ML20087M628

Nonproprietary, Methodology for Addressing Superheated Steam Releases to Ice Condenser Containments

ML20087M628
Person / Time
Site:	Catawba
Issue date:	03/28/1984
From:	WESTINGHOUSE ELECTRIC COMPANY, DIV OF CBS CORP.
To:
Shared Package
ML19268E734	List: ML20087M617 ML20087M625 ML20087M628
References
NUDOCS 8403300119
Download: ML20087M628 (240)

Text

l 4 ~*

, ()

METHODOLOGY FOR ADDRESSING SUPEPHEATED STEAM RELEASES E -

ICE CONDENSER CONTAINENTS Purpse ,

Tne purpse of this report is to document the information presented on March 19, 1984 in a meeting with the U.S. NRC Containment Systems Branch on the status of progress made in addressing the confirmatcry item on the Catawba Nuclear Plant Safety Evaluation Report. This confirmatory item deals with the effects of superheated steam generator mass and energy releases following main steamline break accidents. Attachment 1 includes the list of attendees at the meeting and the overhead slides covered in the Westinghouse presentations.

Technical presentations were made describing the modeling of the steam generator and heat transfer frcm the uncovered tube bundle during the steam generator ,

blowdown alcng with a description of the containment model and transient response. A proposed plan of action was also presented and discussed with the Staff. In accordance with that plan, this report represents the first milestone in the proposed plan of action. As cor:rnitted to in the meeting, the appendices present proprietary infcrmation which relates to the specifics of the models and sensitivities that were not directly addressed in the meeting.

Attachment 2 is an explanation of, and refers to, the overhead slides (Figures) presented at the March 19 meeting. -

4 L

e 8403300119 840328 PDR ADOCK 05000413 A PDR v-

j h I

1

+

1 t

l I

t ATTACHFINT'1 m

M*

8 ..

e O

l t

l l

I

[

3 i

j

s ALs ~ - , .s/ i r/ g.+ .

'11uh4 D / R . l- au , .

IL.E Th _ L . '~. %.ac. AL .,9 ., . '.,

ko-a . C eg o O - No.,

y v W, Al. JA &B'O dR.

NEC /DL

• l2 '. O .* 3/M12P r=-- Dcau /xPD J.' L. L i rr ie*

@ pa c w . 4- 5+W

% .V. TouAG w' coc. E AR S AsrT7

,0i*C./JL

~?d $e a W y vu.s+g. G Arc.,y b5.Lwe =

M u clee,; ~~ So.4 S*

F1y' booc 4 bJ A . Li k. W tr cle<>< 3,1-f<h L . 2. A/cwhiru- .E .vak -

M c.A cen w iv.,ctu < .G Ge +y r 7 s

iwoc m A ti mco , y e( m y.-

S .D Afe.M e D u.I<. e. , C< s ~ Ev.- ~. .,

S 5 &*llv D *.de , b<r:7 E y:sw:

x - % m La

7. E. cc,us 7us

$6/Cri LICEdshtc firiff , ECa*:

E. AMELI k .MQe.A c.ws,J Q ,a a M sa.]. -

W. O. C ,.,,& . .

T V A, G ,.3 : .. ,

• b es op S. I. l.In of k- 0.u. , KJ ,,- f..tJ.- k,eb.- So! '

W NI~ O f/!![ *

& tc)Q rn _,

g ara / suru rery

'2 H. o~.e J .6 /r24KWr-- /Ub& C..+ @

• %4 Q Q

.l. Rak. 6 - #RC/CSB C. Li 115 C,4 i.' G M. Af0Noaceset ef/AC/C ST3 m

4

- . . _ _ _-- - - - - - - _ - _ _ . _ _ _ _ - _ _ - _ _ . . - - - _ _ _ . _ - _ _ _ _ _ _ _ . - _ _ _ _ _ _ - _ - - _ _ - - _ _ - - - _ _ _ _ _ - - _ . --A.-. - - _ . . . _ . _ - . - - - - -

L..

. LOFTRAR ESDEL

! STEAM GENERATOR SEcoblDARY ,

i STEAu n.oW .

f .

y SATORAmED  ;

l *. *

- -

• 4 STEAA gg.steg

-; ' . ' ~ '

i

.31 hVl.

Q' ,sxtua ,, m,aw en

\

4 *

' \< s  % .

O g'%'Jiu , -ig(.:. - " .

v -

~-m,:, .

. l w  % .

b

• StMGLE NCDE , 2 gtEGacd MCOEl l

• MO NEAT TRAdSFER. To SATueATED STEAM. REGlog ,

. l I

• HEAT TlitAMSFER TC SATORATED i WATER REGicd 15 MODIFIED Rl4it.

Tues 'uencen#estt*

1 l FIGURE 1 .

l

\ .

4

, s  ;

L l

LOFTRAM MODEC

. suPERHEAT HEAT TRAMSMR l

! SUPERHEATEb STEAM FL.cW '

SEcoHDAR.T k SUPERHEAT HCCES l a\ 7 0.WPERMEAT N M E HQDES

~

WA6% .

2%

t.r emCTIVE ^

UNCOVERY LEV 5iL ,

~

~ !'

i

' '

• VARI ABLE. SOPER.%E.AT .

HODES MODEL CoHSTART PRtMARM -TEMP 7 RATURE

=

IM. SOPt.RMEAT 12.W Gtcd ASSOMTt FCS.

HE.AT TRANsppt. CALCVLATICMS

• C ALCV LATED 5OMWMF AT HEAT TRAMS 74.R ACCCOMTE.b FCR. IM

, pft, mas.T Tit.AMS1EWT

- nme2'_ - - . -

t ,.

TUBE UNC0VERY LOFTRAN SUPERHEAT MODEL

.860 FT2 BREAK AT 102'PC POWER 1.2000  :  :  :  :  :

G 5 1.0000 -- LOOP 1 2

S y .80000 -- --

E a .60000 -- --

a u

w

> .40000 - -

O Z

2 E

a

.20000 -- --

S LOOP 2 0.0 8 8 8 8 8 8 o 6 6 6 6 e e d S 8 8 8 8 TIME (SEC) 4 FIGURE 3 m

, MASS BLOWDOWN LOFTRAN SUPERHEAT MODEL

.860'FT2 BREAK AT 102 PC POWER 2000.0  :  :  :  :  :

1750.0 -

i b

v>

150b.0 -- --

N 1250.0 -- --

s S

z 1000.00 -

w G 750.00 -- --

M E

O 500.00 -- ---

250.00 - - --

0.0 O O

O O O O O O O '

O O O O b b b O

b O

. O O O O O - cu m r to e TIME (SEC)

I FIGURE 4 I

s I

ENERGY RELEASE LOFTRAN SUPERHEAT MODEL

.860 FT2 BREAK AT 102 P,C POWER 1300.0 1275.0 -- --

3 SUPERHEATED STEAM s

1250. 0 - - --

a_

c- 1225.0 - - --

3 ~

A

$ 1200.0 -

~ --

z SATURATION s

5 1175.0 -- --

z

<C W 1150.0 - - --

w S 1125.0 -- --

1100 0

. y k g k- d g e 5 5 5 5 5 5 6 2 8 8 8 8

~

TIME (SEC)

FIGURE 5 i

TEMPERATURE TRANSIENTS LOFTRAN SUPERHEAT MODEL

.860 FT2 BREAK AT 102 PC POWER 650.00  :  :  :  :  :

'1 600.00 ',

0

$ 550.00 - --

c.

8-6 i

SG TUBE INLET 5e 500.00 -- (PRIMARY SIDE) --

m g -

Q~

• b 450.00 -- -

y SUPERHEATED STEAM o.

h 400.00 -- --

r 350.00 - - --

SATURATION 300.00 '

8 8 8 8 8 8 o d d d d d d 6 3 8 2 3 '8 8 TIME (SEC) l

! FIGURE 6

. e 00'009 lg  ;

e.

'l 00'005 1 .- _ _

j 00*000 0

W w

i E

= <,

S! b - e 00*00E 5

~ 1 I E

~ x o'

5 u ,_

$o a z at ta D <

$5 x 00'002 05 Y

zx s

u_ \

o J

00'001 a

j. i 0*0 o

e o

e e

e o

o e

e o

e Z 0 2

2 O

$ e ee (WE7/01E) Ad'1VHitJ3 FIGURE 7 .

O58 -

l O58 O5? )

C E

S

(

E A M I

T N

0 N P S A M I

R -,

1 T R l

l OS8 A F T P

M O O O L N CE T

PA M R UW RO TL OF Od2 NK

- A NE '

AR N RB T N F s O

L OdS f N

\

• od 0 0 0 0 0 .

0 $

0 0 0 0 0 0 0 0 0 0 5 0 5 0 0 2 2 1 1 5 0 uJ s r }"

1 1 =: >o :I 7g Ya"

?eAw i

k EFFECTS OF ANALYSIS ASSUMPTIONS INITIAL STEAM GENERATOR INVENTORY AUXILIARY FEEDWATER FLOWRATE FEEDWATER SYSTEM FAILURES 1

PROTECTION SYSTEM EPR0RS l

t e

A FIGURE 9

=- -- . . - - .

ADDITIONAL MODEL CONSIDERATIONS

~

LIQUID-STEAM. INTERACTION IMPROVED STEAM HEADER MODEL

~

~

! HEAT TRANSFER THROUGH TUBE WRAPPER TEMPERATURE DROP IN PRIIiAnY SUPERHEAT NODES 4

OPTIONAL VOID CORRELATIONS 5 <

i i-O t

FIGURE 10

. , e i o SHIELD SU/LDING DCME' ~

. / .

j

.,fl l

. g .  :

N (El829.C

'--STEEL CONTAINufMT VE33EL .

-!  ? :.

f/75 TCH POLAR CRANE Y.A.sF Spyggy *

( . PCLAR CRANE .*

BUILDING '

M tLL.-- ,

g

. i i 3

. t ,

TCP CF D n A pTCP OF CRANE RA/L *

1. 39'F/N C L /N.

)

(EL 794.4,ECX & 2 i WJA"2 / EL 79712 { EL 711.$

ICE LCACINC *; ,

't .,

^ Y)

^

CYCLCNE

• 2 h AIR

. &='T'78RIDCE CRANE VENTILATION a .. ,_- 2-'. g,ygg,pg UNIT'~;~'

w-e.n vNiTs Q lWl' <.0 ;A TCP OFICE SED _,

.y EL 780.0 %_,;,- .

f orvroEp gepplER

_, _ ,- . EL 774.43 % _ .

,y .x. ,,, .

~ *

.' CRANE gg 744.23-L*

• N f 1,

/C.,. .

'. y w,444 .* g

.:y:..3.l~

CONDENSER & ~. .

}

e

,. v , ., < :cNrA.NueNr .

s,

.. , m . - , L M A, ,

srEAu e ..

\; . . aENERAroR t,

s

. v : 1 -

CoNrnotocoooivi .

L:dsi.s$ '

.. M/55/LE JN/ ELD , I  : i. nsj .*

, , _ ,gg gggg I l EL 737.23

{

c'

'\ ,

, ,QT ' ' ' ' " " ' * '-,s i IJ. , f-, EL nr.is A- -

y \ /L .. .

!.f

. ~

\, - l l p ,

s4

_. j rR1i I,- .

i

.i L e

<,,,, a .. v. . n u ..

lg+!, a sAaniER ACCUuulATOR-1.'

u d,"(ggcro"

• pyup Is.o.

,o. ..

.~ .

s'% w ,. . . .

. gm

• W

(:

f* , ,

C# ACE .t .. I h.

ii

• EL ICS.0 %

- ~ , - -, , ,

. Vi

,p

~

f l } p- .k 00R EL 702.0a  !!

g.: ,

. p ,.*

7 { } .

y, VENT FAN sg:mEL 395.0~ ;g. . f.*} , { g U *' *-

EL 69J.0 i

{l q'.,.; . 4,.e, ).

y, , . , , , , , ,

4 EQPT

~

• _. , . a m e, n

] le i ..*

, . e .q {a- - 1

• i ,

3 .

.. .. .q'. ..

• w h

,- EL 679.78 s*.e -

"'~~~ ha \*

liit .*

C

'.. 3:n . _.- . . * * *

=; , y 48 '-

6.,a ,. . .,. . ,. . . ,,1 .

.~

.ti , .

e lql % ' REACTOR -.*.

i

. . is

• STEEL UNEp .. . .

.. . .y.- .

k ANCNOR soLT * ., ,,

. .i, ,

,j . / // : 0.304d tr7

1. EL 45/.69 .i e'

. ,l

-J .

. ,1 r

saun /. . ... .

a.im #

d"' - Reactor Building Elevation FIGURE 11

=

. l l

~ L ._

SPENT

% ,yyytypy gggg qb :'" ,g'.Zq

(" ,Y .I FUEL, Pir .i .,.  %. ,v.; <

s.d f  : s .~. a . .; , .% . e.- '_f.3 FUEL TRANSTEp h . - P!PE CHASE --

{qY SYSTEM .f. ' 'r .t; . e. <.. . <;.; f.

&l , ^-l . 6 ., . .. kl . *A .

'5 ll N - 47 1, e. ..

EL E9J.0 e- , ,7

, , , , ,, ,,- ] , ,,,

~

• ^ '

.. REACTOR UPPER v. s I .' >

. ,~ . . s

• . , l ~ ,, INTERNALS STORAG[ -

p EL _c, 5 crop

\,pg, 690.88 ' .1: l, ;(,.,. , y_

, }, .. = . _. .y

/ .s. .a . .

O f,.k"_. ,'

p~.

. 'Nr - or COOLANT .

l)

...A..

w n;5 s% .- . ' /

pudi: a;

.* /

.Q . , %L OWER\

• i.

%. Q.. .

.. e ..

/NTERNALS

~

• a . -

, . TORAG -.,

.E *I' .

~.- ~

C k") W  ?'.',Y' .<^ $ .

f.

O

~~

~

_'f

%q . ' . .

-= & -

~

4* 0 l&;g,7ga

~?% .

6.* -s

(,

,3 1-J STEAM GEN (4)

., , y ,yy y. . *e

• ' ., 4 ,

~.-- REA CTOR ..

,,]

i STEAM [. y _< pagS5ygg .,,

(L. -

f .- ,

-*1 ,

~

.a , ._ ESSEL e. CRANE WALL

}

C'"'n==r:_'

. EL ,. . p'

.c C>. ' W.

s'- TC2J5 .

FEED. .-. . ~

• g WATER -e EL 679.78 ~

^ *VENTIL A TION

' r.

RELIEF ,. l f.

4_

'O Nx ,

.r*  % ;f UN/M (4)

- - - - - _ _ ,. j.

' ' ^

. PRESSURI.Z.E.R. . . . ' . .e--

. . i,'5. .

s/ -e CUMULATORS M

• l': W'? k TABLE EAL

~ .

, /*

,, M. '

.......g /EL 693.0 HEA T EXCHANGERS . ,,

~

& ggg j \. .* .

VENTILATION UN/TS '(2) w .. -

5*

• *

• PERSONNEL LOCx y

• ,,n. ,.

Figure 3.8.3-7 Plan-Lo.er Compartment k FIGURE 12

r 1 9 1 1 I

LOTIC-3 CONTAINMENT CODE

• 4 N0DE CONTAINMENT MODEL l CONDENSATE /REVAPORIZATION MODELS LARGE BREAK (TOTAL REVAPORIIATION)

SMALL BREAK (CONVECTIVE HEAT FLUX) i

• WALL HEAT TRANSFER MODEL l

i

• MODELS SUMP-RECIRCULATION SYSTEM l

3 i

! FIGURE 13 l

1

, w-,- , ,1 - ..- -

" " e v v " 1 M

• s r

l i

l f

I i

I A A Scrays A A

, ,. .3 3 i  : " wall Upper Comaartment I

i ms ,ns,l A T

i

.m. , h, 4  ; U.C. ,

Drain .

e

, l.

~~-

, ~

! Ice *s ' 's G

ice ' Concenser m,h

, a a

- One way flow wall  ;

, tnrougn open area L, '

. on operatino decx Ccncensate & A Ice Melt lm,h 3 3

) ,

I.C. Sumo _

• a ' q' a '

I.CTi Oraini ,

v

• l i * ' * **

I RC5 b r=- = x #10w '

• s' s l m , n, s o accumulato lOead-Ended m'n Ccmoarwent ,

a a release, et I '. L:wer Ccmoarcent boil-of#)

, Air Addition Taole 8

.' od " 'I C I Heat Accition Tacie L.C. '

'Orain " sumo wall

i. - -- i_ _ --

U.C. -

• 7 .

Drain I = L.C. Sume

~

TIGURE 3.3 M.A55 AlfD CIERGY TI.0W DIAG M FOR THE COM2ARU.E::T5 FIGURE 14 O

• =

LOTIC METHOD OF SOLUTION SOLVES CONSERVATION OF MASS, ENERGY, AND M0 MENTUM FOR UPPER, LOWER, AND ICE CONDENSER REGIONS ONCE NEW LOWER COMPARTMENT CONDITIONS ARE DETERMINED, COMSERVATION EQUATIONS,ARE SOLVED FOR THE DEAD-ENDED COMPARTMENT AND FOR THE FLOW RATE BETWEEN THE TWO COMPARTMENTS e

FIGURE 15 i

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ _ . _ . _ _ _ . . . . _ :__. _ . _ b

TYPICAL CONTAINENT TETERATURE TRANSIENE (DRAINSf0DELLED)

1. LEVEL REACHED IN S.G.

-J U. . SPRAY ~

[ DECK FANS ON T J REDUCTION IN I.C. DRAIN FLOW f

e l

0 100 200 300 LOO '500 600

. TIfE (SEC.)

FIGURE 16

i FDDIFICAfl0NS TO THE CONTAINTET PODEL WALL HEAT TRANSFER MODEL C0tNECTIVE HEAT FLUX MODEL ICE C0f0ENSER DRAIN 10 DEL DEAD ENDED C0ffARTENT t0 DEL 4

4 4

FIGURE 17

t *

. 4 i

WALL HEAT TRANSFER CODEL .

1 ORIGINAL LOTIC 50 DEL l qy "= hWyami(rZAT - TVsil) l ItDIFIED LOTIC MODEL

//

C0f}g g y. ' $ ~

,, cony ,tg )

h  %-

.t .

• COnD = f( %,, a/

i . hc ,,, = f(T T Wolf j SAT 1

L =RT T) Well 1 Snr O

FIGURE 18 m ___.___ _ _ __x___-

. y i

COWECTIVE HEAT FLUX FDEEL ORIGINdL LOTIC MODEL i

i

_Scono _

I

~

l Cono bfg }rorat byg (l' f X I '

fDDIFIED L0rIC MODEL h cono = A<o%

hry

=

Sroraz hpg [! +X, u bene [Noix'T gro rns.

5 FIGURE 19 g

4

• f j i

IE CONDBiSER DRAINS

-APFROXIAf%TELY 20 ICE CONDNESER DRAINS

-DRAIN ELEVATION IS ABOUT I40 FEET FROM FLOOR

-DRAIN PIPE IS 1 FOOT IN DIAMETER

-FOR TYPICAL lblB TRANSIENT, DRAIN FLOW VARIES FROM I4000 LB/S TO 500 LB/S FIGURE 20

= - - _ _ - - . _ _ _ . - -

ICE CONDENSER DRAIN F0 DEL i

-CONDENSATION OCCWS AT THE SWFACE' 0F THE STREAM

-FLOW IS WELL MIXED Q = h A AT

-MODEL AS A WALL AT A CONSTANT TEMPERATURE' a

-A IS THE SmFACE AREA 0F THE STREAM q

i

-h IS A CONDENSING TYPE HEAT TRANSFER COEFFICIENT FIGURE 21

_1_'_.

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

I l

. t a CALCULATION OF THE STRENi FLOW AREA 1 .

e P = n (P x L) = a o (P x yO) = 800 p WHERE P IS THE PERIETER OF THE STREN4 4

~

N I.

! I

'// ,

Y IT-

y '

/ b

- I.

f( ... . ._3 L \

h' u

6 f y'2.ff $ *($]

. a l

l FIGURE 22 L

~

I

. . . 1 1

O BDDIFIED LOTIC DRAIN FUDEL o

-WALL WITH A VARIABLE AREA g=hcono A(% -T,,)

Butx 4

hr* 230 F l

. I i

l FIGURE 23

.1

+ .

1 i

f i

1 i

DEAD BCED C&PARTENT iCDt.L i

4 e

.I.

5 '

I

(

i i

)

I

.l J

. I j MODIFIEDLOWERC&P/RTIET l 1

I

_i._.. .

4 e .

.e b

1 4

0 e

t*

. FIGURE 24 f

-- .- - , , , . - 4 y, 9

o l

CATAWBA RESULTS .

102% POWER 0.86 FT BREAK MAXIMLN APd FLOW FSAR HEAT SINKS MAXIMLN S.G. INITIAL MASS T'%X=

f 324*F (LO)ER (al'ARifB(T) .

6 0

?

' FIGURE 25

. 7..g .

?,

T l

ADDITIONAL MJDEL CONSIlERATIONS i

i I

J ,

a o 1 -WALL HEAT TRANSFER lODEL l 4

e-T

., -DRAIN MODEL 1 .

-DEAD ENDED C0ffART?ENT MODEL e

'I 4

4 1

i e

D -

i, h

I FIGURE 26~

f T r

i

.i 1 q . . .

ATTACHENT 2 ,

i OUTLIE OF,1HE REPORT  !

t ,

, 1 I. Introduction  !

i II. Mass & Energy Release Modeling i

III. Containment Modeling

.; IV. Action Plan 1

V. Appendix i

VI. References 1

4 i

i i

t e

. ~

2 1

4 .

i e

I e

l.

l l .

!. _., . - _ . , . . _ . .- , , _ , , - ,. - i ,, r , e s . , ,r . , ~ . , ,. ~ .-+- -, ,

e 4

I. Introduction 1

! During the Containment Systems Branch review of dne Westinghouse topical report, '

! " Mass and Energy Releases Following a Steam Line Rupture",WCAP-8822 i (Proprietary) the Staff noted that heat transfer to steam from the uncovered portion of the steam generator tube bundle was unaccounted for and questioned '

the effect upon the calculated mass / energy release and the subsequent effect en the centainment temperature response. Westinghouse responded in a letter to the 4

Staff (NS-EPR-2563, February 14, 1982, E.P. Rahe to J. R. Miller) that it had determined the impact of the effect by conservatively treating the maximun amount of superheat to be the difference between the primary coolant temperature and the steam temperature. The letter noted that there would be an 4

insignificant effect en dry type containments and that, based on the i conservative model used, there would be an expected increase in containment

temperature for ice condenser type containments. In the Centainment Systems Branch Safety Evaluation Reports on the topical report and the Catawba Plant Safety Evaluation Report, the Staff required that a more refined steam line

{

break analysis be performed to determine the effect on containment temperature

which might impact the environmental qualification envelope used for safety j related equipment.

Since that time, Westinghouse has investigated the effects of tube bundle heat' transfer from the viewpoint of a more refined modeling approach. Subject to the final review and approval of the NRC Staff, the efforts and results obtained to

date indicate that there is little Dnpact on the containment response from the effects of the additional tube bun,dle heat transfer to stecm.

l ~

I 4

?

e 4

I I

F L -

II. Mass and Energy Release Modeling A. LOFTRAN Computer Code Mass / energy releases are calculated using the LOFTRAN code. LOFTRAN is a FORTRAN language, digital computer code, developed to simulate transienc behavior in a multi-loop pressurized water reactor system. The program simulates neutron kinetics, thermal hydraulic conditions, pressurizer, steam generators, reactor coolant pumps, and control and protection systems. Up to four independent loops may be modeled. LOFTRAN is used for analysis of non-LOCA transients and is documented in Reference 3.

The model of importance to blowdown calculations is the steam generator model. The primary side contains multiple nodes to model the tube bundle.

The standard LOFTRAN steam generator secondary side model, (Figure 1), is effectively a one node, two region model of saturated steam and water.

Heat transfer is assumed to occur only to saturated water. If tube uncovery occurs the amount of surface area available for heat transfer is accordingly reduced. The LOFTRAN code incorporates a more detailed steam generator model which is used to predict tube bundle uncovery.

1 B. LOFTRAN Hodel fx Superheated Steam The LOFTRAN code has been modified to account for heat transfer to steam from the uncovered tube bundl,e region. (Figure 2). In the modified version of LOFTRAN, all heat transfer occuring in the uncovered region is assumed to add superheat to the steam exiting the steam generator. The primary side temperature in the uncovered tube region is conservatively assumed to remain constant through the nodes which are uncovered. In reality, there will be a drop in temperature due to heat removal to the secondary side, but this is expected to be small due to the low specific heat capacity of the steam and due the high primary side flew rate.

The heat transfer coefficient used in the uncovered tube region is discussed in the Appendix. This correlation bases the heat transfer on the difference between the tube wall surface temperature and the bulk steam temperature in the region. In the LOFTRAM modification, the conservative assumption is made that no credit is taken for either a primary film heat transfer resistance or a tube metal heat transfer resistance. Therefore, the wall surface temperature of the tube is assumed equal to the primary fluid temperature.

The modified versien of LOFTRAN autcmatically determines the proper number

'of steam generator nodes for the superheat region of steam in the generator. The variable node capability is applied to both the primary and-secondary side. At each time step during the tube uncovery, the modified LOFTRAN code makes a general evaluation of the uncovered tube region (e.g.

steam flow rate, uncovered tube heat transfer area, estimated heat transfer coefficient, etc.) and determines the number of nodes to be used in the l subsequent calculations. The total heat transfer for.the uncovered tube j region is determined and accounted for in the primary temperature transient.

I L

+ ,

calculation. The superheat/ tube uncovery modeling is applicable to all l steam generators.

Figures 3 through 6 show typical results for a 0.86 ft steamline break from 102 percent power using the modified version of LOFTRAN. Figure 3 shows the fraction of tube uncovery versus time with uncovery of Loop 1 (faulted) starting at 152 seconds into the transient. At approximately 300 seconds, the uncovery transient reaches an equilibrim point where the steam flow out of the steam generator matches the auxiliary feedwater flow into the steam generater. Additionally, the tube uncovery transient for Loop 2 (non faulted) is plotted but shows no tube uncovery for the entire transient. Figure 4 presents the steam flow transient for this case.

Figure 5 includes plots of both the superheated steam enthalpy and the saturation enthalpy for the Loop 1 steam generator. Figure 6 includes the Loop 1 temperatures for the steam generator tube inlet (primary side),

steam exit temperature (superheated steam), and the saturation temperature fcr the steam pressure.

C. NOTRUMP Model Comparison The NOTRUMP computer code (Reference 4) was used to verify the LOFTRAN modeling of superheat. The computer code was originally developed to analyze transients of secondary systems with two-phase conditions. In the past, it has been used to analyze various transients in the primary and secondary coolant systems. .NOTRUMP has recently undergone major revisions to enable it to medel ncn-equilibrim nodes (i.e., separate liquid temperature and steam tempera'ture modeling). Using NOTRUMP, the steam generator can be broken down into sufficient nodes to model the nonequilibrim effects of the steam generator, as well as the tube region during uncovery. NOTRUMP can model all modes of heat transfer associated with a steamline break transient, including heat transfer from the uncovered tubes to the superheated steam and the feedback effects between the primary and secondary sides. The two phase mixture level calculation accounts fer primary to secondary heat transfer and the swell associated -

with rapid depressurization of the steam generator during the blowdown.

A comparison of LOFTRAN and NOTRUMP blowdown results-is presented in Figures 7 and 8. The mass releases shown in Figure 8 show excellent agreement. The LOFTRAN prediction of superheat enthalpy is slightly higher than NOTRUMP, ~while the predicted time of tube uncovery is somewhat later.

NOTRUMP shcus a chugging effect during the uncovery phase.of the blewdown.

This is believed to be in part due to oscillations in the flow link between

~

the dcwncemer regien and the steam dome region. (The flow link is .the drain path fcr the moisture separatcrs to the downcomer region.) With the flow direction tcwards the downcomer, superheated steam goes into the downcemer region and is condensed. This alternates with a flashing of a portion of the water volume in the downcomer region. This raises the pressure of the downcomer, resulting in a flow reversal in the link with saturated steam from the downcomer mixing with the superheated steam in the

-deme. This mixing results in the variations in the superheat enthalpy seen in Figure 7. Although LOFTRAN does not show the enthalpy variation since-the detailed modeling of the downcomer and dome are- not included, the overall agreement with NOTRUMP is very good.

D. Effects Of Analysis Assunptions The effects of superheated steam are dependent upon the occurrence and extent. of tube uncovery. The major parameters affecting tube uncovery are:

initial steam generator inventory, auxiliary feedwater flowrate, assumed feedwater system failures, and protecticn system errors. Variatiens in these parameters are in the process of being evaluated fer their effects on the containment temperature response (Figure 9).

Refinements in the mass and energy release modeling (Figure 10), are being evaluated and several areas show a potential for reducin6 the degree of superheat being generated. Some of these areas are:

- Evaluation of liquid-steam interactions such as the phenomenon of tube support plate flooding and heat transfer across the tube wrapper from

. the superheated steam to the auxiliary feedwater flcwing down outside the tube wrapper. .

- A mere detailed steam header model in LOFTRAN.

- Modeling temperature drops in the primary superheat nodes.

e Evaluating other void correlations for use in predicting tube uncovery.

e 1

i 1

1

o

III. Containment Mode, ling '

j A. Description of Containment i

j The general phenomena taking place inside an ice condenser containment during a i steamline break transient can be described utilizing a typical ice condenser j elevation drawing (Figure 11). Steam is discharged to the main (or lower) compartment where heat is removed by the internal structures, steam flow to the

, ice condenser, and the ice condenser drain water. The dead ended compartments are the regicns which are located below the ice condenser and outside the ' crane

, wall (Figure 12). Air is discharged from the main compartment to the dead ended compartment and ice condenser so that the resulting steam to air ratio is.that region is much higher.than in dry containments. At ten minutes following the centainment hi-2 signal, deck fans are actuated which direct air flow from the upper compartment to the dead-ended compartments. Most of the safety related equipment is located in the dead-ended compartments although some equipment and cabling are located in the main compartment.

B. Containment Models 1 Figure 13 outlines the major models and assumptions utilized in the LOTIC-3 '

containment code. In the currently approved version of LOTIC-3 documented in Reference 5, four distinct regions of the containment are modeled; the lower compartment, the dead-ended compartment, the ice condenser, and the upper compartment. Two condensate /revaporization models are used depending on the size ,

of the break. For large steamline breaks, 100% condensate revaporization is.

assumed. For small steamline breaks, a convective heat flux model is used which calculates partial revaporization during the transient. 'lhe wall heat transfer c

medel utilizes the Tagami heat transfer correlation for condensation heat transfer and the convective heat flux model derived from the work of Sparrow

! (Reference 6) which calculates'the convective heat transfer for small steamline breaks. The sump recirculation system is only modeled for the large break LOCA transient containment response.

Figure 14 shows the four regions modeled with'the ' mass and energy ficws that can be assumed in the analysis. The Catawba nuclear plant does not have lower compartment sprays and they are not modeled in the analysis. Superheat heat transfer is conservatively assumed to be zero for the steamline break containment analysis. In the model described in Reference 5, wall heat transfer _.

is not modeled in the dead-ended compartments although these regions do contain i structures which will remove heat. The analysis does include the upper. .

1' compartment sprays, flow through the ice condenser, deck fan flow, and flow-to the dead-ended compartments. -

LOTIC-3 solves the conservation of mass, energy, and momentum' equations.for.

upper, lcwer, and. ice condensor regions (Figure 15). After the new lower _

compartment conditions are determined, conservation' equations are solved for the dead ended compartment' and the flow rate between the compartments is determined..

Figure 16 presents a typical steamline break containment temperature transient 1 that is calculated using superheated steam blowdowns from the LOFTRAN code and.

the modeling of ice condenser drains as a. heat' removal source. . The-transient-shows that initially the containment temperature increases rapidly during the.

__ . _ , _ _ . _. _. _ _ , _ - , ,~ , _ ,u _ _, ,,,

blowdown. When the upper compartment sprays actuate there is a slight decrease in the main ecmpartment tempera;ure. The temperature then rises slowly until i ice condenser drain flew decreases to the point at which time the temperature begins to rise again (approximately 250 seconds). This rise in containment temperature coincides with the steam generator tubes uncovering at 152 seconds and the maximum superheat occurring at approximately 250 seconds. The steam generator level stablizes whe'n the auxiliary feed'.aater flow is equal to the steam discharge at approximately 300 seconds. The containment temperature then starts decreasing with decreasing decay heat. At ten minutes, the deck fans

_ actuate which results in a rapid decrease in containment temperature.

C. LOTIc-3 code Modifications Four modificatiens have been incorporated in the LOTIC-3 containment model which are (Figure 17);

1) wall heat transfer model

2) convective heat flux model

3) ice condenser drain model

4) dead-ended compartment model D. Wall Heat Transfer - - - - -

The modificaticn to the wall heat transfer model is described in Figure 18. In the LOTIC-3 model, only condensation heat transfer, utilizing a Tagami heat transfer coefficient and a temperature difference between the wall and saturatien, was previously modeled.' The modification includes a convection term '

with a conservative convection heat transfer coefficient and a temperature difference between the containment atmosphere and an appropriate interface temperature. Tne Appendix presents a more detailed description of this model.

E. Convective Heat Flux The modification to the convective heat flux model is described in Figure 19. A term has been added to the convective heat flux model to account for the feedback effect from including a convectivo term in the wall heat transfer model. The Appendix presents a more detailed description of this model.

F. Ice Condenser Drain Model In an ice condenser containment there is approximately twenty drains exiting frcm the ice condenser into the lower ccmpartment at an elevation of about forty feet above the compartment floor. The drain pipes are one foot in diameter.

The drain flowrate is calculated by the LOTIC-3 containment code. For _ a typical small steamline break transient the drain flowrate varies' from approxiraately 4000lbm/seeto500ltm/secduringthegimeframeofinterest. Tne temperature of the drain water is approximately 130 F (Figure 20).

Figure 21 presents the assumptions and the basic model used to estimate the heat removal ft om the lower compartment atmosphere to the ice condenser drain water.

It is conservatively assumed that the drain water stream does not_ break up prior to reaching the floor even though many of the drains have equiptent and structures located belcw them. Therefore, heat transfer-is assumed to occur at

. .- e the stream surface only. It is also assumed that the stream surface temperature is at the saturation tenperature of the containment.

The heat transfer to the stream is:

q=hA/LT where h = condensation heat transfer coefficient A = surface area of the stream AT = appropriate temperature difference The calculation of the heat transfer surface area is described in Figure 22.

In order to model the drains in LOTIC-3, the drains are modeled as a wall heat sink with a surface at a constant temperature (see Figure 23). Currgntly,in the version of LOTIC-3, the surface temperature is assumed to be 230 F which is close to the containment saturation temperature. The drain surface area is calculated at two points in time during the transient; early in time with a high flowrate and later in ttne with a low flowrate. To ensure conservatism in the area calculation a 10% reduction of the surface area was assumed.

As described previously (Figures 14 & 15), the LOTIC-3 containment model did not account for wall heat removal in the dead-ended compartments. To obtain a 1 conservative estimate of the tenperature transient in the dead ended ccmpartment, the heat sinks located in the dead ended compartment regi'on along.

with the heat sinks in the lower compartment are modeled in a combined volum.e (see Figure 24). This " modified" lower compartment model is used to determine a conservative dead-ended compartment tenperature transient.. Since the Icwer compartmer.t will be notter than th'e dead-ended compartment, this methodology results in a higher temperature in the dead-ended compartment then kould be expected.

G. Transient Results With the modifications described for LOFTRAN and LOTIC-3, the previous FSAR limiting case for Ca*awba was reanalyzed to determine the impact of superheated steam. The case selected is a 0.86 square foot break at 10j5 power (Figure 25).

The peak lower containment temperaturc for this case is 324 F. This temperature is calculated for the lower compartment only. It is expected that the -

dead-ended compartment temperature will be significantly lower.

In addition to the model modifications incorporated in LOTIC-3, Westinghouse is pursuing further improvements in the areas noted on Figure 26. One area is in the wall heat and mass transfer models. Since condensation is a mass transfer

, - type phenomena, the heat and mass transfer should be linked. This' approach has been used in Reference 7.

An bnproved drain model is also being investigated. This improved model will .

calculate the drain surface area as a function of flowrate. - It will'also calculate the average temperature rise of the drainwater. This model will more accurately represent the actual phenomena in the containment.

t .

(

e - - - +

WESTINGHOUSE PROPRIETARY CLASS II  !

V. Appendix 4

WESTINGHOUSE STEAMLINE BREAK ELOWDOWN AND CONTAINMENT ANALYSIS METHODOLOGY The following sections describe the Westinghouse methodology for determining the containment response fer a steamline break incorporating the effects of superheated steam. These sections describe in detat.1 changes from the methodologies described in References 1 and 5.

I. Steamline Rupture Mass / Energy Blowdown Analysis A. LOFTRAN and MARVEL Computer Modeling Mass / energy releases can be calculated using either the LOFTRAN code (Reference 3) or the MARVEL code (Reference 8). The LOFTRAN code is used for ncn-LOCA FSAR accident analyses. The MARVEL code was specifically developed for assymmetric transients such as steamline breaks. These two codes are very similar because they were developed in an interrelating l fashion and much of the modelIing is common to -both codes. The MARVEL code j was used in the development. of Reference 1 because LOFTRAN at that time was a lumped model which was used for symmetric loop transients. -Furthermore, for steamline break analysis purposes, MARVEL' contains a model fer water entrainment. However, the current version of. LOFTRAN is a multiloop version which also contains a water entrainment model. With the development of a multiloop version of LOFTRAN and the inclusion of an i entrainment model, the use of MARVEL has been generally discontinued. This enables the use of LOFTRAN as a single system analysis code for non-LOCA transient analyses. LOFTRAN.is used in the analyses presented here.

The model of importance to blowdown calculations is the steam generator -

model. The primary side _ of the ' steam generator contains multiple nodes to model the tube bundle fer both the modified version of LOFTRAN and MARVEL.

Heat transfer calculations from the primary to secondary side are identical in the.two codes, although the methods for initializing the heat transfer .

resistances are slightly different. The secondary side .is effectively a 4 one node, two region model of saturated steam and water.- Heat transfer is assumed to occur to saturated water. . If tube uncovery is predicted, the z

amount of surface area available for heat transfer is reduced. -

Both codes contain a-detailed steam generator'model which is usea to predict tube uncovery. This model calculates 'the. liquid volume in.dne steam generator shell and acgognts for the detailed steam generator geometry. The [ l correlation is used in both codes to.

predict the voiding. in the tube region, although the correlation is i- - . modified for use in-LOFTRAN. In MARVEL, tube .uncovery is calculated based i

• e i ytr- , y q 1 gy- - ~ -t- ,*-

t-+ y - vg g e ~

g

. - = - . _ . -_ .- _ .- _ ..

t- .o s i I

on comparisen with the actual water level and the height of the tube bundle. In LOFTRAN, the user specifies either a water volume in the steam generator corresponding to tube uncovery, or a void fraction in tne riser section of the steam generator st which tube uncovery begins.

Both ccdes have similar models accounting for reverse heat transfer, thick metal heat transfer, feedline flashing, and safety injection system operation. Auxiliary feedwater ficw can be input as .a fraction of nominal feedwater flow, although LOFTRAN has an additional _ capability to model auxiliary feedwater flow as a separate system. For analysis of double

! ended ruptures, MARVEL accounts for the volume of steam in the piping downstream of the steam generators in the blowdown calculations. In LCFTRAN, this consideration is added on to the blcwdown mass and energy results by hand. For split ruptures, which the analysis presented here

addresses, the steam piping masses are handled identically in both codes.

In sumary, LOFTRAN and MARVEL are very similar codes, and either can be

, used to calculate mass / energy blowdowns. To demonstrate this, a comparison

of the blowdowns fer a typical case is presented in Figures A.1 and A.2.

l Figure 1 presents the mass release rate for a .86 ft2 split rupture from l 102% power. For this case, Figure A.2 shows the saturated steam .enthalpy as a function of tDne. This blowdown is typical of results used in FSAR '

analyses prior to the modification noted in this report for the LOFTRAN code. As can be seen from the figures, the results are extremely close..

B. LOFIRAN Model for .Superheated Stem As mentioned previously, the LOFTRAN code has been modified to model heat

transfer which may occur in the uncovered tube bundle region. This effect is modeled in both the faulted and intact loops. In the modified version of LOFTRAN, all heat transfer occurring in the uncovered region is assumed to add superheat the steam exiting the steam generator. The tenperature of

, the primary coolant flowing through in the uncovered tube region mode is conservatively assumed to remain constant. Realistically there would. be a drop in temperature due to heat removal to the secondary side, but this will be small due to the low specific heat capacity of the steam and due the high primary side flow rate.

The heat transfer coeffic)3gt uged in the uncovered tube region is' based on the [ *

]g' . The heat transfer coefficient (U)-is calculated by *the following expression: _

a,c

-t a

4 l

This correlation is presently used f 5'superheated. forced convection heat-transfer by.the [ ] ' computer codes. Additionally,-

. _ _ _ _ - _ _ _ _ _ _ _ _ _ - - - - ._-- _ _ _ _ . _-__=_____________--__.__-__:___=_____ - - _ = _ _____ =

i this correlation is based upon the heat transfer from the surface of the tube wall to the average bulk temperature of the steam. In the LOFTRAN modification, no credit is taken for either a primary film heat transfer resistance er a tube metal heat transfer resistance. Therefore,the wall temperature of the tube is conservatively assumed equal to the primary fluid temperature.

'l (1) l i

'e 0

I e

?

, - . 9- - . c - - _. p ,

,e .,

Tne modified version of LOFTRAN automatically selects the proper number of 1 steam generator nodes for the superheat region of steam in the generatcr.  !

l The variable node capability is applied to both the primary and secondary side. At each time step during the tube uncovery, the modified LOFTRAN code makes a general evaluation of the uncovered tube region (e.g. steam flow rate, uncovered tube heat transfer area, estimated heat transfer coefficient, etc.) and determines the number of nodes to be used in the subsequent calculations. Each node is evaluated to determine the steam temperature exiting the node with a convergence criteria that is based upon the total number of nodes used. The exit steam temperature of one node is used as the inlet steam temperature of the next node.

The heat transfer calculation to determine the outlet temperature of the node is based upon the following expression:

Q = UA*(Tpri-(Tout + Tin 2) * "s' s Tout-Tin}

where Q = Heat transfer to the steem S'

U= J 3

Tpd = Primary node temperature out = Steam node outlet temperature T in = Steam nod'e inlet temperature-M3 = Mass flowrate of the steam, Cs = Heat capacity of the steam A = Heat transfer area in the node including both hot and cold leg sides of the tube bundle The total heat transfer for the uncovered tube region is determined and accounted for in the primary temperature transient.

C. Blowdown Sensitivity to Plant Conditions The effects of superheated steam are dependent upon the occurrance and extent of tube bundle uncovery. Parameters affecting tube uncovery are:

initial steam generator inventcry, break size, auxiliary feedwater ficwrate, and the single failure assumed.

The initial steam generater inventory depends upon the measurement errors associated with steam generator level and upon initial power level. Steam generator mass increases with decreasing power, thus, breaks intitiating from low power levels will result in later tube uncovery. -

Larger break sizes result in faster blowdown of the steam generator and -I earlier tube uncovery.

L i

, 1 i .

Large auxiliary feedwater flowrates only delay tube uncovery, but will also cause the final equilibrium steam generator level to be higher. This equilibrium conditicn corresponds to the point when the break flow rate is equal to the auxiliary feedwater ficw rate.

The single failure assumed in the transient may impact the amount of water supplied to the steam generator. Auxiliary feedwater runcut will increase the amount of water supplied to the steam generator. Failure of the feedwater isolation valve will also cause extra water to be supplied to the generator as the additional mass between the isolation valve and the check valve flashes to the generator.

am

• m P

• f

. 4 .

II. Containment Analysis ,

A. Wall Heat Transfer Model d

The original LOTIC-3 wall heat transfer model is based on the stagnant Tagami heat transfer correlation. That is,

• q"=hTAGAMI(TSAT-TWALL) 2

1. N h(TAGAMI, MAX)=72 BTU /hr-ft F o hTAGAMI = 2 + 50 MSTEXi AIR This model was developed for saturated steam in the presence of large amounts of non-condensable gases. In the lower compartment of an ice condenser, most of the air is swept out of the lower compartment through the ice condenser and into the upper compartment. Therefore, after about 30 seconds, there-is almost no non-condensables in the lower compartment. Typical values for the condensation of pure steam are in the range of 1000 to 3000 Btu /hr-ft2 aF (Ref. 5). The correlation used in the modified LOTIC-3 code is in extension of the Tagami correlation fer nearly pure steam, t

q"=hCOND (TSAT-TWALL)

/M hcond = 2+50 MSTUM AIR h(cond, max)

• A ma::imum value of [ Ja,c was chosen as a conservatively lcw condensing heat transfer coefficient in a nearly pure steam environment.

In addition to this modific~ation, an additional term is needed to account for the convective heat transfer from the superheated steam to the condensate film.

This convective heat transfer is dependent upon whether there is condensation occurring on the walls. If condensation is occurring, the correlation used is:

conv =hconv(Tbulk-Tsat where:

If the wall temperature increases to above the saturation temperatu're then the convective currents will be reduced such that the correlation-used is 9"conv=hconv(Tbulk-T y ,11)'

where:

[

j a,c ,

2

' _ ^ -

Th'us in summary, if Twa11<Tsat **"

[ j a,c If Twall > Tsat, then the correlation used is:

[ j a,c S

I e

B.- Convective Heat Flux Model When the containment atmosphere is superheated, the containment temperature is a strong function of the anount of steam mass in the atmosphere. Thus the amount of mass condensed en the heat sink surfaces is a key parameter. The actual amcunt of condensate fermed is -

N h eond

• 9cond f8 Unfortunately, with the use of a heat transfer correlation based only on test data (such as Tagami or Uchida), only the total heat transfer coefficient is obtained. This total heat transfer coefficient includes both the condensation heat transfer and the convective heat transfer. Based on the work of Sparrow (Reference 6), the Westinghouse Convective Heat Flux model in the original LOTIC-3 code calculates the ratio of the convective heat transfer to the condensation heat transfer. Therefere the calculation of the amount of mass condensed is

[ j a,c In the modified LOTIC-3 model, the, amount of superheat convection is calculated.

The amount of convective heat transfer at saturation is not known explicitly in this model. Therefere, in the modified LOTIC-3 code the original convective heat flux model will be used to calculate the fraction of convective heat transfer for saturated conditions. The actual correlation is

. s3c where, (q /q is determined frcm original convective heat flux model and q conv is eggg)38 bunt of convective heat transfer colculated in the wall heatEFEXSfbrmodel In summary, the modified LOTIC-3 model is consistent with the original LOTIC-3 model in its calculation of the mas condensed. The only difference is that in the modified LOTIC-3 code, the amount of superheat convective heat transfer is known explicitly, while in the original LOTIC-III model, only the ratio of convective heat transfer to condensation heat transfer is known.

4

,o* .,

IV.

References:

-1.. Land, R. E., " Mass and Energy Releases Following A Steam Line Rupture" WCAP-8822 (Proprietary) September, 1976 and WCAP-8859 (Non-Proprietary).

2. NS-EPR-2563, February 14, 1982, E. P. Rahe of Westinghouse to J. R. Miller, NRC, " Additional Information on WCAP-8822".

3. Burnett, T. W. T., et al. , "LOFTRAN Code Description," WCAP-7907, June, 1972 (Proprietary).

4. Meyer, P. E., and Kornfilt, J., "NOTRUMP - A Nodal Transfer Small Break and General Network Code," November,1982, WCAP-10079 (Proprietary) and WCAP-10080 (Non-Proprietary).

5. Hsieh, T. and Liparulo, N. J., " Westinghouse Long Term Ice condenser Containment Code - LOTIC-3 Code," February, 1979, WCAP-8354-P-A Sup. 2 (Proprietary),WCAP-8355-NP-A (Non-Proprietary).

6. Sparrow, E. M., Minkowycz, W. J., and Saddy, M., " Forced Convection Condensation in the Presence of Noncondensables and Interfacial Resistar.ce", Int. J. Heat Mass Transfer, Volume 10, 1967

7. Corradini, M. L., " Turbulent Condensation on a Cold Wall in the Presence of' a Non-condensable Gas" Nuclear Technology Vol. 64, pp 186 - 195, February, 1984. ,

8. Krise, R. and Miranda, S., " MARVEL - A Digital Computer Code for Transient Analysis of a Multiloop PWR System," Novenber,1977, WCAP-8843 (Proprietary) and WCAP-8844 (Non-Proprietary).

. 9. McCabe, W. L., an'd Smith, J. C., " Unit Operatens of Chemical ^ Engineering",

3rd Edition, 1976.

~

e s

s s.-

. 1 LOFTRAN - MARVEL COMPARIS0N

.860 FT2 BREAK AT 102 PC POWER 2000.0  :  :  :  :

f

/

1750.0 h

d e

1500.0 - -

s S

~

1250.0 - -

2 S --

"1 1000.00 - -

w --

M 750.00 - -

x 5

c= 500.00 - -

MARVEL -9 t' LOFTRAN --

250.00 - - .

- 12/05/83 0.0 6 6 6 E o 5 -o o

O o a o o d d & 'd & d o

O o a o o

- cu en .e n w a.

TIME (SEC)

.O 1

FIGURE A.1

_____-_a_-_______-_______________--__---__--___-____ - _ _ - _ - _ . - _ _ _ _ _ _ - _ - __. - . _ - _ _

. ._ _ - . . _. = .. . . - - -  :. ..

9

.- t. ,

LOFTRAN - MARVEL COMPARIS0N

.860 FT2 BREAK AT 10'2 PC POWER .

1300.0  :  :  :  :  :

- 1275.0 -- --

5 s

2 a 1250.0 - - --

1225.0 - -

0FTRAN 1200.0 _

5 1175.0 - -

1 i' 1150.0 -- --

m S 1125.0 -- --

1100.O i i i i i c O O O O O O.-

O b b O

b O

b - O O O

. O O, O - cu m n e-TIME (SEC)

FIGURE A.2

'1 l

l

c b >

METHODCLOGY FOR ADDRESSING SUPERHEATED STEAM RELEASES LC ICE CONDENSER OCNTAINFENTS Purpose ,

The purpose of dais report is to document the information presented on March 19, 1984 in a meeting with the U.S. NRC Containment Systems Branch 'on the status of progress made in addressing the confirmatory item on the Catawba Nuclear Plant Safety Evaluation Report. This confirmatory item deals with the effects of superheated steam generator mass and energy releases following main steamline break accidents. Attachment 1 includes the list of attendees at the meeting and the overhead slides covered in the Westinghouse presentations.

Technical presentations were made describing the modeling of the steam generator and heat transfer from the uncovered tube bundle during the steam generator blowdown along with a description of the containment model and transient response. A proposed plan of action was also presented and discussed with the

, Staff. In accordance with that plan, this report represents the first milestone in the proposed plan of action. As committed to in the meeting, the appendices present proprietary infonnation which relates .to the specifics of the models and sensitivities that were not directly addressed in the meeting.

Attachment 2 is an explanation of, and refers to, the overhead slides (Figures)

p. resented at the Fbrch 19 meeting. -

O m

O

- ~ . y

f 1

N 0

ATTACHEllI 1 m*

n i

.1

(

a .

/]Hudoc 01- 3/'7/84 .

'11ans <-V A. L a u , .

is .E n < . L. ~ 7~ u A ~,9.i.s ds-g . cm - . 3 - xa ,

- y u K', Al. *TA&TavR.

AIKc /DL Dage./wpp

- [T. o . 3/.Mt2P r=.- s+w

a. 1. L, ri a @ swas 4.\/. To t.g \C Q Goc. E AR S AFf T7

~.Y [ )(e u lh"C. //L b $. Lo tte W N uCLL-4R. $ hF0TY 6fooc 4 W u u c.l.<.6,. @

1+1F

& J,1!<rh fA. LiL. Hv c(<<> <-

L . 2. JAwhese- h' .ua/m bb C-A 05/f lsd ft.bclu r- Se hiy T- 7 Iwoc ac( Q mod , fpy m}('. .

Ak4,J v Du.I< e. , De :: Em --~.--

S .D S. 5 Mo-llv bde.,><r:j E y:w:

7t/st, N.o sa.am L ea.,.s y Garr

~~T. 2. Otsus

$6ACtl LtCEd. sin c. .5'rnff , ECis':

E. AMLLt C.'b(q'wc.%a q a - <.amt. sa.y. .

T V A, Gr.3: .: , ' ' bn oy W. O. Ccook . .

Qsc.,cbe So!

• S. 2. t/eh t- Q.sa. , S a k ,,- fa.L<b.

W Ns'~D .S l'!![

h k 4)Q r>r _,

ara /secun r. wary

2.M.Ow.e s J% SNAR4tcc=7E-A.:AA- c+0

" \'

v' k '% G b c %.

?'l.R,k.W- /RC /cSB C. Li N f C 4 i.t a A. Mohdouceso v$c/ CST 3 9

J,;LOFTRAM MODEL i

l STEAM GENERATOR SEcov1DARY ,

STEAM FLOW .

s 4

p SATORATED i

. .. STEAu gastw

. '.. .i

...  : n ssruex,su

<.. Y,,d_ ym passg f/ . . - -

Ac L

s

. .?*,

\ 't 4 " 2.

g. _ 's s + l A

• StMGI E McDG , 2 gtEGacd MODEl

• MC HEAT TRAdSFER TO SATueATED STEAM.. REGlou

• HEAT TRAMSFER Tc SATORATED  !

WATER RE.Gted iS MODtF1ED Rll4il TVSE 'UNCcN ERN '

l FIGURE 1 .

e

..___________._____..,__________________j . ,

' ~ ~

LOFTWAfT MobEE s .

SUPERHEAT HEAT TRANSFTR .

SUPERHEATEb STEAM FL.CW '

SEccMDAR.T

{ .- - n .as

-p>

\ ,f ,sov,ensaAT poets q l h ,

'9 M6 e :As ~

' ~ a:.rt-t=CT W E ' '

~ UN COVERY LEVEL ,

, p.

1- .

i _

I . .

A A v im Suom .

HODES MODEL l

k coHSTART PRtm TEMPERATuss l m. suPan.asAT p.ssieu AssupEd Fop.

64E.AT T9.AHSFa.S cALcuLATtous

• C ALCO LA%D SOPERWE AT HEAT asna Accovara ven. in

, pst.sa As.Y T RAn ws.HT ~

cwe2_ ..

7 ..

,,,--_,y--.--r- - . . 2 _- .. . -.;;-

g, TUBE UNC0VERY LOFTRAN SUPERHEAT MODEL

.860 FT2 BREAK AT 102'PC POWER 1.2000  :  :  :  : l G

E 1.0000 -- LOOP 1

~~

z S

s y .80000 -- --

E a .60000 - - --

o E

w

> .40000 -- --

O Y

2

$ .20000 - - --

S LOOP 2 0.0 8 8 8 8 o e s s s s s s " N O t TIME (SEC) 9 FIGURE 3

t MASS BLOWDOWN LOFTRAN SUPERHEAT MODEL

.860 FT2 BREAK AT 102 PC POWER 2000.0  :  :  :  :

f --

1750.0 --

i n

d 1500.0 --

R S

1250.0 -- --

2 3

"I 1000.00 - -

w G 750.00 - -

M E

m 500.00 - -

250.00 -- ---

0.0 i i i i i o e o o -

o o o o b b b b b

. O O O o o O o .-e tu (r) -r LO CO TIME (SEC)

FIGURE 4 l

a l l

ENERGY RELEASE LOFTRAN SUPERHEAT MODEL

.860 FT2 BREAK AT 102 P,C POWER t

1300.0 '

- 1275.0 - - --

3 SUPERHEAT.:.D STEAM y -w-g 1250.0 - - --

g 1225.0 - - --

T w

~

1200.0 - - --

- SATURATION s

5 1175.0 - - --

=

id', 1150.0 - - --

v S 1125.0 -- --

1100.0 8 8 8 8 8 8 o e 6 s s 6 s s S 8 8 8 8 TIME (SEC)

FIGURE 5 l

l

4 2 .

TEMPERATURE TRANSIENTS LOFTRAN SUPERHEAT MODEL

. 860 FT2 BREAK. AT 102 PC POWER 650.00  :  :  :  :

600.00 --

E ,

EI tE 550.00 - --

c.

8.

I i u SG TUBE INLET (PRIMARY SIDE) e 500.00 - - --

Gg A e r - .

< i-

• w 450.00 -- --

5$ SUPERHEATED STEAM m -

O.

h 400.00 -- --

b 350.00 -- --

SATURATION 300.00 i

8 8 8 8 8 8 o d d d d d 5 l

d '3 8 8 8 8 l .

l TIME (SEC) i l

FIGURE 6 i

r k8e

- :g

^

~

O.8n

,~

' 8, C

E S

(

- E M

I T

N 0 =

S I

p 1 8m R A A R P

M T O F C O L

P YP l

(

M L U A RH TT N /

O E 8

.U ' - h

.' #. 6 90.d8- l 'f"' *,.. \*REA CTOR p-p'. .

o

,s CbOLA NT

\w g 4 .g --  :

- i s ,

.s pyg,.pj ,,,, -

4- 4

..h..

.g.

% .* ..)

. , ' %L OWER\

i. Q. ..

e*

/ .

INTERNALS -/

.i RAC #

. .*  ?.:' '

Cy . ':, ' .. ,l :o." .g

= .q

.D

~ , .

~*

n A ..

.h ~y a- y.eam

. GA TE* .b.*' y,7 ~

, STEAM GEN (4)

. s MAIN ,'a~fJ I

_ f fgf3y g y STEAM ' , , ((

G, ;,

- f '

0 h ,[ .' ~ 7

• FEED-c>.

.'. 's': 7C2.15 g 4 WA TER.

Lv EL 679.78

' ' -VENTIL A TION RELIEF .

/

..?. . [ ..

\s*

TANK ,

. r'  % . ' UNITS (4)

/,. ..

7 ,'. PRESSURIZER -

. r. -

'e- , . ,. w 5 : .e - .

l ACCUMULATORS (4)

'O.***' GEAL

' .f Q..N ~

TABLE t

^

• • .p *

......ab f EL gg3.0 HEAT EXCHANGERS' - ' ) (6 . s ,. 4  ;

VENTILATION UNITS '(2) w... et ro0.og ' .i-1

', . * , PERSONNEL LOCK Figure 3.8.3-7 Plan-Lo.er Com.partrent

%=

-d FIGURE 12

J ,

! LOTIC-3 CONTAINMENT CODE

• 4 N0DE CONTAINMENT MODEL i

i

• CONDENSATE /REVAPORIZATION MODELS LARGE BREAK (TOTAL REVAPORIZATION)

SMALL BREAK (CONVECTIVE HEAT FLUX)

• WALL HEAT TRANSFER MODEL

• MODELS SUMP RECIRCULATION SYSTEM i

k i

i e

l l

l. -

o t.

FIGURE 13

-- - ,. - - , _..m.,y.$ , . -y-, . _. , ., ,,7._ ,,,,. , ,,__ ,.

s r

I t

t, t

i i

I 4 A Sorays [

A A t

3 , _.3 ,

= Qwall Upper Comcarment  :

l I i

m,h i A T  !

5 s1 i U.C. -

m3 ,h34 ,

Orain

=  ! Ice *s ' 's A

ice

-  ; i Concenser m'h a a

- One way flow hall - -

i , throuch open area

,  ; on operating decx Concensate & A Ice Melt lm,h -

5 m,n 5l e I.C. Sumo a a

• I.CTi Oraini  ;

Y e

' l ^

I

"'"s ' * ** "l ow ,

I ^

- A , h- (RCS break lCead-Ended 3 o accumulacc Cemcarment i'n a a ' L:wer Cc=oarcent release , c.

r boil-of#)

~-

_ Air Addicion Taole 8

o crn,ic _l ueat Accition Table

. L.C. i

c aal'

'Orain sumo

-_.i L___u U.C. r 7 r

Drain L L.C. sume FIGURE 3.3 MASS AND CIERGY FLOiJ DL*sGR&.

FOR THE COMPAROL","r5 i ,

FIGURE 14 l

e'

.- a .

f i

LOTIC METHOD OF SOLUTION SOLVES CONSERVATION OF MASS, ENERGY, AND MOMENTUM.

FOR UPPER, LOWER, AND ICE CONDENSER REGIONS ONCE NEW LOWER COMPARTMENT CONDITIONS ARE DETERMINED, COMSERVATION EQUATIONS,ARE SOLVED FOR THE DEAD-ENDED COMPARTMENT AND FOR THE FLOW RATE BETWEEN THE TWO COMPARTMENTS 4

l'.

i l

1 FIGURE 15 l

TYPICAL CONTAINENT TEFFERATURE TRANSIENT (DRAINS MODELLED)

1. LEVEL REACHED IN S.G'.

-J U.C. SPRAY ~

[ DECK FANS ON t J REDUCTION IN 1.C. DRAIN FLOW e

b '

h ,

B..~

t 0 100 200 300 400 500 600

. TIfE (SEC. )'

e FIGURE 16 . .

L

- o.

1 I

l l

I i.

FDDIFICAfIONS TO THE CONTAINtEE F0 DEL.

]

WALL HEAT TRANSFER MODEL C0tNECTIVE HEAT FLUX MODEL ICE CONDENSER DRAIN MODEL DEAD ENDED COMDARTIENT t0 DEL ~

J l

t l- .

FIGURE'17 , ,

l ,

, . , . . - . ,y , , . , = r ,

WALL HEAT TRANSFER F0 DEL ORIGINAL L0flC motel

=h kgami (TSAT ~T'dsII]

ftDIFIED LOTIC MODEL Cong g y- W "Ik w,jf &

=$ '7AM l cono  %,,. /

- C0n y Wolf j ggy L =fR.,,,L1 .

FIGURE 18 h i rim um i.

\

CONVECTIVE HEAT FLUX F0 DEL ORIGINdL LOTIC MODEL

}rorngl' '

kcono = Scono hig A,,

L i + x _l

. fDDIFIED LOTIC NDEL

. y = S cono bry

=

% hpgrat [! t %][g

~

cono

!<.~I%~%)~

gYOinL

~

i l

~ FIGURE 19

~

V - e

ICE CONDENSER DRAINS

-APPROXIN MTELY 20 ICE CONDNESER DRAINS

-DRAIN ELEVATION IS ABOUT I40 FEET FROM FLOOR-

-DRAIN PIPE IS 1 FOOT IN DINETER

-FOR TYPICAL MSLB TRANSIENT, DRAIN FLOW VARIES FROM 14000 LB/S TO 500 LB/S I

d l

l l

q FIGURE 20 .

j n . .a -

,vn ,. ,,, p-- ~ - , , e , - y

ICE CONDENSER DRAIN FDDEL

-CONDENSATION OCCURS AT THE SURFACE'0F THE STREAM

-FLOW IS WELL MIXED Q = h A AT 4DDEL AS A WALL AT A CONSTANT TEMPERATWE

-A IS THE SURFACE AREA 0F THE STREAM i

-h IS A CONDENSING T/PE HEAT TRANSFER COEFFICIENT

)

i j .

FIGURE.21

~ ,

CALCULATION OF THE STRFM4 FLOW AREA f = n (f x[] = 20 (Px YO) = 800 P WHERE F IS THE PERIMETER OF THE STREN4 1

1, s

t

/ ,/' Y s , ,

b

%: ,/ ;

i(~ . .. . ,_3

,e h

a

f %.m 5 1

i l

r

'. FIGURE 22 I

l-

. _ ~ - . . _ _ _ ._

4 l

FDDIFIED LOTIC DRAIN M.) DEL 1

i l

-WALL WITH A VARIABLE AREA i

4 i

t

= bcono b (Teutx ~TSg) .

i i

)

Bulx i

i-i l hr ~ 230 F i

1 1

i i

I .

( -

i 1

r 7 - ~

cFIGURE 23

" .c

ee ,

e DEAD ENDED C&PARTiMT f0Dti e

e u

9 I

I i

NOIFIEDLOWERC@PfTTET j i

l

. l ._.  ;

e 9

4 e i 1

i 1

+ j FIGURE 24 '

l CATAWBA RESULTS

- 102% POWER

- 0.86 FT BREAK

- MAXIMUM AFW FLOW-

- FSAR HEAT SINKS

- f%XIMLN S.G. INITIAL MASS T'%X=

f 324*F (M (DPARlfBT)

+

FIGURE'25t s _ _..

i s

's

l - ',

, l l'

(

i l ADDIT 10fML PUDEL CONSIDERATIONS f

(

i j -WALL EAT TRANSFER MODEL 4

l

-DRAIN MODEL .

4

-DEAD ENDED C0ffARTtENT MODEL 4

+

i l

I 4

i f

e F

=

.- j 1

1 II r

~-

FIGURE-26 .

. u: .s-e -

- - . .-. . - _ . - _ _ . = . __ -.__--. .- __ - . . ~. _ .- .. .. _

1 i

ATTACHENT 2 ,

OUILIE OF THE REPORT l

f a

l I. Introduction

! i 1

II. Mass & Energy Release Modeling

III. Containment Modeling IV. Action Plan V. Appendix l VI. References l

i 4

=

0

)

i i 1

e-e

.- p _tf- _ y y y e e -

i I

I. Introduction During the Containment Systems Branch review of the Westinghouse topical report,

" Mass and Energy Releases Following a Steam Line Rupture",WCAP-8822 (Proprietary) the Staff noted that heat transfer to steam from the uncovered portion of the steam generator tube bundle was unaccounted for and questioned the effect upon the calculated mass / energy release and the subsequent effect on the centainment tenperature response. Westinghouse responded in a letter to the Staff (NS-EPR-2563, February 14, 1982, E.P. Rahe to J. R. Miller) that it had determined the impact of the effect by conservatively treating the maximum amount of superheat to be the difference between the primary coolant temperature and the steam temperature. The letter noted that there would be an insignificant effect en dry type containments and that, based en the conservative model used, there would be an expected increase in containment temperature for ice condenser type containments. In the Centainment Systems Branch Safety Eveluation Reports on the topical report and the Catawba Plant Safety Evaluation Report, the Staff required that a more refined steam line break analysis be performed to determine the effect on containment temperature which might impact the environmental qualification envelcpe used for safety related equipment. .

Since that time, Westinghouse has investigated the effects of tube bundle heat-transfer from the viewpoint of a more refined modeling approach. Subject to the final review and approval of the NRC Staff, the efforts and results obtained to date indicate that there is little impact on the containment response from the effects of the additional tube bun,dle beat transfer to steam.

9 0

e O

i

II. Mass and Energy Release Modeling A. LOFIRAN Computer Code Mass / energy releases are calculated using the LOFTRAN code. LOFTRAN is a FORTRAN language, digital computer code, developed to simulate transient behavior in a multi-loop pressurized water reactor system. The program simulates neutron kinetics, thermal hydraulic conditions, pressurizer, steam generators, reactor coolant pumps, and control and protection systems. Up to fcur independent loops may be mcdeled. LOFTRAN is used for analysis of non-LOCA transients and is documented in Reference 3.

The model of importance to blowdown calculations is the steam generator model. The primary side contains multiple nodes to model the tube bundle.

The standard LOFTRAN steam generator secondary side model, (Figure 1), is effectively a one node, two region model of saturated steam and water.

Heat transfer is assumed to occur only to saturated water. If tube uncovery occurs the amount of surface area available for heat transfer is accordingly reduced. We LOFTRAN code incorporates a more detailed steam generator model which is used to predict tube bundle uncovery.

B. LOFIRAN Hodel fcr Superheated Steam h e LOFTRAN code has been modified to account for heat transfer to steam from the uncovered tube bundle region. (Figure 2). In the redified version of LOFTRAN, all heat transfer occuring in the uncovered region is assumed to add superheat to the steam exiting the steam generate . The primary side temperature in the uncovered tube region is conservatively assumed to remain constant through the nodes which are uncovered. In reality, there will be a drop in temperature due to heat removal to the l secondary side, but this is expected to be small due to the low specific heat capacity of the steam and due the high primary side flow rate.

The heat transfer coefficient used in the uncovered tube region is discussed in the Appendix. This correlation bases the heat transfer on the difference between the tube wall surface temperature and the bulk steam temperature in the region. In the LOFTRAN nodification, the conservative assumption is made that no credit is taken for either a primary film heat transfer resistance er a tube metal heat transfer resistance. Therefore, the wall surface temperature of the tube is assumed equal to the primery fluid temperature.

The modified version of LOFTRAN automatically determines the proper number of steam generator nodes for the superheat region of steam in the generator. The variable node capability is applied to both the primary and secondary side. At each time step durin6 the tube uncovery,-the modified LOFTRAN code makes a general evaluation of the uncovered tube region (e.g. I steam flow rate, uncovered tube heat transfer area, estimated heat transfer c0 efficient, etc.) and determines the number of nodes to be used in the subsequent calculations. The total heat transfer for the uncovered tube region is determined and accounted for in the primary temperature transient

calculatten. The superheat/ tube uncovery modeling is applicable to all steam generators.

Figures 3 through 6 show typical results for a 0.86 ft steamline break from 102 percent power using the modified version of LOFTRAN. Figure 3 shows the fraction of tube uncovery versus time with uncovery of Loop 1 (faulted) starting at 152 seconds into the transient. At approximately 300 seconds, the uncovery transient reaches an equilibrium point where the steam flew out of the steam generator matches the auxiliary feedwater flow into the steam generater. Additionally, the tube uncovery transient fcr Loop 2 (non faulted) is plotted but shows no tube uncovery fcr the entire transient. Figure 4 presents the steam flow transient for this case.

Figure 5 includes plots of both the superheated steam enthalpy and the saturatien enthalpy for the Loop 1 steam generator. Figure 6 includes the Loop 1 temperatures for the steam generator tube inlet (primary side),

steam exit temperature (superheated steam), and the saturation temperature for the steam pressure.

C. NOTRUMP Model Comparison h e NOTRUMP computer code (Reference 4) was used to verify the LOFTRAN modeling of superheat. The computer code was originally developed to analyze transients of secondary systems with two-phase conditions. In the past, it has been used to analyze various transients in the primary and secondary coolant systems. .NOTRUMP has recently undergone major revisions to enable it to model non-equilibriun nodes (i.e., separate liquid temperature and steam tempera'ture modeling). Using NOTRUMP, the steam generator can be broken down into sufficient nodes to model the nonequilibriun effects of the steam generator, as well as the tube region during uncovery. NOTRUMP can model all modes of heat transfer associated with a steamline break transient, including heat transfer from the uncovered tubes to the superheated steam and the feedback effects between the primary and secondary sides. The two phase mixture level calculation accounts for primary to secondary heat transfer and the swell associated with rapid depressurization of the steam generator during the blowdown.

A comparison of LOFTRAN and NOTRUMP blowdown results is presented in Figures 7 and 8. The mass releases shown in Figure 8 show excellent agreement. h e LOFTRAN prediction of superheat enthalpy is slightly higher than NOTRUMP,'while the predicted time of tube uncovery is somewhat later.

NOTRUMP shows a chugging effect during the uncovery phase.of the blowdown.

This is believed to be in part due to oscillations in the flow link between the dcwncomer regien and the steam dome region. (The flow link is the drain path for the moisture separaters to the downcomer region.) With the flow direction towards the downcemer, superheated steam goes into the dcwncomer region and is condensed. This alternates with a flashing of a portion of the water volume in the downcomer region. This raises the pressure of the downcomer, resulting in a flow reversal in the link with saturated steam from the downcomer mixing with the superheated steam in the dome. This mixing results in the variations in the superheat enthalpy seen in Figure 7. Although LOFTRAN does not show the enthalpy variation since the detailed modeling of the downcomer and dome are not included, the overall agreement with NOTRUMP is very good.

. ', l i

D. Effects Of Analysis Assumptions The effects of superheated steam are dependent upon the occurrence and extent of tube uncovery. The major parameters affecting tube uncovery are:

initial steam generator inventcry, auxiliary feedwater flowrate, assumed feedwater system failures, and protection system errors. Variations in these parameters are in the process of being evaluated fer their effects on

~

the containment temperature response (Figure 9).

Refinements in the mass and energy release modeling (Figure 10), are being evaluated and several areas show a potential fer reducing the degree of superheat being generated. Some of these areas are:

- Evaluation of liquid-steam interactions such as the phenomenon of tube support plate flooding and heat transfer across the tube wrapper from

. the superheated steam to the auxiliary feedwater ficwing down outside the tube wrapper.

- A mere detailed steam header model in LOFTRAN.

- tbdeling temperature drops in the primary superheat nodes.

4 Evaluating other void correlations for use in predicting tube uncovery.

I

l o e l

l III. Containment Mode, ling A. Description of Containment The general phenomena taking place inside an ice condenser containment during a l steamline break transient can' be described utilizing a typical ice condenser elevation drawing (Figure 11). Steam is discharged to the main (or lower) compartment where heat is removed by the internal structures, steam flow to the ice condenser, and the. ice condenser drain water. The dead ended compartments are the regions which are located below the ice condenser and outside the crane wall (Figure 12). Air is discharged from the main compartment to the dead ended compartment and ice condenser so that the resulting steam to air ratio is that region is much higher than in dry containments. At ten minutes following the containment hi-2 signal, deck fans are actuated which direct air flow from the upper compartment to the dead-ended compartments. Most of the safety related equipment is located in the dead-ended compartments although some equipment and cabling are located in the main compartment.

B. Containment Podels Figure 13 outlines the major models and assumptions utilized in the LOTIC-3 containment code. In the currently approved version of LOTIC-3 documented in Reference 5, four distinct regions of the containment are modeled; the lower compartment, the dead-ended compartment, the ice condenser, and the upper compartment. Two condensate /revaporization models are used depending on the size of the break. For large steamline breaks, 100% condensate revaporization is assumed. For small steamline breaks, a convective heat flux model is used which calculates partial revaporization during the transient. The wall heat transfer model utilizes the Tagami heat transfer correlation for condensation heat transfer and the convective heat flux model derived from the work of Sparrow (Reference 6) which calculates the convective heat transfer for small steamline breaks. The sump recirculation system is only modeled for the large break LOCA transient containment response.

Figure 14 shows the four regions modeled with'the ' mass and energy flows that can be assumed in the analysis. The Catawba nuclear plant does not have lower compartment sprays and they are not modeled in the analysis. Superheat heat transfer is conservatively assumed to be zero for the steamline break containment analysis. In the model described in Reference 5, wall heat transfer is not modeled in the dead-ended compartments although these regions do contain structures which will remove heat. The analysis does include the upper compartment sprays, flow through the ice condenser, deck fan flow, and flow to the dead-ended compartments.

LOTIC-3 solves the conservation of mass, energy, and momentum equations for upper, lower, and ice condensor regions (Figure 15). After the new~Icwer compartment conditicns are determined, conservation equations are solved for the-dead ended compartment and the flow rate between the compartments is determined.

Figure 16 presents a typical steamline break containment temperatm a transient that is calculated using superheated steam blowdowns from the LOFThAN code and the modeling of ice condenser drains as a heat removal source. The transient shows that initially the containment temperature increases rapidly during the

1 l

, ., .1 blowdown. When the upper compartment sprays actuate there is a slight decrease in the main compartment temperature. The temperature then rises slowly until ice condenser drain flow decreases to the point at which time the temperature begins to rise again (approximately 250 seconds). This rise in containment temperature coincides with the steam generator tubes uncovering at 152 seconds and the maximum superheat occurring at approximately 250 seconds. The steam generator level stablizes whe'n the auxiliary feedwater flow is equal to the steam discharge at approximately 300 seconds. The containment temperature then starts decreasing with decreasing decay heat. At ten minutes, the deck fans actuate which results in a rapid decrease in containment temperature.

C. LOTIC-3 code Modifications Four modificatiens have been incorporated in the LOTIc-3 containment model which are (Figure 17);

1) vall heat transfer model

2) convective heat flux model

3) ice condenser drain model

4) dead-ended compartment model D. Mall Heat Transfer - - - -

The modificatien to the wall heat transfer model is described in Figure 18. In the LOTIc-3 model, only condensation heat transfer, utilizing a Tagami heat transfer coefficient and a temperature difference between the wall and saturation, was previously modeled,. The modification includes a convection term '

with a conservative convection heat transfer coefficient and a temperature

! difference between the containment atmosphere and an appropriate interface temperature. The Appendix presents a more detailed description of this model.

E. Convective Heat Flux The modification to the convective heat flux model is described in Figure 19. A term has been added to the convective heat flux model to account for the feedback effect from including a convective term in_ the wall heat transfer model. The Appendix presents a more detailed description of this model.

F. Ice Condenser Drain Model In an ice condenser containment there is approximately twenty drains exiting from the ice condenser into the lower compartment at an elevation of about forty feet above the compartment floor. The drain pipes are one foot in diameter.

The drain flowrate is calculated by the LOTIC-3 containment code. For-a typical i small steamline break transient the drain flowrate varies from approximately l 4000 lbm/see to 500 lbm/sec during the gimeframe of interest. The temperature  !

of the drain water is approximately 130 F (Figure 20).

Figure 21 presents the assumptions and the basic model used to estimate the heat-removal from the lower compartment atmosphere to the ice condenser drain water.

It is conservatively assumed that the drain water stream does not break up prior to reaching the floor even though many of the drains have equipment and structures located belcw them. Therefore, heat transfer is_ assumed to occur at-

the stream surface only. It is also assumed that the stream surface temperature is at the saturation temperature of the containment.

The heat transfer to the stream is:

q=hAAT wher'e h = condensation heat transfer coefficient A = surface area of the stream AT = appropriate temperature difference The calculation of the heat transfer surface area is described in Figure 22.

In order to model the drains in LOTIC-3, the drains are modeled as a wall heat sink with a surface at a constant temperature (see Figure 23). Currgntly,in the versien of LOTIC-3, the surface temperature is assumed to be 230 F which is close to the containment saturation temperature. The drain surface area is calculated at two points in time during the transient; early in time with a high flowrate and later in time with a low flowrate. To ensure conservatism in the area calculation a 10% reduction of the surface area was assumed.

As described previously (Figures 14 & 15), the LOTIC-3 containment model did not account for wall heat removal in the dead-ended compartments. To obtain a conservative estimate of the temperature transient in the dead ended compartment, the heat sinks located in the dead ended compartment region along with the heat sinks in the lower compartment are modeled in a ccmbined volume (see Figure 24). This " modified" lower compartment model is used -to determine a conservative dead-ended compartment temperature transient. Since the icwer ccmpartment will be hotter than th'e dead-ended compartment, this methodology results in a higher temperature in the dead-ended compartment then would be expected.

G. Transient Results With the modifications described for LOFTRAN and LOTIC-3, the previous FSAR limiting case for Catawba was reanalyzed to determine the impact of superheated steam. The.caseselectedisa0.86squarefootbreakat10g5 power (Figure 25).

The peak lower containment temperature for this case is 324 F. This temperature is calculated fer the lower compartment only. It is expected that the-dead-ended compartment temperature will be significantly lower.

In addition to the"model modifications incorporated in LOTIC-3, Westinghouse is pursuing further improvements in the areas noted on Figure 26. One area is in the wall heat and mass transfer models. Since condensation is a mass transfer type phenomena, the heat and mass transfer should be linked. This approach has been used.in Reference 7.

An improved drain model is also being investigated. . This improved model will calculate the drain surface area as a function of flowrate. It will also calculate the average temperature rise of the drainwater. This model will mere-accurately represent the actual phenomena in the containment.

4 i

WESTINGHOUSE PROPRIETARY CLASS II V. Appendix

WESTINGHOUSE STEAMLINE BREAK ELOWDOWN AND CONTAIN!ENT ANALYSIS METHODOLOGY The following sections describe the Westinghouse methodology for determining the containment response fer a steamline break incorporating the effects of superheated steam. These sections describe in detail changes from the 4 methodologies described in References 1 and 5.

1 I. Steamline Rupture Mass / Energy Blowdown Analysis A. LOFTRAN and MARVEL Computer Modeling Mass / energy releases can be calculated using either the LOFTRAN code (Reference 3) or the MARVEL code (Reference 8). The LOFTRAN code'is used for non-LOCA FSAR accident analyses. The MARVEL code was specifically developed for assymmetric transients such as steamline breaks.~ These two codes are very similar because they were developed in an interrelating fashion znd much of the model'ing is common to both codes. The MARVEL code was used in the development of Reference 1 because LOFTRAN at that time was a lumped model which was used for symmetric loop transients. Furthermore, for steamline break analysis purposes, MARVEL contains a model fcr water entrainment. However, the current version of LOFTRAN is a multiloop version which also contains a water entrainment model. With the development of a multiloop version of LOFTRAN and the inclusion of an j entrainment model, the use of MARVEL has been generally discontinued. This enables the use of LOFTRAN as a single system analysis code for non-LOCA l

transient analyses. LOFTRAN is used in the analyses presented here.

The model of hnportance to blowdown calculations is the steam' generator model. The primary side of the steam generator contains multiple nodes to l model the tube bundle for both the modified version of LOFTRAN and MARVEL. I Heat transfer calculations from the primary to secondary side are identical in the two codes, although the methods for initializing the heat transfer resistances are slightly different. The secondary side is effectively a i one node, two region model of saturated steam and water. Heat transfer is' i assumed to occur to saturated water. If tube uncovery is predicted, the -1 stount of surface area available for heat transfer is reduced.-

Both codes contain a detailed steam generator model which is used to predict tube uncovery. This model calculates the liquid volume in the -

steam generator shell and acgognts for the detailed steam ' generator-.

geometry. The [ ] correlation is used in both codes to-predict the voiding in the tube' region, although the correlation is modified for use in LOFTRAN. In MARVEL, tube uncovery is calculated based

• b

~ - ~

• s wn e , - .

.m

m-.ea.  : cA, 2 -n a J a. a ,' u --.i-s 4..a.m-_

i on comparison with the actual water level and the height of the tube bundle. In LOFTRAN, the user specifies either a water volume in the steam generator corresponding to tube uncovery, or a void fraction in the riser section of the steam generator &t which tube uncovery begins.

Both codes have similar models accounting for reverse heat transfer, thick metal heat transfer, feedline flashing, and safety injection system

operation. Auxiliary feedwater flow can be input as .a fraction of nominal feedwater flow, although LOFTRAN has an additional capability to model auxiliary feedwater flow as a separate system. For analysis of. double ended ruptures, MARVEL accounts for the volume of steam in the piping

,' downstream of the steam generators in the blowdown calculations. In 3

LOFTRAN, this consideration is added en to the blewdown mass and energy results by hand. For split ruptures, which the analysis presented here addresses, the steam piping masses are handled identically in both codes.

In summary, LOFTRAN and MARVEL are very similar codes, and either can be used to calculate mass / energy blowdowns. To demonstrate this, a comparison of the blowdowns fer a typical case is presented in Figures A.1 and A.2.

Figure 1 presents the mass release rate for a .86 ft2 split rupture from 102% pcwer. For this case, Figure A.2 shows the saturated steam enthalpy as a function of time. This blewdown is typical of results used in FSAR analyses prior to the modification noted in this report for the LOFTRAN code. As can be seen from the figures, the results are extremely close..

B. LOFTRAN Model for Superheated Stem I

As menticned previously, the LOFTRAN code has been modified to model heat transfer which may occur in the uncovered tube bundle region. This effect is modeled in both the faulted and intact loops. In the modified versicn

. of LOFTRAN, all heat transfer occurring in the uncovered region is assumed to add superheat the steam exiting the steam generator. The temperature of i the primary coolant ficwing through in the uncovered '.ube region mode is conservatively assumed to remain constant. Realistically there would be a drop in temperature due to heat removal to the ~ secondary side, but this will be small due to the low specific heat capacity of the steam and due

the high primary side flow rate.

1 Theheattransfercoeffic}gt,ugedintheuncoveredtuberegionisbasedon the [ '

] ' . The heat transfer coefficient (U) is calculated by the following expression:- _

<- a,c i

9 This correlation is presently.used'f transfer by the [ -

]5'Buperheated computer; forcedLeonvection codes. Additionally, heat

-,w ...,.'A q - - w v e en ye -a y g

1 l

, .; i l

this correlation is based upon the heat transfer from the surface of the l tube wall to the average bulk temperature of the steam. In the LOFTRAN

! modification, no credit is taken for either a primary film heat transfer resistance or a tube metal heat transfer resistance. Therefore,the wall temperature of the tube is conservatively assumed equal to the primary fluid temperature.

l

~

A, C (1) f e

a e

me e

I 1

I L- - -

u___._.___________._____.______.__- _ .__ __1_ _ ._____._.)

i i

The modified version of LOFTRAN automatically selects the proper number of l steam generator nodes for the superheat region of steam in the generator.  !

The variable node capability is applied to both the primary and secondary side. At each time step during the tube uncovery, the modified LOFTRAN code makes a general evaluation of the uncovered tube region (e.g. steam flow rate, uncovered tube heat transfer area, estimated heat transfer coefficient, etc.) and determines the number of nodes to be used in the subsequent calculations. Each node is evaluated to determine the steam temperature exiting the node with a convergence criteria that is based upon the total number of nodes used. The exit steam temperature of one node is used as the inlet steam temperature of the next node.

The heat transfer calculation to determine the outlet temperature of the node is based upon the following expression:

Q = UA*(T 1-(TougTin)/2) = Ms*C 3 *(Tout -Tin) where Q = Heat transfer to the steam U= 1 O

Tpri = Primary node temperature Tog = Steam node outlet temperature T in = Steam nod'e inlet temperature' M3 = Mass flowrate of the steam C3 = Heat capacity of the steam A = Heat transfer area in the node including both hot and cold leg sides of the tube bundle The total heat transfer for the uncovered tube region is determined and accounted for in the primary temperature transient.

C. Blowdown Sensitivity to Plant Conditions The effects of superheated steam are dependent upon the occurrance and extent of tube bundle uncovery. Parameters affecting tube uncovery are:

initial steam generator inventory, break size, auxiliary feedwater flowrate, and the single failure assumed.

The initial steam generator inventory depends upon the measurement errors associated with steam generator level and upon initial power level. Steam generator mass increases with_ decreasing power, thus, breaks-intitiating from low power levels will result in later tube uncovery.

Larger breag sizes result in faster blowdown of the steam generator and earlier tube uncovery.

I

Large auxiliary feedwater flowrates only delay tube uncovery, but will l also cause the final equilibrium steam generator level to be higher. This I equilibrium condition corresponds to the point when the break flow rate is l equal to the auxiliary feedwater flow rate. i l

The single failure assumed in the transient may impact the amount of water I supplied to the steam generator. Auxiliary feedwater runout will increase the amount of water supplied to the steam generator. Failure of the feedwater isolation valve will also cause extra water to be supplied to the generatcr as the additional mass between the isolation valve and the check valve flashes to the generator.

0 e

4

+

e e

+ .

II. Containment Analysis ,

A. Wall Heat Transfer Model J

The original LOTIC-3 wall heat transfer model is based on the stagnant Tagami heat transfer correlation. That is, .

q"=h AGAMI(TSAT-TWALL}

~

hTAGAMI = 2 + 50 MSTEAM " AIR h(TAGAMI, MAX)=72 BTU /hr-ft - F This model was developed for saturated steam in the presence of large amounts of non-condensable gases. In the lower compartment of an ice condenser, most of the air is swept out of the lower compartment through the ice condenser and into the upper compartment. Therefore, after about 30 seconds, there is almost no non-condensables in the lower compartment. Typical values ror de condensation of pure steam are in the range of 1000 to 3000 Btu /hr-ft2 gF (Ref. 5). The correlation used in the modified LOTIC-3 code is in extension of the Tagami correlation fer nearly pure steam.

q"=hCOND (TSAT-TWALL) hcond = 2+50 MSTEAM " AIR h(cend, max)

• A maximun value of [ l a,c was chosen as a conservatively low condensing heat transfer coefficient in a nearly pure steam environment.

In addition to this modific'ation, an additional term is needed to account for the convective heat transfer from the superheated steam to the condensate film.

This convective heat transfer is dependent upon whether there is condensation occurring on the walls. If condensation is occurring, the correlation used is:

com ucombul[s where: a,c If the wall temperature increases to above the saturation temperature then the convective currents will be reduced such that the correlation used is 9"conv=hconv(Tbulk-Twall) where:

[

ja,c

Th'us in summary, if Twal1<Tsat men

[ 3a ,c If Twall > Tsat, then the correlation used is: -

[ j a,c 5

6 i B. Convective Heat Flux Model When the containment atmosphere is superheated, the containment tenperature is a strong function of the amount of steam mass in the atmosphere. Thus the amount of mass condensed on the heat sink surfaces is a key parameter. The actual amount of condensate formed is -

"cond

• 9cond fg Unfortunately, with the use of a heat transfer correlation based only on test data (such as Tagami or Uchida), only dae total heat transfer coefficient is obtained. This total heat transfer coefficient includes both the condensation heat transfer and the convective heat transfer. Based on the work of Sparrow (Reference 6), the Westinghouse Convective Heat Flux model in the original LOTIC-3 code calculates the ratio of the convective heat transfer to the condensation heat transfer. Therefore the calculation of the amount of mass condensed is

[ ja,c A .

In the modified LOTIC-3 model, the, amount of superheat convection is calculated.

The amount of convective heat transfer at saturation is not known explicitly in this model. Therefore, in the modified LOTIC-3 code the original convective heat flux model will be used to calculate dne fraction of convective heat transfer for saturated concitions. The actual correlation is .rc s s

where, (q /q is determined from original convective heat flux model and q conv is eggg)$$. bunt of convective heat transfer celculated in the wall heatEPSESf$rmodel In summary, the modified LOTIc-3 model is consistent with the original LOTIC-3 l codel in its calculation of the mas condensed. The only difference is that in the modified LOTIC-3 code, the amount of superheat convective heat transfer is known explicitly, while in the original LOTIC-III mcdel, only the ratio of convective heat transfer to condensation heat transfer is known. l 6

9 O

s **. ,*. ,

IV.

References:

'1.. Land, R. E., " Mass and Energy Releases Following A Steam Line Rupture" WCAP-8822 (Proprietary) September, 1976 and WCAP-8859 (Non-Proprietary).

2. NS-EPR-2563, February 14, 1982, E. P. Rahe of Westinghouse to J. R. Miller, NRC, " Additional Infcrmation on WCAP-8822".

3. Burnett, T. W. T., et al. , "LOFTRAN Code Description," WCAP-7907, June, 1972 (Proprietary).

4 Meyer, P. E., and Kornfilt, J., "NOTRUMP - A Nodal Transfer Small Break and General Network Code," November,1982, WCAP-10079 (Proprietary) and WCAP-10080 (Non-Proprietary).

5. Hsieh, T. and Liparulo, N. J., " Westinghouse Long Tem Ice Condenser Conteinment Code - LOTIC-3 Code," February, 1979, WCAP-8354-P-A Sup. 2 (Proprietary),WCAP-8355-NP-A (Non-Proprietary).

6. Sparrow, E. M., Minkowycz, W. J., and Saddy, M., " Forced Convection Condensation in the Presence of Noncondensables and Interfacial Resistance", Int. J. Heat Mass Transfer, Volume 10, 1967.

7. Corradini, M. L., " Turbulent Condensation on a Cold Wall in the Presence of a Non-condensable Gas" Nuclear Technology Vol. 64, pp 186 - 195, February,'

1984. ,

8. Krise, R. and Miranda, S., " MARVEL - A Digital Computer Code for Transient Analysis of a Multiloop PdR System," November,1977, WCAP-8843 (Proprietary) and WCAP-8844 (Non-Proprietary).

9. McCabe, W. L., and Smith, J. C., " Unit Operatons of Chemical Engineering",

3rd Edition, 1976.

4 6

a. " . .%

1 LOFTRAN - MARVEL COMPARISON

.860 FT2 BREAX AT 102 PC POWER 2000.0  :  :  :  :

f

~

1750.0\ /-

n d

w 1500.0 --

s 3 ~

1250.0 - -

u 3 -

"a 1000.00 -

w ~

G 750.00 - -

z

' < LOFTRAN ~~

Eca 500.00 -- MARVEL --+ /

~

~

250.00 --

12/05/83 0.0 i i i i i C

o O

O O O O e & 5 5 'd 5 o

o o

a e o e, n w e

- cu m TIME (SEC) i 1

l FIGURE A.1

= _ .

A l

LOFTRAN - MARVEL COMPARIS0N

.860 FT2 BREAK AT 10'2 PC POWER .

1300.0 .'  :  :  : .  :

e 1275.0 - - --

l 4 o

y 1250.0 - - --

1225.0 - - --

OFTRAN 1200.0 - -

N ..

= MARVEL - ~

~

5 1175.0 - - --

=

W 1150.0 - - --

m S 1125.0 -- --

1100 0 o e o o

e o a o o o o o .

o o o e o e o

. e o a .o -o o o - ~ m a w w ,

1

~

TIME (SEC) 1 i

FIGURE A.2.

c- _

e/

' ~ "

e METHODOLOGY FOR ADDRESSING SUPERHEATED STEAM RELEASES IR ICE CONDENSER CONTAINFENTS Purpose ,

The pur;ose of this report is to document the information presented on March 19, 1984 in a meeting with the U.S. NRC Containment Systems Branch en the status of progress made in addressing the confirmatory item on the Catawba Nuclear Plant Safety Evaluation Report. This confirmatory item deals with the effects of superheated steam generator mass and energy releases following main steamline break accidents. Attachment 1 includes the list of attendees at the meeting and the overhead slides covered in the Westinghouse presentations.

Technical presentations were made describing the modeling of the steam generator and heat transfer from the uncovered tube bundle during the steam generator blowdown along with a description of the containment model and transient response. A proposed plan of action was also presented and discussed with the Staff. In accordance with that plan, this report represents the first milestone in the proposed plan of action. As committed to in the meeting, the appendices present proprietary infcrmation which relates to the specifics of the models and sensitivities that were not directly addressed in the meeting.

Attachment 2 is an explanation of, and refers to, the overhead slides (Figures)

p. resented at the March 19 meeting. - .

e R

9

+

{

(

' E

% 4) 0 ATTACHFINf 1 L.

k O

i l

i Q et

/] %d.,- w o ;- 3/t7/8f .

' Ffc.e H $ t,*/ .D k n 4.U *, .

7-~ k A .,9 .i . s

/ % ,E T h L.'

ka y .

C < ~'T * *

• 3 h '* '

u y

~

}(, Al fA8200R.

AlKC /DL p '. o . .sanwn zwa./~n

@ p a eea '.a- S+ &r?'

J L. Li r7 ii

%.V. Tos.gAC ti c o c E A R S A Ft.T 7' 4 Le //L

?

~.Y'y'~~~ j'e 7 e W N u u. e t. G AF&'rv bS. Love U M u cl.co,.

, t*\ ? M oc 4 ft? . Li,' . .

R/ t r c (c ,> , Jel<ry L . 2. JAwnrsc- .W .vaha --

ft b c les.r- S u & l'y bb C. A D5M _LL! .

r- 7

/

I woc .,cA Q MoQ , p'f ( m}f'.

S .D A/<4.r,J.,v- D u.l<- c. , D e s;= 5 m .- . ,

S 5 &*Sv b de.,.b<r:j E y:r<.w:

cc,us 7w, x a m Lca.,s ,fr.u-

- r. E.

56ted LICEd.s/nG .Tr.4ff , EE. P E. ArttL 0.h(q'uc.ua Q a - <.uw. sa p~y .

W. D.

T V A, G9: ./ 3 *

• Dn ,jn Crouch .

N./<,. f,.Id.- Jic.,cb A5 '

p. f. ,vy4 P O.n. ,

.fA/wam w , ur p .rmt g .v r a / a ue w .< r.werr

!- '?. H. O n .e

.Sh/,24tcW' ruAA- C + 8' JJh \(s. ,fDr ,Nc%._

" U-i U'l. R,k.L- AG /cS/3 C. Li Npc 4 i.tc A. Madoacesca v$c/csv3

STE.AM GENERATOR SEcoWDARY ,

STEAM F't.oW .

a

.* *y SATORATED STEAM RE6ted

~.- ...

',<l-

, SATURATED ~

'__.. ~

. . WATER macg ff , . . .

4

.ij". .

9 i

• StMGLE NCDE , 2 It,EGled MCOEn

• MO NEAT TRAdSFER. TO SATueATED STEAM. REGtcM

• HEAT TRAMSFER, To SATVRATED WATER RE.Gten nS MODtF1ED Rl4it.

TVSE 'UNCCN ERT ' -

FIGURE 1 .

• O

~

{ (

SUPERHEAT HEAT TRAMSFER 1

SUPEAHEATEb STEAM FLcW

' ' N DAM ee .- eems p_ j f sop _enweAT 9ects 3 .

'9NSe s' .

. - +.:.r.< ? *

t. r t-t: C T I V E '

~ UN'COVERY LJiN1iiL

~

~

~'

>- i .

a

\ # _

A A vAa,nss- -

HoDES MODEL l

• CoHSTART PRtuAm TEMPERATURE tu. suesansAT 1assieu Assuutt Fos.

447 AT NSFast. cal.COLATICMS .

'

• C ALCO t.,AT4.D 509E12 MEAT HEAT vytAss74W, AcccONTEb , FCR. IM

, pga.uAAny TEAdStENT ciavRE 2 '__ ..

z---; 3 _; ----

1

- e l

~

TUBE UNC0VERY LOFTRAN SUPERHEAT MODEL

~

.860 FT2 BREAK AT 102 PC POWER 1.2000  :  :  : l l G

1.0000 --

E LOOP 1 5

p y .80000 --

E a .60000 --

c:

w

> .40000 - -

0 M

2

$ .20000 -- --

S LOOP 2 0.0 8 8 o s s s s s s d S 8 8 8 TIME (SEC) 6 FIGURE 3

4

, MASS BLOWDOWN LOFTRAN SUPERHEAT MODEL

.860 FT2 BREAK AT 102 PC POWER f

2000.0  :  :  :  :

[

1750.0 k

n d 1500.0 -- --

4 R

S 1250.0 -- --

3 S

"z 1000.00 - - --

w

$ 750.00 -- --

M E

CD 500.00 -- --

250.00 -- --

0.0 O O O O O O O O O

• O O O O b O

. O O O O O O " (\J (0 -t LO CO

, TIME 4SEC) i

.. FIGURE 4

o .

~

ENERGY RELEASE LOFTRAN SUPERHEAT MODEL

.860 FT2 BREAK AT 102 PC POWER 1300.0

- 1275.0 - - --

3 SUPERHEATED STEAM S

1250. 0 - - --

e_

g 1225.0 - - --

T w

$ 1200.0 -

I SATURATION 5 1175.0 - - --

=

id! 1150.0 -- --

m S 1125.0 - - --

1100.0 g

o 8

8

<u 8

m 8, 8 o

8 m

TIME (SEC)

FIGURE 5

4

• . I l

TEMPERATURE TRANSIENTS LOFTRAN SUPERHEAT MODEL

. 860 FT2 BREAK AT 102 PC POWER G50.00  :  :  :

1 600.00 -4 --

U SI tg 550.00 - --

c.

8, SG TUBE INLET u.-

Eo m y 500.00 --

(PRIMARY SIDE) --

t2 ~

• $z 450.00 - -

SUPERHEATED STEAM m

o.

I g 400.00 - - --

350.00 -- --

SATURATION 300.00 ' ' '

8 8 8 8 8 8 o

d- 5 cu 8

m 5

. .:r 5

in 5

e TIME (SEC)

FIGURE 6 l

1 k8m

~

k8o I

~ k8,

)

C E

S t

E M

I T

N _

0 _

S .

I R

1 A

k8m A R P

M T O F C O L

PY P IA M L U A RH ~

TT g O NE 8~

N NK AA R ER T

F B

O L

8-M A-V -

- Ay o 0 0 0 0 0 0 0 0 0. ,

0 5 0 5 0 5 0 5 0 0 7 5 2 0 7 5 2 0 1 1 1 3 2 2 2 2 1 1 1 1 1 1 1 1 1 1

_ 2$3 O Es.- 5 aa lE" G

. e$w s$o s s$a )

C E

S

(

E M

I T

N 0 N P S A M I

R -. T R

R l

l a$m A F T P

M O O O L N CE T

PAR M

UW RO TL -

OF s8eu NK NE AR RB A

--\

T F s

\

O L

8 $-

• o 0 0 0 0 0 0 ~ #

0 0 0 0 0 0 0 0 0 5 0 5 0 0 2 2 1 1 5 awsx3 =: =$ 0<z %wE t a~

2sAw L !

I i

EFFECTS OF ANALYSIS ASSUMPTIONS INITIAL STEAM ~ GENERATOR INVENTORY AUXILIARY FEEDWATER FLOWRATE

FEEDWATER SYSTEM FAILURES PROTECTION SYSTEM ERRORS t

/

I .

FIGURE  ; q w J M v n -

v v nq %>

ADDITIONAL MODEL CONSIDERATIONS f

LIQUID-STEAM INTERACTION IMPROVED STEAM HEADER MODEL HEAT TRANSFER THROUGH TUBE WRAPPER TEMPERATURE DROP IN PRIMARY SUPERHEAT N0 DES OPTIONAL VOID CORRELATIONS

(

i I

-FIGURE 10 L

L -.

o SHIELD BUILDING DCME Q

, / .

. .j. .

'~~

.,[/,1 e -

'--STEEL CONTAINHENT vt33EL h El829.04

-> P -

Sy,ggy **

f175 TON POLAR CRANE $.A

• BUILDING { . POLAR CRANE .*

MILL m

• l

.. n n .*-

TCP OF CECX N TCP Of CRANE RAIL *

• 'RRING LIN*

ICE LQAC/NG ,; fEL 796.63 &f} mk^ '_5fEL 797/2 / El 731.5

/.e A

VENT T10N * *.

CYCLCNE -

• H

• g y ,IR ygg,yg R

Js,0n I

$8R/DCE CRANE N st*-6 *R 3..

UNIT 3 l

TCP OF ICE BED l l '~ *. e ,*-

" I'0 * % c . ; ;. . , .y f o'vicEn e^RR!ER , u ?To.dsm . . _ .

,.,.4 ,

. e' .. .. . .. .. .

s .',

~

]j- ~

. . #CAANC u ygg.gf-.-

• l 8-:. .: 3l~ *'

/CE 19: :i '. gggg4 ., g

'#"#'" ~ '

,. f, n

q 35 :

\ n~{g'l i

l

< cNrA.N-ENr a,:.:v.. .

STEAM .,

• . GENERATOR t

, s .! .. CONTROL A00 CRIVE. .

L:d 2.  :

' (,' .. M/53/LE $NIELD ,  : (.si: j **

_ , , , , j . EL H 7.22 '

'\' ,

Q31

' ' ' ' " ' ' ' - "s-

< v, ] ,-a nr.u ; -

y \ /K I > .

,,,0 , .

1 lsG} R aw X .3 MiNL..

fi?!I.i.{.,i

i. .

+ _ .

sARRiEn a g (;gCry, .

,,.o.

.0 lg. e ACCU-vuroa-w n euu, -

.7. ; ..

c. s% .m .

U -*

C# ACE e- 1 -

( !

EL 705.0 %

~ -ennsrernn

'.t> "

f l }f F*LOOR EL 7C2.0 9

n

, 4 t,::

• 1 *

• }f{m EL695.0~: 'a " *

.,4f ' '

\ ,. y.

? ,,,, VEN7 FAN i W EL 692.0 1",)

Q . .?

~

L' N "II *

• w-l\

l; t,, o U!.

lY ,

4 EQPT

. l . e .k G 1y l * - .\ .

. .g ..,

-l ;,. .

.n . . s *.g . g 0 ,

l

• Y f,7

,- " EL 679.78

" l0 " '

• e.E..

. , , . , ;l;-

g-

• P J==ra. .. y , .. l

. - .. srEa uNis

.Na eicr6 .

al .' .

. o /$

a_

'f . .

N ANCNOR BCLT .* i .e l

,; . i s+ = o.som m e

(EL 65/.69

, .saus.). . . .

Reactor Building Elevation R'- -

FIGURE 11 l l

i

l l

I

. \

f

'7 ,.

% L.

SPENT

%A UX/LM RY BLDG % 'b 'lr' .

FUEL, Pir 4 b' 4 '. I

,v.l

. ', .A..e n . 4.'

l .'sl e, ,

FUEL TRANSFER P/PE CHASE qc SYSTEM W ,;Q.'*&:* ., - .t; . <n . r. 9;.4 _,E.

.y.,

W, 4 6.,..a.

.-l

.ks .

e:

.g

.,.-f-

.M . .

EL 693.0 & ,

l

,,,{ ]

', c .

}g

.v.>~

REA CTOR UPPER INTERNALS STORAG[

s v..s I . .

s

'e l , -

.i g, ,'Cj . h CTOR

%*R.E. A. 3

,

• p' O .

o . . . . r '. '. a.-

~

.y' p~.

CoblA,NT -

4- t f..A.. __W s' ! .h. C.. '" \ PUklPJ (d

• N .r .Q ., %L OWERs *:. Q.~

, . ' /NTERNALS

.g .

4 , ..i, ..< 6 .

~

l *

~

~[n ' Y/

~ -

STEAM GEN (4)

• . r, f E 7

• f ', .m~  : :GA TE' * .**: *'

U ...- _

ygpg ,j 4 j ..*

Q,A,4 ",e 1

REA CTOR

<: ppgssung

-r --

STEAM ., ._ .'.

.ll. CRANE WALL

.A

' *' 4 _ . r_i-- L's, _. , ESSEL ,.

' I EL h

.' , W. *' ';*'- .

-( 702J5

.Q FEEQ , O.

WATER. -s EL 679.78

• ~

' ' 'VENTIL A TION

^

• r. ggtigy ,

('

y,_

'O TANK , D . r;

% [ UNITS (4)

., -.~ .

..~. '

. PRESSUR1Z _,i

,; i,'$.. i '. Y ? ,l5EAL

/ ,[iCCUMULATORS (4) 1

' , .. M .'

~

' .f .....q" s

k TABLE 'f' l .

f, EL 693.0 MEAT EXCHANGER'S - F \

-N

a .- ,. L EL 700.0J *. i j '

VENTILATION UNITS *(2)y _,

• PER$0NNEL LOCK

\.i'

~

V/

(,'

Figure 3.8.3 7 Plan-Lo.er Compartment T:-

FIGURE 12

LOTIC-3 CONTAINMENT CODE 4 N0DE CONTAINMENT MODEL CONDENSATE /REVAPORIZATION MODELS LARGE BREAK (TOTAL REVAPORIZATION)

SMALL BREAK.(CONVECTIVE HEAT FLUX).

• WALL HEAT TRANSFER MODEL

• MODELS SUMP RECIRCULATION SYSTEM

~)

l l

a.

l FIGURE 13 l

s' r

i r

1 i

1 I

i I

[l

~

A Sorays a 4 4 P 3 , .g 3 I

Qwall ll Upper Comoartment I

i m,h t A g-T s s) i U.C.

m.,h34  !

Orain -

f  ; I

! Ice *s' 's 4 m'h 1ce i Concenser a a j . One way flow SaH - -

l i

, through open area L,  ; on operatino deck Condensate & M Ice Melt l mg , h, I.C. Sumo m,n a a '

I.C7 Oraini  ;

Y v

• I '

^ ^

I ^ (RCS breax -low, "s' s "

+ L ,h-3 o accumul at; l Dead-Ended Compart.ent m'n

a a ' L

wer Comoartment release , ct boil-of'

.q

~-

_ Air Addition Tao)le

' 'I C Heat Acdition Tabla l L.C. ,,

i' rc wan

'Orain " sumo

' ' ]

l U.C. L-r i .

i Orain ) - L.C. Sume FIGURE 3.3 MAS 3 MD CIERGY FI.0W DIAGRAM FOR THE COMPARU.C, is FIGURE 14 O

w:

i l

LOTIC METHOD OF SOLUTION 4

SOLVES CONSERVATION OF MASS, ENERGY, AND MOMENTUM FOR UPPER, LOWER, AND ICE CONDENSER REGIONS ONCE NEW LOWER COMPARTMENT CONDITIONS ARE DETERMINED, CONSERVATION EQUATIONS,ARE SOLVED FOR THE DEAD-ENDED

, COMPARTMENT AND FOR THE FLOW RATE BETWEEN THE TWO COMPARTMENTS l

l l

l FIGURE 15 l

L _

TYPICAL CCMAINENT TBPERATURE TRANSIElfl' (DRAINSTODELLED)

1. LEVEL REACHED IN S.G.

-J U.C. SPRAY ~

[ DECK FANS ON T j REDUCTION IN I.C. DRAIN FLOW b '

h ~

B.

I 0 100 200 300 L100 500 . 600 l ,

. TIME (SEC.)

, FIGURE 16 l

1 MDIFICAfl0NS TO THE CONTAINTNT V0 DEL WALL HEAT TRANSFER MODEL CONVECTIVE HEAT FLUX MODEL I l

ICE CONDENSER DPAIN f0 DEL DEAD ENDED COMDARTfENT MODEL e

u FIGURE .

. . 4 WALL HEAT TRANSFER t0 DEL ORIGINAL L0flC MODEL

=h Tagami (r - r SAT All

)

PbDIFIED LOTIC MODEL

//

CO/;g i g / CO^ V w/h ' )

h cono =f( % ) ..

Con y f Wall j T Te = R r%Il / Snr r>

I l

l l

FIGURE 18 1 l

1

i CONVECTIVE HEAT FLUX F0 DEL ORIGINdL LOTIC MODEL h cono = Scono h,y Jrornf_t_

J,,,  !._______)

+x fbDIFIED LOTIC MODEL

' ~

h ij)cono = Scom hyy = 3rora1[!

h, p + X,,,][i e,, (g- y g] -

gforal _

l FIGURE-19

> l I

ICE CONDENSER DRAINS

-APPROXIN%TELY 20 ICE CONDNESER DRAINS

-DRAIN ELEVATION IS ABOUT I40 FEET FROM FLOOR

-DRAIN PIPE IS 1 FOOT IN DIAMETER

-FOR TYPICAL MSLB TRANSIENT, DRAIN FLOW VARIES FROM1 4000 LB/S TO 500 LB/S

?

i f

s i

i

( 1 l

l

! FIGURE 20-

e e

ICl- CONDENSER DRAIN f0 DEL

~

-CONDENSATION OCCURS AT THE SURFACE OF THE STREAM

.-FLOW IS WELL MIXED q = h A AT

-EDDEL AS A WALL AT A CONSTANT TEMPERATLRE

-A IS THE SURFACE AREA 0F THE STREAM

-h IS A CONDENSING TYPE HEAT TRANSFER COEFFICIENT l

FIGURd 21.

i I . _ . - _ __ , , _._ , , _. .

r i

e

• e e i

J 1

f CALCULATION OF THE STRENi FLOW AREA

{

1 3

)

P = n (Pxi) = ao (ex vo) = 800 p WHERE P IS THE PERIMETER OF THE STREN4 b

4 8

I J U

///*

T

- /. ,.

y j  : V ,-'l

, __. m

!' f g-. , ___3

! L II.

h

. _, c Q

FIGURE 22

_ m , - - . .m- a >-.a .a -+e- u = n n- - .w.. s- _ sg. a 6

O e M e

F 4

FDDIFIED LOTIC DRAIN M) DEL i

f P 4

WALL WITH A VARIABLE AREA 4

1 -

cono b {Tem ~ [sg)

I

}=

Tg ,, .

L7"230 F 1

6 a

(

j O

, FIGURE 23 _

a O I c-m.'s._ .

DEAD ENDED COPPARTENT FDDt-L i

L0ER C0FPARTFENT DEAD ENDED C0FF i l i

I I

i P001FIED LOWER COPPfTIEff j l I

.i .

-FIGURE'24

I i CATAWBA RESULTS t

4

- 102% POWER l

0.86 FT BREAK MAXIMLN AFW FLOW FSAR HEAT SINKS

~

i MAXIMLN S.G. INITIAL f%SS

{ .

T = 32Li'F MAX (LOWER COWAR~ LENT) 4 4

d S

V i

FIGURE 25;

-t w w e t- m -

$ r* ~ = "&9--- "-W*7-M' *r e e t' * * '*'

i 1

1-J

- ADDIT 10tML M) DEL CONSIDERATIONS 2

4 WALL HEAT TRANSFER MODEL I

DRAIN MODEL .

s DEAD ENDED COlPARTfENT MODEL '

i

)

1 i

4 e

t

+

l s

t I

i I

1 5

• o i

e FIGURE 26' - -

2

,w

.+e --w --r, zn--~--e-.-y-,,...we, ,,y-- , y +. Ewe e y g-ee g e we . .v-, ,wt,, , , -yr-p gv > #,,, g<- n y- y-,- ,.y- + - - r

l i

ATTACHENT 2 , . .

OUTI.IE OF THE REPORT-1 I. Introduction II. Mass & Energy Release Modeling III. Containment Modeling i

IV. Action Plan '

V. Appendix VI. References i

• 4 e m i

i O

l f

4 A

9

-m

. g vT T -, y  ? --w<* m y -

. s l

I. Introduction During the Containment Systems Branch review of the Westinghouse topical report,

" Mass and Energy Releases Following a Steam Line Rupture",WCAP-8822 (Proprietary) the Staff noted that heat transfer to steam from the uncovered portion of the steam generator tube bundle was unaccounted for and questioned the effect upon the calculated mass / energy release and the subsequent effect on the ccntainment temperature response. Westinghouse responded in a letter to the Staff (NS-EPR-2563, February 14, 1982, E.P. Rahe to J. R. Miller) that it had determined the impact of the effect by conservatively treating the maximum amount of superheat to be the difference between the primary coolant temperature and the steam temperature. The letter noted that there would be an insignificant effect en dry type containments and that, based en the contervative model used, there would be an expected increase in containment temperature for ice condenser type containments. In the Containment Systems Branch Safety Evaluation Reports en the topical report and the Catawba Plant Safety Evaluation Report, the Staff required that a more refined steam line break analysis be performed to detennine the effect on containment temperature which might impact the environmental qualification envelope used for safety related equipment.

Since that time, Westinghouse has investigated the effects of tube bundle heat transfer from the viewpoint of a more refined modeling approach. Subject to the final review and approval of the NRC Staff, the efforts and results obtained to date indicate that there is little impact on the containment response from the effects of the additional tube bun,dle heat transfer to steam.

9 e

a D

. 1 II. Mass and Energy Release Modeling A. LOF1TIAN Computer Code  !

Mass / energy releases are calculated using the LOFTRAN code. LOFTRAN is a FORTRAN language, digital computer code, developed to simulate transient behavior in a multi-loop pressurized water reactor system. The program simulates neutron kinetics, thermal hydraulic conditions, pressurizer, steam generators, reactor coolant pumps, and control and protection systems. Up to four independent loops may be modeled. LOFTRAN is used for analysis of non-LOCA transients and is documented in Reference 3.

The model of importance to blowdown calculations is the steam generator model. The primary side contains multiple nodes to model the tube bundle.

The standard LOFTRAN steam generator secondary side model, (Figure 1), is effectively a one node, two region model of saturated steam and water.

Heat transfer is assumed to occur only to saturated water. If tube uncovery occurs the amount of surface area available for heat transfer is accordingly reduced. The ' 0FTRAN code incorporates a more detailed steam generator model which is used to predict tube bundle uncovery.

B. LOFTRAN Model for Superheated Steam The LOFTRAN code has been modified to account for heat transfer to steam from the uncovered tube bundl,e region. (Figure 2). In the modified version of LOFTRAN, all heat transfer occuring in the uncovered region is assumed to add superheat to the steam exiting the steam generator. The primary side temperature in the uncovered tube region is conservatively assumed to remain constant through the nodes which are uncovered. In reality, there will be a drop in temperature due to heat removal to the secondary side, but this is expected to be small due to_the low specific heat capacity of the steam and due the high primary side flow rate.

The heat transfer coefficient used in the uncovered tube region is discussed in the Appendix. This correlation bases the heat transfer on the difference between the tube wall surface temperature and the bulk steam

. temperature in the region. In the LOFTRAN modification, the conservative assumption is made that no credit is taken fcr either a primary film heat transfer resistance or a tube metal heat transfer resistance. Therefore, the wall surface temperature of the tube is assumed equal to the primary fluid temperature.

The modified version of LOFTRAN automatically determines the proper number of steam generator nodes for the superheat region of steam in the  !

generator. The variable node capability is applied to both the primary and i secondary side. At each time step durin6 the tube uncovery, the modified '

LOFTRAN code makes a general evaluation of the uncovered tube region (e.g.

steam flow rate, uncovered tube heat transfer area, estimated heat transfer coefficient, etc.) and determines the number of nodes to be used in the subsequent calculations. The total heat transfer for the uncovered tube region 'is determined and accounted.for in the primary temperature transient

e calculation. The superheat/ tube uncovery modeling is applicable to all 1 steam generators.

Figures 3 through 6 show typical results for a 0.86 ft steamline break

, from 102 percent power using the modified version of LOFTRAN. Figure 3 shows the fraction of tube uncovery versus time with uncovery of Loop 1

(faulted) starting at 152 seconds into the transient. At approximately 300 seconds, the uncovery transient reaches an equilibrim point where the steam flow out of the steam generator matches the auxiliary feedwater flow into the steam generator. Additionally, the tube uncovery transient for Loop 2 (non faulted) is plotted but shows no tube uncovery for the entire

transient. Figure 4 presents the steam flow transient for this case.-

Figure 5 includes plots of both the superheated steam enthalpy and the saturation enthalpy for the Loop 1 steam generator. Figure 6 includes the Loop 1 temperatures for.the steam generator tube inlet (primary side),

j steam exit temperature (superheated steam), and the saturation temperature j for the steam pressure.

1 C. NOTRUMP Model Comparison The NOTRUMP computer code (Reference 4) was used to verify the LOFTRAN modeling of superheat. The computer code was originally developed to analyze transients of secondary systems with two-phase conditions. In the past, it has been used to analyze various transients in the primary and-secondary coolant systems. .NOTRUMP has recently undergone major revisions to enable it to model non-equilibrim nodes (i.e., separate liquid temperature and steam tempera'ture modeling). Using NOTRUMP, the steam generator can be broken down into sufficient nodes to model the nonequilibrim effects of the steam generator, as well as the tube region during uncovery. NOTRUMP can model all modes of heat transfer associated with a steamline break transient, including-heat-transfer from.the uncovered tubes to the superheated steam and the feedback effects between the primary and secondary sides. The two phase mixture level calculation accounts for primary to secondary heat transfer and the swell associated with rapid depressurization of the steam generator during the blowdown.

A ' comparison of LOFTRAN and NOTRUMP blowdown results is presented in Figures 7 and 8. The mass releases shown in Figure 8 show excellent 3

agreement. The LOFTRAN prediction of cuperheat enthalpy'is ~slightly higher than NOTRUMP, while the predicted time of tube uncovery is somewhat later.

NOTRUMP shows a chugging effect during the uncovery phase.of the blowdown.

This is believed to be in part due to oscillations in the flow link.between the downcomer region and the steam dome region. ~ (The flow link is the

. drain path for the moisture separators to the downcomer region.) . With the flowidirection towards the downcomer, superheated' steam goes cinto the

downcomer region and is condensed.;- This alternates with a flashing'of a i portion of the water 1 volume in the.dcuncomer region. This raises the pressure of the downcomer, resulting in a flow reversal in the link with - _

saturated steam from the downcomer mixing with the superheated. steam in;the.

dome._ This mixing results in the variations in the superheat enthalpy'seen in Figure 7. Although'LOFTRAN does not show the enthalpy variation =since r

the detailed modeling of-the-downcomer and. dome are not included, the overall agreement with NOTEUMP is very good.'

e.

s .

o o

D. Effects Of Analysis Assumptions The effects of sucerheated steam are dependent upon the occurrence and extent of tube uncovery. The major parameters affecting tube uncovery are:

initial steam generator inventory, auxiliary feedwater flowrate, assumed feedwater system failures, and protection system errors. Variations in these parameters are in the process of being evaluated for their effects on the containment temperature response (Figure 9).

Refinements in the mass and energy release modeling (Figure 10), are being evaluated and several areas show a potential for reducing the degree of superheat being generated. Scme of these areas are:

- Evaluatien of liquid-steam interactions such as the phenomenon of tube support plate flooding and heat transfer across the tube wrapper from

. the superheated steam to the auxiliary feedwater f1cwing down outside the tube wrapper.

A mere detailed steam header model in LOFTRAN.

- Fbdeling temperature drops in the primary superheat nodes.

- Evaluating other void correlations for use in predicting tube uncovery.

W l

l'

l l

l III. Containment Mode, ling A. Description of Containment The general phenomena taking place inside an ice condenser containment during a steamline break transient can be described utilizing a typical ice condenser

~

elevation drawing (Figure 11). Steam is discharged to the main (or lower) compartment where heat is removed by the internal structures, steam flow to the ice condenser, and the ice condenser drain water. The dead ended compartments are the regions which are located belcw the ice condenser and outside the crane wall (Figure 12). Air is discharged from the main compartment to the dead ended compartment and ice condenser so that the resulting steam to air ratio is that region is much higher than in dry containments. At ten minutes following the containment hi-2 signal, deck fans are actuated which direct air flow from the upper compartment to toe dead-ended compartments. Most of the safety related equipnent is located in the dead-ended compartments although some equipment and cabling are located in the main compartment.

B. Containment Models Figure 13 outlines the major models and assumptions utilized in the LOTIC-3 containment code. In the currently approved version of LOTIC-3 documented in Reference 5, four distinct regions of the containment are modeled; the lower ccmpartment, the dead-ended compartment, the ice condenser, and the upper compartment. Two condensate /revaporization models are used depending on the size of the break. For large steamline breaks,100% condensate revaporization is assumed. For small steamline breaks, a convective heat flux model is used which calculates partial revaporization during the transient. The wall heat transfer model utilizes the Tagami heat transfer correlation for condensation heat transfer and the convective heat flux model derived from the work of Sparrow (Reference 6) which calculates the convective heat transfer for small steamline breaks. The sump recirculation system is only modeled for the large break LOCA transient containment response.

Figure 14 shows the four regions modeled with the ' mass and energy flows that can be assumed in the analysis. The Catawba nuclear plant does not have lower compartment sprays and they are not modeled in the analysis. Superheat heat transfer is conservatively assumed to be zero for the steomline break containment analysis. In the model described in Reference 5, wall heat transfer is not modeled in the dead-ended compartments although these regions do contain structures which will remove heat. The analysis does include the upper compartment sprays, flow through the ice condenser, deck fan flow, and flow to  !

the dead-ended compartments.

LOTIC-3 solves the conservation of mass, energy, and momentum equations for upper, lower, and ice condensor regions (Figure 15). After the new lower compartment conditions are determined, conservation equations are solved for the dead ended compartment and the flow rate between the compartments is determined.

Figure 16 presents a typical steamline break containment temperature transient that is calculated using superheated steam blowdowns from the LOFTRAN code and the modeling of ice condenser drains as a heat removal source. The transient shows that initially the contaircent temperature increases rapidly during the

O o

blowdown. When the upper compartment sprays actuate there is a slight decrease in the main compartment temperature. The temperature then rises slowly until ice condenser drain flow decreases to the point at which time the temperature begins to rise again (approximately 250 seconds). This rise in containment temperature coincides with the steam generator tubes uncovering at 152 seconds and the maximum superheat occurring at approximately 250 seconds. The steam generator level stablizes whdn the auxiliary feedwater flow is equal to the steam discharge at approximately 300 seconds. The containment temperature then starts decreasing with decreasing decay heat. At ten minutes, the deck fans

_ actuate which results in a rapid decrease in containment temperature.

C. LOTIC-3 code Modifications Four rndificatiens have been incorporated in the LOTIc-3 containment model which are (Figure 17);

1) wall heat transfer model

2) convective heat flux model

3) ice cendenser drain model

4) dead-ended compartment model D. Mall Heat Transfer - - - --

The modificatien to the wall heat transfer model is described in Figure 18. In the LOTIC-3 model, only condensation heat transfer, utilizing a Tagami heat transfer coefficient and a temperature difference between the wall and saturation, was previously modeled..' The modification includes a convection term '

with a conservative convection heat transfer coefficient and a temperature difference between the contafnment atmosphere and an appropriate interface temperature. The Appendix presents a more detailed description of this model.

E. Convective Heat Flux The modification to the convective heat flux model is described in Figure 19. A tenn has been added to the convective heat flux model to account for the

• feedback effect from including a convective term in the wall heat transfer model. The Appendix presents a more detailed description of this model.

F. Ice Condenser Drain Model In an ice condenser containment there is approximately twenty drains exiting from the ice condenser into the lower compartment at an elevation of about forty feet above the compartment floor. The drain pipes are one foot in diameter.

The drain flowrate is calculated by the LOTIc-3 containment code. For a typical small steamline break transient the drain flowrate varies from approximately 4000lbm/seeto500lbm/secduringthe$1meframeofinterest. The temperature of the drain water is approximately 130 F (Figure 20).

Figure 21 presents the assumptions and the basic model used to estimate the heat removal from the lower compartme% atmosphere to the ice condenser drain water.

It is conservatively assumed that the drain water stream does not break up prior to reaching the floor even though many of the drains have equipnent and' structures located below them. Therefore, heat transfer is assumed to occur 1-

the stream surface only. It is also assumed daat the stream surface temperature is at the saturation tenperature of the containment.

The heat transfer to the stream is:

q=hA/LT wher'e h = condensation heat transfer coefficient A = surface area of the stream AT = appropriate temperature difference The calculation of the heat transfer surface area is described in Figure 22.

In order to model the drains in LOTIC-3, the drains are modeled as a wall heat sink with a surface at a constant temperature (see Figure 23). Currgntly,in the version of LOTIC-3, the surface temperature is assumed to be 230 F which is close to the containment saturation tenperature. The drain surface area is calculated at two points in time during the transient; early in time with a high flowrate and later in tbne with a low flowrate. To ensure conservatism in the area calculation a 10% reduction of the surface area was assumed.

As described previously (Figures 14 & 15), the LOTIC-3 containment model did not account for wall heat removal in the dead-ended compartments. To obtain a conservative estimate of the temperature transient in the dead ended ccmp. etment, the heat sinks located in the dead ended compartment region along with the heat sinks in the lower compartment are modeled in a combined volume (see Figure 24). This " modified" lower compartment model is used to determine a conservative dead-ended compartment tenperature transient. Since the lower compartment will be hotter than th'e- dead-ended compartment, this methodology results in a higher temperature in the dead-ended compartment then would be expected. -

G. Transient Results With the modifications described for LOFTRAtl and LOTIC-3, the previous FSAR limiting case for Catawba was reanalyzed to determine the impact of superheated steam. Thecaseselectedisa0.86squarefootbreakat10g5 power (Figure 25).

The peak lower containment temperature for this case. is 324 F. This tenperature is calculated for the lower compartment only. It is expected that the -

dead-ended compartment temperature will be significantly lower.

In addition to the 'model modifications incorporated in LOTIC-3, Westinghouse is pursuing further bnprovements in the areas noted on Figure 26. One area is in the wall heat and mass transfer models. Since condensation is a mass transfer type phenomena, the heat and mass transfer should be linked. This approach has been used in Reference 7.

An unproved drain model is also being investigated. This improved model will calculate the drain surface area as a function of flowrate.- It will also calculate the average temperature rise of the drainwater. This model will more accurately represent the actual phencmena in the containment.

- . . . - __ - __. - _ - - - - . . . . - . . . . . - ~ . . - _ . . .

4 WESTINGHOUSE PROPRIETARY CLASS II l V. Appendir WESTINGHOUSE STEAMLINE BREAK  !

.i BLOWDOWN AND CONTAINMENT ANALYSIS METHODOLOGY The following sections describe the Westinghouse methodology for determining the  !

containment response fer a steamline break incorporating the effects of

> superheated steam. These sections describe in detail changes from the j i methodologies described in References 1 and 5.  ;

I. Steamline Rupture Mass / Energy Blowdown Analysis A. LOFTRAN and MARVEL Computer Modeling Mass / energy releases can be calculated using either the LOFTRAN code 1 (Reference 3) or the MARVEL code (Reference 8). The LOFTRAN code is used for non-LOCA FSAR accident analyses. The MARVEL code was specifically developed for assymetric transients such as steamline breaks. These two codes are very similar because they were developed in an interrelating fashion and much of the model'ing is comon to both codes. The MARVEL code was used in the development. of Reference 1 because LOFTRAN at that time was a lumped model which was used for symmetric loop transients. Furthermore, for steamline break analysis purposes, MARVEL.contains a.model for water entrainment. However, the current version of LOFTRAN is a multiloop version which also contains a water entrainment model. With the development of a multiloop version of LOFTRAN and the inclusion of an entrainment model, the use of MARVEL has been generally discontinued. This enables the use of LOFTRAN as a single system analysis code for non-LOCA transient analyses. LOFTRAN is used in the analyses presented here..

t. The model of importance to blowdown calculations is the_ steam generator model. The primary side-of the. steam generator' contains multiple nodes to model the tube bundle for both the modified ' version of LOFTRAN and MARVEL.

Heat transfer calculations from the primary to secondary side are identical' in the two codes, although the methods for, initializing the heat transfer resistances are slightly.different. The secondary side is effectively a one noce, two. region'model of saturated steam and water. . Heat transfer is

assuned to occur to saturated water. If tube uncovery is predicted, the amount of surface area available for heat, transfer is reduced.

Both codes contain a detailed steam generator model which is used to predict' tube uncovery. This model calculates the liquid volume inithe

-steam generator- shell and acgnts for the detailed steam generator geometry. The [ .1 correlation is used in both codes to i predict the voiding in the' tube region, although the correlation is modified for use in LOFTRAN. In MARVEL, tube uncovery is calculated based D e

. .z on comparison with the actual water level and the height of the tube bundle. In LOFTRAN, the user specifies eitner a water volume in the steam generator ccrresponding to tube uncovery, or a void fraction in the riser section of the steam generator at which tube uncovery begins.

Both codes have similar mcdels accounting for reverse heat transfer, thick metal heat transfer, feedline flashing, and safety injection system operation. Auxiliary feedwater flow can be input as .a fraction of nominal feedwater flow, although LOFTRAN has an additional capability to model auxiliary feedwater flow as a separate system. For analysis of double ended ruptures, MARVEL accounts fer the volume of steam in the piping downstream of the steam generators in the blowdown calculations. In LCFTRAN, this consideration is added en to the blewdown mass and energy results by hand. For split ruptures, which the analysis presented here addresses, the steam piping masses are handled identically in both codes.

In summary, LCFTRAN and MARVEL are very similar codes, and either can be used to calculate mass / energy blowdowns. To demonstrate this, a comparison of the blowdowns for a typical case is presented in Figures A.1 and A.2.

Figure 1 presents the mass release rate fer a .86 ft2 split rupture from 102". power. For this case, Figure A.2 shows the saturated steam enthalpy as a function cf time. This blewdown is typical of results used in FSAR analyses prior to the modification noted in this report for the LOFTRAN code. As can be seen from the figures, the results are extremely close..

B. LOFTRAN Model for Superheated Stean As mentiened previously, the LOFTRAN code has been modified to model heat transfer which may occur in the uncovered tube bundle region. This effect is modeled in both the faulted and intact loops. In the modified version of LOFTRAN, all heat transfer occurring in the uncovered region is assumed to add superheat the steam exiting the steam generator. The temperature of the primary coolant ficwing through in the uncovered tube region mode is conservatively assumed to remain constant. Realistically there would be a drop in temperature due to heat removal to the secondary side, but this will be small due to the low specific heat capacity of the steam and due the high primary side flow rate.

Theheattransfercoeffich'r ged in the uncovered tube region is based on the [ The heat transfer coefficient (U) is calculated by'the following expression:~ ,,,

<- a,c This correlation is presently used fo5'guperheated forced convection heat transfer by the_[ ] computer codes. Additionally,

. .a this correlation is based upon the heat transfer from the surface of the tube wall to the a'verage bulk temperature of the steam. In the LOFTRAN modification, n6 credit is taken f.cr either a primary film heat transfer 4 resistance er a tube metal heat transfer resistance. Therefore,the wall temperature of the tube is conservatively assumed equal to the primary fluid temperature.

g

'J (1) e J

e

=*

e

- 0 ,

e e

l

\

l f

~

Tne modified version of LOFTRAN automatically selects the proper number of steam generator nodes for the superheat region of steam in the generator.

, The variable node capability is applied to both the primary and secondary side. At each time step during the tube uncovery, .the modified LOFTRAN code makes a general evaluation of the uncovered tube region (e.g. steam flow rate, uncovered tube heat transfer area, estimated heat transfer coefficient, etc.) and determines the number of nodes to be used in the

subsequent calculations. Each node is evaluated to determine the steam temperature exiting the node with a convergence criteria that is based upon the total number of nodes used. The exit steam temperature of one node is used as the inlet steam temperature of the next nede.

The heat transfer calculation to determine the outlet temperature of the node is based upon the following expression:

Q = UA*(Tpri-(Tout + Tin)

• Ns*Cs *(T out Tin) -

ubere Q = Heat transfer to the steam U= lM T 1 = Primary node temperature Tout = Steam node outlet temperature T in = Steam nod'e inlet temperature' M3 = Mass flowrate of the steam x

C3 = Heat capacity of the steam A = Heat transfer area in the node including both hot and cold leg sides of the tube bundle The total heat transfer for the uncovered tube region is determined and accounted for in the primary temperature transient. i C. Blowdown Sensitivity to Plant Conditions

\

l

The effects of superheated steam are' dependent upon the~occurrance and  !

extent of tube bundle uncovery. . Parameters affecting tube. uncovery are: 1

- initial steam generator inventory, break size, auxiliary feedwater . l

~ flowrate, and the single failure assumed.  ;

The initial steam generator inventory depends upon the measurement errors associated with steam generator level and upon-initial power level. Steam generator mass increases with decreasing power, thus, breaks intitiating from low power levels will result in.later tube uncovery.

, Larger break ~ sizes result in faster blowdown of the steam generator and -

. earlier tube uncovery, i

0-f /

-r w e e- -

etv--y < .*e e + * * , = - + 9 e-' - ew* *C- re P

1

. s- .

Large auxiliary feedwater flowrates only delay tube uncovery, but will also cause the final equilibriun steam generator level to be higher. This equilibrium condition corresponds to the point when the break flow rate is equal to the auxiliary. feedwater flow rate. l

\

The single = failure assumed in the transient may impact the amount of water j supplied to the steam generator. Auxiliary feedwater runout will increase the amount of water supplied to the steam generator. Failure of the feedwater isolation valve will also cause extra water to be supplied to dne generator as the additional mass between the isolation valve and the check

~

valve flashes to the generator.

I i

+

.I l

l I

l 1

l

II. Containment Analysis ,

A. Wall Heat Transfer Model The original LOTIc-3 wall heat transfer model is based on the stagnant Tagami heat transfer correlation. That is,

• q"=hTAGAMI( SAT ~ WALL}

hTAGAMI = 2 + 5014 STEAM " AIR h(TAGAMI, MAX)=72 BTU /hr-ft - F This model was developed for saturated steam in the presence of large amounts of non-condensable gases. In the lower compartment of an ice condenser, most of the air is swept out of the lower compartment through the ice condenser and into the upper compartment. Therefore, after about 30 seconds, there is almost no non-condensables in the lower compartment. Typical values for the condensation of pure steam are in the range of 1000 to 3000 Btu /hr-ft2 oF (Ref. 5). The correlation used in the modified LOTIc-3 code is in e:: tension of the Tagami correlation fer nearly pure steam.

q"=hCOND (TSAT-TWALL) hcond = 2+50 MSTEAM " AIR h(cond, max)

• A mnximum value of [ l a,c was chosen as a conservatively low concensing heat transfer coefficient in a nearly pure steam environment.

In addition to this modific'ation, an additional term is needed to account for the convective heat transfer from the superheated steam to the condensate film.

This convective heat transfer is dependent upon whether there is condensation occurring on the walls. If condensation is occurring, the correlation used is:

" "y h conv IIbulk-T3 ,g) where:

If the wall temperature increases to above the saturation temperatu're then the convective currents will be reduced such that the correlation used is 9"conv=hconv(Tbulk-Twall) t l where:

i

! [ ja,c ,

Th'us in sunnary, if T w,11<T sat then

-[ j a,c If Twall > Tsat, then the correlation used is:

[ j a,c 6

0

i

'B . Convective Heat Flux Model i

When the containment atmosphere is superheated, the containment temperature is a strong function of the amount of steam mass in the atmosphere. Thus the amount of mass condensed on the heat sink surfaces is a key parameter. The actual amount of condensate fonned is -

"cond

• 9cond h7g Unfortunately, with the use of a heat transfer correlation based only on test data (such as Tagami or Uchida), only the total heat transfer coefficient is obtained. This total heat transfer coefficient includes both the condensation heat transfer and the convective heat transfer. Based on the work of Sparrow (Reference 6), the Westinghouse Convective Heat Flux model in the original LOTIC-3 code calculates the ratio of the convective heat transfer to the condensation heat transfer. Therefore the calculation of the amount of mass condensed is

[ ja,c In the modified LCTIC-3 model, the, anount of superheat convection is calculated.

The amount of convective heat transfer at saturation is not known explicitly in this model. Therefore, in the modified LOTIC-3 code the original convective heat flux model will be used to calculate the fraction of convective heat transfer for saturated conditions. The actual correlation is s,c 4

where, (q /q is determined from original convective heat flux model and q conv is eggg)$$ bunt of convective heat transfer calculated in the wall heatEFEESfbrmodel In summary, the modified LOTIC-3 model is consistent with the original LOTIC-3 model in its calculation of the mas condensed. The only difference is that in the modified LOTIC-3 code, the amount of superheat convective heat transfer is known explicitly, while in the original LOTIC-III model, only the ratio of convective heat transfer to condensation heat transfer is known.

6

, .. o

. . . =

IV.

References:

1.. Land, R. E., " Mass and Energy Releases Following A Steam Line Rupture" WCAP-8822 (Proprietary) September, 1976 and WCAP-8859 (Non-Proprietary).

2. NS-EPR-2563, February 14, 1982, E. P. Rahe of Westinghouse to J. R. Miller, NRC, " Additional Infcrmation on WCAP-8822".

3 Burnett, T. W. T., et al. , "LCFTRAN Code Description," WCAP-7907, June, 1972 (Proprietary).

4 Meyer, P. E., and Kornfilt, J., "NOTRUMP - A Nodal Transfer Small Break and General Network Code," November,1982, WCAP-10079 (Proprietary) and WCAP-10080 (Non-Proprietary).

5. Hsieh, T. and Liparulo, N. J., " Westinghouse Long Term Ice Condenser Containment Code - LOTIC-3 Code," February, 1979, WCAP-8354-P-A Sup. 2 (Proprietary),WCAP-8355-NP-A (Non-Proprietary).

6. Sparrow, E. M., Minkowycz, W. J., and Saddy, M., " Forced Convection Condensation in the Presence of Noncondensables and Interfacial Resistance", Int. J. Heat Mass Transfer, Volume 10, 1967

7. Corradini, M. L., " Turbulent Condensation on a cold Wall in the Presence of a Non-condensable Gas" Nuclear-Technology Vol. 64, pp 186 - 195, February, 1984. ,

8. Krise, R. and Miranda, S., " MARVEL - A Digital Computer Code fcr Transient Analysis of a Multiloop PWR System," November,1977, WCAP-8843 (Proprietary) and WCAP-8844 (Non-Proprietary).

9. McCabe, W. L., and Smith, J. C., " Unit Operatons of Chemica1' Engineering",

3rd Edition, 1976.

O e

LOFTRAN - MARVEL COMPARIS0N

.860 FT2 BREAK AT 102 PC POWER 2000.0  :  :  :

1750.0

/f h

d e

1500.0 - -

s 9 --

1250.0 - -

x S --

"1 1000.00 --

t.a --

(7> 750.00 - -

x c: 500.00 - - MA R V E L --y /

LOFTRAN --

cc 250.00 - -

12/05/83 0.0 & 4 i & .i O

0 O

O o o o O b b b b b O O O o o o

~ N m .:r D tO O

TIME (SEC) .

1 i

FIGURE A.1

~

.. o 1

l LOFTRAN - MARVEL COMPARIS0N

.860 FT2 BREAK AT 102.PC POWER .

1300.0  :  :  :  : .  :

, - 1275.0 - - --

5 s

e a,., 1250. 0 -

1225.0 - - --

0 RAN 1200.0 - - __

5 1175.0 - - --

x .

W 1150.0 - - --

w S 1125.0 -- --

1 i

l 1100.0 i  ; i i i c O O O O O O O b b b b b O

1

. O O O O O 1 O - N M e O O 1

TIME (SEC)

FIGURE A.2

i l

8 l 4 .,

8  %)

METHODOLOGY FOR ADDRESSING SUPERHEATED STEAM RELEASES IR ICE CONDENSER CONTAINFENTS Purpose ,

The purpose of this report is to document the information presented on March 19, 1984 in a meeting with the U.S. NRC Containment Systems Branch on the status of progress made in addressing the confirmatory- item on the Catawba Nuclear Plant ,

Safety Evaluation Report. This confirmatory item deals with the effects of superheated steam generator mass and energy releases follcwing main steamline break accidents. Attachment 1 includes the list of attendees at the meeting and the overhead slides covered in the Westinghouse presentations.

Technical presentations were made describing the modeling of the steam generator and heat transfer from the uncovered tube bundle during the steam generator blowdown along with a description of the containment model and transient response. A proposed plan of action was also presented and discussed with the Staff. In accordance with that plan, this report represents the first milestone in the proposed plan of action. As committed to in the meeting, the appendices present proprietary infermation which relates to the specifics of the models and sensitivities that were not directly addressed in the meeting.

Attachment 2 is an explanation of, and refers to, the overhead slides (Figures)

p. resented at the Fbrch 19 meeting.

Y e

C t

• 43 ATTACHFEllT' 1 m*

e l

i 1 -

i 1

/]Huda-w o >- 3//7/8f .

4 .62 .,

' /1c_e u g. L'/ .Z:L./ .

iL.E nu L '~.- 7.~.ae A~9.i s hla g . O<.g a ,3- Ha, u

- -- y

}{, A/ 3~ A GEOUR.

WEC /DL

' l2 '. O .* 3/MtzP F--. Du:U /"Pb a' L. Li rr a

@ paura- s++rr'

% .V. To u -sc E coc sAe SurT7 Ah" c. /1L

~! f )(e a bS. Lotte W NuwL GAFGYY

, M, F 65coc 4 V nud e

?A.Lss i

' W /4 e-IG i- Jir-!<y L . 2. A/cdini"- h' .Uaz,6,,:,e.

f b C-A D U M .kl., it) rela r- /<< Ct' r~y r r %x.q. og moo , pg( my. -

DnI<.e. , C<s::~ 5m - ~.

S .D Afe. M v A 5 $r'llv b dc., ><r:7 [y :sw:

7% u,- %s , L-

- c. E. ca.ua Sited Locu,smc .trgy , Ec;P:

E. AntLt C.'b(q' uc.% 1 Q a amt.sa=y .

• T V A, G ,sy; ..' , *

• b n op W. O. Cr.au& .

S. ?. .v e.6 e 0.u. , K e, t..t< k- s?,c.,cb< a.! -

Ni D SME

~

W

.$~ / ibuam _,

g "rs / sura recry

2.H.Ow.e J .SER4KWC- Mbb G&0' t,

'9A4 Q ,\ C.

?'l.R,k.M-

/R O.CSB NGC 4 cc C. Li A. Mah.0.caceso Aelc/c sr3 9

4

~

LOFTRAR MCOEL l STEAM GENERATOR SECONDARY ,

STEA6A Ft oW .

~.

~ .

,- ,._ g SATOR.ATED i

. . , . , , . .* d STEAM RE6ted sxwaaren '

~

e Y.,a

~

~'. .

w g m m s g ff . . - -

Z(

l f_ ~

l

.t v

~-9.ny ..:.

(:. . .

- . .y ,,

• StMGLE NCDE , 2 It.Esicd MCOEL

• WO HEAT TRAdSFER. To SATueATED STEAR REGleM l

• HEAT TRAMSFER To SATORATED WATE.R RE.Gncd iS MODLF1ED R:: Sit.

. TVSE 'UNCcN E1ECT ' .

FIGURE 1 -

l 80 g

m -

.~. ,

~~

l4 . .

LOFTI(Abi MODEE~

suPERHEAT HEAT TRANSMR l

I suPERHEATEb STEAM Fl.cW SEcoHDART SuPERHEAT Mote.S m\ gfos, PRIMARY opensaAT acces l

3

'9N8 8

.h}i.s -

t.rr = CTIVE '

UTCOVERY LEVEL ~

• VARt A1!R.E. 50PER.ME.AT '

McDES MODEL

= coHSTART PRtMART TEMPERAiululi tu. Supt.RMEAT 125.Gion AssuMEb FCB.

HE.AT NSFE3a. cALcutAttcHS .

• C ALCVLATED SOMW*1AT HP.AT vgAss72R AcccONTE,b FCR. IM l , ppuAART TRAd1inWT ciant 2 '_ - . --

TUBE UNC0VERY LOFTRAN SUPERHEAT MODEL

.860 FT2 BREAK AT 102 PC POWER 1.2000  :  :  :  :  :

G E 1.0000 - -

LOOP 1

-~

E p

y .80000 -- --

E M

a .60000 -- --

S 5

> .40000 -- --

8 '

8 2

a .20000 -- --

m S LOOP 2 0.0 8 8 8 8 8 o s s s s s s s S 8 8 3 8 8 TIME (SEC)

FIGURE 3

1

• e e

d

, MASS BLOWDOWN LOFTRAN SUPERHEAT MODEL

.860 FT2 BREAK AT 102 PC POWER 2000.0 '  :  :  :  :

i f

l 1750.0 i

n d

m 1500.0 - - --

I N m

1250.0 -- --

2 i S "z 1000.00 - - --

w

$ 750.00 - - --

x E m 500.00 - - --

^

250.00 - - --

0.0 o o

o o

-oo o o o o -

o o O b b d S 8 8 8 8 TIME (SEC)

I 1

a FIGURE 4 v

m w- e r- ww g9

i l

i l

1 i

ENERGY RELEASE LOFTRAN SUPERHEAT MODEL

.860 FT2 BREAK AT 102 P,C POWER 1300,0  :  : -:  :

1275.0 - - --

3 SUPERHEATED STEAM s

a 1250. 0 - -  ;-

g 1225.0 - - --

Y w

$ 1200.0 -

~

c SATURATION

+

5 1175.0 - - --

z id 1150.0 - - --

m S 1125.0 - - --

1100.0 8 8 8 8 8 8 o d 5 5 5 5 5 6 2 8 8 8 8 TIME (SEC)

FIGURE 5

l 1

TEMPERATURE TRANSIENTS LOFTRAN SUPERHEAT MODEL

.860 FT2 BREAK AT 102 PC POWER 650.00  :  :  :

1 600.00 -J --

U E

$ 550.00 - --

c.

8-k t

SG TUBE INLET 5e 500.00 -- (PRIMARY SIDE) --

r ,

• b 450.00 - -

5$ SUPERHEATED STEAM E~

h 400.00 - - --

E 350.00 -- --

SATURATION 300.00 8 8 8 8 8 8 a a a 5 5 5 5 6 3 8 8 3 8 8 TIME (SEC)

FIGURE 6

~

q' 8w

^

8e I

8- )

C E

S I

I E

M I

T N ~

' 0 -

S I f 4 )

k8m R A M A R U P

M T R O F T C O O I

I L N P YP I M L U A RH TT f O NE k 8~

N N K AA R ER T B F

O L

/- E.8-V-

9o 0 0 0 0 0 0 0 0 5 0 5 0 5 0 5 0 7 5 2 0 7 5 2 0 2 1 1 1 1 3 2 2 2 t

. I 1 1 1 1 1 1 1

?8 ta@tE -

3oErn" O

, O58 O58 O5? 3 C

E S

(

E M

I T

N 0 N P S A M ,

I R -

, R T R l

l Sa2 A F T P

M O O O L N CE T

PAR M

UW RO TL OF O&2 NK A

NE AR N RB T

F O

L 8aS

\

• O

\

od -

0 0 0 0 0 0 . /

0 0 0 0 0 0 0 0 0 0 5 0 5 0 0 2 2 1 1 5 0 naJEx$

u g<* g  :<r

Ng i on ;

3eg*

l i

1 l

EFFECTS OF ANALYSIS ASSUMPTIONS i

l INITIAL STEAM ~ GENERATOR INVENTORY AUXILIARY FEEDWATER FLOWRATE FEEDWATER SYSTEM FAILURES

> PROTECTION SYSTEM, ERRORS E

t FIGURE 9

l

. s. . .

1 ADDITIONAL MODEL CONSIDERATIONS t

i LIQUID-STEAM INTERACTION i IMPROVED STEAM HEADER MODEL 9

HEAT TRANSFER THROUGH TUBE WRAPPER e

TEMPERATURE DROP IN PRIMARY SUPERHEAT NODES 4

OPTIONAL V0ID CORRELATIONS 1

i O

t t

i FIGURE 10 S

y ,..r --

, , m rr _ , w - ., , f. .# A ,

o -

o SM/ ELD BUILDING DCME Q

. . /

.j. .

.,[,1

-. . , g -

'~$7 EEL CONTAINMENT vf33EL . H 829.0c N;

f 'I' IO" #0LA# '#^"'Y.A P 3'HIELD BUILD /NG *

( . POLAR CRANE .

WALL m

• e, x x .*-

TOR OF QECX TOR OF CRANE RAIL

• 3PRING LIN.

h^' 2 ' ffL 797/2 ICE CCACING ';

f fEL 796.63 J } l "

/ G 791.3 l.e CYCLONE

"'"

• U .*

L> AIR VENTILATIQM g

'Qf V~ '*I1l BR/DCE CRANE .

,_ ', N,ggu

.,. Q ps.od "N dt* 6 *R U"II"( ,

\

f . uMIT3 TQR CF ICE BEC -

l l '. .e ,e-EL 740.0 %_. p (0!v10ER BARRIER ,

Et 778.ggy , , _

.. . ., e .. e. . j .

. %-~ cR.<xe

.. Et ru.,E-

- N --

y-

'p ICE 293:.:l .

wjz4 ,, g r:ft.,.3 1 ; .

CONCENSER -&. .l, l

\ ,-

e y y uCNrA.NueNr .

aqu5.se 8

e q' .

, p:I:\

l STEAM ..y .,

.CENERATOR t

• CONTROL ROC DRIVE, j

• 1. * #/3J/LE JN/ELO -*

sq i. : .. ,  : i s : .s as)

=o.--- KE CONC l '

.Qi '**"

p, i J. ) (ELisus m'a

\ L I. ,

jXR il I sQ -w .1 M i N L

. 5?!I.s((:

.  :. .n i ...

i -

' ' ' _G #' # dI BARRIER

' '*0 "

m ; gnn'

'#'0"

• N7 .;

Rvun .,

ACCuNuuron-kD

. , . n .- .

P*  % > . .* O -*

CRADE \ *. I tl  % L .

'$ }f F. .LOOR EL 102.0 EL 705.0 '%

- n.., _ .

e

,p

.c { }

~

_f l

..l

• ?

0 a.

I?

,,,, y VENT FAN r"*

j,.y<,a }.

Q ,/ '. t Gi:li;h EL 695.0-; ., k4 .e *g, D l{  : ,*,

{. & ECPT a ass.0 g -

. y '.'

,"G - ' 1  : b' * . . . . .-

. . .i *:

' l .*  !' .

.e. '* ...

]j  ; . f. .,

ee .*. .

'J'.o " g 'f,

."d.;

. Y s" EL 479.78 *-

- I*

p .

., g u . . .

J==m. . . y , .l .*.'

. , /5

-Q

-- Lsrfu ukiR .

. .MafiCr0R - ~

1 . . -

=

4. .

% ANCNon actT .. 1 ..

l

,] .

I ff" = 0 3 0 ' 8

j a ssl.ss y .

. su'".R! , ,-

Reactor Building Elevation g"% - ,

FIGURE 11

l

'[

1. 1. W '

%A UX/LM RY BLOG m '.h r .

,; ,'b,' ~4 .,.'.s FUEL. Pir  %. ~ .. .; ,

.4. . 4-. _.

f .;, .. .

w

..c.

-) &

FUEL TRANSFER

~ - PIPE CHASE * '

<.e. .*. u SYSTEM .%.t' .' . : W. . u. .

~W :e 6 .,,a. '.&, , ' O.$

EL ESJ.0 e- ,, ,

k,c,, ,

.a -.-

~

.. REACTOR UPPER **

v. s *:-

. *

• s 1

.p. '

• b

> INTERNALS SMRAGE  %

^

EL 690.gg. .~%n j'. .

~

YREACTOR n' _

p 'r.*

• r s, ,-

v. a . , .s-

. pj ,- .'

9y . %~ %:' . * , CO OL A,N r ,

PULlP.: lJ)

.q- s '* : :\ W /! ~ '"

' ....A... A Q. // '.

% ' v' .G '

f'LINTERNALS OWER\

• i.

s *-

"' s

% e','s, SmRAGE y [g, ,-

4.. 41 cy  :.- .

..i........, .i.

..y o . ,.

' jf is Ahh \

,5 *

.T 2- a. - .

.':CA TE' r, r STEAM GEN (4)

'E

,m '.*;'

.' , a 0 ( J

's. . ' '

MAIN K

,A,,f ' f.;e

. g

' i 1 ') e -_ REACTOR

'".. pagSSupg

~

r STEAM I  ;'

Q:. '

• ~~.7.,_ ',* - - l g .llq:

~

ESSEL < CRANE WALL r:,,'Wj k.

,. gt , p.

.c C>. - .

s': 702J5 ..

FEED- ,

WATER. Lv EL 679.78

' 'VENTILA TION

- ~

._ , RELIEF /

~

ANK v' '% .' UNITS (4)

Y. ,

.. PRESSURIZF' ~-y

. i l., * $Y ' }

• CUMULATORS (4)

- ~, ,

1 / o grj -

6 ..... .;.

EL 693.0

/ s HEAT EXCHANGERS R .,_ . ) (EL '7009]'W \ .

)

VfNTILATION UNITS *(2)?.-} ,.i' _ _

W l

,.e... * * . PERSONNEL LOCK Figure 3.8.3 7 Plan.t.o.er Compartment

'm -

• FIGURE 12

t LOTIC-3 CONTAINMENT CODE

• 4 N00E CONTAINMENT MODEL

• CONDENSATE /REVAPORIZATION MODELS LARGE BREAK (TOTAL REVAPORIZATION)

SMALL BREAK (CONVECTIVE HEAT FLUX)

• WALL HEAT TRANSFER MODEL

• MODELS SUMP RECIRCULATION SYSTEM O

8.

l l

l l FIGURE 13 1

I

s s r \

I 1

i i

i I

A Sorays [

a 4 4 t

, , - .3 , _

Swall Upper Compartment i I i*

' t m,h i A Y I 5 sI i U.C. .

m3 ,h34  !

Drain .

l .  !

! tee ms ' 's Gje, t Concenser m'h a a

j One way flow all - -

, through open area i

L,  ; on operating deck Condensate & 9 Ice Melt lm,h r 3

m,n 5I s I.C. Sumo a a

• I.CTi ,

Oraini  ;

Y i

l (RCS break #10w ,

• s ' "s l Dead-Ended h' h3

- accumu)a.-

w Ccmoartment m'n a a ' L:wer Compartment release, c.

t boil-oM)

~

lq drn,1c

~-

-Air Addition Taole Heat Acdition Table t

2

! L.C. U i'

r

'Orain sumo j wall l 5 -

• L- - d \

U.C. -

Orain j 'L.C.SumE )

M. ASS AND ENERGY FLOW DIAGRAM FIGURE 3.3 t

FOR THE COMPARU.ENTS FIGURE 14 d

LOTIC METHOD OF SOLUTION SOLVES CONSERVATION OF MASS, ENERGY, AND MOMENTUM FOR UPPER, LOWER, AND ICE CONDENSER REGIONS ONCE NEW LOWER COMPARTMENT CONDITIONS ARE DETERMINED, i CONSERVATION EQUATIONS, ARE SOLVED FOR THE DEAD-ENDED

, COMPARTNENT AND FOR THE FLOW RATE BETWEEN THE TWO COMPARTMENTS l

i FIGURE 15

. . .= .

TYPICAL CONTAINFENT T8PERATURE TRANSIENT (DRAINSf0DELLED)

1. LEVEL REACHED IN S.G.

-J U. . SPRAY - -[ DECK FANS ON t /

REDUCTION IN I.C. DRAIN FLOW

• ~

e 0

-b a>

9 I

0 100 200 300 E 500 600

. TIME (SEC.)

FIGURE 16

. l 1

i A

i MODIFICAfl0NS TO THE CONTAINfBT PDDEL 4

WALL HEAT TRANSFER MODEL l CONVECTIVE HEAT FLUX MODEL i

J.

4 ICE CONDENSER DRAIN MODEL r.

1 i

! DEAD ENDED COMPARTENT MODEL t

4 .

4

.l l

1

.. 1

-FIGURE'17 l

w n

. 4

. WALL HEAT TRANSER F0 DEL CRIGINAL L0flC MODEL 1

n"= h (Tur -

} ragami Tv.it )

1 M DIFIED l.0 TIC MODEL i*bcono :( Nr ' I ]

• w.it bc *'v ~tk Ier]

h =f(n%

Cono 1,_ , , ),

hc ,,,, = f(T Waif I T)

SAT L = fW.,, , x,)

FIGURE 18 k

M "

C0iWECTIVE HEAT FLUX F0 DEL ORIGINAL LOTIC MoDEL kcono - 3cono h,y

}rorn(1 4,, +x) l l

l ttDIFIED LUTIC MoDEL y = Scom =

Troen[! + X,,,][i h he.~I yr.rm.% - %l cono py hpg -

FIGURE 19 w w w -

ICE CONDENSER DRAINS

-APPROXIAMATELY 20 ICE CONDNESER DRAINS

-DRAIN ELEVATION IS ABOUT 40 1 FEET FROM FLOOR i

-DRAIN PIPE IS 1 FOOT IN DIAMETER

-FOR TYPICAL ftiLB TRANSIENT, DRAIN FLOW VARIES FROM I 4000 LB/S TO 500 LB/S

'l e

1 i-FIGURE 20.-

a+ v e e e c , e v +

1 ICE CONDENSER DRAIN PODEL

-CONDENSATION OCCLRS AT THE SURFACE' 0F THE STREAM

-FLOW IS WELL MIXED i

Q = h A AT

-FbDEL AS A WALL AT A CONSTANT TEMPERATURE

-A IS THE SURFACE AREA 0F THE STREAM

-h IS A CONDENSING TYPE HEAT TRANSFER COEFFICIENT -

l FIGURE 21

, e

• W - am A t' - A ,& - A r -=L, --x - + u, h

f 1

. e i

l

! l

}.

CALCULATION OF THE STREN4 FLOW AREA B = n (Pxi) = ao ('ex yo) = 800 p WHERE P IS THE PERIMETER OF THE STREN4

\;

/

i/ ~ --/ -

. x, .

b

,, - ,-- l l(- - -- .. . ._ 3 r

h

.x 4

Mm '

4 t

D l

1 1

FIGURE 22'

-oe-a- - - -- ,y .n , - - . -r ~mr, w-- p--.n,-- , + ,,,-r r, s-- y a-

- . l

. .o I

FDDIFIED LOTIC DRAIN M) DEL L

a .

-WALL WITH A VARIABLE AREA evu< -Tsar) 1

% = eh "o A (T N$$

t 4

4 hr= 230 F 1

i s

l-I f

s FIGURE 23

[ .

-, -,w - --, y y . c., r

1 i

DEAD ENDED COPPARTiUlT f0 Del 9

I I

I 100lFIED LOWER COPPfTIBIT _ q I .

L_.. .

e 4

e FIGURE 24

l CATAWBA RESULTS

- 102% POWER

- 0.86 FT BREAK

- MAXIMtfi AFW FLOW 4

- FSAR HEAT SINKS

- t%XIMlli S.G. INITIAL MASS

. T. = 32I4*F i%X (LOWER C0fPARlfBT) t i

h I

a 9

r i

FIGURE 25- .

4 s

- - , -v.r , n - + - - - r 4

,. 1 s

l i

A i

ADDITIONAL M) DEL CONSIDERATIONS k

I I

1, <

-WALL HEAT TRANSFER MODEL-t i

b i

! -DRAIN 10 DEL i

-DEAD ENDED COMPARDUT t0 DEL e

l i

a 3

1 l

e 'i 1

1 FIGURE 26- -

I i .- I

l .

f i

l ATTACHENT 2 ,

L QUTLIE OF THE' REPORT i

I. Introduction II. . Mass & Energy Release Modeling

~

I III. Containment Modeling 1

IV. Action Plan V. Appendix VI. References-a i

i e e, 6

4 A

i l

4 4

l 6

1. 5

+ r - p e. .,e. m4 r 1F * " " e*e- - - N &Wa-We # 1W' 5 e t-=r-4-6W--+M*4+= w rm---r-*W-"U#w -tv* 'e'

I. Introduction During the Containment Systems Branch review of the Westinghouse topical report,

" Mass and Energy Releases Following a Steam Line Rupture",WCAP-8822 (Proprietary) the Staff noted that heat transfer to steam from the uncovered portion of the steam generator tube bundle was unaccounted for and questioned the effect upon the calculated mass / energy release and the subsequent effect en the containment temperature response. Westinghouse responded in a letter to the Staff (NS-EPR-2563, February 14, 1982, E.P. Rahe to J. R. Miller) that it had determined the hnpact of the effect by conservatively treating the maximum amount of superheat to be the difference between the primary coolant temperature and the steam tenperature. The letter noted that there would be an insignificant effect on dry type containments and that, based on the conservative model used, there would be an expected increase in containment temperature for ice condenser type containments. In the Centainment Systems Branch Safety Evaluation Reports on the topical report and the Catawba Plant Safety Evaluation Report, the Staff required that a more refined steam line break analysis be performed to determine the effect en containment temperature which might impact the environmental qualification envelope used for safety related equipment.

Since that time, Westinghouse has investigated the effects of tube bundle heat' transfer from the viewpoint of a more refined modeling approach. Subject to the final review and approval of the NRC Staff, the efforts and results obtained to date indicate that there is little Dnpact on the containment response from the effects of the additional tube bun,dle heat transfer to steam.

G 9

• • 4 1

II. Mass and Energy Release Modeling A. LOFTRAN Computer Code Mass / energy releases are calculated using the LOFTRAN code. LOFTRAN is a FORTRAN language, digital computer code, developed to simulate transient behavior in a multi-loop pressurized water reactor system. The program simulates neutron kinetics, thermal hydraulic conditions, pressurizer, steam generators, reactor ecolant pumps, and control and protection systems. Up to fcur independent loops may be modeled. LOFTRAN is used for analysis of non-LOCA transients and is documented in Reference 3 The model of importance to blowdown calculations is the steam generator model. The primary side contains multiple nodes to model the tube bundle.

The standard LOFTRAN steam generator secondary side model, (Figure 1), is effectively a one node, two region model of saturated steam and water.

Heat transfer is assumed to occur only to saturated water. If tube uncovery occurs the amount of surface area available for heat transfer is accordingly reduced. The LOFTRAN code incorporates a more detailed steam generator model which is used to predict tube bundle uncovery.

B. LOFTRAN Hodel for Superheated Steam The LOFTRAN code has been modified to account for heat transfer to steam from the uncovered tube bundle region. (Figure 2). In the modified version of LOFTRAN, all heat transfer occuring in the uncovered region is assumed to add superheat to the steam exiting the steam generator. The primary side temperature in the uncovered tube region is conservatively assumed to remain constant through the nodes which are uncovered. In reality, there will be a drop in temperature due to heat removal to the secondary side, but this is expected to be small due to the low specific heat capacity of the steam and due the high primary side flow rate.

, The heat transfer coefficient used in the uncovered tube region is discussed in the Appendix. This correlation bases the heat transfer on the difference between the tube wall surface temperature and the bulk steam

- temperature in the region. In the LOFTRAN modification, the conservative asstrnption is made that no credit is taken fer either a primary film heat transfer resistance or a tube metal heat -transfer resistance. Therefore, the wall surface temperature of the tube -is assumed equal to the primary fluid temperature.

i The modified version of LOFTRAN automatically determines the proper number-of steam generator nodes for the superheat region of steam in the generator. The variable node capability is ~ applied to both the primary and secondary side. At each time step during che tube uncovery, the modified LOFTRAN code makes a general evaluation of the uncovered tube region (e.g.

steam flow rate, uncovered tube heat transfer area, estimated heat transfer coefficient, etc.) and determines the number of nodes to be used in the subsequent calculations. The total heat transfer for the uncovered tube region is determined and accounted for in the primary temperature transient

calculation. The superheat/ tube uncovery modeling is applicable to all steam generators.

Figures 3 through 6 show typical results for a 0.86 ft steamline break from 102 percent power using the modified version of LOFTRAN. Figure 3 l shows the fraction of tube uncovery versus time with uncovery of Loop 1 1 (faulted) starting at 152 seconds into the transient. At approximately 300 seconds, the uncovery transient reaches an equilibrium point where the steam flow out of the steam generator matches the auxiliary feedwater flow into the steam generator. Additionally, the tube uncovery transient for Loop 2 (non faulted) is plotted but shows no tube uncovery for the entire transient. Figure 4 presents the steam flow transient for dais case.

Figure 5 includes plots of both the superheated steam enthalpy and the saturation enthalpy for the Loop 1 steam generator. Figure 6 includes the Loop 1 temperatures for the steam generator tube inlet (primary side),

steam exit temperature (superheated steam), and the saturation temperature for the steam pressure.

C. NOTRUMP Model Comparison The NOTRUMP computer code (Reference 4) was used to verify the LOFTRAN modeling of superheat. The computer code was originally developed to analyze transients of secondary systems with two-phase conditions. In the past, it has been used to analyze various transients in the primary and secondary coolant systems. .NOTRUMP has recently undergone major revisions to enable it to model non-equilibrium nodes (i.e., separate liquid temperature and steam tempera'ture modeling). Using NOTRUFP, the steam generator can be broken down into sufficient nodes to model the nonequilibrium effects of the steam generator, as well as the tube region during uncovery. NOTRUMP can model all modes of heat transfer associated with a steamline break transient, including heat transfer from the uncovered tubes to the superheated steam and the feedback effects between the primary and secondary sides. The two phase mixture level calculation accounts for primary to secondary heat transfer and the swell associated with rapid depressurization of the steam generator during the blowdown.

A comparison of LOFTRAN and NOTRUMP blowdown results is presented in Figures 7 and 8. The mass releases shown in Figure 8.show excellent agreement. The LOFTRAN prediction of superheat enthalpy is slightly higher than NOTRUFP, 'while the predicted time of tube uncovery is somewhat later.

NOTRUMP shows a chugging effect during the uncovery phase.of the blowdown.

This is believed to be in part due to oscillations in the flow link between the dcwncomer regicn and the steam dome region. (The flow link is the drain path for_ the moisture separators to the downcomer region.) With the f1cw direction towards the downcomer, superheated steam goes into the downcomer region and is condensed. This alternates with a flashing of a portion of the water volume .in the downcomer region. This raises.the pressure of the downcomer, resulting in a flow reversal in the link with saturated steam from the downcomer mixing with the superheated steam in the dome. This mixing results in the variations in the superheat enthalpy seen in Figure 7. Although LOFTRAN does not show the. enthalpy variation since the detailed modeling of the downcomer and dome are not included, the overall agreement with NOTEUFP is very good.  ;

1

i o

D. Effects Of Analysis Assunptions The effects of superheated steam are dependent upon the occurrence and extent. of tube uncovery. The major parameters affecting tube uncovery are:

initial steam generator inventory, auxiliary feedwater flowrate, assumed feedwater system failures, and protection system errors. Variations in these parameters are in the process of being evaluated for their effects on the containment temperature response (Figure 9).

Refinements in the mass and energy release modeling (Figure 10), arc being evaluated and several areas show a potential for reducing the degree of superheat being generated. Some of these areas are:

- Evaluation of liquid-steam interactions such as the phenomenon of tube support plate flooding and heat transfer across the tube wrapper from

. the superheated steam to the auxiliary feedwater flcwing down outside the tube wrapper.

- A more detailed steam header model in LOFTRAN.

- Modeling temperature drops in the primary superheat nodes.

- Evaluating other void correlations for use in predicting tube uncovery.

i 1

l 9

l l

l III. Containment Mode, ling A. Description of Containment The general phenomena taking place inside an ice condenser containment during a l steamline break transient can be described utilizing a typical ice condenser elevation drawing (Figure 11). Steam is discharged to the main (or lower) compartment where heat is removed by the internal structures, steam flow to the ice condenser, and the ice condenser drain water. The dead ended compartments are the regicns which are located below the ice condenser and outside the crane wall (Figure 12). Air is discharged from the main compartment to the dead ended compartment and ice condenser so that the resulting steam to air ratio is that region is much higher than in dry containments. At ten minutes following the containment hi-2 signal, deck fans are actuated which direct air flow from the upper compartment to the dead-ended compartments. Most of the safety related equipnent is located in the dead-ended compartments although some equipment and cabling are located in the main compartment.

B. Containment Models Figure 13 outlines the major models and assumptions utilized in the LOTIC-3 containment code. In the currently approved version of LOTIC-3 documented in

, Reference 5, four distinct regions of the containment are modeled; the lower compartment, the dead-ended compartment, the ice condenser, and the upper- 1 compartment. Two condensate /revaporization models are used depending on the size of the break. For large steamline breaks, 100% condensate revaporization is assumed. For small steamline breaks, a convective heat flux model is used which calculates partial revaporization during the transient. 'Ihe wall heat transfer model utilizes the Tagami heat transfer correlation for condensation heat

, transfer and the convective heat flux model' derived from the work of Sparrow (Reference 6) which calculates the convective heat transfer for small steamline breaks. The sump recirculation system is only modeled for the large break LOCA transient containment response.

Figure 14 shows the four regions modeled with'the mass and energy flows that can be assumed in the analysis. The Catawba nuclear plant does not have lower compartment sprays and they are not modeled in the analysis. Superheat heat transfer is conservatively assumed to be zero for the steamline break containment analysis. In the model described in Reference 5, wall heat transfer is not modeled in the dead-ended compartments although these regions do contain-structures which will remove heat. The analysis does include the upper compartment. sprays, flow through the ice condenser, deck fan flow, and flow to the dead-ended compartments.

LOTIC-3 solves the conservation of mass, energy, and momentum equations for' upper, lower, and ice condensor regions (Figure 15).: After the new' lower compartment. conditions are determined, conservation equations are solved for the' dead ended compartment and the flow rate between the compartments is determined.  ;

Figure 16 presents a typical steamline break containment temperature transient' that is calculated using superheated steam blowdowns from the LCFTRAN code and the modeling of ice condenser drains as a heat removal source. The transient shows' that initially the containment temperature increases rapidly during the u

= , . , , - -w- v-e-

blowdown. When the upper compartment sprays actuate there is a slight decrease in the main compartment temperature. The tenperature then rises slowly until ice condenser drain flow decreases to the point at which time the tenperature begins to rise again (approximately 250 seconds). This rise in containment temperature coincides with the steam generator tubes uncovering at 152 seconds and the maximum superheat occurring at approximately 250 seconds. The stean ,

generator level stablizes whe'n the auxiliary feedwater flow is equal to the l steam discharge at approximately 300 seconds. The containment tenperature then starts decreasing with decreasing decay heat. At ten minutes, the deck fans

_ actuate which results in a rapid decrease in containment temperature.

C. LOTIC-3 Code Modifications Four modifications have been incorporated in the LOTIC-3 containment model which ar'e (Figure 17);

1) wall heat transfer model

2) convective heat flux model

3) ice cendenser drain model

4) dead-ended compartment model D. Wall Heat Transfer - -

The modificatien to the wall heat transfer model is described in Figure 18. In the LOTIC-3 model, only condensation heat transfer, utilizing a Tagami heat transfer coefficient and a tenperature difference between the wall and saturation, was previously modeled,.' The modification includes a convection term '

with a conservative convection heat transfer coefficient and a temperature difference between the containment atmosphere and an appropriate interface temperature. The Appendix presents a more detailed description of this model.

E. Convective Heat Flux The codification to the convective heat flux model is described in Figure 19. A term has been added to the convective heat flux model to account for the feedback effect from including a convective term in the wall heat transfer model. The Appendix presents a more detailed description of dais model.

F. Ice Condenser Drain Model In an ice condenser containment there is approximately twenty drains exiting from the ice condenser into the lower compartment at an elevation of about forty feet above the compartment floor. The drain pipes are one foot in diameter.

The drain flowrate is calculated by the LOTIC-3 containment code. For a typical small steamline break transient the drain flowrate varies from approximately 4000 lbm/see to 500 lbm/sec during the gimeframe of interest. The temperature of the drain water is approximately 130 F (Figure 20).

Figure 21 presents the assumptions and the basic model used to estimate the heat removal from the lower compartment atmosphere to the ice condenser drain water.

It is conservatively assumed that the drain water stream does not break up prior to reaching the floor even though many of the drains have equipment and structures located below them. Therefere, heat transfer -is assumed to occur at

. . , a

the stream surface only. It is also assumed daat the stream surface temperature is at the saturation tenperature of the containment.

The heat transfer to the stream is:

q=hAJLT j where h = condensation heat transfer coefficient A = surface area of the stream AT = appropriate tenperature difference 3

The calculation of the heat transfer surface area is described in Figure 22. ,

In order to model the drains in LOTIC-3, the drains are modeled as a wall heat i sink with a surface at a constant tenperature (see Figure 23). Currgntly,in the version of LOTIC-3, the surface temperature-is assumed to be 230 F which is close to the containment saturation tenperature. The drain surface area is calculated at two points in time during the transient; early in time with a high

> flowrate and later in time with a low flowrate. To ensure conservatism in the i area calculation a 10% reduction of the surface area was assumed.

As described previously (Figures 14 & 15), the LOTIC-3 containment model did not account for wall heat removal in the dead-ended compartments. To obtain a

{ conservative estimate of the tenperature transient in the-dead ended compartment, the heat sinks located in the dead ended compartment region along j with'the heat sinks in the lower canpartment are modeled in a combined volume (see Figure 24). This " modified" lower compartment model is used to determine a conservative _ dead-ended compartment tenperature transient. Since the lower ccmpartment will be hotter than th*e dead-ended compartment, this methodology results in a higher tenperature in the dead-ended compartment then would be expected, r

G. Transient Results With the modifications described for LOFTRAN and LOTIC-3, the previous FSAR limiting case -for Catawba was reanalyzed to determine the impact of superheated steam. The case selected is a 0.86 square foot break at 10j5 power (Figure 25).

. The peak lower containment temperature for this case is 324 F. This temperature

- is calculated for the lower compartment _only. It is expected that the -

dead-ended compartment temperature will be significantly lower.

In addition to the 'model modifications incorporated. in LOTIC-3, Westinghouse is pursuing further improvements in the areas noted on Figure 26.. .One area _ is in

~

I

- the wall heat and mass ' transfer models. Since condensation is a mass transfer type phenomena, the heat and mass transfer should be linked. This approach has  ;

, been used in Reference 7.

An bnproved drain model' is also being investigated. This improved model will'

- calculate the drain surface area as a function of flowrate. It will also i calculate the average temperature rise of the drainwater. This model will mere-accurately represent the actual phencmena in the containment.

4 f

i i

8., ,- . s,- ., .n.- -- , .- ,

l l

WE3TINGHOUSE PROPRIETARY CLASS II V. Appendix ,

WESTINGHOUSE STEAMLINE BREAK BLOWDOWN AND CONTAINMENT ANALYSIS METHODOLOGY The following sections describe the Westinghouse methodology for detennining the containment response for a steamline break incorporating the effects of superheated steam. These sections describe in detai.1 changes from the methodologies described in References 1 and 5.

I. Steamline Rupture Mass / Energy Blowdown Analysis A. LOFTRAN and MARVEL Computer Modeling Mass /energ'  ; releases can be calculated using either the LOFTRAN code (Reference 3) or the MARVEL code (Reference 8). The LOFTRAN code is used for non-LOCA FSAR accident analyses. The MARVEL code was specifically developed for assymmetric transients such as -steamline breaks. These two codes are very similar because they were developed in 'an interrelating -

fashion and much of the model'ing is common- to both codes. The MARVEL code l

was used in the development. of Reference 1 because LOFTRAN at that time was 1

a lumped model which was used for symmetric loop transients. Furthermore, l for steamline break analysis purposes, MARVEL contains a model for water

! entrainment. However, the current version of LOFTRAN is a multiloop _

i version which also contains a water 'entrainment model. With the

development of a multiloop version of LOFTRAN and the inclusion of an f entrainment model, the use of MARVEL has been generally discontinued. .This enables the use of LOFTRAN as a single system analysis code for non-LOCA transient analyses. LOFTRAN is used in the analyses presented here.

The model'of importance to blowdown calculations is the steam generator model. The primary side of the steam generator contains multiple nodes to

, model the tube bundle for both the modified version of LOFTRAN and MARVEL.

Heat transfer calculations from the primary to secondary side are identical in the two codes, although the methods for initializing the heat . transfer resistances are slightly different. The secondary side is effectively a-one node, two region model of saturated steam and water. ' Heat transfer is assumed to occur tm saturated water. jIf tube uncovery is predicted, the amount of surface area :available for heat transfer is reduced.-

LBoth codes contain a detailed steam generator model which is used to ,

predict tube uncovery. This model. calculates the liquid volume inithe

~

steam generatcr shell and' acgognts for' the detailed ' steam generator geometry. -The [ .1 : correlation is-used in both codes to.

. predict the voiding in the tube . region, although the correlation is :

modified for use in LOFTRAN._ In MARVEL, -tube uncovery is calculated based. 1 l

4 8

- + - ,_r

, - y

on comparison with the actual water level and the height of the tube bundle. In LOFTRAN, the user specifies either a water volume in the steam i generator corresponding to tube uncovery, or a void fraction in the riser j section of the steam generator t which tube uncovery begins.

Both codes have similar models accounting for reverse heat transfer, thick i metal heat transfer, feedline flashing, and safety injection system I operation. Auxiliary feedwater flow can be input as .a fraction of nominal feedwater flow, although LOFTRAN has an additional capability to model auxiliary feedwater flow as a separate system. For analysis of double ended ruptures, MARVEL accounts for the volume of steam in the piping downstream of the steam generators in the blowdown calculations. In LOFTRAN, this consideration is added on to the blewdown mass and energy results by hand. For split ruptures, which the analysis presented here addresses, the steam piping masses are handled identically in both codes.

In summary, LOFTRAN and MARVEL are very similar codes, and either can be used to calculate mass / energy blowdowns. To demonstrate this, a comparison of the blowdowns for a typical case is presented in Figures A.1 and A.2.

Figure 1 presents the mass release rate for a .86 ft2 split rupture from 102% power. For this case, Figure A.2 shows the saturated steam enthalpy as a function of time. This blowdown is typical of results used in FSAR analyses prior to the modification noted in this report for the LOFTRAN code. As can be seen from the figures, the results are extremely close..

B. LOFTRAN Model for Superheated Steam As mentioned previously, the LOFTRAN code has been modified to model heat i transfer which may occur in the uncovered tube bundle region. This effect 4 is modeled in both the faulted and intact loops. In the modified version

] of LOFTRAN, all heat transfer occurring in the uncovered region is assumed

to add superheat the steam exiting the steam generator. The tenperature of j the prigary coolant flouina through in the uncovered tube region mode is conservatively assumed to remain constant. Realistically there would be a drop in temperature due to heat removal to the secondary side, but diis will be small due to the low specific heat capacity of the steam and due the high primary side flow rate.

The heat transfer coeffic49g g uged in the uncovered tube region is based on the [

• The heat transfer coefficient (U) is calculated by the.following expression: _

a,c s

This correlation is' presently used f

. transfer by:the [ =]5'Buperheated forced Additionally,

' computer codes. convection heat

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

this correlation is based upon the heat transfer from the surface of the tube wall to the average bulk temperature of the steam. In the LOFTRAN modification, no credit is taken for either a primary film heat transfer resistance or a tube metal heat transfer resistance. Therefore,the wall temperature of the tube is conservatively assumed equal to the primary fluid temperature.

4

~~

~'

A, C

(1) e O

og i

4 e

1 9

9 4

y- 3w.a , ~p = .x- y -

., ye - - -y -

w-- ~r,w

l The modified version of LOFTRAN automatically selects the proper number of steam generator nodes for the superheat region of steam in the generator.

The variable node capability is applied to both the primary and secondary side. At each time step during the tube uncovery, the modified LOFTRAN code makes a general evaluation of the uncovered tube region (e.g. steam flow rate, uncovered tube heat transfer area, estimated heat transfer coefficient, etc.) and determines the number of nodes to be used in the subsequent calculations. Each node is evaluated to determine the steam temperature exiting the node with a convergence criteria that is based upon the total number of nodes used. The exit steam temperature of one node is used as the inlet steam temperature of the next node.

The heat transfer calculation to determine the outlet temperature of the node is based upon the following expression:

Q = UA5(Tpri-(Tout + Tin) )*N s s*(Tout -Tin) where Q = Heat transfer to the steam U= l M

Tpri = Primary node temperature Tout = team node oudet temperature T in = Steam nod'e inlet temperature' M3 : Mass flowrate of the steam Cs = Heat capacity of the steam A = Heat transfer area in the node including both hot and cold leg sides of the tube bundle The total heat transfer fer the uncovered tube region is determined and accounted for in the primary temperature transient.

C. Blowdown Sensitivity to Plant Conditions The effects of superheated steam are dependent upon the occurrance and extent of tube bundle uncovery. Parameters affecting tube uncovery are:

initial steam generater inventory, break size, auxiliary feedwater flowrate, and the single failure assumed.

The initial steam generator inventory depends upon the measurement errors i associated with steam generator level and upon initial power level. Steam i generator mass increases with decreasing power, thus, breaks intitiating i from low power levels will result'in later tube uncovery.

Larger break sizes. result in faster blowdown of the steam generator and earlier tube uncovery.

i Large auxiliary feedwater flowrates only delay tube uncovery, but will ,

also cause the final equilibrium steam generator level to be higher. This  !

equilibrium condition corresponds to the point when the break flow rate is equal to the auxiliary feedwater flow rate.

The single failure assumed in the transient may impact the amount of water supplied to the steam generator. Auxiliary feedwater runout will increase the amount of water supplied to the steam generator. Failure of the feedwater isolation valve will also cause extra water to be supplied to the generator as the additional mass between the isolation valve and the check valve flashes to the generator.

e ne e

G

. o II. Containment Analysis ,

A. Wall Heat Transfer Model The original LOTIC-3 wall heat transfer model is based en the stagnant Tagami heat transfer correlation. That is, 9"*kAGAMI(TSAT-TWALL) 2o

/H h(TAGAMI, MAX)=72 BTU /hr-ft - F hTAGAMI = 2 + 50 MSTEAM AIR This model was developed for saturated steam in the presence of large amounts of non-condensable gases. In the icwer compartment of an ice condenser, most of the air is swept out of tne lower compartment through the ice condenser .and into the upper compartment. Therefore, after about 30 secends, there is almost no non-condensables in the lower compartment. Typical values fer the condensation of pure steam are in the range of 1000 to 3000 Btu /hr-ft2 oF (Ref. 5). The correlation used in the modified LOTIC-3 code is in extension of the Tagami correlation fer nearly pure steam.

q"=hCOND (TSAT-Tyggg) hcond = 2+50 MSTEAM " AIR h(cond, max)

• A maximum value of ( Ja,c was chosen as a conservatively low condensing heat transfer coefficient in a nearly pure steam environment.

In addition to this modific'ation, an additional term is needed to account fcr the convective heat transfer from the superheated steam to the condensate film.

This convective heat transfer is dependent upon whether there is condensation occurring on the walls. If condensation is occurring, the correlation used is:

com ucombuds where: ae If the wall temperature increases to above the saturation temperature then the convective currents will be reduced such that the correlation used is 9"conv=hconv(Tbulk-Tg ,11) wnere:

}a,C

[

6

i

.: o Thbs in streary, if Twall<T3,g men

[ 3a ,c If Twall > Tsa , then the correlation used is: -

[ ja,e e

e

.. o l

l

'B . Convective Heat Flux Model 1

i l When the containment atmosphere is superheated, the containment temperature is a I strong function of the amount of steam mass in the atmosphere. Thus the amount of mass condensed on the heat sink surfaces is a key parameter. The actual amount of condensate formed is -

"cond

• 9cond/hg Unfortunately, with the use of a heat transfer correlation based only on test data (such as Tagami or Uchida), only the total heat transfer coefficient is obtained. This total heat transfer coefficient includes both the condensation heat transfer and the convective heat transfer. Based on the work of Sparrow (Reference 6), the Westinghouse Convective Heat Flux model in the original LOTIC-3 code calculates the ratio of the convective heat transfer to the condensation heat transfer. Therefore the calculation of the amount of mass condensed is

[ ja,c J-In the modified LOT 1C-3 model, the, amount of superheat convection is calculated.

The anount of convective heat transfer at saturation is not known explicitly in this raodel. Therefore, in the modified LOTIC-3 code the original convective heat flux model will be used to calculate the fraction of convective heat transfer for saturated conditions. The actual correlation is .sc 3

where, (q /q is determined from original convective heat flux model C "V ic cggg)S$ bent of convective heat transfer colculated in the wall and q heatEPSXSfNrmodel j In summary, the modified LOTIC-3 model is consistent with the original LOTIC-3 model in its calculation of the mas condensed. The only difference is that in the modified LOTIC-3 code, the annunt of superheat convective heat transfer is known explicitly, while in the original LOTIC-III model, only the ratio of l convective heat transfer to condensation heat transfer is known.

e 4

4 m [g

IV.

References:

-1.. Land, R. E., " Mass and Energy Releases Following A Steam Line Rupture" WCAP-8822 (Proprietary) September, 1976 and WCAP-8859 (Non-Proprietary).

2. NS-EPR-2563, February 14, 1982, E. P. Rahe of Westinghouse to J. R. Miller, NRC, " Additional Information on WCAP-8822".

3. Burnett, T. W. T., et al. , "LOFTRAN Code Description," WCAP-7907, June, 1972 (Proprietary).

4 Meyer, P. E., and Kornfilt, J., "NOTRUMP - A Nodal Transfer Small Break and General Network Code," November,1982, WCAP-10079 (Proprietary) and WCAP-1C080 (Non-Proprietary).

5. Hsieh, T. and Liparulo, N. J., " Westinghouse Long Term Ice Condenser Containment Code - LOTIC-3 Code," February, 1979, WCAP-8354-P-A Sup. 2 (Proprietary),WCAP-8355-NP-A (Non-Proprietary).

6. Sparrow, E. M., Minkowycz, W. J., and Saddy, M., " Forced Convection i Condensation in the Presence of Noncondensables and Interfacial Resistance", Int. J. Heat Mass Transfer, Volume 10, 1967.

7. Corradini, M. L., " Turbulent condensation on a Cold Wall in the Presence of a Non-condensable Gas" Nuclear Technology Vol. 64, pp 186 - 195, February, 1984. ,

8. Krise, R. and Miranda, S., " MARVEL - A Digital Computer Code for Transient Analysis of a Multiloop PWR System," November,1977, WCAP-8843 (Proprietary) and *rtCAP-8844 (Non-Proprietary).

, 9. McCabe, W. L., and Smith, J. C., " Unit Operatons of Chemica1' Engineering",

3rd Edition, 1976.

G l

=

.. o l

l l

LOFTRAN - MARVEL COMPARIS0N

.860 FT2 BREAK AT 102 PC POWER 2000.0  :  :  :

f

/

1750.0

\

0 1500.0 - -

C -

5-1250.0 - -

2 S --

"r1000.00 - -

t.a

, G 750.00 --

4 x

c: 500.00 - -

MARVEL ----y /

LOFTRAN --

co .

250.00 - -

' IUCS/83

' i 0.0 .o o i

o o o o

o o

O o o o a b b b 'b b b o

o e o o o

~ m a sn w o -

TIME (SEC) ,

FIGURE A.1

. . = = -

__= _. . . - . . - - . _ .

.* 4

. .e i

1 1

LOFTRAN - MARVEL COMPARISON

.860 FT2 BREAK AT 10'2 PC POWER 1300.0  :  :  : .  :

1275.0 - - --

S N

D

.a 1250.0 - - --

w 1225.0 - - --

0FTRAN 1200.0 - - ..

~

W 5 1175.0 - - --

z

<C W 1150.0 - - --

M

'S .1125.0 - -

1100.0 4 4 1 4 i c Q O O O O O O O. O O O O O

. O O O O O O I o - cu m .:r to - e TIME (SEC) 1 w.

! FIGURE A.2 i-.. . . . . . . . . . .

~,*

9 ty METHODCLOGY FOR ADDRESSING SUPEPHEATED STEAM RELEASES IQ ICE CONDENSER CONTAINENTS Purpse ,

The purpose of dais report is to document the infcrmation presented on March 19, 1984 in a meeting with the U.S. NRC Containment Systems Branch 'on the status of progress made in addressing the confirmatory item on the Catawba Nuclear Plant Safety Evaluation Report. This confirmatory item deals with the effects of superheated steam generator mass and energy releases following main steamline break accidents. Attachment 1 includes the list of attendees at the meeting and the overhead slides covered in the Westinghouse presentations.

Technical presentations were made describing the modeling of the steam generator and heat transfer from the uncovered tube bundle during the steam generator blowdown along with a description of the containment model and transient response. A proposed plan of action was also presented and discussed with the Staff. In accordance with that plan, this report represents the first milestone in the proposed plan of action. As committed to in the meeting, the appendices present proprietary information which relates to the specifics of the models and sensitivities that were not directly addressed in the meeting.

Attachment 2 is an explanation of, and refers to, the overhead slides (Figures)

p. resented at the Ibrch 19 meeting. -

e e

S

i 1 0

}

4 4) l O

ATTACHMENT' 1 m*

e,,

e D

O

/)%d- e o ;- 3//7/84 .

' hf c.ms U/ ZL. I~- ao iL .E n <. t.. .

7~ r A~9 .t .'>

l da n .

c < g a . ., - xa.,

p u

)(,ll. JA &TOO K '

WA'c/DL

' [2 '. O .' 3rMt2P F-- .Duile. /nPD

,.x 4. 11 rr a W ,vaara s+ w i

n.v. -

ro u .xc e. coce ue socry

~! [ l(e u di"C./JL '

b$. Lolm W N uct.L-AA. S MCT Y P\F &c4 VJ ktud e @ ( ,,

- 8A. l n i W Hv c(ca,. J,;-rty

%$tML $ }lgff,yg, b . $.

M c.scun .w_,iv.,ctu , n ce+y '

F 7

~~

~~~~l teacc .~. .c( bd MCC , f'f n.\ m -

S .D A/<.M ,-- Dsh.c. , Des: ~ Em -as. ,

A. 5 Me'$v b 4, ><r:j~ Ey :rw:

r. E. to.us 7e s x - % s ,.G -

$6ACtl LtClu. sinc : .firlf*f* , EE.'n*:

Z. AttiLi

-k.MQ4.A c.h.3 t_.J ,J . odrwt. Sas] .

W. D. c,.,yd . . T V A, E.):- , " bo sjn S. ?. .'v':43 b Q.tr, , G kw f..L L- 4e,cbe 9,5 *

$ AI l4)Q m W , N t"~ D SME ~

0. M. Oso .c g x7~a / ducsu.e 14Fe?"7

) .'$ f4 R 4 /L. f o <-

/U10& C.'N S'

~

K4 Q C

.?'l.R,k.W-

1. RC/CSB C. Li MC C /c c cs M. Malr4.Ocacesa A/$c/c sv3

= -

LOFTRAM MODEL .

, STEAM GENG.RAToR SEcoWDARY ,

STEAM FLOW .

,~ . .

,- ,_ y SAvue.ATED i!

. .,- , ,,. .* d STEAM REGled !

. . ... ~. . . .

' ~

• • . SATVRATED ' l

_- ',<**e..

usATER Rastc4 l y.

Ac

.I ...... .

y . 2~ .

l l

3

• StMGLE MCDG , 2 ft.EGled MCOEn

• WO HEAT TRAdSFER TO SATueATED STEAM. REGtcM i -

1

• HEAT TRAMSFER. Tc SATORATED WATER REGtcM iS MODIF1ED Rl4EL TVSE 'UNCcN ERN '

FIGURE 1 .

~

~. ,

LOFTRAM M6 BEL SUPERHEAT HEAT TRANSMR suPEAHEATED STEAM FLcW N DAD

~

] -

suessaaar accas

\ ^ ,A NPI,INEbT sn,-

IMES

/

+

. 81'C., . -

t:.PP d CTIVE '

~ 'UN'COVERY WL i:

VAnt Aea_s sueensaAT ~

HODES MODEL l

. copsvART P Rt W 5et TENPERATuRE tu. suesansAT SEstou Assuutt Feg.

HEAT TRANSFu.s. cALcuLATNDMS .

• C ALCOLA%D SOMSME AT HEAT TRANSMR AccoVMTED FeR. IM

, pst.sa Any T rad % W.HT m " 2 .

TUBE UNC0VERY LOFTRAN SUPERHEAT MODEL

. .860 FT2 BREAK AT 102 PC POWER 1.2000  :  :

G

< ~~

E 1.0000 -- LOOP 1 8

p --

W .80000 --

E s --

a .60000 - -

S 5

.40000 - -

8 '

8

$ .20000 - -

S LOOP 2 0.0 8 8 8 8 8 8 e s s s s s e d S 8 8 3 8 8 TIME (SEC) e FIGURE 3

MASS BLOWDOWN LOFTRAN SUPERHEAT MODEL

.860 FT2 BREAK AT 102 PC POWER 2000.0  :  :  :

f "

1750.0 --

k b

m 150b.0 - -

N 1250.0 - -

n S

"x 1000.00 - "

w G 750.00 - -

M E

o 500.00 - -

250.00 -- "

0.0 i i i i i o o o o -

o o o o b b o b

o

. O O o o o - m m a w w

( ,

TIME (SEC) 1 FIGURE 4 i 1

l

i ENERGY RELEASE LOFTRAN SUPERHEAT MODEL

.860 FT2 BREAK AT 102 PC POWER 1300.0  :  :  :

- 1275.0 - - --

3 SUPERHEATED STEAM S

1250. 0 - - --

a_

s 5 1225.0 - - --

8 ~

A

$ 1200.0 -

~

c SATURATION s

5 1175.0 -- --

=

ld 1150.0 -- --

m S 1125.0 -- --

1100 0

. y y y y y g o e d 6 6 6 6 6 S 8 8 3 8 8 TIME (SEC)

FIGURE 5 1

TEMPERATURE TRANSIENTS LOFTRAN SUPERHEAT MODEL

.860 FT2 BREAK AT 102 PC POWER 650.00  :  :

1 600.00 -J --

??

$ 550.00 - --

c.

SG TUBE INLET

8. w-5e w y 500.00 - -

(PRIMARY SIDE) --

g

• $ 450.00 --  :-

z SUPERHEATED STEAM E~

h 400.00 - - --

t 350.00 - - --

SATURATION 30~0.00 o

8 8

cu 8

m 8

-r 8

n 8

w TIME (SEC)

FIGURE 6 1

8ag 4

8a8 .

Od )

C E

S

(

E M

I T

N 0

S I Y I )

o on R A M A R U P T R M F T O

C O O L N P YP l

MU A L

RH -

TT ,

O NE O o8 N

NK AA E R R T B F

O L

ko V,

9o 0 0 0 0 0 0 0 0 0 0 5 0 5 0 5 0 5 0 0 7 5 2 0 7 5 2 0 1 1 1 1 3 2 2 2 2 1 1 1 1 1 1 1 1 1

=

$0W8 Edh5 2eE"

.- . e i

o 0:t

• 009 00'00G i

00*004 0

M E

5 z a ,

w < z "

5 o.

%5 u_ s 00'00E E o C J Z 8wr a.

r <"

=) 2 do o 's e

00'002 zi u z <w i

<x Em LL '

o

_a 00'001

/

/ *

/

t l 0*0 o e o o o o ' ~

a d d d o 8 8

~

8 8

8 m o

~

(33S/W81) 31ve 301J ssvw mv3ue 1viot FIGURE 8

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

1 EFF CTS'0F NALYSIS ASSUMPTIONS INITIAL STEAM' GENERATOR INVENTORY AUXILIARY FEEDWATER FLOWRATE i

FEEDWATER SYSTEM FAILURES PROTECTION SYSTEM, ERRORS f

I I

i FIGURE 9-

~'

s >

m

ADDITIONAL MODEL CONSIDERATIONS LIQUID-STEAM INTERACTION

- IMPROVED STEAM HEADER MODEL HEAT TRANSFER THROUGH TUBE WRAPPER TEMPERATURE DROP IN PRIMARY SUPERHEAT N0 DES OPTIONAL V0ID CORRELATIONS

. c 4

} 'E -

4 FIGURE 10 j' i

.- . e o

SHIELD SU/LDlHG DCML*'

Q

. I

• l i{t$.-

~!!

A

, ' ~3 TEEL CoNrAINMENr VE3SEL . .

EL 829.C4 \

I+7 55 i

.i

• f f/73 rCN PCLAR CRANEY,,0 P Syyggy

/ - POLAR CRANE .*

• i BUILOlhG

• WALL m -

p

.. i

_i.

)

TCP of DECX

[EL794.43

)

Mb'*si l h'?

fEL 797u TCP 0F CRANE RAIL

/

-399tNG LIN.

EL 731.3 ICZ loa 0 LNG '; , ,_ -- ~ X^ )..

j'E AIN CYCLONE '" ' ' ~ *

• VENTILAr/CN A,! ,_2-,

.2f Q]~4~,'h .pso BR/0CE CRANE st* 6 *R uMWm[ .

l

. gAppging ygg73 l .

TOP ofICE BED '

l \\~

• ., ,e ,.-

EL 780.0 K .. I .

fotyt0ER BARRIER gg 77g,gg ,_

, , , y ,,

3- .. ..- .

• I## #

/CE .2: s s: '

wg ,,

P: Pa2I .'

CONOENSEX --p % ,

'r . f\ \ \ '. t ECNTA*NMENT .

s: 3: '

e a .

(>:y:\ ..::y .

1 . STEAM ,

. . GENERATOR l .'

. o l s -

CoMroot Roo Oswe, r , .

tys).s:'

• G ,. Mt33lLE $NIELD . l .shss si

/g( I Y.

ice Coxo

- 1 1 " "' .h

.Q5i .;s < ,

/ ,  ; ,-a ..u.a s  ;

!; \ . . . . . . .

f ..

P L.

'l

- l l X a .1 r:. . !j ;

gal l . l.l., .

sis'?t,* *- L n QQ i =

~

.vL (g;LIy,

,.. 0 ACCU-vuroR+ , , rv-R .

_ in c.

.. x- .. 3 ceA0e . . .

n. t. .

Et !!s;*2. -

.

?.

N: -

T I

) [ .'!"" " * -

q .'

k ,. L 9 vfNr nN g( ) -.f.:;m:::nEL ds5.0-(.,;*

6 ..

\ . *; , fjf ? q ),.p.'.TF i.p .' ;< ' . 4' Earr n es.r.o ,

e F

S

, . .p

, (r; . '- 1 Q-L' pj . . . .

4 ,.. '.,*, l Y ..

.? .* . . s .g-

& j N .*.

Y s" EL 679.78 .

'. , 3: t. .

-1 ._ a;-

.'y-l M- . .

• =- .

e y,t a

.* . . Y'

. . n ' e %. ' STEE. L 'L/Iv5Ry.... )ENCrbR . .

• * . . . . . s ,. .

% ANCNon actr . . . g n ..

l . I 1 -

I s+ = o. son m l .

,a ssi.s, '.l .

)

3u-6. . . . ..

i -

Reactor Building Elevation

{ .

FIGURE 11

7-l 1

I w

l

% 1 SPEW ..  % ,y,,gypy gggg ) , .

FUEL Pir ,- '

. -4 .e'.

f

.:s:.s. .;,

'.s . . . .

. . .i:

3.L

.w'.r

'N. '

PIPE CHASE - - 'W q

FUEL TRANSFER

.y . < . < -.

s,.

SYSTEu

., ;l '&- .,4;.4-

. s: .

• g; s '

-t .

.s 4- ,

e. , .- .

l

k' 1. f ' -

s'

.~. .

EL 693.0 . . , , .'d  ;' -) % >..

- I

~

.' . . .l:*

• REA CTOR UPPER v. s I

• s

• s *

• ', INrERNALS STORAGE gg l N

^

l*' s 90.gg. * '. ( ,% f. . . r"

.' - ' \ N'REAcr0R p_

(M .'

~

.l, .

. C5 0.2 A,N T ,

4- ,* Gl . g* ~~

e . ., " put;p; (J)

. / '.

4-

. V...m.. .

-*  % *e .Q '. p* .L OWER\ *6 Q. ..

• /NTERNALS ' s-RACES -.,

s o'

? .' '

.'.;'.~.-

• f 6

._ ~

3 s , .r ..,

• h, .' .

0' $ )f ' ':( 1.\

T: , . ': i I_. '- Q. p,,

~~

g .n-na ", MAIN

.4 'care- '( 4

, SrEAu CEN (4)

,'\A4 e e 9

, , > s.1 ' u REACrQR .

e.

STEAM . .

,1 <: paggsygg

,, _3 f,.' _,,

ESSEL -

CRANE WALL

"~ T , ." z w/ y .s f ,,

k'g .w ..

O,  : .

v . '

• et O l Q -

l,'

O. '-.% 's 702J5 .Q .

FEED: ,

WATER. -w EL 679.73 , I, j ^ 'VENr/L A TION

' V.

. . RELIEF .

'O TANK , . r'  % .Y UNITS (4)

=- --

~' '

PRESSURIZER '/

h i;.

,. ?

l

' , '. ~:~. _m  : ;ye - ' ?. ^

[

~.;-' if. 4 .s ,

• HEAL ~-

/ e U"ULATORS I') '

l,

. ; TABLE

,y . . % ~.'$ s......sg ,

,, EL 692.0 HEAT EXCHANGERS ~l EL 700.0J *

, N' 'a .

e. , ' i' VfNr/LAr10N UNITS '(2) _

~'_* ,.. .. *

• s PERSONNEL LOCK W

Figure 3.8.3-7 Plan-Lo.er Compartment k FIGURE 12

LOTIC-3 CONTAINMENT CODE i

4 N0DE CONTAINMENT MODEL CONDENSATE /REVAPORIZATION MODELS LARGE BREAK (TOTAL REVAPORIZATION)

SMALL BREAK (CONVECTIVE HEAT FLUX)

• WALL HEAT TRANSFER MODEL MODELS SUMP RECIRCULATION SYSTEM O

S e

FIGURE 13 9 -

8 I \

ti t

5 4

1 l

! l l

Sorays  ;

a a a A ,

, ,-.3 3

Qwall Upper Compartment I

g-

• s, hsl f i u.C.

m3 ,n3 4  !

I.:

Drain .

4 ~  ! Ice *s' 's 1ce Concenser m,n

-  ; I 3 a

. One way flow Sall -

f{6

, through ocen area L, 3 on operating deck

- Condensate & A >

Ice Melt lm'h sI s

I.C. Sumo

, ma ' 'a '

I .C. i Oraini j s

v _

i i

^ ^ * *'

(RCS break #10w, Dead-Ended *s s _

l

~53.h3 accumu1a~.

Comoartment m,n l

a a

• L:wer Comoarr.ent release, cs boil-off Air Addition Tao)le s lo drn,1c  ! we at Acdition Tabia

! L.C. i' C r- o wall 1

'Orain sumo I

U.C. - 1. - -- - h7 f

r Orain L.C. Sume

' l FIGURE 3.3 MASS MD ENERG'l Fi.0W DIAGRAM FOR THE CCMPARU.ENTS FIGURE 14

• l

i LOTIC METHOD OF SOLUTION SOLVES CONSERVATION OF MASS, ENERGY, AND MCMENTUM FOR UPPER, LOWER, AND ICE CONDENSER REGIONS ONCE NEW LOWER COMPARTMENT CONDITIONS ARE DETERMINED, CONSERVATION EQUATIONS, ARE SOLVED FOR THE DEAD-ENDED j

COMPARTMENT AND FOR THE FLOW RATE BETWEEN THE TWO COMPARTMENTS

{

FIGURE 15 i

TYPICAL CONTAINEVf TElPERATURE TRANSIENT (DRAINSf0DELLED) l l

1. LEVEL REACHED IN S.G.

-J U. . SPRAY ~

[ DECK FANS ON t /

REDUCTION IN I.C. DRAIN FLOW I e

- Qi l b '

B .,

f 0 100 200 300 f00 500 600

. TIE (SEC.)

4 FIGURE 16.

4

.i i

! +

1 k

.i PODIFICAfl0NS TO THE CONTAINfBT MODEL i

WALL HEAT TRANSFER MODEL

(

CONVECTIVE EAT FLtjX MODEL s

j ICE CONDENSER DRAIN K) DEL i

i

, DEAD ENDED COMDARTENT MODEL 1

l i.

1 .

a i

i 4

I-

' FIGURE 17-r l

r l

,. 4 1

h WALL HEAT TPANSFER PODEL ORIGINAL L0rIC McCEL i

//

q=h y Taganti (i SAT

-i All)

PbDIFIED LDTIC MODEL

~

Cong g y- y,jj WV ulk

. ef Cono m a,,. /

1 hcon v

= f(T~.,, , i,,)

k = f W . ,, , L ,3 9

i FIGURE 18

CONVECiIVE HEAT FLUX FDDEL-ORIGINdL LOTIC MODEL hcono = Scono h,y

},or,{i,g}

j,,, -

t J

ItDIFIED LOTIC MODEL

' ha,,,( q . 7;,,)

=

lE)ConD = h,A '*%3ror1[y p, p y,,,] [ , _

p p

r. cn .

t e

9 e

FIGURE 19

i l

ICE CONDB1SER DRAINS APPROXIAMATELY 20 ICE CONDNESER DRAINS i

f DRAIN ELEVATION IS ABOUT I40 FEET FROM FLOOR

-DRAIN PIPE IS 1 FOOT IN DIAMETER

-FOR TYPICAL & LB TRANSIENT) DRAIN FLOW VARIES FROML 4000 LB/S To. 500 LB/S l

D

- FIGURE 20-'

s b- k 3

)

• O e

e ICE CONDENSER DRAIN FDDiiL CONDENSATION OCCURS AT THE SURFACE'0F THE STREAM

~

FLOW IS WELL MIXED Q=hAAT

-EDDEL AS A WALL AT A CONSTANT TEMPERATURE'

-A IS THE SURFACE AREA 0F THE STREAM

-h IS A CONDENSING. TYPE HEAT TRANSFER COEFFICIENT

-FIGURE'21.

4 4 . _ . . .

O 9

,8 O CALCULATION OF THE STRENi FLOW AREA B = n (P x L) = ao le x vo) = 800 P WHERE P IS THE PERIE TER OF T E STREN4 w

\

i 4

.p s .. / ._

t

,(-- - .

, _)

~9 I-h -

. x 1 FIGURE 22

_ _-_-_--------l--- " - - "

, .1 i

E DIFIED LOTIC DRAIN NDEL-i t -WALL WITH A VARIABLE AREA e

e e

=hcuo A CTeuw -Tsg)

$U N j '

a br* 230 F

.) -

4

.I -

r o

i e

' b i

I 4 -

-FIGURE 23 .

a ,

% . k

- p , . u -  :. . ,, .- -

l

- a 1

DEAD ECO C00'PART?ENT 00Dti 4 IDER C0FPARTPBIT DEAD ENDED C0FF 4

, I i

I I  !

N01FIED LOWER 00PPfTIBIT l l- .

4

. i .__ . . .

)

FIGURE-24 1 _

g m -2 .- # g .%

..- o CATAWBA RESULTS

- 102% POWER

- 0.86 FT BREAK

- MAXIMtE AFW FLOW

- FSAR HEAT SINKS

- MAXIMLN S.G.-INITIAL MASS T'%X=

i 32I4*F (LOWER (RPARRB(T)-

4 e

' FIGURE 25

, -s-

. . . , , . . . _ _ _ _ _ LL_.----_=E-------------E---- --

i f

f 1

i ADDIT 10fML M) DEL CONSIDERATIONS 1,

f'

~

4 I -

WALL HEAT TRANSFER MODEL b

DRAIN MODEL . .

DEAD ENDED COWARTfENT MODEL i

4 i

1 i

) , .

h t

I i

1 I'

. i A

~

t ,

)

1 , .

-FIGURE 326-y d ap g. vwe' g ,w A 2 ---see .. ., 1 u- rw i ** *. 7f,,^+. .ay r-.- 5 yd e. p 4p-y )

ATTACHENT 2 ,

i 00TLIE OFlmE REPORT i

j I. Introduction II. Mass & Energy Release Modeling i

III. Containment Modeling IV. Action Plan V. Appendix 1

VI. References i

G e .

4 J

e 1

i e

l

. , , - , - . - - , _ . . _ , - . . .,, J,- .-- .......,,.-.-..

. . ,,, , , #. . [.. .. .. ,.. . . , . , , _, ,

l I. Introduction I

l l During the Containment Systems Branch review of the Westinghouse topical report, {

l

" Mass and Energy Releases Following a Steam Line Rupture",WCAP-8822 i (Proprietary) the Staff noted that heat transfer to steam from the uncovered portion of the steam generator tube bundle was unaccounted for and questioned the effect upon the calculated mass / energy release and the subsequent effect en the containment temperature response. Westinghouse responded in a letter to the Staff (NS-EPR-2563,e c bruary 14, 1982, E.P. Rahe to J. R. Miller) that it had determined the impac; of the effect by conservatively treating the maximum amount of superheat to be the difference between the primary coolant temperature and the steam temperature. The letter noted that there would be an insignificant effect c7 dry type containments and that, based on the conservative model used, there would be an expected increase in containment temperature for ice cordenser type containments. In the Containment Systems Branch Safety Evaluation Reports on the topical report and the Catawba Plant Safety Evaluation Report, the Staff required that a mere refined steam line break analysis be pe.' formed to determine the effect on containment temperature which might impact the environmental qualification envelope used for safety related equipment.

Since that time, Westinghouse has investigated the effects of tube bundle heat-transfer from the viewpoint of a more refined modeling approach. Subject to the final review and approval of the NRC Staff, the efforts and results obtained to date indicate that there is little impact on the containment response from the effects of the additional tube bundle heat transfer to steam.

~

t O

e 1

I s

II. Mass and Energy Release Modeling A. LOF1RAN Computer Code

' Mass / energy releases are calculated using the LOFTRAN code. LOFTRAN is a FORTRAN language, digital computer code, developed to simulate transient behavior in a multi-loop pressurized water reactor system. The program simulates neutron kinetics, thermal hydraulic conditions, pressurizer, steam generators, reactor coolant pumps, and control and protection systems. Up to four independent loops may be modeled. LOFTRAN is used for analysis of non-LOCA transients and is documented in Reference 3 The model of importance to blowdown calculations is the steam generator model. The primary side contains multiple nodes to model the tube bundle.

The standard _LOFTRAN steam generator secondary side model, (Figure 1), is effectively a one node, two region model of saturated steam and water.

Heat transfer is assumed to occur only to saturated water. If tube uncovery occurs the amount of surface area available for heat transfer is accordingly reduced. The LOFTRAN code incorporates a more detailed steam generator model which is used to predict tube bundle uncovery.

B. LOFTRAN Model fcr Superheated Stean The LOFTRAN code has been modified to account for heat transfer to steam from the uncovered tube bundle region. (Figure 2).- In the modified version of LOFTRAN, all heat transfer occuring in the uncovered region is assumed to add superheat to the steam exiting the steam generator. The primary side temperature in the uncovered tube region. is-conservatively assumed to remain constant through the nodes which are uncovered. In reality, there will be a drop in temperature due'to heat removal.to the secondary side, but this is expected to be small due to~the low specific heat' capacity of the steam and due the high primary side flow rate.

The heat transfer coefficient used in the uncovered tube region is discussed in the Appendix. This correlation bases the heat transfer on the-difference between the tube wall surface temperature' and the bulk steam temperature in the region. In the LOFTRAN modification, the conservative asstrnption is made that no credit is taken for either a primary film heat transfer resistance or a tube metal heat transfer resistance. Therefore, the wall surface temperature of the tube is assumed equal to the. primary fluid temperature.

_ The modified version of.LOFTRAN automatically determines the proper- number

~

of. steam generator nodes for;the superheat region of steam in the -

generator. . . The variable node capability is applied to both the primary-and secondary side. At .each time step durin6 the tube uncovery, the modified -

~

LOFTRAN' code makes a general _ evaluation'of the uncovered-tube region (e.g.

~

steam flow rate, uncovered tube heat transfer area, estimated heat transfer coefficient,' etc.) and determines the number of nodes to_ be used in _the '

. subsequent calculations < The total: heat' transfer for the uncovered tube-region is determined and accounted for in the primary temperature transient -

e k-- ,.-.

calculation. The superheat/ tube uncovery modeling is applicable to all steam generators.

Figures 3 through 6 show typical results fer a 0.86 ft steamline break from 102 percent power using the modified version of LOFTRAN. Figure 3 shows the fraction of tube uncovery versus time with uncovery of Loop 1 (faulted) starting at 152 seconds into the transient. At approximately 300 seconds, the uncovery transient reaches an equilibrium point where the steam flow out of the steam generator matches the auxiliary feedwater flow into the steam generator. Additionally, the tube uncovery transient fer '

Loop 2 (non faulted) is plotted but shows no tube uncovery for the entire transient. Figure 4 presents the steam flow transient for this case.

Figure 5 includes plots of both the superheated steam enthalpy and the saturation enthalpy for the Loop 1 steam generator. Figure 6 includes the Loop 1 temperatures for the steam generator tube inlet (primary side),

steam exit temperature (superheated steam), and the saturation temperature for the steam pressure.

C. NOTRUMP Model Comparison The NOTRUMP computer code (Reference 4) was used to verify the LOFTRAN modeling of superheat. The computer code was originally develcped to analyze transients of secondary systems with two-phase conditions. In the past, it has been used to analyze various transients in the primary and secondary coolant systems. .NOTRUMP has recently. undergone major revisions to enable it to model non-equilibriun nodes (i.e., separate liquid temperature and steam tempera'ture modeling). Using NOTRUMP, the steam generator can be broken down into sufficient nodes to model the nonequilibrium effects of the steam generator, as well as the tube region during uncovery. NOTRUMP can model all modes of heat transfer associated with a steamline break transient, including heat transfer from the uncovered tubes to the superheated steam and the feedback effects between the primary and secondary sides. The two phase mixture level calculation accounts for primary to secondary heat transfer and the swell associated with rapid depressurization of the steam generator during the blowdown.

A comparison of LOFTRAN and NOTRUMP blowdown results is presented in Figures 7 and 8. The mass releases shown in Figure 8 show excellent agreement. The LOFTRAN prediction of superhe st enthalpy is slightly higher than NOTRUMP, "while the predicted time of tube uncovery 'is somewhat later.

NOTRUMP shows a chugging effect during the uncovery phase-of the blowdown.

This is believed to be in part due to oscillations in the flow link b'etween ,

the downcomer regicn and'the steam dome region. (The flow link is the  !

drain path for the moisture separators to the downcomer region.) With the flow direction towards the downcomer, superheated steam goes into the downcomr re51on and is condensed. . This alternates with a flashing of a l portion of .le water volume in the downcomer region. This raises the  ;

pressure of the downcomer, resulting in a flow reversal in the link with '

saturated steam from the downcomer mixing with the superheated steam in. the dome. This mixing results in the variations in the superheat enthalpy seen in Figure 7. Although LOFTRAN does not show the enthalpy variation since the detailed modeling of the downcomer and dome are not included,xthe overall agreement with NOTRUMP is very good.

D. Effects Of Analysis Assumptions The effects of superheated steam are dependent upon the occurrence and extent. of tube uncovery. The major parameters affecting tube uncovery are:

initial steam generator inventory, auxiliary feedwater flowrate, assumed feedwater system failures, and protection system errors. Variations in these parameters are in the process of being evaluated for their effects on the containment temperature response (Figure 9).

Refinements in the mass and energy release modeling (Figure 10), are being evaluated and several areas show a potential for reducing the degree of superheat being generated. Some of these areas are:

- Evaluation of liquid-steam interactions such as the phenomenon of tube support plate flooding and heat transfer across the , tube wrapper from

. the superheated steam to the auxiliary feedwater flowing down outside the tube wrapper.

- A more detailed steam header model in LOFTRAN.

- Modeling temperature drops in the primar,r superheat nodes. l

- Evaluating other void correlations for use in predicting tube uncovery.

t I

III. Containment Mode, ling A. Description of Containment The general phenomena taking place inside an ice condenser containment during a steamline break transient cari be described utilizing a typical ice condenser elevation drawing (Figure 11). Steam is discharged to the main (or lower) compartment where heat is removed by the internal structures, steam flow to the ice condenser, and the. ice condenser drain water. The dead ended compartments are the regions which are located below the ice condenser and outside the crane wall (Figure 12). Air is discharged from the main compartment to the dead ended compartment and ice condenser so that the resulting steam to air ratio is that region is much higher than in dry containments. At ten minutes following the containment hi-2 signal, deck fans are actuated which direct air flow from the upper compartment to the dead-ended compartments. Most of the safety related equipment is located in the dead-ended compartments although some equipnent and cabling are located in the main compartment.

B. Containment Models Figure 13 outlines the major models and assumptions utilized in the LOTIC-3 containment code. In the currently approved version of LOTIC-3 doctznented in Reference 5, four distinct regions of the containment are modeled; the lower compartment, the dead-ended compartment, the ice condenser, and the upper compartment. Two condensate /revaporization models are used depending on the size of the break. For large steamline breaks, 100% condensate revaporization is assumed. For small steamline breaks, a convective heat flux model is used which calculates partial revaporization during the transient. The wall heat transfer model utilizes the Tagami heat transfer correlation for condensation heat transfer and the convective heat flux model derived from the work of Sparrow (Reference 6) which calculates the convective heat transfer for small steamline breaks. The sump recirculation system is only modeled for the large break LOCA

transient containment response.

Figure 14 shows the four regions modeled with'the ' mass and energy flows that can be assumed in the analysis. The Catawba nuclear plant does not have lower compartment sprays and they are not modeled in the analysis. Superheat heat transfer is conservatively assumed to be zero for the steamline break

containment analysis. In the model described in Reference 5, wall heat transfer is not modeled in the dead-ended compartments although these regions do contain structures which will remove heat. The analysis does include the upper-compartment sprays, flow through the ice condenser, deck fan flow, and flow to the dead-ended compartments.

LOTIC-3' solves the conservation of mass, energy, and momentum equations for upper, lower, and ice condensor regions _(Figure 15). After the new lower compartment conditions are determined, conservation equations are solved for the dead ended compartment and the flow rate between the compartments .is determined.

i Figure 16 presents a typical steamline break containment temperature transient i that is calculated using superheated steam blowdowns from the LOFTRAN code and the modeling of ice condenser drains as a heat removal source. The transient shcus that initially the: containment temperature increases rapidly during the

.- o i i

blowdown. When the upper compartment sprays actuate there is a slight decrease j in the main compartment temperature. The temperature then rises slowly until i ice condenser drain flow decreases to the point at which time the temperature l begins to rise again (approximately 250 seconds). B is rise in containment i temperature coincides with the steam generator tubes uncovering at 152 seconds and the maximum superheat occurring at approximately 250 seconds. The steam generator level stablizes whe'n the auxiliary feedwater flow is equal to the steam discharge at approximately 300 seconds. The containment temperature then starts decreasing with decreasing decay heat. At ten minutes, the deck fans actuate which results in a rapid decrease in containment temperature.

C. LOTIC-3 code Modifications Four modifications have been incorporated in the LOTIC-3 containment model which are (Figure 17);

1) wall heat transfer model

2) convective heat flux model

3) ice condenser drain model

4) dead-ended compartment model D. Wall Heat Transfer -- -

The modification to the wall heat transfer model is described in Figure 18. In the LOTIC-3 model, only condensation heat transfer, utilizing a Tagami heat transfer coefficient and a temperature difference between the wall and saturation, was previously modeled. ' he modification includes a convection term '

with a conservative convection heat transfer coefficient and a temperature .

difference between the containment atmosphere and an appropriate interface temperature. The Appendix presents a more detailed description of this model.

E. Convective Heat Flux Tne modification to the convective heat flux model is described in Figure 19. A term has been added to the convective heat flux model to account fer the feedback effect from including a convective term in the wall heat transfer model. The Appendix presents a more detailed description of this model.

F. Ice Condenser Drain Model In an ice condenser containment there is approximately twenty drains cxiting from the ice condenser into the lower compartment at an elevation of about forty feet above the compartment floor. The drain pipes are one foot in diameter.

h e drain flowrate is calculated by the LOTIC-3 containment code. For a typical small steamline break transient the drain flowrate varies from approximately 4000 lbn/see to 500 lbm/sec during the gimeframe of interest. The temperature of the drain water is approximately 130 F (Figure 20).

Figure 21 presents the assumptions and the basic model used to estimate the heat removal from the lower compartment atmosphere to the ice condenser drain water.

It is conservatively assumed that the drain water stream does not break up prior to reaching the floor even though many of the drains have equipnent and structures located-below them. Therefore, heat transfer is assumed to occur at

I i

the stream surface only. It is also assumed that the stream surface temperature is at the saturation temperature of the containment.

The heat transfer to the stream is:

q=hA/LT wher'e h = condensation heat transfer coefficient A = surface area of the stream AT = appropriate temperature difference The calculation of the heat transfer surface area is described in Figure 22.

In order to model the drains in LOTIC-3, the drains are modeled as a wall heat sink with a surface at a constant temperature (see Figure 23). currgntly,in the version of LOTIC-3, the surface temperature is assumed to be 230 F which is close to the containment saturation temperature. The drain surface area is calculated at two points in time during the transient; early in time with a high flowrate and later in time with a low flowrate. To ensure conservatism in the area calculation a 10% reduction of the surface area was assumed.

As described previously (Figures 14 & 15), the LOTIC-3 centainment model did not account for wall heat removal in the dead-ended compartments. To obtain a conservative estimate of the tenperature transient in the dead ended compartment, the heat sinks located in the dead ended compartment region along with the heat sinks in the lower compartment are modeled in a ecmbined volume (see Figure 24). This " modified" lower compartment model is used to determine a conservative dead-ended compartment tenperature transient. Since the lower compartment will be hotter than th'e dead-ended compartment, this methodology results in a higher tenperature in the dead-ended compartment then would be expected.

G. Transient Results With the modifications described for LOFTRAN and LOTIC-3, the previous FSAR limiting case for Catawba was reanalyzed to determine the unpact of superheated steam. The case selected is a 0.86 square foot break at 10g5 power (Figure 25).

The peak lower containment temperature for this case is 324 F. This temperature is calculated for the lower compartment cnly. It is expected that the -

dead-ended compartment temperature will be significantly lower.

In addition to the model modifications incorporated in LOTIC-3, Westinghouse is

~

pursuing further improvements in the areas noted on Figure 26. One area is in the wall heat and mass transfer models. . Since condensation is a mass transfer type phenomena, the heat and mass transfer should be linked. 'D11s approach has been used in Reference 7.

An Dnproved drain model is also being investigated. This improved model will calculate the drain surface area as a function of flowrate. It will also calculate the average temperature rise of the drainwater. This model will mere accurately. represent the actual phencmena in the containment.

i l

WESTINGHOUSE PROPRIETARY CLASS II V. Appendix WESTINGHOUSE STEAMLINE BREAK BLOWDOWN AND CONTAINMENT ANALYSIS METHODOLOGY The following sections describe the Westinghouse methodology for determining the containment response for a steamline break incorporating the effects of superheated steam. These sections describe in detail changes from the methodologies described in References 1 and 5.

I. Steamline Rupture Mass / Energy Blowdown Analysis A. LOFTRAN and MARVEL Computer Modeling Mass / energy releases can be calculated using either the LOFTRAN code (Reference 3) or the MARVEL code (Reference 8). The LOFTRAN code is used for non-LOCA FSAR accident analyses. The MARVEL code was specifically developed for assymmetric transients such as steamline breaks. These two codes are very similar because they were developed in an interrelating fashion and much of the model'ing is common to both codes. The MARVEL code was used in the development _ of Reftrence 1 because LOFTRAN at that time was a lumped model which was used for sputetric loop transients. Furthermore, for steamline break analysis purposes, MARVEL contains a model for water entrainment. However, the current version of LOFTRAN is a multiloop version which also contains a water entrainment model. With the development of a multiloop version of LOFTRAN and the inclusion of an entrainment model, the use of MARVEL has been generally discontinued. This enables the use of LOFTRAN as a single system analysis code for non-LOCA transient analyses. LOFTRAN is used in the analyses presented here.

. The model of importance to blowdown calculations is the steam generator model. The primary side of the steam generator contains multiple nodes to model the tube bundle fcr both the modified version of LOFTRAN and MARVEL.

Heat transfer calculations from the primary to secondary side are identical in the two codes, although the methods for initializing the heat transfer resistances are slightly different. The secondary side is effectively a one node, two region model of saturated steam and water. Heat transfer is assumed to occur to saturated water. If tube uncovery is predicted, the amount of surface- area available for heat transfer is reduced.

Both codes contain a detailed-steam generator model which is used to predict tube uncovery. This model calculates the liquid volume in.the steam generater shell and acgognts fcr the detailed steam generator geometry. The [ ] correlation is used in both codes to predict the voiding in the tube region, although the correlation is modified for use in LOFTRAN. In Mt.RVEL, tube uncovery is calculated based

• 1 on comparison with the actual water level and the height of the tube bundle. In LOFTRAN, the user specifies either a water volume in the steam generator corresponding to tube uncovery, or a void fraction in the riser section of the steam generator at which tube uncovery begins. 3 Both codes have similar models accounting for reverse heat transfer, thick metal heat transfer, feedline flashing, and safety injection system operation. Auxiliary feedwater flow can be input as a fraction of nominal feedwater flow, although LOFTRAN has an additional capability to model auxiliary feedwater flow as a separate system. For analysis of double ended ruptures, MARVEL accounts for the volume of steam in the piping 1

downstream of the steam generators in the blowdown calculations. In LOFTRAN, this consideration is added on to the blcwdown mass and energy results by hand. For split ruptures, which the analysis presented here addresses, the steam piping masses are handled identically in both codes.

In summary, LOFTRAN and MARVEL are very similar codes, and either can be used to calculate mass / energy blowdowns. To demonstrate this, a comparison of the blowdowns for a typical case is presented in Figures A.1 and A.2.

Figure 1 presents the mass release rate fcr a .86 ft2 split rupture from 102% power. For this case, Figure A.2 shows the saturated steam enthalpy as a function of time. This blowdown is typical of results used in FSAR analyses prior to the modification noted in this report for the LOFTRAN code. As can be seen from the figures, the results are extremely close..

B. LOFIRAN Model for Superheated Steam As mentioned previously, the LOFTRAN code has been modified to model heat transfer which may occur in the uncovered tube bundle region. This effect is modeled in both the faulted and intact loops. In the modified version of LOFTRAN, all heat transfer occurring in the uncovered region is assumed to add superheat the steam exiting the steam generator. The temperature of the primary coolant flcwing through in the uncovered tube region mode is conservatively assumed to remain constant. Realistically there would be a drop in tenperature due to heat renoval to the secondary side, but this will be small due to the low specific heat capacity of the steam and due the high primary side flow rate.

The heat transfer coefficiggt

• uged in the uncovered tube region is based on the [ Ja' . The heat transfer coefficient (U) is calculated by the following expression: _

a,c F

This correlation is presently used fo5,8uperheated forced convection heat transfer by the [- ] computer codes. Additionally,

r

, l

. l l

this correlation is based upon the heat transfer from the surface of the tube wall to tne average bulk temperature of the steam. In the LOFTRAN modification, no credit is taken for either a primary film heat transfer resistance or a tube metal heat transfer resistance. Therefore,the wall temperature of the tube is conservatively assumed equal to the primary fluid temperature.

~ s, c -

- (1) 19 e

e O

e

- s

- s.

e G

G n ~

O

_-.mm_ _ _ _ - _ _ _ _ - - . _ _ _ ~ . _

._._...__h.__im._2__m .,__ _ - ,

The modified version of LCFTRAN automatically selects the prcper number of steam generator nodes for the superheat region of steam in the generator.

The variable node capability is applied to betn the primary and secondary side. At each time step during the tube uncovery, the modified LCFTRAN code makes a general evaluation of the uncovered tube region (e.g. steam flow rate, uncovered tube heat transfer area, estimated heat transfer coefficient, etc.) and determines the number of nodes to be used in the subsequent calculations. Each node is evaluated to determine the steam temperatura exiting the node with a convergence criteria that is based upon the total number of nodes used. The exit steam temperature of one node is used as the inlet steam temperature of the next node.

i The heat transfer calculation to determine the outlet temperature of the node is based upon the following expression:

Q = UA*(T 1-(T + Tin)

• Ns*Cs *(T out-Tin) where Q = Heat transfer to the steam U= l3'2 Tpri = Primary node temperature Tout = Steam node outlet temperature Tin = Ste m n de inlet temperature M3 = Mass flowrate of the steam C3 = Heat capacity of the steam A = Heat transfer area in the node including both hot and cold leg sides of the tube bundle The total heat transfer for the uncovered tube region is determined and accounted for in the primary temperature transient.

C. Blowdown Sensitivity to Plant Conditions The effects of superheated steam are dependent upon the cccurrance and extent of tube bundle uncovery. Parameters affecting tube uncovery are:

initial steam generator inventory, break size, auxiliary feedwater flowrate, ano the single failure assumed. -

The initial steam generator inventory depends upon the measurement errors associated with steam generator level and upon initial power level. Steam generator mass increases with decreasing power, thus, breaks intitiating from low power levels will result in later tube uncovery.

Larger break sizes result in faster blowdown of the steam generator and l earlier tube uncovery. l l

1. 4 1

f:

l .* - e Large auxiliary feedwater flowrates only delay tube uncovery, but will also cause the final equilibrium steam generator level to be higher. This equilibrium condition corresponds to the point when the break flow rate is equal to the auxiliary feedwater flow rate.

The single failure assumed in the transient may Lupact the amount of water supplied to the steam generator. Auxiliary feedwater runout will increase the amount of water supplied to the steam generator. Failure of the feedwater isolation valve will also cause extra water to be supplied to the j generator as the additional mass between the isolation valve and the check valve flashes to the generator.

t t

a ep O

J D 1

e T F T

II. Containment Analysis ,

A. Wall Heat Transfer Model The original LOTIC-3 wall heat transfer model is based on the stagnant Tagami heat transfer correlation. That is, q"=hTAGAMI(TSAT-Tyggg)

1. N hTAGAMI = 2 + 50 MSTEAM AIR h(TAGAMI, MAX)=72 BTU /hr-ft - F This model was developed for saturated steam in the presence of large amounts of non-condensable gases. In the lower compartment of an ice condenser, most of the air is swept out of tne lower compartment through the ice condenser and into the upper compartment. Therefore, after about 30 seconds, there is almost no non-condensables in the lower compartment. Typical values for the condensation of pure steam are in the range of 1000 to 3000 Btu /hr-ft2 oF (Ref. 5). The correlation used in the modified LOTIC-3 code is in extension of the Tagami correlation fer nearly pure steam.

q"=hCOND (TSAT-TWALL) hcond = 2+50 MSTEAM " AIR h(cond, max)

• A maximtzn value of ( Ja,c was chosen as a conservatively lcw condensing heat transfer coefficient in a nearly pure steam environment.

In addition'to this modific'ation, an additional term is needed to account for the convective heat transfer from the superheated steam to the condensate film.

This convective heat transfer is dependent upon whether there is condensation occurring-on the walls. If condensation is occurring, the correlation used is:

where:

m m ul[s a,c If the wall temperature increases to above the saturation temperatu're then the convective currents will be reduced such that the correlation used is 9"conv=hconv(Tbulk-Twall) where:

[

Ja ,c

.' '~

Th'us in sumary, if Twal1<Tsat men

[ j a,c If Twall > Tsat, then the correlation used is:

( j a,c 9

1

l 1

'B. Convective Heat Flux Model i When the containment atmosphere is superheated, the containment temperature is a strong function of the amount of steam mass in the atmosphere. Thus the amount

- of mass condensed.cn the. heat sink surfaces is a key parameter. The actual amount of condensate formed is -

j

! "cond

• 9cond/hgg Unfortunately, with the _use of a heat transfer correlation based only on test data (such as Tagami or Uchida), only the total heat transfer coefficient is obtained. This total heat transfer coefficient includes both the. condensation heat transfer and the convective heat transfer. Based on the work of Sparrow (Reference 6), ths Westinghouse Convective Heat Flux model in the original LOTIC-3 code calculates the rativ of the convective heat transfer to the condensation heat transfer. Therefore the calculation of the amount of mass condensed is

-( ja,c 4

J~

In the modified LOTIC-3 model, the anount of superheat . convection is calculated.

The ' amount of convective heat transfer at saturation is not known explicitly in this model. Therefore, in the modified LOTIC-3 code the original convective heat flux model will be used to calculate the fraction of convective heat i transfer for saturated conditions. 'The actual correlation is - s,c where, (q /q is determined from original convective heat flux model and q conyt3c ggg)$$ bunt of convective heat transfer colculated in the wall-heatEF9Xsfar-model In summary, the modified LOTIC-3 model is consistent with the original LOTIC-3

.mo ed l in its calculation of the mas condensed. 'Dae only difference is usat in the modified LOTIC-3 code, the amount of superheat convective heat transfer is known ' explicitly, while in the~ original LOTIC-III model, only the ratio off convective heat transfer to condensation heat. transfer sis known.

9 4

4 e

i b @

f r y m e , w

• 4 IV.

References:

1. . Land, R. E., " Mass and Energy Releases Follcwing A Steam Line Rupture" WCAP-8822 (Proprietary) September, 1976 and WCAP-8859 (Non-Proprietary).

2. NS-EPR-2563, February 14, 1982, E. P. Rahe of Westinghouse to J. R. Miller, NRC, " Additional Information on WCAP-8822".

3. Burnett, T. W. T., et al. , "LOFTRAN Code Description," WCAP-7907, June, 1972 (Proprietary).

4. Meyer, P. E., and Kornfilt, J., "NOTRUMP - A Nocal Transfer Small Break and Gereral Network Code," November,1982, WCAP-10079 (Proprietary) and WCAP-10080 (Non-Proprietary).

5. Hsieh, T. and Liparulo, N. J., " Westinghouse Long Term Ice Condenser Containment Code - LOTIC-3 Code," February, 1979, WCAP-8354-P-A Sup. 2 (Proprietary),WCAP-8355-NP-A (Non-Proprietary).

6. Sparrow, E. M., Minkowycz, W. J., and Saddy, M., " Forced Convection Condensation in the Presence of Noncondensables and Interfacial Resistance", Int. J. Heat Mass Transfer, Volume 10, 1967.

7. Corradini, M. L., " Turbulent Condensation on a Cold Wall in the Presence of a Non-condensable Gas" Nuclear Technology Vol. 64, pp 186 - 195, February, 1984. ,

8. Krise, R. and Miranda, S., " MARVEL - A Digital Computer Code fer Transient Analysis of a Multiloop PWR System," November,1977, WCAP-8843 (Proprietary) and WCAP-8844 (Non-Proprietary).

9. McCabe, W. L., and Smith, J. C., " Unit Operatons of Chemica1' Engineering",

3rd Edition, 1976.

LOFTRAN - MARVEL COMPARISON

.860 FT2 BREAK AT 102 PC POWER 2000.0  :  :  :

f

/J --

1750.0

\

0 e

1500.0 --

s

$~

1250.0 - -

2 3 --

"1 1000.00 - -

w --

G 750.00 - -

x c: 500.00 - - MA R V E L --y /

LOFTRAN --

cc 250.00 - -

' 12/05/83 0.0 O o O O O O

O O

O O O -O O O O O O O C O O O O O O cn (D LO O e N .:t T!ME (SEC) .

FIGURE A.1

^ -

. . . = - -. _-~_.---- =--

.. -j

, LOFTRAN - MARVEL COMPARISON

.860 FT2 BREAK AT 10'2 PC POWER .

1300.0  :  :  :  : .  :

- 1275.0 - - --

5 N

D a 1250.0 - --

1225.0 - -

0 RAN ..

1200.0 __

b--

5 1175.0 - -

=

<C W 1150.0 - - --

M S 1125.0 - -

1100.O i _i i i i a O O O O O O O b b .b b b O O

. O O O O

_O ~ N m ..- LD D

. TIME (SEC)

FIGURE A.2 v ~ s ymc