ML20205R327

Insp Rept 99900003/99-201 on 990322-25.Violations Noted. Major Areas Inspected:Activities Related to Manufacture & Testing of Nuclear Fuel & Related Components

ML20205R327
Person / Time
Issue date:	04/21/1999
From:	NRC (Affiliation Not Assigned)
To:
Shared Package
ML20205R304	List: ML20205R299 ML20205R317 ML20205R327
References
REF-QA-99900003 NUDOCS 9904220292
Download: ML20205R327 (9)

Text

.

U.S. NUCLEAR REGULATORY COMMISSION OFFICE OF NUCLEAR REACTOR REGULATION Report No.: 99900003/1999201 Organization: GE Nuclear Energy P.O. Box 780 Wilmington, NC

Contact:

Caroline Roda Manager, GE-NE Quality Nuclear Activity: Nuclear fuel assemblies and related components for BWRs.

Dates: March 22-25,1999 Inspection Team: Robert L. Pettis, Jr., IQMB/ DISP Dr. Shih-Liang Wu, SRXB/DSSA Approved by: Richard P. Correia, Chief Reliability and Maintenance Section Quality Assurance, Vendor inspection, Maintenance and Allegations Branch Division of Inspection Program Management Office of Nuclear Reactor Regulation Enclosure 2 9904220292 990421 PDR GA999 EMV0DE 99900003 PDR I

1

1 INSPECTION

SUMMARY

From March 22-25,1999, representatives of the U.S. Nuclear Regulatory Commission (NRC) conducted a performance-based inspection of the activities at the Wilmington, North Carolina, facility of GE Nuclear Energy (GE-NE). In conducting this inspection, the team emphasized technically directed observations and evaluations of GE-NE activities related to the manufacture and testing of nuclear fuel and related components.

As the technical bases for the inspection, the team relied upon the following:

Part 21,

• Notification of Failure to Comply or Existence of a Defect," as defined in Title 10 of the Code of FederalRegulations (10 CFR) 10 CFR Part 50, Appendix B, " Quality Assurance Criteria for Nuclear Power Plants and Fuel Reprocessing Plants" GE-NE Quality Assurance Program Description (NEBO-11209).

1.1 Violations Violation 99900003/1999201-01 was identified during this inspection and is discussed in Section 3.2 of this report.

1.2 Nonconformances No nonconformances were identified during this inspection.

1.3 UnrescIved items Additional information is required to determine whether the issue discussed in Section 3.4 of this report is an acceptable item, a nonconformance, or a violation. As a 1 result, unresolved item 99900003/1999201-02 was identified during this inspection. {

2 STATUS OF PREVIOUS INSPECTION FINDINGS No previous findings were identified during the last inspection (99900003/95-01).

3 INSPECTION FINDINGS AND OTHER COMMENTS 3.1 Backoround The NRC inspection team reviewed the activities performed by laboratory analysts in the ,

Chemet Laboratory. The Chemet Laboratory is an integral part of the in-process '

release of raw materials, process control, and final release of finished products and consisted of a metallurgical and wet laboratory. During the inspection, the team primarily focused its attention on the activities performed in the wet chemical laboratory. I i

I l

l 2

I The wet chemical laboratory performs a wide variety of chemical and physical analyses on uranium dioxide powder and pellets, and zirconium fuel cladding. The LECO hydrogen analyzer is used to measure the hydrogen release from uranium dioxide fuel pellets and zirconium fuel cladding while the Centorr furnace is used to determine fuel pellet densification.

3.2 LECO Hydroaen Analyzer

a. Inspection Scoce The inspection team reviewed the implementation and operation of the LECO hydrogen analyzer which is used to determine the hydrogen content of zirconium alloys and uranium dioxide products for compliance with GE-NE and customer specification requirements.

b. Observations and Findinas The inspection team reviewed Quality Notice (QN) QN F-Q-2006, " Qualification of LECO RH404 Hydrogen Analyzer," Revison 1, dated July 30,1993, which documents the basis for using this equipment for hydrogen analysis. A QN is a multi-purpose quality planning document which is primarily used for qualification, evaluation, and special or project planning purposes and is part of GE-NE's quality assura7ce program. The  !

analyzer is used to perform hydrogen tests for both zirconium fuel cladding and ceramic fuel pellets. The test results are then compared to GE-NE and customer specification requirements.

During the review of the LECO hydrogen analyzer, the inspection team requested a computer search of recent calibration constants (known as a 706 report) which are maintained in the Laboratory Material Control System (LMCS) for LECO analyzer No.

174, and covered the period April 23,1996 through March 16,1999. The team's review identified a potential deviation in which a laboratory analyst failed to perform over 20 requi'ed weekly calibrations over the three year period. This was evident since the ((

analyst used the previous week's calibration constant instead of performing the required weekly calibration which uses several carbon steel samples of known hydrogen to calculate the calibration constant which is then input into LMCS. The LECO cannot be operated unless a calibration constant is input weekly i,,tu LMCS. The team also identified an example where the analyst used the same constant for a two and three week consecutive period. A reading of 1.08810 was used for the two week period of February 13 and February 20,1997, and the three week period of March 7, March 14 and March 21,1997. All of these entries into LMCS used the constant of 1.08810 which was calculated by a different analyst for the week of February 6,1997.

1 Historically the calibration constant calculated by LECO ranges between 0.90000 to 1 1.60000 and is input by the analyst directly into LMCS for engineering analysis and quality purposes. It is nearly impossible for LECO to calculate the exact same six digit constant for a two or three consecutive week period. In addition, the team reviewed several " Pellet Hydrogen" worksheets (CL-20 R5, issued January 18,1995) which are used by the analysts to record information obtained during the calibiation and verification process. A review of worksheets for the same analyr t for July 23,1998, i

3  !

l l

o

August 20,1998, and October 19,1998, revealed no information filled-in at the top of the page under " Calibration," which confirmed the team's finding that no calibration was performed for these weeks. GE-NE Chemical, Metallurgical and Spectral (CM&S)

CM&S Analytical Method 1.2.8.3, Revision 11, " Determination of Hydrogen in Zirconium, its Alloys, Uranium Dioxide Sintered Pellets, and Pellets by inert Gas Fusion," dated May 15,1995, requires in Section 13.5, " Calibration and Verification" that LECO be calibrated and verified at a minimum frequency of once per week.

A similar issue regardirig a different analyst was raised internally to GE-NE by an ernployee in late 1998. GE-NE Integrity Case 98-014 reviewed an allegation that an analyst in the wet laboratory, performing a hydrogen test for a zirconium sample using the LECO analyzer, cbtained the same calibration constant as the one calculated for the previous week. GE-NE's review concluded that no employee integrity problem existed since only one calibration W been performed by the analyst in the six month period preceding the data of the concern. However, neither the allegation or the integrity review triggered a review for applicability under 10 CFR 50, Appendix B (e.g.,

CM&S 1.0.0.0, Revision 2, " Quality Assurance Program for Analytical and Testing Laboratories," Section 12, " Deficiencies and Corrective Actions") and 10 CFR Part 21(Part 2;). Such a review is required to evaluate the possibility that other laboratory analysts may have bypassed the cali! cation process which may have resulted in the shipment of nonconforming material to operating nuclear plants.

At the ennelusion of the inspection on March 25,1999, GE-NE initiated Potential Safety Concem 9907 and Corrective Action Request 2170 to evaluate the condition under the reportability requirements of Part 21. Preliminary evaluation results performed by GE-NE identified several fuel pellets which exceeded the GE-NE limit for hydrogen of 1 part per million. The team identified GE-NE's failure to identify and evaluate a potential deviation as violation 99900003/1999201-01.

c. Conclusions The GE-NE Integrity Review only went back six months prior to the date of the alleged calibration incident and did not extend beyond the timeframe or the analyst cited in the allegation. The NRC review of the 706 report, downloaded directly from LMCS, clearly identified a pattern of calibration constants copied from the previous week's calibration activity. The NRC team noted that the analyst responsible for this issue was not the same individual named as the subject of the integrity allegation.

3.3 Centorr Fumace

a. Inspection Scope The inspection team reviewed the implementation and operation of the Centorr densification furnace. The furnace subjects production fuel pellets to a 24-hour thermal simulation of the reactor core conditions to determine densification. During the inspection, the Centorr furnace was not operable.

4

f 1

b. Observations and Findinos The Centorr furnace, located in the wet laboratory, is designed to perform fuel pellet densification tests. During routine fuel production activities, samples of fuel pellets from various production lots are selected and sent to the laboratory for testing. The test results are then compared to GE-NE and customer specification requirements. Before the tests are performed, fuel pellets are measured for initial length, diameter and weight ming a densitometer. A group of control pellets, selected from archive inventories with

, mown densification, are also loaded and mixed with the sample pellets in the rack. The test conditions are established by evacuating the furnace, backfilling with hydrogen, and heating to the appropriate temperature for 24-hours to simulate reactor core thermal conditions. After the test, the same set of data is measured again using the densitometer for comparison with the initial data set to establish the amount of densification. The final results are then stored in LMCS for quality control and engineering analysis.

During operation of the Centorr furnace, an accurate temperature profile is critical to the test results. The temperature profile is established using two (upper and lower) thermocouples and a few control pellets with known densification history. The temperature profile has a very tight tolerance to ensure that the pre-set test conditions are achieved and the thermocouples are required to be calibrated by a GE-NE approved l vendor within one year after their initial test application. Although the thermocouples are the primary means of verifying satisfactory fuel pellet behavior, some control pellets are also used as a backup means of verification during the densification tests. GE-NE QN F-Q 1208, "Densification Furnace Set Point and Temperature Profile," Revision 29, establishes documentation of the temperature profile. Revisions to the QN to document periodic recalibration of the furnace and new set points established for the control thermocouple is required in " Chemet Laboratories Calibration and Operation Instruction,

" Thermal Simulation Testing of Fuel Pellets - Furnace Operation" (No: COI #234.00, Revision 0,

Reference:

CM&S 1.2.20.3, Revision 8, dated October 8,1996), Section 7.0, l

" Furnace Temperature Calibration."

During the summer of 1997, the Centorr furnace was under extensive repair to replace cracked thermal shields and heating elements. The repair, including the thermocouple temperature profile which was needed to re-establish the furnace operating conditions, was documented in Revision 23 of QN F-Q 1208. Revision 23 of the QN expired on June 6,1997 and was administratively extended three months without any documentation to support the basis for the extension. During this period control pellets were included in each furnace run to assure that the densification control values )

remained within LMCS tolerance limits. The two existing thermocouples at the time j were used to measure the furnace temperature on July 15,1997. However, one was l l due for recalibration on May 31,1997, and another on July 15,1997. GE-NE )

documented the use of an overdue thermocouple in a Measurement and Test Equipment Evaluation Report, dated July 31,1997. The overdue thermocouples were subsequently calibrated by JMS Southeast, Inc. with no shift observed in the calibration

! parameters. However, to prevent this situation from recurring, GE NE currently maintains 5 to 6 spare thermocouples in stock. Based on the verified characteristics of the recalibrated thermocouples and the control pellets used for each run, GE-NE j 5

i i

concluded that the temperature profile was accurate and the densification runs for batch 3197 to 4797 were acceptable. GE-NE documented the findings in " Notice of  ;

Measurement Deviation W-97-02," dated September 10,1997. The inspection team )

concluded that there were no anomalies in the fuel densification results as a result of the fumace rebuilding. However, rather than using the required QN to document the furnace temperature profile, the information was entered in the materials and purchasing system (MPAC) which is not authorized for this purpose. The MPAC system )

is a computer database which contains pertinent information related to the furnace operation but is not considered a quality document like a QN.

After establishment of the new temperature profile, the operation of the furnace was obsented below the lower alarm set point during a run on July 17,1997, and documented in a report titled, " Verification Standard Report Out-of-Alarm Limits," which refers to the temperature profile which had just been run with the control pellet results.

GE-NE established the following 5 limits according to statistical analysis for densification runs that use control pellets: (1) upper control limit, (2) upper alarm limit, (3) nominal value, (4) lower alarm limit, and (5) lower control limit. An out-of-alarm (OOA) means the amount of densification from a control pellet falling between upper (lower) alarm limit and upper (lower) control it.mit, and an out-of-control (OOC) means the amount of j densification from a control pellet either higher than the upper control limit or lower than the lower control limit. A control pellet densification resulting in a number between the upper and lower alarm limit is considered acceptable to guarantee an accurate temperature profile.

The NRC inspection team selected for review a July 16,1997, report for batch run 3497 l

which reported the densification results of control pellets in 5 levels of the sample rack.

The top level showed that the data was located between the lower alarm and lower control limits, an OOA, and the bottom level showed data higher than the upper control limit, an OOC. Only the data from the center three levels was located in the acceptable limits, although somewhat biased toward the lower alarm limit. GE-NE attributed the bias to the use of a different population of archive pellets from the usually selected population that resulted in additional variability to skew the results. In order to ensure an accurate temperature profile, GE-NE selected only the center three levels to hold the sample pellets and included at least three control pellets for each run. In doing so, GE- I NE ensured that the test results were backed up by a secondary source (the control pellets).

The team also reviewed a September 2,1997, report for batch run 4597 which reported the results of densification data for 41 production sample pellets and 3 control pellets that were within GE-NE set limits for production pellets (e.g., no data extended beyond the OOA or OOC). The team also noted that the uncertainty band of densification data .

from the LMOS history was drastically reduced after the introduction of the dry j conversion process for uranium dioxide powder, which results in more stable fuel pellets l in terms of densification behavior.

In late 1997, several issues were raised to the GE-NE Power System Ombudsperson regarding irregularities involving operation of the Centorr furnace. The ombudsperson, l with support from the GE-NE staff, is responsible for determining if an alle gation of j misconduct exists and is responsible to form an investigative team to review the issues i 6

i

r

, raised. A formalinternalinvestigation of the alleged concerns, similar to the integrity review performed for the (ECO hydrogen analyzer, was conducted which included a concern regarding thermocouples used beyond their calibration expiration cate. GE-NE concluded that the undocumented extension of the QN and the improper use of utilizing the MPAC system to document the temperature profile was a procedural violation.

Corrective actions included Revision 25 to the ON, dated August 29,1997; replacement of MPAC information with a properly updated revision; and discussions with laboratory staff emphasizing procedural adherence and improved communications.

c. Conclusions The NRC inspection team reviewed the investigative raport and noted that the uncertainty band of densification data from the LMCS history was drastically reduced after GE-NE introduced the dry conversion process for uranium dioxide powder which resulted in more stable fuel pellets in terms of densification behavior. Based on the team's review of the wet laboratory procedures and related documents, the team concluded that the Centorr furnace is acceptable for performing densification tests.

3.4 Review of GE-NE Performed Sucolier Audit

a. Scooe During the NRC review of the rebuilding of the Centorr furnace in mid-1997, it was identified that one of the thermocouples had been sent out to JMS Southeast, Inc. (JMS) for calibration. JMS is listed on GE-NE's Approved Supplier List as a safety-related supplier. The team requested documentation to support the procurement for the calibration service and related documentation to support approval of the vendor in accordance with 10 CFR 50, Appendix B and GE-NE supplier quality requirements.

b. Observations and Findinas The inspection team reviewed GE-NE purchase order (PO) No. 33497066632, dated  :

September 4,1997, to JMS for full calibration and recertification of two Type "C"  !

thermocouples (GE-NE identification Nos. ZO10003 and ZOO 98534) used to determine the thermal profile of the Centorr fumace. The PO required compliance to Part 21 and required the vendor to fumish Certificates of Traceability to the National Institute of Standards and Technology at 1700 degrees centigrade.

GE-NE last audited JMS on March 26,1997, and identified three Corrective Action Requests (CARS). This performance-based audit included a review of the QA Manual j and procedures for compliance to 10 CFR 50, Appendix B, and Part 21. CAR JMS-2, dated March 26,1997, was of particular importance to the NRC inspection team since it identified that JMS's QA program did not procedurally address Part 21or the reporting of defects to customers. Although the lead auditor classified these issues as recommendations, they were documented as CARS for response and tracking purposes within the GE-NE system. GE Wilmington Practices and Procedures No. 60-21, issued November 3,1997, states, in part, in Exhibit 1,

• Procedural Flow," that acknowledgment of the findings are to be returned by the supplier within 21 days of the initiation date.

7

9 However, a review of an April 29,1998, GE-NE Supplier Quality Assurance Program Assessment Form, used to assess annual performance of the supplier, noted that GE-NE is still working with JMS to close the three CARS identified over a year ago. In addition, a review of the JMS Supplier Profile, dated March 25,1999, indicated that an audit needed to be performed prior to May 1999 (almost one year earlier than the scheduled triennial audit due date of March 26,2000). A review of the CARS identified that the lower portion of each form was blank indicating that neither JMS or GE-NE had documented any actions to resolve these issues. A March 31,1997, note to file, prepared by the lead auditor, recommended that JMS be retained as a safety-related source of calibration services and that GE-NE needs to establish controls necessary to meet the requirements of Part 21. This statement and the issuance of CAR-2 appears to indicate that GE-NE must assume the reporting responsibility of Part 21 since JMS is not procedurally equipped to do so.

Although time was limited during this part of the inspection, GE-NE could not produce any dommentation to support satisfactory resolution of the CARS, which have now been unw 3 Ad for over two years. Since the issue could not be resolved during the ins,wi n1 the NRC needs additionalinformation to determine if the issue is an accep4Me item, a nonconformance, or a violation. GE-NE is requested to submit a written response to this item as discussed in the cover letter to this report. As a result, unresolved item 99900003/1999201-02 was identified during this part of the inspection.

c. Conclusions GE-NE could not produce documentation during the inspection to support the resolution of three open CARS identified during GE-NE's March 26,1997, audit of JMS Southeast, Inc. As a result, additional information is required for the inspection team to reach its conclusions. Additionally, should GE-NE accept the Part 21 reporting responsibility for items supplied by JMS, a re-review of JMS's QA program should be performed to ensure that adequate controls exist to identify and resolve nonconforming conditions which may have an impact on calibration services supplied to GE-NE.

4 ENTRANCE AND EXIT MEETINGS During the entrance meeting on March 22,1999, the inspection team met with members of GE-NE management and staff and discussed the scope of the inspection. The team also reviewed its responsibilities for handling proprietary information as well as those of GE-NE. In addition, the team established contact persons within the management and staff of the applicable GE-NE organizations and discussed the results of the inspection with management and staff on March 25,1999.

PARTIAL LIST OF PERSONS CONTACTED C. Monetta Manager, EHS S. Fuller Manager, Fuel & Chemet Lab Quality W. Baker Nuclear Quality Assurance J. Ball Team Leader / Lead Chemist 8

p.

R. Bianchi Senior Engineer J. McLeod Quality Engineer P. Schenck Administrative Specialist H. Fields Laboratory Analyst ITEMS OPENED, CLOSED, AND DISCUSSED ltem Number Iygg Descriotion Opened 99900003/1999201-01 NOV Failure to identify and evaluate a potential deviation related to the lack of calibration of the LECO hydrogen analyzer.

99900003/1999201-02 URI No documentation provided to support resolution of three open CARS identified in GE-NE audit of JMS.

1 3

i l

I i

-