ML19289G118

Discusses Mechanical Condition of TMI Core.Rod Ballooning Leading to Rupture Expected in Highest Power Areas of Many or All Fuel Rods

ML19289G118
Person / Time
Site:	Crane
Issue date:	04/13/1979
From:	Marriott P INDUSTRY ADVISORY GROUP
To:	Ditmore D INDUSTRY ADVISORY GROUP
References
NUDOCS 7906260469
Download: ML19289G118 (8)

Text

a n

! Strictly Private

/

13 April 1979

• E " u Fei 'T b h

To:

D. C. Ditzore c.7 From:

P. W. Itstriott s

e

Subject:

EOWiICAL CDXDITIO3 0F THREE MILE ISLAMD CORE On April 12 the Industry Advisory Group (IAG) requested an independent judg=sut

~

i of the mchanical condition of the Thn e Mile Island (TMI) core, assuming a sequence of events regarding core cooling postulated by expats in the IAg.

A sd=mry of that information, as I un darstood it frca yea in our telecco.

is presented in Attach::ent A.

This ac=u conven our judg=ent for your use in cacparison to others'.

It has not been subjected to indecencent, f atamal review: I leave it to you to handle it accordingly and to put it in proper perspective.

Core Hestup. With the 1.aited and speculative nature-of the sequence of events postulated in Attactzzent A, we have simply assu=ed adequato cooling (clad testperature at saturation) through 116 minutes fol"; owed by convective ccoling to superheated steam untti quenching.

Thermal radiation of peripheral rods to the core barrel would be significant for those rods. but insignificent to the central region of the core except insofar as it would abet natural circulation of steen within the core. Our judg:aentn. unsupported by detailed calculations for this exact sequence. is that the cladding tmperature would incressa to at least 2600F under the flow conditions postulated by IAG.

')possiblyhigher. Assu=ing a ifnited aza.7t of natural circulation of stem e

inside the vesse), the 1 wr powered reg'ans would be heated to similar J'

temperatures.

For simplicity, our core mechanics) considerations postulated a peak cladding tempemture increasing linearly fraz 600F to 2600F from t = 116 minutes to t = 126 minutes, then holding constant at 2f>00F untti quenching, with in-creasing amounts of cladding. reaching 2600F through the transient. Quenchin at t = 176 trinutes would red,ce cladding taperatures to saturation (6) gin seconds. The five-ainutar heatup at t s 195 minutes is probably of swa~h>y im ortance.

This postulation has ths following limitations which my be teportant to the IAS's judgment of the nochanical condition of the core:

(1) It omits the possibility of early (t - 100 to 116 ntnutes) heatup.

and perhaps cladding perforntions, high in the core while good cooling is still taking place at lower elevations. An estimate of this could be made. if IAG's thersol-hydraulic adytsers could speculate on.

vessel inventory and (even battar) void distribution, during the first tasenty minutes or so.

4..

~

(2)

It does not consider at what tico the entire core would become essentially adiabatic and heat to higher tcwwstures.' h.er..

.IAG's postulated event sequence, the bestup sequence we postulated.

K p{ J p Q$g and IAG's estimate of 30-45% of the core's zircaloy b

g corroborate each other approxinstely.

252 m I

7906260 M hmu

?%9f a;

/

D U IC W i"TIV a !G 2<<

Fuel Rod Perfora tion. Assuming a rod in'emal pressure of 400 psi at 20*C, _

,ghe calculated rod internal pressure at %2600*F is M320 psi.

This yields a clad hoop stress of %16.100 psi.

Based on 2ircaloy rupture data the rods 1

wccid be expected to balloon and perforate at cladding temperatures ef N1500*T.

4 al Bssed on General Electric full-scale single-bundle ECCS heat transfer test data, the location of rod ballooning and perforation would be expected to be within + 6 inches of the paak temperature regior, of the rod; the location on any given rod would be rondas within this range.

Ass:. ming clad heatup fras 600*F to %2600*F in ten minutes and constant s

temerature thereaf ter. 6nd using the Baker-Just rate equation, the clad unil cculd be expected to be about 47% cxidized in one-half hour, about 67% oxidized in one hour, and fully oxidized in about two hours.

This calculation considars only external oxidation; the extent crf oxidation can be expected to nearly l

double over a short length in areas where the rods have ballooned and per-q forated, exposing inner clad surfaces to an oxidizing enviroreent.

h Fue1 Rod Balloening.

As noted above, for the reds experiencing the assu:ed y

ele _v_at.ed ta=perature and pressures icosed by the transient, rod ballooning veuid be expected.

The zaximum r.apnitude of the expected ballooning w>old g

! c NTCC%. i.e., the rod initial diasseter would be expected to double.

(This

> stituta is based on AXL 76-121 LWR Safety Rasaarch Program. Quarterly Progress g

MDort July-Septed>er 1976).

Essed on full bundle tests conducted by GE and i

others, coplansr ballooning leading to extensive flow blockage wcrald not be expected; however, as stated above, the ballocning would be expected to be preferentially located within rcughly a one-foot section of the axial locatica of peak cladding ter::perature.

fuel Fod Distortion.

We cammt cremt on the possibility cf rod distortica or bowing curtng the ccre heatup because of cur unfamilf arity with the core mechanical design.

In GE full-scale ECCS heat transfer tests. some bowing of rods did occur at tecperatures several hundred degrees lower than postulated here.

It should be noted that the fuel rods in these tests hsd larger outside dieseter and cladding thickness than TMI's.

The possibility of rod distortica should be considered.

a Clad Embrittlessent and Effect of Quenching.

Due to the clad heetup, significant.

oxidation would ba expected. The orittle behavior of stabilized alpha-phar,a

ircenius oxide would be expected to result in fragmentation under quench condi tions.

The 10CFR 50.46 oxidation lir.it to preclude this condition is 177 for LOCA application. AML* has suggested a limit of 28% under slow quench conditions.

It should be noted that in the experiments discussed by AXL. msay of the rods which were intact following quenching failed during post-tast handling.

Post-test handling failure has also been experienced in fuel rods subjected to similar tenperatures in tests performed by EGaG Idaho. Inc.

The posetlated escunt of cladding oxidation and e=brittler:cnt if present together w:th severe rod distortion, could have resulted in occhanical fa11ttre.

3 of affected rods during the hastup. quenching, or subsequent pressure or J

flow transients.

• Argone kational Laboratory. " Mechanical Prvperties of Zircaloy Containint Q

Oxygen." USMRC Zircaloy Cladding Program Review Meeting. April 25-25. 1975.

4 b.f

= /*'*

hggR*s,'t'e,p Wii di (

t{ ' [t

, ' J I--

,7 e

e e

,L

'i et

' t; Q u u

• n

,g

,5

\\'

3 g

x v.

~

~

6trictly Private

~

ca etusions. Based on these scoping enlaatims the folletting core cechanical cocdition is postulated:

O (1) !!cd balloening Icediq to rupturn expectad !n tis highest-power _

areas of can or all fu21 rods. Ba11 coming not eqccias to be '

coplanar.

. a (2) Clad perforetion in many or all fwl rods resulting from operatica post the rupture capability of Zr.

(3) Clad oxidatica sufficient to cause fregaentatica c:tder gocach conditicas, particularly if aggravated by rod bearing.

P. U. k rriott F

e d

m I

L'u? ' ' '$ %'.b '.". 3 ".N N O.' W? '/'e' L LTY/N*,' GW - s.C'. s' *.e*aU' **M *'A ' WD *

.R /.b d'4'XLC 4'e s*.; W5A w W. ~4 '/k* * '4r

)

s e

9 O-9

~

Strictly Private ATTAQPiENT A g

THREE MILE ISI.Ano-2 CORE HEATUP:

POSTU!.ATED SEQUENCE CF EVENTS

,T:

(Refennce Telecon DC Ditz:nre to PW Rartfott, 4/12/79, 1326 PST)

~

Tire After Event, Min.

Event I

100 "A" prir.ary coolant pugs tripped ("B" poups had been tripped previously) 100-116 Cooling by boiling in subcooled Ifquid and high-densf ty froth 115 D2nsity of fluid in core begins to decrease rapidly. Hot leg coolant temperature begins to shtxt superheat.

116-146 Cooling by low density fmth and (not acch later) steam 146 Unexplained spike in com fluid density 146-176 Mo net inflow or outf1cw of ste:m in vessel (cooling by natural circulation of steam insida vessel) 176 Rapid quenching 176-195 Cooling by boiling in subecoled liquid and high-density froth 195-200 Brief recond heatup 200 Rapid quenching THREE MILE ISLAND-2 CORE HEATUP:

CORE PRESSURE / TIME HISTORY (ReferTnce Telecon DC Dit=cre to PW Marriott. 4/12/79, 1420 PST)

Time After Core Pressure Event, Min, psig 60 1100 75 1045

~

90 1110 105 1000 120 800 135 67s (Iczest)

~

160 eco k-

, 155 1050 A

180 2200

.)

e 252 J

t 4 - 1 1 & 7 'l _

3 D. Tliesin y 'R 4

Estim& oF /wp& ?';.i foy. 5llz h.l g

'~

j o l

2r - /f.6 / eae fi w

. W

,.s Etsis:

Invekoyg/ ekd ox/g) 77, %,cfl ~5)

.;,1..)

T'. -

a.1 of 4'-/-

Er &n

-> +

0 //$ bum - /ho es7%abl to 4 ave.

i release d z.s - 4. o x/o ' pu Eeuivak,&

G

2. s - +. o 2 a--/4 o ryt.

/>

A 9A U. 4 MM 4-/-7f z

epah t G-

//.5 % 2r-N'io rxit

~

@ /000 N I babblefYC5/

s=.)

4-/-7f

' v'ah.t G

/ 5 %

gr-/f a yxa.

N uV8ed k05 ai~ ftvo NBe*F 4-/-7f eganeJat hsi ea

2. 5 % Er-f/2.o '

ru.

h dAass/n g h

3-28 46/

+ ken lefdkn dm - awnd.

be.

- yale 4 te ch'on k bd6[e 2-m waanred

/Bm ff

• R 'facc }7

~

__y_

W 4-l-7f Cf vsfed $

$$e Er-~ //L O rxn DT+'

~

Mer M =M.S%

W c=

4 1. 5 %

@*&O fM D.

a7 au

P Cour sxir 72n.oe u rver O

I 2

3 4

S c

7 3

3 io s2

1. 3 N

sy

\\

~

A 28G AS1 AF9 MS Ah Af1

-]

)

i D

M1 ASG -

2 56 l

l e

2F9 2P7 30.1.

??S r

,I p

2 94 25X Ac5 I

2,Py 2PA 21R 352 k2

.\\ 3 02

' M C, -

SK, l

H 421 Sgr 399 2rg,

m Ki AT, 3cs-AES 222 N.C. -

D4

.a K A.0 A91 Sco 2 93

~

A93 A9G QP9 A21 O:

l p

Ns-

,J

.p

]l R.?

g ygg

~

t(

2 3

+

5 G

7 8

'"~ 5 " --%

to I2 s3 H

Is

$m/ef *~2 ?O'f Cvc bcs.i bal ~ S M *]

'j y )

, {'} ;,

ATTACINENT 81 CORE FLOW BUDCKAGE ASSESSMENT D. C. Ditmore 4/13/79 FROM CORE HEAT R\\ LANCE April 10, 1979 Reactor Conditions:

-- 2800F t

in

=

eore 0

t out = s 285.to 399 F (with two TC's reading belo the core inlet eore and thus likely erroneous.)

0 288.5 F in annular region outside very center of I core exit

's

( Q,- 8.5 F) core 9eore - 3MW 6

Normal operation - 4 pump core flow 137.9x10 lb/hr Single Icop - one pump core flow 1/8 - 1/4 x full core flow * (next page)

Unblocked Core Heat Salance Sincle Loop /Sincie Pump a) 1/3 of full core flow C

Qco re * *cere p atcore BTU 6

(SMW) (3.413x10 Hr. Min.)

o teore - Q,.n7,

=es1.1 F 21 Cp 4,e (137.9x106 lb/hr (~.9 b) 1/4 of full core flow

~,2.20F ateore Blocked Core (Current Condition) Heat Balance - Single Loop / Single Pump Cp Atcore) blocked

(# core Cp atcore) unblocked (W

e re

=

corehlocked (a e re) unblocked S blocked Wcoreunblocked eore) blocked A unblocked at

( zs teore) blocked a,8.50F t

(A eore) unblocked ~1" JAblocked 1-2 F Range

.1} -.24

• * /\\ unblocked 8.5DF 89 - 76% f1'ow blockage

^'

('

89 - 76% core blockage in peripheral re'gion.

Because of high expectation that normal single pump operation core flow is closer to 1/8 of normal four pump operation core flow, the expect ed,yesul't

'G) 1G

/vt i

t. J

ATTACif iEST F4 - Continued O

is closer to 89%.

-1 9

i

• Analyses by T. Hott and B4W indicate ~ 1/8 of the normal four pu:np flow through

}

the core under one pt=p operation, most of the flow by-passing the core and fic ing in reverse n: ode in non-operational loops.

O e

k s

e O

D E

84 m

9 e