NUREG-0619, SE Approving Licensee Request for Relief from NUREG-0619 for Feedwater & Control Rod Drive Line Nozzle Insp for Plant

SE Approving Licensee Request for Relief from NUREG-0619 for Feedwater & Control Rod Drive Line Nozzle Insp for Plant

ML20024J182
Person / Time
Site:	Oyster Creek
Issue date:	10/04/1994
From:	Office of Nuclear Reactor Regulation
To:
Shared Package
ML20024J181	List: NUREG-0619, SE Approving Licensee Request for Relief from NUREG-0619 for Feedwater & Control Rod Drive Line Nozzle Insp for Plant
References
RTR-NUREG-0619, RTR-NUREG-619 NUDOCS 9410110137
Download: ML20024J182 (5)

Text

1

==

r

.A 7

E UNITED STATES NUCLEAR REGULATORY COMMISSION

%.g.

/

WASHINGTON, D.C. 20555 4001

.....f SAFETY EVALUATION BY THE OFFICE OF NUCLEAR REACTOR REGULATION

+

RE0 VEST FOR RELIEF FROM NUREG-0619 FOR r

FEEDWATER AND CONTROL R0D DRIVE RETURN LINE N0ZZlE INSPECTIONS FOR GPU NUCLEAR CORPORATION OYSTER CREEK NUCLEAR GENERATING STATION DOCKET NO. 50-219

1.0 INTRODUCTION

1 In response to cracking discovered at a number of U.S. boiling water reactors (BWRs) in the feedwater (FW) and control rod drive return line (CRDRL) nozzles, the NRC issued NUREG-0619, "BWR Feedwater Nozzle and Control Rod Return Orain Line Nozzle Cracking," on November 13, 1980. NUREG-0619 recommended additional non-destructive testing (NDT) beyond the requirements of the American Society of Mechanical Engineers (ASME) Boiler and Pressure Vessel Code,Section XI, for the FW and CRDRL nozzles to assure adequate l

structural integrity.

Depending upon the plant and component design and configuration, NUREG-0619 recommends ultrasonic (UT), visual (VT), and/or dye penetrant (PT) examinations at varying intervals. These examinations, i

particularly the PT, may result in considerable radiation dose to NDT personnel.

GPU Nuclear Corporation (GPUN), the licensee for Oyster Creek Nuclear Generating Station (0CNGS), proposed to eliminate periodic PT examination of the FW and CRDRL nozzles.

A linear phased-array ultrasonic examination technique would replace the PT examination.

However, the licensee has not proposed the complete elimination of PT inspection for FW and CRDRL nozzles.

If UT examination results indicate the presence of a flaw exceeding the ASME Code allowable crack size, a PT inspection will also be completed in the vicinity of the indication to verify the results. The licensee has also proposed increasing the inspection interval for UT examination of these nozzles beyond the one contained in NUREG-0619 to 10 years.

2.0 BACKGROUND

NUREG-0619 indicates that although state-of-the-art UT techniques at time of issuance were not acceptable [in terms of resolution and sensitivity), future developments and demonstrations of the ability of UT methods to detect small thermal fatigue cracks with acceptable reliability and consistency will allow modification of the stated inspection criteria. Significant advances in automated and computer controlled UT techniques for flaw detection and sizing have occurred since issuance of NUREG-0619. These advances allow certain UT techniques the ability to detect cracking that may have gone undetected by formerly used UT techniques.

9410110137 941004 Enclosure DR ADOCK 050 29

4 A switch from a PT to a UT examination would reduce the man-rem exposure, during inspection of these nozzles.

Although the staff agrees that UT methods may be implemented in place of surface penetrant approaches, extension of the inspection interval for routine PT from 6 refueling cycles to an enhanced UT exam once every 10 years requires consideration of three interrelated factors:

fracture mechanics analysis, inspection techniques and capabilities, and component design.

3.0 EVALUATION In order to analytically justify an extension of an inspection interval, the licensee completed a fatigue crack growth (FCG) analysis in accordance with the methodology in Section XI in the ASME Code.

Qualification testing by the inspection vendor under observation by representatives from Electric Power Research Institute (EPRI) Nondestructive Evaluation (NDE) Center demonstrated that the phase-array UT technique can reliably detect and size flaws in the areas of interest.

In addition, previous modifications to the FW and CRDRL nozzle thermal sleeves mitigate one of the factors leading to the initiation of nozzle cracks.

3.1 Fracture Mechanics Analysis of Nozzles A calculated FCG life is particularly sensitive to the initial size of the flaw assumed in the analysis.

The crack length assumed in the licensee's evaluation corresponds to the size of the smallest crack examined during the performance demonstration of the phased-array UT inspection technique.

The stresses imposed on a postulated nozzle crack in the FCG analyses originate from the thermal, mechanical, and pressure loads in the vicinity of the inner nozzle area.

The primary BWR nozzle loadings previously identified under Generic Technical Activity A-10 were a result of transient thermal conditions due to both localized turbulent fluid mixing at the nozzle exit and large variations in nozzle fluid flow.

The mixing phenomena results in localized high cycle fatigue (low amplitude, high frequency) in the nozzle region.

Transient loads as a result of large changes in flow occur less frequently and result in much higher transient stresses in the material (low cycle fatigue).

The configurations of the thermal sleeves in the FW and CRDRL nozzles at OCNGS feature a double flow baffle arrangement which provides an additional barrier separating nozzle bypass flow and the freestream vessel flow.

The baffle plates are spring loaded against the vessel surface to prevent leakage that could establish currents capable of inducing high cycle transient thermal loadings in the nozzle / vessel wall.

Therefore, the high cycle fatigue loading should have negligible impact on the nozzle loading.

Consequently, the primary transient thermal loads are a result of plant transients (i.e.,

startup, shutdown, and scrams) which significantly change the flow rate through these nozzles.

During these transients, large flow rate variations i

through each nozzle will establish a large temperature gradient across the flow baffles and along the nozzle wall, thus inducing thermal stresses in the

s material.

GPUN determined the nozzle stress fields by performing a three.

dimensional thermomechanical finite element analysis of the nozzle area. The licensee then used the resultant nozzle loads to calculate the applied crack driving force and, subsequently, the fatigue crack growth rate.

During a phone call on June 6, 1994, the licensee stated that the peak nozzle thermal stresses were approximately 90 ksi at the surface decreasing to one-third this value at a depth of one inch. The staff used this applied stress field in an analysis to calculate the crack tip stresses and the applied stress intensity.

The staff also determined crack loading using the expressions for stress intensity in Appendix A to Section XI of the ASME Code.

Cyclic crack loading occurs as a result of the various plant transients (i.e.,

startup, shutdown, and scrams).

There is not a one-to-one correspondence between these plant transients and the number of thermal cycles experienced by a nozzle.

For example, during a single startup-shutdown sequence the FW nozzle region might be subjected to numerous thermal cycles due to changing feedwater flow conditions.

For the analysis submitted by GPUN, the licensee elected to model plant transients according to a generic BWR life cycle proposed by General Electric (GE).' This generic duty cycle consists of 130 startup-shutdown cycles and 349 scrams to low pressure over the 40-year life of the plant.

Based on staff review of the operational history for OCNGS, the generic duty cycle assumed for the structural analysis provided by GPUN underestimates the number of startup-shutdown cycles by a factor of two. However, for the feedwater nozzles, the number of thermal transients within each of these startup-shutdown cycles is approximately three times the number that this plant has been experiencing.

Thus, the net result is that GPUN's assumed cyclic loading for startup-shutdown cycles will overestimate the number of thermal cycles for the feedwater nozzle.

GPUN provided no operational history to compare the actual number of scrams or thermal cycles within each scram to that of the GE generic duty cycle.

Due to this absence of information, the staff analysis relied upon the scram-related assumptions inherent in the GE generic duty cycle to account for transient loading induced by a scram.

Section XI, Paragraph IWB-3611, of the ASME Code requires repair if the end of operating cycle crack length from a FCG analysis exceeds 10 percent of the critical (failure) crack size. Ten percent of the critical crack size corresponds to the maximum allowable crack size for service. The critical crack size assumed for the licensee's nozzle analysis corresponds to the assumption in their evaluation of GE's generic nozzle lil analysis.pd this overall wall thickness of the material. The staff prevt" sly review The staff determined that the use of the end of operating cy' *9 crack length equal to one-tenth the material thickness is acceptable for t m application.

'GPUN letter from R.F. Wilson to J.A. Zwolinski dated November 20, 1985.

2Appendix C to NUREG-0619, "BWR Feedwater Nozzle and Control Rod Drive Return line Nozzle Cracking," November 1980.

The staff assumed the minimum detectable crack size as 0.25 inches in conjunction with the calculated applied nozzle stresses and the generic duty cycle to determine the fatigue life for a postulated crack.

Although the licensee has demonstrated that the inspection technique can detect flaws smaller than 0.25 inches, there is some uncertainty regarding the reliably detectable crack size in this nozzle. Therefore, the staff used a somewhat larger initial crack depth in the FCG analysis to provide margin to cover this uncertainty.

The staff analysis used the FCG rate given in Appendix A to Section XI of the ASME Code relating cyclic crack loading and the crack extension per fatigue cycle.

3 The results submitted by the licensee indicate that approximately 12 years of eneration under conditions similar to GE generic BWR duty cycle would be re

'd for the crack to fully extend beyond the acceptable limit.

Thus, at

.g to GPUN's calculations, total crack growth in the proposed in,~. ion interval of 10 years would be acceptable.

In an independent analysis, taking into account a more realistic operational cycle for OCNGS service conditions and a 0.25 inch initial crack size, the staff calculated that the assumed FW nozzle crack would exceed the allowable crack size in 10.3 years.

Both calculations support an inspection interval of 10 years.

The transient loads in the vicinity of the CRDRL nozzle are not as severe as those calculated for the FW nozzle. The overall methodology for the FCG

' sis for CRDRL nozzle cracks conducted by the licensee is consistent with t.

for the FW nozzle.

Results from this analysis indicate that an assumed minimum crack will not propagate to the allowable limit in a period of 40 years.

The staff's evaluation of GPUN's CRDRL nozzle flaw analysis verified that there is a considerable margin between the proposed 10-year inspection interval and the time required for a crack to grow beyond the allowable flaw size.

One conservatism of the licensee's calculation is that the assumed crack depth in the CRDRL nozzle FCG analysis is nearly twice as large as the detection limit demonstrated in UT qualification tests.

The ability to detect cracks nearly one-half the depth of that assumed in the FCG analysis will further decrease the likelihood that a crack will escape detection and propagate beyond the allowable crack size.

3.2 Ultrasonic Inspection Technique and Capabilities In order to show the proposed UT inspection method can reliably detect and size defects in the area of interest, GPUN and the inspection vendor devoted considerable time and resources to qualify the linear phased-array technique.

GPUN enlisted the assistance of the EPRI NDE Center to monitor demonstration of the capability of linear phased-array technique. The vendor initially carried out performance demonstrations on mock-ups of the FW and CRDRL r,vzzles with electron discharge machined (EDM) notches to simulate flaws. While EDM notches can be accurately placed and sized, their geometry significantly differs from service-induced fatigue cracks.

The vendor's technique 3GPUN submittal f rom J. C. DeVine, Jr. dated April 8,1992.

4 successfully located the EDM notches. Later, GPUN implanted actual thermal fatigue flaws in the mock-ups to more realistically simulate actual flaws that could occur in the installed components.

Due to the limited number of actual fatigue flaws used to demonstrate the ability of the UT technique to detect cracks, there is some uncertainty to the proposed minimum reliable flaw size detection limit. Overall, the inspection vendor demonstrated the capability of the linear phased-array technique to detect and size the thermal fatigue flaws implanted in the nozzle mock-ups consistent with the assumptions in the FCG analysis submitted by the licensee.

3.3 Component Desian In response to NUREG-0619, GPUN made modifications to 0CNGS FW and CRDRL nozzles to mitigate or prevent further thermally induced fatigue cracking. As part of the new design, baffle plates were incorporated to direct the flow of cold feedwater so as to minimize thermal effects. Although the baffle plates and particularly the small flow holes are an integral part of the design, they are inaccessible for periodic inspection of the condition of the plates.

Thus, the licensee cannot verify feedrater continues to flow as the component design assumes.

If the baffle plates deteriorate and flow around the baffles was bypassed allowing cold feedwater to directly contact the vessel wall, thermally induced fatigue cracking could initiate flaws. However, growth of these flaws under bypass flow conditions would be self-limiting since the thermal effect of bypass flow is very small at depths greater than about 0.25 inches.

Since the staff assumed an initial flaw size of 0.25 inches, the inspectability of the baffle plates is not a concern.

4.0 CONCLUSION

S GPUN demonstrated that the linear phased-array technique is an acceptable method for examination of FW and CRDRL nozzles. The staff concludes that the licensee may employ the linear phased-array UT technique in lieu of PT examination for this application.

Based on the analysis supplied by the licensee and an independent staff analysis, uncertainty in the value of a reliably detectable flaw size, and the uninspectable baffle plates, the staff approves the licensee's proposal to perform periodic inspections of the FW and CRDRL nozzles with the linear phased-array UT technique at least once every 10-year period (120 months).

Since the licensee performed the last inspection of the FW and CRDRL nozzles using the qualified linear phased-array technique during the 14R outage in 1993, this date will serve as the beginning of the 10-year period for FW and CRDRL examinations.

Principal Contributor:

K. Battige Date: October 4, 1994