ML17310A350

Forwards Preliminary Rev a of Data from Pvngs,Unit 2 SG Rupture Investigation

ML17310A350
Person / Time
Site:	Palo Verde
Issue date:	06/16/1993
From:	Bradish T ARIZONA PUBLIC SERVICE CO. (FORMERLY ARIZONA NUCLEAR
To:	NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM)
References
102-02537-TRB-J, 102-2537-TRB-J, NUDOCS 9306230094
Download: ML17310A350 (97)

Text

ACCELERATEn DOCUIVIENTDISTRIBUTIONSYSTEM REGULATO INFORMATION DISTRIBUTION STEM (RIDS)

ACCESSION NBR:9306230094 DOC.DATE: 93/06/16 NOTARIZED:

NO DOCKET FACIL:STN-50-529 Palo Verde Nuclear Station, Unit 2, Arizona Publi 05000529 AUTH.NAME AUTHOR AFFILIATION BRADISH,T.R.

Arizona Public Service Co. (formerly Arizona Nuclear Power RECIP.NAME RECIPIENT AFFILIATION Document Control Branch (Document Control Desk)

SUBJECT:

Forwards preliminary Rev A of data from PVNGS,Unit 2

SG rupture investigation.

DISTRIBUTION CODE:

A001D COPIES RECEIVED:LTR ENCL SIZE:

TITLE: OR Submittal: General Distribution NOTES:Standardized plant.

05000529 RECIPIENT ID CODE/NAME PDV LA TRAMMELL,C INTERNAL: ACRS NRR/DRCH/HICB NRR/DSSA/SPLB NUDOCS-ABSTRACT OGC/HDS1 EXTERNAL: NRC PDR COPIES LTTR ENCL 1

1 2

2 6

6 1

1 1

1 1

1 1

0 1

1 RECIPIENT ID CODE/NAME PDV PD TRAN,L NRR/DE/EELB NRR/DRPW/OTSB NRR/DSSA/SRXB OC/LFDCB NSIC COPIES LTTR ENCL 1

1 2

2 1

1 1

1 1

1 1

0 1

1 1

1 D

D R

D NOTE TO ALL"RIDS" RECIPIENTS:

PLEASE HELP US TO REDUCE WASTBI CONTACI'HBDOCUMENT CONTROL DESK, ROOM Pl-37 (BXT. 504-2065) TO ELIMINATEYOUR NAMEFROM DISTRIBUTION LISTS FOR DOCUMENTS YOU DON'T NEEDI TOTAL NUMBER OF COPIES REQUIRED:

LTTR 23 ENCL 21

S.

0 l

~~

f N

f il

Arizona Public Service Company P.O. BOX 53999

~

PHOENIX. ARIZONA85072-3999 102-02537-TRB/JRP June 16, 1993 U. S. Nuclear Regulatory Commission ATI'N: Document Control Desk Mail Station P1-37 Washington, DC 20555

Subject:

Palo Verde Nuclear Generating Station (PVNGS)

Unit 2 Docket No. STN 50-529 Preliminary Steam Generator Information File: 93-056-026 The purpose of this letter is to provide you data from the PVNGS Unit 2 Steam Generator tube rupture investigation.

Please note that this information is in preliminary form and has been marked as such.

A summary list is also enclosed for your information.

Should you have any questions, please contact J. R. Provasoli at (602) 393-5730.

Sincerely, TRB/JRP/ap Enclosure Thomas R. Bradish, Manag r Nuclear Regulatory Affairs cc:

B. H. Faulkenberry C. M. Trammell K. E. Perkins J. A. Sloan i800lz oI, 9306230094 9306i6 PDR ADQCK,05000529 P

PDR

SUMMARY

LIST JUNE 16, 1993 DRAFT A FAILURE MODE INVESTIGATION A.

STATEMENT OF FACTS 1.

EDDY CURRENT TESTING (ECT) 2.

CHEMISTRY 3.

OPERATIONAL REVIEW 4.

FABRICATION 5.

DESIGN 6.

INDUSTRY EVENTS REVIEW 7.

DEPOSITS 8.

TUBE EXAMINATION B.

FAILURE MODE REVIEW

4 0

REV A June 16, 1993 V.

FAILUREMODEINVESTIGATION The PVNGS Nuclear Engineering Department provided the Root Cause: ofFailure Team with an overview ofthe steam generator tube failure and related information such as, steam generator design and operations, industry experience, and receiitiliisteoiy, of the PVNGS steam generators. Through PVNGS's participati'origi thhe';BPdRI Steawmw",",Generator Reliability Project, NED, and Site Chemistry weresable to provide information on'g>~

SGOG's activities regarding steam generators+

p s a A thorough root cause offailure investigation;requires a det'ailed%ampilatlon,of the facts

.~:"

'."i~'."~;W*

. ~~<'M":

related to the component and/or syst'm favilure.'Thats'::rc'oempit ation or "Facts List"was developed initiallyto document'irit'ormation:tthat is:htowntabout the failure. The list was also used to guide decision':making'regparding issues that~require further investigation and troubleshooting. Additional infor'mation,or':":fa'cts obtained:during the troubleshooting and/

or analysis phases of the iiiiestigatiori"hagr'i'::.added~torthe facts list as they are identified/

verified,",Therefo'je, the list of'facteual information regarding the failure does not remain static,;'.during, the cougr's'e.of the evaliiation.,

%~

r(':::::rpThueifacts as'sembuieh'd for this iiiyestigation are categorized by subject in the foHowing Wgrip~r

REV A June 16, 1993 V.

FAIIURE MODE INVESTIGATION(CONT.)

A.

STATEMENTOF FACTS gP~$

1.

Eddy Current Testing

~>",5+jy'?c9,

~

Previous indications from the 1991 outage; W-'.':::-';h'"::j

""<:";:P...,,

- Axialcracking in SG 22 at 01H.

Q gl:g!:.:,,:,,)iQ

- Axial cracking in SG 22 at R117, L54 at 09Hf'+.":3:1" free span.'~':-',;~:,~i..

- Wear at batwings (BW), cold leg comer':eggcrates',"e'ggcrate suapports, vewrmtical c

straps.

- Some identified loose pafts'w'ear.

- Prior to this outage, SG',:'22,had'I'9fi,,tubesipluggh "ed k@

tpe

- Prior to this outageg SG 21'hba'd-,114 txubes plugged.

C)~0>e.:cQQ Q '@:>:.'j xc, 4@%rgx,

- More wear/tubes plugged at'BW'1';~upper eggcrate."P

- Decreased number and"detpth,of'.indicgga'tions wHre observed in 2R3 outage as A'ompeared to the 2R2,outage.

'M,::,;,>>

d'2009o ofSG"21 and 75cfct:::of SGP2 were fulllength bobbin tested during 1991 S

9 S

e e

9

'"':>>',!:100%gof the 01'H;flow distrib'ution plate was inspected using bobbin probe in SG'x22~raxnd SG 21

'7,,-,",;~100%%uo::Ofsefarhee'sPans are'i'above 09H in SG 22 was insPected using bobbin Probe

/educ to KI17,",L54iiiidication.

1 g!>!y

~sSG'"'22!h'adamor'eiwpear indications than SG 21.

4':;.:;.::.

~

Axial":cracking was found in the free span and the upper bundle supports.

"'<N>Peposits (characterized by using low frequency channels) were noted on free span e~k~.ggp eg +r"

'locations.

1

REVA June 16, 1993 V.

FAILUREMODE INVESTIGATION(CONT.)

A.

STATEMENTOF FACTS (cont.)

1.

Eddy Current Testing (cont.)

e~Nsg

~

= Two previous PLP indications found during the 199l":;outagae',have beexnrleclassified as axial indications based on 1993 ECT findings.

"ypP;

~

Allfree span indications and majority:':ofsupposit','.'atxial indicaetisons ww".ereifound inside an arc as depicted in Figure VA.l.ag g ~~:.'.5<

0+/,

~

Axialindications were noted'to'tbe excluusij'els j;,inthe hot,leg side,:.and predominantly on vetticiil!runs.

wi'::,;:

"""i:."':,':Q "tP' One axial indicatdis.O'divas locatebd:atutubse'R131, L46 at BW1+ 19- (i.e. horizontal

~

The imajprity oftreeispan axial indricataiouqs".w'ere associated with linear deposits

~'(fobund with,:low frequencccyjMRPCl,

~The bobbin coilcan not detect'the presence oflinear deposits using current

~>~j""~4'."'>.Ih>>>'.,b>> 3l).,indicationw's,-,identifjed by MRPC were not previously called during analysis of

-".:~~,

~

An in'crreastsed number ofwear indications was found at 08H, 09H, BW1 (i.e., upper

"'"<.,",';,regions ofthe hot leg).

1 t

I

June 16, 1993 V.

FAILUREMODE INVESTIGATION(CONT.)

A.

STATEMENTOF FACTS (cont.)

Feedwater Flow 2.

Chemistry Feedwater flowrate to SG 21 was 1.15% higher<than flowito,80',;2?. during tlie',"..'/

previous operational cycle, but is con'sidered to be n'egligible impact.'"~~:,'::,jW Blowdown Flow Rate

<,'g@

~

Unit 2 had longer periods',,'of. abnormal rat'e,blowdow'n.than.;Units 1 and 3.

~

As much as 80 -,90%,of the!bio'wdowii':flovywa's5eedwater.

~

SG 21 and'SG,22 had equivalent,no'rmal rate'blowdown flows, based on 8/92 data.

~

SG~22:;abnormal rate blowdow<n flowgwas 2x higher than SG 21, based on 8/92 x)@<

~

~

Hideout Return Stiidies PP"'!<i,'.~;~

4"'".X

'.'y~.'.

'"""<I'::.';~~~

~ "':<Jhe av'er'a'ge peak'sodiiii5:::hideout return concentration (1987-1993) was higher

'",";:!':;:,, 44po'rn,SG 22"than SQ~21 (270 vs. 172 ppb). The other units did not show that level

'o'tv<,.arfa'nce;:.betw'een their generators.

kc<jy 5:P Q<,. >,.',

'4p<;"+@<'he av'erage peak sodium hideout return from SG 22 was higher than the other Qj4)5,,SGs. (1987-1993)

I

(

l I

REV A June 16, 1993 V.

FAILUREMODE INVESTIGATION(CONT.)

A.

STATEMENTOF FACTS (cont.)

Hideout Return Studies 2.

Chemistry (cont.)

a g

s

~

The molar ratio calculated from the average'hideout retur'n:peak'":concentrations.of

4,';.'.".t'.".".:...'...,,

Na/Cl were higher in SG 22 than SG t2ly(19.2 vs;','2';9);,,(1987-l993t)

"""'"~j',::.'::~ep

'. gras, ut.".. t:,h

~

The most recent hideout return datan'(1:99,1.-1993) indicatesUnit,2 experienced the Y ~":~

highest return ofsodium;and sulfaate, add thge':lowest,return of chiloride when e

~

MULTEQpHs~were'essentially identical:in',all three, units. The pH range of8.6-

~

The c'oncentration of'lead returned'fr'om Uiiit2 in October 1991 was higher than

'- 'the other iiiiits.(1991-1993!'data)

"gM@,

>~

~

T1ie;concentratiaii:,.qf lead returned from both Unit 2 SG's was essentially identical g: '.(SG 21(was 1'8 granis!SGl22 was 20 grams).

..r:" '.;s "'),

g;

--";,.9~ed onpre~vtoasus sliutdown data, sulfate concentrations were not associated with ere'voice'scihemistiy due to non-prompt return. The most recent hideout return data fromaOn tats I and 3 indicated more prompt hideout return. The increased prompt

~pi:,.sulfate return occurred as the molar ratios had been reduced within the past six

months.

0

REV A June 16, 1993 V.

FAILUREMODE INVESTIGATION(CONT.)

A.

STATEMENTOF FACTS (cont.)

Source Term Study hideout return.

2.

Chemistry (cont.)

+pp~+

~

Source term study estimated 10,600 grams of:sulfate transported:;per year to the,'.P SGs from the condensate demineraliier>>s<(demin's);-'or;:from a coolmg"'w'ite~r:

inleakage of 0.001 gpm with the'!de'rnjgs.bypassed.

~<

),:;:P:~

"""."~.","."::P

~

Makeup water and chemical ii'ejection wer'e'in'significant'sourcesiof sulfate.

~

The Unit 2 demineralizersre'duced sulfate by'>15%.

~

Units 2 and,3'.;:dern in'efBuent s'sulfate sodiu'm.and~chloride quantities were

~

Units 2::and 3 demin efIjluent's sodium~and chloride quantities were approximately

,P.,':>.,the saine (1'2',000 grams,and;:;:2'1.,000,'grams).

Siilfates,d>>o not depress crevice'pH as significantly as chlorides do. Elevated sulfate

's::does riotinecessaoly nn>>ply acidic crevice conditions.

"-"".">p,.

=-'i-':::Ie, Hi'tleout>Return'Su>>lfa't'e

"~~.'.-":.,:.,>'

The"av'erage"concentrations ofsulfate observed (1987-1993) during hideout was 4<!,',:;~. within'10% for all three units.

' gp>>

~

',Generally, the highest concentrations of sulfate have not been seen in prompt III

REV A June 16, 1993 V.

FAILUREMODE INVESTIGATION(CONT.)

A.

STATEMENTOF FACTS (cont.)

2.

Chemistry (cont.)

Resin Intrusion

'cjFaa~,,

~

Afailure of condensate demineralizer service:,vu,essel (SV)j';.'.,Sgioccurred in July".-.'je'991 in Unit 2. Asimilar problem occurred during'February 1992bdue".tta;SVh "E"

"t

~

Sulfate concentrations ine';blowadown samples';did'noet increase followingthe 1991

, event, but sulfate levels did~increase,;in January and February 1992.

~

Due to the very high;sadsorptionaroate,for sulfate, increases in blowdown

'..vcs concentrationsawould not.naeacehsarily'be neoted:";"s The first downpower following these,'.".suspected evenats,::however, did'not indicate any increased sulfate hideout

+return, a

er u

e

~

"~Very small quanti)ies ofresin w'er'e observed during visual inspections of SG 21

"~;-pand SG;.'22'can decks~'.:,:perfofmc't ed during May 1993.

~g.:,::,

">>Q X::.~'~Ttests perfoeimed ontxesin beads obtained during the above-mentioned visual

'inspetctthonhdetei'mined that the resin type could not be determined ianion, cation or inert)'.:but,that the cation functional group (sulfonic) was removed.

u fact

<~j)

REV A June 16, 1993 V.

FAILUREMODE INVESTIGATION(CONT.)

A.

STATEMENTOF FACTS (cont.)

EPRI% acer Test 2.

Chemistry (cont.)

/4o

'4

'~pc:y

~

Based upon these tests, the majority ofimpurities introduc'e'd.:,into the SG remaiii'ed Q~Q'4" '~s"~

during operation. The followinghideoutgractions w'ere,identitte'd;..chloiide;,70%,

C~

X'odium 80%, potassium 89%, sulfate;and calcium approa'ch~>100%+"pP "C48,.""'::::j4

~

Downcomer sample,.impurity',conceritrations were appro'ximately a factor of 6 times higher,:,than hot'~leg:::sample coricentra'tio'ns due'to dilution of the blowdown sample. with feedw'ater. Priorto,::1993, ho't,leg samples were used to control SG

~;:."i:;::;:,Amore coric'entrated and potentially"'aggressive bulk water environment willbe pr'esent'at,,the top>of,the hotleg:side tube bundle.

~~i',:y

~ "~,Based."upon aii'.en'gineering assessment ofhistorical tube leak events, Unit 2 has exp'iiiencedmore condenser tube leaks than the other two units (approximately

>:;::;:::.,>;, two time's the site average of 1.7 reported leaks per year).

45

)

(

REV A June 16, 1993 V.

FAILUREMODE INVESTIGATION(CONT.)

A.

STATEMENTOF FACTS (cont.)

Condenser Leaks (cont.)

2.

Chemistry (cont.)

~

The number of condenser tube leaks was significantly re'duad:;jn',.;1988 after "$>~

lathing was installed to stabilize the tiibipg. Unit-:.2"continuedteo'.eaxeperiencge;.

approximately twice as many tub'e'4eaksas the site ayerageiocf 0.27'sr'ea'ports per year.

~

A condenser tube leak in Uiiit:.2dubrsiiig.,1990,was particularly severe at an

+tilt"'. o.

estimated 150,gpuwm!,'(This data'point not"used in the, above averages.)

~

With the exception, of a'conddeiisei tubae".,'rupture',:evsent which occurred during the Unit;i'!warranty ruri;the condenseriipbe le'ahers were contained by the condensate

. demineraliierrs JAB throne!unhsrhave ope tated in 'accordance with piant procedures, EpRI, and

+o<,"?,,,

7@

~

Mi'tbxee'uiiit5"have operated with essentially identical chemical control programs a4~:a~kgaay since st'artup (ammonia/hydrazine with fullflow condensate polishing).

I i,

REV A June 16, 1993 V.

FAILUREMODE INVESTIGATION(CONT.)

A.

STATEMENTOF FACTS (cont.)

2.

Chemistry (cont.)

g. aeu.

Secondary Chemistry (cont.)

4:.+ 5,

~

A series of enhancements were made to all three~units'o'jedsensate:demineralii~ej operations that resulted in reduced sodium throw@'4p.;.

~

Full bypass operations of the,conesdensate,,deminweralizes wtweurme',mitiataediin Units 1 and 3 during 1993. Unit,2~gequiied its deminpgerwaliz'mrs reinain in,"stewivice due to a small condenser tube leakiduring thee,:1hst cycle.

~

Feedwater pH~was'intcreased sligahtly,.duriiij~1.992'from > 8.8 to > 9.15 to reduce

~

Fee'dwiiler hydrazine'whas incieased"'fiom 20 ppb to > 100 ppb in i992 to reduce

+~~~?

4t

,&the SGs eletctrochemical pocttential."~i:,

"-"'::::.,i. "'"'k

>~ra,

~

Molar ratio chemistry control was initiated in 1992 with significtmt reductions in g'i'molar'tratiru otceumng':in':attttshree units.

4a

%u

"~':::.":;;,:~, "",.:The::generato'rsvp ithiiia unit operate with different chemistries despite having a

'cuo'mwmon,,feeudwater source. SG 41 in Units 1 and 2 operated with the higher impuri'ty"xemoval (higher CPI); whereas in Unit 3, SG 02 operated with the higher

%~:;pi'em oval.

'Q::."..".'p.,

4'47-Q

(

f I

I I

REV A June 16, 1993 V.

FAILUREMODE INVESTIGATION(CONT.)

A.

STATEMENTOF FACTS (cont.)

Secondary Chemistry (cont.)

2.

Chemistry (cont.)

~

Trends in chemistry impurity levels had been consistent withpIant:,operations in'alt three units. With the exception ofthe"Unit 1 condenser. tube le@":duiiiigthey warranty run, there have not been.'any~acute

'secondary ch'emtstry events.

~

The molar ratio trends frotn all':,'three units'indicated chro'nic alkaline chemistry

~

Iron transport.:data "es'timated that>greater" th'~$n,3 po'unds ofiron per day could have accumulated iii"ea'ch stehrn'je'nexator.'Yhat would 'equate (theoretically) to over

~p 5,000'~pounds to date;::Minimal,tube'sheet fouling had been observed by ECI'.

s::R 4

)f

REV A June 16, 1993 V.

FAILUREMODE INVESTIGATION(CONT.)

A.

STATEMENTOF FACTS (cont.)

3.

Operational Review Operationally, the most significant difference between Unit!its steam;generat'ops)and those

""ci!':g "<::,4!,

~~h, ofUnits 1 and 3 was the presence of a level oscillation in Unit'2;::!The,oscillation ocucesuarred

'S y first in steam generator 22, although itlater occurred mp:steam genes'r'ator"2l-;asw~wte ell. The 4%@

'~j:"jg~,,t~

'Q~<':ai~'Qj$ '~bebop apparent cause ofthe level oscillations ~vase a failure of the,.positioner for';steam g'enerator 21's economizer valve (2J-SON-FV,>>.'.1'112);Qtgl,":,"Aj'~,. ',.

%+i': '":t-ip" s

a a

o v p

y e

c g

The followingscenario describes',:tfiet,,effewcet':.of, the failure:

'-"'~".,',.;,:~<

1.

As steam geneiator'.2l'>levei deviated from'.the nom~ viue of 50%, the feedwater control:;system~;,devteloped aens'eerrox sigri~al.

2.,The 21"'econtt omizex valve':failed'"to resp'ond, resulting in the level continuing to e~::,',:',::.;.",;deviate~further',"<and increasingt.the,,tnagnitude of the error signal.

>3".~gThemaiii"feedwate'r'pump,tii'r'bines, which are also driv'en by the error signal, s

g cp would inceiewase or decrease speed to return steam generator 21 to the 50% level

+",.::.:... 4.

Becavugs'e':the main feedwater pumps deliver feed to both steam generators, the Y

"t'ai,';hchange in speed would also vary flow to steam generator 22, causing the observed fevel oscillation. In summary, the feedwater control system "drove" steam

., generator 22 level in its operation to maintain steam generator 21 level.

49 Qz

t 0

REV A June 16, 1993 V.

FAILUREMODE INVESTIGATION(CONT.)

A, STATEMENTOF FACTS (cont.)

ra 3.

Operational Review (cont.)

xaxz e

e The scenario described above onlygapplies wh'en'.'.the:main

%km','<4'5Qi:..".

feedwater PumP high-select, gate in theIfesi:dwater c'ontroi'rejiduu system receives a nominajly higher. signal'>from',the¹l

~<~<~~~

feedwater control.-synesterm cabinet.":::Based on dis'cussion',with the responsible;'.engineers.:the signal fromitrhe;¹1 cabinet was a

i.4i':"ea a,:.t, c kesa g

eÃ4.5 The teamw:,eview'ed 2J-SGI+LR'-:.."1l21, the hard-copy chart recorder output, from

,,., Jun'e:,';:23, 1992, to the,:ISGTR event'date. 2J-SGN-LR-1121 displays the two narrow 4h,.'.,:;::,::,4

'rga'tIge ste4:;gg'e'riqatoz 22~level,htgrwansmit ters (LT-1121 and LT-1122) that input into the

~i:.,:.=""'"'"'~:.';l~< '<i.'-';::::0eedpgater Conti'ol System. The June and July, 1992 recorder output was "nominal."

+~< Tlie",:::.trace,..ag'pyeared;:a's'an almost solid line approximately 1-1/2 to 2% wide. The g

e s

e e

"":>..:;:<(A frequeiicy',';:,of thcese nominal oscillations appeared to be about 40 per hour.

Qp4'P l

I I

I I

I f

REVA June 16, 1993 V.

FAILUREMODE INVESTIGATION(CONT.)

A.

STATEMENTOF FACTS (cont.)

Unit 2 - Steam Generator P2 (cont.)

%~~4+S 3.

Operational Review (cont.)

During blowdown, the magnitude rose to about 2'to,2-1/2% and:the.'frequency dropped

'C':S: Py, Y

+$

?.

to about 20 oscillations per hour. In addition,to operator.':mterviews~the team,reviewed chart recorder output just prior to the'.last.:':outage (2R3, commenced Octob'er 17, 1991) and prior to an unrelated reactor"trjppyhich:.occueedIon Nov'ember 23/1987. The level

. ofperformance was comparabli"..',:to ob'st'vations of all'th'xee, Palo"Verde units in the

~9+

past.

Based on operator interyiews>and'.chaxt recoxder ou'tput, slight level oscillations began to emerg'e'jn. early August'i~1992 (e,g.,'8/4/92)::This included peak-to-peak oscillations of!about'4%. These spikes we'ii',only detectable until August 6, 1992. On September;7, 1992'.the, duty Shift Technical Advisor initiated Condition Report/

"'Dispositi'on;Request,(CRDR) 2<<:,2-0282, identifying the appearance of 5% oscillations.

<!::.,AsiiiiilarCRDR~was writteiion September 12th (2-2-0287) to establish a continuing

'vj-.":"':~~ tr'eiid~A!third CRDR'j'2 2-0286, also dated September 12th, noted the same erratic

"<;.':.:':,'::,,.<,~"," operation>durin'g';high rate blowdown. The return ofthe oscillation was also visible on

~<.'..ithe chart recorder output. The oscillations reached about 5% peak-to-peak. The osillations lessened in magnitude at approximately 1lpm, October 2nd and stopped completely at 3:30pm on October 4, 1992.

REV A June 16, 1993 V.

FAILUREMODE INVESTIGATION(CONT.)

A.

STATEMENTOF FACTS (cont.)

3.

Operational Review (cont.)

Unit 2 - Steam Generator K(cont.)

.:<~>jg At3:00 am, October 8th, the oscillations started'ti.,xeappear,"-and!i'itherbabruptly,

'"ri+

returned to previous high levels at 3:00 pmhtha't same "day,U.'nit:.:Miimtenanyc'e:,xeplaced the actuator on the "Bts heater drain>tank:,level control v'alve (2J.'-'.,EDN-'I V,502, work

~~aPP~-'-:i@~

order 00373300) on November'0!tli;:to correct'h'art,:tvgalve's cycrcting, biitisteam generator 22's level continued to oscillahte,:::,On N!oyember 12th, cieaneout pihs on the control valve Positioner nozzleS:,Were eierca.iSv'ed inagansawtetemPattto imProve Performance, but had littleor no effect':".'A:xeactox tiij',',,frOm aa!ii"',iinrelated::.caauSe, occurred on November>13, 1992. Duonag that oautsagge, both':.Steam generator economizer valves were.";:.recalib'reawted.

Afterstart'-up from"..theg11/13/92 trip>".steam generator 22's level was stable again.

ys e

a

~"A't.,the tiiiie,'lroccasiona1"o'svcIllatitons were observed, but at a lower frequency, typically "rig'!A v4!'+'vt

'ta

'vgt+Vu 8

a w

v s

v

~

,~~rl':bmu q'ci;:::.:,:,:,hottrvs apart. On:.,December 8, 1992, however, the trequency returned to earlier levels.

~~I:.':+

'4,'-:,:':p, Thwe:,';coraiidition perst'sated, with the oscillations gradually increasing in magnitude and

~~";..':;pgs'3~ frequen'cygtByythe'"end of February 1993, peak-to-peak oscillations of 8% were

'i:;',':,,'common. At'the time immediately prior to the tube rupture, steam generator 22's level oscillations were about 3 to 5% peak-to-peak.

w i

REV A June 16, 1993 V.

FAILUREMODE INVESTIGATION(CONT.)

A.

STATEMENTOF FACTS (cont.)

ta4 Unit 1 and 3 Steam Generators thermal generation willbe equivalent to Unit 2's at the time of the tube rupture.

.~s5~%,

3.

Operational Review (cont.)

QW Current recorder charts and interviews in Units I~and 3 did notltigdtcate,that the ott}iehr'"

units had experienced the level oscillationcsfobservedt::inIUnit 2. Tli'eIleveP'rsecorder trace width was a "nominal" 1 to 2.:.wide.:.four both steam generators iii",'e'ace'h unaffected unit. In addition to,.'in'tme'r'v'iews,'sesvsera1,ljmtozical recorde'i taraaces were reviewed and showed no evidence oficy'el oscillations'in:.the pas't'.

Comparing Operatinga

"~Tiinte Cumulative thermal generationhw'hich:'is directly,:related to the steam drawn t'rom the steamfgeneratosxs, was higher';in Unit,2 tlian.,either Units i or 3. As ofMarch 31, 1993,

..thi;;cumulative the'tmal generatioten,:per uiiitwas:

gwg~.g

'cx",:,.

g g s i'-""'.,:.::",;:;,;, ""'c:".'I&Unit.,'2150,594,902 megawatt-hours

"~m:,:.:,;i'.-,:.'Unit':,3"'119,'595,086,,:meiegawatt-hours y~f..<Assumincgmitltat Unit 1 operates at 100% power (3800 megawatts) with no coastdown

'until its scheduled refueling outage, September 4 1993, its thermal generation will have'reached approximately 148,011,150 megawatt-hours. At that time, Unit 1 's 1

I

REV A June 16, 1993 tr V.

FAILUREMODE INVESTIGATION(CONT.)

A.

STATEMENTOF FACTS (cont.)

UnitI and 3 Steam Generators (cont.)

3.

Operational Review (cont.)

w cg'."..

Ifit is assumed that Unit 3 operates at 100% pow'er.(3800 megawawtts)twith no coastdown until its scheduled refueling oui~age, March'Ydpl994, its'ttherm'al 'generation wiHreach aPProximately 151,4234886'tenegawatt;hours>theist.,:Unjt 3's thai mal'eneration would approximately'"e'qual the'samfea'oexposure Unit 2 h'ad;reached at the

'x(x 'g

~+eeyjg~

xa

$y<peQ

'Qtc:grey+

Comparing Steam Gewneerator Seco'i'idary Pr'essure Performance In the past> Unit 2 operators noted thaat.';Unit'2."ts'steam generator secondary pressures 4~"'O)~eg "4L@r

+'~~;

(as measure'd"':in psia) were'lo'wer thtoan Units, 1 or 3. Additionally, Unit 2's steam egenezator ¹2 indicated 3 to 4,psi<lownxexr than steam generator ¹1. Also, Unit 2's Ne:gr,:,x yP.:"'~<;.::.,.:",:::.electrict"altoutput dropp'ed,.8 to 10.m'egawatts during the past operating cycle. The team

'""4!::;.:';< cor"ncluded'tIj'atsthetsecondary>p'ressure observations were normal and were neither lcpxecuwrsors to the:,'tube failure, nor related to the root cause of the tube rupture.

'CH,~:

c rrt c

wy":p

~

sm

'cg>?e~

gpP Ji

REV A June 16, 1993 V.

FAILUREMODE INVESTIGATION(CONT.)

3.

Operational Review (cont.)

Vibration/Loose Parts The failure mode investigation included the possible role ofVibration.and/or loose-,:>@

parts in the cause ofthe tube failure as vibiation/loose'parts alarm's,.'were""received at times in the Unit 2 control room.,On"so'me, occasions, th'e, alarms:...were""classified as "high-rate" (greater than 4 per shift';),The alar'iiis,::c1eared during ab'ri'oiinal rate 4:;::;~

":i::."::K, '4 """,:;::.;.::,'::i'Q blowdowns, however. The Vibr'ation Qi'oup identified'th'a't'::the majority of the alarms'ere on the reactor<coolant pumps sensors"and.not on the steam generators. The Vibration Group aiialyzed'the.:recorded'data, inclii'ding'the potential for signals that originated, in the steam>generator but',::were detecte'd at the distant probes. Based on analysis of:th'e'captured dete'ctor sig'nals, the Vibration Group determined that even if is'ome alarms we're';from the'steam'generators, they did not originate from the level

/~':,'P,'wheie'.,thetiibe ruptur'e'..,"'occurre'dhin"addition, the alarm rate and times did not relate to 4~$;""".W

"~i6(.

+j>a'.)>>

4>>

Cga%:~@'

~<::~W th'e':::fxequeiicy~>of steam.,geiie'rator level oscillations.

g::.:,':,.'+ Altho'u'gb"':a>loose"p'a'it c'ould have been created during a failure of the "B"main

"'i<'><...i".:~i."'eedw'a't'iipump"'discharge check valve during power ascension from the 2R3 outage,

',".';"inorder for"::itto have had an impact on the ruptured tube, any broken parts would have had:',t'o,'travel through the tubes of all three high pressure feedwater heaters. The ygrj vibration and loose parts monitoring equipment provided no indication of loose parts coinciding with the check valve failure and the height and location of the tube rupture also cast further doubt on the loose part as a cause of failure.

REV A June 16, 1993 V.

FAILUREMODE INVESTIGATION(CONT.)

A.

STATEMENTOF FACTS (cont.)

4. Fabrication s

e g v The steam generator tubing operation was performed;I'::an atmospherically"controlled "clean" room at CE's Chattanooga, TN facilitydand consiseted!of the followingsrtceps; installation and optical alignment of the How distribitioxnplates!and 'geggcrate" tube support grids with tubesheet drilling,pattern; insertionsof stubes on a'mow~by r'oYv basis with a gradual assembly of the vertical'sc"ups'p'oit:.grids as tube':rxowss.':,wnere iiiserted; placement, sizing and tube.totubeshe'et>welding soft'uxbesx,,as they were inserted; installation of batwing,wrapper)lie'is, veurtstdttcal grid",crescen¹it'-'~plates, and tube support beams; and explosive expansion (lt'expjansiori'!)'otf tubes;.within the tubesheet full

@c.~cunt

'ew S vast

~r"k The followii'ig:fabricationfa'cts appiy:

"'0,

1. "':fabrication step, by, step details:,taken from "Travelers" within Combustion

"'i~i)'x.,Engirieeriitzg.'s manufacturin'g records are included in the report "APS Unit¹2 g'-,::,:::::i:::::;:a.:,"j!ih.

oFabricatfori'Record:Search" (Attachment V4.1).

':.'::i:< 2.

U.-,boenrdt(tubesg)vbere inserted in the horizontal plane by hand with a four-man tube

'Lc;,...4 w,r handing,crew such that both hot and cold legs are inserted simultaneously.

lt I

I I

I I

REV A June 16, 1993 V.

FAILUREMODEINVESTIGATION(CONT.)

A.

STATEMENTOF FACTS (cont.)

S.

Design The Palo Verde Unit 2 steam generator tube bundle an?dltube"'syupport desigIi~p as as.',?,s:...

ggr<~

predicated on the twin objectives of: 1) Providing adequa'tee!is'uppoerete of the tubes.'i,",:.',,

against design loadings including flowinduced vibration; while";2);:siiiiultaneously'~"'~e;s;%;

Pp promoting adequate secondary side flowby iiiinimizing'tube.-:cto-tubge? s'uppor't'crevices W 0'k~r.".'!!:...

which could be vulnerable to contaminsaenet'.:ceoncenetration and:'scorer'opioean due to out-of-

@cd'Q, specification secondary water":,chemssistryy.

e a

e The following design".,'isla?ed fac?s"ai?i?iy:

'"":::!iQ e

g e

y

~

1.

Palo Verde Unit'-2;tubing'hags',the,following char'acteristics:

4~~::<

4~) g aP;.lt w'iis':produced by'¹randa in Caiiada with a pilgering process.

>p "b.;;...Thesnsteam"'geenerator~resquirs'ecd'1l, 012 U-bend tubes which were 0.75 inch OD e

a g

a e

ee y

'aensd,haadia 0.042'in'ch,wa?H thickness and average 57.75 ft. heated length.

e!:::!a

'~~'..'!:;.,"?~ 'y,:qj':,:,,":,:.The tub'ejriateriaf'is Inconel 600 high temperature annealed SB-163, with an

~d?':,,:,;,ueppser'Immit 'of 55 ksi on room temperature yield strength.

REVA June 16, 1993 V.

FAILUREMODE INVESTIGATION(CONT.)

A.

STATEMENTOF FACTS (cont.)

a.

Primary Tbpt 621.2'F 5.

Design (cont.)

bent:.g< vt)>(t

,e Q.:i.")Aqw 2.

The steam generator operating characteristics which"~wouldljmpact th':.t'ubing are as follows:

4Q.

Sg a

v b

g a

gae$ $

c.

Steam pressure = 1070 PSIA <<g<,

3.

Based on CE experie'n'ce'-gPalo~Vered'ebSG 22:had an',average number offactory 4.

The 08H',and 09H tubeCupports are p'artiai "eggcrate" supports. Tube row 117 is 8;;"",,;'the lowest resow',,'."naumber not fully,.cwaptured by the 09H support, and tube row 66 is thdee<loewestrow number notsIullya captured by the 08H support.

S b S C

e S

a

+:<S.

The Palo'Verde steam generators are inverted U-bend heat eschangers with g>'intesgrralt,econom'i'ze'r on the lower cold leg.

"":;,':'j~ 6.

The Palo.-Verde steam generators, which were designed and build by CE, are p'FpP

"'":;",".'::;~4~currently the only operating units of this design (i.e. System 80).

QgP.."C II I

I I

s REV A June 16, 1993

~

V.

FAILUREMODE INVESTIGATION(CONT.)

A.

STATEMENTOF FACTS (cont.)

5.

Design (cont.)

7.

The flowdistribution plates on both the hot and cold'.,side"are',designed t'o:promote cross flowto sweep deposits from the tubesheet. They'in'ay.,alsso function axs tusbe si,e(xaam jfls 4.c',

'upports but are not required for that purpose.

'~4!,::";:';,"-~,

8.

The tube supports are made of:,409,';gemtic stainless'steel jnaterial.""":-."'i>~~

9.

The hot and cold side flow'.distiibution plates"acrce'made fr'om 405 ferritic stainless e

x

10. Flow distribution. plate..hou'less,,are,dedsc'igned foi';flowe,distribution purposes.

11. Tubes,:.v'arsy in length'&om 550" to 985". )j'+

sfzxtorifices are io'cated at the exti'erne hot side moisture separators for flow distribution and mr'otsture separation performance i62 units).

dsc,i: s.,;i;:,

~j+

i.'d '

c e a

s c

,~~&'~."'-.,+

"'..",:33. The U-becnd! tube suppoits design (segmented type) for Palo Verde is similar to that t:e@v, Q~o'f'SE<Lucie 2,'SanwtOnofre Units 2 and 3, and Waterford. Subsequent units were designed w'ith'a unitized construction that incorporate the best features of the ventilateod corrosion resistant design with the unitized anti-vibration design, nfl p

14.jThe scalloped bars on the 08H and 09H tube supports were made of carbon steel material.

1i 1

i

w 1

June 16, 1993 V.

FAILUREMODE INVESTIGATION(CONT.)

A.

STATEMENTOF FACTS (cont.)

supports were designed accordingly 8

5.

Design (cont.)

+rkt4

15. Cross flow was anticipated in the upper regions o'f~rthe tube";bundle, an'd,:the tube

,.~

16. Flow velocities in the generator were designed to>be'f76% of Connor's"critical velocity for the threshold offluid'elastic,instability. >,

17. The design in the regiori sof,the'partial eggcrastwe""'smc'allosped bars provides:

- Row 117 tubes'1i'ave@ 0.016"<design'raadial clearance as assembled with the

- Row 116 would':,be 1/8 "~ttfi'om.',th',scfa11op,bar.',:li<

+t:

+~gteg+x

~u~@eSPe

- Rows118 tubes have;a'0.016'inches":design"iadial clearance as assembled with the

..."inside of;.the scalloped::bar.

e a

w o

":-::::<,,- Row 119efiills within the 09H:.eggtciate.

e...,:r~.18. Qetneratorps withhbatwings (EW)"cond vertical suPPorts (VS) similar to CE System

'4e":.".",!k g!$0 expser'ienc'ed more'wear at the vertical supports than earlier units but have

,.y!g; ""~::>.;, %':i;::;.

experiencead,.less coi'r'osion damage.

4't"'i:,'s'.$Q, vs';::,;:ev 4e:tO's';.':0'@gap.

<<>~~>4+A'y

";i:::>:.>.:",','",.. 19. D'ess>gn cheaansg'es for Korean plants included some to preclude comer wear and 4jNpe~t.:.app some cha'nges to reduce BW and VS wear, especially BW wear near the central dP'tj~":,cavity.

$';.Y t

I l

t I

I(

REV A June 16, 1993 V.

FAILUREMODE INVESTIGATION(CONT.)

A.

STATEMENTOF FACTS (cont.)

6. Industry Review A search ofIndustry Events on INPO's Nuclear Network>wa's;performed"t'ogdentify steam generator problems reported at other nuclear power"stations'.::.The areas ofsteam generator corrosion, wear, defects, and ruptures we'r'e:,investigadted~tso'ideiitify even't's that were similar to the tube failure that occuired at PVNGS'...None of:the ev'erits identified were identical, howevertsomeiof:themwere similara'saiid'.ps'rovided useful information that aided in thi*,'"analysis'iof the"PVNGS teube failure. The followingis a summary ofthe plantsevents tha't:.were identified and evalutated.

McGuire - (Westiiighouse)-:t"".".~<

'><'A.

<"y '"...

McGuire::Urut 1 was shut':'down in'Marscrh:,l98g,""when a tube rupture occurred in their We'stin'gahouse':Model D2 steam"'generator. The tube was removed and metallurgically examined. A3-inch'long axia1.:'crackpPFas found below the first tube support on the cold a

e a

ae) a t"':leg;ksidetaThe~crack grow'h~rarte",vias determined.

a a

a

'"A'.sec~ond leak o'ccurred.::in McGuire Unit 1 in January, 1992. Review of the ECT data "jt, revealed th'a(t'the,.gi'dication had been missed due to another tube failure during the

~>~>$;.. previous'inspection. Other indications were also found which were missed during the

"'p'r'evious inspection. Ahigh crack growth rate was determined.

lj I

(I

REV A June 16, 1993 V.

F<AII URE MODE INVESTIGATION(CONT.)

A.

STATEMENTOF FACTS (cont.)

McGuire - (Westinghouse) (cont.)

6. Industry Review (cont.)

Qsv.'"4 ry.,@i

QYic, Three tubes were pulled from McGuire Unit 2 during the 1992,refug'cling outage. L'ong axial grooves and 62% to 73% depth defects were found';in the pulled.,tubes';The McGuire evaluation concluded that the',ei'asks were caused by<EGA/SCC'.and were OD Qy erN!.::,.a e initiated. Itwas also concludedstshat':;cmrackStasms'o'ciited'smith g~rooves~dents, and gouges in the tube material were noitnnyal;cNo aadva,erse chemistar'y'ssfactors:%ere found during McGuire's metallurgic'at!tanalyssr's.'By',:using:pusHed tubes and bobbin eddy current "blind"tests, aii00% crack,'detectability';;for cra'cks,that'were 50% or greater through A

~

wall was..achieved.

"""<<~5

'">",':,::;s

"~~,'i+@.'(4?.;,

e

<<'e 5'j":+

r,:"',gtn eg'y

I 1

II

REV A June 16, 1993 V.

FAILUREMODE INVESTIGATION(CONT.)

A.

STATEMENTOF FACTS (cont.)

McGuire - (Westinghouse) (cont.)

1.121 limits.

6. Industry Review (cont.)

V e

v On May 11, 1992 McGuire Unit 41 detected a 235:.GxPD leak!-;The'.:"source ofleaka'genug was a one inch crack with a pinhole locate'drfive inclie's:above the fiist,su'p'port:plate.

The crack was determined to be inkiateChbyd a manufacturer 's bumishiiIgimhrk and propagated by stress corrosion crravcking. A.60tyv'.t'hwrough wall crackyosnxe inch long five inches above the twentieth support platde;:was also loc'ated,'::duringithe investigation. A' a result ofthe inspection ":j,82 tubees'j;cree reemoved from service by plugging. However, few tubes containbegd.:,.indic'attons'::of freesxpane,"cracks sgA majority of the indications were characterized as"'de'nts ordings!!Six tuba'es~Were removed for metallurgical examination'.gxaminatiori'on'f:,thee removed'tubes revealed two tyPes of outer diameter

~(OP) initiaxted axialtxcracking octcuging"inp the tube free-span region ofthe pre-heaters.

fw)tP3Cq

,r,:,'.::";:i'"';gg~3oth'tyt'P'es 'Of,degradaeatiOno weri.:.associated with mechanical deformation ofthe tube A's:ofXuii'e.;1992,'fid',:'tubes had been removed from service at McGuire Unit 2 due to

<'a"::,:::.;grp<ree-spxah crack-'Iiike indications. High residual stresses that exist in the groove regions

'~>'. may have confused the eddy current testing results. Duke Power prepared a Reg.

Guide;1.121 analysis and determined a crack growth rate. Based on this growth rate

~g~P they determined that the unit could run for 12.2 months without exceeding Reg. Guide 0

REV A June 16, 1993 V.

FAILUREMODE INVESTIGATION(CONT.)

A.

STATEMENTOF FACTS (cont.)

6. Industry Review (cont.)

ge o

e

(

Maine Yankee Atomic Power Station (CE) deja~;:;,:,,

Maine Yankee was shut down on December 14,'1,990 when a'<1.4'.OPM.primary to;",:~>>

secondary leak was detected. The leak was identified,'inthe ¹1 steiam.ge'naewr'atorgon December 12, and gradually increased'f'iom,.0006 GPM,to 154'::.:GPM w'lied'the shutdown was performed. The smource ofthe"legate.waas,determined tpo';beau 2 inch long C"":,

axial crack at the apex ofthe:.:~V,'bend'in"the steam gen'erator tube".This location was '

g'.':,'.jQ Q

'%Q described as a "steam'.",blainket region~..;,.;wher'e'ithe batwing supports restricted flow, permitting a steam; void to'four'm';.and,conletawminaiitst'to;,be~deposited on the tube surface.

p Arkansas:;Nwuclear One:(ANP) Untt'2'(CeE) '~P"

<'ANO Unit 2 wacs schutdown'on Mai'ch 9j'1992 when a,25 GPM primary to secondary

/g;"':;"',@leak was.:detected. EQ'.'(Eddy'Current Test) inspection of the steam generator tubes wwa's conduct@dc using an MRPC (Motorized Rotating Pancake Coil) probe. The ECT

'~~,';:,,:,. 4+ape'ctiown results"',.identified the source of the leak as a circumferential crack in a tube

~ ',..:~ at the:.:hot'ieega'.:;expansion transition, near the top of the tubesheet.

~jFQPC~

4

REVA June 16, 1993 V.

FAILUREMODE INVESTIGATION(CONT.)

A.

STATEMENTOF FACTS (cont.)

6. Industry Review (cont.)

Arkansas Nuclear One (ANP) Unit 2 (CE) (cont.)~;-:,~X z&t, Based on the finding of the circumferential crack-;-'a.,l00% MRPC'.'ijrspewction was>"-.'.".~

se:::.eex~

Qy;QQ ~~Q~@K;,.

conducted of the expansion transition locations on thee'boost;leg side::of...both':SG,.s. A 20% MRPC inspection of the expansion',transition on the cold,'leg side':of::one SG was also conducted. Indications (genera11y circuiiifeiential) wer'e,found'o'ii"'the hot leg side 4-

":=-:

of488 tubes. No indications "wer?eefouridhon the'cold legw:::iide. Tubes with MRPC

""'@ax indications were alsoa';:ins'pwected w'ith':a.bobbin:;:prvobewhowever, that probe did not detect most oftlie!ivgutt 'ihdtcatiorns. Thr'ee tub'es'::were'putted for analysis. The analysis determined th'iitsthe chuumferentitd"cmcking was caused by integranuiar stress~corroasxion cracking (IGSCC). <

Do'el.:::Unit 4 (Belgi'u'm)

Do'elyUnit '4..'experienced an".event in the past when a lead object was inadvertently left

'%+~h gC'F::.'::,."..',

"'~!w~ 'tnt;the'steam gernheoratorpie event was reviewed extensively by various industry groups

'~,",,> in agns",:effao'it'rto':.:establish and define the impact lead has on steam generator tubes and

'""v?.

"~;j.;".',vg subsequent'.tube cracking. Doel Unit 4 experienced cracking in the freespan, tube

support, and roll transition regions of the hot leg tubes. The results of the analyses on Doel"Unit 4 were reviewed for applicability to the PVNGS events. (Qo

,li l

ii I

f

~l

,(

i Ir

June 16, 1993 V.

FAILUREMODE INVESTIGATION(CONT.)

A.

STATEMENTOF FACTS (cont.)

EPRI

6. Industry Review (cont.)

aa??~',

y?stn tv?':n In 1991-1992 EPRI contracted with Dominion Engineering tto"invaestigate the extent'nd prevalence off'reespan cracking in ste'am? generadtosrres',",.'.Dominion'n',surve'yroed:.recorded Cjj>;:;.;,?,

vweXp~,

industry events and reviewed their.'dna'tabase,of eddy curient'nresults in orrdter'o ascertain the extent of freespan"'defects. The'si'gnificant events reponited at McGuire and Doel 4 (Belgium) were in'eluded in'.their review. '"<"':j::,':

~j$

Dominion's investigation. found thatafieespan"'iiidications:were not confined to a single steam generator design,nor were!thney singut'ar in nature or cause. Some indications ttgear ggo

?"', 'g??o~tts were long:':,axial cracks that;.were fourid':in?the h'o't leg tubesheet crevices in part depth rolled units. Based on information available, Dominion was able to establish that the craacoits;were induced:,by caustic IGSCC. Other freespan defects were believed to have g::;::i~

v?ebee?n thus':?result::of tube miinufacturing or installation activities. Based on Dominion's

,~"':dg<.,',.::.p>

"":'ii.-.,findiiigs,these,:tripes of.defe'cts have not been significant in number or impact within

>';::;.:;:,h, The scope',.'attd.results ofDominion Engineering's survey are as follows:

V<.'ft!P"

~Y

~ '+>!ll,,':plants containing 59 steam generators were surveyed.

~

15 of 17 plants had identified tubes with free-span OD indications.

~

11 of 17 plants had plugged tubes with freespan OD indications.

a REV A June 16, 1993 V.

FAILUREMODE INVESTIGATION(CONT.)

A.

STATE~MENTOF FACTS (cont.)

gg a

6. Industry Review (cont.)

EPRI (cont.)

a e

u a

e s EPRI was contacted by telephone on 4/18j93,'to discuss;:,the information provided abovue;;BPRI recommended'that.

PVNGS contact Domuiion Engirieetgr'mg':directly:for additional details'aiid information.

e y

u e

e u

In the course ofthe comnvspersation'EPRI,:stateadithat some cracking had been identified at xe

,t)/rt Calvertn'Cliffs,(Subsequevunt:.discussions"'with Dtouminion Engineering found that the

'c'tel;CQ Calvert Cliffs"tcracks were asuswom'ciated with burnishing marks and irregular deposits at

~@~

'"'~"-:g4 EPRI.,statedithwat deposits, diie to solubility changes, had appeared approximately 1/2 g,::;::;b, w'ay,'.up,'::on, steam"'aenerat'or tubes at the Ginna and Surry facilities. EPRI stated that it i;:pter tt::.;:fwaspbrobabieitbat;.tbgose deposits bad been formed during cool down from a viscous NA>,mixof magg'nme'tite and water.

(S~i~'g e

REV A June 16, 1993 V.

FAILUREMODE INVESTIGATION(CONT.)

A.

STATEMENTOF FACTS (cont.)

0 Dominion Engineering

6. Industry Review (cont.)

%$.@gal.

g?)

Dominion Engineering was contacted via telephone, on 4/20/93;:"and.4/25/93. It was:",">

Dominion's opinion that the best information regarding!midspan cr'ackirig"'could be provided by the McGuire facility,(see'.;previous'reference).

Doiiiinionsta'ted that polishing and straightening stresses":,:could'be"':1'0::-I'5;.:KSI (thousand':pounds per square inch) but that scratches could;:not.be ruled.out as. a cause'.:-'.Dominion stated that the problems experienced,,".at.""Doel were',aggravated.by the existence of a lead blanket in It was thejrrecommendation,that-P>VNGS,,cont'a'ct Florida Power and Light (St. Lucie)

"s'j&:y'.Q for'additional":information regarding their, experiences with steam generator tube cracks/indications"..Dominion"stated'that they were aware of eggcrate cracks where

+~line contacts'b'etween a'tub'e,and~eggcrate could develop into a residual crevice. Pulled

+~:,"'::;;p< ":l'A",tubes'(at both"A'kansas'Nuclear One, Unit 2 and St.Lucie confirmed that evaluation.

Dominion'Engineering'was not aware of any detailed evaluations on freespan cracking

"'-'@+~i"- and could.;only'rovide speculation regarding the cause and process.

X<Pj@,

J

REV A June 16, 1993 V.

FAILUREMODE INVESTIGATION(CONT.)

6. Industry Review (cont.)

A.

STATEMENTOF FACTS (cont.)

A

+st:j; >

vIr e,.

Florida Power and Light ss

. ~ "~A~e Based on the recommendation ofDominion Engiiiecring, Florida:Powpers and Lightwas'ss contacted on 4/25/93. FP&LIndustry Review stated: that;St. Lucie..had n'orstIldentified any mid-span indications on theirsti'am';jeneratox tubes..Th'ey.."stated thtat':the cracking Problems occurring at St. Lucie'we,erwegouter,diam'et'er.'stress corrosioii,':.cerrsacking s!llih (ODSCC) and IGSCC at supports. Th'ey'rstated that huiidr'e'ds ofde'fects were present in the eggcrates.

They~1i'i'd~also obser'v'ed.steam:.blanketing on horizontal runs due to the vertical straps acttntg,as steag'm::'.traps~(NOTE; The'System 80 design has ventilating holes in the straps to psrevnent this frtoirithorcc'mng)r PP&L suggested that liftoffofECT t'rom~ocvtdtsatston should be"evcaluated:as totwhether such action could cause the bobbin dtor'iniss an indic'ahtion. FPAL'ls'o.:,stated"that Turkey Point had hundreds ofrandom g"qjebumstsliing points but';";to,;:date, n'o cracking had been observed at mid-span.

~<,",.,

~

Some in4hcatioiis were detectable during ECT by MRPC, but not with the bobbin xs)srgt "4kDreposit indications were recorded in both SG 21 and 22.

Q'es'gs il

I a

June 16, 1993 V.

FAILUREMODE INVESTIGATION(CONT.)

A.

STATEMENTOF FACTS (cont.)

face and 01H.

7.

Deposits (cont.)

,pg sewage ga 4jx'slav,pg

~

Possible loose parts (PLPs) in both Units i and 2 were dtspros4tioned"aas';deposhs in past ontages, several were located high in the SG.

~

Limited MRPC inspections at elevatio'ns below thi;,:.07H identi6eed'at,avery...swmall quantity of free span deposits, located primarily betweaeeii:throe tubesheetrsecondary

~

Axial free span indications';corresp'onded to,.deposit:indications; however, there

%~)h.,

'ieger were also deposit:iiidicationoes',w'shout ideritified'axial Qaws.

a

(

e a

av g

a o

s

~

A secondary'side~vide'o~iiispeeoctton of<he tube,'bundle showed deposit bridging between tube 117-40:,and li'5-40iiiiaSG 22';g)ridging was observed in the video

@Jr'o Q

mast'yectiopii',of tube 127-140,',,

"9,"':,69,,pairs of adjk'cenat tubes+in ceolu'mns were noted as having possible deposit

~:..::.'~ '4ii:::::,':-,;~bridgIng:inSG 2$ :"4:of~7 mid-'"'span axial indications in SG 21 were aligned in these e

a bee e

S e

c w

g a

~ ""iS7'pairsrof adjarcserif tubes in columns were noted as having possible deposit bridging m"SG 22. 19 of 22 mid-span axial indications in SG 22 were aligned in

<egAN<

these s'ets.

gci~~:.'.'sr+

~

'~BCsT measured length ofdeposits varied over the entire range ofthe mid-span. The

~AY median was -8 inches in length. The mean was -10 inches in length.

I J

'L REV A June 16, 1993 V.

FAILUREMODE INVESTIGATION(CONT.)

A.

STATEMENTOF FACTS (cont.)

7.

Deposits (cont.)

~

Based on ECT, 6 of7 mid-span axial indications:ia':.SG 2lqwere aligne'd:,'with a

":y deposit.

~

Based on ECT, 21 of 23 mid-span axiaI. indications:in SG 22"'w'ere'"aligned with a deposit.

~

Lab ICP (ion coupled plasma)..examinatioii'."of-'.loose surface deposit removed from a section of tube 105-156 ':piece 15',"'.,:indicated the"pre'sence":of'lead.

~7'.

~

Visual examination.".,of::,.tube 1':17'line 40 piece 17~at the BWNS lab identified a long ridge df'axIal deposit';IThe ridgeline begaii:.;at. the top of the section and tapere8,of -10" belo'w>the cut. Clu'mps of;:fibers, up to 1/8", typically 1/16" long,

~

~

+ere observed on and eiiibedded'in the deposit. The fibers were identified as 4M~>graphite, 10 mds,thick. The'lab",verified that the deposit was aligned over an axial flaz)SEM/EDS performed<at;.the OD identified maganese, molybdenum, and

"";;:':.,:;ExanjInation"ofp:.cr'oss section from tube 117 line 40, piece 17, confirmed the pi'esence"of'.intergranular corrosion associated with the ridge of deposit observed y'.C,"..',.

duringIth'e'receipt examinations. The area ofIGA/IGSCC extended below the

'w:,"",',deposit ridge in a straight line to the end of piece 17.

gap t

June 16, 1993 V.

FAILUREMODE INVESTIGATION(CONT.)

A.

STATEMENTOF FACTS (cont.)

r Nv.

7.

Deposits (Cont.)

e h

s e

~

Lab examination of the deposits &om 117-40 identified a~loose, powdeiy~surface deposit, that could be scraped offeasily, but.,under thatawas a 2 mil thick deposit that was extremely adherent.

e a

h e

vg

~

X-ray diffraction results of tubetde'posits fmm tube>105!15udxbidettstffiead,the major component as magnetite.

Ql'.:::~"'."),,

~

Pb, S, Ca, Na, Mg, Cu aiid,'.other cwh'emical'constituoeen'ts>were+identified in the deposit removed!toom,the area:oftthe dewfects in tube 105-156 concentrations appear consistent w'ith':industry:experience.'P,,

~

For tube 117-144, 'deposit analyssis:,i'.beii'i'g<p'erformed.

~

Defcects'there OD,:mitiated IGP': and IGSCC. The cracks were axial and often "C>",,'.:!b.

~!d'I:,::X<Th'e.,free Spapn':defeCtSx Were aSSOCiated With ridge-like depOSitS.

';::.)I ~

Cruaeckirigii8'-':at"'.combination ofIGA and IGSCC, with the trend towards IGSCC as

~e<jiw the cra'ck;propagates throughwall.

4+Midspan crack indications tend to be long ()10 inches).

)p"

~

O,D. surface, midcrack and crack tip oxide elemental and compositional analysis have shown the crack crevice chemistry to be alkaline.

REV A June 16, 1993 II V.

FAILUREMODE INVESTIGATION(CONT.)

A.

STATEMENTOF FACTS (cont.)

8.

Tube Examinations (cont.)

v a

v v

g

~

Oxide chemical analysis also indicated the presence':of Pb'.SO4, S

Mn;PCg, Cu and Sl.

yp~

~

Cracking (i.e., near throughwall and tliroughwall),is<associated:;withtlineary ridge-like deposit formation in the upper,,:tube sections.

vt A "'%Wl'-.:-:: <

~

Cracking (i.e., near throughwo "all!send thr'ozuxgahwal1),,is also associartevd with worn 4 """<t""""W surface areas which hav'eiuumeroiisiaxtal driente'd"scniatchesiwithin these areas.

~

Tube bowing wastt'dentified boot(i~b secondmy side visual examination of the steam generatoi'itubes ande'byrlab me'ansuremeiits,"-::y"

~

Tube'ZGA/lGSCC was.'iidentified'in;areas'w'ithout significant worn areas with gs'csratches",;however, the eexstent ofthe attack was not as severe as the worn areas.

~

""'L'ab eddy current:,;data is beingcgoyrrelated to field data.

~'v'!yTube'material.sensitizationrtesting (grain boundary chromium carbide formation)

Chsplayed a(low dejiee of sensitization.

'i% ~

Tube bulk,:,cheniistry data for samples received to date meet specification

"'<.';::,;k Rr:',';:,x requir'ements.

oft e

I lf

'll

REV A June 16, 1993 V.

FAILUREMODEINVESTIGATION(CONT.)

B.

FAILUREMODE REVIEW Based on the facts that were known early in the Engineering RootYCause of<Failure Analysis (ERCFA) process, a failure investigation plan was>developed. The:plan's goal was to ensure that potential failure modes were identified an'd!eyalu'ated. The plati':,%as

>'~a used to identify required data and methods for obtaining the infoitriatiotl'..to.,support or~!:.P refute the failure mode. The potential failure'modes wer'e'ev'aluated base'd on: the.;known

$ '+c,'i.(jgng e>""".'~W.

facts, causes ofsimilar events in industry,:;and,:;.knowledge of";the'behavior of.;Itic*onel 600 tubes in high temperature environme'ritssThe "pla'n:ji'isented as~Figures'!'VB.a. &VB.b.,

the Failure Mode Analysis Chai't'assistedthe:,team"in troubleshooting and performing the required analysis.

4 x4a WjMgy

'4 'I,:i:

~j'par...

'jif:'jig J

(

REV A June 16, 1993 V.

FAILUREMODEINVESTIGATION(CONT.)

B.

FAILUREMODE REVIEW (cont.)

Outside Diameter Stress Corrosion Cracking One potential failure mode, Stress Corrosion Cracking (SCQ), is'iitttiated froiii'.:tileoutside

'Oi:::::.

diameter (OD) or secondary side of the steam generator (ODSCO), This phenomeiihaihas

'"ji:<~'":pP been observed in other pressurized water reactor (PWR)':,jlants that>tuse,."Ine'cone1.600 SG 4g"tv'~

Q>>j."fp~)

'<z>.')':;)4 "4$t<e~a'p tubes and ODSCC is a recognized cause:of SG tube failure. Tiie:,'cause ofODSCC has not been limited to a single factor but isla!c omb hia totnof'various foims and magnhudes of three factors: material type, entronmeaiit<i:and stress.

(Reefer;ton Figure V'B.c.).

/m,.v K

S..'.<'vP SK 5.>

The first factor is the type'.of mateiilaLused.in the:,steam generator tubes. Inconel 600 steam generator tubge:material has be'en shown to'be'suscepiible to ODSCC. The tubes used in the CE System" 80 SG s were high teiiil>erature millannealed and designed to be resistant to ODSC~C!>The P>VNGS.gmenerator htbing<(vas to be examined/analyzed to ensure that the b eh e

z s

z v

""~"'XtX

~

,.<pa':,:, material properties;.of the a~ctualetube<ts were consistent with the specifications for Palo W, YP' s

f a

h

,,j4'>@gb+Vexd>e Inconel 600:Itube material.

~ < ",f" s

f

'I I

I l

s s r REV A June 16, 1993 V.

FAILUREMODE INVESTIGATION(CONT.)

B.

FAILUREMODE REVIEW (cont.)

Material Type (cont.)

.t" ""-'-";~

The thermal treatment process applied to the tube material:can affect;the resistance to intergranular attack and SCC. Therefore, it was important to v'eitfy that the heat treat<

~<<'<. e

+:YYYvs+c process used was proper. Areview ofmetallurgical da'ti::,for the tubces':use'd<:,in<the

~,"<~yP).;~a fabrication ofthe Unit 2 SGs was conducted. Minimum certifie'd'yield stxen<'gth da't'a forthe tubes was also reviewed. The review,';::is fuxothex.,:di'scusssed in S'ec'tioii'VI!E,7.e.'ctual tube material properties, including m'icrostructure characterization, will,be determined as a part of the tube metallurigical,,analysis>"":,",+>

gp Inconel 600'ste'am generator'tube material"is:a ma'terial that is susceptible to ODSCC. A significa'nt factor:iii',the developmeiteof SC<;Cis the environment to which the material is exposed-!The rate and severity of<SCC'c'an be influenced by the pxesence and c

s s

e s

e

':";.::.> co'necentxatio'n;o co'ntaminaen'ts.- The!'paresence ofcrevices or deposits can result in localized

'aeieas ofi'csonceniiated chem'icals or a more severe pH.

Vl I

l 1

I

REV A June 16, 1993 V.

FAILUREMODE INVESTIGATION(CONT.)

B.

FAILUREMODE REVIEW (cont.)

Environment (cont.)

'+4:"':&q

'+5:,:.'.2;;.,XQ5,:;

Chemical contaminants in the SGs are controlled by main'taming"'the.,bulk cheiiiis<try of secondary systems within plant chemistry specifications. BaseCh'on the<bulk chemistiyof

@0P,g~,"",j;.'~<

+,'.;4P Vi4....

+:':.'."<4@m':i,>",::".~

the water within the SG, a prediction ofthe conditions:within cxevic<es::an'd':;deposits can"be derived using computer analysis. A historical review of Unit'2:.plant chemistry da~t'a was conducted. The review included a compari'son.,of;:Unit 2 chem'istry w'ith<thatiofthe other PVNGS Units. Chemical analysis',of deposits oii'the tub'es'>and spectrometry ofcrack surfaces provided insight,into the't'y<p<e;.and coric'entration of ch'emicals within the steam generators.

The rate ofSCC for a given,"..'chemical e'nvn'onmen<tls>also influenced by temperature. The

""<c<:,<;g pressuri2'ed waterxeactors at PVNGS operate%gth a reactor coolant system (RCS) temper'ature th'at is htgher than tnost:.plants%Hot leg temperature does vary with reactor

,>':gower lev'elyThelplants 'a'e,',base lo'aded at 100% power, which corresponds to a hot leg

'"',;:.tempe'rature o&621,'FW

+"j,,', p.

<'<:.'~.g

.;:b'<:.'<>(< ~g

':g..":;~/<'y

+P8w~

P ~

~ g J

au e

h REV A June 16, 1993 V.

FAILUREMODE INVESTIGATION(CONT.)

B.

FAILUREMODE REVIEW (cont.)

Stress Stress levels in the material can also accelerate the crack; jroewth'raretef In assessiiig the effect ofexternal stress, both residual and applied on the crackiiig:in Uiut 2's SGs, several sources of stress were evaluated.

g bag e

e e

g In terms of applied stress, several knoewne';.stree'sssa..scontributors::were"extawmineud IThe first was hoop stress due to the pressure differ'etttial bet'Werc'eunn:;the.'RC5 pressure (isriside tube) and the secondary pressure (outside tube),"',The diffesrential'pressure tis essentially constant about 1200 psi. This stress is kno&.to be'piessent iri'th'e;:,tubes'and has been estimated to be of the order of 10 KgpfrtliousanCh'of pounds pets quar'etinch).

Another sour'ce;.:.,oaf applied stress,,coul'd,be~thetbowing of the tubes 'as a result of restricted growth'i'n the bu'ndle:,due to heat-oup;,or other, phenomena. This circumstance can be evaluathe'd',.analytically"ahndsby physical. observation during secondary inspection of the n

h au a

r g

4':.":;:gTthe:,stress,'::associated'with th'ermal cycling was another potential source that was i'yaluated.'<<Thgat'tyjie:.,of shtress results &om the differential temperature across the tube wall.

~-'The difference!in;.temperature varies as the tube is surrounded alternatively by saturated wateer':;",,two phase mixtures and steam blanket. Based on assumed thermal cycling conditiongs, the maximum stress estimated to occur is 3-5 KSI.

VJ

REV A June 16, 1993 V.

FAILUREMODE INVESTIGATION(CONT.)

B.

FAILUREMODE REVIEW (cont.)

Stress (cont.)

<C@~

Stress levels in the tube material could also be caused by;residual!'stress. Th'at'type of stress could be caused by certain m es ofmateri~ defects such'm'scratches or cold eeXq,,'(@~

'rx working on the outer surface. The presence and degree,:of:residual sttres's"'cxoultLbe 4! t';:~e determined from the metallurgical analysis. ofpulled tubes, '~wj;:,":.~

Cyclical stresses on tube material,carilead to mi'tearial":fatigue. One condition that could create cyclic stress on the tubes is',tthevibrattorono ofthe tubes'."'.Vibratlnal types are torsional and radial, and there are n'iethods of;intdiicing 'vibrteaetion."Original CE analysis ofthe SG does not predict a sig'iiificxant vib'rationalt am'plitudegregsulting from normal operation. SG 22 was known.to have had feedwater leva',el;.oscillations'during the last cycle. Analysis of the potential eKe'cts of the oscillations in,.SG 22,'s level and the type ofvibration in the p: "'.::,::au e,,.,:.'.,",< Theriiiare otliiir;:merechanisms:::as~wxrellthat could create stress on the tubes. For example,

'kSi'.

,.',."lilies~> fi'etting'wear of the!tubes diie to impingement on the support structures or other tubes

%'::(mid-'spa'n)':;:.is:,: joessible,; Yhawt type of wear could be determined by examination of the Ypex";.,br gVV4 0 I 4

e

Fic~v~

Root Cause Investigation Team - SGTR Event Failure Modes Failtire Modes ODSCC Deposit Induced Lockup Cross Flow Supports I

~ Material t I

Stress Environ-ment High Stress During HU/CD Abnormal Flow Vortex Direction Shedding 5 Velocity Abnormal Vibration at Supports Loose Supports at 08H Properties Fatigue Residual Applied Concen.

trating Mechanisms

-: PYISCC

'Bulk Chemistry Crevice Conditions Deposit Conditions (Steam Blanketing)

I

(

Timeat I

Temperature I I

Heat Treat Fabricalion h

Micro-stiucture Vibration Torsional Vibration Radial material

'efect (Scratch)

Material Property (Heat Treatment)

Bowing Hoop/

Pressure Thermal Cyctingr Secondary Stiesses Fretting

~.wear:;,.

Caustic Acidic Pb and Sulphate Accelerant Caustic Acidic Pband Sulphate Accelerant LEGEND:

O Confirmedbased on examination/

testing of pulled tubes D

Analysis required to support 1

I,.=-,-, I Known to be susceptible/factor D

No data has indicated problems

+~@" lnthlsarea

~ 15%5 AOO55

r ~

~

K 1

I I

I I

Root Cause Investigation Team SGTR Event Failure Modes

~ ODSCC Material Properties Stress Environment Heat Treatment

~ ILmAHeat Deal Records:

- WithinSpecs (low end)

- Could nol conelale with tubes for trend

- Yield suength as lav as 36 KSI (less susceptible lo corrosion)

~ Tube examination:

- Yield strength to be determined Microstructure

~ Examined tubes allah:

- ll4crostructure was acceptable, nol ideal

- Nol scnsitifed

- Relatively finc grained Material Composition

~ Hol conlunied a~set:

- 5600spec.

C clic Vibration

~ Evidence ot wear hlotion

~

~ Higher wear in U2 than Ul. U3 (622 > 321)

~ Wear conlacl force and

~

motion wear =

i

~ Normally produces circumlerentialmacks Pressure

~ Approx. 50PSlmaxkalxn

~

~ Considered negtigibte Thermal

~ 42-5.0KSI

~ Level oscitlation could induce cycling

~ SG2>SGIP2)

~ Unable to verify Two-Phase Flow Instability

~ Possible based on analysis

~ HOmetattrxgieat

~

Residual Cold Work

~ Residual tensge is assimlcd on some

~ Deep IGA'SCC associat (blctattuigicat 2)

~ Wear marks indicate presence of cold work Scratches

~ Most have IGAISCC associated(metakrrgbaB

~ Humerous scratches found on tubes Tube Ovalization

~ Not seen in Lab

~ No cracks in curve

• Material Heat Treatment Physical Damage

~Fretlin:

~ Possible indicadonin Smiled areas Physical Damage

~ Wear:

~ Present in supporlr scallop bar

~ Not seenin areas with significant damage.'axbtts

~ loca!iced effect

'ee ) (a teria) Properties A

lied Hoop Stress

~ Approximately 10 12 KSI

~ Known lobe present Bowing

~ Indicated by deposit kcation

~ Video review'Ub observation

~ Can resull from fabricalion

~ Stress factor is minimal Thermal Expansion

~ Designed to expand (analysis)

~ May result in baring

~ Can cause support mls.

alignmenl

~ Evidence ol gross wear in upper region could be related Support Misalignment General Characteristics

~ Reducing

~ pH (Nominal) 9 alkaline

~ High Fe transport

~ Hideoul return SG 22 >.

21 (S024, Ha)

~ Ccnsistenl vifthother Unit

~ bblar rabos > 2

~ Operated as Cu free plant

~ Resin intrusion known Caustic

~ Muheq predicts caustic

~ High Ha 15024 return

~ High prompt Ibr), hi communicative S

~ Some chromkrm depleted on tube fracture faces

~ Thin oxides on crevice (5000 A)

~ Molar ratio much >2 as high as150

~ Cu return as With SO24 indicaSng CaS04 Ppf (outside crevice)

Acidic

~ Acidicpossible ~ BrW cvkfcnce sup porls Hi depletion at surface

~ Not supported by Bulk Chemisby Accelerants

~ Pb IS

~ Cu(bav)

General

=:

Characteristics c High Fe transport

~ Iknimalat bwer Knovm to be inupper region '

Thermal hydraINC model

~ Chemistry concentration

~ Appear lo bc hl pairs bddging'contact

~ Afigned in cotumns Uniform (porous) ridge (tenacious,lpoious)

~ Edge deposit accctcfant conccntrationis2-3 times higher

~ Kdgcs are axhl and i

correspond lo cracks Formation of Crevice

~ SI4>port contact

~ Deposits

~ BOwing bibcs Bulk Chemical Crevice Chemist De itFormation Tem erature

= General Characteristics

~ Known lo be greater than most plants (fh > 621'F)

~ Highesl at TS 605'F

~ v560-582'Fal09H '

Bndgc deposits ctcvalc focal temperature (up to RCS temperature)

Time General Characterist

~ Unit 2 longest continuously run u.

~ Nominal operation lONi for bngcr lim Ihan Units I and 3

~ Most opemlional bi (150x106 MWha.

(Itfh more) x h

' it amass ROOT I

FIGURER~

V. B.c-.

Material Susceptibility Material I 6-00 Anneal Temperature

"::;I:-';,':-j~'-',::::::~Tens I Ie ;-,'.:

.",';:Stresses Environment Neater Temperature LTMA)tubing ed resistance ature Mill Annealed (HTMA) (1800-1875'F)

P

.,',i ow.Temps'rat'ur'e':Mill Anneal (

'.;::..'".';is mo'st'...sus'c'e'p'tible',,:".,',.':-,::,",:,:;.'-':,:,-,

I:-'::;:

0 Thermai treated tubes:increas 0 ABB-CE tubing is High Temper

~ 0

~l