ML20084N914

AO 50-249/1975-31:on 750626,cracks Discovered on Collet Housings of Control Rod Drives 984,883,1032 & 1099. Apparently Caused by Temp Cycles Drive Experiences During Reactor Scram.Analyses Initiated

ML20084N914
Person / Time
Site:	Dresden
Issue date:	07/03/1975
From:	Stephenson B COMMONWEALTH EDISON CO.
To:	James Keppler NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION III)
Shared Package
ML20084N873	List: ML20084N914 ML20084N925 ML20084N947
References
409-75, NUDOCS 8306090110
Download: ML20084N914 (4)

Text

[Ah

• yy oy y31, N lti.'ml PI q.1._oNC.100. 00anois Address Reply to Post O!!ics Box 767 Crnc go. iii nois 0 /

v.

i EES Ltr. ph09-75 Dresden Nucleai Power Station R. R. #1 Morris, Illinois 60450 July 3, 1975 "

r,~;- . j-C . , , .' ' o !. ,

Mr. James G. Keppler, Regional Director Directorate of Regulatory Operation-Regien III ,

, s .-

J,' -/

U. S. Nuclear Regulatory Co=ission

',4 -

799 Roosevelt Road -

'/

Glen Ellyn, Illinois 60137 - -

SU3 JECT: RE?CRT OF AE:C 'T., CCCURuE::E PER SECTICN 6.6.A 0F THE TECE!! CAL SPECIFICATIC: 3 CRACKS IN CO.',7ECL RCD DRI'.'E CCL'E RCUSZ:GS R;ferences: 1) Regulatory Guide 1.16 Rev. 1 Appendix A.

2) Notification of Region III of U. S. Nuclear Regulatory Co=.icsion Telephone: Mr. P. Johnsen, 0800 hours0.00926 days <br />0.222 hours <br />0.00132 weeks <br />3.044e-4 months <br /> on June 27, 1975 -

Telegram: Mr. J. Keppler, 11h5 hours on June 27, 1975

3) Drawing Number R;; port Nu=ber: 50-249/1975-31 R port Date: July 3, 1975 -

/

Occurrence Datet June 26, 1975 -

Facility: Dresden Nuclear Power Station, Morris, Illinois 60450 IDE:'"IFICATICN OF CCCURRE!CE On June 26, 1975 cracks were discovered on the collet housings of control rod drives 984, 883, 1032, 1099 CCNDITICMS FR!OR TO CCCURRECE Unit-3 was shut down in a refueling outage.

DESCRIF:CN OF OCC'JRRE!CE At approximately 0900 hours0.0104 days <br />0.25 hours <br />0.00149 weeks <br />3.4245e-4 months <br /> on June 26, 1975, while overhauling control rod drive 984, a maintenance forc=an noticed a crack in the collet housing sh' ort tube.

8306090110 750904 PDR ADOCK 05000249 P pDR 0 COPY SENT REGION _ .Nm, -

sm.- ,

< ,Ja=:c, G. Ebppler July 3, 1975

r. .

.* .i

• ' kJ

/

'/ The crack was detectable visually, and was confir cd by = cans ,of a dye penetrant test. Co=menwealth Edisen's Operational Analysis Department (CAD) was contacted to evaluate the crack cn fr954, as well as exa.ine control rod drives 883, 1032, and 1099, which were available for scrutiny. A Level II inspector confimed that cracks were present in the cellet housing of each of the four centrol rod drive =echanisms. In each case, the cracks occurred in the collet housing short tube below the water ports, in the area of increased wall thickness.

General Electric's Nuclear Energy Divisien was also centacted on June 26, 1975.

GE examined their test centrol rod d ive mechanists for cracks similar to those experienced en Unit-3 Cracks were f;und that were nearly identical to those on the four centrol rod drives. A proble= evaluation has been established by GE to study this proble= and dete=ine all possible effects.

Between June 27 and 30, 1975, several = ore centrol rod drive =echanisms were exw%ed for cracks in the collet housing area. As of June 30, GE Level II inspectors'have examined is =echanists using the dye penetrant test. Of these 18 =echanisms,11 have displayed sc=e indication of cracking.

DESIG IATION OF A?NS-" CAUSE CF CCCUSRE"CE General Ilectric has observed similar cracks on test drive mechanisms that have

< been scran-cycled 2000 times, and more severe cracking on mechanisms scram-cycled more than 4000 thes. GE was aware that cracking h'ad occurred, but assumed

the problem was associated with the abnor= ally high number of scra= cycles parformed. There was no indicaticn that cracking would develop within the expected lifetime limit of 200 scrats. The cause of the cracking appears to be related to temperature cycles the drive experiences during a reactor scrac. Alternately, if cooling water is lost or restricted, a thermal cycle will occur when the drive is operated with ner=al drive ficw. At present, General Electric, the Oper-ntional Analysis repartcent, and Argenne Natienal Laborator/ are conducting independent antallurgical studies to determine the cause of the cracking. It has not been definitely dete=ined that the dracks are due to ther=al stress.

/

ANALYSIS OF OCCUWCE .

As de=onstrated by 2000 and 4000 scra= cycle tests conducted by GE, the probability of a total collet housing failure is quite remote. Tne collet

~

housing is not a pressure barrier, but acts as a restraint to contain the collet assembly, experiencing a force of approximately 600 psi during withdrawal, and considerably less during scrams and insertions. Tnis stress is vastly less than the yield strength of the cellet housing metal. If the collet housing ware to fail, the possibifity exists that the collet barrel and spring could jam the collet fingers, reducing the scra= speed or preventing the drive frc= being inserted. However, the probability of a number of drives failing simultaneously is extremely unlikely. Should such an impin ible failure occur, highly localized core damage could result from abnomal rod patterns and power levels.

The standby liquid control system would be available to reduce reactivity and main-tcin the reactor in a shutdown condition. All radioactivity would be contained either within the reactor vessel or the standby gas treatment system. There would, bn no danger to plant personnel or the public.

. James G. feppler July 3, 1975 ,

- , f. . O O '

/ The crack was detectable visually, and was confirmed by =e 'ns .of a dyepenetrant

,/

test. Co=cnwealth Edicen's Operational Analysis Department (CAD) wad /

contacted to evaluate the crack cn 493, as well as examine centrol/ red drives c33, 1032, and 1099, which were available for scrutiny. A Level If inspector confimed that cracks were precent in the collet hcusin; of each df the four control red d ive =echanis=c. In each case, the cracks occurred /in the collet housing short tube below the water ports, in the area of increabed wall thickness.

General Electric's Nuclear Energy Division was also contacte$ on June 26, 1975 GE exanined their test centrol red d-ive mechanis=s for cracks similar to those experienced en Unit-3 Cracks were f;und that were n arly identical to those on the fou centrol rod drives. A proble= evaluation'has been established by GE to study this proble. and dete =ine all possible effects.

/

Between June 27 and 30, '975, several core centrol rod d/iite techanisms were '

exn-Med for cracks in the collet housing area. As of June 30, GE Level II inspectors have exa.ined 15 cechanisms using the dye peh.etrant test. Of these 18 mechanisms, 11 have displayed sc=e indication /of cracking.

/

DESIG'!ATIO:: OF APPA .'D~T CAUSE CF CCCL?2E!CE f l

/

l l General Electric nas observed similar cracks on test drive mechanisms that have been scra .-cycled 2:'00 times, and =cre severe crac'r.ing cn mechanisms cera=-cycled more than LOCO times. GE was aware that crackingjhad occu red, but assumed the problem was ascociated with the abnor= ally hikh nu-ber of scra= cycles perfor=ed. There was no indication that cracking would develop within the expected lifetime limit of 200 scra=s. The cause of the/ cracking appears to be related to temperature cycles the drive experiences during' a reactor scra=, Alternately, if cooling water is lost or restricted, a themal cycle will occur when the drive is operated with no =al drive flow. At,present, General Electric, the Oper-atic::a1 Analysis Department, and Argenne National Laboratorf are conducting independent metallugicalstudiestodete=inethecause/ofthecracking. It has not been definitel;' detemaned that the cracks are due to ther=al stress.

/

/

ANALYSIS OF CCCL?2E:0E

/

/ ,

As demonstrated by 2000 and h000 scra= cy/cle tests conducted by GE, the probability of a total collet housing failure is quite remote. The collet housing is not a pressure barrier, but acts as a restraint to contain the collet assembly, experiencing a force of appro'ximately 600 psi during withdrawal,s and considerably less during scra s and insertions. This stress is vastly less than the yield strength of the collet / housing metal. If the collet housing weretofail,thepossibilityexists/thatthecolletbarrelandspringcouldjam the collet fangers, reducing the scra= speed or preventing the drive frc= being l

inserted. However, the probability j of a nu=ber of drives failing simultaneously is extremely unlikely. Sheuld such an imp 3:sible failure occur, highly localized core damage could result /from abnomal rod patterns and power levels.

The stadby liquid control syste=,would be available to reduce reactivity and main-tain the reacter in a shutdown condition. All radioactivity would be contained either within the reacter vessel or the standby gas treatment system. There would, be no danger to plant personnel or the public.

l i

/

. /'

4

. .a.....

JhriesU.Keppler 9 Quly3,1975

.~

COR.;.x. . r. .r-. r.o .r.--- m.

C.:.ediate corrective action has been to initiate analyses of the cracked material.

Future actiens will be determined by the cutecte of the studies now in prodress.

New housin; asse-blics have been received to replace the cracked assemblies.

A fc11cw-up repcrt will te sub.itted as soon as a d.arirdtive cause can be established.

FAII_L?.E DATA There have been no kno'c. failuras of collet housings to date.

The control red drive =echa.is: is manufactured by the General Electric Cc=pany.

This d-ive techanis is the same type used en all boiling water reactors manufactured b/ General Electric since 1967.  !

1 l

,9 b.~

o I L/ '

A (

L//nj '$//V

e nenson l

.m e Superintendent - l l

BBS:s=p ,

FileAGC e