ML17191A823
| ML17191A823 | |
| Person / Time | |
|---|---|
| Site: | Dresden  |
| Issue date: | 08/05/1998 |
| From: | NRC (Affiliation Not Assigned) |
| To: | |
| Shared Package | |
| ML17191A822 | List: |
| References | |
| NUDOCS 9808100324 | |
| Download: ML17191A823 (3) | |
Text
UNITED STATES NUCLEAR REGULATORY COMMISSION WASHINGTON, D.C. 20555-0001 SAFETY EVALUATION BY THE OFFICE OF NUCLEAR REACTOR REGULATION RELATED TO REACTOR PRESSURE VESSEL HEAD STUD *
1.0 INTRODUCTION
FLAW TOLERANCE EVALUATION COMMONWEAL TH EDISON COMPANY DRESDEN NUCLEAR POWER STATION. UNIT 2 DOCKET NO. 50-237 By letter dated April 14, 1998, the Commonwealth Edison Company (ComEd, the licensee) submitt~d, for NRC review, a flaw tolerance evaluation by General Electric Corporation (GE) on postulated flaws on reactor pressure vessel (RPV) head studs for Dresden, Unit 2. to support continued op~ration of this plant with a flawed RPV head stud (No. 81). This issue was originated from the wrongful replacement of one of the two flawed studs (Nos. 52 and 81) detected by the licensee during the in-service inspection (ISi) of the RPV head closure studs on March 8, 1998. After the replacement, the licensee discovered that although Stud No. 52 was correctly replaced, Stud No. 81 was left in place while a sound stud (No. 91) was mistakenly replaced by a new stud. The licensee intended to demonstrate using the fracture mechanics analysis that the unit could be operated with one completely fractured stud. The analysis is contained in the report, GE-NE-523-93-0991.
The staff reviewed GE-NE-523-93-0991, and requested the licensee to substantiate an
- assumption regarding the use of the bolt-up tensioning condition as the most limiting load case in.
its fracture mechanics analysis. GE later determined.that the subject assumption was not valid, and replaced GE-NE-523-93-0991 by another report, GENE-813-01920-58, and provided it by letter to the NRC dated June 19, 1998. This safety evaluation (SE) relates to the latter report.
2.0 EVALUATION 2.1 Licensee GE (the licensee's contractor) reevaluated the determination of the limiting load condition by considering the applied stresses for all load conditions analyzed in the "design stress report" of Dresden, Unit 2, for the closure flange region. They are: bolt-Lip tensioning, leak test, and startup conditions. GE then performed a complete fracture analysis for these three operating conditions. This submittal assumed that one stud is completely fractured, and the remaining 91 studs have an edge crack of depth 0.157 inch corresponding to the detection threshold of the end-shot ultrasonic technique (UT) utilized in the ISi. As in the previous report, GE used the formula for an edge cracked round bar to calculate the applied stress intensity factors (applied K's) due to bending, and the formula for an edge cracked plate to indirectly calculate the applied ENCLOSURE
~~9808100324 980805.
~
POR ADOCK 05000237 I
p POR
-.l --
K due to tension. For the fracture toughness, Kie, GE used 130 ksi(in)%, 140 ksi(in)%, and 174 ksi(in)% for bolt-up tensioning, leak test,* and startup conditions. (The first two values are test based and the third is derived from the Rolfe-Novak-Barsom correlation.) The safety factors for postulated flaws were from Appendix G of the American Society of Mechanical Engineers (ASME) Boiler and Pressure Vessel Code (Code): 1.5 for bolt-up tensioning and leak test and 2.0 for startup.
Using the applied K, the fracture toughness, and the safety factor described above, GE concluded that (1) the limiting allowable flaw is 0.39 inch for the most limiting load condition (start-up), (2) the calculated flaw size at the next inspection is.0.21 inch, and (3) continued operation for Dresden, Unit 2, with. a cracked vessel. head stud is justified because the flaw depth at the next inspection is smaller than the allowable flaw depth; 2.2 NRC Staff 2.2.1 Applied Stress Intensity Factor Calculation
- There is no applied K formula in the literature for the RPV head stud cross section. GE approximated the applied K by l!Sing a formula developed by J. H. Underwood based on the experimentally determined_K values by A. J. Bush for an edge crack on a round bar subjected to three point bending. µsing this formula is appropriate because for the allowable flaw siZe of 0.39 inch, the crack depth is only 6.8 percent of the stud diameter, and the corresponding crack front is far away from the inner hole ofthe stud to be affected by it. For the stud stress due to tension, an engineering approach based on the. ratio of the applied K due to tension to the applied K due to bending frorri an edge cracked plate was utilized. GE's approach of using the ratio is acceptable because in the current application, the difference in the component geometry should not affect the ratio of applied K considerably. The staff estimated the error in applied K due to
- tension using GE's engineering approach to be less than 10 percent. The margin (a predicted crack depth at the next inspection of 0.21 inch vs. an allowable flaw depth of 0.39 inch) is large enough to tolerate.this error.
2.2.2 Stress Corrosion Cracking Growth Rate The stress corrosion cracking (SCC) growth rate used in both the previous and the revised GE reports is 6.6x10 _. in/hour. This rate was derived from the two flawed RPV head studs at Dresden, Unit 2, in 1989, and is acceptable to the staff. When compared to published data, the sec growth rate used by GE falls in the upper band of the data range, and is more conservative than the best-estimate value of the published data.
2.2.3 Fracture Toughness - Kie GE.continued to use a Kie value of 130 ksi(in)% for the stud under bolt-up tensioning condition.
However, two new Kie values have been developed in this revised report for the stud under the leak test and start-up conditions. The Kie value of 140 ksi(in)% for the stud under the leak test condition (193°F to 219°F) was based on test data, and appears to be reasonable to the staff.
For the start-up condition, since the limiting stresses occur late in the heatup phase of startup, the stud temperature is very likely to be in the upper-shelf region. The staff agrees with the use of the Rolfe-Novak-Barsom correlation to derive the Kie in the upper-shelf region based on
'* Charpy energy and yield strength. The Rolfe-Novak-Barsom correlation was developed from test results on 11 steels having yield strengths in the range of 11 O to 246 ksi. This correlation has been substantiated by additional tests by other researchers. Since the yield strength of the stud
_ (155.2 ksi) is well above the minimum_ value of 100 ksi suggested for applying the correlation and the scatter of the test data around the Rolfe-Novak-Barsom correlation line is small, the staff accepts the Rolfe-Novak-Barsom carrelation for the current application. Consequently, the Kee:
value of 174 ksi(in)~ derived from this correlation is acceptable.
2.2.4 Fracture Toughness - Kea versus Kie Appendix G requires the flaw tolerance evaluation procedure for bolting be consistent with that in Welding Research Council Bulletin 175. In the section entitled, "Toughness Requirements for Bolting", Kt~ is the only fracture toughness that is discussed there. Therefore, unlike detected
- flaws, where IWB-3600 of Section XI specifies Kia be used as the fracture toughness in the flaw evaluation for the normal and upset conditions, it is appropriate to use Kee as the fracture toughness in the flaw _tolerance evaluation for the bolting under the normal and upset conditions.
3.0 CONCLUSION
The NRC staff has determined thatthe flaw tolerance evaluation is performed in accordance with the procedures and criteria in the ASME Code and the assumed crack growth rate is adequate for this application. Since the predicted flaw size (0.21 inch at the end of the next inspection) is less than the allowable flaw size (0.39 inch), the staff accepts the licensee's flaw tolerance evaluation, and concludes that the RPV with one flawed head stud is acceptable for continued operation without repairs until the end of the current operating cycle.
Principal Contributor: Simon C. F. Sheng Date:
August 5, 1998