ML20072S958

From kanterella
Jump to navigation Jump to search
Part 21 Rept Re Crane Design Gate Valves Involved in Jan 1994 Incident at Praire Island Nuclear Generating Plant. Historical Records Do Not Contain Info Addressing Total Number of Configuration Mfg by Crane
ML20072S958
Person / Time
Site: Prairie Island  Xcel Energy icon.png
Issue date: 08/30/1994
From: Hornyak R
AFFILIATION NOT ASSIGNED
To:
NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM)
References
REF-PT21-94 NUDOCS 9409140324
Download: ML20072S958 (5)


Text

.

__._m.

m I

i TELEFXDNE (613) 727-2600

/

FOST G7FICE BOX 3339 JOLIET. IL 60434 CRANE NUCLEAR OPERATIONS JOLIET. IL 60431 104 NORTH CHICAGO STREET CRANE VALVES i

August 30,1994 l

Nuclear Regulatory Commission Document Control Desk Washington, DC 20S55

SUBJECT:

1%)TENTIAL 10CFR21 CONDITION l

REFERENCE:

a.) Crane Letter dated August 23,1994 h.) Northern States Power Letter to the

- NRC dated April 18,1994 Gentlemen:

This letter will. serve to adviseyou of our continuing investigation concerning Crane design Gate Valves involved in the January,1994 incident at the Prairie Island Nuclear Generating liant. Details of that incident were contained in the reference (b) correspondence, a copy of which as attached t

We have been in constant communication with Westinghouse Electric's NATD Engineering Technology Department concerning their December,1969, P.O. 546-CAK-116878 BN on which the valves in question were purchasedfrorre Crane Company, in turn, Westinghouse supplied the valves to a number ofJVuclear Utilities, recordsfor which have been retrieved from Crane archives, On August 23,1994, a letter was issued by Crane to those Utilities to whom valves were 7

supplied. A sample copy of the letter is attachedfor your reference and a total list of Utilities to whom the letter was sent is showr; on the attachedpage, t

At this point, we consider the incident a potentially reportable condition based on the NSP evaluation at theirplant. Upon receipt of additional inputfrom the balance of the plants, we will assess the situation to determine a subsequent wurse of action. llistorical records do not contain information addressing the total number cf valves of this configuration i

manufactured by Crane, nor their application throughout the Industry.

9409140324 940830 h

PDR ADOCK 05000282 hq s

l 1

Page 2 As additional information becomes available, we will advise. You may direct any questions to the writer or to our Manager of Developmental Engineering, Mr. Bruce flarry. My direct line is 815-740-7S97 and Bruce can be reached at 815-740-7570.

Sincerely, Ronald F. Ilornyak Manager QNSupport Engineering RFil/cs cc:

J. Carlson F. Bisesto B. Ilarry K. Ilutchinson II. Sandner (Westinghouse)

File Attachments l

t

m..

m TEli,KDNE I4166 F27-2000 FDST G7ECE BOX 3339 JouET.it 60434 CRANE NUCLEAR OPERATIONS 104 NORTH CHICAGO STREET JOUET,IL 60431 CRANE VALVES August 23,1994 Beaver Valley Power Station SEB-3 Duquesne Light Company P.O. Box 4 Shippingport, PA 15077 Gentlemen:

In January of this year a containment isolation valve at the Prairie Island Nuclear Station was damaged as a result of operationalproblems involving maintenance procedures.

Details of the event have been reported to the NRC via Prairie Island's Docket Nos. 50-282 and 50-306 and associated NRCInspection Report Nos. 282/94002 and 30684002.

The valve in question was a Crane 10" Figure 63174 and during the course of the investigation, it was discovered that, contrary to information shown on the valve assembly drawing, shims were installed between the yoke and the adapter plate. The extraordinary conditions experienced by the valve during the event resulted infailure of the weldsjoining the yoke and adapter plate and more catastrophicfailure to the other valve parts.

Crane assisted the Utility in evaluating additional valves of the same design for evidence of the existence of shims andfor the adequacy of the attachment welds where shims did exist.

It was concluded that a substantial safety hazard did not esist at the Prairie Island plant, based on the walk-down that was performed and additional analytical evaluations of the ss-built valve characteristics.

Since that incident, Crane has conducted additional evaluations, including a search of archives stored in a Cranefacility in Pennsylvania. The valves at the Prairie Island plant were part of an order produced by Cranefor Westinghouse Electric in the early 1970's.

Although the records retrieved contained no information to support the addition of shims to the design, the records indicate that valves of the Figure 63174 design were supplied to you on that same Westinghouse order.

Crane requests your cooperation in determining if those valves are in your system. We will be glad to assist you in evaluating the condition of the hardware and in performing any analytical evaluations to determine if safety hazards exist. Valves of the 63174 design were supplied in 10" and 12" sizes.

i

{.

t You have been contacted based upon your affiliation with the Motor Operated Valve User's Group and its Member Koster. We appreciateyour cooperation in channeling this conununication to the appropriate penannel at yourfacility.

Your questions can be directed to the writer or to our Manager ofDevelopment l

Engineering, Mr. Bruce Harry (who performed the initial on-site evaluation at Prairie Island) or Mr. David Dwyer, our Analytical Project Engineer, who has paid subsequent l

visits to the Prairie Island plant. Our general number in Joliet is 815-727-2600.

Very Truly Yours, i

Ni R.F. Hornyak Manager QA/ Support Engineering i

RFH/lik cc:

J. Carlson K. Hutchinson B. Harry D. Dwyer i

Attachment Names / Address of Recipients of August 23,1994 Letter l

Beaver Valley Power Station Kewaunee Nuclear Power Plant SEB-3 Insconsin Public Service Corp.

l Duquesne Light Company N 490 Huy 42 PO Box 4 Kewaunee,IVI 54216 Shippingport, PA 15077 A TTN:

Mr. Neil Morrison A TTN:

Mr. Larry L Limberg Senior Engineer Maintenance Engineer Diablo Canyon Power Plant MC N-50 Pacific Gas & Electric Co.

Public Service Electric & Gas Company PO Box 56 P() Box 236 Avila Beach, CA 93424 Hancocks Bridge, NJ 08038 A TIN:

Mr. Don Bauer ATTN:

Mr. Robert S. Lewis Senior Staff Engineer Corporate Office j

BR SA Tennessee Valley Authority 1101 Market St.

Chattanooga, TN 37402-2801 A TTN:

Mr. Richard G. Simmons Program Manager, Valves Innsbrook Technical Center Virginia Power Cmnpany 5000 Dominion Blvd.

Glen Allen, VA 23060 A TTN:

Ms. Pamela E. Detine lyoject Engineer

l Northam States Power Comparty i

Prairk askend NucJear Gererating Plant i

l 1*C Wamoriade Dr. tn5 w sen. u.w x.na socs0 e

l April 18, 19 %

10 CFR Part 2 Appendix C US Nucleas Regulatory Coe.massion Attn: Document Control Desk

)

unshington, DC ':0555 PRAIRIE ISIAtm NUCLEAR GDTERATING PIJWT Docket Nos. 50-282 License Nos. DPR-42 S0-106 DPR-60 Reply co a Notice of Violation NRC Inspection f.eport Nos. 282/94002(DPJ) and 305/94002(DRP)

I're c e dura i Deficiency Allowi n t r> qpare _ t o a safety-related Valve Your latter of liarch 18, 1994 transmitted :he subject inspection report and violation notica which required a 33 day response, Attached is our responce.

In our responsa, va have r,sde new NRC com:nitments which are identified as such in the attachment as the statements which are in italics,

If you have any questions regarding this response, please contact Jack Leveille (612-388-1121, extension 4662).

/

(oA c'-s.jU-4

\\

z 1.. ru G.r.ecai %an#ger Prairie Island Site cc:

Regional Administrater III, SRC Senior Pesident Inspe cta r, 1GC l

?TR Proj e.ct. Ma na ge r, b"RC i

J. E. Silberg Attach =ent.

RESPONSE To VICL'dION r-G L{D f) $ 0 h b - -

l l

.:a w:=nw m. c=

_oc. 2.M 3.ca

< dt W.I.3iE"$ 01 5319 G Il dC ilUdd U2:UI P0' EI W l

l P.F.Sp0NSE To t'ICLA! ION Zi.e.la thti Criterian J of 10 CpE SC, Appendix 3 requires that' activities affecting quality be prascribed by documented instructions, procedures, or drawings, of a type appropriate to the circumstances and shall be accomplished. in accordance with these instructions, procedures, or

drawings, Contrary to the above. on January 24, 1994 extensive damage to the normally closed outboard containment isolation valve (MV-32181) berveen the containment building sump.and the suction of No. 22 residual heat removal pump occurred when the electrician,- stroking the valve locally frca the motor control concer, continually depressed the close contactor instead of depressing the open contactor. The maintenance procedure used to locally stroke the valve was not appropriata to the circumstances in that: 1) the procedure did not raquire that direct communicacions be established between the electrician and operators i

I during stroking of the valve: 2) expected values for motor current draw were not included; 3) the method oi making up the contactor vas not specified (i.e. the open contactor did noc have to be continually depressed in order to operate the valve); and 4) the procedure required c:rorless human perrornance because of the absence of actuator protective feature s.

l 1

l This is a Severity f.evel IV Violation (Supplement I).

Backereund We consider :his failura to be significant since a risk significant component failad, albeit not while performing in the accident functional mode.

Although the work procedure for cyclin 5 the valve had the deficiencies noted in the violation, it should be noted that the plant electricians have cycled motor valves locally at the breaker using the contactors numerous times as part of zotor~ operated valve (MOV) testing efforts.

Such cycling activities have occurred since the 1980's.

In an effort to reduce the likelihood of the cotor valve and system en5 neers initiated the procedure referenced i

scror, in the violat2on.

In preparing the procedure the engineer had spoken with electrical maintenance personnel experienced in motor operated valve testing l

to ensure that the procedure would be appropriate.

Based on this input, the pro:cdure was developed for cycling K7-32181.

nV-32121 was the tnird valve to be cycled in response to a NRC identified concuen regarding the pocantial for ptessure locking.

The person assigned to perfore the :ask was a journeyman electrician.

Previously in January, two l

different clettricians had used the same procedure satisfactorily on the two l

Unit 1 valves.

nun:

A24 4.

~ 11 63 9 5.);0

~a3E.'03

._00 Wu2Wd

M r'J a K.o 01 911W UDI1 M O Hodd i2:01 ru.. I Tif I

=

b Att edgmen t Page 2 of 5 K* e s o'n for che Violation Two root cause analyses vere initiated - one to deternine the root cause for the cve.nt and the second to datermine the valve failura root cause (this is discussed in the " Corrective S teps Taken and Results Achieved" section). The avent root cause analysis was performed by the on-site Error Reduction Task Forcc (EETF) and is documented in dRTF Report 94-01.

The EATF repoet identified evo primary causes, evo secondary causes. and tvo additional possible causes for tne event. The primary cause was determined to be htunan error. The causes were:

Fritnarv causup - Human __Eryg 1.

Self-checking was not applied to verify that the choice of contactor was correct. or that the intended action was correct.

We electrician depressed the wrong contactor.

2.

The electrician did not have the proper information at the job site to verify whether the valve's circuit was seal-in or not.

L'i chout tais nfernation, the electrician pressed and held the contactor to ensure the valve vould go open for the required 30 seconds.

Holding the contactor in bypasses the torqua switch trip.

l l

<econdary causes - Erronomic l

1.

The open and close contactors in the MCC breaker cubicle were not labeled.

I 2.

  • he work request was somewhat generic in that,it did not specify l

the expected current draw.

Also, the work request did not contain instructions to push and release :.hc contacter, nor did it nention the seal-in featu:e.

Possibl+ contribuc h eausoc 1.

No couanuntentions were established between the MCC breaker cubicle l

and the snot at valv.s.

i l

l 2.

Conseqwaces of potential error were not discussed before starting I

the work.

Correctite Stecs Taken and Results__ Achieved The affected valve was repaired and restored to service on January 26, 19 9 t+.

Prior to cyclin 6 additional =otor operated valves (e.g.. MV-32180, Containment i

trG6CM l

l r

Y - '

~:: =

su 92 m:

=me.c:a s

octar m 9rifra!.%IC8 01 DNggrg31, tio uoaa 13:ct & O Til t

l.

I it Ls<Annem t Pace a or 5 Sump S to 21 EHR Fump) by this method, the causes of the event were identified uni correctivo actions to prevent a similar event were discussed with the plant alectricians. A more detailed procedure was used for the cycling of the next valve, MV-32180, which was cycled successfully on February 6 1994 Prairie Island has developed a videotape intended to emphasize self checking.

This video, "Right f rom the Start, has now been viewed by some, but not all, of the ' plant staff, including the e lectricians.

A method was implemented to label the open and close contactors in motor valve hCC breaker cubicles. To date, the contactors in 142 MCC breaker cubicles have been labeled.

On March 24, 1994, tamporary memos (94-24 and 96-25) vara issued to both I

unics' quarterly surveillance procedures. SP1089 and SP2089 (Residual Heat l

Removal Pumps and suction Valve.s from the Refueling Vater. Storage Tank),

respec tively.

Thase temporary memos are refinements of the procedure used for the M7-32180 cycling.

These refinements were developed durin5 a post-event evaluation by those involved in the event. These procedure changes involved cycling the Sump B valves locally for potential pressura locking.

The i

procedure specified the following additional information/ requirements beyond those specified in the original work request:

(1) precaution to self -check, (2) parform a pce-job briefing, (3) use of headset communication between the electrician at the MCC breaker cubicle and the operator at the valve, (4) identificacion of the expected full load and nameplace locked rotor ampera ge,

(S) caution that che contactor need not be hold In, but only momentarily depressed since it is a scal-in circuit, and (6) verification enat the open and closed contactors are labeled.

The MOV testing engineers were advised to ensure adequate ' instructions are provided for those cases where local operation of an MOV is required.

They vill evaluate their procedures for necessary changes.

A Safety Evaluation revision provides short torm justification of the operability of the Sump B valves based on the calculational methodoloEY af required opening thrusts under pressure locking condition.

These valves will no longer be cycled for pressure locking concerns. A modification is being prepared to codily the valves to prevent pressure locking, j

The Equipment Failure Root Cause evaluation was initiated by the engineering 1 & 002

~1R e ' l~ ' 91..' ' fi d!2 939 526C UNO8.OC5

A WC' B d R &/. d I83 91 C D Ed1 I

l

+

At.tathstant tuo s at s stati'with the assistance of the valve manufacturer (Crane /Aloyco), an

  • 1ndependent engineering firm (Altran) experienced in failure analysis of this type, and the corporate Materials and Spectal Processes department. The vendot analyses are. still in pror,ress.

l Mcchanically the weld failed at the point where the yoke arm was attached to the actuator adapter plate.

This failure location uns not predicted by the l

valve rnanufacturer's weak link analysis. Rather, failure was anticipated in the yoke.

The failure root cause analysis has detemined that the following i

additional factors contributed to the prematura failure:

(1) A shim was installed between the yoke and adapter plate that resulted in a lower failure stress than design.

(2) Ihe shim was not indicated on the fabrication drawings, (3) The shita was not included in the desi n calculations, and 5

i (4) The weld that broke was poor quality as indicated by less-than-design fusion.

Preliminary data from Crane and Altran indicate that the as-built values for the weldments were less than the _ original Vastinehouse spgcification desig values.

These additional lactors ar_e_ apparently due__t;o. inadentrate Crane, quality assuranca and controls and inadequate Westinghcuse oversight.

Crane waQe]manuf acturer and Westinghouse was the supplier.

Ce have determined that the us. built valvo characteristics do not constitute a sntisTalRTa""1 salerfjnurfar4T5f tWe traTire ETind plant applicatton. 11ovaver,

the existence of the shims in valves in different applications in other plants may present a substantial safety hazard. We _have notified Vestinghouse that

..[

they may need to perfom an evaluation for 'l'O CFR Part 21 reporting purposes. I(

Although the torque switch was bypassed due to the actions of the electrician.

h,; E,6N lower setting sinto the g f, y, the tcrque switch settinp could have been set at a plant design differencial pressure is 46 psid rather than the generic 700 psi.d,

& si.gn differential pressu:e.

The discovery of the shim led to expansion of the investigation to all other Crane valves of this type used in safet'/-related applications. These valves Inciude:

containment Sump S to PJR:

12' 32075. 32076. 32077, 32078 (Unit 1) 32178, 32179, 32180. 32181 (Unit 2)

RUST to P.HR Pr.:=ps :

10" 32034, 32085 (Unit 1) 32137, 22188 (Unit 2)

The tour Pl.;$T to RIR furep valves and the other three outside Sump h to PRR

c,oa l'

' 32.

I } -6 12 W)TMO P Q,, M UO ':00 ~ 3 dd T dPO;a'. SIC. 71 Of II SI d Jl l 'lJO 4 PJdd 2 :OI Ud

  • EI TIf

l At40chment r

5 or s i

talve's cere inspected for presence of a shic and weld qualicy. The results of the inspeccions were satisfacrery.

The four inside Sump B to RER valves verc twe itepected at this cime since a unit shutdown vould be needed and it is teliewd that inspection of the remaining valves provides a reasonable epoccation that the valvan inside che contain:none boundary are not different.

Tha torque switch setting for MV-32181 has been adjusted downward to correlace closer to the 46 psid plant design than the higher ganario design differential pressure, h is change had been planned prior to the event occurrence.

A iurther investigation of the maintenance history showed that one of the Unit 1 RUST to RHR pump valves failed in 1975. At that time those four valves were evaluated and revelded as appropriate.

j J

Corrective Steps Tant Will !!e Taken To Avoid Nrther Violations The need for a revision of the planc maincenance procedure wrLcers ' guide vill be reevalesced, by June 1, 1994, in light of the observarlons x::sde during the evaluatica of this evenc.

1

.4 Maincenance seccion procedure vill be compleced, by June L,1994, chac describes in detail che method for cycling a accor valve locally by using open and close concaccors since chis spproach is used La che HOV cescing program.

The remaining centainment sump a valves vill be inspected for presence of che

}

shim and weld quality and the corque svicch seccings vill ha readjusted to i

corralace wich che lower planc design differencist pressure, by July 1.1994 for Unic 1 and July 1,1995 for Unic '

i The videotape "Right fren the Start" has been incorporated into the Cenaral j

u:rployee Training re-qualification progra=, which is presented to all l

persor.ncl badged for access to the plant.

Date Cien Full Compl(ance Vill Be _ Achieved Full cerepliance has been achieved.

to.*2 Ni.~

'N

1. I *.3 6Q 93 Tl=.3 M2, 2[

100' 00 7:ed 9r2t1f45100 01 941SILOll ldo N0d ;

22:01 N, O Tf

. -... -,..-. -,, - -.