ML20199F081

Forwards Nine Technical Points Re MSIV Spring Issue for Info & in Support of Recommendations Transmitted on 860618

ML20199F081
Person / Time
Site:	Fermi
Issue date:	06/19/1986
From:	Dennison D GENERAL ELECTRIC CO.
To:	Vaughan L NRC
References
NUDOCS 8606240188
Download: ML20199F081 (3)

Text

F OM GE 488 9237341 d.19.1984 6t?! P. 2 OMPANY kT S3 E E m s82 6 - f 7 _ gc,4,.g

'Jci: Lan y %ug h am f/

W e w =#= (30/) +92 - 7 3 7 o 0 4"-

(so/) H 2 - 22 e r S umer : Fesai A si v s oei~s Xss.>e

/f&ched dre. -the 9 2?ecAuie n /

porJs t<Rc4 Geny SM bu k- du c/

tAe 4 0s, e wi M I oli sc o ssed' om f

you ye s+ ~hy 'Thi S is b ya ev

~

femala>n a ac) s v,p,oo r h -fAe-..

frecom+mcd/uv 4 ycv / u, / /, e 3 wea ,

@vs'c k Com m whec4 tuns f e /c.c.,,;, ,,i f.o yw y.e.sk eclay .

y4 you ho n auy yuca /M y.n. .

we e call on [4'o2) 9a c"; 3302.

1

) .

s vwvan, a

l xf llfN 5 ' ') .St}IEtf .

GENERAL @ ELECTRIC

. s s m ,, w d.,v, .

h00M GE G88 935F301 c.19.1083 <e?? p, 3 5

Basis -

1. Based on the metallurgical study of the cross section araa of the broken spring, the failure is not fatigue related. The overload test of the inner springs that Fermi conducted at the site showed no additional failures. We believe that any spring failures would occur early in the spring life, and is not induced by the cyclic operation. It is also expected that a spring which is tested at overload would provide satisfactory service at normal load.

2. We believe that any midlife spring failures would be random, and individual.

i

3. In the stateline Station MSIV closure test, it was found that over a large portion of the valve closure, the stea.m flow actually pull the disc into the seating position. A failure of some of the springs will reduce the closing force, but will have a minimum effect on the valve closure time.

4. The Fermi FSAR assumed a MSIV closure tima of 10 second for offsite dosage calculation. The MSIV's are adjusted to c1cse in 3 to 5 seconds. We would expect the closing time to be not more than 6 seconds even if all of the inner springs fail, or three outer spring fail. This would have no offcet on the peak fuel cladding temperature. We would not onrect i to see these many springs to fail at one tiraa in the Eame valve.

S. GE analyzed the potential tilt of the spring fitnge id tp to four inner springs and four outer springs from two yo% tcds failed. The deflection is not enough tc, cause the flance to rub against the yoke rods. If there are more sprit.g f2.'?xres in the same valve, the tilt would be lessar. Besi&ar, the flanges are provided with rollers that would provide a qced wear contact between the flange and the spring flar.gs. m nes ,

we would not expect the tilt of the spring flange ta iahjlit  !

i valve closing motion.

i 6. As the valve closes, the gap between the. spcing cet3..i sculd l increase. Hence we would not expect a broken sprinJ coli to I be lodged between coils of another sprirg and prucnt the l complete seating of the valve disc.

7. If some spring fail, the force on the spring flungc ntic *ec uneven. The seating force is certainly radi.ced. Ibeur the

, actuator contributes over 60% of the clo11r.t f'ot c e , r.c t i accounting the force contributed by tne flow s ;: ( s tr . kc would not expect the broken spring to si gnifica n tly aricct

! the local leak rate test result.

1 l

l 1

ROM GE G88 9257301 0 10.1086 0:03 P. O

8. The broken inner a g was determined by testing to be more brittle than we , have expected for the material. We believe that the  : failures are related to the heat treatment and the procedure. Even though we could not isolated the of material that might be susceptible to thic kind of failure, we do not believe it is generie. As the outer spring is made of a larger diameter rod; ie, a )

different heat of material, we would not expect them to be as brittle or as susceptible to fracture. Of all the MSIV's in used in the BWR's, we have only heard of the failures in Fermi and one broken spring in Peach Bottom. If the problem is really widespread, we would expect to see more failures.

9. At this point, we certainly do not expect any spring to fail imminently.

I However we just could not rule out the possibility that some spring might fail in service. While the effect of spring failure is expected to be acceptable, we feel an overload screening test to be prudent and should bc conducted at a scheduled outage. ,

l 1

• l l

.