ML20244A510

Final Significant Deficiency Rept SDR-L2-89-39 Re Failure of Stainless Steel Midlock Ferrules.Caused by Isolated Stress Corrosion Cracking Not Affecting Other Penetration Ferrules in Penetration 20JX222 Replaced

ML20244A510
Person / Time
Site:	Limerick
Issue date:	06/07/1989
From:	Kowalski S PECO ENERGY CO., (FORMERLY PHILADELPHIA ELECTRIC
To:	NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM)
References
SDR-L2-89-39, NUDOCS 8906120051
Download: ML20244A510 (6)

Text

.. _ _ _ _ _ _ _ _

h c.

PHILADELPHIA ELECTRIC COMPANY 2301 M ARKET STREET

. [

P.O. BOX 8699 F'HIL ADELPHI A. P A.19101 i:t s) e414 502 jl S. J KOW ALSKI I

vict PREsiotMT seu cks an a ssosais a n eses I

U.S. Nuclear Regulatory Conmission 10CFR50.55(e) a Attn: Document Control Desk Washington, D.C.

20555 June 7, 1989

.i Docket No.: 50-353 CPPR-107

SUBJECT:

Limerick Generating Station, Unit 2 -

ql)

Final Significant Deficiercy Report, Failure of Stainless Steel Midlock Ferrules

/

L '

REFERENCE:

Telecon of May 22, 1989, R. J i.ees (PECo) to H. Williams (NRC) 1 I L FILE: QUAL 2-10-2 (SDR'L2-89-39) q Gentlemen:

/.cl '

By the referenced telecon, Philadelphia Electric Company (PECo) Inforry>d e

l,f you of a reportable condition under 10CFR50.55(e) regarding failure of stainless steel mldlock ferrules in a single containment penetration (20dX222).

/

he are reoviding this final report as indicated in the May 22 telecon.

h fl, A

!. I We have performed a review of penetration records and conducted metallurgical evaluations of ferrules from the failed penetration. Additionally, ffrrules

)

b['I[J sarrpled f rom other penetrations were also evaluated by PECo. Conax, the manu-

.,I facturer, has perfonned field investigation artl testing of the Limerick Unit. 2

.y %

midlock ferrules. On the basis of these reviews, we conclude that the failure

<.. i/

of penetration 20dX222 midlock ferrules was due to an isolated incidence of i

y']

utress corrosion cracking not affecting any other penetration ferrules. We have

)

replaced all ferrules in penetration 20dX222, and it successfully passed a leak

'fl rate test. Our evaluation (contained in the enc 1, re) and the completed

( L M, f corrective action resolve this identified signific-sit deficiency.

i_s ill 1, If you have any further questions at this time, please contact us.

l l l"/m,,'a

. ll Sincerely, lI j< ;

6906120051

!)

E906

{DR ADOCK0500g7 jl 3

{ l}l

,/d.btMh

/,

Enclosure

/

4-1

/ l l%

f&A tl tW'/d!m/002

/

i

ii l cc:

W. T. Russell, USNRC, Administrator, Pegion I

[;(fl f

T. J. Kenny, USNRC, LGS Senior Resident Inspector

,,f R. J. Clark, USNRC, LGS Project Manager

./ y' 1 i

/, [(

t f

e ifJ

_1.___.

___.____________2.________

i c

p i

bec:

G. M.'Leitch LGS MC 200 4

8 L. B. Pyrih N2-1 R. J. Lees-N2-1 G. A. Hunger, Jr.

52A-5 Chesterbrook~

~j E. J. Bradley S23-1 W. J. Boyer,_dr.

N2-1 A. S. McAinsh LGS MC SB3-4 M. S. lyer (Bechtol) 1

d. F. O'Rourke N2-1 H. D. Honan N2 R. B. Rock N2-1 P. J. Duca, Jr.

LGS-MC AS-1 M. A. - Miller (Tenera) -

N2-1 D. P. Helker/Conmitment Tracking 52A-5 DAC '.

NG-8

-l

)

l 1

)

'1 l

l

. 1

Enclosure Linerick Generating Station, Unit 2 Significant Deficiency Report Evaluation of Failure of Stainless Steel Midlock Ferrules Description of Deficiency Linerick Unit 2 experienced failure of 10 of 14 midlock ferrules of elec-trical containment penetration 20dX222. These ferrules did not neet leak rate' test criteria due to stress corrosion' cracking (SCC) which caused leaks beyond acceptable limits or linear Indications.

Figure 1 depicts a penetration con-talning midlock ferrules.

Safety Inpilcations Penetration pressure boundary sealing between inside and outside containment is provided by nechanical compression ferrule design. The ferrule is a three-plece Interlocked arrangement' held in place by a threaded midlock cap. The three-plece ferrule consists of a nain seal between the header plate and electrical feedthrough and two seals which are part of a_ leak rate nonitoring and testing system for this nain' seal and the penetration feedthrough end seals.

Fleid Investigation and testing by Conax Buffalo Corporatio'n (the manufacturer) 1 of ferrules at Limerick Unit 2 concluded that the cracks did not' create a-

)

leakage path from inside to outside' contalmnent (Reference.1).

This is because

]

the cracked ferrules were part of the leak rate test systen in all cases and did l

not affect the nain' seal which purpose is to nelntain the contalmnent pressure -

]

boundary. Conax also reported that the cracked-ferrule condition is the first

']

instance of its type in 15 years of use of this ferrule design in thousands of Installations (Reference 2).

l l

4 Corrective Actions All fourteen midiock ferrules of containment penetration 20dX222 were replaced.. In order to confi nn that the failure mechanism was~ caused by -

l SCC two independent laboratories perfonned metallurgical evaluations-of the removed sanples, i

The test laboratory evaluations' by PECo (2 ferrules) and Conax (11 ferrules)

Independently concluded that the midlock ferrules renoved from containment penetration 20dX222 falled due to' SCC (References 3 and 4).

Certain characteristics of these ferrules nede than susceptible to SCC.

For instance, chemical analysis on the ferrules verified that the;naterial is 303 stainless steel (SCC-susceptible reterial), as.specified by the manufacturer. - Also, iron-oxide' dark stains _were present on some ferrules, Indicating that the parts had been wetted at sone tine.

Bechtel and PEco Engineering have reviewed the PEco 'and Conax lab reports and concur that SCC caused the midlock ferrules to fall the leak rate test.

l 1

i

A review of the penetration records T1.e., storage, instal'ation, 1

testing, etc.) did not identify any unique events wh:ch could have re-suited in the failures or in the wetting of the ferrules (Reference 5). A

{

condition did exist where the feedthrough support system through the containment nozzle was out of alignment with the sealing parts of. the header plate by approximately 7" to 10% We do not believe that this. misalignment contributed to the SCC.

Eleven ferrules fr om five additional ' Unit 2 penetrations were renoved an'd sent to the PECo test lab for metallographic examination, chemical analysis, I

~

i and hardness testing'.to determine whether indications of similar SCC existed, Three ferrules were renoved from penetration 20dX230A (the only other wetwell penetration besides 20dX222). The remaining eight ferules were reroved from four drywell penetrations. The ferrules were removed from plugged ports which did not require de-termination and re-termination of electrical connections.

j The plugged port ferrules are installed and treated the same as feedthrough j

ferrules.

]

A sampling inspection approach was utilized using MIL standard 105D. The test lab results indicated zero cracks or defects in all ' sample ferrules examined (Reference 6). The inspection of ferrules from five penetrations of-a total population of 24 penetrations without Identifying any defects consti-tutes an appropriate sample size per MIL Standard'105D.

1 Actions to Prevent Recurrence On the basis of the evaluations of the cracked ferrules and the ferrules I

sampled from other penetrations, we are abic to conclude that the cracking

)

problem is not appilcable to any other penetrations in Units' 1 or 2.

We believe that something unique and random occurred to penetration 20dX222 during its manufacture, shipping, storage or installation which caused SCC.

Normally, the penetration ferrules are kept under a pressurized nitrogen blanket which prevents contamination and corrosion. ' During installation, the blanket may be temporarily renoved per manufacturer procedures. The event which caused contamination and ultimately SCC of the ferrules in penetration 20dX222 cannot be detenuined. However, in the unlikely event that such an occurrence should be repeated, normal surveillance testing of penetration pressure gauges and long-tenn leak rate testing (LLRT) during refueling outages will detect any leaks of ferrule and feedthrough seals.

I l

i t

I; 1

l

___ =

A

A 1

f

.g.

r

. ~,

1 References 1.

Trip report from P. A. Christopher, Conax Buffalo Corporation discussing field investigation of leaking penetration 20dX222 dated April. 27, 1989.

_l 2.

Letter from Thomas P. Schaefer, Conax Buffalo Corporation to Ant i Kar, Bechtel Power Corporation,_ " Electric Penetration No.

20dX222, Limerick-Generating Station," dated May 25, 1989.

3.

Memorandtm from J. F. Delong, PECo to A. S. MacAinsh, PEco,

" Metallurgical Examination of Limerick Unit No. 2, 303 Stainess Steel Ccopression Couplings to Determine the Cause of Failure, Metallurgical Laboratory Note No.89-624," dated May 4,1989.

4.

Transmittal from T. P. Schaefer,_ Conax Buffalo Corporation to Glenn Frank 1in, Bechtel. Construction, Inc., "Fallure Analysis of Three (3) Piece Ferrules," dated May_4, 1989.

]

1 5.

Menorandtm from M. S. Iyer, Bechtel Power Corporation to E. R. Nelson, Bechtel, " Interim Report, PECo Finding 2E-554," dated May 25, 1989.

]

i 6.

Memorandtm to R. B. Rock, PECo, from J. F. Delong, PECo, j

" Metallurgical Examination to Determine the Presence of Cracks in

.J Eleven Conax Ferrels Taken from Limerick Unit No. 2, Metallurgical Laboratory Report 89-624," dated May 30, 1989.

t l

l

)

l l

- L A

g 4

@3i

~

t C $

d 1'

o 5

sgg 55 8

m s

hg

-E 6,

s w

e a

gg.

e g

gis ge E'

en a

e

>g w

e se e

w

.o C

h h

h5 EE k

,J.k T

P3 g

e BEW

$$y, 9

4l#

l

'c Je 3

2 sEEh se 5 8 He:

al 0k ZZ

@' Nb gln, 5lll< jets!ES E!!!

=

+[V 99 Be l E5

} 6.n 9

ll UU 10 ct

2. 2 e

a

+

+

4 Sf.

Qy qq f,/,,

^ '

?#g.j.

w

/7 N

/

s.,

i wm 1

rp lll rh 1

4"'

e AA d

/L\\

L_w

/,\\

F wg 7x w

e i

9

\\

O

~ k..:.hh

.5

/

D ~s R o x

oo g*j

~Q~4 h< 0 37{.L 5:

\\

Q f$

E k f.sg &

E ds\\

Q

/

L v

w f

a v

g ~ad,f ?if U'

'x O

v._c=>

/

\\

x

.s 1.y s.

- f,,\\\\

.t I -

i 1,*_sS, a

,g.

W[ q

, l y a.

//

nl

'>=4 l\\

4 rM

/<7 aW

\\

/

~

~

l

%... g 2

1

.m

,x y=.,

_-~ ~ ~. q;

~. :-(

g.

, & YY

\\f

~

p

,e

/.,

/'tby

.cM===?

)i t;/h, g.

^

s J'?

/

s-2e n.

.5 xO q/m{Q,/

..i 4 9 9 ( k@,

N a

4 R

x e

J $ g ! M=

N'4 g

--- y 3

i 4

=

N' 3

~

=

,/,

__ ' ~._

j.g

'Q g

j1.,

g

)/'

']1;,

.t k

.i O,

j.