ML19220C631

Discusses GRASS-SST Calculations for TMI to Explain Release of 20-40% Total Core Inventory of Fission Gas

ML19220C631
Person / Time
Site:	Crane
Issue date:	04/16/1979
From:	Rest J ARGONNE NATIONAL LABORATORY
To:	Marino G NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES)
Shared Package
ML19220C630	List: ML19220C629 ML19220C631 ML19220C632 ML19220C637
References
NUDOCS 7905110348
Download: ML19220C631 (2)

Text

' us M dEng nein-..-...

RGONNE NATIONAL LABORATORY 9700 Sack CAss Atu. Arpet.tircis 60439 Tdglot 312M72 5026 April 16, 1979 Dr. C. Marino Fuel Behavior Branch RSR/NRC Mail Stop SS1130 Washington, D.C.

20555

Dear George:

This letter su=marizes the GRASS-SST calculations I performed,by NRC request, for assu=ed Three Mile Island accident scenerlos to es ate the fuel te=pera-tures required to explain the release of from 20-40% of.he total core inventory of fission gas.

Two suggested scenarios were used to perform the calculations. The first scenario consisted of an irradiation at 6 G/ft (average core power rating) at an average fuel temperature of 1200*F for 50 full power days followed by (1) a relatively instantaneous reduction in power (1.2% of nominal due to reactor scram) and a fuel cooldown to 550*F occurring in about 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br />'s time, and (2) a heatup of the fuel at a heating rate of 1* F/S. The second scenario was similar to the above, but dif fered in that fuel from approximately 25% of the core which was irradiated for 60 full power days at about 12 KW/ft with an average temperature of 2550*F was considered. The objective of these calcula-tions was to determine at what fuel temperatures GRASS-SST would predict 20-40%

total gas release.

The results of the analysis, which were transmitted verbally by me to you on April 11, are as follows:

1. Predictions made with GRASS-SST for the 6 KW/ft fuel indicate that a maximum of from 5-10% total gas release would occur at fuel temperatures between 4700*F and the fuel melting point (s5160*i).

This result was obtained assuming that no extensive grain-boundry separation occurs in the fuel. A substantial amount of ga s release from the grains was calculated to occur as a result of the heatup, but this gas was trapped on the grain surfaces sad edges, and hence was not released to the exterior of the fuel.

(The predicted fuel swelling due to the retained gas en the grain surfaces and edges was too small to cause appreciable long-range inter 11nkage of the porc,sity).

2. For the 12 KW/ft fuel 20% and 40% total gas release was predicted to occur at fuel te=peratures on the order of 4500* and 4800'F, respectively. Again, this result was obtained asstening that exten-sive grain-boundary separation did not occur.

7905110 N

%W nt UAtRsay dOcvp Arpvt UstRsmcs Assoonm

g

\\(

4

.7

/

W. V. Johnston M 0 4 25 whgre*P is the power in kw/ft H the HTC from clad to steam in BTU /HR-FT -F, and T in degrees F.

The equation is plotted as T versus P in figure l for various values of H.

Note that for an initial rod power of 10kw/ft, and a decay heat fractional power of 0.01 (i.e., P=0.1kw/ft) the. maximum temperature obtainable for H's of 5, 4, 3, 2.5, and 2 are ll60F,1320F,1560F,1770F, and 2080F at steady-state. Pickles 1mer's

~

curve of zircaloy oxidation damage (based on Kassner's work) indicates that to oxidize clad sufficiently in an hour so that it wguld not sur-vive thenr.a1 shock, the temperature would have to be 2010 F.

This requires a HTC less than 2.5 for the TNI conditions. This is an extremely low HTC and will probably only occur in very stagnant steam condtions.

It would appear that any means of assuring a inodest f'ow of steam would have kept the HTC near 5.0 and prevented serious damage to evetn the hottest rods of the core.

.j G. P. Marino Fuel Behavior Research Branch Division of Reactor Safety Research cc:

L. S. Tong D. Hoatson H. Scott R. Van Houten W. Beckner Y. Y. Hsu R. Sherry M. L. Picklesimer S. Fabic Z. Rosztoczy J. Volglewede G. L. Bennett C. Johnson DISTRIBUTION SUBJ CIRC CHRON BRANCH-RF GPM RF.

GPM Cr 98 100 RS R : W ; F_B_,

GPMarino/pg su = = >

4/ /79

.m >

Jac rons na wn uxcu aue t..

.....,... ~................

I

N Weijl i!n!' j:w!

_i

-i ntdq"in !*i ntihnnii I -: :,7.j. !

i. ti I-i.

+ '

t' I I

U.

7hj:

N ii i

I i

t}

i I

!ii!

l*i I.

th l

No,:{. p.j f I.ul hT !

l-

!i!

!l l!'

i!

"Mqian ieb;1 M!

l.! j$./N.,.JlFs: "i.p Eh n

l P. lT !,[}l+i g*

jl T

+

+.;7

l

.; h Ijt,l}. ; +h, ii

".S

},:' i r.

i Iii t

, i il i!.

I,n-f '.i u

'LT W

.7 l

-M ijInM

,1 j

!i

".p:

nJ i

i h ki

.a!

Ii "Ipmlflfp,p L t!

[

'T f

l i

.; n lI !i

'.. i HiTllyfd;I n

!i l

i!

t

!a@ 11nniu,.ap tIU!

l lip I

t

Mm.-i/Hi!

if dp.un HmH H

~ d.

H I

h i

"i I.

ifi

.p n i! u p

I' tii l.i,,::F!ii

,i l

"i.

"h.l.A,!%gn-:

7 n< !

n S' of.

l r

.;;I Il.!!

I i

ii p

l.

ii.N' 7!'

I g ".i lIn r b h

jn

!I!l

[l i

f mI.

i i,..

H.

q. l*

h' {,!

.nt.pvJ

' Y. I<. pup'

!H H

l:

n h.- :!c

"!I ht' n! P I

i.

i.!

H.L'

p..

t

- e 7./m.-i f?u I.

e./+si

. i

}! ! n r'

.]

lI'mn /. i. M,li l

'}

"!h"it i

.i :

f na. hth,.p ij,.

in t

ia hTpi!

+l In n i, L]il

e. i}I j

fi+h1m i

b LaI

!+

i, r

~.

".! rwL d

f,imye l.l

!+ htn 5".

I

'i " i

',d 'h,lp.i:i. p u li f

lIh t i...

!.l i

!4 i.n t i

U 'p.

Unn ! hbM"i

'I

.i]

. I i

' ' h

• 4:t i

t R97~iam7eI

~on'?!Un!n.htM !l4in !!i

..n i I

, '/

l F - "

M. n.... -

~,

!!Jptik

!l i d;

+

" 4 1=N,.J

.!n! IP h}i+

fl n

r i

pnt-j,H kb.nl.

!fl id hu u -

o/.

1 iZ"

" t iaNMp:m!

.F i

i

!lIin r

p i. Y.;: mu.im.j e -- i.S l.

• h+n p.

'. NlU u.q' l.h;T!

! n+ p,i i Ih

.!,[

l 4r. i/" U F. iq,y: =!

d J

i j

4 i i0d. Ej,t.I

.p.

l

.p l

r tfi

c. M i i f m h. k f n,l p

',.h + h,hu.

l Il dr.

-. i i!

1 mi I'p!

pf tn U

]r {H n

~

i; ;h,.p" i

j,

)

. -. '.. {!;!.

IIi!!;l'!

P, :.

+ Y*~n!c n.

"l T

i:~.

t' "I

+

!i c

. n f,'. I.!tl. ;i II!

!)

l!e I n FiU dj p r:

i!*I

!l

!;i t!

+.,. pph - l,Ii.ilil FHy!.l

~F p. j l}ill'I4

~ i.,f 'mf.!.

"[

i I

+!.-

Walh !.!"

pn f "n

~

.t T.

t i

i e..

i.

t il['l'!

Mh H

I 1n "i

! 1 i

j.

f I i TMg 4

i r n!.

l.

h,, n. ol..

t i*!

h i[ig 'wI itjin/ S !+

rl'i n 7

T Milq.}l f

f i

I ii i

aH 1 1t!

Y,y9. iu.M,!!!

I i.

.!+il[ 'It l h

'i i

e: y. - A n IX" O+ P'h. n. ",./

t'..U Id!ti7 ii i N.J,.g.rl.7. f 'ii; !

, h+5

!i

l

>"H!

i i

p.

iF :

i

.j

.I'lll i '

- - n :. N i

?j nJ !

lii i

7 n,,. d{"i l

r f

i; lm i,

P,l

{ h;l

-+

e

.,r

~

l!l

!}

Ny m r-

.. 1r7".=

i i!I i

i 7n. +,.:y

"*"'nq.1'Q i i }ti l

~

l.

", Iy:

1 I

li M[

"I

!l.

i!+'lfn i

IZuon nn,e

+N>

.l

'al*n

,..i l'

,p.

it il

.j %nd

• 7i

!+l*u r.

LM ep.4I,ap.4ipyn.p.

lIl.lj i

4 i!n'iu n i

!.i t! I tn ii t.

ti N,qq n(

i i i

n; I

l t

I i

g i

r

!t!

h1p,..fo;plh[n I'i t

i.

"o mh.H e l

l dprI.(.

r Ll.

"a i 1 :.

i n

r i

' N.

t i

'. m 7.@dmIpy" e4-id[d,..

"n I /"un -.

l a

I I

n}

liMlnH lA4 II

!.I qvw';:r.n a

n II t

i 4ut!

M iI 1r 7.). j /7n7 !'i' t

l'ui"n i

i.

,t7 u

I'

+i pF! '/

il h,, 11!:3d".'iWIAii l

i n.

ti n

"+

D7phyoF.H! hip-l l"

It l i

l p '!'-;nmW t!a, S7 ie L +

i mfo/":!p4'i[ymps.L]+': ! i II, iliLitI li n

ii i!

+} n jil

- i h y

h-t

.j p

vh'.lg:'/:',.N' dl' l lt

!ii i

+

dI, r.

!t hp' H!4n i.'

I.!

i4, H -lr i:W ?.J U,.

H

!h

.i

}l Y',pa!

I!qI. / iq+ui g'. " l.'MlIn'/d*li L

i iilklu ih I*

i n

I...

/ " F, I :

I'p h'. n :

i}!:!

t!

h;'n H'l

'eh f

Ii1 iM;.,j,

4"m,/*,. p4 '

'n iiW

+WW.i!!lI Il+

y ItIf t !

- i e!

l+l i

..i,i l h..!i Dt h r.

q-"

. t N ;'. n j.

l tl!+

!l im l

Nlgj!-

.i[i4, n,u.

a![ 'l'n I

n1 p"/',

a p ;p. [!

l r

-r I

I r/ - iitn-t P '

f."

n!!i i

!Oi i

+

I iiu:4' i

.ie.

p. lpg!

!l+Hu f

I

~

m "9M qFl.hiap,.j;jJ

}

l}

4 H

, i i i v D-imH.

.i d '

l.

l i

.i t

". 4.

In i

p.

n l

': H I

T.f:pfaiun*+:.yd[f,.

I.

i 'i

+N+44

.- ",i 9 -

i,

!d.

f

', - kpg%1 n

i ht} h M

!l i nI ap.h U,.

Da$i : i

• Y ib" fi lu.j

-}ii+

i n

n I

i H,.

T

.i i4".

'ti iY

+.lhj /' j.n.q,.

~!ii ti TIit:

t i. j f ih f j ih..

lj,

. + htlfl j i !c {}8. t t i ir t 9 I I l l . : p lli'! i ' i4! i iH I:t::+. pji,.'~0'Y7:h'WMi!mjj i.!m I j.h+ 4 ld i Un en4 i hN,1, '. j . ~ i n!L.. sb A"M 3[I'l I i; i i I4H dI J '.

iil.!i i..

9 '. d.pp '% -i ' fip '{ji }!+ li. i. r'( [ ij ii oi L i + t ny g d l,]i M[p' i '..ti h d 3 F d f m h ul. T! ! 4;1 i. i,Un.{u "M., N,g 'i

1. mie}!F ?n t -l,g*. i;ii

- 6dr :.i r ri ".n Yj7.S o j,I'.h N p i l+ t i.9pu l } i: g~4

• 4 j!

i l-l H, D.:,uFn H f d nh j} + i

i. j p m ! n n

- i! { ,jiiii +}i !j T y, i 7 'j 4 'dfi l!fl lj j i U b.< m:.. -;#yj u t,it-. -] l r j .i i pi: L,i f i i i t 4 If p<rJEMM.7!m[3u}Tdi3h Ii ]Hn ndy n = j j i+ / :i I i UNt

4!

y d + j U n q: t d: t + . x + r dr i,ii",. 'I: ih I. 1r l o y I.h.hii i',

i l

i t i i+y!'i "nM@i 1!; !I i I. l4L ."".L... N',i " !ir }! {! ,n ii I !. i bII I. t

. b. U I -;

l, t I' .t I. el Iii D. ; .t I, dT h ' [ 4~ j

T n.iNl.l, a

t: i :n. I ii o atiHm ','h.e. iTI t lh l I'I " e t! .g l tgi],!r'y I' -., }T@H i { l l.-[1 g ' "z ,r l [" "ig..f _ I i m I '} %g8 i - 'gk. q,gw 'i% C(g~3 " WE t

ATTACHMENT 6 [v. ns. 'o, UNITE D STATES (f % NUCLEAR REGULATORY COMMISSION S k g,Y.o/ i .' i W ASHIN GTON. D. C. 20555 %, V APR 6 1979 MEtiORANDUM FOR: File FROM: William D. Beckner Separate Effects Research Branch

SUBJECT:

PRELIMINARY CALCULATIONS OF THREE MILE ISLAND CORE DAMAGE Initial attempts have been made to bound the possible damage to the core. During the first 1.8 hours of the transient, evidence exists that the core was covered with saturated fluid with high flow rates. Thus, there is little probability of damage during this time. At 1.8 hours, the main coolant pump was tripped. Superheat in the hot leg and subcooling in the cold leg was noted as well as a pressure drop. This suggests little or no coolant flow and an at least partially voided core. To bound the damage, an assumption of a completely voided core with a heat transfer coefficient of 3.0 Btu /hr-ft'. F (lower limit based on calculations of Y. Y. Hsu on 4/6/79). Under these assumptions and making use of information developed by the Fuel Behavior Research Branch (letter, Marino to Johnston dated 4/4/79), onlyportionsofthecorewithinitialpowersgreaterganabout 11-13 Kwt/ft would be damaged. This represents about 25% of the core volume near the center. This should represerit a good upoe_r bound on core damage,at this time. During later time periods (8-14 hours), similar dry conditions could have lasted for significantly longer periods and steam temperatures may have been higher. Areas with powers as low as 10 Kw/f t could also have been damaged. Up to 40% of the core volume could have been damaged. These numbers are scoping calculations based on a limited amount of information and some assumptions. However, it shows that significant damage was possible. Updated calculations will be performed as more information becomes available. Iu'dh b b' William D. Beckner Separate Effects Research Branch Division of Reactor Safety Research cc: see next page 98 102

Memo to File. cc: - L. S. Tong W. V. Johnston

5. Fabic A. W. Serkiz G. P. Marino Y. Y. Hsu M. L. Picklesirner D. Hoatson K %

98 103

ATTACHMENT 7 Adiabatic Heat-up Rate of a Fuel Rod (Marino) Let P(Z) be the power at location Z in a fuel rod. At location Z for a foot of fuel rod P(Z) kw _ P(Z) kw 3412.9 BTU /hr = 3412.9 BTU /hr/ft ft ft kw 3412.9 P(Z) LT = IH = Z 0; (PC Cp1 + M; Cp;) (M,, Cp " + MUO; Cp UO .0791 BTU /lb

• CP 2 =.0717 BTU /l b, Cp

= M in one foot = - (.01542)

• 1 = 7.469 x 10 ~ f t 3 x 647 lb

~ = 0.483 lb. UO ft-2 Mzr in one foot = 2- (.01792).0265 x 406 = 0.1 lb 12

Thus, i=

3412.9 P(Z) = 80226 P(Z) *F/hr = ?2.28 P(Z) F/sec (. 48 3I. O'7TT +.1 x. 07 91 Thus, 22.28 x 11 x.011 = 2.695 F/sec_ for an 11 kw/ft rod @ 1.l'; decay heat. 0;

• PATPRO VALUE AT 1650*K - above 1800K (2781 F) Cp INCREASES AND DOUBLES AT 3000K ( 5000 F). THUS AT HIGHER T's, T SHOULD DECREASE.

98 104-

ATTACHMENT 8 Adiabatic Heat-up of Fuel Rod (Picklesimer) 27.7 x-106 w/ core

• 36,800 rods = 752.7 w/ rod a

= 0.721 Btu /sec. rod Decay heat available = 5.007 x 10 3 Btu /sec/ inch red 3 UO

= 10.08 g/cc (92.5% density) = 0.3642 lb/in 2

vol/ inch rod = { (.370)2 x 1 inch = 0.1075 in / inch rod 3 mass UO / inch rod = 0.03915 lbs/ inch rod 2 Cp UO = 0.0777 Btu /lb F 2 uh = 3.042 x 103 Btu / inch rod F uG2 Zr clad c = 6.54 g/cc = 0.2363 lb/in3 vol = m x 0.405 x 0.025 = 0.03181 ink Cp = 0.0830 Btu /1b F 0 600 - 1100 K = 0.1315 Btu /lb F @ 1100 - 1300 K = 0.0908 Btu /lb F >l300K rass Zr/ inch rod = 7.514 x 10 3 lb/ inch rod aH = 6.24 x 10' Btu / inch rod F aH = 30.42 x 10 4 Btu / inch rod F U0 2 ~ cH rod = 3.666 x 10 3 Btu / inch rod F aH a/ailable = 5.007 x 103 Btu /sec/ inch rod aH rod = 3.666 x 10 3 Btu / inch rod F F/sec = 1.366 F/sec adiabatic 98 105

ATTACHMENT 9 Calculation of % of Decay Heat Required to Raise Steam at Same Rate vs. the Fuel Rod G. P. Marino e Heist required to raise dx length of fuel rod at position U to AT F is: aH =C aM aT

Thus, p

r r Heat required to raise dx of steam aT divided by AH IS r r 5 s aHs = C AM P s C r "Mr (M and AMr are masses of steam and rod dx long.) aHr p j fuel Rob 5 S 5 C aM =As. dx ps ~ C c / r f p f,. dx:f C 2-r o dx aM =C tr - s p ccc g s s / AHs = A csC AIir~ [ G=.o/79[b U f c m[C r o +C 2r o 0 p f f p ccc k . St.B " As, (.5 2) -,( ) - 1.23 x 10 3 ft2 = area of steam / rod ~ =.52 BTU /lb = density of steam and specific heat of steam c s p I r = (.01542 ft), C =.0717 BTU /lb, C c =.0791 BTU /lb = fuel radius and f P p specific heats =.0265" =.0021 ft, cf = 647 l b/ f t3 = clad thickness and density of fuel a P = 402 l b/ f t3 = density of cladding c AHs - f t 1.23 x 10 3 ft2 x 2.04 lb/ft3 x.52 BTU /lb ~ g n[.0717 BTU 2.377 x 10" ft2 647 lb f t + 2'1.79 x 10 2f t..00221 lb ft . 0791 ~ 402] .0307 = 3.07 x 10 2 = 98 10f;

ATTACHMENT 9 (cont.) Of total power is Pdx kilowatts, then 0.97 Pdx can heat rod while only 0.03 Pdx can heat steam to same temperature. Thus,. 03 Pdx = h (T -T s) = heat flux required to steam r 2nydx '03 " (.03) X P X 3412.9 BTU /hr h = 2nT = -T 2r x.215r :T r 12 b = 909.5P BTV/hr-ft;- F aT If P = 0.1 kw/f t b, 90.95 BTU /hr-ft:- F aT For h = 5, a ai of only 18'F will allow steam to heat up with clad. 98 107

ATTACHMENT 10 Calculation of steady state steam temoerature reached in a reactor of power P at flow rate u. G. P. Marino our [ s ~ } q a d x. m An element of steam dx thick moves upward with velocity u. The heat picked up is /* 2nr dx h (T(x) - Ts(x)) - at aH = o where at is the time spent over du of rod. at = dx/u where idul = idxl u I 2-rdx h (Tr(x) - Ts(x )) du aH = O T But at steady state Q = h (Tr(x) - Ts(x)) = P(x) = F 2:r 2nr where F is the average rod power. U T(u) Tin + I 2-rdx P du = O WSF&r u s gsdxes. A is the steam area / rod ams = Tin + /" P du __ Tin + P O T(u) = = Cp u.A'c" Cp'u.Abe' O b Cps =.52 BTU, u= 7 ft/ min = 420 ft/hr, as = 2.04 lb lb ft3 As = 1.23 x 10-3 ft? 98 100

ATTACHMENT 10 (cont. ) T(u) = Tin + FU X 3412.9 .52X420X1.2 3x 10- k2,04 3 = Tin + 6.23 x 10 P U For F + 6 kw/ft and a decay heat of 1" T = Tin + 373 U 0 If Tin s Tsat s 600 F T = 600 + 373 e 0 at 10 ft, T = 4337 F at 5 ft T = 24650F Thus, steam rising from a boiling core can reach very high temperatures at steady state no matter what h is used. The AT's (Tr-Ts) calculated by my earlier calculation (attachment 5) would add 600*F to the T's computed here. Thus, high T's can easily be achieved in the THI core by essentially heating both steam and rods. As t'ne steam is heated its density is changed which will slightly affect the results. Also if u is larger (say by a factor of 10) the heat up will be only 1/10 of that computed here. We have seen earlier that if the steam gets only 3% of this power its temperature will rise with the rods. Therefore, it seems that adiabatic heating is a good approximation to the rod temperatures in the TMI core. 98 109

ATTACHMENT 11 This note is prepared to sumarize speculations on what might have happened at Three Mile Island (TMI) during the first few hours that led to the initial core damage. Information presently available is not sufficient for rigorous calculation of a mass balance on the water in the TMI system to be made. It is not yet clear how all of the water got out of the TMI system to allow the boil-off and heat-up of the core that occurred. If the steam generators had been full of water on the primary side at the time core heat-up occurred, they would have provided much water to keep the core cool so it must be assumed they were nearly empty. It is not yet clear where all this ma~ss of water went. (Note: B&W at the ACRS presentation of 4/16 was using a much larger " break area" than had been inferred from relief data in the FSAR - if the B&W break area of 0.02 f t2 - 0.025 ft2 is used, it explains the mass balance discrepancy noted above. We do not at this time know the basis B&W used. It is noted, however, that this area is consistent with the inside area of the 2-1/2" pipe on which the relief valve was mounted.) The following sequence of events appears to explain what might have happened: After the initial pressure and temperature transients, the pressurizer relief ficw did not stop. The plant depressurized to about 1350 psig at 6 min. after the accident started. This was the saturation pressure for a hot leg temperature of 585 F. After an initial nressure and temperature rise for a few more minutes, the auxiliary feedwater pump valves were opened at 8 min. and a cooldown at about 200 F/hr at saturation conditions continued out to about 20 min. The pressure then stabilized at 1015 psig (548 F hot leg) for the period until the pump trips occurred at 74 min. and 100 min. During this stabilized period, decay energy added to the primary coolant was being removed through the pressurizer relief and through the steam generators. There does appear to have been a net loss of water mass from the system during this stabilized period. A loss of mass within a fixed volume means that the fraction of the loop volume which was originally liquid was being reduced while the fraction which was steam was being increased. Where was this steam bubble forming? 98 110

2-The geometry of the plant becomes important in answering this question. Note on the figure (attachment 1) that the hot leg from the reactor rises 46'-8-l/2" above the reactor vessel outlet to the inlet at the top of the steam generator. The hot leg temperature - pressure equilibrium will occur in this hot leg pipe at the top of the steam generator. It is at~this point where tne qnversion of water to steam (i.e., boiling ~. or flasning) will occur - at une system high point. The 46 feet of water in the hot leg will prevent bulk boiling in the core below (although some local boiling probably would occur). Since the top of the pressurizer is 7' 8" below the top of the hot leg pipe centerline, it is probable that tne pressurizer is also water-filled (unless the pressurizer heaters are on - which they should not be). The picture during this stabilized phase appears to be one of pumps drawing liquid frem the bottom of the steam generators, delivering it to the reactor vessel cold leg inlet, flowing down the downctmer, then flowing up through the reactor core where it picks up decay heat, then out the hot leg to the top of the steam generator. A side leg is flowing into the pressurizer where liquid is being released by choked flow through the pressurizer relief. At the top of the hot leg, flashing of some hot water to steam is occurring and this two phase mixture enters the top of the steam generators. In the steam generator, the phases separate and the steam phase remains in the top portion of the steam generator and the liquid goes to the bottom of the steam generator where it enters the suction cf the primary pumps to complete the circuit. A large mass flow of water is being circulated by the pumps through the reactor vessel and steam generator where steam is accummulating at a rate dependent upon the mass and energy balances of the system. With time, the level in the steam generator drops (due to the net loss of mass from the system). As the bubble in the steam generator grows and the level drops, the suction head on the pumps is being reduced and as a result, the flow is decreasing. Eventually, the level in the steam generator will get low enough to permit the formation of steam bubbles in the suction line which leads to pump cavitation and was the probable cause of pump vibration that led the operators to trip the "B" loop pumps at 74 min. It is entirely possible that the operators did not realize they were " boiling the steam generators dry" during this period. After the "B" loop pumps are tripped, the "A" loop pumps continue to furnish water to the reactor vessel and on up to the tops of the two hot legs. However, the "B" loop pumps are no longer supporting the difference in head between the low level in the "B" steam onerator and the high level in the hot leg. The "A" loop pump discharge tries to flow in 3 directions: 98 lii

~. - through the reactor vessel and up the "A" hot leg and return to the "A" steam generator; - through the reactor vessel and up the "B" hot leg to the overflow into the "B" steam generator; - around the top of the downcomer, back through the "B" pumps and . into the bottom of the "B" steam generator. The net result of this complex redistribution process appears to be a higher "B" steam generator level and lower "B" hot leg level and probably a lower "A" steam generator level. Pressure changes will also occur due to th changes in levels. Some flashing may also occur. The precise level changes would have to be calculated based on a better knowledge than we now have of the starting volumes where the "B" pumps were turned off. It does appear though, that the core at this time will still be covered and will be cooled by the forced circulation from the "A" pumps. Between the 74 min. point where the "B" pumps were turned off and the 100 min. point where the "A" pumps were turned off, the "A" steam generator level will continue to decrease until the "A" pumps are aiso tripped. When this is done, the differences in level will readjust to seek a comon level in the system. The levels in the hot legs will displace downwards into the reactor vessel and down through the core or through the downcomer vent valves, through the pumps, and into the steam generators. A corinon level will be established in the steam generators and reactor vessel. If that level is below the cold leg pipes, the core will boil the gooi of water in the reactor vessel out through the pressurizer relief. As the level drops below the tcp of the core, the upper part of the core will begin heating up (essentially adiabatically per Marino's calculations). If no water is added to the reacto-vessel, the core level can boil down at a rate of 64 ft/hr initially, dropping to 32 ft/hr at half core height, and to zero at the bottom of the core. N 9 98 l4i m

~ ~. (& e p.o e g O, 8 J' A T, lHih 4 ,,E,, CENERATOR j 5 _.J I e .8 f g* [J y g 2 I-8 i REAC181 ~ I I* _I 79'-Bf 8 ~ l~ ' [Q g 9 l 8 1 8'-o" 5" ] e 1 42'- 8 8 UCh y 'j ~ ~ l 4 g LANT 8 N t _) ) ( ,)~ '4-4-+-)I ( ;h-dj.' b y b I $ga l -jss :+9 4 l 3,..e-

a. o-e l'

29 '- 3 g 1 8 E l I 1 s 6'-O*' ] e 4 [ 1' W_ 1 o V 9'-o" l rN N i g O 4 0 "7 pf.f 98 14 3 REACTOR COOLANT SYSTEll ARRA ymtEE MILE 131 AND NUC1J.AR STATIO _.__., -~~~--- . -. A._._. ; 4

Possible Systm 71ould Distribution Prior To Pump Trip At 1.8 Hours Pump Keeps vessel, Hot Leg And Pressurizer b Pull Whi).e SG Empties. j ~# 3 1000 ft l +, :. " s \\ \\ N \\ \\_ F c* s v-d litAR - \\ A stut: Aft 1 o 1500 ( stACT81 \\ g3 g-3x

3. g

,f_o 'A w 42'- e.. 4.__ __c ) y, s } ~% \\ o A M ,r; LN ~ q s hg 77 ) \\s N ~~ s s \\ 8 d 31'-0" N s s \\ Ytt3 N%,'\\ d ~24'-0* 1000 N N s 6 ~ l'

• • '- ' I 4 js

%\\ e, ( N 1 Q .) [ s' L m q 98 1W. $W5 _______[

f Possible Systec Liquid Distribution A.fter Pur:ro Trio At 1.8 Rourq p. Fluid Runs Back To SG Froc Yessel And Hot Leg. Natural Circulatio.n Blocked. E g /- b p ,, o y,13 f f-s' ~ v - o c a

sui, 1

i stutt& Tit 6, r ,~& s, ~ l ~s ' }, e,t atitter 1 l-3 i-t 1 $e 'b' t a-1 ' g s -o- {g g 3 s-az'-sir void q' i (jv+h 1 \\f w e f,\\ \\ - = =" / l / J 3 r-o-s s / k NfNN ,.I s s ~ 3 as -21 m rt g w-/ m rt /'q s r e a 2._ s r 3 'L 3000 Ct: hL -~ , r

g. o" g,

g wm4 981i5jg3 j '~ .. - - :1 __}}