ML19308A501

RO on 740409:during Filling of Sodium Thiosulfate Tank,Upper Section of Tank Buckled Inwards.Caused by Negative Internal Pressure.Repairs Under Way & Procedures to Prevent Recurrence in Preparation

ML19308A501
Person / Time
Site:	Three Mile Island
Issue date:	04/11/1974
From:	John Miller METROPOLITAN EDISON CO.
To:
References
NUDOCS 7910250704
Download: ML19308A501 (3)

Text

O '

V G) m-

:m

.~-

. - x. ...,:,.n

. n.?.-.m. ~ g.m g

e pa.3.h u,n.m3mp9mmmE9.m4'

. nu: m a # + cf.: h aa gq,m,yusum w %

M Th---. -

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

p}/ 0:,.& ,'yk

~

~

.t Q,:EL= .. .. s .

h o a,

--m

/

/

/,*

*r,,) xussa n < nw'

• i

/

METROPOLITAN EDISON COMPANY sem.v vi ut.mc crmi n:ureur.w:anwourmu/ /

ItC r<n!NG, PENNf ;Y1.VANI A 1MC3 TCLEPHONE 215 - 22') 3G01

'T OFFICE tiox M2 April ll, 197h Dr. D. F. Knuth, Director Directorate of Tsegulatory Operationn g oo D Office or RegulxL5cn United Ot.ates Atomic hergy Comtsiesinn co b'achius. ton, D. C. 205h5 o - r, -

Sub,1ect : Three !ille Inland Unit tin. 1 g . , k -3 Sodiun Thiouulfste Tnnh Bsilure Dochet No. 50-289 Dehl'"Dr. Knuth:

On April 9.19'(h, nn incident occurred which may be considered reportable undert0 CFR 50 55(e). The incident concernn the buckling inwxrd of the ur.pt'r nect. ion or the Godium thiusuirkte tank (iiG-71), due to nunstive

.i nte rna} pressures, while filling the' tank. .

Donnri pt. ion' or,, incident The rnauter plant was in a cold shutdown (partin11y drained wit.h head ren:oved) condition. Final preparations were undervny for loadina ru.1.

The nnaium t.hiosultate tank in n ut.tinless steel tann, 'd feet hign and 6h inchen in diineter, with a 15,000 pilon enpacity. Tt in ritt.ua with a enmhination vacuum terentter/relier valve and a valved vent line. The tank is equipped with heat tracing.

On April 1 and P,197h, preparations vern made t.o fill the tanX vith approxiz<ttely 13,000 gallons of n 304 by weight solution of sodium thinnulratn.

h'eperati nnn included execution or valve lineups, opening thu vent. valve un top of the tank, and loosening n r)ango which covered the end of t,he ves ' . The f.Lange van not, cor.pletely removed. A procedure for the evolution had not been prepared.

The t.ank vss to be filled by pumping, from n mixing t.ank t, ruck via a tempornry connection intn the fill line of the sodium t.hionuirate tank.

On April 3.197h, the tank wau partially fill.ed wit.h 8,000 gallonn of sodium thionulfate without incident, t,he ope) stion heing diucunlinued 3rt.er exhauntinn the nupply of chemica19. on hnnd.

t 7910250 [N f

s

$ h e . C 6 RiiGi m m mg ws.u. ~ -- e--sunmum:r.:::aamn:meaut:santmmmamernm ersamir 021 QF D F Dr. Knutn ,

.s -

, k{ c page 2 April 11, 197h On April 0.1(T(h ndditinnn) cherniculu were received, the required nolution wse saixed in the nuppiinr'u mixinn tank truck, and then nrtt overnight in t.he unheated truek tankn. During the night of April $ 1 Q7h , ai r terr.perature:.

3 dropped to below hD"F.

On April 0, 197h, the valve lineup vnn rechecked and thr tank vent.

valve wmu checked open. At 8 h5 A.14. the tank truck pump was- start.ed and discharge prennurn entablinhed al approx 11ait.ely 8') psi. The sodiu::1 thiosulfete tank fill val ve (E1:V-51) was then opened, Inndiately after the valve was ersched open, the upper hc5 portion of thn noclium thionnifatn tank huckled inward. The valve (33Y-51) vos it:medintely nhut and 1.he pump secured.

F, valuation of Tncident.

1Enediktely subsequent to the incident, the Vent Valve flang:t. which had previounly heten lonnennd, van found to have been nucked tightly aghinst the gnnhet nurfnce.

The tempernture of the sodium thiosulfate uo3ution heing pumped frc:a Lhe Lank Lruck wan found to he npproximntrly h0*F. The solution in the nodtum thioculfhte tank was approximstely 110*F, the clevnted temperature being chte to thr heat tracinc on the tenk. The vapor upace I.empnrature nny have been vell nhnvc 110*F due to tne heat trace on the upper portion of the tank .

The combinaLicu vacuum breaker / relief vnlve vrts removed and found to be in proper nperitting condition.

Based on 1,he above. it han heen concluded thnt the cause of tank failure van the rapid prenmurn reduction ineide the codium thinnuJ rnte tsuk when%hh relatively cold (LO*F) nniution was pwnped intu the relatively' hot (110"Fr ) vapor npace rit the top of the partially filled nodium thiosulfate tank. The fill line connnetion in ??-1/2" below the top of t,he tank and, therefore, directly into the vapor sy1ce.

Calculabiurm show that the prensure change in the Lunk renulting frem the recuetion of the water vapor preneure and from the gas volume change enuld have been hotveen ?.82 psic and 6.26 psig, depending on n*:numed tenperaturn extremes in the vspur space.

The open vent volve was ineffective in equalizing pressure becanne the blank fisage was not, properly removed.

f t. han heen furthnr concluded that this condit.ien could not occur during system operation in conjuncLlun vit.h an F.CCn trnnsient, the vacuum becaher being sufficiently sized to equs31ne preunure during functinnni drsi::down. In support of the above, a corf of T/P 20h/3, Reactor nu11 ding upray .% . tem, functional Tent, previnualy completed satisfnetorily.18 altst:hud.

- - __g na + _..a- _

.g

." QO D '

Dr. Knuth Par.e 3 D

April ll. 3 97h O C.)

ng- -

Cor13e_Qve Action - y ]

With regnrri to repsir of the damaged part. ion of the tanh, ysrinun s3 ternatives have been evniunted. AL f.hin time, it appents that. a plpn rur . .

ree,Lahlinhing the neumetry of the preuent.ly buchied f.nnk plate can be D'ursued by n combination of internal water pressurization nnd rewurk of the plate.

Procedurten for this are hning developed and if thl:. pinn continunn to prove nucceLeru), it appears reanible to hsve this work completed, includiun retentinn and refilling with nodium thiunulfate prior to Mny 8.197h.

In pnrallel with 'the above, an errorL in being purmted with the tank g rsbrientor which vili provide for prersbricatir,n. off uite, or a now tank top nt'etion, i including the present top hend, such Lhnt the damaged portion vould hn removed and the new portion nimply inntallr>d as a unit. Intennive work errurtn k have been authorized which es prnncnt vou.1d permit delivery or the prefahricated )

tank Lop section by April PP, 197h. Danen on thin niternste, it appears reanible to have Lhn tonK vperational on Mny B 397t, thue nupporting the prennnt initial cr1Licolity date.

to r,aintainIt or in better Met-Ed's theevintention drites.in pursue the above plane smaking every errori.

Furtheri.'nvreeLiveActiorj Ln Prevent Recurrence-Plann are being prepared to prec2ntre future incidwnts or this naturr.

These plann are concidevini; the following correctiv meanures:

1. 1*romulgntion, cold liquid intotoan11 pernennel.

vnrmer tank. of.the dangers senociated with pumping a

2. FrumulantiQn, to 613 pntconne], of the necesciLy for Lhnrough reviev of evolutions prior to perfor?r.encte, including requ.irements for Vritten proceduren.

3. RevicV breakerkudsi:',eevnlu8 tion t.anha.

un closed of the ni", inn criterik uned in teutnblishing Yacuum

h. Review And evaluation of the heat trhCe temperature gettings for vnriouu hrr$t treced innhs. Including require:::ent.n for rilling culntion temperaturn reintionuhipn with tank teroperntures. <

Very truly youre. '

( .

\. J'.f,) y; x t,

... h ' '

! . C. Miller T

l Vice President I hc CC: &*. J. P. O'Reilly (with enclosure )

l l

l l

l