;f +g          ,,--,.p,..+*,     *'

qv a ^; *.cf '

                                                                                                                                                                                     .,y,-

           . - - -                   -       --                            ..~     .
                       .O A.sa 9
            .                  LC6iC                                                                     Co**sasessr-                                            l          ors :s,sv
     ;.                                                                     l roe. cated aniv.M (,wtia , ar s), com patte p eep e,revy rwt temrue as Ptaasr seseiFc sytttas cousstema sisvir"savr car

Laud:r sa./ Etsvaaka*N , c.,stsity . O s. swf **e.T b-4 YT CF

                                            /                                    SY~'fLM G4.rEAh.tawcits.TWd meins acce pesato ay                                     pup l :ows me.its.ee, esst F ,wc r.ew cf weg sistem A a.o                                I l.
  ,                                                                              wwar Acteows                 musT         TMt     e6act    To  psapwen      pes      l ExsicM j h tou >b-e qp                           .       Fvut. Tie     **. USED    T&at   *rificA    t pas?  Pa  a cea powsaer

p.a,.o Ts Fai6 vat Pa casis viss To Aehow.erere swsta.m i 84"*" l Fasssea PeisestiLivy ysso,7voe"Eter TO 06 rtari.ws 4 snave s l TF Twt FAsLwat l' Acessiu1nas Fot EACH AC1*ew TW87" g *,sEOs 7b SC #tareserto att ss&meirica.ert-( l

?>

II DeFatsasr. T.5 eso answirica*ir cierv.atas.es veist

Fo.wo, iset sis ftss m4EE ca,s.esAan s.a.a.a t p svo

( 6Eoutso nb ous 64 sTw.es eetee( . 3F ssuur.ca.sr i OlFFE8 castes eaRE Fouwo, TREA1as7w&M As orppgggafr SYSTEme . grecvfs i Den 9t E (,2. svaveMto Fwil PLA*/PS l .s I

                                                                                                                                     \s I                              ' 1,A P*'        42 weis                              l 1'                                                                                                      3I 88*,(en ges         &NPr$ 4 dew $4
                                                                                                  -____-- ~ --                                                                          - -

r t o

o ucTi-* 545T=>S a * ** L 5 *** otvebott" Faubt Tsat T M C01Eb1* F' t.

                         'ow*"#                                                      Fe A.      rwa sy1 fgx                si towP s Tte Gvaarrip.r s aee.v rvsa                                                                                                                                   880'
                         ,eecats                                                  "Ti4 E    Do*=ra,*W AccioEAff" wouchess. rwe areests SNETEM
                                                                              , USE 'TW C FeldLT 'rtta" l'iCDu6At kc 4 s t.                         cawcapT i                                                                                                                                                                     4Louf i                                                                                       *
                                                                                  'fMAT ts Co s ea 5TeWF twiTw rWC Ta P

A 8PS' fA*4&a" Age.P ><.66 VSE 7%#1 TA P A 4 7 r9004.L Te "lW C certarr roesstLC 1 If rea ancs acc.os.sr wcus*>cc, we syste.vs 5.ist or n..,vi m i

t sw A=0 SwifeET s*tsTipos AssetA MA To e't14ed*T E i . n swis e oeisme eiersac $3 3 rge.43 Fan asca enema.a TwE # C888EAsr Asts 3r ss ar'rv,a0 8 kub g x 40**$LES s

f$***T l

ws rsms l Ta4 4.s#

A ssen.# j t'f # # 3 8 lFC86 l: ravv -

res , nees, use, us, t,,,.. mu ,.

                                                                                                                                                                              ...a.,,

s 5 3. O - np,s,s u ncio s.rr i l sit avemeCC 1 i

l 2 .

-(

1 s 1 l l

     .                                                                         l                                                                                    l l
                                                                                                                                                                                . 244
                                                                            .I-
         ----=,-.---.                       -                 .
                                                                                                                                    .          3,,,      .,...,
                                                                                                                                                                        ,.,p.., .p.

       ~
                                                                                                                                                                                           )9sEP

1. 0 G ! C- CeestsE a.rT" dE svLf l

m. . ...,7 ..._,# m ,,,,.

l. ~dtopPS L.t'.T OF i

2. i ARE Cees es. "Tt Fata ca g ewn n% r

                                       .                                    l .F n,.v.,                 .x se c.               , i<,.,,,,,,.~. .rw.a

wm a' e . . ,,~ , o, - .. e .-. ,< > n , r . r ar ru . .or

                                                                             ' ca.m. ass pc4 Tus ret 1new.at eccao.ur saewswet                                                       CL*S553 y

p',ar waert

    .                                a.n eva                                               .aser P t.E             rot hC paae da.saseg                                                                                                                                    .

PLAnse,t Pi.e.,1 & F6 *** *4-NPI, I gpg g gpggy I powe. i,3,3 ci. e +

    .                                                                       I 3W1                          8* 3 c .. ,                 :

y pis e, sw i (seek P6amt e,3. m. fta ,7 Ca*** A M pis s ,s., s (,,y ea asur 3

f.*** ca aes 5 )

l . . .. . : c, . r ., . ~ ., - . ) y u - ,- . -., u F. 8. u,T ., ( . . . . . j r..,. . _ ,-... ... ..... s,sgs. gg.g pg,uv ,,g,s> g,g g,,,,,,. ,,,, , ;,,,,,,,,,,, av aa r e ,,r w are**r***T . Fews.t ,aars noosi.s Fca l R PCs4 l Pwa.rr t i ** S g _ y l~ -a ~~ . ..e 4.**a ES

                                                                                                                       ,, .or . 1, Sow   so)   Foul  g. sac.
                                                                                                                                                        , ~,

sw.Wrs saJ l

                                                                                                                                                                                    .3.,

pp

                                                                               ' A**o nefr&A *wo                                                                                    pg74
                                      .n,     g.g.c g                          I'D.E          F*nAF TEEE ,.ests.t. "D.s s s ecmew eP' ,5tt.C                                        g g ,, ,g i                                                                                    Van.ug s Age SA&Eo OW A.J een AM.a.def.aw OF we8..W3

'e

                                                  .= .                              f,gyA g.wsCE.g EUMM448tso tw *D4e STR7tw

{ $6httedt .tt.80DJt" stand.aJt4 And46T$ss AstotT. TW." l paTa sumaneieso iv rwtr dev=.T iucue para l F8*** PA Aa , A,98 8Er oaTM sweenna.c3 , a.g ros e wo , _ p,iscsi.e.a.n e w. AEpstys. Twg Ste? .sTimmets Ad.8' g Tw Tveica. .auves vaso ew ,wns on,n s.veess . I

                                        , , . , _                                        . .           . .             .,                   - ~                                     am
                                       ., , ,,, =                                   ..-,,.v-.-,...m,..,                                                            .

l o, . w,

                                         . .., .                                   ,,s . . .....                 . . - .. .. .,. , . ~ . m                                          ,,,

a PoA 'D4C stews s.g twse.o.cy g 90 '/. ce=riotsweg satsw

l Pet KDCh. SET op f"At.Wr TREE' Fe0tn 5,6 rMtes sACsf S.7-

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

( I su . ~,, . c ., , ,,, . , .,. . w. . . . w .,

                                                                                     . .us.., ... e..e - ... nse
                                                                                                                                               ... r. 4 < v. a .

3 iT. o.& . esr, w. e s. ; S + 4-t . - rgt.e . .?'

e . .

w.. .. .. . . . _ . , . . a .. s l.

                            /.o4'C                                                                        cc*wewr AstP l                                                                                            ,        a asa t-
                                                                                                              /(,1' 1 N          C.T0J Coutaa.mtur PvPr                A ST. P
                                                                                                      ,                             s                                        3 a# fbi:64 1s.v.s w ow TevV                                      TM b. S E G ut" * '* *3              H es.g pa c MY l                                                                 /

po p P=M '

                                                                                                            <               'N N r e.n useet iE S 6#Lf3 1 svio n.

F m.s atT La. A ej g i. A y a,, g y g,g,, fas es s e

                                                                                             / f\                                'l '.

p ,s. SEavsw.a eTTE4N I,, ",, Yames cassiessoa s em.mo

T 3.,,T,, ,. n., u,uu ua Ce aes= o a43 o l cad c C.*P*

                                ,..,,.m.

ou s -iwatmo ew wc uca m mverseru **" le'aarr C. LDSs E 5 l y we sessr.ws sse esa.am ou a o<resas rr as ,, , ,a wesusw.= v.s.r.sw.s,n.c - - tr w e ! peun. P 6 A nst gg aggieg peggT g ygg, Theg aggan,g gggg gg eygag. (' h ttEs es.a.er c.a as *wn vi.e 7=saa naves rs w.it scow. c= Cbatt ey itPenn#TTV4*% y C hat./f" Ct.Aasta A&E { P L.h a#f c, LASSES REDE P*assO. Wf. Ca uetSED PLA a.T ca Aests , 3e* T1Li ne si sm 6e.y towfpew Spet. apes. PL*mn's werW'W l

                                                                                %.E st eengterT% , 84 vn4Ttd. e

T*st S Pe tiri s. ,

P6Aurs' Ram syrTsaw CHas ArmnefTeca, wovos ALL. l 5eawew<.e s wiv eia, A courws a- wer Aee S Osa/Tt F *t D MS tes,rTheaooars, s onn.a s CendMW s en-f6.marrs Mrs TPacst seCuences,7 sere ss rut ,

                                                                                @ denr M eArd G E F**E M Puma #6 M CdJasta j .

p.re vpdra Asassearta p?TAff e.. _ _ i _ Ct1 -- C"1" *J costa =ererr WE

                                                                                                    ,s          s,               : ss'.',.

a U i . ugp f Den v6d*T s ewai. Fema A L te svi.vs ,, ,,, g PLamer cLArs

                                                                                                     /f l*

l0 ' '

l Hrs.inteser a

                                                                                                          ,,,                                  s==           c us, t
 -(

wo scu ./ n< iye ,' Z , seca.c.e av, cots

                             ...              y
                                                   ~
                                                                 -- L.
                                                                         ~
                                                                                  ~         ~
                                                                                               ~,-  *I
                                                                                                                  ~                 ~~                 ~
                                             . , ,,,7.:.~

7- .s 7

                                                        , . +,      .

__ _ ,, . . .-. Q .

C. . ._ . . . _ . . . . _ . . . . . . . -

Version 100683 f Issue 2.2.7 Effects of Uncertaintites on Estimates of Severe Accident Likelihood

1. Implication of the Issue to Regulatory Questions In the regulatory framework, making decisions based on results with large uncertainties is very difficult. Therefore, in the case of estimating the severe accident likelihoods, the understanding of the uncertainties on the estimates is highly important in terms of identifying the attributes and ways to reduce their values. The position of the issue can effect the answers to the following regulatory questions.

3. How safe are the existing plants with respect to severe accidents?

3.1 What should be considered in this measuremer.t of safety? I 3.2 How do the terms of measurement compare, including uncertainties? 3.4 Using these measurements, how safe are the existing plants?

4. How can the level of protection for severe accidents be increased?

4.1 What types of improvements are available? ,' 4.2 How effective are they? 4.3 What are their costs and side effects?

-                      5. What additional research or information is needed?
!                             5.1 What are the infonnation gaps bearing on the severe accident decision?

5.2 Are data necessary for impelmentation of the decision? r (

For A -t+-M C,/ +8 7,.; _ .

b

    ;                                                                               5.3 Are data necessary for confirmation of the decision?

5.4 Are there specific issues that require more data before they are decided?

6. Is additional protection for severe accidents needed or desirable?

6.2 What is the likelihood that performance could be improved by the alternatives available? 6.3 Are the likely improvements worth the cost?

2. Subissues

                      -    Can all dominant accident sequences be identified?
                      -    How do the uncertaintites on accident progression affect the estimation of the sequence likelihood?
                      -    How do the uncertainties on the human action affect the estimation of the sequence likelihood?

How do the uncertainties on the data affect the estimatior. of the I sequence likelihood?

- How do different statistical approaches used for uncertainty analysis affect the estimation of the sequence likelihood?

How can uncertainties be accounted for in regulatory actions? a

e I s [

3. Approach to Resolution A statement addressing the accident likelihoods is not complete without addressing the uncertainties that can potentially affect the accident sequence frequencies. ASEP identifies four types of uncertainties. First, data i.

l' uncertainties, such as the uncertainties in initiating event frequencies and . component failure probabilities, always contribute to the uncertainty in any estimate of accident likelihood. The second category is completeness un-certainties. This issue addresses whether all important accident sequences

can be indentified. The third category is accident progression uncertainties l

that can affect our perspectives regarding accident sequence success criteria and mitigation potential. The fourth category is human error-related uncertainties. These uncertainties occur since operators can play a very important role in mitigating certain accident sequences and yet our ability to model and quani.ify the human performance is not as good as the treatment g of system hardware failures.

4. Status of Understanding (9/30/83)

!' ASEP has addressed the uncertainties related to accident progression and human factors in an interim report entitled, " Accident Sequence Likelihood Reassess- ! ment," August 1983. Data uncertainties will be addressed in early 1984. The question of completeness in regard to identifying all important accident , sequences will be addressed in the long tem research to be completed in 1985. As of now, ASEP is addressing only the "most likely" dominant sequences identified by the existing PRAs. The accident progression and human-related uncertaintites are addressed first because ASEP believes they might significantly I k affect accident sequence frequencies based on the results of past PRAs and - j '"*g 9

M= 4'e e g , g ,

                                                                                                * =U 9'      .
  - - ---            ~.--.,_z.   - , _ _ . -                               _.

( , m,. . ~_..-_.m__.__..s

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

lb j 4 considerable lead time is required for other NRC programs to investigate these uncertainties so that any resolutions can be factored in to the long term ASEP , research. In the interim report both the PWR and BWR accident progression uncertainties have been grouped into three major categories: i

   -                            1.          Core melt processes

2. Containment failure processes

3. Other accident processes The list of the PWR and BWR accident progression uncertainties are provided 'n the report. These lists of uncertainties have been prioritized, where possible.

    -                     based on the preceived potential effect of each uncertainty on sequence frequencies and system success criteria. Discussions have been held with eth' staff of the NRC SASA program to inforin them of ASEP's perspective on important accident
    -                     progression uncertainties worthy of study in the SASA program. Agreement has been reached that these uncertainties represent important questions that could

{ affect sequence frequencies on modeling and so should be studied by SASA as ',, time and money allow. In fact, work has already begun on many of the uncer-tainties to either resolve the uncertainty or otherwise increase our under-standing in these areas. 1 Important human error-related uncertainties affecting PWR .and BWR accident l

sequences are also listed in the report. In each case, the uncertainty is '

listed and other clarifying coments are provided. These uncertainties are ll ! ;t ( also listed in a prioritized forin with the highest priority given to those I

         -...,.,.,r..._.,        . . . . . . .         .
                                                                                                                                                                ,n
                                                                                                                                . _ _ _ _ _ . _ _ __ i - ir

_ _ _ _- _ r ._ _ ._ _ 1 _ __. _ ___ _ _ _ _

I. . t *

                ,.                                                                                                                 5-
      .(

[ actions important to actual accident conditions and for which the least amount of data exists. Discussions have been held regarding these uncertaintites

     ,                            with staff involved with human factorsprograms. The ability to respond to
     !                            these ASEP identified needs depends on the results of a human factors program planning effort currently underway.

5. NRC Position i

    !                             A firm NRC position can not be made at this time because the research is still in progress.
  .1
             =

6 h o 6 e 1 O e ( il

                                                                                     .a_.                                      _.
                                                                                                                                                     , _ , 1, , , _ , . _ .   ..; _                1 ; , _ .   ..y          _

1, . . i .' a

     ;                                       2.3.1     Design and Operation Changes to Prevent Severe Accidents Including j       ('                                        Sabotage Protection i

1. Implication of the issue to Regulatory Questions #

This issue is directly associated with the fourth regulatory i question: how can the level of protection for severe accidents be

increased? As its name indicates, this issue relates to ways of protecting against severe accidents by the prevention of these i

accidents. This prevention can occur either by improving the use of existing equipment (e.g. by improved operational reliability,

pmcedures, testing and maintenance) or by providing additional

     ;                                             equipment (e.g. dedicated shutdown heat removal systems). Of interest here are both the identification of the various ways of preventing severe accidents and the measurement of the benefit
     ,                                             achieved by these ways, individually and in combinations.
    ,                                      2.      Subissues j                                              -

Measure of benefit (core melt frequency reduction, risk reduction, improvement in defense-in-depth) I i - Generic applicability of measured benefit Relative importance of prevention vs. mitigation

'(

Benefit already achte.ved by TMI-fixes

                                                   -       Impact of uncertainties poorly understood, unknown sequences (e.g. seismic, sabotage) component, common-cause failure data (e.g. human error rates)
                                                   -      Relationship to unresolved safet
 ;                                                        Decay Heat Removal Requirements)y issues (e.g. ySI A-45, Alternate If cost / benefit is about one, are changes mandated or not?

3. Approach to Resolution l$ In order to address this issue, three types of data are needed:

ll (1) prevention options of potential merit; (2) an assessment of the risk from the specific or generic plants (including dominant sequences,

  ,                                                system models, important systems, failure modes, consequences t '

of specific sequences); and (3) cost infonnation on plant modifi-cation options. With this data and measures of associated un-certainties, the core melt frequency or risk reduction reduction benefit can be calculated as well as cost-effectiveness value. In the Severe Accident Risk Reduction Program (SARRP) and with , input from supporting programs, such calculations are being made '

I in an iterative manner, consistent with the phases identified in the Severe Accident Research Plan (SARRD).

FotA-M-W

c. +'t me . . 3 p*O w = ep . g e g

_ - - , - - - - , _ _ e - - - - . - - - - - - - - ,----_.,.._-,-_,,,-,---.-_4-

                                                                                                                           ,-,---%..         _ _ - - _ - , -    .           - - , _ . - . . -      .   , e-  ,, . - , -        --

                                                                                                                 -                                  ~
    'R                    *
  )

(' Phase ! i Two sets of cost-benefit calculations are being performed in the

      !                               SARRP phase I work.       In the first set:

a relatively broad spectrum of prevention options was identified; six existing PRAs (Reactor Safety Study, RSSMAP) were used as the risk assessment base; and

                                      -     very rough cost measures were assessed.

With this supporting data, methods were developed for calculating and displaying risk reduction benefit and cost-benefit, as well as important sensitivities and uncertainties. l ' The second set of phase I calculations will be using: i a narrowed list of options and combinations (with less promising options screened out); an expanded base of risk studies for generic classes of plants. using data on accident sequence likelihoods synthesized from a broader set of PRAs, and consequence data reflecting up-( to-date understandings of the phenomena involved; and more detailed cost estimates, with support from an architect-engineering firm.

    .                                With this data the risk reduction and cost-benefit calculations
     '.                              described abov,e will be redone, on a schedule consistent with the mid-84 severe accident comission paper.

l Phase !! The phase II SARRP evaluation of core melt frequency or risk reduction benefit and cost-benefit of accident prevention options is oriented with the completion of a further evaluation of accident sequence

      '                              likelihoods (which includes consideration of external events) and with the completion of most of the severe accident experimental 1

3 programs. Using this new data on accident probabilities and j, consequences, the cost-benefit, sensitivity, and uncertainty calculations will be redone. ls - 1-l l (~ l. i L e e o *

                                                                                               -e, -.w., ,. . - -     y-,,,   , ,-, , , , - - -- - -.-.                    ,-

r.. . . _ . . . _ -_ _ . _ _ _ _ . _ I( l 4. Status of Understandina j At this time, the potential risk reduction benefit of prevention j options is relatively well understood for a set of specific plants,'

    ;                            based on the RSS and RSSMAP risk assessments. More robust and
   ,                             generic information is not available because of incomplete study i                             of the competing risks of option installation and use, the costs j                            of such options, and the more generic extension of these plant-specific results. Many of these issues will be accounted for in i                             the second set of phase I calculations, either by refined calcu-

lating or by sensitivity studies. Issu'es assessed by these sensi-tivity studies in phase I will be the focus of phase II work.

l 5. NRC Positten ' l From the calculations made to date, it appears that major modifi-cation to prevent severe accidents (e.g. dedicated shutdown heat

    -                            removal systems) will be difticult to justify on the basis of quantitative cost-benefit calculations. Combinations of lesser
   ?                             wndifications, however, can be expected to be marginally cost-effective. In addition, consideration of uncertainties and poorly-quantified sequences (e.g. , seismic events, sabotage),
   .                             or less quantitative (probabilistically) methods (e.g. ensuring defense-in-depth) could lead to reconsnendations for requiring modifications which are, at best,Cet effectivef1Erginally           f

( only quantitative best-estimate analysis was used. 4

6. IDCOR Position ,,

Little is known about the IDCOR analyses at this time. However, general statements have been by IDCOR management that their analyses indicates very low risk from existing LWRs. Because of this assessment and the corolla that no further risk reduction is needed little quantitative eval ion of measures to reduce risk has been made. i 4

  .t
     )
                                   /

4 i ' d 1

                                                                                                 ~     .. . .
            ,,,g..__.._._,_.-..__m.......                             ..               .

                ..       .           ... . . - . . . . . . . . . , . .           . . - - . . - . . . - . .    .. ... .   . - . ~ .. ~ . . ,. .. . . . . . . . -
 ;.}                 ~
    }

o

'l                                                                              ISSUE 2.3.2

[ l IMPROVEMENT IN SEVERE ACCIDENT MANAGEMENT CAPABILITY

1. Implications of the Issue to Regulatory Questions a The issue is how to improve accident management capability and how

  .(                               to assess the costs and benefits associated with such improvements. In
.I                                 this context accident management refers broadly to that ensemble of design, q$                                  operational and human factors modifications intended to improve the licensse's j                             ability to restore the plant to a safe stable state given the iritiation of a 3.4                                 severe accident. Accident management is distinguishable from modifications j                                  intended to prevent accidents, such as an additional train of emergency core
 ?j                                cooling, and from modifications intended to mitigate consequences, such as a lj                                  passive vented containment or distribution of potassium iodiide tablets. The g                                   resolution of this issue is of direct consequence to Regulatory Question 3 --
 .j how can the level of protection against severe accidents be increased -- since T                                improvements in accident management capability will presumably be evaluated
     *(                            against other means of improving safety.

4 ' ,) 2. Subissues What are the success paths to safe stable states? lj How much credit should be given for corrective or innovative } action? What degree of reliance should be placed on computers in terms of q i automatic control or providing guidance to the operator during ! upset conditions?

.)                                  Are there conflicting strategic goals during severe accidents (e.g.,

i preserve the core vs. preserve containment) and how should they be 4 addressed? q Can one confidently assess the value of improvements in accident b management capability? 3 How far beyond the plant boundary does the scope of accident J - management extend?

.y j          '

FeA - 84 '94

;                                                                                                            c.    's o a
    ;l f ~Q.W:7*'T.y    - ,       =?!;7 " '*                                      ~                                        '      ~
 ._.:~L:                      .. . : 2 _ s L . _ - -                  _ . _ .

I. . / ' ;,

2 "v - - . ,3 M

                                                                            ,       7

3. Approach to Resolution i The approach to resolving this issue is only recently beginning to -

I-l evolve. An accident management task force-has been appointed to .,l revise SARP's Accident Management Program Element. This revision-is

   ;                         to be submitted to the Senior Review Group for approval before "'

.)i October, 1983. ,

I ~ j In the meantime a project (B0826) entitled Human Factors Review for SevereAccidentSequenceAnalysiswilhbeinitiatedatORNL. To start, they will, investigate the factors affecting the control room operator's { performance during a BWR ATWS sequence, focusing on operational strate-

   )                         gies for reducing risk in situations involving core damage and beyond.

I Factors to be investigated include: o Human engineering of the control room ; o Written abnormal / emergency operating procedures , o Operator training s o Operator reliability , Note: Considerably more effort than thitask described above would

       .                                   be needed to derive useful insights on accident management prior
                                      ./ to the mid-84 decision. Examination of specific sequence's' of a specific plant are necessary but not sufficient. A broader
                                      ,    perspective is needed and can be gained by reviewing all of the

[ relevent products of the SASA program to date to draw more generic insights. Other SARP programs, such as A1220, Development and Analysis of Vented Filtered Containment should be reviewed for their relevence.

Includedintheconsiderationkmusthethepost-TMIhumanfactors i , requirements in order to judge their effects on the licensee's 4 ability to manage severe accidents. Some development or refinement of methodology, e.g., Operator Action Event Trees, may be required to provide a structured framework for evaluating the information

        !                                  and presenting results.

1

 ,          (
 .is ~
                                                                                                                           ~'

5 / 1 i .

                                                                                                                   ~
          ;,w" p ry, r+ q r. .ypx; ggg;ig::.,..gm: :~7 - -

r _

L . 3

}

l-

4. Status of Understanding j( The. phrase " accident management" emerged from the post-TMI

'j environment, where it was recognized that a more structured approach lj to coping with the unanticipated was needed. The concept originally i established a broad perspective involving not only physical events I transpiring at the plant but also the information flow and off-site l activities during the accident. Since then the scope has apparently narrowed to focus on the role of the operating crew in controlling events at the plant.

 ;                                                              Several developments since TMI have a direct bearing on accident
 .                                                              management and our understanding of it. These include:

o Emergency procedures o Emergency response facilities o Operator training

  ;                                                                    o    Emergency preparedness drills
   '(

t - o Severe accident research - Emergency procedures -- All existing plants are in the process of {

                .                                                converting their written emergency procedures to symptom-oriented
   ,                                                             from event or1ented. The old procedures required the operator to
'j'                                                              correctly diagnose the initiating event based on observables in the i                                                               control room and then select the appropriate procedure for bringing

the plant under control. The newer procedures are much more helpful

   ;                                                             in directing'the operators to the correct actions and place less reliance on their ability to diagnose the exact cause of the disturbance.

j , The current scope of the new procedures, however, is limited to events

 ].                                                              prior to core damage. There is as yet no validated guidance for coping with events beyond that.
   -l                                                            Emergency response facilities --- Substantial upgrades have been made j                                                               in equipment and facilities. Computerized safety parameter display
    ;                                                             systems are being installed in most existing plants to provide the

j. operators with a more compact and meaningful overview of the plant's 4

ik s 4i . a i

          ?'                                                             _     _?                                       -   r -

i. .

4 i-I status than is available from the control room panels. More advanced

!(                           computerized systems, such as EPRI's Disturbance Analysis and Surveillance Systems are still developmental and are years away from widespread implem-
!                            entation.

Technical support centers are being added to most existing plant to

!                            augment the engineering capabilities of the operating crew during an i                            accident. Emergency operations centers are also being set up to coordinate off-site activities and thus allow the operating crew to concentrate on controlling the plant.

Operator training -- Greater emphasis has been placed on training operators to recognize degraded core conditions and to utilize existing plant equipment to control them. The requirements for simulator training have been increased but here are doubts about the fidelity of the simulators to severe accident conditions. The position of shift technical advisor was treated as an interim effort to increase the engineering expertise in the control room. The STA is to maintain an overv1ew of plant status ano is consistent with the human factors perspective of the operator as the system's manager.

s4 i '! Emergency preparedness drills -- Requirements have been implemented to improve 3 emergency plans and exercise them regularly. Protection action guides give-the licensee's staff guidance as to the trigger points for escalating emergency response actions. i l Severe accident research -- The Severe Accident Sequence Analysis (SASA)

}                             program and other related programs are examining the operators' role in i                            severe accidents. Considerable insights have been derived from studying

] specific sequences at specific plants regarding success paths to prevent 1 core damage given the initiation of a risk-dominant sequence. Considerably less effort has gone into evaluating options available to the operators

i once core damage has begun and the effects of those options. Programs i investigating mitigation features such as vented filtered containment have

  !                           occasionally addressed the potential role of the' operator with respect to
  ,                            such features. There has not been however, an attempt to review the
  $f    '

numerous products of these programs from an integral accident management perspective to see what has been learned and what future direction to take.

q

    - q v v:,      ,. --        ., . y- ,             .: ... y.. . .- . ; y . q-; . : . y,?     ;             ._      - g. ,

s . 5 I We have undoubtedly generated much information and activity related to t accident management but it is not clear that we understand the concept l any better than we did before TMI. What appears to be useful is a i systematic effort to define the issues more clearly and review the recent

  ]                              base of regulation, experience and research information prior to the mid-
      !                          84 decision.
  .i o                          5. NRC Position j                              The current NRC objective of improving accident management capability q                              is too vague. It must be defined more clearly and addressed more
  , .l                           integrally. The current body of available information will be assessed
 ,:                              to gauge what the current status of accident managment capability is, l                            including what weaknesses my remain. A preliminary assessment will be f                     '

made using available analytical techniques of the values and impacts associated with correcting any major identified weaknesses. Although it is premature to identify changes in regulations that may be required to improve accident management capability, we recognize this as a likely area in which improvements may be warranted. The importance of accident management to reactor safety has been highlighted by the TMI-2 accident and subsequent research. Further, the costs of improved management are perceived to be small in comparison with changes in plant configurations. The above will be accomplished prior to the mid-84 cecision. 6

1 4 e i' l l-l l I

('

4 l l '.: l i * * *, - - . ~y ;- s- i ;j

                                               ,          ..;; . . - > , -  ,.-s-.     .

g, .,.

                                                                                                                    ,4, ,

              -                                                                        -                 ~         ~                   ~ - ~   -~
        ,         .._._J._.       '...

y .-

i. s g f fE3

.i( j
   -i Issue 2.3.4 1                                                 Changes in Emergency Response Capability 2                       1. Implication of the Issue to Regulatory Questions 1

j This issue involves the degree and extent of emergency planning and ]M response capability. It includes the size of emergency planning j zones and the types of protective actions planned for, as well as 1 the organizational aspects such as training and exercises. For 1 i

     ;                       existing plants, the position on this issue affects primarily the outcome of Question 3--How safe are existing plants with respect b                              to severe accidents? The position on this issue for plant modifi-cations or emergency planning alternatives affects the outcome of k                    Question 4--How can the level of protection for severe accidents be
      ,                        increased?

l Essentially, the issue is How do changes in emergency response affect

    ;                          risk? Since risk is determined, first, by the ability of the plant and operators to prevent the occurrences of severe accidents; second.

N by the ability to diagnose and manage accident sequences to prevent 1

-j                             further degradation, given the occurrence of mishap; third, by the

.i perfomance of key mitigation features, such as containment, to prevent c, large releases of radioactivity, and, finally, by the degree and effi-

 )l cacy of emergency planning and response, it is clear that emergency I
. l,
 ](

i Foi A -g4-iza l C,l5l

 .]i(     .;.-------....--..

z, a . . g. _ .. .: . ,.... ,., ,-.,.:., .g a._;. ,.y p ,

I l i ; (, l l

j planning comes into play only at the end of a long chain of failures. I Nevertheless, many members of the public and not a few decisionmakers a

l view emergency planning and response as vitally affecting the outcomes

  .}                             of Questions 3 and 4.

Key alternatives include the size of the planning zones, and the type i and timing of protective actions to be taken. Increasing the size of the EPZ from 10 to 20 miles, say, would imply that reactors are not very safe, since emergency planning is needed out to very large distances. Reducing the size of the EPZ, on the other hand, would imply that reactors are safer than originally believed. Similarly, requiring an innediate evacuation throughout the EPZ would strongly imply that reactors are not very safe, whereas a reconnendation that no evacuatio'n was necessary would carry the 1:nplicatinn that reactors were safer.

     ,                      2. Subissues o        Size of the planning zones o       Types of protective action to be required o        Timing of protective actions                        -

i~ o Public notification systems

      .                           o       Drills and exercises l
   -}                       3. Approach to Resolution (1) By March 31,1984 j                                     The staff is reviewing current emergency planning regulations based 5                                     upon a reappraisal of the risk using existing source terms. The
    !(

1 6 t

      ), e.-    + = .       -ge,      7..  ..        . . . ,      ,   ,,       ,  j          ..  . g.  ,,,       ..._

i i, ,

e

j. i(

i i potential for revisions to the regulations based on existing risk h estimtes will be identified by the end of calendar year 1983. (ii)AfterMarch 31. 1984

 -                                       A vigorous research effort is underway to reassess severe accident source terms for a number of reactor types. This involves con-tractor studies and analyses by BC), and ORNL together with expert peer reviews, to be followed by a broad-based peer review of the scientific basis of the reassessment. The staff's reassessment of the results of the contractors effort will be documented in NUREG-0956, to be issued "for consent" in June 1984, with final publication by December 1984. The staff's assessment will include the assessment of IDCOR results subject to the timeliness of the transmittal of IDCOR results. Any possible revisions of emergency planning criteria will be scheduled inmediately after revised source term estimates and preliminary results from the broad-based review are available in the summer of 1984.
  ^

4. Status of Understanding Offsite public health risk justifying emergency planning and response 1

   -                              is alnest entirely due to degraded core and core-melt events. Releases from core-melt events have been categorized into three types, referred j                               to as SST1, SST2 and SST3 with SST1 being the largest and least probable
  }                               release, and SST3 being the sallest and nest probable. The estimated i                             probabilities for these releases are 1x10-5, 2x10-5 and 1x10-4 per reactor-(                        year, respectively.
                                                 'yS      ',*                               "   ' ' * *
                                                                                                        '/* 3,

[ ' *! * *!.".]"% $ ? '

                    .,                                           * , ' 5*

1 ' * ! " < .

s, , 24 .'

                                                                                               '~ l
                                                                              ^ ^^~~~        ~
        ..          . _ .        . . L- -   . . .     - - . -

l..c i ! l l

  } (-                                    '

An ongoing reappraisal of reactor risk based on research reports such as I NUREG/CR-2239, using existing source tems, as well as additional experience. l indicate that there is a significant variation of risk within the 10 mile EPZ and that the type and timing of protective actions ought not to be f uniform throughout this region. These studies suggest that the risk is much higher close-in to the reactor, and that priority should be given to planning and protective actions within the first few miles. In addition, i risk studies also indicate that when a core-melt is in the early stages, reactor operators may be unable to predict th= failure or performance of key mitigation features such as containment, and hence may be unable to predict whether a given core-melt will produce a release such as SST1 or SST3.

5. NRC Position The NRC position on this issue for the mid-1984 decision, pending a reassessment of severe accident source tems, is that the plume exposure emergency planning zone should remain fixed at 10 miles, but that a gradation of response is appropriate. Priority should be given to ij residents within the first few miles of a reactor, and since the release

,              magnitude cannot be easily. predicted in the early stages of a core-melt sequence, the conservative position ought to be that actions should
;              be planned for that are consistent with a severe release. For this I              reason, a prompt evacuation should be recommended within the first few I

'? )( i

5-

!(
.                          miles whenever a " general emergency" is declared. This evacuation should be precautionary, based upon internal plant indicators that
;                           indicate the iminence or actual occurrence of fuel overheating, and should not await the actual releases of radiation from the plant.

Beyond the first few miles, the protective action planned for should include prompt sheltering with the imediate institution of a moni-toring program to detect areas of high activity. Upon detection, res! dents in these areas should be promptly relocated. 4 0 0

t 5

                 - - - -        -         e*

y ' s ; p,, ,..q .-,.. '

                                                                                          . . ~         <

                         .~ . _._.____ -_ _.___ _ _ _

i i , 3.D Decision Methodology Irsue Papers ( j

  >                   Definition of Terms                                                                   I I have tried to use the following definitions in preparing position       .           l papers in this issue area:
  .                   Decision Criteria - explicit standards, rationale, benchmarks, guide-lines or philosophy set out in advance which will establish a framework for Comission Decisions. The criteria may be qualitative or quantitative.
  ,                   These criteria will be set out in answering regulatory question 1.

,, Decision Process - the formal process used by the Commission to arrive at a severe accident decision for existing plants. This will most likely be notice and coment rulemaking. The process will be set out in answering regulatory question 2. Decision Aporoach - a deterministic-probabilistic mix of tests, observations, analyses and assessment used to answer regulatory k questions 3 and 4. Decision Techniques - formal decision analysis techniques to assist the decisionmaker in developing decision criteria (regulatory question 1) I1 , and what additional severe accident protection, if any, is needed or desirable (primary question and regulatory question 6). i i

i

~

n

       ~ ~:" ~ ; L                      L~            .: L iX.Rh%?;"?>b.    ._

D'~ _ _ _ _ _ _ .

        .                                            ----:=____.._.;._            - - - - - _ .. . .

jI . ~~ E . . i'* I

  '                                                          ISSUE 3.1.1

(' SELECTION OF DECISION TECHNIQUES

1. Statement of the issue The issue here is which of several available analytical techniques shall be used to support decisionmaking related to serve accidents and how these techniques should be used. There are a number of fonnal techniques for organizing and displaying infonnation to the decisionmaker. These techniques differ not only in tenns of what infonnation is displayed but to some extent in what portion of the 4

decision is dictated, or at least indicated by the techniques. Therefore, at issue is not only how infonnation is to be organized and displayed but to what extent we want to fonnalize the decision process itself and perhaps limit the discretion of the decisionmakers.

2. Subissues

,(

What is the proper role of fonnal decision analysis techniques in the decision process? What techniques are most appropriate given the decisions to be made the decision criteria to be used and the available infonnation. Do the decision analysis techniques facilitate infonned participation

 '                               in the decisionmaking process by interested parties?

t t

3. Implications of the Issue to Regulatory Questions

 !                               It is assumed that the regulatory questions are fundamental and are not dependent on the decisionmaking techniques employed.

l 4. Approach to Resolution A process for making the mid-84 severe accident decision for existing plants will be developed by the staff, discussed with the ACRS and i proposed to the Connission in the fall of 1983. This process, if

 !                                approved by the Comission, will be published in the Severe Accident 1(
!                                                                            Fot A M c,lsS
          *==-;.     -,.-.....f,-     ,       ..                       ,                        ,                         ,
                                                                                  ,.J.f__*5_ WJ' 6,d: __f ,_.______.__;__

        -                                                                                    . . - - - . c;nn 2                      _ ; -- __ ----:            _
                                                                                                                                                                      ^

7; - i:

                                                                                -                                                   2 4
     >                                                                         Policy Statement in early 1984. The process will include, as appropriate,

( analytic techniques designed to standardize and expedite the decisionmaking process. 1

5. Status of Understanding There are several analytical techniques available to assist in decisionmaking. ,

Some typical examples are multi-attribute utility weighted matrix, value/ , 1 impact and analytical hierarchy techniques. In a general sense, all techniques have a comon characteristic. That is, they provide a structured , framework for organizing the information relevant to a decision. Some ! '. decision analysis techniques are specifically designed to facilitate decisionmaking even when there are large uncertainties in the information base. The use of formal analytic techniques serves a valuable purpose in documenting the rationale supporting a particular course of action. The techniques can vary considerably in their specific characteristics, e.g., the degree of quantification, the specification of attributes and the effort and data required to perform a defensible analysis. For any b given issue one method may be preferable over another. Generally the analyst decides which method to use since the method used is usually considered Tess critical than the competence of the analyst and the quality of the available data. However, the Comission may wish to preselect one or more methods in order to standardize and facilitate public scrutiny of the decisionmaking process.

6. NRC Position Formal decision analysis techniques will be used to help organize, analyze and display information pertinent to the mid-84 decision. NRC will encourage the use of such techniques in-house and by the industry. The choice of techniques is considered less important than the care taken in selecting data used in the analysis and in documenting its results. However, care will be taken to select a technique which facilitates decisionmaking in the face of uncertainties, clearly display important information to the l3 decisionmaker and makes the reasoning underlying the analysis clear to 3

k

                  ..~ ..-                           . .                              . . . .                             ,                                                               ,,       , , ,          _
          . _ _ _ _ _ _ . _ _ . , . . _ _ . _ _ _ , , _ _ _ _ _ , . _ , , ____ . .                 , , . _ _ _ _ _ _ _ _   _ _ , _.         _ _ . _ _ , _ _ . ,__,______.,_,,n,-._    __      , _ , . _ _ , .*     , , _ , . _ _

     -          . .       . . . . = . . . ~ . _ . . - . . . _ . . - - - _ . . . . . - __ -. _ . - . . _ . . . . _ _ . ..

3 all interested parties. Parametric displays will be used wherever possible to provide the decisionmaker with a feel for the sensitivity of presented results to various input parameters. Regardless of what

   -                               technique or techniques are chosen the analysis once performed should not dictate the decision no matter how comprehen'                     s ive or defensible it may be. That always remains the prerogative of the decisionmakers.

O I n l . k -

              '=
 )'                       *.
               ..                      .                                                            OCT 4    8 83
 ;f, , ,                                                                                                                       ,

3 ISSUES 3.1.2 & 3.1.3

   .'. (

USE OF COST-BENEFIT ANALYSIS, [ 1. Statement of Issue j How will cost-benefit analysis be used in the decision process? 1 ( j 2. Implication of Issue To Regulatory Questions l This issue has a major impact on regulatory questions 1 and 6. Subissues

 'l
   .                       listed below need to be addressed when developing the decision criteria
  ?                        called for in regulatory question 1.
   ;                   3. Subissues

1. What attributes are to be used in cost-benefit analysis? What

is the basis for this decision?

2. How are these attributes to be quantified and compared?

3. What weights, if any, should be assigned to each attribute?

   .                       4.      Should attributes be quantified on a risk (probabilistic) or a deteministic basis?

5. How should uncertainties be taken into account?
4 4. Approach to Resolution i A Value-Impact Handbook to be published in October 1983 will establish j procedures for perfoming a cost-benefit (value-impact) analysis and dislaying information for the decisionmaker. However, this Handbook

{ y will not address the policy considerations underlying the subissues j described above. The staff will attempt to resolve these policy j ' issues by preparing position papers for Comission review. These papers will be discussed with the ACRS prior to submittal to the 4 Comission.

  ;(      '

5. Status of Understanding It is clear that what attributes should be used and how they should be FotA- t4 -%8 c,[ Gb

                 .I jpt.

O ik . , . ( weighted is a political decision which involves complex policy and judgmental issues. For example, the French have apparently given heavy i emphasis to protection of offsite property in detemining whether to require filtered vented containments for nuclear power plants. In the U.S. there are differences of opinion as to whether onsite damage and costs to the utility should be considered.

6. NRC Position

[ Cost-benefit (or value-impact) analysis provides a convenient method for looking at both the absolute and relative merits of proposed changes to improve plant safety. However. cost-benefit analysis should not be j4

' looked upon as a decisionmaking tool. The proper role of a good cost-benefit analysis is to display for the decisionmaker (and other interested parties) in an organized and standardized fom all of the costs and

!?                        benefits (and their uncertainties) associated with a proposed change j                           in plant design or operation. This infomation should be displayed in parametric form so that the decisionmaker can clearly determine the
  '                        sensitivity of'any conclusions reached to changes in each of the important

!; attributes. The use of any weighting factors should be explicit in the information display. There are several ways of displaying cost-benefit

3 results. These include " net-benefit" and " cost-benefit ratio." Each provides the decisionmaker with different insights. Regardless of

  !                        the method or methods used for displaying the results the infomation display should' include all information on the individual .sttributes.   ,

All assumptions underlying the information should be explicit. i

  '                         Neither value-impact analysis or any other fomal analytic method will be used as a substitute for informed and reasoned judgment to set policy or make severe accident decisions. The Comission will have

[ the option of selecting which costs and benefits will be considered in ' evaluating and comparing modifications, whether any of the costs or benefits should be weighted and what weight should be given to cost-benefit balance in making final decisions on whether (or what type) of additional protection for severe accidents is needed or desirable. I

          \
  =------+..._~__,.,___,

        ~'n              -

ylo 93 t 5 ( iSSur 3.1.2 SELECTION OF ATTRIBtffES TO BE USED IN COST-BENEFIT ANALYSIS -

1. Implication of the Issue to Regulatory Questions This issue assumes that efforts will be made to quantify the safety of existing plants and the advantages and disadvantages of specific

   .                    alternatives for increasing protection against severe accidents. It further l                     assumes that the results of such analyses can affect the Consnission's
   ,                    decisions. The specific issue at hand is which of the many possible measures of benefit and cost (or, more generally speaking, value and impact) should one apply in these analyses. Conceivably, different measures could lead to different conclusions about plant safety and about the relative attractiveness
.;                      of alternatives for improving safety.

2. Subissues What role will value/ impact analysis play in decision-making?

What is the basis for selecting from among the possible attributes? What and how great are the uncertainties associated with each attribute? l' Can these uncertainties be reduced? l* l~ i- 3. Approach to Resolution i- 'i

   }                                                              Considerable effort has been undertaken in the Severe Accident Risk

,j Roduct'on Program to develop a technique to display the tradeoffs in the )j valurs and impacts of possible plant modifications. The attributes used in - j these a1alyses are oriented to a risk-based decision approach. Different attributes are appropriate for a deterininistic decision approach. The measures of safety discussed in the August 5 letter to the ACRS on the decisionmaking process (design intentions, design margin, design and construction errors, operating experience, complexity, etc.) could be used as

    ;(                    attributes in a deterministic decision approach. A new task will be z
   !                                                                                                                                                         Fo/A 8'+ - 128
   ):

GS4 L, , -. ..

                                                                                                                                                                 ; e.                     ,
 ,             ,  ., - = - - - - - , , - - _ - - _ - , - - , . - , - - - - . . , . . - . , - . . , . _ . . _ . - - - . . - - - . . - - - . - - -                   - - - - - - - - - - - - - - - - - .- - - - - - - -

4. .

              ~      -         .
      .[                                                                       2

( undertaken early in FY84 to evaluate these measures of safety. In the same time period. RC-DRA will be responsible for developing a decision approach for the mid44 decision. The approach will describe the attributes to be used in the decision process. The approach will be presented to the SARP Senior Review Group for approval before being presented to the ACRS and the Commissioners.

4. Status of Understandina l Several previous efforts have developed and applied attributes for i use in value/ impact analyses. The list is long and includes quantitative l parameters such as:

o Core melt probability i o Curies iodine released from containment

  !                                o    Release category probability o    Man-rem public exposure

( o Man-rem occupational exposure o Acute health effects ~ i, o latent cancers i o Implementation cost >$ o Cost to NRC. 'I +

   ,                               These attributes may be presented in a determinir, tic form l{                      (fatalities) or probabilistic form (fatalities /yr) depending on the decision

!' approach. i,' Additionally, qualitative measures expressed in relative terms have been j explored. These include, for example: i o Analyzability of design o Regulatory significance I

  ,                                o    Public acceptability.

i t Dozens of measures of risk have been proposed. We know that the f( specific measure of value used can markedly alter the conclusions one might . i ; I bif).h)l A h l, 'i{ ' .

      -     =    .    . - . . . . . -      - . - - -     . . -            .-
         ~....

tofoYhI.3

  ;f ISST 3.1.3
  .                                     SELECTION OF WEIGHTS FOR COST-BENEFIT ANALYSIS
  !              1. Implication of the Issue to Regula'ory Questions e                               As with the selection of attr ibutes (Issue 3.1.2), this issue assumes that structured analytical methods will be used to quantify the i              advantages and disadvantages of specific alternatives for increasing l              protection against severe accidents. The specific issue in this case is the s

relative weights assigned to the various attributes used in the cost-benefit (or value/ impact) analysis. Obviously the weight a particular attribute carries with a decision-maker has a substantial effect on the relative attractiveness of alternatives for improving safety.

2. Subissues

( ,j Who is (are) the decision-maker (s)? , l~- What role will value/ impact analysis play in decision-making?

  ;      .                        What is the basis for weighting attributes?

1 How should nonquantitative attributes be weighted?

 .I
/

l 3. Approach to Resolution

  !                                Most structured decision-analytic methods include provisions for j                weighting individual attributes. As is the case with the selection of

{ attributes, there appears at this point to be no conscious effort to establish

,}                 a comonly accepted set of weights. There exists general agreement that i                public safety should receive highest priority followed well back by l[                  occupational exposure and cost to the industry. Beyond that, there are no j                 generally accepted guidelines nor plans for their development. Such l                considerations tend to be implicit within the decision-maker and the specific

!! decision at hand. Resolution of this issue is unlikely prior to the mid-84 decision

   !,(
   ..              and for some time afterward. This is primarily because assignment of weights Foi A-N-92.s
  ]                                                                                c.l st n... . , .      _ , ,.             - . : . . . ., ,. . ., .: . . . ,. , . ., . , ,c r . .

                                       .            - . . .    =.   =          .. ...      .. ..--  -            .

(. . . I j . . . . - I 2 i(

   ;               is less of a technical issue resolvable through research than it is a j                political issue resolvable through compromise over a longer time frame.

j The IstC/DRA will be responsible for developing a decision approach j for the mid-84 decision including the treatment of weights. The approach will f be presented to the SARP Senior Review Group for approval. The approach will . subsequently be presented to the ACRS and the Commissioners. j 4. Status of Understandina

 }                                    We know, of course, that the conclusions one draws from decision

analysis methods depend s'.rongly on the weights one assigns to the individual

attributes of the decision. We also know that when decision-makers are free I to choose, weights vary significantly according to the personal preferences,

  $                experiences, and interests of th'e decision-maker. In the NRC, the decision-

making responsibility is diffused vertically and horizontally. The weights which are implicitly assigned become, therefore, an indistinguishable mix of personal, traditional, and institutional biases.

 -h From the perspective of a totally objective observer, some bases for establishing relative weights are defensible. For example, attributes clearly specified in mandates such as the enabling legislation for the NRC merit' highest priority. Alternatively, the weights of quantified attributes may

decrease as their uncertainty grows.

  .                                    Efforts to equate dollars with radiation exposure (e.g., the ALARA criterion of $1,000/ man-rem) imply a weighting scheme of sorts. This particular criterion, however, was developed for use in evaluating routine

,. radiological releases and its applicability to considerations of severe l' accidents is untested. . '1 i! ii 5. NRC Position 1 lj At this time there has been no clearly expressed position by the NRC lf as to the relative weighting of attributes considered in severe accident j decision-making. The Commission's guidelines for performing value/ impact ll analysis are silent on the matter. Traditionally, effects on public safety are considered to be of greatest weight, followed by occupational health I ; i

                        . - _  m b b ' -.   ,.: % . .f $    [b    ,   ,
                                                                        .&   _      ..n ,.     *'I

_ . . _ .' 3I .1-[_.___.__

3 f I . f effects. Other important quantitative attributes, such as cost to the industry and cost to the IstC, have no direct bearing on safety. They are.

however, receiving increasing attention and, presumably, increasing weight in j decision-making. The weights assigned to less quantifiable attributes, such I as publi: confidence in the NRC are also difficult to establish.

j The NRC, if it so chooses, could mandate a set of weights prior to j the mid 84 decision assuming the attributes had been identified (see Issue

  ]                          3.1.2). It appears, however, that this would serva little useful purpose as i

long as j o The general agreement persists on the primacy of public risk j over industry costs, and

 .]                                            o        Decision-analytic methods are used in only a supportive role.
    ,                                          Rather, it is appropriate for the EC to take a position which maintains flexibility and strongly defends its discretionary powers on these matters. More specifically, the E C should
   ~

o Confirm the supportive role of value/ impact analyses, o In performing its own analyses, assign weights on a relative basis consistent with prevalent institutional (i.e., NRC) biases, j o Investigate variations in weights to see what effects these might have on the decision at hand, and !(

o Encourage analyses sponsored by others to use a similar j approach. 5 J l, t ! i l. I 1 i a..

                                                       . 7 . . . . . 7 g , , ,i,               , , , .

e, g:. n y .y 7 q. .

                                                                                                                            .,,y.
                                                                                                                               . 1, .

A. '

 )-
  ;                                                                                               DRAFT o                                                                                               MLErnst:psb i

l\ 9/30/83 1ssue 3.1.4 i Role of Safety Goals

1. Implications of Issue to Regulatory Questions At issue is whether and to what extent the staff should consider the Comission's safety goal in making recomendations to the Comission on how to decide the severe accident question. The staff could consider the safety goals with regard to whether and how much additional pro-tection against severe accidents is justified.

2. Subissues

 .                            a.      Should alternative safety goals be considered, since the Commission

(' asked the staff to consider alternative goal formulations during the two-year safety goal evaluation period?

b. What weight should be given to the safety goals, considering the uncertainties surrounding the results of PRAs?

3. Approach to Resolution The Comission's policy statement clearly states that the published j safety goals are not to be used in the licensing process. Also, the

!                                Evaluation Plan states that the safety goals should not be used in t

j regulatory decisionmaking. Rather, they are to be evaluated for a l two-year period, which encompasses the 1984 severe accident decision.

 '                               However, the safety goal evaluation period will be nearing its and at
?

i( Fo r A- 14-94 j

    - - . . . .             . . . , _     ..n..  . . .

9, I i ,. .( f the time of severe accident decisionmaking, and it can be argued that I - a safety goal perspective could be useful in arriving at decisions regarding i iI this important area of regulation.

4. Status of Understanding Safety goals provide a point of reference against which to evaluate the acceptability of risk posed by nuclear power plants. The currently i 'q published safety goals contain both qualitative and quantitative measures of acceptability. The qualitative ones are:

, o Individual members of the public should be provided a level ! of protection from the consequences of nuclear power plant operation such that individuals bear no significant additional ( risk to life and health. 1-1 2 o Societal risks to life and health and from nuclear power plant

operation should be comparable to or less than the risks of

generating electricity by viable competing technologies and should not be a significant addition to other societal risks.

The proposed quantitative guidelines are: 4

o The risk to an average individual in the vicinity of a nuclear power plant of prompt fatalities that might result from reactor accidents should not exceed one-tenth of one percent (0.15) of f

the sum of prompt fatality risks resulting from other accidents to l{

'i                                                  which members of the U.S. population are generally exposed.

(

;i i!

y.. ; :

c. . _

                                                                            -   ;            + ,..  ..      ~

p ___ - _ . . . _ _ . _ - .__ _ sa , -

            ~                                                                                                                            l
                      ~

5 l { l 3 , 1( l f o The risk to the population in the area near a nuclear power .k plant of cancer fatalities that might result from nuclear power

   .$                                plant operation should not exceed one-tenth of one percent (0.15)
   '                                 of the sum of cancer fatality risks resulting from all other
,                                     causes.                              .

I

o The benefit of an incremen,tal reduction of societal mortality !- risks should be compared with the associated costs on the basis of $1,000 per person-rem averted. i o The likelihood of a nuclear reactor accident that results in a large-scale core melt should nomally be less than one in ; 10.000 per year of reactor operation. { i While generally viewed as a step in the right direction, there has been ! significant controversy surrounding the safety goals, and the Connission itself has raised several questions to be evaluated during the two-year !

period. Among the more significant issues are whether there should be l

   -                           a containment performance goal, whether averted financial impacts (offsite                               r l

and/or onsite) should be included as a benefit, whether there should be

   .-                          a quantitative individual risk goal for cancer fatalities, and whether                                   (

2 there should be an ALARA principle. Also, the details of implementing i the goals have yet to be worked out; e.g., action levels and the time frame j perinitted to rectify an unnecessarily risky situation at one or more

plants.

i t .

I 1

        ,'[~n~T'rja[~.;~,*.~'               .-           ? *.        ,
                                                                           .;     -w                  -

i._ . . . _ _ _ _ _ _ =__ _ _ _. .__ h'. ; . !r j., ,. Ji

   ;                         It was the uncertainties surrounding the numerial estimates generated
                           'in PRAs, as well as questions of implementation of the safety goal in
   )                         regulation, that caused the Consission to instruct the staff not to use i                         the safety goals as part of the regulatory process. The Consission desired that the staff only use the safety goals after the fact, to ll  '
   '                         determine whether use of the safety goals might have made a difference

! .I in the decision that was reached. t l l 5. NRC Position '; i The NRC staff should not be guided by the Comission's evaluation plan Ii in making severe accident reconmendations. Any staff recomendations should be evaluated against the safety goals, and a reasonable effort I should be made to also compare the recomendations against alternative goal formulations. The severe accident issue is the most important issue facing the Comission today. and the Comission should be given j the full benefit of any information available to it that would possibly 1 ) add perspective and enhance their decisionmaking. This is not to say that the decision should be controlled in any major way by the PRA 1-i results and the safety goal comparisons. The principal basis for the 1

- decision should be traditional; i.e., a good, fundamental understanding
Y i '

i of plant perfomance; good engineering judgment; defense-in-depth; ade- . quate balancing of prevention vs. mitigation; conformance with the general

   )                                        '

intont and philosophy of NRC's present regulations; and an appropriate l f consideration of costs. The Comission should be provided the comparative

     '.                           analyses against the safety goals and the staff's judgments in this area, 1
 , j [-                           and,then they can decide whether to give much credence to this comparison.

l! j1 ..

          - ~- -                                                                        . -. ..                                                ..

8

                                         - . ; , 7.:77.m . ,p.
                                                                         ;,.     ,  u.          *
                                                                ,L_"            - - ,L        .                   ._-.,__ - -                          .,,_     -

                                               = .. . -     -

1 ISSUE 3.1.5 1 ( BALANCE OF APPROACHES TO ANSWERING THE REGULATORY QUESTIONS

1. Statement of the Issue The issue here is what is the appropriate balance of approaches to

   -                       answering the regulatory questions and in particular regulatory questions 3 (How safe are existing plants with respect to severe accidents?) and 4 (How can the level of protection for severe accidents
   -                        tiincreased?) Approaches here refers to alternatives such as deterministic, teobabilistic and systems assurance methods. This is distinguished from techniques discussed in Issue 3.1.1 for aiding the decisionmaker in making regulatory decisions once the basic regulatory. questions have
  ~

been answered.

2. Implication of the issue to Regulatory Questions The resolution of this issue has direct implications for how the Comission

~'(

decides the severe accident question, for example, whether it remains with

   '                          the traditional deterministic approaches or chooses to rely more heavily on newer approaches. The issue has significant implications for what additional research or information is needed since the different approaches
   -                          require somewhat different sets of technical data. The balance struck can significantly affect the direction of the Severe Accident Research Program.

3. Subissues I What are the strengths and weaknesses of each approach? How can each

   -               -          approach best be used to answer each of the regulatory questions? What are the advantages and drawbacks to the use of each approach for this
   ;                           purpose?

Do the NRC and nuclear industry have enough experience with probabilistic systems assurance methods to use them confidently?

  .]
   !(                                                                                       Fo / A -f4 -124

_j C. / s'? i

   ~7
                                          ,   ...y..          . . . . , , . .   . . , . . .          ,             ,       ,.    ,

r o , 2 4 I What performance criteria will be used to determine whether existing

.(

plants are safe enough? Can the selected approaches be used to measure agains't these criteria? , What information is required and available to implement each approach?

4. Approach to Resolution 3

An approach which is primarily deterministic but which utilize ( the strengths of both deterministic and probabilistic approaches is being recomended to the Comission in draft two of the paper " Severe Accident Decisions for Existing Nuclear Power Plants". This hybrid approach, if approved by the Commission, will be published in the Severe Accident policy Statement in early 1984.

5. Status of Understanding ;

There are several points which bear on the resolution of this issue, not all of which are primarily technical. '

- r

( First, the NRC staff and industry are most comfortable applying detarministic analysis by virture of its traditional use in reactor regulation. It usually provides a well defined scope of considerations and a clear cut indicator of acceptability. The staff is somewhat less comfortable with PRA, although its use has increased rapidly. There is less experience among the staff with the systems assurance techniques used by NASA and D00. Second, both the deterministic and probabilistic methods rely strongly on analytical models of physical phenomena which have never been experienced. These models (e.g., MARCH, MELCOR, CONTAIN, TRAC, etc.) will never be verified to everyone's satisfaction. They will always provide a

;                         source of uncertainty whose significance can only be evaluated using engineering judgment. The probabilistic methods have the further disadvantage of major uncertainties in factors such as external events, i                         and human performance which tends to reduce confidence in their use.
.(

i i 7.....,.

                                                                                            .,,   ;..y.
                                                                                                      .                                        j

           -                                              3 h
                                                                                                        ~.

Third, one of the attractive features of deterministic analysis (the

}(
 .                  DBA approach) has been the existence of performance criteria which help clearly distinguish the acceptable from the unaceptable, e.g.,

2200c F maximum fuel clad temperature or 10 CFR 100 dose limits. However, r,o generally accepted criteria now exist for evaluating reactor systems performance during severe accidents. Fourth, the extent to which regulatory decisions will be generic versus plant-specific must be decided. Available pRA's tell 'us that certain accident sequences, e.g., station blackout or loss of dece.y heat removal, are important risk contributors for all plants studied. The risk from certain other sequences, e.g., ATWS, varies significantly with the class of plant, and in other cases, a specific piant's risk is dominated by an accident sequence (an outlier) specific to that plant's design. Finally, we know that systems assurance analysis has been applied with apparent success by NASA and D0D to comp 1tx systems. It represents a bottoms-up approach to hazards management and does not rely on specifying ( and quantifying particular acciddent sequences as a prerequisite. Its specific techniques (e.g., failure mode and effects analysis, sneak ' circuit analysis) have been used in the nuclear industy though they are ! not a regulatory requirement. They can provide import insights into weaknesses in design and thus opportunities for improving safety. They cannot, however, establish the adequacy of a system without a standard acceptability.

6. iF. Position The principal decision path recommended by the staff is primarily determi-nistic in character. That is to say, it relies primarily upon engineering analysis of t.WR safety and performance; of the response of existing plants to core melt accidents; and of potential performance objectives, hardware,

;                  changes, and operational controls or procedures that could be backfit to
.(                                .

t

                                                       ~              _   - _ - . - -
                                                                                             .. c_  .. __

l improve safety for severe accidents. This deterministic method would also include a cataloging and assessment of the relevant considerations ( j for understanding the hazard that severe accidents pose, including some sense of their likelihood. Because of the primarily deterministic nature of the decision process and the primarily judgmental character of the

                ' final decision, the process would best be _ implemented _by_regu]ations.                 The process would no_t place primary reliance upon the Commission's Safety Goal,                  ,

but would use, quantitative engineering analysis where possible and where supported by data. - Frobabilistic risk assessment will be used to supplement the deterministic approach where justified by data and by a full consideration of uncertainties. pRA will be of value in cataloging and arranging in order of significance the dominant accident sequences and associated containment response for the

                      " internal events" and in providing a convenient additional perspective on risk judgments. PRA has the advantages of most direct comparability             ,

with the Commission's Safety Goal and of direct utility in NRC's required system of Regulatory Analysis for all new generic requirements. It is difficult to a priori prescribe the weight to be given PRA in severe accident decisionmaking. The weight will vary among the technical issues and the extent to which ongoing or past research has reduced uncertainty about them. In suminary, we recommend a deterministic app, roach that most highly values engineering analysis and judgment, supplemented as appropriate by PRA, for < . $ the severe accident decision for exis' ting . plants. t h

        ~+r - . . . .    . . . . , .   .,g.    . , , , , , , , ,    ,,                   , , , ,

OCT 4 1983 3 , ISSUE 3.1.6 lj (

BALANCE OF PREVENTIVE AND MITIGATIVE MEASURES , 1 . l

1. Statement of the Issue ;

Past regulatory practices although based on a " defense-in-depth" phitesophy f have emphasized prevention rather than mitigation of accidents. Plant l design requirements have been derived through the use of design basis l i accidents (DBAs). These are supplemented by requirements intended to

  ]                               ensure safety system reliability (e.g., the single-failure criterion, OA                                                                                              ,

requirements). The use of stric.t risk perspective could provide unlimited

  .j                                                                                                                                                                                                    l
   !                               opportunity, at least in theory, to trade eff prevention vs. mitigation to minimize risk. However, given existing uncertainties which are made                                                                                               .

more obvious by PRA, we are likely to continue to rely on defense-in-I depth with perhaps more emphasis on mitigation.

  !.i
  'I                     2.         Subissues,

1. How confident are we concerning our understanding of potential accident initiators, including imponderable evento such As sabotage and large

       ,                                             earthquakes, and plant response to these initiating events (i.e., where j                                                are we on the learning curve)?                                       ,

2. Where do we draw the line when we refer to " prevention" (e.g., no n,

   ;                                                 core damage. TMI core, core in-vessel, core in containment, containment 4                    .f                            failure)?

i

3. What tradeoffs are we willing to make between maximizing overall

    '                                                 safety performance a'nd maintaining defense-in-depth?

a i 4. What are the relative costs of prevention and mitigation options and I what weight are we willing to give to these cost differences. i i 3. Imolication of the Issue to Regulatory Duestions .

,  j                                   This issue keys directly into regulatory questions 1 and 6. Therefore, early resolution is important,.to orderly progress toward a Connission;                                                                                  l decision on severe accidents.                                                                                                                  .> -             ;
   *1                                                                                                                                                                                '
         ~                                                                                                                           FotA -84 @8                                        ;
                                                   .                                                                                 clste j              .9      - , . - .       ..                  g. us , , . .           4.
                                                                                             .,,.,.pg.    . . . g j

I* ~ "i f*

   * - (                 4.                      Approach to Resolution l
 -                                                It is suggested that this issue be included as part of a series of broad policy papers designed to establish Comission policy on Regulatory Question 1.

5. Status of Understanding The key to resolving this issue is to detemine where we are on the learning curve in our understanding of severe accident initiators j and response of plant functions and systems to these initiators.

i This will require a careful evaluation of op4 rating experience as well as deteministic and probabilistic stue es to assess our level l of uncertainty and how often we are being surprised by what we j learn from actual operating experience. I

6. NRC Position We are not ready to take a position on this issue although it is i[

  .N obvious that uncertainties are sufficiently large that we must continue to rely on both preventive and mitigative features as part                                                                                                                  ,

of our defense-in-depth approach to severe accident safety. Establishing a position on this issue will have to be accomplished in two phases. We should be able to establish early some guidelines the Comission can use to detemine a proper balance given a specific set of cir-cumstances. However, specific decisions on how this balance will be made must await (1) an assessment of where we are on the learning curve (2) answers to regulatory questions 3 and 4 and (3) and assessment of the uncertainties associated with these answers. 4 I5 l1 'e' i . O t

           *= w                                               ... ..g                      .. . . .. , , , , ..,                     ,,            , , , ,               , ,                          ,.   ,
                                    - + .7. . - y . .

l

I - _ . . ._. - - _ . . _ _ _ _ . _ _ _ =~ _ GEP

6 s93

   ,            NOTE TO: James C. tialaro

{ FROM: M. Fleishman I i

In response to your question concerning beyond DBA considerations for the

   !            CRBR, I reviewed the SER and spoke with Cardis Allen, who is in the CRBR Project Office.

The SER specifically states that core melt accidents (or core disruptive

    ,           accidents (CD8))are excluded from the design basis accident (DBA? spectrum (pg.15-5). Engineering judgment was used in choosing the DBAs (i.e.,

those considered to be credible). Durino the LWA hearings, intervenors wanted the hearing board to rule that Es should be considered DBAs. The NRC staff maintained that CDAs should not be DBAs. The board deferred a decision until the CP hearings when they reconsidered the issue. The board is now expected to issue a ruling on this matter in a few months. The main reasons a core melt accident is given more consideration for LMFlFis are:

1. Voigng of coolant from the core decreases the reactivity of an@ LWR but increases reactivity of an LMFBR; and

2. Fuel relocation decreases the reactivity in an LWR whereas it

( may increase reactivity for an LMFBR.

   ;            Because a CDA in an Ll1FBR is potentially quite severe (more so than in an LWR) they have been treated differently than in LWRs. While the CDAs are not part of the DBA spectrum, they have been given a considerable amount of attention. In particular, design features have been included both to (1) mitigate the consequences of a CDA, and (2) provide assurance that a CDA will be very improbable (pgs.15-7 and 15-8). The design criteria that have been used to evaluate the CRBR design in terms of CDAs are consistent with those proposed for LWRs in NUREG-0850 for Zion and Indian Point (pg. A.4-1).

Thus, while the CDA is not considered a DBA, significant design work was done and design features actually incorporated, because of CDA considerations. 4 1. Design modifications were made to the reactor vessel head to ensure that it could accommodate the sodium slug energy of i a CDA.

2. A total of 12 different desion features were either added or upgraded in order to mitigate the consequences of a CDA (pg. A.4-19).

3. 5 special design features were provided to provide assurance that a CDA will be improbable (pg.15-8).

(

                            *     ~ * * * * ' * *    - ,                       e ,e. . . .    ,   ,              ,}}

ML20127F110

Contents

Text

Subject:

Introduction:

SUMMARY

SUMMARY

SUMMARY

SUMMARY

SUMMARY

SUMMARY

SUMMARY

SUMMARY

1.1.2 Title

REFERENCES:

Navigation menu

ML20127F110

Text

Subject:

Introduction:

SUMMARY

SUMMARY

SUMMARY

SUMMARY

SUMMARY

SUMMARY

SUMMARY

SUMMARY

1.1.2 Title

REFERENCES:

Navigation menu

Search