ML20128F136

Safety Evaluation Re Leak on Core Spray in-vessel Annulus Piping.Plant Can Be Safely Operated for One Fuel Cycle W/O Repairing Observed Leak at Listed 1/4-inch Fillet Weld

ML20128F136
Person / Time
Site:	Oyster Creek
Issue date:	02/05/1993
From:	Office of Nuclear Reactor Regulation
To:
Shared Package
ML20128F115	List: ML20128F110 ML20128F136
References
NUDOCS 9302110263
Download: ML20128F136 (4)

Text

.

• ~w

' p3 ?tg hI

+ o UNITUD STATES I

[*

g NUCLEAR REGULATORY COMMISSION

, g a WASHINGTON, D. C 20565 SAFETY EVALUATION BY THE OFFICE OF NUCLEAR REACTOR REGULATIM RElf, TING TO A LEAK ON THE CQM SPRAY IN-VESSEL AtMILUS PIPING GPU NUCLEAR CORPORATION OYSfER CREEK NUCLEAR GENERATING STATION DOCKET NO. 50-212

1.0 INTRODUCTION

During the in-service inspection of the core spray spargers and piping in the current Oyster Creek refueling outage (14R), a leak was discovered at the inlet downcomer piping (6 inches schedule 40 st inless steel) in the core saray system I (lower sparger). The subject d ncomer alping was located in tie annulus between the vessel wall and core shroud. lie leak was found on a 1/4-inch fillet weld (L-3A) of a downcomer piping coupling. The accessibility to the area of the leak was very limited because the leak location was very_'I close to the vessel wall with a clearance of only about three inches. In examining the leak the licensee (GPUN) employed a very small camera (1 inch in diameter), which was mounted on a special fixture, and positioned the camera directly above the weld with additional lighting provided by General Electric's " Firefly", a remotely operated vehicle (R.0.V.). The-licensee reported that the leak was coming from a 1/8-inch round hole in the fillet weld. No other cracking indications were observed around the hole or at other areas of the fillet weld, The licensee also examined the fillet weld (U-3A) at the same relative location in the core spray system 11 (upper sparger).

There was no leak at weld U-3A; however, there were two linear indications.

The licen.ee concluded that these indications were most likely to be construction defects such as splits or tears between weld passes.

There are two independent core spray sparger assemblies in the Oyster Creek reactor vessel. Each sparger contains 56 nozzles and is fed from an independent penetration through the reactor vessel. The piping from the vessel penetration to the shroud penetration is known as the annulus piping.

A couplingewas c used in the installation of the annulus pi)ing with a 1/4-inch fillet weldfis the final fit up weld to be performed in tie field. The leak was observedrin this field fit up fillet weld (L-3A) of core spray system 1.

Early in 1978 a through wall crack was found in System II (upper) sparger, _ and in 1980, additional ;urface indications were observed in System I and 11 sparger:. The leak observed in 1978 was believed to be resulting frc.a intergranular stress corrosion cracking (IGSCC). A total of 10 mechanical clamps were applied at those locations to ensure structural integrity of the spargers. Air tests and/or visual inspections of core spray spargers and piping were performed during each refueling outage. No additional leaks or indications were found until the current refueling outage (14R).

93.)2110263 930205 PDR ADOCK 05000219 p PDR -

2. DISCUSS 10Mi i

The licensee has determined that the most probable cause for the observed leak in the subject fillet weld would be the opening up of weld defects such as slag and/or lack of fusion after approximately 20 years of service with normal cyclic stresses. This was supported by the following observations:  !

(1) The surface appearance of the subject fillet weld was very rough especially in the area with limited accessibility, indicating poor quality of the weld, which would most likely contain some weld defects.

(2) The observed air bubbling rate was high which indicated that the crack path leading to the leak was not tight. If the cracking was caused by fatigue or intergranular stress corrosion cracking (IGSCC), the bubbling rate should be much less than observed.

(3) The calculated cyclic service loads were small and the cracking by fatigue would not be likely to occur. Even if it did occur, its giowth would not be significant.

(4) Intergranular stress corrosion cracking (IGSCC) was not a plausible mechanism because the observed cracking occurred at the weld metalt and not in the areas of heat affected zones.

However, the staff believes that IGSCC could be a potential dominating mechanism for continued crack growth in the subject fillet weld in view of a through wall IGSCC crack was reported in the core spray sparger piping in 1978. The licensee's 1GSCC crack growth calculations showed that the growth of the crack in the circumferential direction would be approximately 3 inche, in a period of 2 years. This calculation assumed the presence of a residual stresses of 20 ksi and an initial 1/2-inch through wall crack. The licensee estimated the leak rate for a 3-inch through wall crack would be less than 1 gpm, which is much less than the margins that are available in the core spray sparger systems.

The licensee reported that because of the weld size and diameter, a 3/16-inch fillet weld at the lower end of the subject coupling joint would be the '

limiting struct. ural weld. Therefore, the subject leaking 1/4-inch fillet weld could loose approximately 31% of its circumference (about 7 inches) and still maintain th9;r'eqbired axial loading capacity of the coupling. The licensee performed finite < element stress analyses of the subject annulus piping un_ der the conditi'ons.of, n'ormal operation and core spray injection. The combined load consisted of deadweight of the pipe including the water, drag loads due to recirculation flow in the annulus, thermal loads, pressure load during core spray injection and an assumed seismic load of 59 acting simultaneously in the three orthogonal directions. The largest shear stress was calculated to be 5.8 ksi which is much lower than the ASME Code allowable of 19 ksi at 550*F for the subject fillet weld.

With an assumption of a complete failure of the subject 1/4-inch fillet weld, the licensee performed an estimation of the relative axial displacement at the

+ my

. i 1

3 coupling jointJresulting from the blowdown loads due to a LOCA at the recirculation pising system. The calculated displacement was 0.5 inch which is well within tie designed nominal engagement of 1 inch for the coupling.

Based on the-installation drawing and measurements from the field video tapes, the licensee reported that the actual engagement was estimated to be about 0.9-inch. This evaluation concluded that the inlet downcomer piping would not be separated under the postulated LOCA condition even if the subject fillet weld was completely failed.

The licensee performed hydraulic analyses to determine the performance of the core spray system. For the bounding case of assuming a complete failure of the 1/4-inch fillet weld with a 0.015-inch gap around the pipe, the calculated flow loss through the assumed defect is small (46 gpm) in comparison to the available margin of about 500 gpm for the core spray System 1. Therefore, the flow loss due to a complete failure of the subject fillet weld would not affect the intended performance of the core spray system.

During a site visit on January 27, 1993, the staff reviewed video tapes regarding the visual examinations performed at this and previous outages to confirm the licensee's reported observations. The staff also requested some additional.information regarding their evaluation submitted on January 19, '

1993. The additional information verbally provided by the licensee was i summarized below:  ;-

(1) Among all the design basis loading conditions considered, the loading condition of core spray injection and seismic would provide the limiting combined load on the annulus piping.

(2) The design basis seismic load is only 2.09 Based on this seismic load, the required length of the 1/4-inch fillet weld for the limiting load condition (core spray injection and seismic) is 9.5 inches.

(3) With an assumption of a complete f ailure of the subject fillet weld, the axial displacement of the coupling due to a combined load of recirculation piping LOCA and seismic (2.0g) was estimated to be 0.52-inch, which is within the nominal 1-inch engagement of the coupling.

(4) The;11'censee brief.ly described how the reported actual engagement (0'.Sijntli the coupling was derived. It was calculated by subtdctfh)(j of4the' length (3.6 inches) of the upper coupling sleeve, which was mqsured on'the video tape, from the dimension (4.5 inches) specipedeoptheGeneralElectric'sinstallationdra91ng.

The licensee should submit to NRC the above adoittonal information in more detail no later than 60 days after plant restart.

4

~ -g i

. 3,0 QNCLUSIS ^

y Based on our reYiew of the licensee's submittals and the additional information as well as the viewing of the video tapes, the staff concludes that Oyster Creek can be safely operated for at least one fuel cycle without repairing the observed leak at the subject 1/4-inch fillet weld. The intended function of the core spray system would be maintained under the worst design basis condition even if the subject weld is assumed to be completely failed.

The licensee should perform detailed visual examinations and air tests of the core spray spargers and annulus piping during the next refueling _ outage to ensure there is no further degradation of the piping system. When nec9ssary, appropriate surface or volumetric examinations should also be performed to -

investigate the extent of alping degradation. A long-term mitigation program of the core spray system siould be consicered if further degradation of the core spray system is observed.

Principal Contributor: W. Koo Dated: February 5, 1993 e 1,

'); *Y '

sOSi--t

[k hs wy;p-t

+

4 hi rp ai