ML20046B362

Session 3.7;Review of SG Tube Rupture Procedure at Borssele Nuclear Power Station.

ML20046B362
Person / Time
Site:	05200002
Issue date:	07/15/1993
From:	Vayssier G NETHERLANDS, GOVT. OF
To:
Shared Package
ML20043H221	List: ML20043H220 ML20046B362
References
NUDOCS 9308040115
Download: ML20046B362 (15)

Text

r j s '

. . EuctosuRE l 7  !

P3 \ oF l(o t

i 4 n6 33 15:u3 FAI -31 n 3334015 DGt ETD 2co' ,

./ /..

~

/

- 3.7 i SESSICN 3.7 i A TEVIEW CF THE STENi GENERATOR TUEE RUPTURE DRDC CURE AT THE  ;

BORS5 ELE NUCLEAR POWER STATICN.

i O. '!ayssier, Nucear Oeoarteent, Ministry of Social Affairs t.. Winters, Netherlands Energy Researen Fcuncation. ,

1. Introcuetion Borssele (screnvm: KCB, Kernc.entrale .E.cr sele) is a ABC PWe 2-loce PWR, ouilt ty Kraftverk Union anc in comercial operation since 1973.

Its original pretectien against ster.m qenerator tube ru ture (5;TR) relied en a cuick depressurization of the reacter coolsnt system lROS), initiated by a N-16 signal in the secondary system snd executtd by the pressurizer spray systems. Safety injection (SI) was not excected to occur, since the pressuri:er level vould not have caen 1:st e:mletely and such SI c'eoended on the ccincident signals Icw oressuri:er :ressure and low pressur::er level.

After the Three File Island fTMI) ac:icent it was cecided not t:

initiste SI any lenger on ecine: cent pres p ri: r signals, but to start II :n Icu crecsurarer pressure sicne.

s a consecuence, EI uss expected to ocrue curing a ECTP. event anc methces were investigatec to rancle the safety :njection phase cf the SOTR-event, inclucing its tirely termination : prevent an eventus1 50-overfill. f 9308040115 930715' PDR ADOCK 05200002 O @{M

h bj d ce } h DGA.Krp 2ago O L6 J3 15:u5 TAI -31 70 3334015 i

- 3.7 Cxceraence with the Doel SGTR lSel;ium) and, more recently,tre Cinna event ' 1;S A ) , raiseo the questien vrether the KCS-trethocs could be .[

a:ecua:3. In accitien, concerns cf the utility arcse whether incor-tant plant periods such as start-uo cr shutdown were protected satisfseterily by an ECt! vhose initiation signal is t ypassed ocring tne 'ine cf low pressure that is inherent to these satustiens.

t was ceciced to revsew the ECTR chenc,mena and procedures fer KCS. ,

I. **e cre.TMI situation f ericinal :esien) -

Both ain steam lines are centinuously monitored for the presence of NIE. 's Soon as this signal appears an imediate reactor scram will folleu. This is ateut 15 sec. after the rupture of the tube, due ~

ts the distance tetveen the break location and the N16-sensor.

About 2C sec. later the turcine is tripped. This turbine trio celev -r extracts a large amount of heat frem the primary system, which l' initiates the depressurizatien r. the RCS. In additien, it keeps the secercary pressure at e Occi'  :/ 31 stance from the secondary safety calve set points.

The N-16 signal also tri. L. yrizer heaters and directs the enemical anc volume cents. ystem (CVCS) injection to the pressurizer

)

s;;sys.

n ace:tlon, the secene of three CVC5-pum:s (one is nornally in. -

Ocerat:en) is started and the letcoun station is c1:se . Hence tre I

spray ueter temperature crecs ractdly to about 50 C. This increases the de:ressurization rste of the RCS. .

As soon as the primary pressure has fallen belev the seconcary l Lefety valve set point,wnich t&kes about 5 rinutes (see aise fi;.3;,

'ne C'.CE-scray action is terminatec and switched t.o norrral in,1ection.

~ a: : n ce te crevert teiling in tre F.CS.

I icv tre faulty SC is isolated and the oower station is brought te

lc Inut:cvn using the antact EC.

~re cert:recus coeratten cf tne T.'CS to; ether witn t .e relatisely lar;e

resstr::e prevent the lesel fr:n reachin; tre low level retpo2n:.

'See slso fig. 4.}

e

$. P3 5cr16- ,

t

'e

.29.u6 33 15:04 TAI -31 70 3334015 DGA KFD Z (.

i l

-3.7. .

f f

i Therefore, safety injection (SI)- will not occur and the incicent is r

treatec alnest c moletely by normal operating systems.

If a power failure would cceur at the moment of turbine trip, the fellowing events would take place:

- Icss of main feed and, hence, initiation of auxiliary feect 3

- loss of reactor c clant pumps (RCP's); ,

- Icss CVCS for about 30 see, since this is the time necessary for -

the CVCS to be powered frem the ciesels.

The incident would be treate: as bercre. Due to the large 4 1 over the ecre, inherent to natural circulation, the CVCS back switch to '

the injection mode vould now fall below the boiling peint in the ho: '

leg. Veicing would occur, but was not expected to aggrevate the sitcatien.

As befere, SI would not occur. Fest rejection is over the SG atnes:herie dumovalves. If a SGTR would occur d';ing startup or shuwd:wn, then tne above mentioned actions are to be initiated manually due to the aosence '

of the N-16 signal.

3. T$e TM!-consecuences and the lessons fr0n the Doel eve *t r oIIowing the TMI-accident it was decided not to use the coincicence of Icw pressuri:er pressure-lcu pressurizer level for SI-initiation any more. The reaeter pr:tection system (P,PS) was changed so as to ce enc only en primary pressure for safety injection. since it was beliesec that the prescurirer level woul not ce relisele curing certaan small leaks in the cressuritrr. Emohesis was put :n subeeiling as a re:essary eencition to sustain retural circulation and to cbtain useful infer at;cn t

frem the pressur17a* 1evel.

Consecuently, the SGTE-scenaric nao to te changed censicerably.

The rev3ew of the SCTR-scenarie snd crece:Gres war acceleratec fell:uin; the cet=ction cf 50 tube cerresien curing *he refuelling period 1*. tre fell :( 1979 anc influencec cy the SOT -event in Coel, Belgium,' 2n "une 1979 (Westingnouse-unit).

fer those leakages that, at nominal pressure or at a nearty pressure level,

cule .. te corpensated cy the CVCS, the pressure vould trigger tte H195 Fressure Fressure Esfety Injectien (hPSI) at 135 bar (pre:eration at 135 bar, at .sl start at 120 bar).

s

. t N 0F lh i

9 u6 ;3 m n9 FAI -31 70 33a4016 DGA TJD 2 0L

- 3.7 ine !:;nals tnat activate the hPSI vould also isolate the cont.inment and trip :ne RCP's, as well as the CVCS, se that all pressurizer sprays vill be lost (see fig.5). The RCP's end the CVCS can be restarted tanually and, Pence, scray espacity vill be regainec.

The HPSI-punca are centrifugal purcs. If the hPSI and the leak flow (fu)1 SOTR' are plotted against pressure, an equilibrium vill te reached at point S, see fig. 1, that is break flow matches injection flow.

If the cressurizer has gone empty, the HPSI will eventually not refill it  ;

since that vould recuire a cepressuritation of the pressurizer in creer to have injection fics exceecing the break flow. The existence of an adequate

~

pressurizer level together with sufficient subcooling is necessary to judge unether the core is sufficiently covereo, which is a necessary condition to shut =f f the nPSI-pumps. This is othervise required to avoid a SC overfill .

as the Dcel transient has learned us.

~

t To regain pressurizer level the pressurizer PCRV is actuated in ciscrete steps. In fig. I we are then at lef t from point S and the injecti:n f1:v exceres the break flow.

Afterhavingregainedpressurizerlevelsneafterhabingettsblishedproper suo00 cling the SI-pums are shut off. l The RCS-pressure, that had stabilired around point S, fig. 1, vill nov drop.

Simultaneously, the pressurizer level will cecrease'. It may be needec to restart the SI-pumps intermittently to restore the pressurizer level.

In the esse of a relatively small SCTR the CVCS may' control the Icska;e, so that a more or less conventional depressurization and cool covn is possible. .

As soon as the RCS-pressure has gone mlov ee secondary safety sal (e set peint, the faulty 50 ein be isolated. Since the SI-oceration has pic3cnged a

the decressurization, a considerable amount of water will have leaked to '

I the faulty SC and the level will have risen aCCordingly.

"m: ::s.;t. in a layer cf relatnely hot vater above me SC : woes, Lee fig. 2. Subsecuent eccidevn cf the DCS will not affect this i layer Oi .ot ester 02nce the steam come :n too of it vill act, th;cuan

+ne la a.c . as a pressurizer en the entire RCS, thus preventing furtner

. j l

l l

l 1

l a The pressur;:er is empty (i.e. full of steam) at a crimary pressure of i l

120 bar, unerest the HPSI shut off head is 110 bar.

I

.f

") h DF h,

~

up 06 93 15:o3 fAI -31 70 3334016 DGO KfD @c

-3.7  !

ceoressuritation, as tre Doel event has shown. The procedures, therefore, have oeen extended tc incluce the cepressurization of the faulty SG. *)

A typical feature of KCB 'is that the reactor protection cystem (RPS) keeps track of the pressure ciffere ce between both SC's. As soon as the difference exceeds 15 bar, then the SG with the lower pressure is isolatec.

This is done to offer orotection ouring a steam or fetcwater line break.

In this cose, however, tre prolonged cool cown of the intact 50, together with the faulty SC remaining on relatively higS pressure would lead the RPS to isolate the intact SG, being the one with the lover pressere.

This would mean a loss of all heat. sink capability. This RPS ecmmand

. must, therefere, timely be bypassed oy the operator. ,

It should be noteo that the N16-actions (activation of oressurizer sprays etc.', partly have been removed. A.o. reactor scram anc turbine trip haven been preservec.

'. Recent review inclusion of the Cinna event The utility was asked to review its SGTR-ptcceoures in the lignt of the Cinne event of January 1982.

05arasteristic enenomens in Cinna were the stuck-ooen failure of the PC V that initially was used to deoressurire the RCS, the steam bucole in the reactor pressure vessel (RPV) head and the everfill of the SG secondary 31ce. Hao the secondary safety valve stuck open, then a serious tituation might have developcc, since en out-of-containment LOCA would have occurred.

n ecdition, the Borssele operators did not feel comfortable in not

, having the autoestic protection of the HPSI available curing the startup er the shutdown period of the reactor, as the low pressure initiation signal then hac to be typersed.

Tre original RPS-situation, i.e. using the coincident pressuriter signals, coulo offer the next accantages:

s. tne fast primary system cecressurization, carriec out bv tne i sorays whien are traggered by the N16-signal, orings about a minimum  !

of primary-to-secondary leakage

]

) *.ote: Stratification is also expectec to ceeur in the intact SC but

, coes not harm since the steam can be extracted. )

l

b CE kb l

i DGA EFD 2 01*

- "9 H6 03 15:10 fAI -31 r3 3334015

- 3.7 6-

c. in ec:itien, the fast cepressuri:ation will result in an ecualisatien of pressure over the leak which lies certainly below ,

the seconcary safety valve setpoint;

c. no safety system actuation will occur which means: ,

- the RCP's are not tripped; fereed circulation will remain available, vnich vill clininate the risk of stesnbumble format:en, unless a loss of offsite power occurs simultanously with the  !

SGTR; this is highly unlikely, since KCB hr a reliable connection with the national grid which, in turn, nas a direct. connection with the Curecean IED kV grid system thst intercenneers .Tany power stations an cifferent ::untries; additienal arguments for the imoortance cf fo ced circulation are presented in sec. 5;

- no automatic containment isolation will occur, so that mere systems are available for accident mitigation and reccVery;

- no automatic isolation of the CVCS: this system is able to ecpe with the ACS ass less and provides the P.CS with boron througnout the transient; I the etsence of HPSI eliminates the risk of SC overfill (together with t'a fast decressurization of the PCS)-and preserves the va'ter erntents of the ?PS! stcrace tanks, unich nay be of use for any unexpected coroli:ation.

t vss realireo that the reesteolishment of the coincidence of pressurizer pressure and pressurizer level as HPSI starting condition would bring disscv&ntages for small breaks in the s',eam Comp of the pressurizer

'or a stuck open PORV, such as TMI).

It was intenceo t: manually start the HPSI-pumos in suen a case. The Berssele licensee arguec that it is more easy to start the HP51 manually curing omall LOCA's in the pressurizer steam space than to steg the MPEI manually during such a complicated transient as the SGTR - such -

a step is necessary to preven t SC overfill sna to initiate RCS cecressurization, as alresoy has been discussed.

t shcule be res312eo that there is a bscu-up of the nanval HPSI-start

uring small " steam-LCCA": another automatic HPSI-starting command is given by the coineicent3 low RCS pressure - high containment pressure (3C0 mm sater ::lsmn , unten is significantly lover than many other PWR's).

The latter signal will be delsyed aceut 7 minutes if a PCRV is stuck ,

ccens :his tine is reovired to blow the pressure relief tank rupture disc.

Frcm regulatcry viewpoint it was recuitec that the Borssele licensee i

V0u'. demonstrate that for all water leakaQes f cm the RCS. saturation scu e et occur ze'ere the pressuri:er level sculd hase cecreasec rel:9

p kh l CF k(c k

iG t.6 03 23:03 TAI -31 70 333401$ DGA EfD 2 00' 5

- 3.7 the Jos level eet p= int, otherwise no autematic protection woule be provicec for many smell LOCA's.

It was f:unc that curing a veter leakage fl.e. not from the pressurizer steam dome, but from any other locatien an the RC3) the low level set point would be reachec at a RCS-pressure of 120 bar, so that ample suecooling exists. If later in the transtent subcooling is Icet, so that the pressurizer level will rise again following the formation of a steam bubble in the RPV-head, HPSI has alreacy initiated oceration. ,

Technically, HPSI can then te reset; however, specific criteria have t been cevelcpec that must be satisfied before HPSI-pumps may be securec.

Among these are an accouate subecoling, together with a suffic2ently high pressurizer level. '

Larger breaks in the SC than a single SCTP, are, therfere, protected by an early HPSI-start. Mcwever, no ceteilec protecures have yet been set up, ner have been required, for such events. It is recognized that tne CVCS spray act2on is not safety grade. Therefore, it may operate outside its agreed and sopreved ccerating range and, thus, it may fail during its operation.

r If it fails after low pressure (e.g. 90 bar) in the RCS elrea'cy has been reacned, then saturst:en may occur before the pressurizer level has sunk belcw the Icw level set point. Operating personnel has teen instructed to manually start HPSI if, for some reason,subcooling :s lost.'*) As already explained the hPSI-pumps may not stopped before sub-cooling and oressurize* 1evel have been restcred.

Generally, cetailed ;rceecures have been estselishec for those situations where the CVC$ does not function proper 3y anc the traFSient has to be handlec by the HPSI.

These procedures are almost icentical to the enes described in sec. 3. i As has been explained in sec. 2 subecoling will be lost af ter a SCTR

, bat occurs simultaneously with a less of cffsite power (LOOP).

In contrast to the old situation, the coerating personnel will now l manually start the HPSI-punos. The reason is that a rising pressLrarer level will ccen up again the CVCS letdown, thereby increasing the mass loss fron the RCS.

• although orcecures are carectec sc as to aveie SC overfill, suen en everfill is a realistic pesssbility curing FPSI coeration and, therefore, the mechanical consequences of SC overfill as uell as ways te crain the full SC have teen investiqated. These natters are still uncer cateustico.

It sopears tnat steam line failure by the veinnt cf the water in tnere, coes

• ct have te de exceeted. nuwever, anw ef fectr, Of .ater hammer are urxno 1

-f ter the THI-evenc . suocociing cisplay was installed in the contreircom.

') CF k h pG4 KTD 2 013 ,

O nc .3 16:10 FAI -31 70 3334015

-3.7 B-

• n tnis . *-cect it must also mentienec that the N?21e of the pressurirer sex;hary .ptsy is ecuipped with a therral sleeve and, thereferee may accept scrvpt tew erature enanges, such as vould occur in the beginning of tne transient af ter the falling pressuriter level has closed the C'JCS-letcoen. In sce2tien, the pressurizer has en internal

• skirt in the ,

r steam scace to prctect the valls against cold spray water.

In aceitional change has been mace in the FPS not strictly related ta the EGTR. fctrerly the HPSI-pumos could only be stopped after the primary pressure had fallen to 7 bar in a LDCA event. In practice this means that i curing any small LOCA the HPSI pumps would run until the storage tanks scult almost he ceolete:. This trears that the cumps are challenged fer late cf sucticn head (*.PSH). Althougn this is within the c'esign beeis cf .

the punos, it increases the failure precability of the puxs, as en inci- ,

cent with a KCS HPSI-purp has shown.

In aedition, it may be ;rucent to preserve some of the HPSI tank contents f

e.g.ifastretch-outofhPSImaynecessaryduringaLCCAthatinhc1hesa .

veter less outside containment, suen as vould occur if a SGTR is followed byartuck-openseconcarysafetyvalbe.

XC3 ras nov changed the RPS So as to be able to shut off HPSI-pumps at i ery pressure below 120 bar.

ri;s. 3 and 4 present a few grapns of how RCS pressure and pressuriter t

letel benave curing a SGTR with effsite pover available.

. toss of nstural circulatien The Eerssele licensee has, curing the latest refuelling peried, inbestigated the syetem's capability cf retural circulatien with one SC cperative and the cther SG isolated. This reflects the situation after a 537R, fcllowing -

th: :C5 ceprecsurization to about E0 tar (tcgether with LOOP).

Natural circulation ever the intact genereter appeared to ce succesful, as excectee. u ovever, it van found that the circulatien over the isolated SG -

could conc to a cerrclete step, in contrast to the general belief that so*.e

culation continues to exist in the isolated SG. Hence, staenent primary ,

sater is found in the isolated loop. A direct consequence is that this

• amcunt of water Vill act be bcrated curing the remainder cf the transient.

'.:culd cower te regainec, and the cperater would restart a PCP, then a reactn ity transient 1;nt fo11cv. It is expectee that this situation is ret unicue to a tve-1:co plant. However, for three and more leops the situa-tian vould be somewhat ecre fsvourable, due to the tra11er water content in the SG U-tubes.

Thefindingsnahebeendrs;nupinareportthatsoonvillbepublishecI

.in the a TFT riport is titled: batural CirculationFover screrele Nuclear Plants author C.. Valserecer ,

-) o i- b us. n6 93 13:06 FAI -31 70 3334016 DGA KTV

- 3.7 An 1 reciste censec;uence is that cohling down with one SC isolated and using natural circulation, as would be desirable after a SCTR together '

with t00P, is not, permitted. The coersters have to restore the offsite cover and restart at least one RCP, before they are allowed to cool down the RCS from the 60 bar-260" level that is reached after isolating the cefective SG.

6 Cenelusions 9 The use of ecIncident pressurizer signals (low pressure-low level) toi with the N-16' actions) will prevent, in many cases, HPSI-operatien during a SGTR event. Hence the CVCS will handle the transient, while the RCP's keep running. There is no danger for SC overfill with all its complications and steam voiding in the RCS is prevented. It is believed that tnis apprcach to the SGTR also makes the accident ecre easy to be handled by the o personnel, Equalization of the pressura difference ever the lesk is ouick snd an early isciation of the defective SG is possible, which minimizes radiological release.

i However, certain leakages frem the pressurizer steam dome are now to e b handled manually, since the pressurizer level is not reliable and may crevent automatic HPSI-start in these situations. An automatic back-up, provided by the containrent pressure set point, together with low PCS creesure, is available.

By-cassing the HPSI-protecticn during sci'eduled low pressure periods, su as start-co/ shutdown, is not longer necessary.

Final depr355urirsticn and cool down under natural circulation conditions

~

with one loop isolated is not allowed, since it may lead to resetivity transients anc/cr voiding in the U-tubes of the isciated SG.

The HPSI-purps can now be stopped, under proper process conditiens, at i FCS cressure below 120 bar. This prevents NPSH-caeage and creserves water L for other purpcses, such as containment spray. In addition, it would help cu:ing evt.of-containment LOCA's, such as would occur during a SGTR plus a stuck-ocen seconcary safety valve.

i 2VRS

[ .j

- h g

O n6 93 15:11 frI .31 in 3334alg pg4 gpD 201J

- 3.7 <

I t f injection flew l injection flow exceecs leeksge fis less than flow leakage flow flev g q g

leak flew l HPSI flow 4

1 I

IID 160 pressure (bar)

Fig. 1. Leak flev versus HPSI flow during a SGTR event (IOC: rupture' ,

M f

i .

t Ba \\ ee %

DGA ETD Zong 9 u6 93 Ig:ui FAI -31 70 333J015 t

- 3,7 - II -

h r  :

e- y .

j i

steam line steam space j

!e\ surface o a ~

het water .

".f-i A feec water line cold Water ,

i .

i Stratification in an almest completely fillec SG. l rig. 2.

5

I

. I DE

- .i LGA KTV 2015 ti9. 6 93 15:12 TAI -31 70 3334016 l

l 4

- 3.7 l t

150 -

I i

- RCS-pressure' c - 120 5O

@j - 100 SO -

i SG-pressure i

60 turbire trip i

I -

35 300 time (see) t rig. 3. #CS-pressure and SG-preshure after a 53TR, Offsite power remains available (isolation of cefectise SG not included in the calculations, will occur manually after the RCS pressure has fallen below 87 bar. t5st is after about 5 minutes). ~

Ecth pre-THI and present situation.

t e

b O or } h '5 k

r DGA KI'D E888 41 u6 :.3 n:ct TAI -31 o 3334016  ;

- !.7 13 -

i 40 e -( 100%)

press. water uc1t.me " 32

.m3) 26- (6C:) r 16o ,

. En (20%)

I 100 time (sec) rig. 4 Water velurr.e in the pressurizer sfter SCTR; cffsite power +

and CVCS renain available - both pre-TMI and present sitt;stion.

b 1

l l

-1 l

l l

l l

4 _

J e'1 O

,C , # . a t h .. Uo +.O 4

9  :

&K g 0 ?. -

' y ,. i % . _

e e

l l t

t r 0 a g

< e S' 8 3 rl e

- - o C"

VI" r dn bm o a a *.

C,G e .

r ul .

ssve 2

0 V ee rp l, i .

1 _

4 #

n P7

. gI

t. l'^ (

sV

• _

(l HC _.

3 g 1_

( _

y <

t _

b r d-I - ta .

• c Is i _

pc Pb 0 Ss t a

cc l PP 41 *'"

w e

~ Riw l i

[ S P' 5 s

\ t m

u f

ny _

7 S w1 1 n

o n i u< t c

nl.e _ ~

n e ' _

e o t r v o

'r i

e t _

F f. e d

j 7o l

l _

r S'tv.c t o

c a a r

e H

5 g g 3

3 i

I

< ' t m n n

.. e _

p no r i _

p Il V 'o nt i

tl no a

os ci

. l h I$ Os \ (f. l J

i DGA KTD 2 017

'9 "6 33 15:13 FLI -31 *0 333J015 '

f o [ 2- tw./.* [/

~m 5. w

//ainh,6,u& Ar a .

.. ..~ ..

t e ._ _ _ _ _ .___ _ _ I I

I c'...

.~.::n.~;

n.n 1 .

I I l -

i J 9.*~ g

. , . . ~

2... .

.,,. -j - . . .

.' M * " g n

T' ~*2 .

' 3-a.~.;.,

1a ai$a i

n . . ., s .u, ~... . s I l

t i

...... g. l g i i~. an- ,.,.4..

. . s.

i g. , '

i i I

_._ _ _ __1 - __ _ __ +j _ __ ._ _ _ _ __ 1 J  ;

-- . . _2

• a., >

(IPfl

!- . . , i 1I . '*

i j

Il.:s r r'-

i

- - - - -- --I ~i ' ~ _] .

asi .

. I

!t

. ~ . .r .'

.c.....

4........ , ,

. . . . . . . ~.

?

I f

^

l i r I

. s l

t .

l f .

u

.~ .

4

.. s.M.., ,

f

'.s q *:* '.*:1 :

• i"." .'/t :

gg i .

, . I .

i t

..A,.....

.~ .-

. l i

4.,

g $p' t. as 1e =c.

y

. .a., . ... e4 .

a