ML20056F356

Safety Evaluation Granting Licensee 930607 Request for Relief from Certain ASME Code Requirements for Temporary Repairs to Svc Water Sys

ML20056F356
Person / Time
Site:	Beaver Valley
Issue date:	08/12/1993
From:	Office of Nuclear Reactor Regulation
To:
Shared Package
ML20056F355	List: ML20056F354 ML20056F356
References
NUDOCS 9308270024
Download: ML20056F356 (4)

Text

- .-- . - - _

ENCLOSURE

[ o g UNITED STATES ,

n NUCLEAR REGULATORY COMMISSION g E WASHINGTON, D. C. 20555

%.....) SAFETY EVALUATION BY THE OFFICE OF NVCLEAR REACTOR REGULATION REdVESTFORRELIEFFROMASMECODEREPAIRRE0VIREMENTS FOR SERVICE WATER EXPANSION JOINTS DUOVESNE LIGHT COMPANY BEAVER VALLEY POWER STATION. UNIT NO. 2 DOCKET NO. 50-412

1.0 INTRODUCTION

By letter dated June 7, 1993, Duquesne Light Company, the licensee, requested relief from the ASME Code Section XI repair requirements .for two leaks in a moderate energy class 3 service water system (SW) at Beaver Valley Power Station, Unit No. 2. The leaks were detected in two 30-inch diameter metal expansion joints (MEJ). There is one affected joint in each of the two independent SW system supply trains. The licensee proposes to apply a temporary, reversible, non-Code repair to each joint until a Code repair can be performed later this year.

Each joint is located in its own independent SW train valve pit which is outside the plant, between the screen house and the plant building. The SW ,

system conveys ambient temperature Ohio river water at a design pressure of 150 psig. Each MEJ is constructed of type 304 stainless steel corrugated bellows with 150 psi rated carbon steel flanges. The bellows are  !

unreinforced; there is no external wire braid. The rest of the SW system is I unlined carbon steel.

The leak on the "B" train MEJ was discovered first. The maximum leakage rate observed since discovery is estimated to be less than 1 gallon per hour.

During the subsequent augmented inspection of the SW system, the leak on the "A" train MEJ was discovered. This joint is the sister joint to the first.

The weep rate for the "A" joint has been observed at one drop per 5 minutes.

The leakage is suspected to be caused by erosion and/or corrosion / pitting in the region of the dissimilar metal weld joining the stainless steel bellows to the carbon steel flange. Type 309 stainless steel weld metal was used. The bellows material fits into the flange in a socket joint like configuration.

However, unlike a socket fitting, the weld was executed from inside the pipe, presumably to preclude creating a wetted crevice in the annulus-between the bellows and the carbon steel flange. The observed leakage on both MEJs is from this crevice. Only one weld on each MEJ is leaking. In both cases,-it is on the discharge end of the joint.

Due to this weld configuration, the flaw cannot be observed visually. The licensee has been una~ o le to detect any flaws by ultrasonic testing (UT).

Thus, the flaws cannot be characterized as to location, type, size, or cause.

Flaw evaluation, using analytical methods such as found in Generic Letter 90-05 is consequently impossible. However, because of the very low leakage rates, the licens,ee has assumed that the flaws are..small.

9308270024 930812 .

PDR ADOCK 05000412 'L P PDR-

' \

4 -

2-4 2

Because of the MEJ locations in the SW trains, the leaks cannot be isolated a for repair without removing their respective SW trains from service. The j licensee has estimated that repair / replacement would require longer than the Limiting Conditions for Operation period (72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br />) for one SW train. Thus, a I unit shutdown is required to effect Code repairs.

As a compensatory measure, the licensee proposed application of a reversible mechanical restraint on each joint. The restraint is a segmented ring (six dogs) which fits between two convolutions of the bellows. The six dogs are i bolted to the flange studs. With this arrangement, the dogs prevent relative movement between the end of the bellows and the attached flange. This restraint would prevent separation of the bellows from the flange in the event i '

of a postulated guillotine break of the bellows to flange weld. This restraint would not significantly affect the functioning of the bellows.

, In addition to the structural reinforcement provided by the six dogs, the 4

licensee proposed application of a glass-epoxy circumferential patch at the bellows-flange junction. This would effectively create an encapsulating fiberglass pipe at the leak location to provide added strength and prevent gross leakage. There would be no effect on the function of the bellows.

These temporary repairs are proposed to be employed until the next refueling 1

outage which is scheduled to begin in September 1993. At that time a Code repair / replacement would be performed along with a root cause analysis. If a planned outage of 30 days or more occurred before then, the joints would be repaired during that outage.

An augmented inspection of the SW system was performed. Two other joints of

the same design and similar operational impact have been identified and were not leaking. Since the cause of the two leaks is uncertain, the licensee i committed to monitoring the other two joints once a week for the duration of

the proposed temporary repairs.

2.0 DISCUSSION Due to the unknown cause of the two observed leaks, a flaw analysis is impossible. In its place, a consequence analysis was performed to investigate the effects of a postulated guillotine break at the expansion joint weld. The consequence analysis included investigation of the ability of the restraining dogs to withstand a guillotine break, piping system reactions, flooding of '

equipment, and loss of flow.

For the postulated guillotine break, seismic and operational loads at the joints were analyzed. Nearby pipe anchors are capable of withstanding all reaction forces involved in a postulated guillotine break. The loads on the expansion joint are minimal. Its only apparent function is to provide for thermal expansion. There are no pipe reaction forces transmitted across it.

The pipe anchors absorb all reaction forces.

, . , , - - - , - - , , -y,-, ., . .# y ..,,-,- .n ,,--,,,,w -

,,-t ,-, , -r ,c,

i

• l From this analysis, the licensee concluded that the forces acting upon a '

presumed flaw in the bellows to flange weld are small. Thus, flaw growth, and increasingly larger leak rates are not anticipated to occur as a result of a stress driven mechanism. Should the cause be erosion or corrosion, the staff I expects that a slow leak rate increase may be anticipated. The licensee has committed to a twice daily visual inspection of the two leaking joints. l Leakage will be estimated and trended. Any leakage beyond the rate of 1 '

gallon per minute (gpm) for either joint will be reported to the NRC. l The six restraining dogs bolted to the flange joint are designed to retain the end of the bellows inside the flange joint. This compensatory action provides .

mitigation for a structural weakness in two ways. First, the six restraining dogs will act to reinforce the weld joint. This will reduce the stress levels  !

at the weld and slow crack growth, assuming a crack-like flaw. Second, should  !

degradation of the joint be due to another mechanism, such as erosion, the restraining dogs will still prevent separation of the bellows from the flange.

This would minimize the loss of flow and flooding of other equipment. '

Additionally, this modification will not prevent the bellows from performing its normal function. l i

i With respect to the application of the leak mitigating fiberglass / resin wrap, the staff cannot determine if it would provide any substantive structural l reinforcement. This is due to the unknowns regarding the shear strength of l the resin to stainless steel bond. Thus, no credit for its application would l be warranted. 1 Given a failure of one of the bellcws welds, the installed " dogs" would limit the amount of leakage to about 300 gpm which would allow approximately 20 i minutes to isolate the leak before the isolation valve became inundated. This  !

would leave the redund ut train to bring the plant to safe shutdown conditions. Even if the bellows was not isolated, the 300 gpm leak rate would not prevent safe shutdawn because only the equipment in the vault would be affected. The flooding conditions in the vault would not cause any of the valves in the vault to change position and isolate safety related cooling loads. The loss of only 300 gpm would not prevent the affected header from receiving adequate flow to perform its safe shutdown function. If it is assumed that both trains' bellows fail simultaneously which is the worst-case scenario, the plant could still be brought to safe shutdown conditions using one or both trains of the system with or without isolation of the leakage. If both bellows are isolated cooling would be lost to the primary component cooling water (CCW) system and the secondary plant CCW system. At Beaver Valley 2, this would not prevent safe plant shutdown. Without isolation, both trains of the SW system could still be used to supply all cooling water loads; however, one train would have to be isolated to affect repairs. Based on the capability of safe plant shutdown being maintained, and the impracticality of performing the Code repairs without imposing the burden of the plant transients associated with an unnecessary plant shutdown and startup, the request for relief should be granted.

l . 4.

1

3.0 CONCLUSION

The staff concludes that the proposed. temporary non-Code repair provides reasonable assurance.of structural integrity for a postulated guillotine break of the expansion joint weld and that the capability of safe plant shutdown is maintained. Based on these findings, and the impracticality of performing the Code repairs without imposing a burden on the licensee, relief should be granted pursuant to 10 CFR 50.55 a(g)(6)(1).

An administrative limit of 1 gpm per joint must not be exceeded without

authorization from the NRC.

l Principal Reviewers:

l G. Hornseth l W. Lefave j G. Edison i

j Date: August 12, 1993 i

)

i 1

l l

l l

. _ . . . . . . . . . . . , _ . . _ _