A,C CELEBRATED . D1 BUTION DEMON STRIDOR SYSTEMh REGULATOR FORMATION DISTRIBUTION S~'EM (RIDS)

ACCESSION NBR:8812130172 DOC.DATE: 88/ll/18 NOTARIZED: NO DOCKET Nt

~ ~ OEEDENDEE SNENT NERONT ELDNI

~INERT agel niIASLO CAN'ION NiT 2

~ RT ~E ~ y,~ s ~ ~ ~

PP DA4Q> QE@„1)V)RETAINENG RLSX STUD DREAKAGE DUE TD ENTERGRANULA

ol

~ &Tl0' ~4 g4tia

OlNLO CANYON NIT l~ RKILTTOS «N»LWD ~

ACNE ~ li»R ~ j~ s e ~

10 9 bbbe 't4 00»IR»»T iamma) SWANN» h Te DINAR»«RD»TDP a er» f EST> ~ Seoe

~ N

000V

~ ~ '+P~ rAy

VOuethRY REPORT

MAT&natl W tm St» TEDi

AHES C. NOLAN, REGULATORY QNPL1ANCE NG1NEKRTI.D~WE»~»

Qe+VolN~

B10874AA

«NNP AETN»R ~

395

~««INTR «RR L«R N»» Dan ~NT D«EA«E INC»ODD a»»a eWST ITS~ O T+Jot"4 4PSDE

ES

M~DR

OJ aSW»TNS»IN»ei DDNDETS ~ i~ R

»DR ~m ~ (+%PIG Mhl+4 HTD E ~v ~ T DNI'I'his

report 1s being voluntarily subi1tted for )nforeation purposes only asdescr1bed in ItesE 19 of Suppltatni Nuaber l to NUREG 1022.

On October 9, 19SS, at Oi30 POT during a preventative aaintenance (PK) inspect1on,both re|a1ning block studs for check valve RHR-2-b740h vere observed to be broken.The four disc are alignient guide pins vere intact. Nanual valve cycling during PHshoved no significant sisal)gnaent or disc-to-body contact. Th1s valve had passedprevious surveillance testing and vas considered operable prior to d1sasseably.

I

The failed studs are aude of ASTN hl93 16 Type il0 SS. fascination of the studsShoved 1ntergranular stress corrosion crack1ng as ihe failure Nechaniss. Siiilarvalves in Un1t 2 vere 11sasseab'led and ihe studs vere rtaaved for

inspection.'o~destructiveexaa1nation of the studs froa the other nine Unit 2 valves shoved no«bnoraalities. Studs in all siailar Unit 2 valves vere replaced vith vendor

'ec~ndedASTN A55i type 630-1100 aaterial.

Radiograph)c testing of the Oe accessible Unit l NR check valves confirmed therevere no broken retaining block studs. Herever, coeputer-enhanced radiographyshored a possible. crack 1n one of the studs. The studs in siailar Unit 1 valvesvill be replaced by the end of the next Unit 1 refueling outage.

2~1esi0065r.

8812130172 881118PDR ADDCR 050003238 PNU

12. 06. 66 12: 26 PM P03

NIIC fere EDDA(DM)

fACILITYNAME I'l

UCENSEE EVENT REPORT (LER) TEXT CONTINUATIONUS, NVCLEAIIIIECULATOAYCOMMIDSION

AffAOYEOOMD NO 3ISOWI04KXDIIIES'/DII

DIOllENTIALrrvrr I e

IVrDIOrr~rQrr I e

DIABLO CANYON UNIT 1TEXT INnsae Meee ~ teerrrerE ~~ HIIC ferrrr ~'el IITl 0 5 0 0 0 3 2 3 8 8 0 4 02 o~ 09

Unit 1 was 1n Hode 1 (Power Operation) and Un1t 2 was in a refueling outagewith all fuel removed from the reactor vessel when the event occurred. BothUn1ts have been operating at various modes and power levels with these brokencheck valve reta1ning block studs.

A. Event:

Valve RHR-2-8740A (BP) (V) was chosen for internal 1nspect1on 1naccordance with the preventative maintenance program administered byHaintenance Procedure (HP) H-51.14, "Check Valve Haintenance Program."This valve 1s located in the Residual Heat Removal (RHR) hot leg in]ectionline immediately adjacent to a piping elbow. This valve is an AnchorDarling model S350H 8-inch swing check valve.

On October 9, 1988, during a Unit 2 refueling outage valve inspection, twobroken retaining block studs were found 1n check valve RHR-2-8740A. Thebroken studs were found after the valve had been manually cycled throughits travel arc twice with no apparent problems. One stud was severedflush with the valve body while the other stud had a stub extending about1-9/16 inches 1nto the retaining block. Four alignment guide pins wereobserved to be in place and intact in the retaining block to valve bodymating surface. The studs were made of ASTH A193 86 type 410 stainlesssteel (SS).

Each Unit has a total of 10 valves of this design installed. All of theUnit 2 valves were disassembled and those Unit 1 valves that areaccessible were examined by radiography (RT).

On October 10, 1988, ultrasonic testing (UT) of the check valve studs fromvalve RHR-2-8740B (BP) (V) showed no discontinuities. Disassembly of therema1ning Unit 2 suspect check valves was 1nitiated. During documentreview, a previous occurrence of a fa1led retaining block stud in a checkvalve of this design due to intergranular stress corros1on cracking(IGSCC) was identified in 1984 during Unit 2 startup testing.

On October 13, 1988, RT of Unit 1 valves RHR-1-8740A 5 B cceeencedutilizing a Cobalt 60 source. The studs from the other nine Unit 2 checkvalves were magnet1c particle fluorescent dye tested (HT) and showed nocracking. Reassembly of the Un1t 2 check valves was completed with avendor-approved alternate material for the studs.

2418S/0065K

NeC f0AM DeeeIDIID>

IIIIC fww &4A(043 I UCENSEE EVENT REPORT (LER) TEXT CONTINUATION

US. IIUCLfAIIllfOULATOIIYCOMMISSIOII

AtfAOVfDOMS IIO 3150&10lfXhhfS ftfl(N

fACILITYIIAMfill DOCKfT IWNNfllQl LfllIIIANfII IQ tAOf Ill

DIABLO CANYON UNIT 1TSCT IK~~ e~ ~~ IYIICForm ~'ill Ill 06000323

YfAII 'Y

8 8

fIOUINTIALNUM IA

0 4

IYNIORgQIJ I0

0 0 3 oF 09

.'//On October 15, 1988, RT of an assembled check valve with a.flawed stud(installed for test purposes) demonstrated that a cracked stud can be seenusing a miniature linear accelerator as an RT source.

On October 17, 1988, the failure mechanism for the studs in valveRHR-2-8740A was confirmed by a materials testing, laboratory to be IGSCC.

On October 19, 1988, RT of Unit 1 valves RHR-1-8740A L B utilizing theminiature linear accelerator was completed. A computer-enhancedradiograph of valve RHR-1-8740B showed a possible crack in one stud closeto the retaining block to valve body mating surface.

B. Inoperable structures, components or systems that contributed to the event:

None.

C. Dates and approximate times for ma)or occurrences:

October 9, 1988 at 0430 PDT: Broken studs were found in RHR-2-8740A.

October 10, 1988 at 0200 PDT: UT of RHR-2-87408 showed nodiscontinuities.

October 10, 1988 at 0900 PDT:

October 13, 1988 at 0300'PDT:

October 15, 1988 at 1800 PDT:

Other similar Unit 2 valves opened forinspection.

RT of RHR-1-8740A 5 B comnences. HT ofUnit 2 studs showed no cracking.

Hockup testing showed RT with a linearaccelerator can detect a broken stud.

October 17, 1988 at 1600 PDT: Failure mechanism confirmed to be IGSCC.

October 19, 1988 at 1400 PDT: A possible crack was identified inRHR-1-8740B retaining block .stud.

D. Other systems or secondary functions affected:

None.

E. Method of discovery:

During a routine outage inspection, the maintenance crew observed unusualmotion in the retaining block of check valve RHR-2-8740A after a fewmanual valve cycles.

2418S/0065K

MAC IDIOM MS@I041I

NRC Fern bbbAIQ4ll ~ UCENSEE EVENT REPORT (LER) TEXT CONTINUATION

Ub NUCLEAR REOULATORY COFbUIbbION

AFFROVED OMb NO bISOW104

EXFIREb, blbl/%

FACILITYNAME I'l

DIABLO CANYON UNIT 1TEXT NF~~ N~ ~ FFFNNNF NIICfam ~'F I lltl o s o o o 3 2 3

LER NIRbbER I~ I

SEOVEHTiAL~IVU

8 8 —0 4

IVSK>4UVU IR

0 0 4 OF 09

INPO SOER 86-03 identified and listed various past failures in checkvalves in many operating nuclear plants. As a result of the concernsidentified in SOER 86-03, EPRI undertook various studies and experimentswhich resulted in an EPRI application guideline for various check valvedesigns. Among the reasons for failure identified in the EPRI guidelinewere location in the piping system and the operating conditions to whichthey are sub)ected.

On the basis of this guideline, PGLE Nuclear Engineering and ConstructionServices (NECS) reviewed all 3-inch and larger safety-related check valvesplus certain Hain Steam and Feedwater check valves in Unit 2 and issued areport to Plant Haintenance requ'esting an inspection of 26 check valves inUnit 2 in accordance with HP H-51.14. Valve RHR-2-8740A was selected aspart of this inspection sample.- The scope of the inspection was to openthe selected 'valves and visually inspect for any broken items, excessivewear, proper alignment, and security of retaining devices.

Check valve RHR-2-8740A is in the RHR recirculation line to the RCS hotleg. In accordance with the inspection program, the cover of this valvewas removed and a visual inspection was made of the internalconfiguration.

During the initial inspection of the valve internals, no abnormalitieswere noted. The disc was rotated to check if it was free and nothingunusual was noted at that time. As the inspection progressed the disc armwas manually operated by swinging it to observe if any binding existed.During this phase of the inspection the mechanic noted unusual play in theretaining block studs and the retaining block. At this point a closerinspection of the retaining block and studs was performed.

The valve internals were again cycled by hand with emphasis placed oninspection of the retaining block and block studs reaction to valvemanipulation. The block was loose and the studs showed signs ofmovement. Upon examination, the studs were found to be broken; the leftstud was sheared in 'the location of the block to valve body connection,and the right stud was broken off inside the retaining block assembly withapproximately 1-9/16 inches protruding from the valve body.

The only abnormalities noted were in the 10 to 12 o'lock position (facingthe valve seat from downstream) outside the in-body seat area, andconsisted of wear marks 3/8 inches wide and estimated to be 1/32 to 1/16of an inch deep (see drawing: Attachment 1). Precise measurements werenot possible due to radiological clothing interference and difficultaccessibility. The wear marks could have been caused by disc-to-bodycontact. The mechanic noted that the retaining block studs hadsignificant corrosion product buildup.

2418S/0065K

HRC fORU bbbklb@i ~

1

NIIC fee 44SA1443 l LlCENSEE EVENT REPORT (LER) TEXT CONTINUATIQN

US. NUCI EAII IIEOULATOIIYCOMMISSION

AffllOVEOOM4 NO SISOWIOSEXfIIIES"4I$1I44

~ fACILITYNAME lll

DIABLO CANYON UNIT 1TEXT IP awe ~ N~. ~ OAINNNINIICArm~ Tl 11TI

OOCKET NUh&EII Ql

o s o o o 3 2 3 8 8

LEII NLNNEII IllSSOVIHTIAL

WVU

01 4

QVNIOHUVU Ii

tAOE ISI

05 oF 09

F. Operator actions:

None.

G. Safety system responses:

None.

f v nA. Imnediate cause:

B.

IGSCC caused reta1ning block studs to fail 1n service.

Root cause:

The manufacturer's 1ncorrect heat treatment of the check valve reta1ningblock studs caused susceptibility to IGSCC. Th1s was determined bymaterial analysis, hardness testing, and microscopic section examinationof the studs in a materials laboratory.

IV. Analysis of Event

A. Computer Assisted Drafting Valve Geometry Analysis:

The Anchor Darling swing check valves 1nvolved in this evaluat1on aredesigned with very close tolerances. This design provides a valve which1s hydraulically similar to a straight piece of pipe. A large diameterbonnet is placed over the body to house the disc when 1t is lifted by theflow. The disc diameter 1s the same as the pipe OD, while the seat ID isequal to the p1pe ID. The valve body is. enlarged only enough to allow thedisc to swing down into the flow and cover the seat.

The effects of degraded hardware were evaluated by use of a computerassisted drafting (CAD) model developed from the manufacturer's orig1nalshop fabricat1on draw1ngs. The resulting model shows that s1nce the studnuts are tack welded to the retain1ng block, there 1s a low probabilitythat loose parts could exist in the valve body which could potentiallyaffect other system components. Study of this model showed that shouldonly one stud fall, the operation of the valve w1ll be unaffected sincethe two guide pins in each retaining block will prevent any movement ormisalignment.

2418S/0065K

HIIC fOIIU SSSAt444 I

IIAC fera SSAAI04$ l UCENSEE EVENT REPORT (LER) TEXT CONTINUATION

US, IIUCLSAII IISOULATOIIYCOMMISSION

AffAOVKOOMS IIO. SI$0-0104CXflllfS; SISI/

fACILITYNAMflll LSII IIUMSEII IlrWOUIrrTr*L „LYrflQrr

~rUV I rr rrVv

DIABLO CANYON UNIT lmnrreaaMeeee~~~frrrCa ~Lrlln

o so o o 32 388 —0 4 06 o~ 09

B.

The CAD model was developed to evaluate the consequence of the worst casefailure of both studs holding the retaining blocks in place. The resultsshow that the most likely effect would be no detectable change 1n theperformance of the valve. This results from the close fabricationtolerances which create a guide to d1rect the disc to travel in its normaldesign path. The four gu1de pins prevent lateral movement. The discswing arm geometry provides for little uplift force until the disk is well'p

out of the flow path. The guide design combined with the weight of thedisc is sufficient to ma1nta1n the location of the assembly. The physicalevidence confirms this evaluation since the disc assembly 1n RHR-2-8740Aremained 1n its des1gn location and apparently functioned properly eventhough both retain1ng block studs failed.

The worst case failure of the valve would be the disc rotating normallyfrom the bottom out of the flow stream. Detachment, 1f postulated, wouldprobably occur at two-thirds of full disk rotation up out of the flowstream. The model .shows that the swing arm would prevent the top of thedisc from rotating back into the flow and the bottom could not movedownstream without .becoming wedged in the valve body up out of the flowstream. No significant flow restriction would result, although the valvemay not reseat.

An evaluation of the flow effects was performed on the RHR Hot LegRecirculation line. Hith one of the RHR check valves postulated to beblocked and the other valve with only a 2(C free flow area, the flowreduction is only 10 to 15'L of normal flow. The RHR flow rate is ad)ustedthrough the Flow Control Valves (FCV). Hith increased fr1ction, due tovalve blockage in the system, the FCV's would open wider to allow moreflow to compensate for the friction increase. The actual flow reductionwould be insignificant. Th1s shows that the reduced transient conditionand the full flow test conditions are comparable, and that the full flowtests are 1ndicative of valve functionality for the reduced transient.

Hydraul1c Analys1s

Even though it has been demonstrated that the valves remain operable w1thbroken studs, it is worth noting that the LOCA analysis 1s suitablyconservative to accomodate some degree of flow blockage (although that isnot postulated in this case).

Based on the existing margins to the peak clad temperature lim1t of10 CFR 50.46 that ex1st in the FSAR Update, Hestinghouse )udges that if asmall break LOCA analyses were performed using the best estimatetechn1que, for 11nes s1x 1nches 1n d1ameter and smaller, with the

2418S/0065K~r+C SOrlv 0deeIII43r

llllC $«e 3QAI04) I UCENSEE EVENT REPORT tLER) TEXT CONTINUATION

US IIUCLTAAALOULATOIIYCOMMOSIOII

AffAOYEOOMa IIO i150&I04fXhAES- StÃl

fACILITYIIAMflll

DIABLO CANYON UNIT lmar W««f~e~~~alCa ~fIIIT>

osooo323

LCII IIIANTA IIITTQlllicT

NNC Fmss 399Al943 I

FACILITYNAME I'l

UCENSEE EVENT REPORT {LER) TEXT CONTINUATIONUS, NUCLEAII hEOULATONY„COMMIEEION

AFFAOYEO OME hO 9IEOWID4B(FIhES 9/91I

DIABLO CANYON UNIT 1TEXT Idaeae ~ e aywrK ~ eJAenelNIIC ferm ~EI IITI o s o o o 3 2 3 8 8

'ICOUChT(ALhVU IA

0 4

EVamhhVU fi

0 0 8 0 9

V.

Based on the analysts of this event, operat1on of DCPP Un1ts 1 and 2 didnot and does not now create an unreviewed safety question and will notadversely affect the health and safety of the public.

A. Immed1ate Corrective Action:

A search was made to locate all valves having the potential for a s1milarstud failure. Ten Anchor Darling valves of this model were 1dentified ineach Unit; 8740A L B in the RHR hot leg in)ection, 8956A through D (BP)(V) for the accumulators, and 8948A through D (BP) (V) for the SI to thecold leg injection. In addition to the inspection performed onRHR-2-8740A, the other nine Un1t 2 Anchor Darling check valves were also1nspected. The studs of the other RHR valve 8740B, the 8948A through D,and the 8956A through D SI valve studs showed no cracking in MTinspection. No wear marks or disc binding were observed in these valves.

A review of maintenance records for Anchor Darl1ng check valves showedthat 1n November 1984 an 1nspection of valve SI-2-89488 following a failedleak check test revealed one of the two retain1ng block studs was broken.The cause of this failure was determined to be IGSCC.

Microscopic examinat1on of a 410 SS stud from valve RHR-2-8740B, aftersectioning and polishing, showed no cracking present. Hardness testing ofthis stud 1ndicated it had correct heat treatment. The hardness ofSI-2-8948A and SI-2-8956A studs indicated they were susceptible to IGSCC,but microscopic examination showed no cracking had occurred in three yearsof service in borated water.

B. Corrective Action to Prevent Recurrence:

VI.

All Unit 2 Anchor Darling check valve retaining block studs have beenreplaced with studs made of vendor recomended ASTM A564 type 630-1100material.

All Unit 1 Anchor Darling check valve retaining block studs w111 bereplaced w1th studs made of vendor receimended ASTM A564 type 630-1100material before the end of the next refueling outage.

A. Failed components:

Check valve RHR-2-8740A, an Anchor Darling S350H 8-inch swing check valve.

2418S/0065K

HWC FOAM 9FIAI949 I

NIIC fere WAUCENSEE EVENT REPORT (LER) TEXT CONTINUATION

US NUCLEAN IIEQULATOIIYCOMMISSION

AffAOVEOOMb NO 31EO&IOSEXflllES'l S I /%

fACILITYNAME III OOCXET HIANEIIIEI LEA NEATEN ISISSOVSNTrAL

NVMSVrtrVrr

NVM te

~AOE Ii}

DIABLO CANYON UNIT 1TEXT II~~ N eetreerE ~~ lfrtC ferrrr ~111111 0 5 0 0 0 3 2 3 8 8 01 4 0 0 9 OF 09

I/-B. Previous LERs on similar events:

None.

C. Similar designs by other manufacturers:

Nineteen Velan check valves were identified (ll tn Unit 1 and 8 in Unit 2)that use 410 SS material for the retaining block studs. These valves arein the Auxiliary Feedwater (AFH) System and Hain Steam (HS) to the turbinedriven AFH pump. Records show that nine of these valves have been

uninspected

or replaced since 1985. The HS valves are full flow tested on a'onthly basis and the valves in the AFH system are tested at each cold

shutdown'. Based on the testing results and on the difference in chemistryfor these valve applications, these valves are not considered to have the .same stress corrosion cracking failure mechanism potential as the ECCScheck valves. There are no entries in NPRDS for retaining block studbreakage for Velan check valves and the manufacturer was unaware of anyfailures of this type.

D. Related documents:

SOER 86-3, issued by INPO on October 15, 1986, provides recomendationsfor a check valve preventative maintenance program which are beingimplemented in a change to HP H-51.14.

2418S/0065K

rrrlC r OIIM SSSAIt@Sr

ATTACHMENT 1

CENTER PONT OF DISK

DISK TO BODYCONTACT POINT

hFROX. AREA OF DAMAGEI ILJ I

ILJ IN BODY SEAT

3/S

SECTION A-A

SCALQ 3/5 a 1 ANCHOR DARLING CHECK VALVE

~ ~

Pacific Gas and Electric Company 77 Beate StreetSan Francisco, CA 94106

4151972.7000

TWX 910 372 6587

James D. Shilfer

Vice President

Nuclear Power Generation

November 18, 1988

PGLE Letter No. DCL-88-281

U.S. Nuclear Regulatory Comm)ssionAttn: Document Control DeskHashington, D.C. 20555

Re: Docket No. 50-323, OL-DPR-82 Docket No. 50-275, OL-DPR-80Diablo Canyon Units 1 and 2

. Licensee Event Report 2-88-014-00 — VoluntaryAnchor Darling Check Valve Retaining Block Stud Breakage Due ToIntergranular Stress Corrosion Cracking

Gentlemen:

PGtEE is submitting the enclosed voluntary Licensee Event Report(LER) regarding Anchor Darling check valve retaining block studbreakage due to intergranular stress corrosion cracking. Thisreport is being submitted for information purposes only, asdescribed in item 19 of Supplement Number 1 to NUREG 1022.

This event has in no way affected the public's health and safety.

Kindly acknowledge receipt of this material on the enclosed copy ofthis letter and return it in the enclosed addressed envelope.Sincerely,

3. D. Sh'rcc: J. B. Hartin

H. H. HendoncaP. P. NarbutB. NortonH. RoodB. H. VoglerCPUCDiablo DistributionINPO

Enclosure

DC2-88-HM-N111

2418S/0065K/DHO/2193