Text

Forwards Evaluation of Licensee Proposed Justification for Continued Operation of Plant Due to Broken Studs in Anchor/ Darling Model S350W,Valve RHR-2-8740A.NRC Concurs W/Util Conclusion That Unit Can Operate Safely W/O Insps
	ML20011C479
Person / Time
Site:	Diablo Canyon
Issue date:	12/20/1988
From:	Cheng C; Office of Nuclear Reactor Regulation
To:	Knighton G; Office of Nuclear Reactor Regulation
Shared Package
ML20011C480	List: ML20011C479; ML20011C609;
References
	IEIN-88-085, TAC-69826, NUDOCS 8812270132
	Download: ML20011C479 (5)
	v • d • e

j3[o*at i

bh h

. UNITED STATES

.i 1

WCLE AM GEGULATORY CO*AMISSION l

3

('

W ASWA"; ^t 0N, D. C. P0%6

.f is, ' " *

DEC 3 e 1998 l

)

MEMOMNDult FOR:

George W. Khighton, Director

)

Project Directorate Y DivC lon of Reactor Projects - lil/lV/V and Special Projects

')

FROM:

C. Y. Cheng, Chief Materiels Engineering Branch Division of Engineering and Systems Technology

SUBJECT:

EVALUATION OF JUSTIFICATION FOR CONTINUED OPERATION OF DIABLO CANY0ti UNIT l'(TAC NO. 69826).

The enclosed evaluation was prepared by the Materials Engineering Branch (ENTB) to provide the results of the staff's review of the Justification for Continued 7

Operation (JCO) dated October 19, 1988, and revised JC0 cated October 20, 1980, submitted by Pacific Gas and Electric (PGME), the licensee for Diablo Canyon Power Plant (DCPP), Unit 1.

The JC0 was necessary, because during a surveillance activity associated with the current DCPP Unit 2 refueling outage, the licensee found broken studs in' Anchor / Darling, Model 5350W check valve No. RHR-2-874CA.

l The hanger bracket studs were of type 410 stainless steel material and the cause of failure was determined to be stress corrosion. When the condition of valve RHR-2-8740A was found, an Event Response Plan was initiated and actions were begun immedictely by the licensee. Also the licensee performed an.10 CFR 50.59 l

evaluation as part of the JC0 and the staff has sumarized it in the enclosed i

ENTE evaluation.

('

The licensee's effort showed that the valve population of concern was ten Anchor / Darling check valves in each of the two DCPP Units. The remainder of the ten identified DCPP Unit 2 Anchor / Darling check valves were inspected by the licensee and no cracks were found.

In addition the licensee replaced all studs. in the ten DCPP Unit 2 Ar.cbor/ Darling check valves. Thus, the licensee I

proposes operation of DCPP Unit I until the next refuelin outage currently scheduled for October 1989, at which time the licensee wi$1 inspect and replace all the studs in the ten identified Anchor / Darling cimck valves.

In addition, if an unscheduled outage occurs for DCPP Unit 1, before the October 1989 refueling outage, the licensee has committed to inspect and replace the studs for valves RHR-1-8740A & B.

If tirae permits the licensee has stated it will inspect valves the SI-1-8956A B, C & D valves. The staff believes that every I

effort shculd be made by the licensee to complete this work if a forced outage s

occurs.

L

~ The staff has already issued Inforniation hotice 80-85 on this subject. A l

Bulletin is currently being prepared by the f;RC staff and any actions so speci-fled in the Culletin would supersede these proposed by the licensee in the JCO.

C. Y. Cheng, Chief lp flaterials Engineering Branch l

Division of Engineering ano Systens Technology CONTACT:

C.D. Sellers X20930 7

f L..'

(9Wa2:4/JW xA l

s-(g e

o

,.o,-

L D George W. Knighton c, 4

M, s

c

Enclosure:

es stated cc:

L. Shao' L

G. Holahan E

J. Richardson R. Herinen I

H. Rood.

F T. Miller T. McLellan C. Sellers c

ll 1

? -

5 I'

s h-

\\

i f ' '. r -

i

)

l' b.

k[ -l "'

ji.

j 1

c PACIFIC GAS AllD ELECTRIC DIABLO CANYON UNIT 1 EVALUATION OF LICENSEE'S PROPOSED JUSilf! CAT!0ft FOR C0hTINUED OPERATION Of UNIT 1 1.

LICENSEE BASIS FOR CONT!tlVED OPERATION Letters proposing Justification for Continued Operation (JCO) dated October 19, 1966 and a revised JC0 dated October 20, 1988 were submitted by Pacific Gcs and Electric. (PGAE), the licensee for Diablo Canyon Power Plant (DCPP), Unit 1.

The JC0 war pruposed because during a surveillance activity associated with the i

current DCPP, Unit 2 refueling outage, the licensee had found broken hanger bracket studs in an Anchor / Darling, model S350W check valve RHR-2-8740A.

The hanger bracket studs were of ASTil A193 B6 (A151 Type 410 stainless steel) and the l

cause of the failure was determined to be stress corrosion.

When the conditior, of valve RHR-2-8740A was found, Event Response Plan 88-008 was initiated and actions were begun immediately by the licensee. These actions encompassed a full area of investigation to determine the potential extent of the noted condition and to assess the impact of such stud failures on DCPP Unit I and 2 plant safety and operation. Key activities incluoed:

record review to establish the popula-tion of potentially affected valves and their history; possible impact on vahe performance; review of the in. pact on valve operation and on required safety function; valve internals physical geometry studies using CAD (computer assisted drafting) techniques; and a program to develop a method of radiographic non.

destructive examination.

LICEllSEE ANAL YS!S The licensee's results from these efforts showed that the valve population of concern was ten Anchor / Darling check valves in each Unit.

There were two eight inch valves 6740A and B in the RHR line to hot leg recircul6 tion and eight ten inch valves (874BA-D) in the cold leg injection lines from the accumulators. All valves were within the containment building. These valves had complete material records and test / maintenance records which showed consistent reliable performance including valve RHR-2-8740A.

The failure mechanism appeared to have occurred early in plant life because valve SI-2-89488 was found to have had a cracked stud in 1984 prior to plant operation. The crack surface had corrosion product buildup.

So it was con-cluded the failure occurred during earlier layup periods.

Valve RHR-2-8740A had corrosion products on the fracture surfaces of the failed studs and evidence of side root cracking from the fracture surface into the stud body.

A red contaminant (most likely rust) was found deposited in RHR-2-8740A.

This would have come from poor layup conditions not operational chemistry.

No failed studs were found in SI-2-8956A-D. These valves have seen over three years of borated water service.

The licensee also indicated that any failed studs on Unit I nest likely failed early on in startu) flushing, testing, and layup periods, their surveillance test program has ciallenged these failed studs multiple times with no double s

}

7 I

' j stud failures detected. Also, if a contaminant is the cause, they do not expect i

J operational water chemistry to accelerate SCC.

However, a computer enhanced radiographic analysis of two valves in what was j

eensidered to be the worst environn.ent of the 10 valves, RHR-1-8740AAB, showed a possible indication of a crack in one of the studs.

l The 1_icenseo proposed the JC0 until the next refueling outage, scheduled for October 1989, and has proposed to inspect and replace all the studs in the ten DCPP, Unit 1 Anchor / Darling Valves.

In addition if an unscheduled outage for DCPP, Unit 1 occurs before the October 1989 refueling outage; the licensee will i

N p et and replace the studs for valves RHR-1-8740A & B, and if time permits the licensee will do the same for the SI-1-8956A, B, C & D valves. A aulletin is currently being prepared to provide a solution to the problem and will super-sede the JCO.

The licensee has performed a 10 CFR 50.59 evaluation as part of the submitted JC0 which is summarized below:

1.

The licensee claims that the potential effects of degraded valve disc retaining block studs have been ev61uated under an Event Response Plan.

The capability of the valves were evaluated for normal plant operation and for required function in an accident mode.

The valves are passive compo-nents and the condition of the retaining block studs has no impact on i

their ability to remain seated and retain pressure as designed for normal i

plant operation, as evidenced by surveillance test program results. Also the condition of the retaining block studs does not affect the pressure boundary or the ability of the valves to open when required in response to e

an accident. CAD studies have shown that there will be no loose parts outside of the valve casing to affect other equipments. The valves there-fore will not increase the probability of occurrence or the consequence of an accident or malfunction of equipment important to safety previously analyzed in the safety analysis report.

2.

The licensee further claims that the valves are in the closed position during normal plant power operation (Modes 1, 2 and 3). The valves will provide proper function in the closed mode as documented through test.

This function of providing inter-systeni pressure protection and inter-systen. LOCA protection is not impacted as the condition of the studs was shown to have no bearing on the closed function of the valve.

There are no other functions of the valves associated with normal plant at power cperation, consequently the studs de not have an effect on continued normal safe operation of the plant.

Likewise continued plant operation a

with the valves providing their normally closed function does not increase the possibility for an accident or malfunction of a different type than any evaluated previously in the safety analysis report.

3.

The licensee also stated that the valves are expected to open when re-quired to fulfill their initial safeguards function, as demonstrated by periodic testing. The licensee admits to the possible failure to resent

-,e

--a

n F..

c, g,

-3.

j i

after initial opening. Analysis shows that renating is the ir.ost likely result, which is supported by the test records.

However the potential consequences of failurt to rescat have been evaluated and found to be acceptable. A situation could bc postulated where the valve disc / seat orientation is lost.

In the worst case, a complete orientation loss leading to a loose valve disc was postulated and evaluated. The evalu-ation addresses the anticipated effects on the performance of the ECCS and the impact on the margin of safety.

In addition, the Probabilistic' Risk Assessment (FRA) calculates a negligible change (using quite conser-vative assumptions) in risk factor by allowing continued operation of DCPP l

Unit 1.,

The results of the analysis, records evaluation, inspections of j

Unit 2 valves, and Unit I radiographs provide confidence of the operability l

of the Unit I valves. The continued operation of the plant does not reduce J

l the margin of safety as defined in the basis for any Technical Specification.

111 STAFF REVIEW l-Based on the following factors:

1)

The licensee's evaluation;-

)

i 2)

The absence of any evidence of check valve malfunction with failure of the j

hanger bracket studs 3)

The probability of similarity of conditions of other plants for which there is currently no requirement for action relative to similar valves, i

The staff concurs with the licensee's conclusion that DCPP Unit I can operate L

safely and not create an unreviewed safety question without inspecting valves 1

RHR-1-8740A&B, SI-1-8940A, B C. & D, and SI-1-8956A, B, C, & D until the next refueling outage scheduled for October 19E'.

In addition, continued operation s

will not adversely affect the health and safety of the public, i

I L

1

'g

i use:sas evsen nspott asni 122334 1

ll

..i

.w caNfon Unn i eisie>eo). n

, i o.,,

y.i

. ami.

igg, gggigg enam eacx svuo==st = 70 m = uw

i....

... e

.. m.m_ - -

e m c = on unn i e, s, e, e, em ii. on....

..iiid one iii in sie sisitie,ei i,

..._-...s...

Y 01010 D e== -

NEN L"Jg*,pgsggiggg,,

,. p,, (..

' *n~.

.m

.e

. IhMES C. ROLAN ttalLATORY CDIPLIANCE ENGIN!(4

m. I, 5Ffr77PD l

IP

,i l

e

ty.t'

t:5.*tt'

?.T

l.==

eggf,ge

.gy;;,9;i 3 9 lj _"!$

)

i 74}AA t1 Ojt 3j i

jij jjj 3

3 i

i_ aaa 33a L

i aii ii i se.... e...... e....

.r...

y:C.%.

n..

i op e a, l

This report is belns voluntarily subattted for information purposes only as described in 1 tem 19 of Supplement Number 1 to NUREG 1022.

On October 9,1988, at 0430 pDT during a preventative estatenance (PN) inspection.

both retaining block studs for check valve RNR-2-87404 were observed to be broken.

The four disc are alignment guide pins were intact. Manual valve cyc1tne during PN l

showed no $10nificant misalignment or disc-to-body contact. This valve had passed previous surveillance testing and was considered operable prior to disassembly.

i The failed studs are made of ASTM A193 B6 Tspe 410 $$. Es6mination of the studs showed intergranular stress torreston trackung as the failure mechanism. Steller valves in Unut I were disassembled and the studs were removed for inspection.

Nondestructive esamination of the studs from the other nine Unit t valvcs showed no l

abnormalities. Studs in all sin 11er Unit # valves were rap 1tted with vendor receanended AlfM A564 type $30-1100 anterla1.

l Radiographic testing of the two accessible Unit 1 RNR theck valves confirmed there were no broken retaining block studs. tiovever. computer-enhanced radiography showed a possible crack in one of the studs. The studs in sistlar Unit 1 valves will be replaced by the end of the nest Unit I refueling outage.

2418S/0065K 78/2/s. W gp 12 CE.et 12:26 FM PC3

~. ~

4

+

sp.

n near.mo w e,e

' UCENSEE EVENT CEPORT ILERI TEXT CONYINUATION waovee en =o sim w q

n e ei.

i

..w.

. una.. e, o.

aret.

'at' DIA8LO CANYON UNIT 1 e is t e le lej 312 l 3 818 O ll 14

) 10 0l2 or 019 wi I,

' Initial canditiant

+

j Unit l'was in Mode 1 (Power Operation) and Unit 2 was in a refueling outage l

with all fuel removed from the reactor vessel when the event occurred. Both Units have been operating at various modes and power levels with these broken.

j check valve retaining block studs.

!!. Deterintion of Event i

A.

Event:

Valve RHR-2-8740A (BP) (V) was chosen for internal inspection in accordance with the preventative maintenance program administered by Maintenance Procedure (MP) N-51.14, " Check Valve Maintenance Program."

This valve is located in the Residual Heat Removal (RNR) hot leg injection l

lit.e immediately adjacent to a piping elbow. This valve is an Anchor Darling model 5350N 3-inch swing check valve.

On October 9,1988, during a Unit 2 refueling outage valve inspection. two l

broken retaining block studs were found in check valve RHR-2-8740A. The broken studs were four:1 after the valve had been manually cycled through its travel are twice with no apparent problems. One stud was severed flush with the velve body while the other stud had a stub extending about J

1-9/16 inches into the retaining block.

Four alignment guide pins were observed to be in place and intact in the retaining block to valve body e

mating surface. The studs were made of ASTM A193 56 type 410 stainless steel (SS).

Each Unit has a total of 10 valves of this design installed. All of the l

Unit 2 valves were disassembled and those Unit L valves that are accessible were examined by radiography (RT).

l On October 10, 1988, ultrasonic testing (UT) of the check valve studs from valve RHR-2-87408 (8P) (V) showed no discontinuities. Disassembly of the remaining Unit 2 suspect check valves was initiated. During document review, a previous occurrence of a failed retaining block stud in a : heck valve of this design due to intergranular stress corrosion cracking (IGSCC) was identified in 1984 during Unit 2 startup testing.

r On Octroer 13, 1988 RT of Unit 1 ulvec RHR-1-8740A & 8 connenced utiliz'.ng a Cobalt 60 source. The studs from the other nine Unit 2 check valves were magnetic particle fluorescent dye tested (NT) and showed no cracking. Reassembly of the Unit 2 check valves was completed with a vendor-approved alternate material for the studs.

k.

2418S/0065K

=.

~.

1 i

.m.. e, UCENSEE EVENT CEPORT ILERI TEXT CDNTINUATCN apaevio e= =o sies =

ems eve

..w, one.n ausia e, 6e.

e.,

' e m.ss mn 0 $ 14 DIA8LO CANYON UNIT 1 o Is le f e le l31213 8 l8

)10 0l3 of 019 von v e a==,,.,.

j V

l l

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 seen using a miniature linear accelerator as an RT source.

1 J

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

J On October 19, 1988 RT of Unit I valves RHR-1-8740A & 8 utilizing the miniature )inear accelerator was completed. A coayuter-enhanced

]

radiograph of valve RHR-1-87408 showed a possible crack in one stud close i

to the retaining block to valve body mating surface.

j 8.

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

None.

C.

Dates and approximate times for major occurrences:

i October 9,1988 at 0430 PDT:

Broken studs were found in RHR-2-8740A.

1 October 10,1988 at 0200 PDT:

UT of RHR-2-87408 showed no discontinuities.

October 10, 1988 at 0900 PDT:

Other similar Unit 2 valves opened for i

inspection.

October 13, 1988 at 0300 PDT:

RT of RW-1-8740A & 8 commences. MT of Unit 2 studs showed no cracking.

October 15, 1988 at 1800 PDT:

Mockup testing showed RT with a linear accelerator can detect a broken stud.

l October 17, 1988 at 1600 PDT:

Failure mechanism confirmed to be IGSCC.

October 19,1988 at 1400 PDT:

A possible crack was identified in RHR-1-87408 retaining block stud.

D.

Other systems or secondary functions affected:

None.

E.

Method of discovery:

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

24185/0065K g,.

_w._

%,/

i es m m m n e s. em E'

ap UCEN$tt EVENT REPORT (LER) TEXT CONTINUAT60N a==oveo e

.o sim w enus eve t

..m.

m n.

ese enamin e.

u s a r,n

=

0 10 0 14 o' Olg 0 p 14 DIA8LO CANYON UNIT 1 o 16 le l* l e l 31213 818 un n

.an

.ar a am.. m i

INP0 $0ER 86-03 identified and listed various past failures in check i

valves in many operating nuclear plants. As a result of the concerns t

identified in 50ER 86-03, EPRI undertook various studies and experiments which resulted in an EPRI application guideline for various check valve l

designs. Among the reasons for failure identified in the EPRI guideline j

were location in the piping systes and the operating conditions to which they are subjected.

On the basis of this guideline, PG&t Nuclear Engineering and Construction f

Services (NECS) reviewed all 3-inch and larger safety-related check valves plus certain Main Stea.a and Feedwater check valves in Unit 2 and issued a report to Plant Maintenance requesting an inspection of 26 check valves in

+

Unit 2 in accordance with MP N-51.14. Valve RHR-2-8740A was selected as part of this inspection sample. The scope of the inspection was to open the selected valves and visually inspect for any broken items, excessive wear, proper alignment, and security of retaining devices.

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

i During the initial inspection of the valve internals, no abnormalities were noted. The disc was rotated to check if it was free and nothing i

unusual was noted at that time. As the inspection progressed the disc arm was manually operated by swinging it to observe if any binding existed.

i During this phase of the inspection the mechanic noted unusual play in the i-retaining block studs and the retaining block. At this point a closer inspection of the retaining block and studs was performed.

The valve internals were again cycled by hand with emphasis placed on inspection of the retaining block and block studs reaction to valve manipulation. The block was loose and the studs showed signs of movement. Upon examination, the studs were found to be broken; the left stud was sheared in the location of the block to valve body connection, and the right stud was broken off inside the retaining block assembly with approximately 1-g/16 inches protruding from the valve body.

The only abnormalities noted were in the 10 to 12 o' clock position (facing the valve seat from downstream) outside the in-body seat area, and consisted of wear marks 3/8 inches wide and estimated to be 1/32 to 1/16

)

of an inch deep (see drawing: Attachment 1).

Precise measurements were I

not possible due to radiological clothing interference and difficult accessibility. The wear marks could have been caused by disc-to-body contact. The mechanic noted that the retaining block studs had l

significant corrosion product buildup.

2418S/0065K g,o.. mu

r:

.i 3

J e

l en een-es ens a n no: n n en,sa m e UCENSEE EVENT REPORT ILER) TEXT CONTINWAftDN menevee ew no e m.eim.

l enmis eve l

j

.. m mass n.

ses n anne.. ei j

.. y......

I' 01114 DIABLO CANYON UNIT 1 e is te le le l 31213 8 18

) 10 015 o' 01 9 on n e

se.., =>

j L

i F.

Operator actions:

None.

I G.

Safety system responses:

~

None, d

t III. Cause of Event A.

Immediate cause:

j IGSCC caused. retaining block studs to fail in service.

8..

Root cause:

The manufacturer's incorrect heat treatment of the check valve retaining block studs caused susceptibility to IGSCC. This was determined by material analysis, hardness testing, and microscopic section examination of the studs in.a materials laboratory.

IV.. Analysis of Event-

]

A.

Computer Assisted Drafting Valve Geometry Analysis:

4 The Anchor Darling swing check valves involved in this evaluation are designed with very close tolerances.. This design provides a valve which.

t is hydraulically similar to a straight piece of pipe. A large diameter bonnet is placed over the body to house the disc when it is lifted by the flow. The disc diameter is the same as the pipe 00, while the seat ID is:

equal to the pipe ID. The valve body is enlarged only enough to allow the disc to swing down into the flow and cover the seat.

l I

The effects of degraded hardware were evaluated by use of a computer.

assisted drafting (CAD) model developed from the manufacturer's original shop fabrication drawings. The resulting model shows that since the stud nuts are tack welded to the retaining block, there is a low probability that loose parts could exist in the valve body which could potentially affect other system components. Study of this model showed that should only one stud fail, the o>eration of the valve will be unaffected since the two guide pins in eac) retaining block will prevent any movement or misallgnment.

)

24185/0065K g...

en w so.

v2 mu.. s, UCENSEE EVENT CEPORT 6LER) TEXT CENT 6NUAT60N '

m w,e eno eim..

emais eve j

l-

...n

.. uu...

]

u.a

"!c.

'm?:

0114 H 10 016 o' 019 l

DIABLO CANYON UNIT 1 e 18 le le le 131213 818 mn.

n

.ar a mu.nm

)

The CAD model was developed to evaluate the consequence of the worst case failure of both studs holding the retaining blocks in place. The results show that the most likely effect would be no detectable change in the performance of the valve. This results from the close fabrication 1

tolerances which create a guide to direct the disc to travel in its normal design path. The four guide pins prevent lateral movement. The disc L

swing are geometry provides for little uplift force untti the disk is well up out of the flow path. The guide design combined with the weight of the disc is sufficient to maintain the location of the assembly. The physical evidence confirms this evaluation since the disc assembly in RHR-2-8740A q

l remained in its design location and apparently functioned properly even L

though both retahing block studs failed.

The worst case failure of the valve would be the disc rotating normally from the bottom out of the flow stream. Detachment, if postu ated, would probably occur at two-thirds of full disk rotation up out of the flow stream. The model shows that the swing arm would prevent the top of the l

disc from rotating back into the flow and the bottom could not move downstream without becoming wedged in the valve body up out of the flow r

stream. No significant flow restriction would result, although the valve j

may not reseat.

An evaluation of the flow effects was performed on the RHR Hot Leg Recirculation line. Mith one of the RHR check' valves postulated to be blocked and the other valve with only a 201 free flow area, the flow reduction is only 10 to 15% of normal flow. The RHR flow rate is adjusted through the Flow Control Valves (FCV). With increased friction, due to valve blockage in the system, the FCV's would open wider to allow more flow to compensate for the friction increase. The actual flow reduction would be insignificant. This shows that the reduced transient condition l-and the full flow test conditions are comparable, and that the full flow tests are indicative of valve functionality for the reduced transient.

8.

Hydraulic Analysis Even though it has been demonstrated that the valves remain operable with broken studs, it is worth noting that the LOCA analysis is suitably conservative to accommodate some degree of flow blockage (although that is not postulated in this case).

Based on the existing margins to the peak clad temperature limit of 10 CFR 50.46 that exist in the FSAR Update, Nestinghouse judges that if a small break LOCA analyses were performed using the best estimate technique, for lines six inches in diameter and smaller, with the 2418S/0065K g.

4 I

.ne se,.

va austsie as: wen, essa,eion UCENSEE EV8NT REPORT ILER) TEXT CONTINUATION an.noveo e=,e sim 1

= =.....

j

..6....

== n

6..

E

h!% f'

"J.!Lt:

DIABLO CANYON UNIT 1 o is le le le l 3l213 818 0 l1 l 4 D l0 0l 7 or 0] 9 assumption of a failure of one accumulator to discharge, the resulting peak clad temperatures would not exceed the limit. The assumption of no discharge from one accumulator is conservative as this results in discharge of only two accumulators into the RCS, since one accumulator is already assumed to be lost due to the initiating pipe break event.

Based on analysis performed for many plants and accepted by the NRC, leak-before-break (LBB) has been demonstrated for the reactor coolant loop piping and large branch lines attached to the loop (down to and including 8-inch lines). Therefore, the LOCA assumed for this evaluation is the rupture of a line six inches or smaller attached to the loop, although J

e evaluations show that even for smaller lines, the mode of failure would 1

still be LLB. For this case, the accumulators would still be required to

?

operate, but at a reduced flow, and with lower loads on the check valves in the discharge line.

The flow conditions which check valves 8948A-D and 8956A-D would experience during the above assumed best estimated LOCA case has been calculated by Westinghouse to be a peak flow of approximately 7.300 gpm.

The analysis of the hydraulic forces acting on the valve internals shows that the disc will lift to its full flow position with the flow used in the flow test (2000 gpm through 8948) and that additional flow does.not result in increased lifting of the valve disc. The disc floats above the flow. Consequently, the valve disc and are force balance during flow testing and are representative of valve forces during the maximum expected LOCA flows.

Based on an evaluation by Westinghouse, it is believed that the guide pins are sufficient to assure proper operation of the valve disc. An analysis of the hydraulic forces acting on the valve internals confirms the capability of these pins (two per block, four per valve) to withstand the Ioads imposed during opening of the valve and durinn full flow operation.

This demonstrates that the capability of the retainLnq block studs is not required to maintain the valve disc in the open posit'on. The disc assembly is not expected to lift off of the guide pins and proper seat / disc orientation is maintained. The worst failed condition of the studs has been assumed in the analysis, i.e. both retaining block studs were assumed to be broken at the retaining block / valve body interface.

Inspection data obtained to date indicate stud failure locations which vary from the block / body interface to points higher up inside the retaining block. A failure of the stud inside the retaining block leaves a stub portion of the stud which would assist in maintaining disc orientation and which could provide additional load carrying capacity.

241BS/0065K g...

i one see een wa.wettia metaiteet eenese.en l

UCENSEE EVENT CEPORT (LER) TEXT CONWNUADON an om en w ec o d

e==ws...

I

.m n..

sec.n w. e,

E!r."

"J.it:

1 0 ll l4 DIABLO CANYON UNIT 1 e 15 le le f e l 31213 8 18

> 10 018 o' 019 1

mnu anc an.v.m 4

Based on the analysis of this event, operation of DCPP Units 1 and 2 did 3

not and does not now create an unreviewed safety question and will not adversely affect the health and safety of the public.

V.

Corrective Actinna j

A.

Immediate Corrective Action:

i l

A search was made to locate all valves having the potential for a similar 1

stud failure. "en Anchor Darlini valves of this model were identified in each Unit: 8740h 1 B in tne unx act laa *=*a'tiaa aos6A througn v ten fui rar +ka mer"=iilatprs, and 8948A through D (BP) m far +ha u rhe i

wa cold leg injection.

In ad31~tton to tne inspection performed on

~

l.

RHR-2-8740A, the other nine Unit 2 Anchor Darling check valves were also inspected. The studs of the other RHR valve 87408, the 8948A through D, i

and the 8956A through D SI valve. studs showed no cracking in NT inspection. No wear marks or disc binding were observed in these valves.

A review of maintenance records for Anchor Darling check valves showed that in November 1984 an inspection of valve S!-2-89488 following a failed leak check test revealed one of the two retaining block studs was broken.

The cause of this failure was determined to be IGSCC.

t a

Microscopic examination of a 410 SS stud from valve RHR-2-87408, after i

sectioning and polishing, showed no cracking present. Hardness testing of this stud indicated it had correct heat treatment. The hardness of SI-2-8948A and SI-2-8956A studs indicated they were susceptible to IGSCC, but microscopic examination showed no cracking had occurred in three years L

of service in borated water.

l B.

Corrective Action to Prevent Recurrence

l' All Unit 2 Anchor Darling check valve retaining block studs have been replaced with studs made of vendor recommended ASTM A564 type 630-1100 material.

All Unit 1 Anchor Darling check valve retaining block studs will be replaced with studs made of vendor recommended ASTM A564 type 630-1100 material before the end of the next refueling outage.

VI. Additional Information A.

Failed components:

Check valve RHR-2-8740A, an Anchor Darling S350M 8-inch swing check valve.

24185/0065K 5*4.' *** **"

g '

v.

a m...

UCENSEE EVENT CEPORT H.ER) TEXT CONTINWA700N emieveo

,e ee.eim Smhil e'eie

.mm n.

see.n

. ei o...,

'ea*

h!.u."

'J.P.n i

0 $ l4 4 10 019 or 019 DIABLO CANYON UNIT 1 e is le j e le l 31213 88 1

som u-.

- sar

.,.u.

8.

Previous LERs on similar events:

None.

C.

Similar designs by other manufacturers:

Nineteen Velan check valves were identified (11 in Unit I and 8 in Unit 2) that use 410 SS material for the retaining block studs. These valves are in the Auxiliary Feedwater (AFN) System and Main Steam (MS) to the turbine driven AFW pump. Records show that nine of these valves have been inspected or replaced since 1985. The MS valves are full flow tested on a f

monthly basis and the valves in the AFW system are tested at each cold shutdown. Based on the testing results and on the difference in chemistry for these valve applications, these valves are not considered to have the same stress corrosion cracking failure mechanism potential as the ECCS check valves. There are no entries in NPRDS for retaining block stud breakage for Velan check valves and the manufacturer was unaware of any failures of this type.

D.

Dear Mr. Davis:

The purpose of this letter is to provide you with information concerning recently observed degradation of A 193 Grade B6 Type 410 stainless steel retaining block studs in Darling Valve and Manufacturing Company Clear Waterway check valves installed at the Cook Nuclear Plant. The observed condition did not result in any check valve failures, and we'have determined that the condition was not reportable under Title 10 CFR or our technical specifications (T/Ss). However, because degradation of the type observed at the Cook Nuclear. Plant has been of general industry interest in the past (e.g., INPO Significant Operating Experience Report [SOER) 86 03), we have elected to submit this voluntary report. A summary of the observed condition and actions we have taken is provided below.

===.

Background===

In conjunctio'n with the performance of other maintenance on 8" Darling Clear Waterway swing check valve (2 SI 151W) installed in the low pressure emergency core cooling system (ECCS), an inspection of the valve internals was performed in accordance with the maintenance program that we established in response to INPO SOER.86 03.

During this inspection, one of the two retaining block studs was found broken and the other cracked. A diagram of the valve type in question is provided in Figure 1.

The retaining block studs (Part No. 11542-61 5) retain the blocks (Part Nos. 11542 60/60 1) that hold the valve disc assembly in place. As a result of this finding, the corresponding check valve (2 SI 151E) in the redundant low pressure ECCS train was oI hn I yiMd2 8268l[p s Pg.

T

L:

i J

i e.

2-AEP:NRC:1054

)

Mr. A. B. Davis

)

inspected. Again, one of the two retaining block studs was found broken and the other cracked. Discovery of this second instance prompted the expansion of the inspection to all Unit 2 Darling check valves of the same design as those in which the degraded studs were found. There are 12 valves of this design installed in the ECCS and RNR systems in each unit at Cook Nuclear Plant.

All of these valves are classified as pressure isolation valves (PIVs) and leak tested in accordance with our IST valve program.

They are:

o (4) 10" check valves at the accumulator outlet (SI 166L1, L2, L3, L4) o (4) 10" check valves ECCS injection to cold legs (SI 170L1, L2, L3, L4) o (2) 8" check valves low pressure ECCS

($1 151E & V) o (2) 8" check valves normal RHR (RH 133, 134)

L Figure 2 provides a simplified flow diagram which identifies the locations of these check valves in either unit at Cook Nuclear Plant.

i Soon after the decision was made to initiate the inspection of L

Unit 2 check valves however, Unit 1 went from power operation to hot shutdown (Mode 4) due to an unrelated event. As a result, a s

decision was made to immediately inspect all Unit l' check valves of this design accessible in Mode 4.

In Unit 1, the only check valves accessible for inspection during the Mode 4 forced outage were 1 SI-151E, 1 SI 151W, 1 SI 166L1, and 1 SI 166L4 The Unit 1 inspections found one broken stud in each of the check valves installed in the low pressure ECCS (1 SI 151E & W) and stud material with an appearance not typical of Type 410 stainless steel in each of the two accessible accumulator outlet valves (1 SI-166L1, & L4).

'ThecontinuingUnit2inspectionsidoneff[edone.additionalcheck valve with one cracked stud (2 SI-166LA), and one valve (2 S1 166 L1) in which, although the studs were intact, the stud material did not have the appearance typical of Type 410 stainless steel, the material specified on the valve drawing for the retaining block studs. Both valves are located on the accumulator outlet.

A summary. of inspection results for the valves inspected in both Unit 1 and Unit 2 is provided in Table 1.

Actions Resulting from Check Valve Inspections In each of the cases discussed above, the cracked or broken studs, or studs of a material having an appearance not typical of f

w-i a

.l k

l

.. a l

g 3

AEP:NRC:1054 i

Mr. A. 3. Davis l

g Type 410 stainless steel were replaced with A.193 Crede 48 stud

)

material. This new stud material.is recommended by-the valve j

manufacturer for this application, in addition, maintenance job orders were initiated to replace all A.193 Crade 86 Type 410 J

stainters steel: studs with the new A.193 Crade 88 material j

re&ardless of whether any degradation is currently evident. This 1

action is also in accordance with the valve manufacturer's j

recommendation. To date, retaining block studs in 10 of the 12.

~.

1 Unit 2 check valves and 4 of the 12 Unit 1 check valves have been j

replaced with the new A.193 Crade 58 stud material. The studs in the remaining Unit 2 check valves will be replaced during the 1

current steam generator repair project outage. The studs in the remaining Unit I check valves are to be replaced during the next scheduled outage, with the possible exception of those installed in the low pressure injection lines to the cold legs (1.SI.170L1, j

L2, L3 and 1/ ).

Service conditions for these valves may not be conducive to the type of stud degradation observed in the other systems inspected. Westinghouse, who supplied the check valves under the original NSSS contract, and Darling, the valve manufacturer, were advised of the inspection findings discussed above. Westinghouse is conducting metallurgical evaluations to i

determine the root cause of the stud degradation.

Evalustion of Safety Significance With regard to the evaluation of the safety significance of our valve inspection findings, the following key factors were considered 1)

The check valves in their as.found condition had not failed, nor was valve operability impaired.

-l 2)

Inadvertent pressurization of a low pressure ECCS system is precluded since in each case where a check l

valve with potentially degraded studs was found, at t

least two valves were available to prevent back leakage from the reactor coolant system.

3)

The check valves would have performed,their intended function (i.e., opened) in the event of a 1hCA regardless of whether the retaining block studs had completely failed.

4)

Of the 10 valves inspected on Unit 2, three showed degradation of the retaining block studs and one was found to have questionable stud material. The corresponding Unit 1 valves which see the same service conditions as Unit 2 were inspected and studs replaced with the new stud material. No anomalies were observed in the remaining Unit 2 check valves inspected.

i

,, e

,i4-l l

Mr. A. B. Davis 4

AEP:NRC 1054 5)

The Unit 2 valves in the. degraded condition had.

functioned successfully in passing the required flow during Mode 5 or 6 operation at the beginning of the, l

steam generator repair outage. Successful operation'of these valves during this evolution is equivalent to passing the full flow test normally performed to confirm valve operability, Unit 1.was returned to service on September 15, 1988, and the actions discussed above to replace retaining block stude on both units have commenced.

This document has been prepared following Corporate procedures which' incorporate a reasonabis set of controls to ensure its accuracy and completeness prior to signature by the undersigned.

l i

Sincerely, 2

't M.

. Al ich Vice president idp

{

cc:

D. H. Williams, Jr.

l W. G. Smith, Jr. - Bridgman R. C. Callen G. Charnoff A. B. Davis NRC Resident inspector Bridgman i

G. Bruchmann i

h Og o

i i

t

(.

i TABLE 1

SUMMARY

OF ANCHOR / DARLING CHECK VALVE INSPECTIONS Retaining Block Valve size Service Location Stud Condition 1

Unit l' 1

1 51 151E 8'

Safety Injection (SI)

One Broken Stud To Hot Legs 1 SI 151W 8"

SI To Hot Legs one Broken Stud 1 SI 166L1 10" Accumulator Outlet QuestiongbleStud Material 1.S1 16614 10" Accumulator Outlet Questionable Stud Material i

Unit 2 2 RH 133 8"

Residual Heat Removal OK (RRR) To Cold Leg 2 RH 134 8"

RHR To Cold IAg OK i

2 SI 151E 8"

SI To Hot Legs One Broken Stud

~

3 One Cracked Stud 2 SI 151W 8"

SI To Hot Legs one Broken Stud One Cracked Stud 2 SI 166L1 10" Accumulator Outlet Questionable Stud Material

?

2 - SI-166L2 10' Accumulator Dutigt OK 2-SI-166L3

'10" Accumulator Outlet OK

~

2 SI-166LA 10" Accumulator Outlet One Cracked Stud l

2 SI-170L1 10" Low Pressure Injection OK To Cold Leg L

i 2 SI 170L4 10" Low Pressure Injection OK l

To Cold Leg 1'

l l

1 t

3 e'

f,F i'.

?

Table 1 Notes:

f 1)

A broken stud is a stud that has completely sheared into two parts.

In each case where a broken stud is reported, the break occurred 6t or near the plane of j

the interface between the valve body and the retaining j

block, j

2)

-Questionable stud material. refers to studs that looked shiny instead of having the black appearance typical of l

Typ6 410 stainless steel, the material listed on the valve drawings for the retaining block studs.

It appears that the "as found" material is either Type 304 or 316 stainless eteel.

l 3)-

A cracked stud is a stud that has partially sheared but f

has not parted into two pieces.

?

t c

t t

I e

i 1

s

/

l l

l

o u i

i ne! g.

iprev.viv

o.,yI s

.-n i

m 1

r I!ll lf hq g

i "l "I N I l

4 9

" * " '"u'

! i i n l i llnu w.

..=.L-)1 s.

i n+

t

1 I

1 i

u i

[

l !

l l I

,' f i L'

~., '.

!jl,il'[g.L i

g li u

l !!! !

l li" !

I It 1 '

rrg d ! ! i m!),,il

.Il l l!: r u u

l 1

. i.!..

i.i.

i. il

...i..i.i.

n

.n.

i

...m q.;

u n H I I

1 a t..,!It!III Idl il!If lM J

11 fp

=+

/

i l

t

)

. =mv,,.

i w.- =,

i S

g il sxxyp, I

[

i l

y F

(

/

~-

i a

I l \\

l

~,f c h"r;; i@.-- = l ly f

7 g

i i

g

..t-4g p

II ::l t

. i. ::

g H.

If i 11 11 U. -

..e

~it.

y q

vf tt!

g j,

,-l i i j'

Wg

i!

1l i

t-

-}

r s

i rW De E

x \\ ii t

r-i h

h'.

V/////rgfVjA f

t'

(__-

l H h%m j

I 1

i w

. s, C,\\ g

's s..

, g'

\\

'j

,g cc

/

i.__ O

,/

~

\\

,/

~,-

}!j!

, \\

... =

v 1

i i

se l

l l

t

,1 t

,t u

.e e

. : 'y.

ii n i e I

L1 1l : :

' l D

9 r

p, a;t w

i e

c lihr ' t1 N

i

}g,,

~

1.

g!

.f '{.' C l"';

m..If

)

h,

,[ i,

,-- 1 p

u s-rh o

0

{l s

f f

N

,' w}; V 4E

< > p' 8

9 l

x r'

O A

.,1 j

ir y

\\'

o b

j,!

plli

.i

. f-i p, "

_[

i l,

(l?

r 7

s.

I,,

g c F H II'j

.m ll I.

l 7

llig M

cx 1

L

., v h

g c

e

,-u a

,v

.I

.E L

.f s

r I

nt r

l i i

til Ii -

lis h

'l J

8*

Ij

.+

v.

-}Tl t-1 f,t e

r.

.::. :: :.~':.:"::

i

~1 l_c 11

) y v,f

I Ii---

l,t i

0

ji

.i il L

r i

5-i

~<!

1,

-w 1

-t,

i,,

_t r,

s i

.i

~..

l 1

is a

~

g 3

i k'

r; f

li'

~

I i

l

) gy,

i g

t s m a..

V..

,1 i.

,(

i

.,p..

I III 3

i-l A M I l1

,2

.m. -.

i!

,lT,,:

n i

t lqr:Ft:

7 Q

l s

,e

. mon, is a

uu i

I

l

. n m iw e.,

6 i

.=,o, i e (=..o e..

.l L.lCENSEE EVENT REPORT (l.ERI r

l m.......

. = >..

I AEll A. Fll!P AIRlCR NUCl! At POWit Pt ANT e 1. i o i o i o i s i s i s i losj e i) i.i...

H1'(.! l t'RiilS t. lilROT11.l: VAIAT. H01.1 S BROKl:N I)lT.10 ISil'Kol'l:R lil:NI 1 RI'NDti:N1

..e::c.

r.:

l.

.iiioioie, i i f

tj l,, I l tl i I

llI d l 0l 3

$ll Olh.' la 8l 7

~

0 i l ieie,0, i i

. ~

l

=

n,.,

~

n,

=

e,i o,n 2.c g._,,,2

..,i...,

.....n.

....n.

. n.....

Voi.t Ni Any e

... n

..n.....

i' kont.H i la ki.R, MAIN 11:NAsci. St l'cRINTLNol:N1 3l i 5 34d iiii 0 i

3 84 i

.u.

.n...n,..i

tp*,;;j;.

...,.. '. n..

3; 6

.;3;;5

.g.,*;,.;; s

.~.

nei.

h h t l l ai Y

r i s n ni ii n l

I l i I l l i

l i i l i i

-..............n.....

o..

e.,

.g......

7...

, emnee w,

m.,

7.o i

i n. x,....

.......~n..

l A maintenance inspection of the High Pressure Coolant injection (llPCI) (EIIS Code BJ) Terry Turbine during the 1987 j

refueling outage had discevered broken holts associated with the turbine throttir valves lifting beam.

The broken bolts did not make the turbine inoperable beenuse two dowel pins and two i

remaining bolts kept the lifting beam intact.

Since the two remaininn bolts were also cracked, an inoperable HPCI r.ystem l

could have occurred if this condition had not been discovered.

The corrective action included replacement of the bolts, ennnoltation with the vendor and metallurgical examination cf the boltn which indiceted failure by stresn corrosion cracking of improperly heat treated bolts.

Failure of lil'C1 coincident with small loss of coolant accidents is within the range of accidents considered in Final Safety Anulysis.

There have nrit ber.n any similar 1.ERs involving bolting tallure due to si ress corre.sion cracking atul improper heat t re. t i ng.

[

i

.., +.. ;. n.,; ;... p.;. ;

. :s..-.

?._,.

..,5,.

i.:.; m.~ ;.-

a an. >

i m.u.

l LICENSL EVENT REPORT ILER) TEXT CONTINUATION

=**d e *.* ao s ' * *

3

...n.

e u.........

. in,

o..

r....

d2.*

I'41E l

J Aufl A, ill!P A1Ritt Riittlet POWIR PL ANI 0 l 013 01 1 0 !.'

08 0!3 op lolol3 3l3 37

.m

=c The liigh Pressure Coolant Injection (IIPCl) (Ells Code BJ) i Terry Steam Turbine received a eemplete tear down and internal inspection during the 1987 refueling traintenance outege.

1)urirn l

this inspection, six (6) of the eight (8) holt s which hold t he throttic valve lifting beam together were found broken.

The remaining two (2) bolts were badly cracked.

Two (2) dowel pins and the two (2) remaining bolts were maintaining the lift beam intact.

If the lifting beam had not remained intact, turbine control would have been lost causing the HPCI system to be inoperable.

All beam bolting were replaced with specification conforming men cri al.

The vendor was on-site during the oserhaul and vendor ent neering was informed of the bolting failure.

The cracked i

bolts were analy:cd to detert ine the failure mechanism.

f A metallurgical evaluatior, of three failed ASTM A193, Grade l

Bb (Type 410 stainless steel) bolts was conducted to establish the most probable cause(s) of the cracking of the bolts.

Fractographic and metallogrcphic examinations showed that the holts failed by an intergranular failure rnode, most likely stress-corrosion cracking.

Chemical analysis confirrned that the chemical composition of the bolts met the material specification requirements.

However, hardness measurements showed that the bolts were much harder than expected.

This hip,her hardness is believed to have been a major contributor to the cause of

failure, l.aboratory tempering studies conducted at the minimum

~)

tempering temperature snd time (1100 F for 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br />) specified for the ASTM A19.1 Grade B6 bolts resulted in a drop in hardness of over 10 points Rockwell C, from the upact 30's for the as-received bolts to the mid-20s for tie laboratory retempered boltn.

I Copper was detected by electron dispersive X-ray analysis during examination of the thrend root regions in the scann1ng electron microscope.

It is believed that the copper may have come from a copper bearing antiseizure compound used on the boltn.

The use of such compounds have been shown to cause localize d pitting corrosion, which in turn, rnay have acted as the origin :ites of the stress-corrosion cracks in Type 410 stait,less steel.

It is recotrnended that when bolts are heat treated as specif.ed for ASTM A193, Grade B6 bolts the use of a lubricant such is a non-metal bearing petroleum jelly be used as an antir.einure compound.

This recommendation for bolt lubricants will be incorporated in the llPCI maintenance procedures.

.4...

LICEN$tt EVENT REPORT (LER) TEXT CONTINUATION

.a e.io m =o ** e*

~..... -....

t

" t M.* '

"" *Ji J4ut$ 8,illlPittitt RVttt At P0wlt Platt q 0j 3 gll Oli e' 09 e

oe 333 g;

nn

- =

The Licensee will inspect the lif t liern bolts during, the next refueling outarc r,cheduled in 198r.,

Safety consequences and implications of failure of the HPCI system is within the envelope of accidents and transients considered in the Final Safety Analysis Report.

Failure of HPCI coincidence with demand for the syste.n as a result of a small loss of coolant accident when the plant is operatiny, could recuire operation of the Autonatic Depressurization System (E!!S Cocle AD) to reduce reactor pressure to within the range of 1.ow i

Pressure Core Spray (Ells Code BM) and Low Pressure Coolant Injection (E!!S Code BO).

t There have not been any similar 1.ERs involving bolting failure due to stress corrosion cracking and improper heat treatment.

1 t

l 9

2 'c ~ ~

.a

O-Wr,,>*-

4:. A #:ee..a i j, Capt toes Demog 760 set

u a e. %. ai i,.....

.,.ipwi e.s. 'sa, o. a.

3, l

Radforal J. Conver w l

" ' " ' " " ~ "

> NewYorkPower 4# Authority I

L f

June 24, 1987 L;

JAFP 87-0503 p.

1 United States Nuclear Regulatory Commission

. Document Control Desk t

Washington, D.C. 20555

REFERENCE:

DOCRET NO. 50-333 LICENSEE EVENT REPORT:

87-003-01

' l i

Dear Sirs 4

'C Enclosed please find referenced Licensee Event Report in accordance with 10CFR50.73.

If there are any. questions concerning this report, please contact Mr. Robert Baker at 315-349-6201, j

l Verytrujyy rs.

,/-

/

g.w1 d RADFORD J. CONVERSE 1

RJCikbinan i

CC:

USNRC, Region I (1)

- t INPO Records Center, Atlanta, Ga. (1)

American Nuclear Insurers (1)

Internal Power Authority Distribution NRC Resident Inspector Document Control Center LER/OR File l.

I l

'y*

e, - i.

i x

,e

.y'.

,q,,

4

,, _. -

v t e Letter Sequence Other
TAC:69826, (Open)
Initiation Acceptance... Administration Questions Answers Results Approval... Other: ML20011C479
MONTHYEAR1988-12-20 [Table View]

ML20011C479

Text

SUBJECT:

Enclosure:

Dear Mr. Davis:

SUMMARY

REFERENCE:

Navigation menu

Search