ML17191B354
| ML17191B354 | |
| Person / Time | |
|---|---|
| Site: | Dresden  |
| Issue date: | 05/18/1999 |
| From: | NRC (Affiliation Not Assigned) |
| To: | |
| Shared Package | |
| ML17191B353 | List: |
| References | |
| NUDOCS 9905260045 | |
| Download: ML17191B354 (4) | |
Text
UNITED STATES NUCLEAR REGULATORY COMMISSION WASHINGTON, D.C. 20555-0001 SAFETY EVALUATION BY THE OFFICE OF NUCLEAR REACTOR REGULATION RELATED TO JET PUMP RISER EXAMINATION RESULTS AND FLAW EVALUATION FOR DRESDEN NUCLEAR POWER STATION. UNIT 2 COMMONWEAL TH EDISON COMPANY DOCKET-NO. 50-237
1.0 INTRODUCTION
By letter dated July 10; 1998, the Commonwealth Edison Company (ComEd, the licensee) submitted, for NRC review, its jet pump examination results and the associated flaw evaluation for the flaws detected at Dresden, Unit 2. The in-vessel*visual inspection (IWI) of the 10 jet pumps was conducted during the D2R15 refueling outage in accordance with the guidelines of*
report BWRVIP-41, "BWR Jet Pump Assembly Inspection and Flaw Evaluation Guideline~." The.*
scope of the inspection also included five welds (designated as RS""1 to RS-5) on each jerpunip riser (JPR). The inspection results indicated that a 1.5-inch indication exists along the heat
.. affected zone of the.riser elbow at the RS-1 weld* of jpR 15/16. The licensee submitted an analytical fla\\Y evaluation to demonstrate *that the unit could be operated without repair for two fuel cycles.
. 2.0. BACKGROUND Based on the detected flaw length of 1.5 inches on the JPR, the licensee assumed that the
. indication was a through.,wall crack and performed a flaw evaluation. The licensee employed the limit ioad analysis consistent with the latest Appendix C (1996 Addenda) of Section XI of the American Society ofMechanical Engineers (ASME) Boiler and Pressure Vessel Code (Code).
The limit load analysis used a safety factor of 2.77 for Normal and Upset and 1.39 for Emergency and Faulted conditions, a Z-factor for shielded metal arc welds (SMAW), -and a botJnding intergranular stress corrosion cracking (IGSCC) growth rate of 5X10-5 inch/hour. The normal load includes dead weight, hydraulic loads, flow induced vibration,* and thermal loads.
The faulted load includes the normal load plu_s two additional loads: safe shutdown earthquake inertia and fluid drag during a postulated suction side recirculation line break loss-of-coolant accident (LOCA).
- Com Ed also considered the fatigue crack growth due to flow induced vibration (FIV). under the normal condition and determined it to be insignificant.
The licensee added the crack growth corresponding to two fuel cycles, or 32,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> of operation (0.8 inch per crack end per cycle), to the detected crack length (1.5 inches), and obtained the final crack length (4.7 inches). On the other hand, the licensee calculated the allowable crack length through limit load analysis and found its value to be 17.9 inches. Since.
the predicted crack -length at the end of two fuel cycles is less than the allowable crack length
~905260045 990518 I
PDR ADOCK 05000237 G
- - )i with an adequate margin, the licensee concluded that the observed flaw in the weld RS-1 of JPR 15/16 does not pose a.safety concern for the next two fuel cycles.
- The licensee estimated the leakage to be 11 gallons per minute (gpm) using the Bernoulli equation for incompressible flow, where the flow loss coefficient was assumed to be 1.0 and the crack area was assumed to be rectangular. The Bernoulli equation (Equation 1) is as follows:
- where Q i!? leakage, C is flow loss coefficient, A is crack area, *P 0is fluid density, and ~P.is
- pressure difference across the pipe. The crack opening displacement that was used in the
- crack area (A) calculation was based_ on linear elastic fracture mechanics (LEFM) analysis from
- EPRI Report NP-2472, Volume 2.
3.0 EVALUATION 3.1 Limit Load Analysis of the 1996 Addenda The licensee performed a flaw evaluation for the detected flaw at the RS-1 weld of JPR 15/16.
The details of the evaluation methodology can be found in the report, GE-NE-523-B13-01869-054, "Jet Pump Riser Weld Flaw Evaluation Handbook for Dresden Unit 2 arid Unit 3," attached to the' licensee's submittal. The flaw evaluation was based on the limit load analysis of the latest.
Appendix C (1996 Addenda) of Section XI of the ASME Code.. P.reviously, the Code required the user to set the pipe diameter in the Z-factor equations to 24 inches for any pipe having a diameter less than 24 inches. The 1996 Addenda has removed this additional conservatism based on recent results f~om extensive pipe fracture experiments on austenitic welds using the submerged arc weld (SAW) and SMAW welding processes as documented in NU REG/CR-
. 4878.* The staff has accepte~ the limit load analysis of the 1996 Addenda.
3.2 Flaw Evaluation The staff evaluated the licensee's flaw evaluation and determined that the limit load analysis, which gave an allowable flaw length of 17.9 inches, meets the rules of the ASME Code (1996 Addenda) and, therefore, is acceptable. The initial flaw length* is 1.5 inches, which includes the
. uncertainty associated with EVT-1 examinations using a ruler (BWRVIP-03); After adding two cycles of IGSCC growth of 3.2 inches, the predicted flaw length becomes 4. 7 inches. Since the predicted flaw length at the end of two cycles (4.7 inches) is less than the allowable flaw length (17.9 inches), the staff has accepted the licensee's flaw evaluation.
- Also, to further support the determination that FIV does not contribute to fatigue growth, the licensee provided additional information in a letter dated November 2, 1998. This information indicated that the FIV crack length corresponding to the threshold ti.K of 5 ksi(in))i is 6.5 inches.
Since the predicted crack length after two cycles is 4.7 inches, less than the threshold FIV crack length of 6.5 inches (FIV would contribute to crack growth after the crack is grown to 6.5 inches) the staff agrees with the licensee's conclusion that th~re will be no fatigue growth due to FIV during these two cycles.
3.3 Leakage Calculation Evaluation Although Bernoulli flow is a slrnple model for the flow through the crack, the licensee used the equation very conservatively. First; the flow loss coefficient was assumed to be 1.0. Hence, the pressure losses due to phase change, area change, and friction loss associated with surface roughness were neglected. This would overestimate the leakage as indicated by Equation 1.
Second, using a rectangular crack area is more conservative than using the elliptic crack area based on LEFM. Since th.e predicted crack length is much smaller than the critical crack length based on limit load analysis, it is appropriate to use LEFM to predict the crack opening
- displacement. -.Using the larger rectangular crack area, again, overestimated the leakage as indicated by Equation 1. Basecj on the above, the staff has accepted the licensee's leakage calculation methodology.
3.4. Leakage Effects Evaluation Jet pump assemblies were not designed to meet the ASME Code.. However, they are classified as safety-related components since the structural integrity of the jet pump assembly is relied upon for assuring the ability to reflood the core, up to two-thirds core height; following a design basis accident (OBA) LOCA. An additional safety funct.ion of the jet pump assemblies at
- Dresden, Unit 2, is to provide a flow path for low.pressure c_oolant injection.(LPCI) flow into_the core.
ComEd's evaluation of the effect of potential leakage through th~ cr.ack indication assumed a
- through-wall crack with a length equal to predicted flaw size at the end of two fuel *cycles~ The predicted flaw size of 4. 7 inches and a width of Jess than 0.5 mil resu.lted in a tot_al leakage* of 11 gpm. Although this leakage is not significant w.ith regards_ to total recirculation flow, reduction of core cooling capability due to the leakage must be* considered.
ComEd considered the decrease in LPCI flow during the two most limiting DBA's, single failure LPCI (SF-LPCI) and single failure diesel generator (SF-DG). The bounding SF-LPCI case for' Dresden, Unit 2, is a recirculation suction line break iri combination with a failure of the LPCI loop select logi¢. The only low pressure emergency core cooling. system (ECCS) pumps available for the SF-LPCI case are two core spray pumps. The SF-LPCI case assumes that the loop select logic selects the*:broken loop such that all four LPCI pumps are pumping out the break. In this case, jet pump leakage is not a coricem since LPCI is assumed to be lost. The staff notes that the SF-LPCI case results in a peak clad temperature (Pen of 2021 degrees Fahrenheit. According to the licensee, this PCT is 122 degrees Fahrenheit.higher than the SF-DG case without the potentialleakage from the crack indication on the jet pump riser.
For the SF-DG event, reduction in ECCS flow is caused by the failure of the LPCI minimum flow valve to close. The only low pressure ECCS pumps available for the SF-DG case are two LPCI pumps and one core spray pump. With the minimum flow valve open, the degradatiQn of LPCI flow could be 706 gpm for the SF-DG case. According to the licensee, the SF-DG case PCT
.. - ---~-__,... -
... --- r------* ---
-~
I'.
.e
-.4 -
without the potential leakage from the indication in the jet pump riser is 1899 degrees *
- Fahrenheit.
With the potential leakage from the crack indication (11 gpm) and the 706 gpm LPCI degradation, the estimated PCT for the SF-DG case was 1900 degrees Fahrenheit. Similar to the discussion above, the.combination of leakage from the indication in the jet pump riser and the SF-DG case, the resulting PCT is approximately 121 degrees Fahrenheit lower than the PCT resulting from the SF-LPCI case.* Therefore, the staff concludes that potential leakage from the jet pump riser crack indication will not have an impact on the limiting PCT design basis case for the next two fuel cy.cles.
4~0 CONCLUSIONS The staff has reviewed the licensee's submittals. The staff determined that the flaw evaluation meets the rules of the ASME Code and the ass.urned crack growth rate is adequate for this.
application. Since the predicted final flaw length at the end of two cycles (4.7 inches) is far less than the allowable flaw length (17.. 9 inches) from the limit load analysis, the staff determines that Dresden, Unit 2, can be operated without repair of the identified flaw for two fuel cycles. The staff concludes that the licensee's proposal to reexamine the flaw according to the guidelines and schedule of BWRVIP-41 is acceptable. The staff also concludes that the licensee's leakage evaluation is adequate and that operation in the proposed manner for two fuel cycles with the calculated jet pump leakage meets the requirements of 10 CFR 50.46 and is acceptable.
Principal Contriqutors: c: F. Sheng K. Kavanagh Date:
May 18, 1999