ML20207A853

Safety Evaluation Supporting Operation for Full Fuel Cycle W/O mid-cycle Insp for Crack Growth

ML20207A853
Person / Time
Site:	Brunswick
Issue date:	11/05/1986
From:	Office of Nuclear Reactor Regulation
To:
Shared Package
ML20207A848	List: ML20207A846 ML20207A853
References
NUDOCS 8611110341
Download: ML20207A853 (3)

Text

_ _ _ _ _

c

• p us

'o UNITED STATES g

E o NUCLEAR REGULATORY COMMISSION l

?, wasumoTow, o.c.2oses

. ,1

• • ,,,,,* SAFETY EVALUATION BY THE OFFICE OF NUCLEAR REACTOR REGULATION RELATING TO TWO CRACKED N0ZZLE-TO-SAFE-END WELDS CAROLINA POWER AND LIGHT COMPANY l

BRUNSWICK UNIT 2 DOCKET NO. 50-324

1.0 INTRODUCTION

During the Brunswick Unit 21985 refueling outage, cracks were found in the Inconel 182 butter of two recirculation outlet nozzle-to-safe-end welds.

These two welds were not repaired because the cracks in the butter were small and shallow. The staff identified some concerns regarding the licensee's (Carolina Power and Light Company, CP&L) flaw detection and sizing, and crack growth evaluation of the two unrepaired welds. In the SER assessing whether the flawed piping was suitable for an additional fuel cycle of operation, we recommended that the licensee perfonn a mid-cycle inspection to ensure that no excessive crack growth has occurred in the unrepaired welds. Our concerns that were identified in the SER are briefly summarized below:

1. Based on the past experience at other facilities, we considered the detection and sizing of axial cracks in the two unrepaired welds may not have been very accurate and reliable.

2. Because of limited test data, we questioned the use of a plateau crack growth rate to calculate the crack growth in Inconel 182.

3. We were concerned that cracks in the Inconel 182 butter could extend into the reactor vessel nozzle from continued plant operation.

CP&L presented additional infonnation to address our concerns and to support their contention that a mid-cycle inspection was not needed in their submittal dated July 15, 1986.

2.0 EVALUATION CP&L provided additional information regarding the ultrasonic inspection i performed on the two unrepaired welds during the 1985 refueling outage. CP&L reported that the quality of this inspection was better than those performed in the past. More specifically the licensee identified the following two items as factors that would afford more accurate and reliable inspection results:

1. The outside surface of the weld, including the Inconel 182 butter, was machined flush with the pipe and finished to better than 124 rms.

This surface finishing allowed a more accurate sizing of the cracks because the transducer was not decoupled and the accessibility was not limited during ultrasonic inspection.

G

I , .

2. State of-the art inspection techniques, developed under a General Electric /EPRI contract were used for the subject examination. Ten different search units with a variety of frequency (1 MHZ to 4 MHZ),

angle / mode (45 degrees shear, O degree L and 0 to 70 degrees RL) and focal distance (30 to 60 mm) were used to confirm and characterize the indications.

Fatigue is considered the most likely mode of crack growth in the low alloy steel nozzle material _because the low alloy steel is not considered susceptible to IGSCC. GE performed a fatigue crack growth calculation assuming 20 startup/

shutdown events in one fuel cycle. The calculated fatigue crack growth was less than 0.2 mil in one fuel cycle. GE also performed calculations for a limiting axial crack length in the low alloy steel material based on an arrest fracture toughness of 200 ksi (in)g The results of the analysis are that the throughwall axial crack would not become unstable until its length exceeds 30 inches. Since the crack would be so long before instability would occur, a large safety margin against unstable rupture exists since the flaw would be

, identified from leakage well before it reached 30 inches.

GE provided detailed crack growth data for Inconel 182 developed by GE and others. The presented crac for crack growth of 4.5x10inch g growth

/ hour data up to appeared a K value of to 42 support ksi (in)afor boundigg 0.2 value ppm oxygenated water. This bounding value is considered conservative because the data were obtained from load c Therefore, GE used a plateau crack growth rate of 5x10-gntrolled inch / hour toCT tests. the end-of-cycle calculate

crack size in the two unrepaired welds, assuming the crack growth rate would not increase much at larger stress intensity factor (K). The calculated final crack size at the end of an 18-month (12000 hours) period is well below the Code allowable flaw depth of 1.55 inches.

The staff has reviewed the above stated information and has determined the following :

I

1. The ultrasonic inspection perfonned on the two unrepaired welds was a much better and thorough inspection than that performed in the past.

Therefore, the inspection results are considered as reliable as a current state of the art in ultrasonic examination would permit.

I

2. We still have reservations regarding GE's crack growth calculation using a plateau crack growth rate. However, our concern in this matter is greatly reduced because of the additional confidence we have gained regarding reported crack sizes in the two unrepaired welds. The reported crack sizes are short and shallow axial flaws.

Therefore, even if these cracks did grow throughwall, they would not I impact the structural integrity of the nozzle to safe-end weld and the loss of coolant would be small.

3. Based on the results of GE's analysis, it appears that any crack i growth in the low alloy steel nozzle material resulting from either IGSCC or fatigue would not be significant.

l t _____

.y .

4.0 CONCLUSION

Based on the above evaluation, we conclude that Brunswick Unit 2 can continue to operate a full fuel cycle of 12000 hours without the need of performing a mid-cycle inspection.

Principal Contributor: W. Koo Date: Noveanber 5,1986 i

- - - - -- , ._, _ _ .