ML20199E298

From kanterella
Jump to navigation Jump to search
Insp Rept 99901300/97-01 for Framatome Technologies,Inc,On 970519-0930.No Violations Noted.Major Areas Inspected: 10CFR21 Program & Reactor Vessel Weld Chemistry Data
ML20199E298
Person / Time
Issue date: 01/28/1998
From: Correia R
NRC (Affiliation Not Assigned)
To:
Shared Package
ML20199E236 List:
References
REF-QA-99901300 99901300-97-01, 99901300-97-1, NUDOCS 9802020123
Download: ML20199E298 (17)
v • d • e


Text

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

4 l U.S. NUCLEAR REGULATORY COMMISSION i

OFFICE OF NUCLEAR REACTOR REGULATION 4

Report No: 99901300/97 01

. Organization: Framatome Technologies, Inc.

4 Location: 3315 Old Forest Road-l Post Office Box 10935 l Lynchburg, Virginia 24506-0935

Contact:

K.E. Moore, Manager

, Materials & Structural Analysis Unit l (804) 832-3277 4 Nuclear Industry Engineering and material support services to owners / operators of Activity: -

nucleai power plants with nuclear steam supply systems or reactor pressure vessels designed and supplied by portions of the i_ former Babcock & Wilcox Company purchased by Framatome.

Dates: Onsite: May 19-21,1997, continuing review through September 1997 Team Leader: Stephen D. Alexander, Reactor Enginaer, HQMB

Technical Lead
Barry J. Elliot, Senior Materials Engineer, EMCB Inspectors: - Meena K. Khanna, Materials Engineer, EMCB
Andrea D. Lee, Materials Engineer, EMCB Kenneth J. Karwoski, Materials Engineer, EMCB Approved by
Richard P. Correia, Chief Reliability and Maintenance Section Quality Assu'ance, Vendor Inspection and Maintenance Branch Division of Reactor Controls and Human Factors Enclosure 9802020123 900128 PDR GA999 ENV*****

99901300 PDR

1 INSPECTION

SUMMARY

On May 19-21,1997, the NRC conducted an inspection at the Lynchburg, Virginia, engineering facilities of Framatome Technologies, Inc. (FTI), to review information rele'ed to the fabrication by Babcock & Wilcox Company (B&W) of reactor pressure vessels (RPVs) for domestic nuclear power plants. B&W Nuclear Technologies, who maintained this information, was purchased by the Framatome Technologies Group (FTG) and became FTG's affiliated company, FTI. The inspectors found some discrepancies and learned that there were some data that had not been '

reported in B&W topical reports. The inspectors identified the need for additional information from FTl that was provided in separate letters to the NRC subsequent to the onsite portion of the inspection. After completing their review of the inspection data and subsequent vendor submittals, the inspectors were able to determine that even when taking previously unreported data into account, most of the chemistry factors in the NRC's Reactor VesselIntegrity Database are still conservative. Licensees have been notified of the few cases involving low-copper weld wire heats in which inclusion of additional data could affect chemistry factors. They are expected to assess the impact of the additional data and take appropriate action in accordance with applicable regulations.

With regard to FTI's program for compliance with 10 CFR Part 21, the inspectors found that FTI's procedures adopted pursuant to 10 CFR 21.21(a) had some deficiencies that amounted to a minor violation. The inspector also reviewed FTl records relating to 10 CFR Part 21 evaluations and found one instance in which FTl did not meet the time requirements for notification of affected licensees or purchasers in accordance with 10 CFR 21.21(b). In accordance with the NRC's enforcement policy as promulgated in NUREG-1600, the procedural deficiencies and the missed time requirement (because of minor safety significance) will be treated as a minor violation of 10 CFR 21.21 and no notice of violation will be issued.

[

The inspection bases were as follows:

l

  • Appendix B," Quality Assurance Criteria for Nuclear Power Plants and Fuel Reprocessing Plants," to Part 50 of Title 10 of the Code of Federal Regulations (10 CFR Part 50),
  • 10 CFR 50.60, " Acceptance Criteria for Fracture Prevention Measures for Lightwater Nuclear Power Reactors for Normal Operations"
  • 10 CFR 50.61, " Fracture Toughness Requiremerts for Protection Against Pressurized Thermal Shock Events"

o Generic Letter 92-01, Revision 1, and its Supplement 1

~

2 STATUS OF PREVIOUS INSPECTION FINDINGS No previous findings relating to FTl or B&W were addressed during this inspection.

3 INSPECTION FINDINGS AND OTHER COMMENTS 3.1 10 CFR Part 21 Program

a. Inspgetion Scope In conducting a routine review of FTl's program for compliance with 10 CFR Part 21, the inspector reviewed the latest effective revision, Revision 10, dated April 16,1997) of Framatome Technologies Group (FTG), Inc. (FTI's parent company), Corporate Policy Statement 0401, " Reporting of Defects and Noncompliances Conceming Substantial Safety Hazards," and the latest effective revision (Revision 26, also dated April 16,1997) of FTl Administrative Procedure (Quality Assurance) 1707-01, " Processing Safety Concerns." In addition, the inspector reviewed FTI's records of safety concern evaluations back through 1992.
b. Qbservations and Findinas The inspector found that Policy Statement 0401 indicated that it was applicable to Framatome Technologies Group (FTG), Inc., and its affiliated companies, including FTI.Section IV.A of Policy Statement 0401 stated that it was FTG's policy that FTG and its affiliated companies should " adopt appropriate procedures" to accomplish the requirements of 10 CFR 21.21(a),

which it listed correctly. Section 21.21(a) of 10 CFR Part 21 sets forth the requirements for the content of procedures that are required to be adopted pursuant to the regulation.Section IV.A of 0401 also required adopting procedures to comply with 10 CFR 21.21(b) regarding informing '

affected licensees or purchasers of deviation or failures to comply. While 10 CFR Part 21 does not at the present time explicitly require this provision in procedure 1 adopted pursuant to the regulation, it may in the near future, and, more importantly, it does iequire that Section 21.21(b) be followed when applicable; therefore, the inspector found the inclusion of this provision to be prudent.Section IV.B also prudently specified the requirements for compliance with 10 CFR 21.21(c) governing the NRC reporting process, although this section of Part 21 is also not presently required to be covered by Part 21 procedures. Likewise,Section IV.C of 0401 covered the provisions of 10 CFR 21.31 regarding procurement documents and also 10 CFR 21.6(b), alternative posting requirements, including the prescribed notice as an attachment to the policy statement.

Administrative Procedure 1707-01 contained the detailed provisions for implementing FTG Corporate Policy Statement 0401 and complying with Part 21 requirements, it assigned responsibilities for meeting the various Part 21 requirements by title and delineated the mechanics, including forms, etc., for processing what it called Preliminary Reports of Safety Concerns, or "PSCs." Paragraph Vlli.A defined a safety concem in general terms, broad enough to encompass most things that might be considered deviations or failures to comply as defined in Part 21. However, the inspector identified several weaknesses in the procedure.

First, the procedure did not relate safety concerns to deviations or failures to comply or explain 3

deviations or failures to comply as defined in Part 21 (or state that included with safety concerns, specifically, are deviations and failures to comply) such that knowledgeable staff could decide if a given safety concern actually constituted a deviation or failure to comply in a

" basic component" (also not explained) as defined by Part 21, that has been delivered or offered for use at an NRC-licensed facility, and thus requires evaluation in accordance with 10 CFR 21.21(a)(1). The procedure also referred to time limits in Part 21 for processing safety concems when the term safety concem is not used in Part 21.

Second, the procedure required an evaluation of safety concems, but did not explain that the evaluation should determine (a) if the safety concern was a deviation or failure to comply as defined in Part 21, and (b)if the deviation was a " defect" as defined in Part 21 or if the failure to comply was associated with a " substantial safety hazard," also as defined in Part 21. The ins-pector was concerned that without reference to or explanation of the terms used in Part 21 (and also in Policy Statement 0401), Procedure 1707-01 might allow some deviations or failures to comply to go unevaluated. For example, certain departures from technical requirements in procurement specifications (deviations) might not be recognized as a safety concern as defined in the procedure. Hence, those deviations might not be properly evaluated to determine if they constitute defects, i.e., deviations that could creato substantial safety hazards or lead to exceeding a license technical specification safety limit.

Third, the procedure stated that safety concerns that were to be transferred to customers would be transferred after management concurrence with the " final report" instead of within five working days of determining that FTl would not be able to perform the 10 CFR 21.21(a)(1) eva!uation as required by 10 CFR 21.21(b). Presuming "the final report" to mean the report of the group performing a 21.21(a)(1) evaluation of a PSC (what should be related to a deviation or failure to comply), the language of the procedure effectively allows information on deviations and failures to comply to remain unreported to affected licensees or purchasers until the 60-day evaluation time limit has expired, or the evaluation time as extended by an interim report when FTl may have known it was not going to perform an evaluation much earlier in the process. If the procedure had made it clear that at any time after the initiation of a PSC, one disposition of the PSC could be a transfer of the information to the customer, and at the point when that determination is made (documented, for example, by some kind of memo or entry in the PSC file on that date), the five-working-day time period starts, then compliance with 10 CFR 21.21(b) would be better assured.

The inspectors review of FTI's records of evaluations of safety concerns revealed that Rosemount Instruments had sent a letter to Bailey Meter Company (a B&W/McDermott subsidiary) dated August 31,1992, informing its customer of a problem with Rosemount differential pressure transmitters (static pressure effect on span). As a result of a Bailey Meter letter to B&W Nuclear Technologies (as FTl was then known), dated October 15,1992, informing B&W of the problem, B&W initiated PSC 4 92 on October 29,1992 recommending that B&Ws affected customers, the Tennessee Valley Authority (for Bellefont), the Washington Public Power Supply System (for WNP-1) and Consumers Power (for Midland), be notified of the problem (this would have been a 10 CFR 21.21(b) notification). Although the B&W Part 21 procedure corresponding to FTI's 1707-01 in effect at the time, BWNT-1707-01, Revision 23, dated October 1,1992, generally reflected the revision to 10 CFR Part 21 that became effective on October 29,1991, the procedure contained essentially the same language as FTI's 1707-01 4

regarding transferring safety concern information to customers. Apparently as a result, the file for PSC 4 92 contained another memorandum, dated November 19,1992, indicating that the concern would be transferred to customers and distributhg the preliminary evaluation report in house for comment. It requested that comments on the customer notification be returned in one week. Comments were received and the final evaluation report was routed for engineering and QA concurrence on December 1,1992. The PSC was finally closed out on December 9, 1992, and the affected utilities (as well as those with non-affected, B&W-designed operating plants) were notified of the concern by Letter ESC-1047, dated December 16,1997. The notification of affected licensees and purchasers occurred within five working days of management concurrence with the final evaluation report as prescribed by the procedure.

However, this was not consistent with the requirement of 21.21(b) (or its intent) because affected licensees or purchasers were not notified of the deviation within five working days of the time (November 19,1992) at which the file clearly indicated that the cognizant staff within BWNT had determineo that the PSC would not be evaluated.

c. Conclusions Procedure 1707-01 referenced Policy Statement 0401 as well as Part 21 itself and NUREG-0302 (contains some interpretation guidance). However, Procedure 1707-01 was the only working procedure adopted pursuant to 10 CFR Part 21 and was supposed to stand alone.

Therefore the inspector concluded that with the weaknesses identified, the procedure might not always be effective in ensuring compliance with 10 CFR 21.21(a), and, although compliance with 10 CFR 21.21(b) is not currently required to be covered by procedures, the language of the procedure if followed without reference to the regulation could allow violation of 10 CFR 21.21(b). Accordingly, the inspector determined that these weaknesses constituted a minor violation of 10 CFR 21.21(a). In accordance with the NRC Enforcement Policy promulgated in NUREG 1600, the matter was bought to the cttention of FTl and is discussed here in the report, but no Notice of Violation is being issued.

With regard to the failure of B&W in 1992 to notify affected purchasers of a deviation (that B&W did not intend to evaluate) within the time limit prescribed by 10 CFR 21.21(bt this occurrence would constitute a Level-IV violation of 10 CFR 21.21(b), except that in this particular case, the deviation was of minor safety significance. This was because (1) the affected plants of BWNT's affected purchasers were not operating, (2) the file indicated that Rosemount had also informed affected utilities, and (3) Bailey Controls had determined that the problem had a minor impact on safety. Therefore, in accordance with the criteria in NUREG-1600, it is considered a minor violation and no notice of violation will be issued.

3.2 Reactor Vessel Weld Chemistry Data

a. Inspection Scope in order to obtain and validate weld chemistry data from available sources for domestic B&W-fabricated, submerged-arc welds in pressurized water reactor (PWR) and boiling water reactor (BWR) vessels, the inspectors reviewed information at FTl to:

P 5

1, Establish the traceability for weld chemistry data points by comparing them to the weld metal qualification (WMQ) test reports, chemistry laboratory work requests, or other fabrication records.

2. Review B&Ws fabrication records to verify general consistency with topical reports that were submitted to the NRC.
3. Obtain all available data in order to assess the impact of including these data in best-estimate chemistry determinations.

Regulatory Guide (RG) 1.99, Revision 2, " Radiation Embrittlement of Reactor Vessel Materials," provides a method for calculating the amount of radiation embrittlement. RG 1.99, Revision 2, indicates that the amount of radiation embrittlement is dependent upon the amount of copper, nickel, and neutron fluence. It also provides a method for determining the amount of radiation embrittlement from surveillance data. In topical reports BAW 1500 and BAW-2121P, the B&W Owners Group (B&WOG) reported the chemical composition of the Linde 80, submerged-arc welds used in fabricating the 177-Fuel Assembly OG reactor vessels BAW-1500 and BAW-2121P identified the chemical composition of welds fabricated using the following weld wire heat numbers: 8T1762,299L44,72105,406L44,8T1554,821T44,61782, 71249,72102,72442,72445,1P0962, T29744,8T3914,1P0661, and 1P0815.

As a result of the NRC Staff review of licensee responses to Generic Letter 92-01, Revision 1, a comprehensive database was developed to compile and record summaries of the materials properties of the reactor vessel beltline materials for each plant. This database is called the Reactor VesselIntegrity Database (RVID). It should be noted that the data in the RVID came from pressurized thermal shock and pressure-temperature limit assessments, from surveillance capsule reports, and in responses to staff requests for additional information (RAls) as well as in responses to generic letters. All of the heats identified above are included in the RVID. In addition, a few heats that were identified in the RVID, but not reported in BAW-1500 or BAW-2121P, include what are called " low-copper" weld wire heats, comprising heat numbers 442002, 442011, H4498, 31401, and 1084-18. Low-copper welds are, by definition, ones in which the weld wire had no copper coating. There may be trace amounts of copper present from other sources, but none from weld wire coatings.

As part of the process of verification of the RVID, the Materials and Chemical Engineering Branch (EMCB), Division of Engineering, Office of Nuclear Reactor Regulation, identified the need to review the chemistry data for the heats of submerged-arc wald wire material used by the B&W in fabricating RPVs used in domestic nuclear power plants. Accordingly, at FTl, the EMCB inspectors reviewed the recorded data for weld wire heats 299L44,72105,71249, 61782, 72445, 406L44, 72442, 821T44, 8T1554, 72102, 8T3914, T29744, 8T1762,1 P0962, 1P0661 1P0815, 442002, 442011, H4498, 31401, and 1084-18.

b. Findinas and Observations The chemical composition data used by FTl to determine the best-estimate chemistry for each heat came from four potential sources. These sources included: (1) nozzle beltline dropouts, (2) surveillance weld samples, (3) WMO tests. and (4) reactor vessel beltline welds (e.g.,

6 t

) .. ..

m Midland) The inspectors verified that FTl collected and reviewed all available material chemistry data, including results of the previous studies performed on weldments manufactured by B&W. While performing this rev;ew, the inspectors noted a weakness regarding the traceability of the recorded test data. As a result of discussions with the inspectors, FTl agreed to provide documentation that contained information to assist the inspectors in establishing / verifying traceability of each of the heats to its respective recorded test data.

The inspectors noted that the data and calculations for the majority of the heats, were recorded in an acceptable manner. However, the inspectors identified some data inconsistencies in FTI's records. The inspectors found that FTl had not transferred some of the data to its " master list,"

when determining the best-estimate chemistry, in addition, the inspectors noted that FTl had not identified suspect data, where applicable. During the review of " raw" data from original laboratory or weld metal qualification test reports, the inspectors considered some data points to be suspect, but these had not been identified as such by B&W or FTI. For example, data from tests on Zion Unit 2 weld WF 209-1, specimens W-40A and W-40B, had been omitted from the suspect data section in FTl's " master list." The inspectors noted that many weld wire heats had suspect data. Therefore, the inspectors requested detailed information which explained why the data were considered to be suspect and whether the data should have been included in determining the best-estimate chemistry.

Subsequent to the on site portion of the inspection, FTl submitted the additional information in its letter INS-97-2262, dated June 6,1997. FTl addressed suspect data, provided traceability documentation, included the additional data, and determined the effect of the additional data on the mean values of copper and nickel.

By letter INS-97-2450, dated June 19,1997, FTl submitted the available data for the non- ,

copper-coated weld wire heats used with Linde 80 flux in fabricating the low-copper welds. In the enclosure to that letter, FTl included the best-estimate chemistry values for low-copper l welds. The inspectors had requested the information because of apparent discrepancies (i.e.,

values of copper too high for ostensibly low-copper welds) they found in some of the reported values for welds which had been identified as having low-copper content values. Also, in this letter, FTl committed to provide the additional data to the members of the B&WOG.

On July 10,1997, as stated above, a conference call was held between FTl and the inspectors to discuss: traceability, correction factors, misprints and/or data not matching the " master list,"

and reasons for not including missing or suspect data in determining the best-estimate chemistry. These issues were resolved during the conference call and by FTI's letter, dated July 10,1997.

Detailed Observations and Findinas for Each Coooer-Coated Weld Wire Heat of Interest Table 3, included after the discussions of copper-coated weld wire heats below, compares the amount of copper and nickel in Linde 80 welds, fabricated with copper-coated weld wire, that were reported (1) by the licensees, (2) in B&WOG Reports BAW-1500 and BAW-2121P, and (3) from FTI's evaluation of all traceable data. FTI's estimates of the chemical composition were derived from the average of all"non-suspect" traceable data and from a coil-weighted average for copper. The coil-weighted average is determined by taking the sum of the products 7

__n

of the average amount of copper from each sample end the number of coils used in the fabrication of the sample, and dividing this sum by the number of coils to fabricate all sample welds. Note that FTl's estimates of percent copper using the coil-weighted averages were less than or equal to the best estimate values reported by licensees, except for weld wire heat numbers 821T44,72442, and 72445; For these heats, the coil-weighted average was 0.01 percent greater than the FTl best estimate values derived from the average of all the data.

Where the coil-weighted averages exceed the best-estimate values derived from averaging all data, the NRC will notify affected licensees and request that they assess the impact on a plant-specific basis.

b.1 Heat 8T1762 FTI's best-estimate chemistry for weld wire heat 8T1762, based on averaging the individual data points, was 0.19 weight percent for copper and 0.55 weight percent for nickel. The inspectors independently assessed the data and verified th'at FTils best-estimate chemistry calculations were correct. Nevertheless, the copper and nickel values reported by licensees and contained in the RVID (0.20 and 0.55 wt-% respectively) were greater than or equal to the average from all traceable data. Therefore, there were ultimately no concerns with weld wire heat 8T1762, b.2 Heat 299L44 The inspectors identified additional data for weld wire heat 299L44 that resulted from FTI-sponsored "round-robin' testing. The round-robin testing was conducted by several laboratories on specimens from the same welds to evaluate error sources associated with analysis methods and practices, including equipment, calibration, analyst perception and judgement where applicable, data reduction, and laboratory practices. FTl committed to include the data for these subject welds in its assessment of best-estimate chemistry. FTl submitted this information to the NRC with its June 6,1997, letter. The inspectors verified that the additional data were added to the existing data for weld wire heat 299L44 and found the values to be acceptable. Although the inspectors noted that the standard deviation for the mean copper remained unchanged, and the standard deviation for the mean nickel for heat 299L44 increased slightly (from 0.03 to 0.04 weight percent), the inspectors confirmed that the copper and nickel values reported by licensees for 299L44 and contained in the RVID were greater than or equal to the average from all traceable data. No further concems were identified with this heat.

b.3 Heat 72105 The inspectors identified additional data that had not been reported for weld wire heat 72105.

Also, review of the supplemental information (original data) for this heat provided by FTl in its July 10,1997, letter revealed several apparent typographical errors in the B&W topical report (BAW-1500). For example, BAW-1500, Exhibit B-6 listed the copper content for weld WF-209-1 (heat 72105) as 0.40 weight percent. The retest value for copper reported in Table C-2 of BAW-1500 was listed as 0.49 weight percent. The WMQ test report indicated that the correct value is 0.40 we,ight percent. Among the data not previously reported were suspect / rejected data for weld wire heat number 72105. The chemical analyses of three Zion Unit 2 samples 8

4 fabncated with weld wire from this heat were rejected since it was believed that the specimen was notched in the base metal resulting in erroneous chemistry values (Zion 2 WF-209-1: 1 specimen with ID of W-40A, and 2 specimens with ID of W-408). Several WMQ test data points were considered suspect and, therefore, the data were rejected, including: (a) the nickel value for WF-70 WMQ from Mt. Vemon WMQ Test Lab No. 6595; (b) the copper content for WF-113 from Mt. Vemon WMQ Test Lab No. 7277; (c) the nickel content for WF-209-1 from Mt.

Vernon WMQ Test Lab No.10029; and (d) the copper and nickel content for WF-353 from Mt.

Vernon WMQ Test Lab No.14433. A total of 5 copper data points and 6 nickel data points were rejected for weld wire heat number 72105.

However, the data reported for heat 72105 in FTI's letter dated July 10,1997, accurately reflected the data identified by the licensee and the inspectors. FTI's be?t-estimate for the chemistry of heat 72105 was based on averaging the individual data points. The inspectors verified that 'he calculations were correct. This resulted in 0.32 weight percent for copper and 0.57 weight percent for nickel.

The inspectors also assessed the best-estimate chemistry, for heat 72105, by averaging the mean value from the individual sources of data. The 12 sources of data, if the data points discussed above are exciuded, include: (1) WMQ test for weld WF-70, (2) WMQ test for weld WF-113, (3) WMQ test for weld WF-209, (4) WMQ test for weld WF-209-1, (5) nozzle belt dropout for Midland 1, (6) beltline weld for Midland 1, (7) reactor vessel surveillance program (RVSP) from Oconee 2, (8) RVSP from Oconee 3, (9) RVSP from Crystal River 3, (10) RVSP from Midland 1, (11) RVSP from Zion 1, and (12) RVSP from Zion 2. The resultant values from this "mean of the means" approach are listed in Table 1. The inspectors also assessed the impact of including the data points excluded from the database (discussed above) on the "mean of the means" chemistry. For this case, another source of data was added (i.e., the WMQ test for weld WF-353). These values are also listed in Table 1. The results indicate that the copper and nickel values reported in RVID for heat 72105 are conservative. Therefore, the inspectors had no further concerns with weld wire heat 72105.

TABLE 1: MEAN of MEANS for HEAT 72105 Suspect Data Excluded Suspect Data included

! Copper Nickel Copper Nickel Mean 0.323 0.578 0.301 0.561 Standard Deviation 0.049 0.018 0.083 0.043 (sample) b.4 Heat 821T44 FTl provided a comprehensive listing of chemistry data for weld wire heat 821T44. FTI's est' mates of the percent copper, using the coil-weighted averages, were less than or equal to the best-estimate values reported by the licensees, except for weld wire heat numbers 821T44, 72442, and 72445. For these weld wire heats, the coil weighted average was 0.01 weight 9

percent greater than the best-estimate values derived from the average of all the data.

However, the inspector's review of the data for weld wire heat 821T44 indicated that the copper and nickel values reported by licensees and contained in the RVID are greater than or equal to the average from all traceable data. Therefore, the slightly higher coil-weighted average for copper notwithstanding, there were ultimately no concerns with wald wire heat 821T44.

b.5 Heat 72442 FTl provided a comprehensive listing of chemistry data for weld wire heat 72442. Although for 72442, the percent copper by coil-weighted average was 0.01 weight percent greater than the best estimate value from averaging 611 the data, the inspector's review of the data for weld wire heat 72442 indicated that the copper and nickel values reported by licensees and contained in the RVID for 72442 are equal to the average from all traceable data. Therefore, there were no concerns with weld wire heat 72442.

b.6 Heat 72445 FTl provided a comprehensive listing of chemistry data for weld wire heat 72445. FTI's best-estimate copper content values, derived from the average of all data, as well as by coil-

,g weighted average as stated above, was found to be 0.01 weight percent greater than the best-estimate values reported by the licensees, for weld wire heat number 72445. However, for this b heat, the average amount of nickel was less than that reported by licensees. The lower value for nickelin this case offset the higher copper values such that the chemistry factor remained the same. Therefore, there were no concerns with heat 72445.

b.7 Heat 71249 The inspectors identified additional data for weld wire heat 71249 resulting from the round-robin testing. FTl committed to include the data in its assessment of best-estimate chemistry. FTl submitted the information to the NRC with its July 10,1997, letter. Also included with this letter were data points from several sources cited in a letter from Florida Power & Light Company (FP&L), dated February 10,1984. However, these data were not included in the database for this weld because they could not be verified or traced. These data included: (a) 5 data points from Westinghouse, (b) at least 2 data points from the Heavy Section Steel Technology (HSST) program, (c) 1 data point from Point Beach 1 (WCAP 8743), (d) 1 data point from Zion 2 (SECY 82-465), (e) 1 Oconee data point (SECY 82-465), and (f) 1 data point from Ginna. Of the 5 Westinghouse data points, some may be duplicates of the values listed in the June 30,1977, and April 11,1977, letters from FP&L included and referenced in the Jun+ G,1997, FTl 'etter.

, The inspectors noted that the data points from Point Beach 1 and Ginna n.ay have been duplicates of the WMQ test report data from weld SA-1101; although this could not be verified with the available data. However, the possible duplication in this case is inconsequential because the data were not used.

The four copper and nickel values for heat 71249 reported from the National Bureau of Standards (NBS) analysis of specimen 62W (Oconee 1, SA-1101) are the mean of four analyses (i.e., there are four groups of data, each with 4 analyses, for a total cf 16 measure-ments). However, the chemistry values from these four groups may not have been corrected 10

_ _ _ _ . . . J

for calibration procedures and for spectralinterferences. Nevertheless, the overall average of copper and nickel of the 16 measurements took these " correction" factors into consideration as indicated in a July 24,1985, letter from John A. Norris of NBS to Evan Morgan of B&W. These corrections to the " raw measurements" appeared to be insignificant, as evidenced by a comparison of the average value of copper from the " raw measurements"(0.172 weight percent) to the " corrected measurements" (0.170 weight percent).

A total of 17 copper and nickel data points were not included in the best-estimate chemistry for heat 71249 because the WMQ test reports were not available. These included 7 specimens from analysis of HSST- 62W [62W-309,62W-359,62W-202. 62W 223 (2 specimens),62W 276 (2 specimens)) and 10 specimens from analysis of HSST-61W [61W 232,61W-276,61W-246, 61W-225,61W- 222,61W 270 (4 specimens), and 61W-234).

The inspectors noted that not all of the available data were listed in the FTl letter dated July 10, 1997, although the letter contained all of the data where traceability could be established. FTI's best-estimate for the chemistry of heat 71249 was based on averaging the individual data points. The inspectors verified that the calculations for these values were correct (0.24 weight percent for copper and 0.61 weight percent for nickel). In addition, the inspectors assessed the impact on the best-estimate chemistry of including the previously rejected data points. The result was the same mean value of copper, a lower mean value for nickel, and a higher sample l standard deviation for both.

i The inspectors also assessed FTl's best-estimate chemistry for heat 71249 by averaging the mean values from the individual sources of data. The 10 sources of data, when the data points discussed above are excluded, included: (1) WMQ test for weld SA-1094, (2) WMQ test for weld SA-1101, (3) WMQ test for weld SA-1229, (4) WMQ test for weld SA-1344, (5) WMQ test for weld SA-1706, (6) WMQ test for weld SA-1769, (7) RVSP from Westinghouse, (8) RVSP from Turkey Point 3, (9) RVSP from Turkey Point 4, and (10) nozzle belt dropout (NBD) from Oconee 1. The resultant values from this "mean of the means" approach are listed in Table 2.

The inspectors also assess 0d the impact of including the data points excluded from the database (discussed above) on the "mean of the means" chemistry. For this case, two sources of data were added (i.e., RVSP from the HSST program and a NBD from the HSST program).

These values are also listed in Table 2.

TABLE 2: MEAN of MEANS for HEAT 71249 1

Suspect Data Excluded Suspect Data Included Copper Nickel Copper Nickel Mean 0.234 0.591 0.234 0.592 Standard Deviation 0.049 0.036 0.048 0.036 (sample)

FTl's best-estimate nickel content values were less than or equal to the values reported by the licensees, except for weld wire heat number 71249. However, for this heat, the average 11

amount of copper was less than that reported by affected licensees, which resulted in a lower chemistry factor. In addi ion, the mean copper content for heat 71249, determined from the WMQ test data and the .10Zzle belt dropout data, were significantly lower than the mean determined from the RVSP. This indicated that there was considerable variability in the copper content within this heat of material. However, the results indicated that the copper and nickel values reported in the RVID for heat 71249 are conservative.

The inspectors confirmed that FTl added the data from Florida Power & Light (FP&L), for weld

= wire heat 71249 to the data tables, as appropriate.

b.8 Heat 72102 For weld wire heat 72102, the WMQ test report for weld SA-1187 indicated that the WMQ test failed. The reason the test failed was not detailed. Although the copper and nickel values reported for this WMQ test were used to estimate the mean value for the total population of welds using weld wire heat 72102, the values were low to medium range and had a negligible effect on the mean. For heat 72102, FTI's best-estimate chemistry, based on averaging the individual data points, was found to be 0.21 weight percent for copper and 0.59 weight percent for nickel. Nevertheless, the inspectors confirmed that the copper and nickel values for weld wire heat 72102 reported by licensees and contained in the RVID are greater than or equal to the average from all traceable data. Therefore, there were no further concerns with this heat.

b.9 Heat T29744 The inspectors identified additional data for weld wire heat T29744 during the inspection.

However, the additionalinformation on T29744 in FTI's letters of June 6 and July 10,1997, indicated that the best-estimate mean copper and nickel values did not change significantly with the additional data. Also, FTl added weld identification 63W to the existing data for heat 299L44, but the inspectors determined that the best-estimate mean copper and nickel values remained unchanged with the addition of the data from weld identification 63W. The inspectors confirmed that the copper and nickel values reported by licensees and contained in the RVID for T29744 were greater than or equal to the average from all traceable data. Therefore, there were no concems with this weld wire heat.

b.10 Heat 1P0661 The inspectors identified additional data for weld wire heat 1P0661 during the inspection. The inspectors found that the best-estimate chemistry was not conservative when comparing it to the previous data for these weld wire heats. The inspectors reviewed the new information in FTI's letters dated June 6 and July 10,1997, and concluded that the best-estimate mean copper and nickel values did not change significantly with the additional data for weld wire heat 1P0661. The inspector's review of the data for weld wire heat 1P0661 indicated that the copper I and nickel values reported by licensees and contained in the RVID are greater than or equal to l the average from all traceable data. Therefore, there were no further concerns with this heat.

12

b.11 Heats 406L44, 8T1554, 61782,1 P0962, 8T3914 1 P0815 FTl provided a comprehensive listing of chemistry data for these weld wire heats. The inspector's review confirmed that the copper and nickel values reported by licensees and contained in the RVID for these heats were greater than or equal to the average from all traceable data. Therefore, the inspectors had no concems with these weld wire heats.

TABLE 3: CHEMICAL. COMPOSITION DATA for LINDE 80 WELDS FABRICATED With COPPER-COATED WELD WIRES Heat Chemical Chemical FTl Best-Estimate of FTl Estimate Number Compos?. ion Composition from Chemical Composi- of Chemical Reported by BAW-1500 and tion from Average of Composition Licensees and BAW-2121P All Data from Coll-Compiled in Weighted RVID Average

%Cu %Ni %Cu %Ni %Cu %Ni %Cu 299L44 0.35 0.68 0.35 0.68 0.34 0.68 0.34 72105 0.35 0.59 0.35 0.59 0.32 0.57 0.33 406L44 0.31 0.59 0.31 0.59 0.29 0.58 0.26 821T44 0.24 0.63 0.24 0.63 0 24 0.63 0.25 61782 0.25 0.54 0.25 0.54 0.24 0.54 0.24 71249 0.26 0.60 0.26 0.61 0.24 0.61 0.22 72442 0.24 0.60 0.24 0.60 0.24 0.60 0.25 72445 0.21 0.59 0.21 0.59 0.22 0.58 0.22 8T1554 0.18 0.63 0.18 0.63 0.16 0.61 0.16 72102 0.25 0.63 0.23 0.63 0.21 0.59 0.21 T29744 0.29 0.68 0.29 0.68 0.27 0.68 0.27 8T3914 0.18 0.64 0.18 0.64 0.18 0.64 0.18 8T1762 0.20 0.55 0.20 0.55 0.19 0.55 0.19 __

1P0815 0.17 0.52 0.17 0.52 0.17 0.52 0.17 1P0962 0.21 0.64 0.21 0.64 0.21 0.64 0.21 1P0661 0.19 0.63 0.17 0.64 0.17

  • No data were reported to the NRC for this heat.

13

c Detailed Observations and Findinasfor Non-Cocoer-Coated Weld Wire Heats of Interest Table 4 compares the amount of copper and nickelin Linde 80 welds, fabricated with non-copper coated weld wire, that were reported by the licensees and documented by FTI's evaluation of all traceable data. The chemical composition database for non-copper coated weld wires he been distributed by FTl to the licensees with RPV welds fabricated using these heats of we4 uire. Licensees are expected to address the effect of the data on their pressure-versus tempurature operating limit curves and pressurized thermal shcck (PTS) assessments in accordance with applicable regulations.

TABLE 4: CHEMICAL COMPOSITION DATA for LINDE 80 WELDS FABRICATED with NON-COPPER-COATED WELD WIRES Heat s emical FTl Best Estimate of FTl Estimate of Number Composition Chemical Composi- Chemical Com-Reported by tion from Average of position from Licensees and All Data Coil-Weighted Compiled in RVID Average

%Cu %Ni %Cu %Ni %Cu 442002 0.030 0.460 0.029 0.680 0.059 0.050 0.620 442011 0.030 0.630 0.033 0.688 0.032 0.030 0.650 '

H4498 0.030 0.500 0.042 0 460 0.042 31401 0.180 0.540 0.193 0.576 0.193 0.190 0.560 0.230 0.570 1084-18 0.040 0.600 0.038 0.600 0.038 b.12 Heat 442002 The inspectors noted that FTI's best-estimate copper value for low-copper weld wire heat 442002 (0.029 wt-%) was less than that reported by the licensee (0 030 wt %), but the best-estimate nickel value (0.680 wt-%) was significantly greater than the 0 460 wt-% reported by the licensee. The reported copper values for heat 442002 were not considered suspect; however, best-estimate chemistry values were obtained for all of the low copper welds for completeness. To assess this information, the inspectors computed the resultant chemistry factors using the RVID program. The chemistry factor associated with the licensee-reported values was 41.0. However, the chemistry factor associated with the best-estimate values was less than that reported by the licensee, i.e.,39.6 because copper is much more heavily weighted in the calculation. Therefore, there were no further concerns with this heat.

14

.d

b.13 Heat H4498 The inspectors identified some discrepancies in the reported copper values for low copper weld wire heat H4498; and requested additionalinformation. FTl provided the NRC with a summary of the best estimate chemistry for the low-copper welds with its June 19,1997, letter. The copper value for H4498 was slightly higher than previously reported values, but the welds still reportedly contained very low copper. The reported copper values for heat H4498 were not identified as suspect. The inspectors determined the chemistry factor for H4498 associated with the data reported by licensees using the RVID program (41.0) and compared it to the chemistry factor associated with FTI's best estimate copper and nickel values # rived from the average of all data (56.7). However. FTi stated that the affected weld for this heat was in a lower fluence location in the Brakiwood vessel, and was not the controlling (limiting) material / weld for that vessel. Therefore, the apparently significant change in the chemistry factor would have no impact on the operating limits for the Braidwood vessel. There were no further concerns with this heat.

b.14 Heat 442011 The inspectors noted that FTI's best-estimate copper value for heat 442011 (0.033 wt %) was more than those reported by the licensee (0.030 wt %), and the best-estimate nickel value at 0.688 wt % was also greater than the 0.630 and 0.650 wt-% values reported by the licensee. '

The data for heat 442011 were not considered suspect. To assess this information, the inspectors computed the resultant chemistry factors using the RVID program for comparison.

The chemistry factor computed for the licensee _-reported values was 41.0 and the chemistry factor computed for the best-estimate values was 44.9. Although the copper valuer. are very low compared to those from copper-coated weld wire, they would be expected to have a -

calculable effect on operating limits because, according to FTI, heat 442011 was used in the center circular (beltline) weld, the controlling material in the reactor vessels in Braidwood i

Station. FTl confirmed that the affected licensee, Commonwealth Edison Company, was promptly informed of the information and had reported it was assessing the impact an operating limits as required by regulations. The licensee will be expected to make the appropriate submittals to the NRC associated with any resultant changes, b.15 Heat 1084-18 The inspectors identified some discrepancies in the reported copper values for the low copper weld wire heat 1084-18 and requested additional information, in a subsequent submittal, FTl provided a summary of the best-estimate chemistry for the low-copper welds. The inspectors noted that only one data point was identified by FTl for low-copper weld wire heat 1084-18.

The best-estimate mean values for copper and nickel for this weld wire heat were recorded as 0.038 and 0.600 wt-%, respectively. However, the copper and nickel values reported by licensees and contained in the RVID were greater than or equal to the average from all traceable data. Therefore, there were no concerns with this heat.

15

4 b.16 Heat 31401 1 The inspectors noted that the copper values repotted by licensees and listed in the RVID for low-copper weld wire heat 31401 were unusually high for an ostensibly low-copper heat; although still not as high as those typical of copper-coated weld wire. The nickel values were not unusual. The inspectors requested additionalinformation which was provided in a subsequent FTl submittal. FTI's best estimate values for both copper and nickel were consistent with the values in the RVID, falling in the middle of the range. Therefore, there were no further concerns with this heat,

c. Conclusions The inspectors concluded that the data compiled by the B&WOG in topical reports BAW-1500 and BAW-2121P were not complete and not fully traceable. However, the inspectors were provided with sufficient information during and following the inspection to conclude that (1) adequate traceability was eventually established for most of the weld chemistry data, (2) consistency was verified between B&Ws fabrication records and topical reports, and (3) most of the information in the RVID (at least chemistry factors) remained conservative even when all available data were included in the best-estimate chemistry determination. The exceptions have been appropriately dispositioned as discussed in the individual weld wire heat sections above.

It should also be noted that FTl informed the B&WOG Reactor Vessel Working Group, Rochester Gas and Electric, and Tennessee Valley Authority, by letter INS-97-2526, dated June 30,1997, of the results of their efforts prompted by the NRC inspection. This letter included summary tables of the chemical composition assessments for the Linde 80 weld metals, as reported to the NRC. The chemical composition data were presented using three different assessments: (1) currently reported best-estimate values, (2) best-estimate values based on total population of all traceable data, and (3) coil-weighted average for weld copper content. FTI's review of the data assessments indicated that the weld compositions determined prior to the inspection were consentative for most of the weld wire heats. For the copper-coated wire heats, where the estimates were non-conservative, FTl determined the effect to be inconsequential (i.e., greater by only 0.01 weight percent). The inspector's review of these data confirmed that the copper and nickel values for copper-coated weld wire heats reported by licensees and contained in the RVID are greater than or equal to the average from all traceable data, except for we!d wire heat numbers 72445 and 71249. However, since the chemistry factor values (used by the licensees h developing operating limits), based on the average of all l data for these weld wire heats are less than or equal to the chemistry factor values in the RVID, the chemistry data for copper-coated weld wire heats in the RVID are acceptable.

For low-copper heats, FTI's best-estimate values for copper or nickel or their associated chemistry factors were either conservati';e with respect to previous reported values or were not significant because of not being in the controlling welds with one exception. For heat 442011, although it contained very low copper, the coppar value and resultant chemistry factor was greater than that previously reported and was in the controlling material. FTl confirmed that the 16

affected licensee, Commonwealth Edison Company, was promptly informed of the information and had reported that it was assessing the impact on operating limits as required by regulations. The licensee will be expected to make the appropriate submittals to the NRC associated with any resultant changes.

The staff assessment evaluated chemistry data, it did not evaluate the impact of surveillance data on embrittlement. Licensees that use survaillance data to determine the amount of embrittlement will need to assess the FTl chemistry data. The NRC will notify the affected licensees.

4 ENTRANCE and EXIT MEETINGS At the entrance meeting on May 19,1997, the NRC inspectors discussed the scope of the inspection, outlined the areas to be inspected, and established interfaces with cognizant FTl management and staff. At the exit meeting on May 21,1997, the inspectors discussed their findings and observations and obtained a commitment from FTl to submit the remainder of the cF nistry data and other information requested by the inspectors that FTl was unable to pwvide during the onsite portion of the inspection.

5 PARTIAL LlST OF PERSONNEL CONTACTED Ell K.E Moore, Manager, Materials and Structural Analysis Unit L.B. Gross, Advisory Engineer M.J. DeVan, Metallurgical Engineer R.J. Schomaker, Project Manager Richard Rawlings, Quality Assurance Audits Manager Stephen Fyfitch, Senior Supervisor, Materials Group Jim Taylor, Licensing Manager NBC Jack Strosnider, Chief, Materials and Chemical Engineering Branch (EMCB)

Keith Wichman, Chief, Materials Integrity Section, EMCB ITEMS OPENED, CLOSED, AND DISCUSSED None 17

!

Retrieved from "https://kanterella.com/w/index.php?title=ML20199E298&oldid=3336385"