ML20214X263

From kanterella
Jump to navigation Jump to search
Slide Presentation Entitled, Limited Application of Latin Hypercube Sampling for NUREG-1150
ML20214X263
Person / Time
Issue date: 12/09/1985
From: Haskin F
NRC
To:
Shared Package
ML20213E209 List:
References
FOIA-87-113, FOIA-87-60, RTR-NUREG-1150 NUDOCS 8706160404
Download: ML20214X263 (16)
v • d • e


Text

O' LIMITED APPLICATION OF LATIN HYPERCUBE SAMPLING FOR NUREG-1150 F. ERIC.HASKIN DECEMBER 9, 1985 l

l l

kDO'I60fj487061o SH0Lt.h07-60 PDR

- . . sk2

LIMITED APPLICATION OF LATIN HYPERCUBE SAMPLING FOR NUREG-1150 BACKGROUND OBJECTIVES: -

FRONT-END APPROACH ILLUSTRATIVE EXAMPLE RISK EQUATION OCP APPROACH t

LATIN HYPERCUBE SAMPLING WHAT IS IT?

WHY CONSIDER IT FOR NUREG-1150?

EXAMPLE OF LIMITED APPLICATION ADVANTAGES / DISADVANTAGES PARALLEL DEVELOPMENTS l

l 1

l O

e ww -&m-m --,g-----cr- w-+%-wg -aww- gmy -- -- ew,-ye -pg --erywe4-- 4w"w --wwe-ag w w qw e ,e,-eyey-w--weg,79- w e, eyr=-% m+*W-e ---,k*rwe- ir----

NUREG-1150 SENSITIVITY / UNCERTAINTY ANALYSES ---.

OBJECTIVES: -

4

1. PROVIDE A GOOD ENGINEERING PERSPECTIVE ON MODELING ASSUMPTIONS THAT DRIVE THE ANALYSIS (SENSITIVITIES)

~

l 2. PROVIDE A " REASONABLE" ENVELOPE IN WHICH THE ACTUAL VALUE WOULD LIKELY BE FOUND AND WHICH NEED NOT BE EXPRESSED IN TERMS OF FORMAL STATISTICAL BOUNDS (UNCERTAINTY) e e

e

GENERAL APPROACHES TO SENSITIVITY / UNCERTAINTY ANALYSES

1. BOUNDING METHODS
2. SINGLE PARAMETER VARIATIONS ABOUT A BASE CASE
3. MULTIPLE PA'RAMETER VARIATIONS USING MONTE CARLO AND SAMPLING DISTRIBUTIONS .

6 i

s l

e a

b

-- , ,..r_, - - - - - - -- w .. -.,~--..r. ,.w,. -

+,,-,,.-ww------

APPROACH TO SENSITIVITY / UNCERTAINTY ANALYSIS FOR CORE MELT FREQUENCY 'M II ' b)

~

F CM

=F CM D

1. USE LOG-NORMAL DISTRIBUTIONS TO CHARACTERIZE UNCERTAINTIES IN XD (DATA UNCERTAINTIES) .
2. SELECT A BASE CASE VECTOR FOR Xg (MODELING PARAMETERS) 4 6

1 f

i l

l t

l e

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

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

y ,., - , -,---w--7 ._.-,-,#w-,,,y .- , - - - -

le _'-v+e Mv *r--'w'"'ww'C-v"' -"vv""~~' ' ' ' --v *'- - * - ' ~ * ~ ' ' ' - - '

' - -M CORE MELT FREQUENCY DISPLAY

~

( __

C B A< -

( __

MOST -

PESSIMISTIC SENSITIVITY CASE BASE CASE MOST OPTIMISTIC SENSITIVITY A. RANGE OF MEANS CASE B. "5/95" -- BASE CASE ONLY C. "5/95" -- EXTREMES OF SENSITIVITY CASES se 4

ONE-AT-A-TIME SENSITIVITIES AN ILLUSTRATIVE EXAMPLE R=X1+X2+X3+X4+X5 ASSUME ALL THE X; ARE MODELING PARAMETERS (SENSITIVITY ISSUES)

ASSUME X g CAN TAKE ON ONE OF TWO VALUES:

ZERO OR ONE X X X X X N 1 2 3 4 5 BASE CA3E 1 0 1 0 1 3 SENSITIVITY 1 0 0 1 0 1 2 SENSITIVITY 2 1 1 1 0 1 4 SENSITIVITY 3 1 0 0 0 1 2 SENSITIVITY 4 1 0 1 1 1 4 SENSITIVITY 5 1 0 1 0 0 2 LONCLUSION: R=3 i 1 O

e

~

' /s

(

i V//,//

Y/ /,///// /,'/,'i

, f,'/f

~

// / / / ,

i -

W,/ / /it6,'44 .

/,f/ / /

1 /

< k?$ b,,' h/ E/h .

3- / / /3 ', Ipy.

l I' '

2- /

l

/s l/

l m # b h b id b  ? b,/ n b 1 2 h I h R = A1 + r.2 + X3 + A4 + X5 e

ILLUSTRATIVE EXAMPLE (CONT.)

R=X1+X2+X3+X4+X5 Xj=0OR1 ' '

' ~

POSSIBLE COMBINATIONS X X X Xq X R 1 2 3 5 ~

1- 0 0 0 0 0 0 2 0 0 0 0 1 1 3 0 0 0 1 0 1 4 0 0 0 1 1 2 5 0 0 1

  • 0 0 1 6 0 0 1 0 1 2 7 0 . 0 1 1 0 2 '

8 0 0 1 1 1 3 9 0 1 0 0 0 1 10 0 1 0 0 1 2 11 0 1 0 1 0 2 12 0 1 0 1 1 3 13 0 1 1 0 0 2 14 0 1 1 0 1 3 15 0 1 1 1 0 3 16 0 1 1 1 1 4 17 1 0 0 0 0 1 18 1 0 0' 0 1 2 19 1 0 0 1 0 2 20 1 0 0 1 1 3 21 1 0 1 0 0 2 22 1 0 1 0 1 3 23 1 0 1 1 0 3 24 1 0 1 1 1 4 25 1 1 0 0 0 2 26 1 1 0 0 1 3 27 1 1 0 1 0 4 28 1 1 0 1 1 4 29 1 1 1 0 0 3 30 1 1 1 0 1 4 31 1 1 1 1 0 4 32 1 1 1 1 1 5 25 OR 32 *78%) 0F THE POSSIBLE COMBINATIONS ARE IN THE INTERVAL (2,4) DEFINED BY ONE-AT-A-TIME SENSITIVITIES FROM THE BASE CASE.

e

ILLUSTRATIVE EXAMPLE'(CONT.)

R=X1+X2+X3+Xq+X5 .

Xg = 0 OR 1 _

RB=1+0+1+0+1 FOR THIS SIMPLE EXAMPLE, THE ENVELOPE OF ONE-AT-A-TIME SENSITIVITIES COVERS THE VALUES OF R ASSOCIATED WITH 25 0F THE 32 POSSIBLE COMBINATIONS (78% COVERAGE).

R = 3 1 1 (78% COVERAGE)

IN GENERAL, COVERAGE WILL DEPEND ON THE NUMBER OF DRIVING INPUT PARAMETERS.

ALTERNATIVE EXAMPLE:

R = 3X1 + 3X 2 + X3+Xq+X5 Xg = 0 OR 1 RB=1+0+1+0+1 FOR THE SAME BASE CASE, ONE-AT-A-TIME SENSITIVITIES YIELD R = 5+/-3 (94% COVERAGE)

S .

e O

COVERAGE R=Xy+X2 + ... + Xg Xg = 0 OR 1 ONE-AT-A-TIME TWO-AT-A-TIME ~

SENSITIVITY SENSITIVITY COVERAGE COVERAGE K

1-2 100% 100% ,

3-4 87.5% 100%

5-6 78.1% 96.9%

7-8 71.1% 93.0%

9-10 65.6% 89.1%

11-12 61.2% 85.4%

13-14 57.6% 82.0%

15-16 54.6% 79.0%

17-18 51.9% 76.2%

19-20 49.7% 73.7%

i 29-30 41.6% 63.8%

39-40 36.4% 57.0%

49-50 32.8% 52.0%

IMPORTANT: THE COVERAGE DEPENDS ON THE EQUATION FOR R AS WELL AS ON THE NUMBER OF INDEPENDENT PARAMETERS!

e I

F f

OBSERVATIONS

1. FOR PRACTICAL PROBLEMS, COVERAGE CAN SELDOM BE CALCULATED.
2. HOWEVER, THE CONCEPT OF A REASONABLE RANGE IS INTUITIVELY LINKED TO A. COVERAGE t

B. DEGREE OF BELIEF

3. IF ALL POINTS IN INPUT SPACE ARE EQUALLY LIKELY, COVERAGE IS EQUIVALENT TO " DEGREE OF BELIEF."
4. TWOMETHODSOFDERIVINGkREASONABLERANGEVIA SENSITIVITY CASE:

t A. HEURISTIC B. BY DESIGN 2

5. BOTH METHODS RELY ON ENGINEERING JUDGMENT.

i i

G g

-, .- ,,-.--,n,. ,e. , , , , - , - , - , - , , - - , , , , , , , , , . . . . _ - , , , , . _ _ , , _ .,,-----_._n, ,- _ n - ,.--,, r-,,

. 'T

. )

RISK E00ATION RISK, - [ [ N 0 7 g ,1 7 )

N , yg - C. M W j I J l SYtSOL DEFINITION SOURCE FRE0 7

FRE00ENCY OF CORE MELT ACCIDENT SE00EEE I ASEP. IN COORDINATION WITH SARRP CRMP 3,y PROBABILITY OF CONTAIR1ENT RELEASE SARRP CONTAIW1ENT EVENT ANALYSIS N0DE J. GIVEN ACCIDENT SE00ENCE I FP, FISSION PRODUCT SOURCE TERM FOR BIN TO BCL STCP CALCULATIONS, BINNING WHICH SE00ENCE I WITH RELEASE MODE J MEETINGS, AND SARRP ANALYSES IS ASSIGNED g W , ,) MAGNITUDE OF CONSE00ENCE K. GIVEN MACCS AND CRAC2 CAL.CULATIONS CONS SOURCE TERM FP O

IJ 0

  • gI

OPTIMISTIC - CENTRAL - PESSIMISTIC (OCP)

RISK UNCERTAINTY ANALYSIS

1. DEVELOP OPTIMISTIC, CENTRAL, AND PESSIMISTIC ESTIMATES OF FREQ g ,

CRMP gg FP B

IJ ,

l CONSg (FP g IJ

2. PERFORM ONE-AT-A-TIME, TWO-AT-A-TIME, ETC., SENSITIVITIES ABOUT THE CENTRAL ESTIMATE.

t i

I l

1 i

j 2

4

{ .

- LAN M FATAL _ITY/ YEAR . -

Itf, d-  :

1(f, .

" d, 4

" SOURCE VARIAllONS

  1. ! i Sequence Pt+wTai-yf gf, g 2 Sequence Data oc _
o. . .

3 Containment Release Mode IVebs.

w  :

a. m 4 Release Fraction 10-*.
  • 5 Consequence x .

t O n

  • an

~"g ec x

3 m 4 ~

m;

~

4 5* ~

b K

. an c . .

an 3=

16' n  : - -- -

- =  :- .

4 ,

g *,

g, . . . .;

e 9

nousut

! Smott SOURCES .

i l

l i

l > .

\ ,

O O

CRITICISMS OF OCP APPROACH

~

1. CONSISTENCY ISSUE DIFFERENT APPROACHES HAVE BEEN TAKEN TO DEFINE OCP-ESTIMATES FOR VARIOUS STAGES OF PROBLEM.
2. DIFFICULT TO ASSESS REASONARIENESS EACH BAR IS A' DIFFERENCE BETWEEN TWO CONDITIONAL POINT ESTIMATES WITH NO QUANTIFICATION OF THE DEGREE OF BELIEF.
3. QUALITATIVE SENSITIVITIES EVALUATION OF RELATIVE IMPORTANCE OF DRIVING FACTORS IS QUALITATIVE. .

e k.

WHAT IS LATIN HYPERCUBE SAMPLING?

l

  • A CONSTRAINED MONTE CARLO SAMPLING SCHEME WHIC:t PERMITS OUTPUT DISTRIBUTIONS TO BE ESTIMATED WITH FAR FEWER SAMPLES THAN REQUIRED BY ORDINARY MONTE CARLO.
  • FOR PURPOSE OF THIS PRESENTATION, CONSIDER IT A FORM OF EXPERIMENTAL DESIGN TO BE DEMONSTRATED BY AN EXAMPLE.
  • FOR MORE INFORMATION CONSULT IMAN, R. L., AND SHORTENCARIER, M. L., A FORTRAN 77 PROGRAMANDUSER'SGUIDhFORTHEGENERATIONOFLATIN

.HYPERCURE AND RANDOM SAMPLES FOR USE WITH COMPUTER MODELS, NUREG/CR-3624, SAND 83-2365 (MARCH, 1984).

4 8

e J

WHY CONSIDER LATIN HYPERCUBE APPROACH?

r:

1. OBJECTIVES: DRIVING FACTORS & REASONABLE RANGE
2. TIME, COST AND MANPOWER CONSTRAINTS A. LIMIT TO 10-15 IMPORTANT ISSUES B. LIMIT To 20-30 SENSITIVITY CASES
3. LIMITED QUANTITATIVE INFORMATION BASE:

A. FEW QUANTITATIVE ESTIMATES AVAILABLE FOR SOME ISSUES B. NEED OPTION OF ASSIGNING NONEQUAL WEIGHTS TO DISCRETE ESTIMATES.

E 1

l l

EXAMPLE OF LATIN HYPERCUBE FOR DISCRETIZED INPUT PARAMETERS 10 ISSUES, EACH HAS 2 LEVELS '~"~5 20 SAMPLES SAMPLING WEIGHTS Xy X 2

X' Xq X X 6

X X Xg X 3 5 7 8 10 LEVEL 0 0.1 0.2 ' O 3 0.4 0.5 0.6 0.7 0.8 0.9 0.33 LEVEL 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.67 NUMBER OF SAMPLES AT EACH LEVEL X X X X X X X Xg X 1 2 3 X.4 5 6 7 8 10 LEVEL 0 2 4 6 8 10 12 14 16 18 7 LEVEL 1 18 16 14 12 10 8 6 4 2 13 e

EXAMPLE OF LATIN HYPERCUBE FOR DISCRETIZED INPUT PARAMETERS 10 ISSUES, EACH HAS 2'f.EVELS 20 SAMPLES SAMPLING WEIGHTS ,

Xy X X X X X X X Xg X 2 3 4 5 6 7 8 10 LEVEL 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 '0.9 0.33 LEVEL 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.67 NUMBER OF SAMPLES AT EACH LEVEL SAMPLE X X X X X X X X Xg X 3 2 3 4 5 6 7 8 10 1 1 1 1 1 1 0 0 1 0 0 2 1 1 1 1 1 0 1 0 0 1 3 1 1 1 1 1 1 0 0 0 1 4 1 1 1 0 0 0 0 0 0 0 5 0 1 1 0 1 0 0 0 0 1 6 1 1 1 0 ,

0 1 0 0 0 1 7 1 1 1 0 1 0 1 0 0 0 8 1 1 0 1 1 0 0 0 0 1 9 1 0 1 0 0 1 1 0 1 0 10 1 1 1 1 0 0 0 0 0 0 11 1 1 0 0 1 0 0 1 0 1 12 1 1 0 1 1 0 1 0 1 0 ,

13 0 1 1 1 0 1 0 0 0 1 14 1 1 0 0 0 1 0 0 0 1 15 1 1 1 1 1 0 1 0 0 1 16 1 1 - 1 1 0 0 0 1 0 1 17 1 0 1 1 1 1 0 0 0 0 18 1 1 0 1 0 1 1 0 0 1 19 1 0 0 1 0 1 0 1 0 1 20 1 0 1 0 0 0 0 0 0 1 1

MAXIMUM CORRELATION COEFFICIENT BETWEEN INPUTS IS .51 l

hm _

EXAMPLE 10F LATIN HYPERCUBE FOR DISCRETIZED INPUT PARAMETERS 10 ISSUES, EACH HAS 2 LEVELS 20 SAMPLES SAMPLING WEIGHTS '~~-

X X X Xq X X X X Xg X 1 2 3 5 6 7 8 10 LEVEL 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7~ 0.8 0.9 0.33 LEVEL 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.67 R=X1+X2+X,3+Xq+X5+X6+X7+X8+Xg+X10 .

EXAMPLE RESULTS EXACT LHS R WT. WT.

0 0.00.01 0.0000 1 0.0026 0.0000 2 0.0206 0.0000 3 0.0850 0.1000 4 0.2010 0.1500 5 0.2847 0.4000 6 0.2435 0.2000 7 0.1229 0.1500 8 0.0346 0.0000 9 0.0048 0.0000 10 0.0002 0.0000 1.0000 1.0000 AVG. = 5.1700 5.1500 4

i j 't i

POPULATION DENSITY OA-1 f .

i O.3-l 0.2 -

s 0.1 - .

( .

0.0 I I I I I I l I l l g y 0 1 2 3 4 5 6 7 8 9 M 1

i .

.k 1

--- m p-%_ ..-- .--- ___m_- a- - . . ______-,w , _ ._A_ _m, _. -h--

_ .. _ & - - _ -C L e. m f *.

'1 i

i i LHS ESTIM ATE {N=20}

O.4-j 0.3-1 0.2-i

. 0.1 -

i 0.0 I I I I I I I I I I l 3 l;

0 1 2 3 4 5 6 7 8 9 m i

t l

4

\ .

p ._ , _ _ . . _ ,_. ----, . _---------m _ y r - _ _ _ _ .m.._-_ .._mm_a .,u. #..a . - .-.-4; _

-_;___m .m._

. _ ,__$ x ~a.

POPULATION DENSITY (SOLID)

SAMPLE ESTIMATE (DASH), N=20 OA- r---- :

0.3 - i  !

0.2- - - - - - - -

O.1 - ----

0.0 i i i i i i i i , i , ,

0 1 2 3 4 5 6 7 8 9 M 1

EXAMPLE OF LATIN HYPERCUBE FOR DISCRETIZED INPUT PARAMETERS 7 ISSUES WITH 3 LEVELS

_1 ISSUES WITH 2 LEVELS -

10 ISSUES

~

20 SAMPLES SAMPLING WEIGHTS Xy X X Xq X X X X Xg X 2 3 5 6 7 8 10 LEVEL 1 0.3 0.2 0.2 0.3 0.2 0.33 0.25 0.5 0.5 0.5 LEVEL 2 0.4 0.6 0.6 0.4 0.6 0.33 0.5 0.5 0.5 0.5 LEVEL 3 0.3 0.2 0.2 0.3 0.2 0.33 0.25

.. NUMBER OF SAMPLES AT EACH LEVEL Xy X X Xq X X X X Xg X 2 3 5 6 7 8 10 LEVEL 0 6 4 4 6 4 7 5 10 10 10 LEVEL 1 8 12 12 8 12 7 10 10 10 10 LEVEL 3 6 4 4 6 4 7 5 i

i e

EXAMPLE OF LATIN HYPERCUBE FOR DISCRETIZED INPUT PARAMETERS.(CONT.)

SAMPLE Xy X X Xq X X X X Xg X 2 3 5 6 7 8 10 1 2 2 3 2 2 1 2 2 2 1 2 3 2 2 3 3 2 2 2 2 2 3 3 2 2 3 2 3 1 1 1 2 4 3 2 2 1 1 1 2 1 2 1 5 1 2 3 1 3 1 2 1 1 2 6 2 3 ,3 1 1 3 2 1 1 l' 7 2 2 2 1 2 2 3 2 1 1 8 3 2 1 2 2' 1 1 1 1 2 9 3 1 3 2 2 3 3 1 2 1 10 1 3 2 3 2 1 1 2 1 1 11 2 2 1 1 . 3 2 1 2 2 1 12 1 2 1 2 3 2 3 1 2 1 13 1 2 2 2 2 2 1 1 2 2 14 1 2 2 1 2 3 2 2 2 2 15 2 2 2 3 2 2 3 2 1 2 16 3 3 2 2' 2 2 2 2 2 2 17 2 1 2 3 2 3 2 2 1 1 18 2 3 1 2 2 3 3 1 2 2 19 1 1 2 3 1 3 2 2 1 2 20 2 1 2 2 1 1 2 1 1 1 MAXIMUM CORRELATION COEFFICIENT BETWEEN INPUTS IS .32 o

O

} WHYNbTUSEAFRACTIONALFACTORIALDESIGN?

  • WOULD NOT PERMIT AN ARBITRARY NUMBER OF RUNS - FOR THE EXAMPLE, WOULD HAVE TO BE FRACTION OF 73 3 2 , I.E., 6,

~ 12, 18, 36, ...

  • WOULD REQUIRE EQUAL WEIGHTS FOR EACH VALUE OF AN ISSUE.

e

  • WOULD NOT PERMIT CORRELATIONS TO BE INDUCED BETWEEN i ISSUES.

s 4

e f

S I

I

mm-4,--%-p-~-r 4--9 w4s- -emy,-p.--er-wmm-.,-em,me--,,,- -er------w-mov--gne-m,we~m,-ww .-, y-,.s.-no-we-w w. w-w-'emnr-we Ne men.~--'-~'--~m'

s CAN INDUCE CORRELATIONS IF REQUIRED i

UNCORRELATED

~

0.2 3 -

1 2 1 X 0.6 2 -

4 4 4 R=0 0.2 1 -

1 2 1 1 1 1 1 2 3 0.3 0.4 0.3 X

f CORRELATED' O.2 3 -

4 X 0.6 2 -

2 8 2 R = 0.82 0.2 1 -

4 1 l l 1 2 3 -

l 0.3 0.4 0.3 l

X e

M

2 STEPS IN LIMITED APPLICATION OF ATIN HYPERCUBE SAMPLING ASEP/SARRP WITH AfD OF REVIEW GROUPS

' ~-- -- .

  • DEFINE ISSUES (10-15 PER PLANT)
  • SELECT A SET OF DISCRETE OUTCOMES FOR EACH ISSUE (2-4 PER ISSUE)
  • ASSIGN WEIGHTING FACTORS TO EACH OUTCOME (0.1 TO 0.9)
  • INDICATE CORRELATIONS BETWEEN ISSUES (CONDITIONAL WEIGHTING FACTORS) ,

4 ASEP/SARRP ONLY

  • PERFORM LATIN HYPERCUBE SAMPLING
  • PERFORM RISK AND RISK REDUCTION CALCULATIONS FOR EACH SAMPLE (20-30 SAMPLES)
  • DISPLAY RESULTS, EVALUATE SENSITIVITIES (REGRESSION ANALYSIS, ETC.)

6 e

=

W su --

-%w - m- m + -w-w e- mg m - , , w em--a mm- --

w-,--' o-- - -

LIMITED APPLICATION OF LATIN HYP'ERCUBE SAMPLING ADVANTAGES: .

1. ABLE TO WORK WITHIN IMPOSED LIMITATIONS. _
2. PROVIDES CONDITIONAL ESTIMATES OF A. RELATIVE SENSITIVITIES B. REASONABLE RANGE
3. FOCUSES ATTENTION ON MOST IMPORTANT ISSUES.

i l

l l

-rq -ewy--g-- -- - - - ---- y- ,---- v ,--, ,-,

-c e

LIMITED APPLICATION OF LATIN HYPERCUBE SAMPLING

~

~

. DISADVANTAGES:

1. DISCRETIZATION AND WEIGHTING OF INPUT PARAMETER SPACE REQUIRES ENGINEERING JUDGMENT.
2. RESULTS ARE CONDITIONAL UPON A. DISCRETIZATION AND WEIGHTING OF INPUT PARAMETER SPACE.

B. LIMITED NUMBER OF ISSUES CONSIDERED.

3. SEPARATE CALCULATIONS BASED ON RESULTS OF EXISTING STCP RUNS WILL BE REQUIRED TO QUANTIFY SOME ISSUES.

~

4. CANNOT DO A POST-ANALYSIS REWEIGHTING IF CORRELATIONS ARE INDUCED BETWEEN ISSUES OR IF VERY LOW WEIGHTS ARE ASSIGNED TO SOME VALUES.

l l

m 9

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

% /c PARALLEL DEVELOPMENTS .

1. IT Mal BE POSSIBLE TO COMPUTE ALL VALUES OF FREQ ,

i CRMP i a, FPIJ, AND CONSK (FP B ) CORRESPONDING TO THE ..

IJ POSSIBLE COMBINATIONS OF DISCRETIZED INPUTS. THIS WOULD PERMIT LARGER SAMPLES AND REWEIGHTING.

2. TEMAC (TOP EVENT MATRIX ANALYSIS CODE)

A. WILL REPLACE SEP CODE B. MORE FLEXIBLE, HANDLES LARGE PROBLEMS C. WILL BE USED FOR PEACH BOTTOM FRONTEND SENSITIVITY / UNCERTAINTY ANALYSIS.

3. PRA UNCERTAINTY EVALUATION PROGRAM (PRUEP):

WILL PERFORM A COMPREHENSIVE LATIN HYPERCUBE SENSITIVITY / UNCERTAINTY ANALYSIS FOR PEACH BOTTOM TC.

4. QUEST 2 AT BROOKHAVEN.

2 4

5 e

~ : ,

l lae:eq , ,b,'

ll \l~ /.i.*.'Yl'

g

,y A UNITED STATES NUCLEAR REGULATORY COMMISSION -

/f 3 WASHINGTON, D. C. 20655 r,, , hs

\o o...

/

MAY S Gd5 -

MEMORANDUM FOR: Themis P. Speis, Director -

Division of Safety Technology FROK: Zoltan R. Posztoczy, Chief Research and Standards Coordination Branch Division of Safety Technology SUECECT: MINUTES OF,THE NRC/IDCOR MEETING ON MARCH 26, 1985 heetirr Fernrary The purpose of the meeting was (1) to discuss the resolution of outstandir:g technical issues related to severe ac.cidents, and (2) to outline the proposed IDCOR program for 1985. The meeting was opened'by T. Speis who outlined the purpose of the meeting and the status of the program. Tony Buhl acknowledged the difficulties of achieving full agreemeht and preposed that each issue be resolved to the point where we can proceed with decisionmaking. Ecb Henry ci:: cussed a schedule producing a proposed methedclocy for the comparison and evaluatice cf plants by the June-July 1985 pericd and an evaluation of the trethedcic5y en a few selected plar.ts by the el d cf CY 85. Presenting first a prcpcsed approach to resolving the technical issues, Z. Rcsztoczy proceeded as chairman of the meeting-Seven of the eighteen cutstanding technical issues were discussed. Agreement was reached on the approach to be taken with six of the issues. The agreed on resolution of these issues is summarized belcw under "Resoluticn of Technical Issues". One issue, hydrogen generation within the reactor vessel, will need further discussion.

The 1985 IDCOR program, as it was described at the meeting, will develop a methodology permitting the application of reference plant results to other operating plants and will apply its methodology to selected plants. Initial NRC reaction to the IDCOR '85 program was very positive.

A list of attendees is enclosed. Copies of handcuts are available from the HRR contact, James Watt, RSCB (X28279).

Resolution of Technical Issues Summ= ries of the resolution or status of the seven issues discussed at the meeting are provided below. ,

,p3 3 !!

h_Ja m

U ay e&]), ~ff\ -

pr c;x 0 99 .

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ EJ/2 n

e \

\

.o .

1 l

  • ~ ~

MAY S 1985 l l

- Is' sue al Fission Product Release Frior to Vessel'Failurc '-l IDCOR agrees with the proposed fiRC position on this issue,. Thct is, IDCOR will:  !

1. Put in a new fuel relocation slump model and user input "eutectic"~

terperature, thereby dropping the recuiratrcr.t ir. tirir current accel that fuel only'begins to relocate after it reaches 3100'K.

2. Retain their fission prcduct release model which is based upon the oxidation of UO., They cite a curve which shows very good agreement with the releas6 used in BMI-2104 which is based on ORNL data.
3. They will put in a "Te-retention by unoxidized Zircaloy" model a la the BMI-2104 mooel and will, therefore, release Te during the ex-vessel stage.

The above changes, if implemented and checked out prcperly, should resclve this issue. .

NRC: G. Karino IDCOR: J. R. Gabor

- Issue #2 Recirculaticn of Coolant in Reactor Vessel In-vessel and ' primary system natural circulation cf superheated steam may effect accident prcgression: increasing H producticr., decreasir.g primary system retentior., potentially altering pri$ary system failure location, and altering ccre melt timing, quantity and bulk temperature.

The staff and IDCOR agreed on the above-cited phenomenology, and also that the above concerns appear limited to PL'R high pressure secuences. Ecth parties agreed that at pressures of the order of 1000 to 1200 psia, typical of the small break LOCA, natural circulation would be less significant en the everall accident progression. It remains to be seen how important these effects are at pressures in excess of 2000 psia,

, typical cf a station blackout scenario.

. The staff proposed that this issue be addressed and resolved by sensitivity studies in the impacted areas. ICCOR basically agreed with this proposal, however, ICCOR ir.dicated that they intend to tnelytically eddress this is:Le for the short term and will perforn calculations with consequential steam generator tube rupture only if the analysis indicates pcter.tial tube rupture. The success cf the IDCOR approach depends on the development of credible models in a timely manner.

I '

KRC
J. Rosenthal

! IDCCR: M. A. Kenton

--m_ .m,-_--.,.--,.__-_-,__.e.--- _

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

MAY S 1985 Issue #3 Release Model for Control Rod Material ,

IDCOR and NRC are in substantial agreement on the path t'o' resolution of this issue for PWR control rod material (Ag/In/Cd). ,

1. New data becomin'g available from in-pile experiments ever the next six months will be used to further assess which contrcl rsd release model is more physically representative.
2. Sensitivity / uncertainty analyses to explore aerosol concentration effects would be useful.

Regarding BWR control mater'ials (B C), IDCOR does not believe significant 4

amounts of steam will be available following fuel relocatier tc react with Bt C reds because of blockage and channel box confinement. This is ccnsistent with IDCOR's positicn en H, generation in BVRs. It was agreed thct NRC would provide IDCOR with res61ts of the ORNL (SASA) scoping analyses.on effect of B C - steam rea.ctions durir.g BWR accident sequence as

- they become available 4before the June 1985 meeting. Data from ORHL out-cf-pile core melt release experiments with B4 C control rods will be available after June 1985.

. NRC: P.. Silberberg IDCOR: R. E. Henry

'- Is:ue fi4 Fission Product and Aerosol Deposition in the Primary System Although the NRC staff continues to be concerned about the validity of the IDCOR empirical aerosol model for the range of conditions and phenomena applicable to RCS fission product and aerosol transport, IDCOR and ARC agreed on the following path to resolution:

1. IDCOR will continue to explore the appropriate range of aerosol model decay constants recuired to test the applicability of tFe medel for RCS analyses. Results will be sent to NRC in July 1985.
2. IDCOR will assess the empirical model with a detailed (mechanistic) aerosol model.
3. IDCOR will continue to compare the empirical model against applicable

, experimental data.

Finally, the NRC staff believes it would be useful at some point to make a clear comparison of the IDCOR model with the TRAP-MELT 2 code on an appropriate RCS benchmark analysis.

. NRC: M. Silberberg -

IDCOR: M. Epstein/R. Henry l

r- -- ,, . , - - - . ,--,,.,w-,,,,,.-w.- -

. MAY S 1983

- Issue 65 In-Vessel Hydrogen Generation 1 '- - -

, For both BWRS and PWRs, the IDCOR and NRC predictions of in-vessel hydrogen generation differ in the amount cf hydrogen produced, the timing of the production and the mechanisms which affect the prcduction. ICCOR predicts lowe- hydrogen production due to steam starvation caused by fuel channe?

blockage, and also due to a cucher of mecherisms wtich ter.d to inhibit zirconium-steam reactions, once the zirconium has liquified. The PFC mchl predicts er;htnced tydroger preducticn efter the start of fuel relocation.

NRC is currently investigating two mechanisms, in-vessel natural recirculation and B,C cccbustinn, which could further enhance b.ycrogen production. Finall), there are operater actions which coulc also increase the zirconium-water-interaction.

Given the NRC estimates of in-vessel hydrogen producticr., prcr.pt hydrogen-burn failure of an ice condenser containment is pessible.

Overpressure failure of a B1
R Mark I is also possible for large LCCA sequences. The NRC position is that .these failure modes cannot be ruled cut based en in-vessel hydrogen production totals. A pcssitie path to resolution is to bound the probcbility of such events based en both hydrogen production and hydroger corrbustion phencmena.

IDCOR suggested that NRC accept the channel blockage model, based on its conformance with experimental tests and its ability to calculi.te the 2

observed hydrogen production rates for the Three mile Island accident. FRC insisted that evider.cs er.ists tc refute the channel blockage tocei. b F.C also pointed out that the TMI hydrogen results can be predicted without invoking channel blockege.

No consensus on resolution of this issue was reachec. however, NRC has agreed to assess the possibility of comparisons of March 2 and SCDAP with i- the TMI results.

l

, NRC: R. Barrett l IDCOR: M. A. Kenton -

--Issue #6 Core Melt Progression and Vessel Failure Both NPC and IDCOR have developed simplified models for core melt progression and vessel failure, basec largely on separate effects and engineering judgment. NRC has an ongoing effort to produce a mere mechar.istic code, MELPRCG, which will be available by the end of 1985.

IDCOR has announced plans to incorporate a more mechanistic melt progression l model in the MAAP code, with completien expected by the sunn.er of 1585.

r l While the NRC recognizes the importance of more mechanistic ac& ling, we are l not cor.fident that in-versel relt progressicn phencnera car te relitbly ,

predicted. There are no integral test data, nor are there likely to be any _

in the near future. Furtherecre, core relocation phencrena ars dependent en

- operator actions and other stochastic effects.

  • ~-,-e .e--, ------.-.--.---m e. . m.,_ , - - - , ,..-,

ww , --,.w---.-m,+ -----------y

.3 MAY 9 1985 Core melt progressior phencment can affect many important aspects of the -%

accident sequence, including hydrogen production, direct heating, and steam explosions. They also impact the composition of core debris and therefore the release of refractory fission products due to core-concrete interacticr.s.

hRC asserted that (1) melt progression is not well understood. (2) melt-prograssion and vessel failure phenomena shcult nct be invcked tc resclve containment f ailure issues and ex-vessel fissier prcduct release, and (3) such issues shculo be addressed from other perspectives.

IDCOR agreed with these conclusions.

NRC: R. Barrett IDCOR: M. Kenton/J. Gabor , ,

- Issue #7 Alpha-Mode Failure by In-Vessel Steam Explcsion A special review group (Steam Explosion Review Group) was cppointed to review tl.is issue. it.e review group consisted of experts f rcm universities, natientl. laboratories, the nuclear in.dustry and one member each from the United Kingdom and. Germany. The findings of the review group are docutented in NUREG-1116. "A Review of the Current Understanding of the Potential for Contair. ment Failure Arising from In-Vessel Stear Explosicns".

Eoth ICCOR and NRC agree with the conclusions of the review group, namely (1) Gecurrence of a steam explosior cf sufficient energetics which could lead to alpha-mode containment failure has a low probability.

(2) There is a need for a continuing steam explcsion research program which would improve our understanding of certain aspects of steam explosion phenomenology and which woulc be directed toward providing confirmation of the above conclusion.

NDC is sponstring a research program which is consistent with the review groups recomrr.endation. No action is expected from IDCOR until new research results are available. At that time IDCOR should review the results.

Generic trplicability of IDCOR Methodology The IDCOR approach to generic applicability was outlined by Ecb Henry. This included e description of the general approach which is divided into two parallel efforts, the first being an approximate assesstent of the core damage prevention capability and the second an evaluation of the release to the environment. The concept for assessing the core camcge prevertion capability is to address the influence of major differences in designs from those of the reference plants. A similar approach is used fcr the primary system and containment response through MAAP analyses.

The BWR methods develeptrent for core damage prever.tien was prtstr.teo by -

Ed Burns of Delian Corporation and included a descriptier cf the generalized -

approach, functional level event trees and a specific exarrple of functior.al event trees along with nodal, questions for the trees. Mike Hitchler of

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

MAY s 1985 L'estinghouse presented the PWR methodology for core damage prevention with ~-

the approach being directly related to the extensive information developed on Zicn and the other plant specific and generic results developed by the NRC arid the indestry. Lastly, Marc Kenton and 6'eff Cator of Fauske & Associates reviewed the approach to individual plant analyses for containment response using the HAAP analyses of the IDCCR referer.ce plants. These ar.alyses would concentrate en the differences in the plant design with respect to the reference plant and the influence these differer.ces vce'c have on the release of fissicr. products to the environment.

Z Ywn S R

c. .

Zoltan P.. Roszteczy, Criff Research and Stancards Coordination Brarch Division of Stfety Technology

Enclosure:

List cf Atter. dees e

9 e

1

4 List of Attendees ~ '- : -

NRC/IDCOR Meeting 3/26/85 '

J. Rosenthal NRC 492-9447

- R. Barrett NRC 492-7591 M. Silberberg NRC 427-4737 G. Marino NRC 427-4270 M. Fontana ENERGEX (615)481-3300 J. Raulston TVA (615)632-3063 R. Henry . FAI (312)323-8750 M. Leverett EPRI (415)855-2936 T. Buhl ENERGEX (615)481-3000 Z. Rosztoczy NRC 492 4221 T. Speis I;RC 492-7517 i R. Bernero NRC 492-7373 S. Shelly .

UC.S (202)296-5600 L. Connor NRC CALENDAR' 22S-6553 R. Borsum B&W 951-3344 P. Morris WESTINGHOUSE (412)374-5490 M. Hitchler WESTINGHOUSE (412)374-5070 M. Kenton FAI J..Gabor FAI (312{323-8750 (312,323-0750 M. Plys FAI (312)323-8750 4

1. Burns DELIAN (408)446-4242 J. Carter. ENERGEX (615)d81-3300 M. Epstein FAI (312)323-8750 J. Mitchell NRC 427-4614 D. Paccione LILCO (516)929-6300(X3619)

, D. Giessing DOE 353-3692 S. Kirslis NRC 492-8997 J. Read NRC 492-8301 M. Cunningham NRC 443-7984 P. Cybulskis BATTELLE (614)424-7509

5. Bassett NRC 427-4281 L. Chan NRC 427-4715
A. Wang NRC (202)634-3267 J. Watt NRC 492-8279 R. Wright NRC 427-4717

. W. Pratt BNL (516)282-2630 T. Theofanous PURDUE UNIVERSITY (317)494-5757 i

o e

9 6

. - - - - . , , , . -, - , - . - , - - - - - - - - - . - - - - - - - - - , - - - - - - - ---,--,--,,-.e.-c,<r- . - . - - , - - - - v - - - -

3 4'.', *

  • M [ .(eI - *

~

s j * '

,ya .

I t

c1.,

9-23 ', g ,

If! U sf to be about fourteen types of cavity configurations which will be itemized and described as follows. , fl .

Nj 10 All the PWR plants owned by IDCOR participant utilities were cate- !F

.l' gorized in two manners for ready reference. The first grouping is an alphabet-ical listing of the plants, so they were accordingly categorized by plant  ?

designation, NSSS vendor, containment type, reactor cavity type, A/E, and ..

[

utility as detailed in Table 9.5. Since preliminary information also indicated .}

that the reactor cavity groupings of these plants Was also substantially related to the A/E of the containment, these plants were also categorized with j:Q respect to NSSS vendor and A/E as indicated in Table 9.6. 4 These plants as categorized in Table 9.6 yield about 14 types of h geometries, and these are detailed in Table 9.7. Each of these types are briefly described as follows.

7- f W

9.6.2.1 Type A ,j i

.- Nij }

. o This configuration, Fig. 9.5, is representative of Braidwood 1 & 2  ? i t c Byron 1 & 2, and Zion 1 & 2 all of which are Westinghouse /Sargent & Lundy j i plants. This configuration would allow debris dispersal during a high pres- .

j sure blowdown of the primary system as pr'edicted in the Zion Probabilistic i Safety Study [9.4] and confirmed by .the ANL [9.2,9.3] and SNL experiments 'f

, [9.6]. The containment structures that could possibly entrap debris after it M l was ejected from the reactor cavity and instrument tunnel would be the lower side of the seal table room and two of the nearby walls.

f4 l

I

'y .

9.6.2.2 Type B qi{

, a i

This configuration, Fig. 9.6, is representative of Seabrook, Indian [

Point 2 & 3, and Trojan which are Westinghouse / United Engineers and Constructors [

and Bechtel plants. This geometry would also not retain much debris after high pressure melt ejection. It also has an additional manway which would

]

t serve as a vent during such an event. The debris that would be ejected and lg propelled upward from this type cavity would impact upon the bottom of the 'd "

floor of the' seal table room. There is somewhat more surface area for the '

p RlL? 5

.. o q

a 9-29 'jl h

is a

so 4

I

-%g\

N . J' C

3

\, S, i

\ @

L

> \ -

3 g

m / 8 l

/ 3 .

/ I

/ 3 8

U g 4 W

L L

- .'l elW'in C *,.

O

  • ese l N 'I a W. I

< Ii '

. .1;-

a.

h

, W f h '

. l, l

w

?

I I l

$a

- . - , - - - - - - -,,--.--,-----w------ -, , _n,,-

. 9-30

\ o

\

3

\.

e . , s

/ 8

. /

/  :

7 b

M-

$ 5 -

e

$ t s- -

/ . A z

.a E

l . &

l c

f

,. **s M - -

.j

9-31  :

debris to impact upon than for a type A configuration and thus more potential for aerosol generation.

.m.---

9.6.2.3 Type C

>- i This configuration. Fig. 9.7, is representative of Sequoyah 1 & 2.

Catawba 1 & 2. McGuire 1 & 2. Diablo Canyon 1 & 2. Vogtle 1 & 2, and Watts Bar ,

1 & 2. These plants all. have a Westinghouse NSSS but a variety of containment f

A/Es (TVA, Duke, Pacific Gas & Electric, Southern Services, and Bechtel). .

This type has a reduced potential fo'r direct entrainment and an increased l potential for retaining a' considerable amount of debris at steps and standoff regions away from the main gas flow. The manway into this reactor cavity would serve as a vent during such an event. The impingement locations for an'y ejected debris from the reactor cavity from which an aerosol could perhaps be  !

, produced would be under' the floor containing the seal table and on the edges  !'

of some of the adjacent walls.

9.6.2.4 Type D This configuration, Fig. 9.8, is representative of Millstone 3, Beaver Valley 1 & 2. Ginna Harris 1. North Anna 1 & 2 Robinson 2, and Surry 1 & 2. All these plants have a Westinghouse NSSS, and the containment A/Es were Stone & Webster, Gilbert & Associates, and Ebasco. This configuration is expected to retain essentially all of the debris. Most of the debris would be anticipated to initially accumulate at the sump end of this reactor cavity type.

1 9.6.2.5 Type E This is the configuration of South Texas 1 & 2 which are Westinghouse /

Bechtel plants, Fig. 9.9. Little debris will escape this reactor cavity i i

because the instrument tubes are individually sealed through the concrete wall )!

in the reactor cavity and the manway represents a tortuous path to the upper i ,;

compartments in the containment. The manway continues around in a circular -

pattern for about 45' from the door location illustrated in Fig. 9.9 after j which it starts up a flight of stairs. . It is the horizontal surface area of b i

w

9-32 [,

i

-N  ::

o t

\

% ._ /

4

+

////////////////a .

i Fig. 9.7 Type C (Sequoyah) lower reactor cavity configuration.

l l

I ,

i l

i -- . . -

g s @9& -"

. ii 9-33 . j:

l -

- f.& .

e t

. i

. l

.l 3

  1. ,.-------=====% ,

/ \ b .

c

/ -

> \

\

8 I e I x

\

gr ~

/  %  !.

N, ----

u

% f a  :! l g-  :

. i ;

E  !'

i 1

L  ;

C  !

3 '.[ .

i A

O )

y 5 c

.o a

1' W I

- t I

z w

b a e , g I'

gh.

f e i M f

. I

\1 en r I,

s ( '

.u

i it 1;.

T't n;

el i h

t

.. _ _ _ _ _ _ _ _ _ - _ _ _ _ _ _ _ - _ _ _ _ _ _ _ _ _ n

9 34 'p

. T 7

,/" f-NB Wl/ I t

/

Y / S ~~ ~~ ~ ~~~ ~1

/j tl /

/  %=== ==Y /

I Z '

///// ///////// \ l

?  !!! .

r"F l .

Fig. 9.9 Type E (South Texas) lower reactor cavity configuration.

l l

r, l

', 9-35 -

the roof in this stairwell compartment that provides nearly all the impaction i area for the hypothetical generation of an aerosol that could then possibly be i carried up further into the containment. -

l 9.6.2.6 Type F i i

This configuration, Fig. 9.10, is representative of Calvert Cliffs 1 l

& 2. Arkansas 2 Millstone 2, and Palisades which are.all Combustion Engineer- l ing/Bechtel plants. Most,all of the debris would escape through the annulus I between the reactor vessel and the biological shield. Most of the ejected l l debris would then impact on the missile shield over the control rod drive -

mechanisms and be accumulated in the refueling pool. (

l 9.6.2.7 Type G This is the configuration of. Oconee 1, 2, 3 which are Babcock &

3 Wilcox/ Duke-Bechtel plants, Fig. 9.11. Not much core debris is expected to

]

escape from this reactor cavity since the instrument tunnel is dead ended.

The instrument tubes are individually sealed as they pass through separate ,1 penetrations in the floor before they enter the seal tank. Hence, significant h fractions of the debris inventory could not migrate up into the containment $

via this path. The s 0.76 m diameter access passage to the reactor cavity is f

brick filled but without any mortar. If this is opened by the blowcown, debris would be distributed onto the containment floor and would not provide for significant direct heating. n 9.6.2.8 Type H N n

i' This configuration, Fig. 9.12, is representative of Summer 1, Maine [

Yankee, Palo Verde 1, 2, 3. WNP 3, Farley 1 & 2, Prairie Island 1 & 2, and j Turkey Point 3 & 4. These plants have Combustion Engineering & Westinghouse [

NSSS, and the containment A/Es were Stone & Webster, Bechtel, Ebasco, Southern Services Gilbert & Associates, and Pioneer, Table 9.7. Little debris is '

expected to escape from'this configuration partially because of the trapping of such material at the sump end of the lower cavity. The upper wall adjacent to where debris would accumulate at the sump is also chambered thus provided E

Y-

f A_...

. 9-36 #

1 .

~

N ,e,

__sNN

\\

%x E

\\ h 11 N

N 2 i

11 N t ff N  :

//

// N

\

U 5

l

/ / \

N

- --====___ -=_-m#

- n/ a x

'q j .

f e

S e

D

9-37 E i

l d

M .

e o,

UO e-o oe d 3o S a a

>o 2

\\\\\\\\\\\k\\\ \\\\\\\\\\\\\  :.:

l-r-

y--- ----- --

gg ,

i I il I '

i a Il3 I ls Ii

> lli g

l u

u l Q l --

i

11' C t i'

lI I . s 2 i il I I T 2 l; t _, _ _ _'b _ __ .- _ _ LI- - -

  • E g j{

~

S e l

, . i

.m y 1:

E E.*c 3

uU pums ll tn G , .

cm

..c ."Q eo

\ '

o L o n 5

3, ff l?

I i

l

' I l..

n 1 ..i.

i l

..g .

9-38

-f.

,. n s y

l

) *

!a 'N ,

d\\\\\ i, 0

/ .

> \ a N..,.___.__..__/ )

s &

j

( \

.ua .

!"""hNNNN\ .

9

. ~..... ..

9-39 I

i additional flow area for the gas / vapor flow to negotiate the turns without ---i entraining the heavier debris, j l

. i 9.6.2.9 Type ! ,

t This configuration. Fig. 9.13, is representative of St. Lucie 1 & 2, Point Beach 1 & 2, and Waterford 3. These plants have Westinghouse and Combustion Engineering NSSS, and the. containment A/Es were Bechtel and Ebasco, Table 9.7. Not much core debris is expected to migrate to the upper regions 5 i

of the containment primarily because of the tortuous path involved in this configuration. Furthermore, the flow area enlarges as the upper floor is approached, reducing the velocity of the debris.

9.6.2.10 Type J jl This is the configuration of San Onofre 2 & 3 which are Combustion Engineering /Bechtel plants with top-mounted in-core instrumentation, Fig.

9.14. A considerable amount of core material is expected to escape from this reactor cavity through the cooling ducts and up around the RPV. Most of the ,

debris would probably pass up around the RPV since the total cooling duct flow i area is only about 40% of that around the RPV. Consequently, most of the ll debris impaction area for possible aerosol generation would be on the main 1 coolant lines just outside of the RPV before they penetrate the concrete  !

biological shield and also on the missile shield over the control rod drive mechanisms. ,

1 9.6.2.11 Type X l $

1 This configuration, Fig. 9.15, is representative of Arkansas Nuclear  !

One and WNP 1 which have Babcock & Wilcox NSSS and containments designed by }!

Bechtel and United Engineers & Constructors, respectively Table 9.7. Not j much core material is expected to escape from this type even with the large volume of the instrument tunnel because it is dead ended where the instrument I tubes pass th'ough r individually sealed penetrations in the floor supporting the seal tank. The most likely debris migration path in such an event would l l

l 1

f .

u _ .w 7 : . -

43 9-39 I

additional flow area for the gas / vapor flow to negotiate the turns without j entraining the heavier debris.

{

9.6.2.9 Type ! --

l This configuration, Fig. 9.13, is representative of St. Lucie 1 & 2, ,

Point Beach 1 & 2, and Waterford 3. These plants have Westinghouse and

]

Combustion Engineering NSSS, and the. containment A/Es were Bechtel and Ebasco,  ;

Table 9.7. Not much corg debris is expected to migrate to the upper regions l

of the containment primarily because of the tortuous path involved in this configuration. Furthermore, the flow area enlarges as the upper floor is approached, reducing the velocity of the debris. ,

9.6.2.10 Type J .

This is the configuration of San Onofre 2 & 3 which are Combustion i Engineering /Bechtel plants with top-mounted in-core instrumentation, Fig. I 9.14. A considerable amount of core material is expected to escape from this

[

reactor cavity through the cooling ducts and up around the RPV. Most of the y debris would probably pass up around the RPV since the total cooling duct flow j area is only about 40% of that around the RPV. Consequently, most of the h i debris impaction area for possible aerosol generation would be on the main f

coolant lines just outside of the RPV before they penetrate the concrete  !

biological shield and also on the missile shield over the control rod drive mechanisms.

9.6.2.11 Type X R

p, l This configuration, Fig. 9.15, is representative of Arkansas Nuclear 4 One and WNP 1 which have Babcock & Wilcox NSSS and containments designed by {

Bechtel and United Engineers & Constructors, respectively, Table 9.7. Not b much core material is expected to escape from this type even with the large ,

volume of the instrument tunnel because it is dead ended where the instrument f

tubes pass through individually sealed penetrations in the floor supporting l the seal tank. The most likely debris migration path in such an event would  ;

' 9-40 g

- *j

.+ -

5 _

\\\>\\\ (

s s iN t N

q s N s s N - ~,3 l  :

$ s

\ lRPVl \

s t j s 's/ '

\\\

i Fig. 9.13 Type I (St. Lucie) lower reactor cavity configuration.  ;

~.;

iT

'. 9-41  !! h

\V I

ii[

- fif l e il i g---- ._.

3 .g ,

/ N .:ll

,/,

~

r ,. 9, .;

~

's ' y- sq

}

,[l lI 1

I J RPVF

\

h I

l h

l

~ ~

s ,s ~

l

\% #  % ,,,, / / ': 1.!

N / *

%- --- -/  :

3 3 3 j '

N5 N N N t N,'r 's N

\o cr \ ' ' ' ' - N N l N ,

N s a

\ t\ ') I L\S \ s ii N N , i N s 1 klIk x h N  !

5 '~ /

x mmsm i b

s.

Id Fig. 9.14 Type J (San Onofre 2 & 3) lower reactor cavity configuration. f l

]

I 8

' 9-42 .g!

!?

.- - j

/

, is sr /

INCORE IN S T R U.

ROOM f/N i

i

~

Y ,**%

- - s 7~,

V//91 ( 7u,

/ < <

f '/

~

f '/'

,- ,'""l' } W/Y /

, l l g j

} l !l lN C O R E INSTRU.

/

/

%wdV 19 ROOM l

\

/

/

l Fig. 9.15 Type K (Arkansas 1) lower reactor cavity configuration. , _

^

l l

l I

9-43 a;

be up around the RPV where the impaction area would be that imposed by the q!'

main coolant lines between the RPY cuter wall and concrete biological shield ,

as well as on the missile shield. .

f!

9.6.2.12 Type L -

,i i'

This is the configuration of Bellefonte 1 & 2 which are Babcock &

Wilcox/ Tennessee Valley Authority plants Fig. 9.16. Even though the instru-ment tunnel and the ventilating ducts.are shown in the same plane for simplic-ity in Fig. 9.16, they are actually rotated about 120' from each other as p

illustrated in Fig. 9.17. Some core material would be anticipated to escape  :

from this configuration. Most is expected to, traverse up around the RPV and also out through the outboard cooling duct, Fig. 9.16. Not much is anticipated h to migrate through the smaller flow area, inboard cooling duct, or through the  !'

large volume, instrument tunnel because it is sealed where the instrument l tubes pass through the floor supporting the seal tank. The impaction of core d debris would most likely occur on the structure above the outboard cooling r.

duct exit and on the main coolant line in the annulus surrounding the RPV. j 9.6.2.13 Type M k

~

N.i.

This configuration, Fig. 9.18, is representative of Calloway 1, f'

" Comanche Peak 1 & 2. Wolf Creek, and Davis Besse 1. These plants have '

' Westinghouse and Babcock & Wilcox NSSS and containments designed by Bechtel, Gibbs & Hill, and Sargent & Lundy Table 9.7. This particular reactor cavity p1.

arrangement is expected to retain a good fraction,of the core material ejected from the RPV. Most of this debris is anticipated to accumulate in the corner

[

i' opposite the RPV which contains the sump. The debris retention capabilities -,

are enhanced by the manway which could serve as a vent and the chamfered wall adjacent to the sump which would create more flow area to relieve the gas / vapor flow during the blowdown.

9.6.2.14 Type N

.. "g ,

,i This is the configuration of Yankee Rowe which is a Westinghouse / Stone j

& Webster plant, Fig. 9.19. It is expected that a reasonable fraction of the II i

3

a

$12

'u.

,:~.

l g

4I

'1

)sEEE:

r yyD '

l.

N N ss A / / i Ns$qs s

\

\x n o

8 \

i t

a NN i r

u g

\ t f

n o

c y

c t v\\s?s h \N d u \

\

i v

a c

r

\ .

a\

n t o

\ \ i l

o

\ c a

e

\ o r

\\\ \Nx l CM \

r e

w o

lv

!i

\ l V \

)

P \

t o

e n

R l e \

f e

d. Ili n l

l

,I n \ B e

u \ (

_ \\k t t

n\

e

\

L T

e p

y NNN \

ks e m r u otr \

\ 6 1

9

_ cs nn \ i g

ii

\ F _

sNN\N\

Y I

l k [sN\\\x _

%j \wNN\\\s'M\L

- s ,'s ' '

\\ 's

_A s s >

. 4 .lj;: . ii l j,

l .-

1 9-45 -

.i i

~-}i= i, e

\ k I .-

//NNN jj l

l" f

.,e g i, : e I s

, 3,,y.-----

/

- -, \_ . .

s / .

% # Ventilating

\ # - ,#

l l Ducts i I;

l

\ incore In s t r um e n t a tioll1

}l i I Tunnel ql

/ . /

i. .
n/u /# y t l i l 9{;

ti!

Fig. 9.17 Orientation of in-core instrumentation tunnel and ventilating ducts in Bellefonte 1 & 2. ,,

\!- *

e. t lil 41 I

!f i

! I

. 9-46

.~

e

/ Y

/

/ / '

9/Z2 s - / 3 EZ j j

/ I l l

/ ,,

' I )

/ < l l

l f/

!\ j l '

i

%d l )

'/

fr==A Fig. 9.18 Type M (Calloway) lower reactor cavity configuration.

, s.

  • ~

9-47

[

El  %

' / / /

/  ;

y'

/ . / > >

//////

/,/,///

Y ls f/ ls f c

  1. s

/ 6 n e--- 1 =4

<_ _s > 9 z 4 f V ,7'"f

f. .g/li jl p_; ,/7 sy /

/ \J /  : 9

/ +,/ 1 1

i r

concrete supports /

/

'~~~ ~

~ ~- -- 0.51 m dia, manhole c

Fig. 9.19 Type N (Yankee Rowe) lower reactor cavity configuration.-  !

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

Retrieved from "https://kanterella.com/w/index.php?title=ML20214X263&oldid=3295169"