ML17037C468

Letter Regarding a Stress Assisted Corrosion Problem and an Enclosed Safety Evaluation by the Office of NRR - Supporting Amendment to License No. DPR-63 and Changes to the Technical Specifications Inoperable Control Rod ..

ML17037C468
Person / Time
Site:	Nine Mile Point
Issue date:	09/23/1975
From:	Lear G Office of Nuclear Reactor Regulation
To:	Rhode G Niagara Mohawk Power Corp
References
Download: ML17037C468 (20)

Text

2 V'EP 2 3>g75 Docket No.50-220 Niagara mohawk Power Corporation ATTN: Hr.Gerald K.Rhode Vice President-Engineering 300 Erie Boulevard Hest Syracuse, New York 13202 DISTRIBUTION:

NRC PDR Local PDR Docket ORBft3 Rdg KRGoller TJCarter OELD 016E (3)GLear JGuibert SATeets SVarga DEisenhut TBAbernathy ACRS (16)RSilver Gray file extra cps (5)Gentlemen:

NINE NILE POINT UNIT 1 Cracks have been detected in the collet housings of the control rod drives at Dresden Unit 3, Browns Ferry 1, and Vermont Yankee.The problem appears to be a stress assi.sted corrosion problem that may be generic to most boiling water reactors.In light of this experience, we believe thpt appropriate changes to technical speci.fications for thi.s type reactor are needed that vill prohibi.t extended operation with immovable rods.Accordingly, unless you inform us in writing within 20 days of the date of this letter that you do not agree, with this course of action, including your reasons, we plan to initiate steps to issue the enclosed change to the technical spec'ifications of your facili.ty.

A copy of our related safety evaluat'ion on this matter-is.enclosed.Sincerely, George Lear, Chief Operating Reactors Branch I'f13 Divi.sion of Reactor Licensing

Enclosure:

1.Technical Specifi.cations 2.Safety Evaluation OPPICC~dURNAM28P'ATE 3P'JG ibert:kmf gi y~i7S RSilyer/7S ORB03 GLear 9I~~LZs PDI2D hEC.318 (Rev.9 53)hXCM 0240 4 U, d, OOVCRNMCNT PRINTINO OPPICCI IOTA 42d Idd

~0'l Niagara Mohawk Power Corporation CC: Arvin E.Upton, Esquire LeBoeuf, Lamb, Leiby$MacRae 1757 N Street, N.W.Washington, D.C.20036 Anthony Z.Roisman, Esquire Berlin, Roisman 5 Kessler 1712 N Street, N.W.Washington, D.C.20036 Dr.William Seymour, Staff Coordinator New York State Atomic Energy Council New York State Department of Commerce 112 State Street Albany, New York 12207 Oswego City Library 120 E.Second Street Oswego,.New York 13126

()t t.l I f LIMITING CONDITION FOR'PERATION SURVEILL'ANCE RE UIREMENT (2)Reactivity margin-stuck control rods Control rods which cannot be moved with control rod drive pressure shall be considered inoperable.

Inoperable control rods shall be valved out of service, in such positions that Specification 3.1.1 a(l)is met.In no case shall the number of non-fully inserted rods valved out of service be greater than six during power operation.

If this specification is not met, the re'actor shall be placed in the cold shutdown condition.

If a partially or fully withdrawn control rod drive cannot be moved with drive or scram'pressure the reactor shall be brought to a shutdown.condition within 48 hours5.555556e-4 days <br />0.0133 hours <br />7.936508e-5 weeks <br />1.8264e-5 months <br /> unless investigation demonstrates that the cause of the failure is not due to a failed control rod drive mechanism collet housing.b.Control Rod i'withdrawal (1)The control rod shall be coupled to its drive or completely inserted and valved out of service When removing a control rod drive for inspection, this requirement does not apply as long as the and all other operable rods fulLy inserted.(2)Reactivity margin-stuck control rods Each partially or fully withdrawn control rod shall be exercised at least once each week.This test shall be performed at least once per 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> in the event power operation is continuing with two or more inoperable control rods or in the event powe~operation.

is continuing with one fully or partially withdrawn rod which cannot be moved and for which control rod drive mechanism damage has'not been ruled out.The surveillance need not be completed within 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> if the number of inoperable rods has been reduced to less than t&o and if it has been demonstrated thai control rod drive mechanism collet housing failure is not the cause of an immovable control rod.b.Control Rod Withdrawal (1)The coupling integrity shall be verified for each withdrawn control rod by either: (a)Observing the drive does not go to the overtravel position, or 27

'I BASES FOR 3.1.1 AND 4.1.1 CONTROL ROD SYSTEM maximum contribution to shutdown reactivity.

lf it is valved out of service in a non-fully in-serted position, that position is required to be consistent with the shutdown reactivity limita-tion stated in Spec'fication 3.1.1 a(l),>>hich assures the core can be shut down at all times with control rods.The allowable inoperable rod patterns will be determined using informaiton obtained in the startup test program supplemented by calculations.

During initial startup, the reactivity condi-tion of the as-built core will be determined.

Also, sub-critical patterns of widely separated withdrawn control rods will be observed in the control rod sequences being used.The observa-tions, together with calculated strengths of the strongest control rods in these patterns will comprise a set of allowable separations of malfunctioning rods.During'the fuel cycle, similar observations made durihg any cold shutdown can be used to update and/or increase the allowable patterns.The number of rods permitted to be valved out of service could be many more than the six allowed by the specification, particularly late in the operating cycle;however, the occurrence of more than six could be indicative of a generic problem and the reactor will be shut down.Placing the reactor in the shutdown condition inserts the control rods and accomplishes the objective of the specifications on control rod operability.

This operation is normally expected to be accomplished

~>>'ithin ten hours.The weekly control rod exercise test serves as a periodic check against deter-ioration of the control rod system.Experience with this control rod drive system has indicated that weekly tests are adequate, and that rods which move by drive pressure will scram when required as the pressure applied is much higher.Also if damage within the control rod drive mechanism and in particular, cracks in drive internal housings, cannot be ruled out, then a generic problem affecting a number of drives cannot be ruled'out.Circumferential cracks resulting from stress assisted intergranul'ar corrosion have occurred in the collet housing of drives at several Bus.This type of cracking could occur in a number of-drives and if the cracks propagated until severance of the collet housing occurred, scram could be prevented in the affected rods.Limiting the period of operation with a potentially severed rod and requiring increased surveillance after detecting one stuck rod will assure that the reactor will not be operated with a large number of rods with failed collet housings.Control Rod Withdrawal (1)Control rod dropout accidents as discussed in Appendix E*can lead to significant core damage.If coupling integrity is maintained, the possibility of a rod dropout accident is eliminated.

The overtravel position feature provides a positive check as only uncoupled drives may reach this position.Neutron instrumentation response to rod movement provides an indirect verifica-tion that the rod is coupled to its drive.Details of the control rod drive coupling are given in Section IV.B.6.1.*

• FSAR 34

~l i' I UNITED STATES NUCLEAR REGULATORY COMMISSION WASHINGTON, D.C.20555 SAFETY EVALUATION BY THE OFFICE OF NUCLEAR REACTOR REGULATION SUPPORTING AMENDMENT TO-LICENSE NO.DPR-63 AND CHANGES TO THE TECHNICAL SPECIFICATIONS INOPERABLE CONTROL ROD LIMITATIONS NIAGARA MOHAWK POWER CORPORATION NINE MILE POINT UNIT 1 DOCKET NO.50-220 INTRODUCTION r On June 27, 1975, Commonwealth Edison Company (CE)informed NRC that cracks had been discovered on the outside surface of the collet housings of four control rod dr ives at Dresden Unit 3.The cracks were discovered while p'erforming maintenance of the control rod drives;the reactor was shutdown for refueling and maintenance.

In a letter dated July 3, 1975, CE i.nformed us that i.f the cracks propagated until tip'collet housing failed, the af fected control rod could not be moved~"), In a meeting with representatives of General Electric (GE)and CE the NRC staf f was advi.sed that furCher inspections revealed cracks in 19 of the 52 Dresden 3 control rod drives inspected, in one spare Dresden 2 control rod drive, in one Vermont Yankee spare control rod drive and in two GE test drives'.In a report dated July 30, 1975, after (3)additional rod drives were inspected CE stated that cracks had been found in 24 of 65 drives inspected (".Recently, the Tennessee Valley Authori.ty reported that cracks were found in the colletihousing of (1)Telegram to J.Keppler, Regi.on III of the NRC, June 27, 1975, Docket No, 50-249'2)Letter from B.B.Stephenson, Commonwealth Edison Company Co James G.Keppler, U.S.Nuclear Regulatory Commission, July 3, 1975, Docket No.50-249./(3)Memo from L.N.Olshan, Division of Te<<hnical Review (DTR)to T.M.Novak, DTR,"Meeting on Cracks Found in Dresden 3 Control Rod Drive Collet Retainer Tubes," July 18, 1975.(4)Letter from B.B.Stephenson, Commonwealth Edison Company to James G.Keppler, U.S.Nuclear Regulatory Commission, July 30, 1975, DockeC No.50-249.

4~~l

>~2 r.r r I~seven of nineteen drives inspected at: Browns Ferry 1 and Vermont Yankee found cracks in the.collet housing of 4 of 10 control rod drives inspected.

Because a number of control rod drives have been affected, because complete failure of the drive collet housing could prevent scram.of Che affected rod;and because we do not consider exi.sti.ng license requirements adequat:e in view of the collet housing cracks experienced, we have concluded that the Technical Specifi,cations should be changed for those reactors with control rod drive designs susceptible to collet.housing cracks.The change should assure that reactors which could be affected would not be operated for ext:ended periods of time with a cont:rol rod which cannot be moved.DESCRIPTION The control rod drive is a hydraulically operated uni.t made up primari.ly of pistons, cylinders and a locking mechanism to hold the movable part of the drive at the desired position.The movable part: of the drive includes an'index Cube with circumferbntial, grooves located six inches:apart.The collet assembly which serves as the index tube locking mechani.sm contains fingers whi.ch engage a groove in the index tube when the drive is locked in posi.tion.

In addition to the collet, the collet assembly includes a return spring, a guide cap, a collet retainer tube (collet housing)and collet pi.ston seals.The collet housing surrounds the cqllet and spring assembly.The collet housing is a cylinder with an upper section of wall thickness 0.1 inches.and a lower sect:ion with a wall thickness of'bout 0.3 inches.The cracks occurred on the outer surface of the upper thin walled section near t: he change in wall t:hickness.

1.Consequences of'racking The lower edges of Che grooves in the index tube are t:apered, allowing index tube insertion wit:hout mechanically opening the colleC fingers, as they can easily spring outward.(If the collet housing were to fail completely at the reported crack location,.

the coil collet spring could force the upper part of the collet housing and spring retainer upward, to a location where the spring and spring retainer would be adjacent to the collet fingers.The clearance between the collet fingers and the spring when in this location will not permit the collet fingers to spring out of the index tube groove.This would lock the index tube in this position so that the control rod could not be inserted or wi.thdrawn.

I

The failure of up to six conf rol rods to operate has previously been-evaluated and the Technical Specifications presently allow up to six rods to be inoperable.

If.more than six rods are inoperable or if the scram reactivity rate is too small or if shutdown reactivi.ty requirements are not met, the existing Technical Specifications require the reactor to be brought to a cold shutdown conditi.on.

Reactor power operation with up to six rods inoperable would not involve a new hazards consideration nor would it endanger.the health and safety of the public.2.Probable Cause o f Cracking The cause of the cracking appears to be a combination of thermal cycling and intergranular stress corrosion cracking.The thermal cycling results from insertion and scram movements-.

During these movements hot reactor water is forced down along the outside of the collet housi.ng, whi.le cool water is flowing up the insi.de and out of flow holes in the housing.='hese thermal cycles are severe enough to yield the material, leaving.a high residual tensi.le stress on the outer surface.The collet housing materi,al i.s type 304 austeniti.c stainless steel.The lower portion of the collet housing has a thicker wall and i.ts inner surfaqe is nitrided for wear resistance.

In 1960-61, si.milar drives usi.ng high hardness 17-4 PH material for index tubes and other parts were found to have developed cracks.The problem caused GE to switch to ni.trided stainless steel-The nitridi.ng process involves a heat treatment in the 1050 F to 1100 F range, which sensitizes the entire collet housing, making it susceptible to oxygen stress corrosion cracki.ng.

The cooling water used in the drives is aerated water.This water contains sufficient oxygen for stress'orrosion to occur in the sensiti.zed material i.f it is subjected'o the proper combination of high stresses and elevated temperatures.

We believe that the'racking is caused by a combination of thermal fatigue and stress, corrosion.

GE has determined that both full stroke inserti.on and scram will cause hi.gh thermal stress.The cracks are completely intergranular and extensively branched, indicating that corrosion is a major factor.The type of theimal cycling, plus the, buildup of corrosion products in the cracks be-tween cycles probably results in a ratcheting action.This"is also indicated by the"bulged" appearance of the cracks on the OD.

1 (\~

f~3.Probability of Early Failure I We believe that the cracking is progressive and is cycle dependent~Although the details of the cracking process are still not clear, we have not identified any mechanism that would cause rapid cracking with progression to complete ci.rcumferential failure.The axi.al loads on the housi.ngs are very low at all times so that through wall cracks would have to progress at least 90%around the circum'ference before there, would be concern about a circumferential fai.lure.Although one housing at Dresden 3 had three cracks which nearly joined around the circumference, no cracks at Dresden 3 were through wall and none of the housings exami.ned approached the degree of cracking necessary for failure.The collet housing has three flow holes in the thin section equally, spaced around the circumference.

The observed cracks have been confined primarily to the areas below and between the holes and near thy area where the wall thickness of the collet housing changes.Since all the cracks except those located at the change in wall thickness are fairly shallow and since those at the change in wall thickness are largely confined to the ci.rcumferential area between holes, the net strength of the cracked housings is sti.ll far greater than necessary to perform their function.A test drive at GE that had experienced over 4000 scram'cycles had a more extensive developed crack pattern.Although Che sati.sfactory experience with this cracked test housing is encouraging, its performance may not be correlated directly to thaC of drives in service, as this test drive was subjected to lower temperatures, and possi.bly less severe thermal cycles than could be encountered in actual service.The cracks were first noticed on Che test drive after about 2000 cycles-many more cycles Chan the cracked housings at Dresden 3 had experi.enced.

The chance that a large number of collet housing would fail completely at about the same time is very remote.This is primari.ly true because the distributions of failures by cracking mechanisms such as stress corrosion and fatigue are not linear functions.

That i.s, failure is a function of'og time or Log cycles.Distribution of fai.lures of simi.lar specimens generally follow a log normal pattern, wi:th one to two orders of magnitude in time or cycles between failures of the fi.rst and failures of the last specimen.As'o collet housing has yet fai.led, we are confident that there would be very few, if any, failures during the next time period corresponding to the total service life to date.

Cj I.~~I 4.Changes to Technical Specifications I Existing limiting conditions of operation" allow operation to continue with up to six inoperable control rods.Existing surveillance requirements'specify that daily surveillance of the condition of all fully or partially wi.thdrawn rods would not have to begin until two rods are found inoperable.

We do not consider that these existing limiting condi.tions of operation and surveillance requirements sufficiently limit the possibility of operating for an extended period of time wi.th a number of rod drive mechanisms which cannot be moved.We have therefore'oncluded that the Technical Specifi-cations should be changed as di.scussed below.(a)One stuck control rod does not create a significant safety'oncern.However, if a rod cannot be moved and the cause of, the fai.lure cannot be determined, the rod could have a failed collet housing.A potentially failed collet housing would be indi.cative of a problem whi.ch could eventually affect the scram capabili.ty of more than one control rod.Since the cracks appear to be of a type which propagate slowly, it is highly unlikely that a second control rod would experience a failed collet housing wi.thin a short peri.od of time after the first failure.Therefore, a period of time of 48 hours5.555556e-4 days <br />0.0133 hours <br />7.936508e-5 weeks <br />1.8264e-5 months <br /> can be allowed to determine the cause of fai.lure.This period is considered long enough to determine, if the cause of fai.lure is not in the drive mechani.sm, yet short enough to be reasonably assured that a second collet fai.lure'oes not occur.Therefore Section 3.1.1.a(2)(Reactivi,ty Hargin Stuck Control Rods)should be expanded to require that if a control rod cannot be moved during normal operation, testing or scram, the reactor shall be shutdown wi.thin 48 hours5.555556e-4 days <br />0.0133 hours <br />7.936508e-5 weeks <br />1.8264e-5 months <br /> i.f the reason that it cannot be moved cannot be shown to be due to causes other than a failed collet housing.I I (b)Ef a control rod drive cannot be moved, t)>e cause of the stuck rod mi.ght be a problem affecti.ng other rods.To ensure prompt detection of any addi.tional control rod drive failures which could prevent movement, Section 4.1.1.a(2) should be expanded to require surveillance every 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> of all partially and fully withdrawn rods if one rod drive is found to be stuck.Unti.l permanent corrective measures are taken to resolve the potential for stuck control rods due to failed collet housings, we believe that these addi.ti.onal speci.fi.cations provide reasonable assurance that an unacceptable number of control rod collet housing will not fail during IJ bi~~

operation.

Upon completion of the'investigations being performed by GE, addition'al corrective actions may permit revision of these requirements.

CONCLUSION We have concluded, based on the considerations di.scussed above, that: (1)there is reasonable assurance that the health and safety of the public will not be endangered by operation in the proposed manner, and (2)such activities will be conducted in compliance with the Commissi.on's regulations and the'ssuance of this amendment will not be inimi.cal'o the common defense and security or to the health and safety of the public.Date: SEP 2 3 i975

~~~~4 1, f~